1191 files changed, 42201 insertions, 28691 deletions

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 94f7a711bf8d..1461652921be 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -15,6 +15,9 @@ menuconfig DRM
 	select I2C_ALGOBIT
 	select DMA_SHARED_BUFFER
 	select SYNC_FILE
+# gallium uses SYS_kcmp for os_same_file_description() to de-duplicate
+# device and dmabuf fd. Let's make sure that is available for our userspace.
+	select KCMP
 	help
 	  Kernel-level support for the Direct Rendering Infrastructure (DRI)
 	  introduced in XFree86 4.0. If you say Y here, you need to select
@@ -225,6 +228,7 @@ source "drivers/gpu/drm/arm/Kconfig"
 config DRM_RADEON
 	tristate "ATI Radeon"
 	depends on DRM && PCI && MMU
+	depends on AGP || !AGP
 	select FW_LOADER
         select DRM_KMS_HELPER
         select DRM_TTM
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 6bf6cfaea3f1..13ebb1f71e49 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -56,7 +56,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
 	amdgpu_gmc.o amdgpu_mmhub.o amdgpu_xgmi.o amdgpu_csa.o amdgpu_ras.o amdgpu_vm_cpu.o \
 	amdgpu_vm_sdma.o amdgpu_discovery.o amdgpu_ras_eeprom.o amdgpu_nbio.o \
 	amdgpu_umc.o smu_v11_0_i2c.o amdgpu_fru_eeprom.o amdgpu_rap.o \
-	amdgpu_fw_attestation.o
+	amdgpu_fw_attestation.o amdgpu_securedisplay.o
 
 amdgpu-$(CONFIG_PERF_EVENTS) += amdgpu_pmu.o
 
@@ -71,7 +71,7 @@ amdgpu-y += \
 	vi.o mxgpu_vi.o nbio_v6_1.o soc15.o emu_soc.o mxgpu_ai.o nbio_v7_0.o vega10_reg_init.o \
 	vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o navi10_reg_init.o navi14_reg_init.o \
 	arct_reg_init.o navi12_reg_init.o mxgpu_nv.o sienna_cichlid_reg_init.o vangogh_reg_init.o \
-	nbio_v7_2.o dimgrey_cavefish_reg_init.o
+	nbio_v7_2.o dimgrey_cavefish_reg_init.o hdp_v4_0.o hdp_v5_0.o
 
 # add DF block
 amdgpu-y += \
@@ -97,6 +97,7 @@ amdgpu-y += \
 	tonga_ih.o \
 	cz_ih.o \
 	vega10_ih.o \
+	vega20_ih.o \
 	navi10_ih.o
 
 # add PSP block
@@ -170,7 +171,8 @@ amdgpu-y += \
 # add SMUIO block
 amdgpu-y += \
 	smuio_v9_0.o \
-	smuio_v11_0.o
+	smuio_v11_0.o \
+	smuio_v11_0_6.o
 
 # add amdkfd interfaces
 amdgpu-y += amdgpu_amdkfd.o
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index ab4ac3b2651e..1af2fa1591fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -88,6 +88,7 @@
 #include "amdgpu_gfx.h"
 #include "amdgpu_sdma.h"
 #include "amdgpu_nbio.h"
+#include "amdgpu_hdp.h"
 #include "amdgpu_dm.h"
 #include "amdgpu_virt.h"
 #include "amdgpu_csa.h"
@@ -106,6 +107,7 @@
 #include "amdgpu_gfxhub.h"
 #include "amdgpu_df.h"
 #include "amdgpu_smuio.h"
+#include "amdgpu_hdp.h"
 
 #define MAX_GPU_INSTANCE		16
 
@@ -178,6 +180,7 @@ extern uint amdgpu_smu_memory_pool_size;
 extern uint amdgpu_dc_feature_mask;
 extern uint amdgpu_dc_debug_mask;
 extern uint amdgpu_dm_abm_level;
+extern int amdgpu_backlight;
 extern struct amdgpu_mgpu_info mgpu_info;
 extern int amdgpu_ras_enable;
 extern uint amdgpu_ras_mask;
@@ -285,7 +288,7 @@ enum amdgpu_kiq_irq {
 
 #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
 #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
-#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */
+#define MAX_KIQ_REG_TRY 1000
 
 int amdgpu_device_ip_set_clockgating_state(void *dev,
 					   enum amd_ip_block_type block_type,
@@ -577,7 +580,8 @@ enum amd_reset_method {
 	AMD_RESET_METHOD_MODE0,
 	AMD_RESET_METHOD_MODE1,
 	AMD_RESET_METHOD_MODE2,
-	AMD_RESET_METHOD_BACO
+	AMD_RESET_METHOD_BACO,
+	AMD_RESET_METHOD_PCI,
 };
 
 /*
@@ -607,7 +611,6 @@ struct amdgpu_asic_funcs {
 	/* invalidate hdp read cache */
 	void (*invalidate_hdp)(struct amdgpu_device *adev,
 			       struct amdgpu_ring *ring);
-	void (*reset_hdp_ras_error_count)(struct amdgpu_device *adev);
 	/* check if the asic needs a full reset of if soft reset will work */
 	bool (*need_full_reset)(struct amdgpu_device *adev);
 	/* initialize doorbell layout for specific asic*/
@@ -890,6 +893,7 @@ struct amdgpu_device {
 	/* For pre-DCE11. DCE11 and later are in "struct amdgpu_device->dm" */
 	struct work_struct		hotplug_work;
 	struct amdgpu_irq_src		crtc_irq;
+	struct amdgpu_irq_src		vline0_irq;
 	struct amdgpu_irq_src		vupdate_irq;
 	struct amdgpu_irq_src		pageflip_irq;
 	struct amdgpu_irq_src		hpd_irq;
@@ -920,6 +924,9 @@ struct amdgpu_device {
 	/* nbio */
 	struct amdgpu_nbio		nbio;
 
+	/* hdp */
+	struct amdgpu_hdp		hdp;
+
 	/* smuio */
 	struct amdgpu_smuio		smuio;
 
@@ -1002,6 +1009,12 @@ struct amdgpu_device {
 	bool                            in_suspend;
 	bool				in_hibernate;
 
+	/*
+	 * The combination flag in_poweroff_reboot_com used to identify the poweroff
+	 * and reboot opt in the s0i3 system-wide suspend.
+	 */
+	bool 				in_poweroff_reboot_com;
+
 	atomic_t 			in_gpu_reset;
 	enum pp_mp1_state               mp1_state;
 	struct rw_semaphore reset_sem;
@@ -1201,8 +1214,10 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define amdgpu_asic_read_bios_from_rom(adev, b, l) (adev)->asic_funcs->read_bios_from_rom((adev), (b), (l))
 #define amdgpu_asic_read_register(adev, se, sh, offset, v)((adev)->asic_funcs->read_register((adev), (se), (sh), (offset), (v)))
 #define amdgpu_asic_get_config_memsize(adev) (adev)->asic_funcs->get_config_memsize((adev))
-#define amdgpu_asic_flush_hdp(adev, r) (adev)->asic_funcs->flush_hdp((adev), (r))
-#define amdgpu_asic_invalidate_hdp(adev, r) (adev)->asic_funcs->invalidate_hdp((adev), (r))
+#define amdgpu_asic_flush_hdp(adev, r) \
+	((adev)->asic_funcs->flush_hdp ? (adev)->asic_funcs->flush_hdp((adev), (r)) : (adev)->hdp.funcs->flush_hdp((adev), (r)))
+#define amdgpu_asic_invalidate_hdp(adev, r) \
+	((adev)->asic_funcs->invalidate_hdp ? (adev)->asic_funcs->invalidate_hdp((adev), (r)) : (adev)->hdp.funcs->invalidate_hdp((adev), (r)))
 #define amdgpu_asic_need_full_reset(adev) (adev)->asic_funcs->need_full_reset((adev))
 #define amdgpu_asic_init_doorbell_index(adev) (adev)->asic_funcs->init_doorbell_index((adev))
 #define amdgpu_asic_get_pcie_usage(adev, cnt0, cnt1) ((adev)->asic_funcs->get_pcie_usage((adev), (cnt0), (cnt1)))
@@ -1221,6 +1236,7 @@ bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev);
 int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 			      struct amdgpu_job* job);
 void amdgpu_device_pci_config_reset(struct amdgpu_device *adev);
+int amdgpu_device_pci_reset(struct amdgpu_device *adev);
 bool amdgpu_device_need_post(struct amdgpu_device *adev);
 
 void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index 8155c54392c8..2e9b16fb3fcd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -903,10 +903,11 @@ void amdgpu_acpi_fini(struct amdgpu_device *adev)
  */
 bool amdgpu_acpi_is_s0ix_supported(struct amdgpu_device *adev)
 {
+#if defined(CONFIG_AMD_PMC) || defined(CONFIG_AMD_PMC_MODULE)
 	if (acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0) {
 		if (adev->flags & AMD_IS_APU)
 			return true;
 	}
-
+#endif
 	return false;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index db96d69eb45e..c5343a5eecbe 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -47,12 +47,8 @@ int amdgpu_amdkfd_init(void)
 	amdgpu_amdkfd_total_mem_size = si.totalram - si.totalhigh;
 	amdgpu_amdkfd_total_mem_size *= si.mem_unit;
 
-#ifdef CONFIG_HSA_AMD
 	ret = kgd2kfd_init();
 	amdgpu_amdkfd_gpuvm_init_mem_limits();
-#else
-	ret = -ENOENT;
-#endif
 	kfd_initialized = !ret;
 
 	return ret;
@@ -696,86 +692,3 @@ bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd)
 
 	return adev->have_atomics_support;
 }
-
-#ifndef CONFIG_HSA_AMD
-bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm)
-{
-	return false;
-}
-
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
-{
-}
-
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo)
-{
-	return 0;
-}
-
-void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
-					struct amdgpu_vm *vm)
-{
-}
-
-struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f)
-{
-	return NULL;
-}
-
-int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm)
-{
-	return 0;
-}
-
-struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev,
-			      unsigned int asic_type, bool vf)
-{
-	return NULL;
-}
-
-bool kgd2kfd_device_init(struct kfd_dev *kfd,
-			 struct drm_device *ddev,
-			 const struct kgd2kfd_shared_resources *gpu_resources)
-{
-	return false;
-}
-
-void kgd2kfd_device_exit(struct kfd_dev *kfd)
-{
-}
-
-void kgd2kfd_exit(void)
-{
-}
-
-void kgd2kfd_suspend(struct kfd_dev *kfd, bool run_pm)
-{
-}
-
-int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm)
-{
-	return 0;
-}
-
-int kgd2kfd_pre_reset(struct kfd_dev *kfd)
-{
-	return 0;
-}
-
-int kgd2kfd_post_reset(struct kfd_dev *kfd)
-{
-	return 0;
-}
-
-void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry)
-{
-}
-
-void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd)
-{
-}
-
-void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint32_t throttle_bitmask)
-{
-}
-#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index ea391ca7f2f1..a81d9cacf9b8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -94,11 +94,6 @@ enum kgd_engine_type {
 	KGD_ENGINE_MAX
 };
 
-struct amdgpu_amdkfd_fence *amdgpu_amdkfd_fence_create(u64 context,
-						       struct mm_struct *mm);
-bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm);
-struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f);
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo);
 
 struct amdkfd_process_info {
 	/* List head of all VMs that belong to a KFD process */
@@ -132,8 +127,6 @@ void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev,
 void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev);
 void amdgpu_amdkfd_device_init(struct amdgpu_device *adev);
 void amdgpu_amdkfd_device_fini(struct amdgpu_device *adev);
-
-int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm);
 int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
 				uint32_t vmid, uint64_t gpu_addr,
 				uint32_t *ib_cmd, uint32_t ib_len);
@@ -153,6 +146,38 @@ void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd);
 int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev,
 					int queue_bit);
 
+struct amdgpu_amdkfd_fence *amdgpu_amdkfd_fence_create(u64 context,
+								struct mm_struct *mm);
+#if IS_ENABLED(CONFIG_HSA_AMD)
+bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm);
+struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f);
+int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo);
+int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm);
+#else
+static inline
+bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm)
+{
+	return false;
+}
+
+static inline
+struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f)
+{
+	return NULL;
+}
+
+static inline
+int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo)
+{
+	return 0;
+}
+
+static inline
+int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm)
+{
+	return 0;
+}
+#endif
 /* Shared API */
 int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
 				void **mem_obj, uint64_t *gpu_addr,
@@ -215,8 +240,6 @@ int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd,
 					struct file *filp, u32 pasid,
 					void **vm, void **process_info,
 					struct dma_fence **ef);
-void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
-				struct amdgpu_vm *vm);
 void amdgpu_amdkfd_gpuvm_destroy_process_vm(struct kgd_dev *kgd, void *vm);
 void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *vm);
 uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm);
@@ -236,23 +259,43 @@ int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
 		struct kgd_mem *mem, void **kptr, uint64_t *size);
 int amdgpu_amdkfd_gpuvm_restore_process_bos(void *process_info,
 					    struct dma_fence **ef);
-
 int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd,
 					      struct kfd_vm_fault_info *info);
-
 int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
 				      struct dma_buf *dmabuf,
 				      uint64_t va, void *vm,
 				      struct kgd_mem **mem, uint64_t *size,
 				      uint64_t *mmap_offset);
-
-void amdgpu_amdkfd_gpuvm_init_mem_limits(void);
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo);
-
 int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd,
 				struct tile_config *config);
+#if IS_ENABLED(CONFIG_HSA_AMD)
+void amdgpu_amdkfd_gpuvm_init_mem_limits(void);
+void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
+				struct amdgpu_vm *vm);
+void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo);
+#else
+static inline
+void amdgpu_amdkfd_gpuvm_init_mem_limits(void)
+{
+}
 
+static inline
+void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
+					struct amdgpu_vm *vm)
+{
+}
+
+static inline
+void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
+{
+}
+#endif
 /* KGD2KFD callbacks */
+int kgd2kfd_quiesce_mm(struct mm_struct *mm);
+int kgd2kfd_resume_mm(struct mm_struct *mm);
+int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm,
+						struct dma_fence *fence);
+#if IS_ENABLED(CONFIG_HSA_AMD)
 int kgd2kfd_init(void);
 void kgd2kfd_exit(void);
 struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev,
@@ -266,11 +309,68 @@ int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm);
 int kgd2kfd_pre_reset(struct kfd_dev *kfd);
 int kgd2kfd_post_reset(struct kfd_dev *kfd);
 void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry);
-int kgd2kfd_quiesce_mm(struct mm_struct *mm);
-int kgd2kfd_resume_mm(struct mm_struct *mm);
-int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm,
-					       struct dma_fence *fence);
 void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd);
 void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint32_t throttle_bitmask);
+#else
+static inline int kgd2kfd_init(void)
+{
+	return -ENOENT;
+}
 
+static inline void kgd2kfd_exit(void)
+{
+}
+
+static inline
+struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev,
+					unsigned int asic_type, bool vf)
+{
+	return NULL;
+}
+
+static inline
+bool kgd2kfd_device_init(struct kfd_dev *kfd, struct drm_device *ddev,
+				const struct kgd2kfd_shared_resources *gpu_resources)
+{
+	return false;
+}
+
+static inline void kgd2kfd_device_exit(struct kfd_dev *kfd)
+{
+}
+
+static inline void kgd2kfd_suspend(struct kfd_dev *kfd, bool run_pm)
+{
+}
+
+static inline int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm)
+{
+	return 0;
+}
+
+static inline int kgd2kfd_pre_reset(struct kfd_dev *kfd)
+{
+	return 0;
+}
+
+static inline int kgd2kfd_post_reset(struct kfd_dev *kfd)
+{
+	return 0;
+}
+
+static inline
+void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry)
+{
+}
+
+static inline
+void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd)
+{
+}
+
+static inline
+void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint32_t throttle_bitmask)
+{
+}
+#endif
 #endif /* AMDGPU_AMDKFD_H_INCLUDED */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
index 4763bab7a4d0..62aa1a6f64ed 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
@@ -23,7 +23,6 @@
 #include "amdgpu_amdkfd.h"
 #include "gc/gc_10_1_0_offset.h"
 #include "gc/gc_10_1_0_sh_mask.h"
-#include "navi10_enum.h"
 #include "athub/athub_2_0_0_offset.h"
 #include "athub/athub_2_0_0_sh_mask.h"
 #include "oss/osssys_5_0_0_offset.h"
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
index 50016bf9c427..fad3b91f74f5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
@@ -24,7 +24,6 @@
 #include "amdgpu_amdkfd.h"
 #include "gc/gc_10_3_0_offset.h"
 #include "gc/gc_10_3_0_sh_mask.h"
-#include "navi10_enum.h"
 #include "oss/osssys_5_0_0_offset.h"
 #include "oss/osssys_5_0_0_sh_mask.h"
 #include "soc15_common.h"
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index e440af37dde8..9f6b299cbf74 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -26,6 +26,7 @@
 #include <linux/sched/task.h>
 
 #include "amdgpu_object.h"
+#include "amdgpu_gem.h"
 #include "amdgpu_vm.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_dma_buf.h"
@@ -1160,7 +1161,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	struct sg_table *sg = NULL;
 	uint64_t user_addr = 0;
 	struct amdgpu_bo *bo;
-	struct amdgpu_bo_param bp;
+	struct drm_gem_object *gobj;
 	u32 domain, alloc_domain;
 	u64 alloc_flags;
 	int ret;
@@ -1228,19 +1229,14 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s\n",
 			va, size, domain_string(alloc_domain));
 
-	memset(&bp, 0, sizeof(bp));
-	bp.size = size;
-	bp.byte_align = 1;
-	bp.domain = alloc_domain;
-	bp.flags = alloc_flags;
-	bp.type = bo_type;
-	bp.resv = NULL;
-	ret = amdgpu_bo_create(adev, &bp, &bo);
+	ret = amdgpu_gem_object_create(adev, size, 1, alloc_domain, alloc_flags,
+				       bo_type, NULL, &gobj);
 	if (ret) {
 		pr_debug("Failed to create BO on domain %s. ret %d\n",
-				domain_string(alloc_domain), ret);
+			 domain_string(alloc_domain), ret);
 		goto err_bo_create;
 	}
+	bo = gem_to_amdgpu_bo(gobj);
 	if (bo_type == ttm_bo_type_sg) {
 		bo->tbo.sg = sg;
 		bo->tbo.ttm->sg = sg;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 306077884a67..6107ac91db25 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -112,6 +112,7 @@ int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev)
 union igp_info {
 	struct atom_integrated_system_info_v1_11 v11;
 	struct atom_integrated_system_info_v1_12 v12;
+	struct atom_integrated_system_info_v2_1 v21;
 };
 
 union umc_info {
@@ -209,24 +210,42 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 		if (adev->flags & AMD_IS_APU) {
 			igp_info = (union igp_info *)
 				(mode_info->atom_context->bios + data_offset);
-			switch (crev) {
-			case 11:
-				mem_channel_number = igp_info->v11.umachannelnumber;
-				/* channel width is 64 */
-				if (vram_width)
-					*vram_width = mem_channel_number * 64;
-				mem_type = igp_info->v11.memorytype;
-				if (vram_type)
-					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+			switch (frev) {
+			case 1:
+				switch (crev) {
+				case 11:
+				case 12:
+					mem_channel_number = igp_info->v11.umachannelnumber;
+					if (!mem_channel_number)
+						mem_channel_number = 1;
+					/* channel width is 64 */
+					if (vram_width)
+						*vram_width = mem_channel_number * 64;
+					mem_type = igp_info->v11.memorytype;
+					if (vram_type)
+						*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+					break;
+				default:
+					return -EINVAL;
+				}
 				break;
-			case 12:
-				mem_channel_number = igp_info->v12.umachannelnumber;
-				/* channel width is 64 */
-				if (vram_width)
-					*vram_width = mem_channel_number * 64;
-				mem_type = igp_info->v12.memorytype;
-				if (vram_type)
-					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+			case 2:
+				switch (crev) {
+				case 1:
+				case 2:
+					mem_channel_number = igp_info->v21.umachannelnumber;
+					if (!mem_channel_number)
+						mem_channel_number = 1;
+					/* channel width is 64 */
+					if (vram_width)
+						*vram_width = mem_channel_number * 64;
+					mem_type = igp_info->v21.memorytype;
+					if (vram_type)
+						*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+					break;
+				default:
+					return -EINVAL;
+				}
 				break;
 			default:
 				return -EINVAL;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
index 6333cada1e09..cfb1a9a04477 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
@@ -155,7 +155,7 @@ static bool amdgpu_read_bios_from_rom(struct amdgpu_device *adev)
 	u8 header[AMD_VBIOS_SIGNATURE_END+1] = {0};
 	int len;
 
-	if (!adev->asic_funcs->read_bios_from_rom)
+	if (!adev->asic_funcs || !adev->asic_funcs->read_bios_from_rom)
 		return false;
 
 	/* validate VBIOS signature */
@@ -291,7 +291,7 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 			continue;
 
 		status = acpi_get_handle(dhandle, "ATRM", &atrm_handle);
-		if (!ACPI_FAILURE(status)) {
+		if (ACPI_SUCCESS(status)) {
 			found = true;
 			break;
 		}
@@ -304,7 +304,7 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 				continue;
 
 			status = acpi_get_handle(dhandle, "ATRM", &atrm_handle);
-			if (!ACPI_FAILURE(status)) {
+			if (ACPI_SUCCESS(status)) {
 				found = true;
 				break;
 			}
@@ -348,7 +348,8 @@ static bool amdgpu_read_disabled_bios(struct amdgpu_device *adev)
 	if (adev->flags & AMD_IS_APU)
 		return igp_read_bios_from_vram(adev);
 	else
-		return amdgpu_asic_read_disabled_bios(adev);
+		return (!adev->asic_funcs || !adev->asic_funcs->read_disabled_bios) ?
+			false : amdgpu_asic_read_disabled_bios(adev);
 }
 
 #ifdef CONFIG_ACPI
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 594a0108e90f..3e240b952e79 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -98,8 +98,7 @@ static int amdgpu_cs_bo_handles_chunk(struct amdgpu_cs_parser *p,
 	return 0;
 
 error_free:
-	if (info)
-		kvfree(info);
+	kvfree(info);
 
 	return r;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index c34be9f612c8..43059ead733b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -35,6 +35,7 @@
 #include "amdgpu_dm_debugfs.h"
 #include "amdgpu_ras.h"
 #include "amdgpu_rap.h"
+#include "amdgpu_securedisplay.h"
 #include "amdgpu_fw_attestation.h"
 
 /**
@@ -356,7 +357,7 @@ static ssize_t amdgpu_debugfs_regs_pcie_read(struct file *f, char __user *buf,
 	while (size) {
 		uint32_t value;
 
-		value = RREG32_PCIE(*pos >> 2);
+		value = RREG32_PCIE(*pos);
 		r = put_user(value, (uint32_t *)buf);
 		if (r) {
 			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
@@ -423,7 +424,7 @@ static ssize_t amdgpu_debugfs_regs_pcie_write(struct file *f, const char __user
 			return r;
 		}
 
-		WREG32_PCIE(*pos >> 2, value);
+		WREG32_PCIE(*pos, value);
 
 		result += 4;
 		buf += 4;
@@ -1669,6 +1670,8 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 
 	amdgpu_rap_debugfs_init(adev);
 
+	amdgpu_securedisplay_debugfs_init(adev);
+
 	amdgpu_fw_attestation_debugfs_init(adev);
 
 	return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_list,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 168f3adabf57..6447cd6ca5a8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -81,7 +81,6 @@ MODULE_FIRMWARE("amdgpu/navi10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/navi14_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/navi12_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/vangogh_gpu_info.bin");
-MODULE_FIRMWARE("amdgpu/green_sardine_gpu_info.bin");
 
 #define AMDGPU_RESUME_MS		2000
 
@@ -930,6 +929,18 @@ void amdgpu_device_pci_config_reset(struct amdgpu_device *adev)
 	pci_write_config_dword(adev->pdev, 0x7c, AMDGPU_ASIC_RESET_DATA);
 }
 
+/**
+ * amdgpu_device_pci_reset - reset the GPU using generic PCI means
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Resets the GPU using generic pci reset interfaces (FLR, SBR, etc.).
+ */
+int amdgpu_device_pci_reset(struct amdgpu_device *adev)
+{
+	return pci_reset_function(adev->pdev);
+}
+
 /*
  * GPU doorbell aperture helpers function.
  */
@@ -1434,10 +1445,8 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev,
 		amdgpu_device_resume(dev, true);
 
 		dev->switch_power_state = DRM_SWITCH_POWER_ON;
-		drm_kms_helper_poll_enable(dev);
 	} else {
 		pr_info("switched off\n");
-		drm_kms_helper_poll_disable(dev);
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 		amdgpu_device_suspend(dev, true);
 		amdgpu_device_cache_pci_state(pdev);
@@ -2550,11 +2559,11 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 	if (adev->gmc.xgmi.num_physical_nodes > 1)
 		amdgpu_xgmi_remove_device(adev);
 
-	amdgpu_amdkfd_device_fini(adev);
-
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
+	amdgpu_amdkfd_device_fini(adev);
+
 	/* need to disable SMC first */
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if (!adev->ip_blocks[i].status.hw)
@@ -2669,7 +2678,8 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 {
 	int i, r;
 
-	if (!amdgpu_acpi_is_s0ix_supported(adev) || amdgpu_in_reset(adev)) {
+	if (adev->in_poweroff_reboot_com ||
+	    !amdgpu_acpi_is_s0ix_supported(adev) || amdgpu_in_reset(adev)) {
 		amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
 		amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 	}
@@ -3036,7 +3046,7 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
 #endif
 	default:
 		if (amdgpu_dc > 0)
-			DRM_INFO("Display Core has been requested via kernel parameter "
+			DRM_INFO_ONCE("Display Core has been requested via kernel parameter "
 					 "but isn't supported by ASIC, ignoring\n");
 		return false;
 	}
@@ -3119,7 +3129,10 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
 	 */
 	adev->gfx_timeout = msecs_to_jiffies(10000);
 	adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout;
-	if (amdgpu_sriov_vf(adev) || amdgpu_passthrough(adev))
+	if (amdgpu_sriov_vf(adev))
+		adev->compute_timeout = amdgpu_sriov_is_pp_one_vf(adev) ?
+					msecs_to_jiffies(60000) : msecs_to_jiffies(10000);
+	else if (amdgpu_passthrough(adev))
 		adev->compute_timeout =  msecs_to_jiffies(60000);
 	else
 		adev->compute_timeout = MAX_SCHEDULE_TIMEOUT;
@@ -3722,14 +3735,15 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 
 	r = amdgpu_device_ip_suspend_phase1(adev);
 
-	amdgpu_amdkfd_suspend(adev, !fbcon);
+	amdgpu_amdkfd_suspend(adev, adev->in_runpm);
 
 	/* evict vram memory */
 	amdgpu_bo_evict_vram(adev);
 
 	amdgpu_fence_driver_suspend(adev);
 
-	if (!amdgpu_acpi_is_s0ix_supported(adev) || amdgpu_in_reset(adev))
+	if (adev->in_poweroff_reboot_com ||
+	    !amdgpu_acpi_is_s0ix_supported(adev) || amdgpu_in_reset(adev))
 		r = amdgpu_device_ip_suspend_phase2(adev);
 	else
 		amdgpu_gfx_state_change_set(adev, sGpuChangeState_D3Entry);
@@ -3806,7 +3820,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 			}
 		}
 	}
-	r = amdgpu_amdkfd_resume(adev, !fbcon);
+	r = amdgpu_amdkfd_resume(adev, adev->in_runpm);
 	if (r)
 		return r;
 
@@ -4208,6 +4222,8 @@ bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev)
 		case CHIP_NAVI14:
 		case CHIP_NAVI12:
 		case CHIP_SIENNA_CICHLID:
+		case CHIP_NAVY_FLOUNDER:
+		case CHIP_DIMGREY_CAVEFISH:
 			break;
 		default:
 			goto disabled;
@@ -4457,6 +4473,46 @@ static void amdgpu_device_unlock_adev(struct amdgpu_device *adev)
 	up_write(&adev->reset_sem);
 }
 
+/*
+ * to lockup a list of amdgpu devices in a hive safely, if not a hive
+ * with multiple nodes, it will be similar as amdgpu_device_lock_adev.
+ *
+ * unlock won't require roll back.
+ */
+static int amdgpu_device_lock_hive_adev(struct amdgpu_device *adev, struct amdgpu_hive_info *hive)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+
+	if (adev->gmc.xgmi.num_physical_nodes > 1) {
+		if (!hive) {
+			dev_err(adev->dev, "Hive is NULL while device has multiple xgmi nodes");
+			return -ENODEV;
+		}
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			if (!amdgpu_device_lock_adev(tmp_adev, hive))
+				goto roll_back;
+		}
+	} else if (!amdgpu_device_lock_adev(adev, hive))
+		return -EAGAIN;
+
+	return 0;
+roll_back:
+	if (!list_is_first(&tmp_adev->gmc.xgmi.head, &hive->device_list)) {
+		/*
+		 * if the lockup iteration break in the middle of a hive,
+		 * it may means there may has a race issue,
+		 * or a hive device locked up independently.
+		 * we may be in trouble and may not, so will try to roll back
+		 * the lock and give out a warnning.
+		 */
+		dev_warn(tmp_adev->dev, "Hive lock iteration broke in the middle. Rolling back to unlock");
+		list_for_each_entry_continue_reverse(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			amdgpu_device_unlock_adev(tmp_adev);
+		}
+	}
+	return -EAGAIN;
+}
+
 static void amdgpu_device_resume_display_audio(struct amdgpu_device *adev)
 {
 	struct pci_dev *p = NULL;
@@ -4570,20 +4626,36 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 			DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress",
 				job ? job->base.id : -1, hive->hive_id);
 			amdgpu_put_xgmi_hive(hive);
+			if (job)
+				drm_sched_increase_karma(&job->base);
 			return 0;
 		}
 		mutex_lock(&hive->hive_lock);
 	}
 
 	/*
+	 * lock the device before we try to operate the linked list
+	 * if didn't get the device lock, don't touch the linked list since
+	 * others may iterating it.
+	 */
+	r = amdgpu_device_lock_hive_adev(adev, hive);
+	if (r) {
+		dev_info(adev->dev, "Bailing on TDR for s_job:%llx, as another already in progress",
+					job ? job->base.id : -1);
+
+		/* even we skipped this reset, still need to set the job to guilty */
+		if (job)
+			drm_sched_increase_karma(&job->base);
+		goto skip_recovery;
+	}
+
+	/*
 	 * Build list of devices to reset.
 	 * In case we are in XGMI hive mode, resort the device list
 	 * to put adev in the 1st position.
 	 */
 	INIT_LIST_HEAD(&device_list);
 	if (adev->gmc.xgmi.num_physical_nodes > 1) {
-		if (!hive)
-			return -ENODEV;
 		if (!list_is_first(&adev->gmc.xgmi.head, &hive->device_list))
 			list_rotate_to_front(&adev->gmc.xgmi.head, &hive->device_list);
 		device_list_handle = &hive->device_list;
@@ -4594,13 +4666,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 	/* block all schedulers and reset given job's ring */
 	list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
-		if (!amdgpu_device_lock_adev(tmp_adev, hive)) {
-			dev_info(tmp_adev->dev, "Bailing on TDR for s_job:%llx, as another already in progress",
-				  job ? job->base.id : -1);
-			r = 0;
-			goto skip_recovery;
-		}
-
 		/*
 		 * Try to put the audio codec into suspend state
 		 * before gpu reset started.
@@ -4738,7 +4803,7 @@ skip_recovery:
 		amdgpu_put_xgmi_hive(hive);
 	}
 
-	if (r)
+	if (r && r != -EAGAIN)
 		dev_info(adev->dev, "GPU reset end with ret = %d\n", r);
 	return r;
 }
@@ -4788,7 +4853,13 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
 						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3);
 		} else {
-			if (speed_cap == PCIE_SPEED_16_0GT)
+			if (speed_cap == PCIE_SPEED_32_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN4 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN5);
+			else if (speed_cap == PCIE_SPEED_16_0GT)
 				adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
 							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
 							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3 |
@@ -4808,7 +4879,13 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 			adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
 						   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2);
 		} else {
-			if (platform_speed_cap == PCIE_SPEED_16_0GT)
+			if (platform_speed_cap == PCIE_SPEED_32_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN5);
+			else if (platform_speed_cap == PCIE_SPEED_16_0GT)
 				adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
 							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
 							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3 |
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
index f764803c53a4..48cb33e5b382 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
@@ -926,8 +926,10 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 				       struct drm_file *file_priv,
 				       const struct drm_mode_fb_cmd2 *mode_cmd)
 {
-	struct drm_gem_object *obj;
 	struct amdgpu_framebuffer *amdgpu_fb;
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *bo;
+	uint32_t domains;
 	int ret;
 
 	obj = drm_gem_object_lookup(file_priv, mode_cmd->handles[0]);
@@ -938,7 +940,9 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 	}
 
 	/* Handle is imported dma-buf, so cannot be migrated to VRAM for scanout */
-	if (obj->import_attach) {
+	bo = gem_to_amdgpu_bo(obj);
+	domains = amdgpu_display_supported_domains(drm_to_adev(dev), bo->flags);
+	if (obj->import_attach && !(domains & AMDGPU_GEM_DOMAIN_GTT)) {
 		drm_dbg_kms(dev, "Cannot create framebuffer from imported dma_buf\n");
 		return ERR_PTR(-EINVAL);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 06dbb612a53a..47e0b48dc26f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -40,6 +40,7 @@
 #include <linux/dma-buf.h>
 #include <linux/dma-fence-array.h>
 #include <linux/pci-p2pdma.h>
+#include <linux/pm_runtime.h>
 
 /**
  * amdgpu_gem_prime_mmap - &drm_driver.gem_prime_mmap implementation
@@ -151,9 +152,13 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	if (attach->dev->driver == adev->dev->driver)
 		return 0;
 
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0)
+		goto out;
+
 	r = amdgpu_bo_reserve(bo, false);
 	if (unlikely(r != 0))
-		return r;
+		goto out;
 
 	/*
 	 * We only create shared fences for internal use, but importers
@@ -165,11 +170,15 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	 */
 	r = __dma_resv_make_exclusive(bo->tbo.base.resv);
 	if (r)
-		return r;
+		goto out;
 
 	bo->prime_shared_count++;
 	amdgpu_bo_unreserve(bo);
 	return 0;
+
+out:
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+	return r;
 }
 
 /**
@@ -189,6 +198,9 @@ static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf,
 
 	if (attach->dev->driver != adev->dev->driver && bo->prime_shared_count)
 		bo->prime_shared_count--;
+
+	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 }
 
 /**
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index b4ea67e12ada..b26e2fd1c538 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -132,8 +132,12 @@ uint amdgpu_pg_mask = 0xffffffff;
 uint amdgpu_sdma_phase_quantum = 32;
 char *amdgpu_disable_cu = NULL;
 char *amdgpu_virtual_display = NULL;
-/* OverDrive(bit 14) disabled by default*/
-uint amdgpu_pp_feature_mask = 0xffffbfff;
+
+/*
+ * OverDrive(bit 14) disabled by default
+ * GFX DCS(bit 19) disabled by default
+ */
+uint amdgpu_pp_feature_mask = 0xfff7bfff;
 uint amdgpu_force_long_training;
 int amdgpu_job_hang_limit;
 int amdgpu_lbpw = -1;
@@ -777,6 +781,10 @@ uint amdgpu_dm_abm_level;
 MODULE_PARM_DESC(abmlevel, "ABM level (0 = off (default), 1-4 = backlight reduction level) ");
 module_param_named(abmlevel, amdgpu_dm_abm_level, uint, 0444);
 
+int amdgpu_backlight = -1;
+MODULE_PARM_DESC(backlight, "Backlight control (0 = pwm, 1 = aux, -1 auto (default))");
+module_param_named(backlight, amdgpu_backlight, bint, 0444);
+
 /**
  * DOC: tmz (int)
  * Trusted Memory Zone (TMZ) is a method to protect data being written
@@ -789,9 +797,9 @@ module_param_named(tmz, amdgpu_tmz, int, 0444);
 
 /**
  * DOC: reset_method (int)
- * GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco)
+ * GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco, 5 = pci)
  */
-MODULE_PARM_DESC(reset_method, "GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco/bamaco)");
+MODULE_PARM_DESC(reset_method, "GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco/bamaco, 5 = pci)");
 module_param_named(reset_method, amdgpu_reset_method, int, 0444);
 
 /**
@@ -1085,6 +1093,8 @@ static const struct pci_device_id pciidlist[] = {
 
 	/* Renoir */
 	{0x1002, 0x1636, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RENOIR|AMD_IS_APU},
+	{0x1002, 0x1638, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RENOIR|AMD_IS_APU},
+	{0x1002, 0x164C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RENOIR|AMD_IS_APU},
 
 	/* Navi12 */
 	{0x1002, 0x7360, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_NAVI12},
@@ -1092,6 +1102,7 @@ static const struct pci_device_id pciidlist[] = {
 
 	/* Sienna_Cichlid */
 	{0x1002, 0x73A0, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
+	{0x1002, 0x73A1, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 	{0x1002, 0x73A2, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 	{0x1002, 0x73A3, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 	{0x1002, 0x73AB, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
@@ -1263,7 +1274,9 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
 	 */
 	if (!amdgpu_passthrough(adev))
 		adev->mp1_state = PP_MP1_STATE_UNLOAD;
+	adev->in_poweroff_reboot_com = true;
 	amdgpu_device_ip_suspend(adev);
+	adev->in_poweroff_reboot_com = false;
 	adev->mp1_state = PP_MP1_STATE_NONE;
 }
 
@@ -1305,8 +1318,13 @@ static int amdgpu_pmops_thaw(struct device *dev)
 static int amdgpu_pmops_poweroff(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+	int r;
 
-	return amdgpu_device_suspend(drm_dev, true);
+	adev->in_poweroff_reboot_com = true;
+	r =  amdgpu_device_suspend(drm_dev, true);
+	adev->in_poweroff_reboot_com = false;
+	return r;
 }
 
 static int amdgpu_pmops_restore(struct device *dev)
@@ -1341,11 +1359,12 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 	adev->in_runpm = true;
 	if (amdgpu_device_supports_atpx(drm_dev))
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
-	drm_kms_helper_poll_disable(drm_dev);
 
 	ret = amdgpu_device_suspend(drm_dev, false);
-	if (ret)
+	if (ret) {
+		adev->in_runpm = false;
 		return ret;
+	}
 
 	if (amdgpu_device_supports_atpx(drm_dev)) {
 		/* Only need to handle PCI state in the driver for ATPX
@@ -1398,7 +1417,6 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 		amdgpu_device_baco_exit(drm_dev);
 	}
 	ret = amdgpu_device_resume(drm_dev, false);
-	drm_kms_helper_poll_enable(drm_dev);
 	if (amdgpu_device_supports_atpx(drm_dev))
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
 	adev->in_runpm = false;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
index 51cd49c6f38f..24010cacf7d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
@@ -146,7 +146,7 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
 	size = mode_cmd->pitches[0] * height;
 	aligned_size = ALIGN(size, PAGE_SIZE);
 	ret = amdgpu_gem_object_create(adev, aligned_size, 0, domain, flags,
-				       ttm_bo_type_kernel, NULL, &gobj);
+				       ttm_bo_type_device, NULL, &gobj);
 	if (ret) {
 		pr_err("failed to allocate framebuffer (%d)\n", aligned_size);
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c
index 7c6e02e35573..8d1ad294cb02 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c
@@ -47,10 +47,9 @@ typedef struct FW_ATT_RECORD
 	uint16_t AttFwIdV2;              /* V2 FW ID field */
 	uint32_t AttFWVersion;           /* FW Version */
 	uint16_t AttFWActiveFunctionID;  /* The VF ID (only in VF Attestation Table) */
-	uint16_t AttSource;              /* FW source indicator */
-	uint16_t RecordValid;            /* Indicates whether the record is a valid entry */
-	uint8_t  AttFwTaId;              /* Ta ID (only in TA Attestation Table) */
-	uint8_t  Reserved;
+	uint8_t  AttSource;              /* FW source indicator */
+	uint8_t  RecordValid;            /* Indicates whether the record is a valid entry */
+	uint32_t AttFwTaId;              /* Ta ID (only in TA Attestation Table) */
 } FW_ATT_RECORD;
 
 static ssize_t amdgpu_fw_attestation_debugfs_read(struct file *f,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index a5c42c3004a0..b443907afcea 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -269,8 +269,8 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 		resv = vm->root.base.bo->tbo.base.resv;
 	}
 
-retry:
 	initial_domain = (u32)(0xffffffff & args->in.domains);
+retry:
 	r = amdgpu_gem_object_create(adev, size, args->in.alignment,
 				     initial_domain,
 				     flags, ttm_bo_type_device, resv, &gobj);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index cd2c676a2797..8e0a6c62322e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -193,15 +193,16 @@ static bool amdgpu_gfx_is_multipipe_capable(struct amdgpu_device *adev)
 }
 
 bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev,
-					       int pipe, int queue)
+					       struct amdgpu_ring *ring)
 {
-	bool multipipe_policy = amdgpu_gfx_is_multipipe_capable(adev);
-	int cond;
-	/* Policy: alternate between normal and high priority */
-	cond = multipipe_policy ? pipe : queue;
-
-	return ((cond % 2) != 0);
+	/* Policy: use 1st queue as high priority compute queue if we
+	 * have more than one compute queue.
+	 */
+	if (adev->gfx.num_compute_rings > 1 &&
+	    ring == &adev->gfx.compute_ring[0])
+		return true;
 
+	return false;
 }
 
 void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index 6b5a8f4642cc..72dbcd2bc6a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -380,7 +380,7 @@ void amdgpu_queue_mask_bit_to_mec_queue(struct amdgpu_device *adev, int bit,
 bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, int mec,
 				     int pipe, int queue);
 bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev,
-					       int pipe, int queue);
+					       struct amdgpu_ring *ring);
 int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, int me,
 			       int pipe, int queue);
 void amdgpu_gfx_bit_to_me_queue(struct amdgpu_device *adev, int bit,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h
new file mode 100644
index 000000000000..43caf9f8cc11
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __AMDGPU_HDP_H__
+#define __AMDGPU_HDP_H__
+
+struct amdgpu_hdp_funcs {
+	void (*flush_hdp)(struct amdgpu_device *adev, struct amdgpu_ring *ring);
+	void (*invalidate_hdp)(struct amdgpu_device *adev,
+			       struct amdgpu_ring *ring);
+	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+	void (*update_clock_gating)(struct amdgpu_device *adev, bool enable);
+	void (*get_clock_gating_state)(struct amdgpu_device *adev, u32 *flags);
+	void (*init_registers)(struct amdgpu_device *adev);
+};
+
+struct amdgpu_hdp {
+	const struct amdgpu_hdp_funcs		*funcs;
+};
+
+#endif /* __AMDGPU_HDP_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 024d0a563a65..7645223ea0ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -195,6 +195,10 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	if ((ib->flags & AMDGPU_IB_FLAG_EMIT_MEM_SYNC) && ring->funcs->emit_mem_sync)
 		ring->funcs->emit_mem_sync(ring);
 
+	if (ring->funcs->emit_wave_limit &&
+	    ring->hw_prio == AMDGPU_GFX_PIPE_PRIO_HIGH)
+		ring->funcs->emit_wave_limit(ring, true);
+
 	if (ring->funcs->insert_start)
 		ring->funcs->insert_start(ring);
 
@@ -295,6 +299,11 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	ring->current_ctx = fence_ctx;
 	if (vm && ring->funcs->emit_switch_buffer)
 		amdgpu_ring_emit_switch_buffer(ring);
+
+	if (ring->funcs->emit_wave_limit &&
+	    ring->hw_prio == AMDGPU_GFX_PIPE_PRIO_HIGH)
+		ring->funcs->emit_wave_limit(ring, false);
+
 	amdgpu_ring_commit(ring);
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
index dcd9b4a8e20b..dc852af4f3b7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
@@ -205,3 +205,48 @@ restart_ih:
 	return IRQ_HANDLED;
 }
 
+/**
+ * amdgpu_ih_decode_iv_helper - decode an interrupt vector
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: ih ring to process
+ * @entry: IV entry
+ *
+ * Decodes the interrupt vector at the current rptr
+ * position and also advance the position for for Vega10
+ * and later GPUs.
+ */
+void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
+				struct amdgpu_ih_ring *ih,
+				struct amdgpu_iv_entry *entry)
+{
+	/* wptr/rptr are in bytes! */
+	u32 ring_index = ih->rptr >> 2;
+	uint32_t dw[8];
+
+	dw[0] = le32_to_cpu(ih->ring[ring_index + 0]);
+	dw[1] = le32_to_cpu(ih->ring[ring_index + 1]);
+	dw[2] = le32_to_cpu(ih->ring[ring_index + 2]);
+	dw[3] = le32_to_cpu(ih->ring[ring_index + 3]);
+	dw[4] = le32_to_cpu(ih->ring[ring_index + 4]);
+	dw[5] = le32_to_cpu(ih->ring[ring_index + 5]);
+	dw[6] = le32_to_cpu(ih->ring[ring_index + 6]);
+	dw[7] = le32_to_cpu(ih->ring[ring_index + 7]);
+
+	entry->client_id = dw[0] & 0xff;
+	entry->src_id = (dw[0] >> 8) & 0xff;
+	entry->ring_id = (dw[0] >> 16) & 0xff;
+	entry->vmid = (dw[0] >> 24) & 0xf;
+	entry->vmid_src = (dw[0] >> 31);
+	entry->timestamp = dw[1] | ((u64)(dw[2] & 0xffff) << 32);
+	entry->timestamp_src = dw[2] >> 31;
+	entry->pasid = dw[3] & 0xffff;
+	entry->pasid_src = dw[3] >> 31;
+	entry->src_data[0] = dw[4];
+	entry->src_data[1] = dw[5];
+	entry->src_data[2] = dw[6];
+	entry->src_data[3] = dw[7];
+
+	/* wptr/rptr are in bytes! */
+	ih->rptr += 32;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
index 3c9cfe7eecff..6ed4a85fc7c3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
@@ -30,6 +30,18 @@
 struct amdgpu_device;
 struct amdgpu_iv_entry;
 
+struct amdgpu_ih_regs {
+	uint32_t ih_rb_base;
+	uint32_t ih_rb_base_hi;
+	uint32_t ih_rb_cntl;
+	uint32_t ih_rb_wptr;
+	uint32_t ih_rb_rptr;
+	uint32_t ih_doorbell_rptr;
+	uint32_t ih_rb_wptr_addr_lo;
+	uint32_t ih_rb_wptr_addr_hi;
+	uint32_t psp_reg_id;
+};
+
 /*
  * R6xx+ IH ring
  */
@@ -53,6 +65,7 @@ struct amdgpu_ih_ring {
 	bool                    enabled;
 	unsigned		rptr;
 	atomic_t		lock;
+	struct amdgpu_ih_regs	ih_regs;
 };
 
 /* provided by the ih block */
@@ -75,5 +88,7 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv,
 			  unsigned int num_dw);
 int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
-
+void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
+				struct amdgpu_ih_ring *ih,
+				struct amdgpu_iv_entry *entry);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
index bea57e8e793f..afbbec82a289 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
@@ -444,7 +444,8 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev,
 	} else	if (src_id >= AMDGPU_MAX_IRQ_SRC_ID) {
 		DRM_DEBUG("Invalid src_id in IV: %d\n", src_id);
 
-	} else if (adev->irq.virq[src_id]) {
+	} else if ((client_id == AMDGPU_IRQ_CLIENTID_LEGACY) &&
+		   adev->irq.virq[src_id]) {
 		generic_handle_irq(irq_find_mapping(adev->irq.domain, src_id));
 
 	} else if (!adev->irq.client[client_id].sources) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 3c37cf1ae8b7..64beb3399604 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -173,8 +173,6 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags)
 		switch (adev->asic_type) {
 		case CHIP_VEGA20:
 		case CHIP_ARCTURUS:
-		case CHIP_SIENNA_CICHLID:
-		case CHIP_NAVY_FLOUNDER:
 			/* enable runpm if runpm=1 */
 			if (amdgpu_runtime_pm > 0)
 				adev->runpm = true;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h
index e62cc0e1a5ad..7c11bce4514b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h
@@ -57,7 +57,6 @@ struct amdgpu_nbio_funcs {
 	u32 (*get_pcie_port_data_offset)(struct amdgpu_device *adev);
 	u32 (*get_rev_id)(struct amdgpu_device *adev);
 	void (*mc_access_enable)(struct amdgpu_device *adev, bool enable);
-	void (*hdp_flush)(struct amdgpu_device *adev, struct amdgpu_ring *ring);
 	u32 (*get_memsize)(struct amdgpu_device *adev);
 	void (*sdma_doorbell_range)(struct amdgpu_device *adev, int instance,
 			bool use_doorbell, int doorbell_index, int doorbell_size);
@@ -89,6 +88,7 @@ struct amdgpu_nbio_funcs {
 	int (*ras_late_init)(struct amdgpu_device *adev);
 	void (*enable_aspm)(struct amdgpu_device *adev,
 			    bool enable);
+	void (*program_aspm)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_nbio {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 599c9a132eb6..984dcf5a475e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -893,7 +893,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 		return -EINVAL;
 
 	/* A shared bo cannot be migrated to VRAM */
-	if (bo->prime_shared_count) {
+	if (bo->prime_shared_count || bo->tbo.base.import_attach) {
 		if (domain & AMDGPU_GEM_DOMAIN_GTT)
 			domain = AMDGPU_GEM_DOMAIN_GTT;
 		else
@@ -907,10 +907,16 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 
 	if (bo->tbo.pin_count) {
 		uint32_t mem_type = bo->tbo.mem.mem_type;
+		uint32_t mem_flags = bo->tbo.mem.placement;
 
 		if (!(domain & amdgpu_mem_type_to_domain(mem_type)))
 			return -EINVAL;
 
+		if ((mem_type == TTM_PL_VRAM) &&
+		    (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS) &&
+		    !(mem_flags & TTM_PL_FLAG_CONTIGUOUS))
+			return -EINVAL;
+
 		ttm_bo_pin(&bo->tbo);
 
 		if (max_offset != 0) {
@@ -926,7 +932,6 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	if (bo->tbo.base.import_attach)
 		dma_buf_pin(bo->tbo.base.import_attach);
 
-	bo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	/* force to pin into visible video ram */
 	if (!(bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS))
 		bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
@@ -979,6 +984,7 @@ error:
  */
 int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain)
 {
+	bo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	return amdgpu_bo_pin_restricted(bo, domain, 0, 0);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 523d22db094b..839917eb7bc3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -36,6 +36,7 @@
 #include "psp_v12_0.h"
 
 #include "amdgpu_ras.h"
+#include "amdgpu_securedisplay.h"
 
 static int psp_sysfs_init(struct amdgpu_device *adev);
 static void psp_sysfs_fini(struct amdgpu_device *adev);
@@ -249,7 +250,7 @@ psp_cmd_submit_buf(struct psp_context *psp,
 {
 	int ret;
 	int index;
-	int timeout = 2000;
+	int timeout = 20000;
 	bool ras_intr = false;
 	bool skip_unsupport = false;
 
@@ -282,7 +283,7 @@ psp_cmd_submit_buf(struct psp_context *psp,
 		ras_intr = amdgpu_ras_intr_triggered();
 		if (ras_intr)
 			break;
-		msleep(1);
+		usleep_range(10, 100);
 		amdgpu_asic_invalidate_hdp(psp->adev, NULL);
 	}
 
@@ -563,7 +564,7 @@ static int psp_asd_load(struct psp_context *psp)
 	 * add workaround to bypass it for sriov now.
 	 * TODO: add version check to make it common
 	 */
-	if (amdgpu_sriov_vf(psp->adev) || !psp->asd_fw)
+	if (amdgpu_sriov_vf(psp->adev) || !psp->asd_ucode_size)
 		return 0;
 
 	cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL);
@@ -1315,8 +1316,12 @@ static int psp_hdcp_terminate(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
-	if (!psp->hdcp_context.hdcp_initialized)
-		return 0;
+	if (!psp->hdcp_context.hdcp_initialized) {
+		if (psp->hdcp_context.hdcp_shared_buf)
+			goto out;
+		else
+			return 0;
+	}
 
 	ret = psp_hdcp_unload(psp);
 	if (ret)
@@ -1324,6 +1329,7 @@ static int psp_hdcp_terminate(struct psp_context *psp)
 
 	psp->hdcp_context.hdcp_initialized = false;
 
+out:
 	/* free hdcp shared memory */
 	amdgpu_bo_free_kernel(&psp->hdcp_context.hdcp_shared_bo,
 			      &psp->hdcp_context.hdcp_shared_mc_addr,
@@ -1462,8 +1468,12 @@ static int psp_dtm_terminate(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
-	if (!psp->dtm_context.dtm_initialized)
-		return 0;
+	if (!psp->dtm_context.dtm_initialized) {
+		if (psp->dtm_context.dtm_shared_buf)
+			goto out;
+		else
+			return 0;
+	}
 
 	ret = psp_dtm_unload(psp);
 	if (ret)
@@ -1471,6 +1481,7 @@ static int psp_dtm_terminate(struct psp_context *psp)
 
 	psp->dtm_context.dtm_initialized = false;
 
+out:
 	/* free hdcp shared memory */
 	amdgpu_bo_free_kernel(&psp->dtm_context.dtm_shared_bo,
 			      &psp->dtm_context.dtm_shared_mc_addr,
@@ -1642,6 +1653,175 @@ int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 }
 // RAP end
 
+/* securedisplay start */
+static int psp_securedisplay_init_shared_buf(struct psp_context *psp)
+{
+	int ret;
+
+	/*
+	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
+	 * physical) for sa ta <-> Driver
+	 */
+	ret = amdgpu_bo_create_kernel(psp->adev, PSP_SECUREDISPLAY_SHARED_MEM_SIZE,
+				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
+				      &psp->securedisplay_context.securedisplay_shared_bo,
+				      &psp->securedisplay_context.securedisplay_shared_mc_addr,
+				      &psp->securedisplay_context.securedisplay_shared_buf);
+
+	return ret;
+}
+
+static int psp_securedisplay_load(struct psp_context *psp)
+{
+	int ret;
+	struct psp_gfx_cmd_resp *cmd;
+
+	cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL);
+	if (!cmd)
+		return -ENOMEM;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+	memcpy(psp->fw_pri_buf, psp->ta_securedisplay_start_addr, psp->ta_securedisplay_ucode_size);
+
+	psp_prep_ta_load_cmd_buf(cmd,
+				 psp->fw_pri_mc_addr,
+				 psp->ta_securedisplay_ucode_size,
+				 psp->securedisplay_context.securedisplay_shared_mc_addr,
+				 PSP_SECUREDISPLAY_SHARED_MEM_SIZE);
+
+	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+
+	if (ret)
+		goto failed;
+
+	psp->securedisplay_context.securedisplay_initialized = true;
+	psp->securedisplay_context.session_id = cmd->resp.session_id;
+	mutex_init(&psp->securedisplay_context.mutex);
+
+failed:
+	kfree(cmd);
+	return ret;
+}
+
+static int psp_securedisplay_unload(struct psp_context *psp)
+{
+	int ret;
+	struct psp_gfx_cmd_resp *cmd;
+
+	cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL);
+	if (!cmd)
+		return -ENOMEM;
+
+	psp_prep_ta_unload_cmd_buf(cmd, psp->securedisplay_context.session_id);
+
+	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+
+	kfree(cmd);
+
+	return ret;
+}
+
+static int psp_securedisplay_initialize(struct psp_context *psp)
+{
+	int ret;
+	struct securedisplay_cmd *securedisplay_cmd;
+
+	/*
+	 * TODO: bypass the initialize in sriov for now
+	 */
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	if (!psp->adev->psp.ta_securedisplay_ucode_size ||
+	    !psp->adev->psp.ta_securedisplay_start_addr) {
+		dev_info(psp->adev->dev, "SECUREDISPLAY: securedisplay ta ucode is not available\n");
+		return 0;
+	}
+
+	if (!psp->securedisplay_context.securedisplay_initialized) {
+		ret = psp_securedisplay_init_shared_buf(psp);
+		if (ret)
+			return ret;
+	}
+
+	ret = psp_securedisplay_load(psp);
+	if (ret)
+		return ret;
+
+	psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
+			TA_SECUREDISPLAY_COMMAND__QUERY_TA);
+
+	ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__QUERY_TA);
+	if (ret) {
+		psp_securedisplay_unload(psp);
+
+		amdgpu_bo_free_kernel(&psp->securedisplay_context.securedisplay_shared_bo,
+			      &psp->securedisplay_context.securedisplay_shared_mc_addr,
+			      &psp->securedisplay_context.securedisplay_shared_buf);
+
+		psp->securedisplay_context.securedisplay_initialized = false;
+
+		dev_err(psp->adev->dev, "SECUREDISPLAY TA initialize fail.\n");
+		return -EINVAL;
+	}
+
+	if (securedisplay_cmd->status != TA_SECUREDISPLAY_STATUS__SUCCESS) {
+		psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
+		dev_err(psp->adev->dev, "SECUREDISPLAY: query securedisplay TA failed. ret 0x%x\n",
+			securedisplay_cmd->securedisplay_out_message.query_ta.query_cmd_ret);
+	}
+
+	return 0;
+}
+
+static int psp_securedisplay_terminate(struct psp_context *psp)
+{
+	int ret;
+
+	/*
+	 * TODO:bypass the terminate in sriov for now
+	 */
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	if (!psp->securedisplay_context.securedisplay_initialized)
+		return 0;
+
+	ret = psp_securedisplay_unload(psp);
+	if (ret)
+		return ret;
+
+	psp->securedisplay_context.securedisplay_initialized = false;
+
+	/* free securedisplay shared memory */
+	amdgpu_bo_free_kernel(&psp->securedisplay_context.securedisplay_shared_bo,
+			      &psp->securedisplay_context.securedisplay_shared_mc_addr,
+			      &psp->securedisplay_context.securedisplay_shared_buf);
+
+	return ret;
+}
+
+int psp_securedisplay_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
+{
+	int ret;
+
+	if (!psp->securedisplay_context.securedisplay_initialized)
+		return -EINVAL;
+
+	if (ta_cmd_id != TA_SECUREDISPLAY_COMMAND__QUERY_TA &&
+	    ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC)
+		return -EINVAL;
+
+	mutex_lock(&psp->securedisplay_context.mutex);
+
+	ret = psp_ta_invoke(psp, ta_cmd_id, psp->securedisplay_context.session_id);
+
+	mutex_unlock(&psp->securedisplay_context.mutex);
+
+	return ret;
+}
+/* SECUREDISPLAY end */
+
 static int psp_hw_start(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
@@ -2116,6 +2296,11 @@ skip_memalloc:
 		if (ret)
 			dev_err(psp->adev->dev,
 				"RAP: Failed to initialize RAP\n");
+
+		ret = psp_securedisplay_initialize(psp);
+		if (ret)
+			dev_err(psp->adev->dev,
+				"SECUREDISPLAY: Failed to initialize SECUREDISPLAY\n");
 	}
 
 	return 0;
@@ -2166,6 +2351,7 @@ static int psp_hw_fini(void *handle)
 
 	if (psp->adev->psp.ta_fw) {
 		psp_ras_terminate(psp);
+		psp_securedisplay_terminate(psp);
 		psp_rap_terminate(psp);
 		psp_dtm_terminate(psp);
 		psp_hdcp_terminate(psp);
@@ -2230,6 +2416,11 @@ static int psp_suspend(void *handle)
 			DRM_ERROR("Failed to terminate rap ta\n");
 			return ret;
 		}
+		ret = psp_securedisplay_terminate(psp);
+		if (ret) {
+			DRM_ERROR("Failed to terminate securedisplay ta\n");
+			return ret;
+		}
 	}
 
 	ret = psp_asd_unload(psp);
@@ -2313,6 +2504,11 @@ static int psp_resume(void *handle)
 		if (ret)
 			dev_err(psp->adev->dev,
 				"RAP: Failed to initialize RAP\n");
+
+		ret = psp_securedisplay_initialize(psp);
+		if (ret)
+			dev_err(psp->adev->dev,
+				"SECUREDISPLAY: Failed to initialize SECUREDISPLAY\n");
 	}
 
 	mutex_unlock(&adev->firmware.mutex);
@@ -2589,11 +2785,10 @@ static int parse_ta_bin_descriptor(struct psp_context *psp,
 
 	switch (desc->fw_type) {
 	case TA_FW_TYPE_PSP_ASD:
-		psp->asd_fw_version	   = le32_to_cpu(desc->fw_version);
+		psp->asd_fw_version        = le32_to_cpu(desc->fw_version);
 		psp->asd_feature_version   = le32_to_cpu(desc->fw_version);
-		psp->asd_ucode_size	   = le32_to_cpu(desc->size_bytes);
+		psp->asd_ucode_size        = le32_to_cpu(desc->size_bytes);
 		psp->asd_start_addr 	   = ucode_start_addr;
-		psp->asd_fw                = psp->ta_fw;
 		break;
 	case TA_FW_TYPE_PSP_XGMI:
 		psp->ta_xgmi_ucode_version = le32_to_cpu(desc->fw_version);
@@ -2620,6 +2815,11 @@ static int parse_ta_bin_descriptor(struct psp_context *psp,
 		psp->ta_rap_ucode_size     = le32_to_cpu(desc->size_bytes);
 		psp->ta_rap_start_addr     = ucode_start_addr;
 		break;
+	case TA_FW_TYPE_PSP_SECUREDISPLAY:
+		psp->ta_securedisplay_ucode_version  = le32_to_cpu(desc->fw_version);
+		psp->ta_securedisplay_ucode_size     = le32_to_cpu(desc->size_bytes);
+		psp->ta_securedisplay_start_addr     = ucode_start_addr;
+		break;
 	default:
 		dev_warn(psp->adev->dev, "Unsupported TA type: %d\n", desc->fw_type);
 		break;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index da250bc1ac57..cb50ba445f8c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -30,6 +30,7 @@
 #include "ta_xgmi_if.h"
 #include "ta_ras_if.h"
 #include "ta_rap_if.h"
+#include "ta_secureDisplay_if.h"
 
 #define PSP_FENCE_BUFFER_SIZE	0x1000
 #define PSP_CMD_BUFFER_SIZE	0x1000
@@ -40,6 +41,7 @@
 #define PSP_HDCP_SHARED_MEM_SIZE	0x4000
 #define PSP_DTM_SHARED_MEM_SIZE	0x4000
 #define PSP_RAP_SHARED_MEM_SIZE	0x4000
+#define PSP_SECUREDISPLAY_SHARED_MEM_SIZE	0x4000
 #define PSP_SHARED_MEM_SIZE		0x4000
 #define PSP_FW_NAME_LEN		0x24
 
@@ -171,6 +173,15 @@ struct psp_rap_context {
 	struct mutex		mutex;
 };
 
+struct psp_securedisplay_context {
+	bool			securedisplay_initialized;
+	uint32_t		session_id;
+	struct amdgpu_bo	*securedisplay_shared_bo;
+	uint64_t		securedisplay_shared_mc_addr;
+	void			*securedisplay_shared_buf;
+	struct mutex		mutex;
+};
+
 #define MEM_TRAIN_SYSTEM_SIGNATURE		0x54534942
 #define GDDR6_MEM_TRAINING_DATA_SIZE_IN_BYTES	0x1000
 #define GDDR6_MEM_TRAINING_OFFSET		0x8000
@@ -298,12 +309,17 @@ struct psp_context
 	uint32_t			ta_rap_ucode_size;
 	uint8_t				*ta_rap_start_addr;
 
+	uint32_t			ta_securedisplay_ucode_version;
+	uint32_t			ta_securedisplay_ucode_size;
+	uint8_t				*ta_securedisplay_start_addr;
+
 	struct psp_asd_context		asd_context;
 	struct psp_xgmi_context		xgmi_context;
 	struct psp_ras_context		ras;
 	struct psp_hdcp_context 	hdcp_context;
 	struct psp_dtm_context		dtm_context;
 	struct psp_rap_context		rap_context;
+	struct psp_securedisplay_context	securedisplay_context;
 	struct mutex			mutex;
 	struct psp_memory_training_context mem_train_ctx;
 };
@@ -380,6 +396,7 @@ int psp_ras_trigger_error(struct psp_context *psp,
 int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
+int psp_securedisplay_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 
 int psp_rlc_autoload_start(struct psp_context *psp);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index c136bd449744..1fb2a91ad30a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -846,7 +846,7 @@ static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev,
 	if (amdgpu_dpm_allow_xgmi_power_down(adev, true))
 		dev_warn(adev->dev, "Failed to allow XGMI power down");
 
-	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW))
+	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_ALLOW))
 		dev_warn(adev->dev, "Failed to allow df cstate");
 
 	return ret;
@@ -1518,7 +1518,7 @@ static int amdgpu_ras_badpages_read(struct amdgpu_device *adev,
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data *data;
 	int i = 0;
-	int ret = 0;
+	int ret = 0, status;
 
 	if (!con || !con->eh_data || !bps || !count)
 		return -EINVAL;
@@ -1543,12 +1543,12 @@ static int amdgpu_ras_badpages_read(struct amdgpu_device *adev,
 			.size = AMDGPU_GPU_PAGE_SIZE,
 			.flags = AMDGPU_RAS_RETIRE_PAGE_RESERVED,
 		};
-		ret = amdgpu_vram_mgr_query_page_status(
+		status = amdgpu_vram_mgr_query_page_status(
 				ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM),
 				data->bps[i].retired_page);
-		if (ret == -EBUSY)
+		if (status == -EBUSY)
 			(*bps)[i].flags = AMDGPU_RAS_RETIRE_PAGE_PENDING;
-		else if (ret == -ENOENT)
+		else if (status == -ENOENT)
 			(*bps)[i].flags = AMDGPU_RAS_RETIRE_PAGE_FAULT;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 1dd040166c63..19d9aa76cfbf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -30,6 +30,7 @@
 #define EEPROM_I2C_TARGET_ADDR_VEGA20		0xA0
 #define EEPROM_I2C_TARGET_ADDR_ARCTURUS		0xA8
 #define EEPROM_I2C_TARGET_ADDR_ARCTURUS_D342	0xA0
+#define EEPROM_I2C_TARGET_ADDR_SIENNA_CICHLID   0xA0
 
 /*
  * The 2 macros bellow represent the actual size in bytes that
@@ -62,7 +63,8 @@
 static bool __is_ras_eeprom_supported(struct amdgpu_device *adev)
 {
 	if ((adev->asic_type == CHIP_VEGA20) ||
-	    (adev->asic_type == CHIP_ARCTURUS))
+	    (adev->asic_type == CHIP_ARCTURUS) ||
+	    (adev->asic_type == CHIP_SIENNA_CICHLID))
 		return true;
 
 	return false;
@@ -100,6 +102,10 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 	case CHIP_ARCTURUS:
 		return __get_eeprom_i2c_addr_arct(adev, i2c_addr);
 
+	case CHIP_SIENNA_CICHLID:
+		*i2c_addr = EEPROM_I2C_TARGET_ADDR_SIENNA_CICHLID;
+		break;
+
 	default:
 		return false;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
index 1a612f51ecd9..b644c78475fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
@@ -166,7 +166,7 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		     unsigned int max_dw, struct amdgpu_irq_src *irq_src,
 		     unsigned int irq_type, unsigned int hw_prio)
 {
-	int r, i;
+	int r;
 	int sched_hw_submission = amdgpu_sched_hw_submission;
 	u32 *num_sched;
 	u32 hw_ip;
@@ -258,8 +258,7 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 	}
 
 	ring->max_dw = max_dw;
-	ring->priority = DRM_SCHED_PRIORITY_NORMAL;
-	mutex_init(&ring->priority_mutex);
+	ring->hw_prio = hw_prio;
 
 	if (!ring->no_scheduler) {
 		hw_ip = ring->funcs->type;
@@ -268,9 +267,6 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 			&ring->sched;
 	}
 
-	for (i = DRM_SCHED_PRIORITY_MIN; i < DRM_SCHED_PRIORITY_COUNT; ++i)
-		atomic_set(&ring->num_jobs[i], 0);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
index 7112137689db..56acec1075ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
@@ -197,6 +197,7 @@ struct amdgpu_ring_funcs {
 	void (*soft_recovery)(struct amdgpu_ring *ring, unsigned vmid);
 	int (*preempt_ib)(struct amdgpu_ring *ring);
 	void (*emit_mem_sync)(struct amdgpu_ring *ring);
+	void (*emit_wave_limit)(struct amdgpu_ring *ring, bool enable);
 };
 
 struct amdgpu_ring {
@@ -242,11 +243,7 @@ struct amdgpu_ring {
 	struct dma_fence	*vmid_wait;
 	bool			has_compute_vm_bug;
 	bool			no_scheduler;
-
-	atomic_t		num_jobs[DRM_SCHED_PRIORITY_COUNT];
-	struct mutex		priority_mutex;
-	/* protected by priority_mutex */
-	int			priority;
+	int			hw_prio;
 
 #if defined(CONFIG_DEBUG_FS)
 	struct dentry *ent;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
new file mode 100644
index 000000000000..834440ab9ff7
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
@@ -0,0 +1,176 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ *
+ */
+#include <linux/debugfs.h>
+#include <linux/pm_runtime.h>
+
+#include "amdgpu.h"
+#include "amdgpu_securedisplay.h"
+
+/**
+ * DOC: AMDGPU SECUREDISPLAY debugfs test interface
+ *
+ * how to use?
+ * echo opcode <value> > <debugfs_dir>/dri/xxx/securedisplay_test
+ * eg. echo 1 > <debugfs_dir>/dri/xxx/securedisplay_test
+ * eg. echo 2 phy_id > <debugfs_dir>/dri/xxx/securedisplay_test
+ *
+ * opcode:
+ * 1：Query whether TA is responding used only for validation pupose
+ * 2: Send region of Interest and CRC value to I2C. (uint32)phy_id is
+ * send to determine which DIO scratch register should be used to get
+ * ROI and receive i2c_buf as the output.
+ *
+ * You can refer more detail from header file ta_securedisplay_if.h
+ *
+ */
+
+void psp_securedisplay_parse_resp_status(struct psp_context *psp,
+	enum ta_securedisplay_status status)
+{
+	switch (status) {
+	case TA_SECUREDISPLAY_STATUS__SUCCESS:
+		break;
+	case TA_SECUREDISPLAY_STATUS__GENERIC_FAILURE:
+		dev_err(psp->adev->dev, "Secure display: Generic Failure.");
+		break;
+	case TA_SECUREDISPLAY_STATUS__INVALID_PARAMETER:
+		dev_err(psp->adev->dev, "Secure display: Invalid Parameter.");
+		break;
+	case TA_SECUREDISPLAY_STATUS__NULL_POINTER:
+		dev_err(psp->adev->dev, "Secure display: Null Pointer.");
+		break;
+	case TA_SECUREDISPLAY_STATUS__I2C_WRITE_ERROR:
+		dev_err(psp->adev->dev, "Secure display: Failed to write to I2C.");
+		break;
+	case TA_SECUREDISPLAY_STATUS__READ_DIO_SCRATCH_ERROR:
+		dev_err(psp->adev->dev, "Secure display: Failed to Read DIO Scratch Register.");
+		break;
+	case TA_SECUREDISPLAY_STATUS__READ_CRC_ERROR:
+		dev_err(psp->adev->dev, "Secure display: Failed to Read CRC");
+		break;
+	default:
+		dev_err(psp->adev->dev, "Secure display: Failed to parse status: %d\n", status);
+	}
+}
+
+void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct securedisplay_cmd **cmd,
+	enum ta_securedisplay_command command_id)
+{
+	*cmd = (struct securedisplay_cmd *)psp->securedisplay_context.securedisplay_shared_buf;
+	memset(*cmd, 0, sizeof(struct securedisplay_cmd));
+	(*cmd)->status = TA_SECUREDISPLAY_STATUS__GENERIC_FAILURE;
+	(*cmd)->cmd_id = command_id;
+}
+
+static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __user *buf,
+		size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private;
+	struct psp_context *psp = &adev->psp;
+	struct securedisplay_cmd *securedisplay_cmd;
+	struct drm_device *dev = adev_to_drm(adev);
+	uint32_t phy_id;
+	uint32_t op;
+	int i;
+	char str[64];
+	char i2c_output[256];
+	int ret;
+
+	if (*pos || size > sizeof(str) - 1)
+		return -EINVAL;
+
+	memset(str,  0, sizeof(str));
+	ret = copy_from_user(str, buf, size);
+	if (ret)
+		return -EFAULT;
+
+	ret = pm_runtime_get_sync(dev->dev);
+	if (ret < 0) {
+		pm_runtime_put_autosuspend(dev->dev);
+		return ret;
+	}
+
+	if (size < 3)
+		sscanf(str, "%u ", &op);
+	else
+		sscanf(str, "%u %u", &op, &phy_id);
+
+	switch (op) {
+	case 1:
+		psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
+			TA_SECUREDISPLAY_COMMAND__QUERY_TA);
+		ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__QUERY_TA);
+		if (!ret) {
+			if (securedisplay_cmd->status == TA_SECUREDISPLAY_STATUS__SUCCESS)
+				dev_info(adev->dev, "SECUREDISPLAY: query securedisplay TA ret is 0x%X\n",
+					securedisplay_cmd->securedisplay_out_message.query_ta.query_cmd_ret);
+			else
+				psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
+		}
+		break;
+	case 2:
+		psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
+			TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
+		securedisplay_cmd->securedisplay_in_message.send_roi_crc.phy_id = phy_id;
+		ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
+		if (!ret) {
+			if (securedisplay_cmd->status == TA_SECUREDISPLAY_STATUS__SUCCESS) {
+				memset(i2c_output,  0, sizeof(i2c_output));
+				for (i = 0; i < TA_SECUREDISPLAY_I2C_BUFFER_SIZE; i++)
+					sprintf(i2c_output, "%s 0x%X", i2c_output,
+						securedisplay_cmd->securedisplay_out_message.send_roi_crc.i2c_buf[i]);
+				dev_info(adev->dev, "SECUREDISPLAY: I2C buffer out put is :%s\n", i2c_output);
+			} else {
+				psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
+			}
+		}
+		break;
+	default:
+		dev_err(adev->dev, "Invalid input: %s\n", str);
+	}
+
+	pm_runtime_mark_last_busy(dev->dev);
+	pm_runtime_put_autosuspend(dev->dev);
+
+	return size;
+}
+
+static const struct file_operations amdgpu_securedisplay_debugfs_ops = {
+	.owner = THIS_MODULE,
+	.read = NULL,
+	.write = amdgpu_securedisplay_debugfs_write,
+	.llseek = default_llseek
+};
+
+void amdgpu_securedisplay_debugfs_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+
+	if (!adev->psp.securedisplay_context.securedisplay_initialized)
+		return;
+
+	debugfs_create_file("securedisplay_test", S_IWUSR, adev_to_drm(adev)->primary->debugfs_root,
+				adev, &amdgpu_securedisplay_debugfs_ops);
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h
new file mode 100644
index 000000000000..fe98574748f4
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h
@@ -0,0 +1,36 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ *
+ */
+#ifndef _AMDGPU_SECUREDISPLAY_H
+#define _AMDGPU_SECUREDISPLAY_H
+
+#include "amdgpu.h"
+#include "ta_secureDisplay_if.h"
+
+void amdgpu_securedisplay_debugfs_init(struct amdgpu_device *adev);
+void psp_securedisplay_parse_resp_status(struct psp_context *psp,
+		enum ta_securedisplay_status status);
+void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct securedisplay_cmd **cmd,
+		enum ta_securedisplay_command command_id);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
index f80716179968..792d20261846 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
@@ -21,7 +21,7 @@
  *
  */
 
-#if !defined(_AMDGPU_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#if !defined(_AMDGPU_TRACE_H_) || defined(TRACE_HEADER_MULTI_READ)
 #define _AMDGPU_TRACE_H_
 
 #include <linux/stringify.h>
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
index 0e43b46d3ab5..46449e70348b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
@@ -122,6 +122,9 @@ struct ta_firmware_header_v1_0 {
 	uint32_t ta_dtm_ucode_version;
 	uint32_t ta_dtm_offset_bytes;
 	uint32_t ta_dtm_size_bytes;
+	uint32_t ta_securedisplay_ucode_version;
+	uint32_t ta_securedisplay_offset_bytes;
+	uint32_t ta_securedisplay_size_bytes;
 };
 
 enum ta_fw_type {
@@ -132,6 +135,7 @@ enum ta_fw_type {
 	TA_FW_TYPE_PSP_HDCP,
 	TA_FW_TYPE_PSP_DTM,
 	TA_FW_TYPE_PSP_RAP,
+	TA_FW_TYPE_PSP_SECUREDISPLAY,
 };
 
 struct ta_fw_bin_desc {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 8b989670ed66..e2ed4689118a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -1170,7 +1170,7 @@ int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
 	int r, i;
 
 	r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &bo, NULL, (void **)&msg);
 	if (r)
 		return r;
@@ -1202,7 +1202,7 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
 	int r, i;
 
 	r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &bo, NULL, (void **)&msg);
 	if (r)
 		return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 2d51b7694d1f..5da04d45b637 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -560,10 +560,14 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev)
 static void amdgpu_virt_update_vf2pf_work_item(struct work_struct *work)
 {
 	struct amdgpu_device *adev = container_of(work, struct amdgpu_device, virt.vf2pf_work.work);
+	int ret;
 
-	amdgpu_virt_read_pf2vf_data(adev);
+	ret = amdgpu_virt_read_pf2vf_data(adev);
+	if (ret)
+		goto out;
 	amdgpu_virt_write_vf2pf_data(adev);
 
+out:
 	schedule_delayed_work(&(adev->virt.vf2pf_work), adev->virt.vf2pf_update_interval_ms);
 }
 
@@ -571,8 +575,8 @@ void amdgpu_virt_fini_data_exchange(struct amdgpu_device *adev)
 {
 	if (adev->virt.vf2pf_update_interval_ms != 0) {
 		DRM_INFO("clean up the vf2pf work item\n");
-		flush_delayed_work(&adev->virt.vf2pf_work);
 		cancel_delayed_work_sync(&adev->virt.vf2pf_work);
+		adev->virt.vf2pf_update_interval_ms = 0;
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index d2de2a720a3d..c89b66bb70e2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -473,6 +473,9 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man,
 	for (i = 0; pages_left >= pages_per_node; ++i) {
 		unsigned long pages = rounddown_pow_of_two(pages_left);
 
+		/* Limit maximum size to 2GB due to SG table limitations */
+		pages = min(pages, (2UL << (30 - PAGE_SHIFT)));
+
 		r = drm_mm_insert_node_in_range(mm, &nodes[i], pages,
 						pages_per_node, 0,
 						place->fpfn, lpfn,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index 541ef6be390f..659b385b27b5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -324,7 +324,7 @@ static void amdgpu_xgmi_sysfs_rem_dev_info(struct amdgpu_device *adev,
 
 struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev)
 {
-	struct amdgpu_hive_info *hive = NULL, *tmp = NULL;
+	struct amdgpu_hive_info *hive = NULL;
 	int ret;
 
 	if (!adev->gmc.xgmi.hive_id)
@@ -337,11 +337,9 @@ struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev)
 
 	mutex_lock(&xgmi_mutex);
 
-	if (!list_empty(&xgmi_hive_list)) {
-		list_for_each_entry_safe(hive, tmp, &xgmi_hive_list, node)  {
-			if (hive->hive_id == adev->gmc.xgmi.hive_id)
-				goto pro_end;
-		}
+	list_for_each_entry(hive, &xgmi_hive_list, node)  {
+		if (hive->hive_id == adev->gmc.xgmi.hive_id)
+			goto pro_end;
 	}
 
 	hive = kzalloc(sizeof(*hive), GFP_KERNEL);
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c
index 921a69abda55..5b90efd6f6d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c
@@ -27,7 +27,6 @@
 #include "athub/athub_2_0_0_offset.h"
 #include "athub/athub_2_0_0_sh_mask.h"
 #include "athub/athub_2_0_0_default.h"
-#include "navi10_enum.h"
 
 #include "soc15_common.h"
 
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
index 66c183ddd43e..7b1b18350bf9 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
@@ -26,7 +26,6 @@
 
 #include "athub/athub_2_1_0_offset.h"
 #include "athub/athub_2_1_0_sh_mask.h"
-#include "navi10_enum.h"
 
 #include "soc15_common.h"
 
diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c
index 13737b317f7c..4d6832cc7fb0 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik.c
@@ -1251,13 +1251,22 @@ static void kv_restore_regs_for_reset(struct amdgpu_device *adev,
 	WREG32(mmGMCON_RENG_EXECUTE, save->gmcon_reng_execute);
 }
 
-static int cik_gpu_pci_config_reset(struct amdgpu_device *adev)
+/**
+ * cik_asic_pci_config_reset - soft reset GPU
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Use PCI Config method to reset the GPU.
+ *
+ * Returns 0 for success.
+ */
+static int cik_asic_pci_config_reset(struct amdgpu_device *adev)
 {
 	struct kv_reset_save_regs kv_save = { 0 };
 	u32 i;
 	int r = -EINVAL;
 
-	dev_info(adev->dev, "GPU pci config reset\n");
+	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
 
 	if (adev->flags & AMD_IS_APU)
 		kv_save_regs_for_reset(adev, &kv_save);
@@ -1285,26 +1294,6 @@ static int cik_gpu_pci_config_reset(struct amdgpu_device *adev)
 	if (adev->flags & AMD_IS_APU)
 		kv_restore_regs_for_reset(adev, &kv_save);
 
-	return r;
-}
-
-/**
- * cik_asic_pci_config_reset - soft reset GPU
- *
- * @adev: amdgpu_device pointer
- *
- * Use PCI Config method to reset the GPU.
- *
- * Returns 0 for success.
- */
-static int cik_asic_pci_config_reset(struct amdgpu_device *adev)
-{
-	int r;
-
-	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
-
-	r = cik_gpu_pci_config_reset(adev);
-
 	amdgpu_atombios_scratch_regs_engine_hung(adev, false);
 
 	return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c
index da37f8a900af..307c01301c87 100644
--- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c
@@ -194,19 +194,30 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
-	if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) {
-		wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
-		/* When a ring buffer overflow happen start parsing interrupt
-		 * from the last not overwritten vector (wptr + 16). Hopefully
-		 * this should allow us to catchup.
-		 */
-		dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
-			wptr, ih->rptr, (wptr + 16) & ih->ptr_mask);
-		ih->rptr = (wptr + 16) & ih->ptr_mask;
-		tmp = RREG32(mmIH_RB_CNTL);
-		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
-		WREG32(mmIH_RB_CNTL, tmp);
-	}
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	/* Double check that the overflow wasn't already cleared. */
+	wptr = RREG32(mmIH_RB_WPTR);
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 16). Hopefully
+	 * this should allow us to catchup.
+	 */
+	dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+		wptr, ih->rptr, (wptr + 16) & ih->ptr_mask);
+	ih->rptr = (wptr + 16) & ih->ptr_mask;
+	tmp = RREG32(mmIH_RB_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32(mmIH_RB_CNTL, tmp);
+
+
+out:
 	return (wptr & ih->ptr_mask);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c
index ffcc64ec6473..9810af712cc0 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c
@@ -294,7 +294,7 @@ static int dce_virtual_get_modes(struct drm_connector *connector)
 	static const struct mode_size {
 		int w;
 		int h;
-	} common_modes[21] = {
+	} common_modes[] = {
 		{ 640,  480},
 		{ 720,  480},
 		{ 800,  600},
@@ -312,13 +312,14 @@ static int dce_virtual_get_modes(struct drm_connector *connector)
 		{1600, 1200},
 		{1920, 1080},
 		{1920, 1200},
+		{2560, 1440},
 		{4096, 3112},
 		{3656, 2664},
 		{3840, 2160},
 		{4096, 2160},
 	};
 
-	for (i = 0; i < 21; i++) {
+	for (i = 0; i < ARRAY_SIZE(common_modes); i++) {
 		mode = drm_cvt_mode(dev, common_modes[i].w, common_modes[i].h, 60, false, false, false);
 		drm_mode_probed_add(connector, mode);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index ba1086784525..45d1172b7bff 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -38,7 +38,6 @@
 #include "smuio/smuio_11_0_0_offset.h"
 #include "smuio/smuio_11_0_0_sh_mask.h"
 #include "navi10_enum.h"
-#include "hdp/hdp_5_0_0_offset.h"
 #include "ivsrcid/gfx/irqsrcs_gfx_10_1.h"
 
 #include "soc15.h"
@@ -71,6 +70,11 @@
 #define GB_ADDR_CONFIG__NUM_PKRS__SHIFT                                                                       0x8
 #define GB_ADDR_CONFIG__NUM_PKRS_MASK                                                                         0x00000700L
 
+#define mmCGTS_TCC_DISABLE_gc_10_3                 0x5006
+#define mmCGTS_TCC_DISABLE_gc_10_3_BASE_IDX        1
+#define mmCGTS_USER_TCC_DISABLE_gc_10_3            0x5007
+#define mmCGTS_USER_TCC_DISABLE_gc_10_3_BASE_IDX   1
+
 #define mmCP_MEC_CNTL_Sienna_Cichlid                      0x0f55
 #define mmCP_MEC_CNTL_Sienna_Cichlid_BASE_IDX             0
 #define mmRLC_SAFE_MODE_Sienna_Cichlid			0x4ca0
@@ -99,6 +103,10 @@
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid			0x1580
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid_BASE_IDX	0
 
+#define mmGOLDEN_TSC_COUNT_UPPER_Vangogh                0x0025
+#define mmGOLDEN_TSC_COUNT_UPPER_Vangogh_BASE_IDX       1
+#define mmGOLDEN_TSC_COUNT_LOWER_Vangogh                0x0026
+#define mmGOLDEN_TSC_COUNT_LOWER_Vangogh_BASE_IDX       1
 #define mmSPI_CONFIG_CNTL_1_Vangogh		 0x2441
 #define mmSPI_CONFIG_CNTL_1_Vangogh_BASE_IDX	 1
 #define mmVGT_TF_MEMORY_BASE_HI_Vangogh          0x2261
@@ -115,6 +123,9 @@
 #define mmVGT_ESGS_RING_SIZE_Vangogh_BASE_IDX    1
 #define mmSPI_CONFIG_CNTL_Vangogh                0x2440
 #define mmSPI_CONFIG_CNTL_Vangogh_BASE_IDX       1
+#define mmGCR_GENERAL_CNTL_Vangogh               0x1580
+#define mmGCR_GENERAL_CNTL_Vangogh_BASE_IDX      0
+#define RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh   0x0000FFFFL
 
 #define mmCP_HYP_PFP_UCODE_ADDR			0x5814
 #define mmCP_HYP_PFP_UCODE_ADDR_BASE_IDX	1
@@ -160,6 +171,9 @@
 #define mmGCVM_L2_CGTT_CLK_CTRL_Sienna_Cichlid          0x15db
 #define mmGCVM_L2_CGTT_CLK_CTRL_Sienna_Cichlid_BASE_IDX	0
 
+#define mmGC_THROTTLE_CTRL_Sienna_Cichlid              0x2030
+#define mmGC_THROTTLE_CTRL_Sienna_Cichlid_BASE_IDX     0
+
 MODULE_FIRMWARE("amdgpu/navi10_ce.bin");
 MODULE_FIRMWARE("amdgpu/navi10_pfp.bin");
 MODULE_FIRMWARE("amdgpu/navi10_me.bin");
@@ -3237,7 +3251,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3_vangogh[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_EXCEPTION_CONTROL, 0x7fff0f1f, 0x00b80000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_ADDR_CONFIG, 0x0c1807ff, 0x00000142),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL, 0x1ff1ffff, 0x00000500),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Vangogh, 0x1ff1ffff, 0x00000500),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL1_PIPE_STEER, 0x000000ff, 0x000000e4),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_0, 0x77777777, 0x32103210),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_1, 0x77777777, 0x32103210),
@@ -3324,6 +3338,7 @@ static void gfx_v10_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume);
 static void gfx_v10_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start, bool secure);
 static u32 gfx_v10_3_get_disabled_sa(struct amdgpu_device *adev);
 static void gfx_v10_3_program_pbb_mode(struct amdgpu_device *adev);
+static void gfx_v10_3_set_power_brake_sequence(struct amdgpu_device *adev);
 
 static void gfx10_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue_mask)
 {
@@ -3768,9 +3783,6 @@ static void gfx_v10_0_check_gfxoff_flag(struct amdgpu_device *adev)
 		if (!gfx_v10_0_navi10_gfxoff_should_enable(adev))
 			adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
 		break;
-	case CHIP_VANGOGH:
-		adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
-		break;
 	default:
 		break;
 	}
@@ -4480,8 +4492,7 @@ static int gfx_v10_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
 		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
 		+ ring->pipe;
-	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring->pipe,
-							    ring->queue) ?
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
 			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	r = amdgpu_ring_init(adev, ring, 1024,
@@ -4926,8 +4937,15 @@ static void gfx_v10_0_tcp_harvest(struct amdgpu_device *adev)
 static void gfx_v10_0_get_tcc_info(struct amdgpu_device *adev)
 {
 	/* TCCs are global (not instanced). */
-	uint32_t tcc_disable = RREG32_SOC15(GC, 0, mmCGTS_TCC_DISABLE) |
-			       RREG32_SOC15(GC, 0, mmCGTS_USER_TCC_DISABLE);
+	uint32_t tcc_disable;
+
+	if (adev->asic_type >= CHIP_SIENNA_CICHLID) {
+		tcc_disable = RREG32_SOC15(GC, 0, mmCGTS_TCC_DISABLE_gc_10_3) |
+			      RREG32_SOC15(GC, 0, mmCGTS_USER_TCC_DISABLE_gc_10_3);
+	} else {
+		tcc_disable = RREG32_SOC15(GC, 0, mmCGTS_TCC_DISABLE) |
+			      RREG32_SOC15(GC, 0, mmCGTS_USER_TCC_DISABLE);
+	}
 
 	adev->gfx.config.tcc_disabled_mask =
 		REG_GET_FIELD(tcc_disable, CGTS_TCC_DISABLE, TCC_DISABLE) |
@@ -5691,7 +5709,7 @@ static int gfx_v10_0_cp_gfx_load_pfp_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->nbio.funcs->hdp_flush(adev, NULL);
+		adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_PFP_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
@@ -5769,7 +5787,7 @@ static int gfx_v10_0_cp_gfx_load_ce_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->nbio.funcs->hdp_flush(adev, NULL);
+		adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_CE_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_CE_IC_BASE_CNTL, VMID, 0);
@@ -5846,7 +5864,7 @@ static int gfx_v10_0_cp_gfx_load_me_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->nbio.funcs->hdp_flush(adev, NULL);
+		adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_ME_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
@@ -6215,7 +6233,7 @@ static int gfx_v10_0_cp_compute_load_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->nbio.funcs->hdp_flush(adev, NULL);
+		adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_CPC_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
@@ -6523,8 +6541,7 @@ static void gfx_v10_0_compute_mqd_set_priority(struct amdgpu_ring *ring, struct
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
-		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring->pipe,
-							      ring->queue)) {
+		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring)) {
 			mqd->cp_hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH;
 			mqd->cp_hqd_queue_priority =
 				AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM;
@@ -7192,6 +7209,9 @@ static int gfx_v10_0_hw_init(void *handle)
 	if (adev->asic_type == CHIP_SIENNA_CICHLID)
 		gfx_v10_3_program_pbb_mode(adev);
 
+	if (adev->asic_type >= CHIP_SIENNA_CICHLID)
+		gfx_v10_3_set_power_brake_sequence(adev);
+
 	return r;
 }
 
@@ -7377,8 +7397,16 @@ static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 
 	amdgpu_gfx_off_ctrl(adev, false);
 	mutex_lock(&adev->gfx.gpu_clock_mutex);
-	clock = (uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER) |
-		((uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER) << 32ULL);
+	switch (adev->asic_type) {
+	case CHIP_VANGOGH:
+		clock = (uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh) |
+			((uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh) << 32ULL);
+		break;
+	default:
+		clock = (uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER) |
+			((uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER) << 32ULL);
+		break;
+	}
 	mutex_unlock(&adev->gfx.gpu_clock_mutex);
 	amdgpu_gfx_off_ctrl(adev, true);
 	return clock;
@@ -7812,6 +7840,20 @@ static void gfx_v10_cntl_power_gating(struct amdgpu_device *adev, bool enable)
 		data &= ~RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
 
 	WREG32_SOC15(GC, 0, mmRLC_PG_CNTL, data);
+
+	/*
+	 * CGPG enablement required and the register to program the hysteresis value
+	 * RLC_PG_DELAY_3.CGCG_ACTIVE_BEFORE_CGPG to the desired CGPG hysteresis value
+	 * in refclk count. Note that RLC FW is modified to take 16 bits from
+	 * RLC_PG_DELAY_3[15:0] as the hysteresis instead of just 8 bits.
+	 *
+	 * The recommendation from RLC team is setting RLC_PG_DELAY_3 to 200us(0x4E20)
+	 * as part of CGPG enablement starting point.
+	 */
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && adev->asic_type == CHIP_VANGOGH) {
+		data = 0x4E20 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh;
+		WREG32_SOC15(GC, 0, mmRLC_PG_DELAY_3, data);
+	}
 }
 
 static void gfx_v10_cntl_pg(struct amdgpu_device *adev, bool enable)
@@ -7873,6 +7915,7 @@ static int gfx_v10_0_set_powergating_state(void *handle,
 		break;
 	case CHIP_VANGOGH:
 		gfx_v10_cntl_pg(adev, enable);
+		amdgpu_gfx_off_ctrl(adev, enable);
 		break;
 	default:
 		break;
@@ -9169,6 +9212,31 @@ static void gfx_v10_3_program_pbb_mode(struct amdgpu_device *adev)
 	}
 }
 
+static void gfx_v10_3_set_power_brake_sequence(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX,
+		     (0x1 << GRBM_GFX_INDEX__SA_BROADCAST_WRITES__SHIFT) |
+		     (0x1 << GRBM_GFX_INDEX__INSTANCE_BROADCAST_WRITES__SHIFT) |
+		     (0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT));
+
+	WREG32_SOC15(GC, 0, mmGC_CAC_IND_INDEX, ixPWRBRK_STALL_PATTERN_CTRL);
+	WREG32_SOC15(GC, 0, mmGC_CAC_IND_DATA,
+		     (0x1 << PWRBRK_STALL_PATTERN_CTRL__PWRBRK_STEP_INTERVAL__SHIFT) |
+		     (0x12 << PWRBRK_STALL_PATTERN_CTRL__PWRBRK_BEGIN_STEP__SHIFT) |
+		     (0x13 << PWRBRK_STALL_PATTERN_CTRL__PWRBRK_END_STEP__SHIFT) |
+		     (0xf << PWRBRK_STALL_PATTERN_CTRL__PWRBRK_THROTTLE_PATTERN_BIT_NUMS__SHIFT));
+
+	WREG32_SOC15(GC, 0, mmGC_THROTTLE_CTRL_Sienna_Cichlid,
+		     (0x1 << GC_THROTTLE_CTRL__PWRBRK_STALL_EN__SHIFT) |
+		     (0x1 << GC_THROTTLE_CTRL__PATTERN_MODE__SHIFT) |
+		     (0x5 << GC_THROTTLE_CTRL__RELEASE_STEP_INTERVAL__SHIFT));
+
+	WREG32_SOC15(GC, 0, mmDIDT_IND_INDEX, ixDIDT_SQ_THROTTLE_CTRL);
+
+	WREG32_SOC15(GC, 0, mmDIDT_IND_DATA,
+		     (0x1 << DIDT_SQ_THROTTLE_CTRL__PWRBRK_STALL_EN__SHIFT));
+}
+
 const struct amdgpu_ip_block_version gfx_v10_0_ip_block =
 {
 	.type = AMD_IP_BLOCK_TYPE_GFX,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 37639214cbbb..84d2eaa38101 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -29,6 +29,7 @@
 
 #include "amdgpu.h"
 #include "amdgpu_gfx.h"
+#include "amdgpu_ring.h"
 #include "vi.h"
 #include "vi_structs.h"
 #include "vid.h"
@@ -1923,8 +1924,7 @@ static int gfx_v8_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
 		+ ring->pipe;
 
-	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring->pipe,
-							    ring->queue) ?
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
 			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_RING_PRIO_DEFAULT;
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	r = amdgpu_ring_init(adev, ring, 1024,
@@ -4442,8 +4442,7 @@ static void gfx_v8_0_mqd_set_priority(struct amdgpu_ring *ring, struct vi_mqd *m
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
-		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring->pipe,
-							      ring->queue)) {
+		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring)) {
 			mqd->cp_hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH;
 			mqd->cp_hqd_queue_priority =
 				AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM;
@@ -6847,6 +6846,66 @@ static void gfx_v8_0_emit_mem_sync_compute(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0x0000000A);	/* poll interval */
 }
 
+
+/* mmSPI_WCL_PIPE_PERCENT_CS[0-7]_DEFAULT values are same */
+#define mmSPI_WCL_PIPE_PERCENT_CS_DEFAULT	0x0000007f
+static void gfx_v8_0_emit_wave_limit_cs(struct amdgpu_ring *ring,
+					uint32_t pipe, bool enable)
+{
+	uint32_t val;
+	uint32_t wcl_cs_reg;
+
+	val = enable ? 0x1 : mmSPI_WCL_PIPE_PERCENT_CS_DEFAULT;
+
+	switch (pipe) {
+	case 0:
+		wcl_cs_reg = mmSPI_WCL_PIPE_PERCENT_CS0;
+		break;
+	case 1:
+		wcl_cs_reg = mmSPI_WCL_PIPE_PERCENT_CS1;
+		break;
+	case 2:
+		wcl_cs_reg = mmSPI_WCL_PIPE_PERCENT_CS2;
+		break;
+	case 3:
+		wcl_cs_reg = mmSPI_WCL_PIPE_PERCENT_CS3;
+		break;
+	default:
+		DRM_DEBUG("invalid pipe %d\n", pipe);
+		return;
+	}
+
+	amdgpu_ring_emit_wreg(ring, wcl_cs_reg, val);
+
+}
+
+#define mmSPI_WCL_PIPE_PERCENT_GFX_DEFAULT	0x07ffffff
+static void gfx_v8_0_emit_wave_limit(struct amdgpu_ring *ring, bool enable)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t val;
+	int i;
+
+	/* mmSPI_WCL_PIPE_PERCENT_GFX is 7 bit multiplier register to limit
+	 * number of gfx waves. Setting 5 bit will make sure gfx only gets
+	 * around 25% of gpu resources.
+	 */
+	val = enable ? 0x1f : mmSPI_WCL_PIPE_PERCENT_GFX_DEFAULT;
+	amdgpu_ring_emit_wreg(ring, mmSPI_WCL_PIPE_PERCENT_GFX, val);
+
+	/* Restrict waves for normal/low priority compute queues as well
+	 * to get best QoS for high priority compute jobs.
+	 *
+	 * amdgpu controls only 1st ME(0-3 CS pipes).
+	 */
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		if (i != ring->pipe)
+			gfx_v8_0_emit_wave_limit_cs(ring, i, enable);
+
+	}
+
+}
+
 static const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
 	.name = "gfx_v8_0",
 	.early_init = gfx_v8_0_early_init,
@@ -6930,7 +6989,9 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
 		7 + /* gfx_v8_0_ring_emit_pipeline_sync */
 		VI_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + /* gfx_v8_0_ring_emit_vm_flush */
 		7 + 7 + 7 + /* gfx_v8_0_ring_emit_fence_compute x3 for user fence, vm fence */
-		7, /* gfx_v8_0_emit_mem_sync_compute */
+		7 + /* gfx_v8_0_emit_mem_sync_compute */
+		5 + /* gfx_v8_0_emit_wave_limit for updating mmSPI_WCL_PIPE_PERCENT_GFX register */
+		15, /* for updating 3 mmSPI_WCL_PIPE_PERCENT_CS registers */
 	.emit_ib_size =	7, /* gfx_v8_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v8_0_ring_emit_ib_compute,
 	.emit_fence = gfx_v8_0_ring_emit_fence_compute,
@@ -6944,6 +7005,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_wreg = gfx_v8_0_ring_emit_wreg,
 	.emit_mem_sync = gfx_v8_0_emit_mem_sync_compute,
+	.emit_wave_limit = gfx_v8_0_emit_wave_limit,
 };
 
 static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_kiq = {
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 5f4805e4d04a..65db88bb6cbc 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -38,7 +38,6 @@
 #include "gc/gc_9_0_sh_mask.h"
 
 #include "vega10_enum.h"
-#include "hdp/hdp_4_0_offset.h"
 
 #include "soc15_common.h"
 #include "clearstate_gfx9.h"
@@ -53,6 +52,7 @@
 
 #include "asic_reg/pwr/pwr_10_0_offset.h"
 #include "asic_reg/pwr/pwr_10_0_sh_mask.h"
+#include "asic_reg/gc/gc_9_0_default.h"
 
 #define GFX9_NUM_GFX_RINGS     1
 #define GFX9_MEC_HPD_SIZE 4096
@@ -2228,8 +2228,7 @@ static int gfx_v9_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
 		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
 		+ ring->pipe;
-	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring->pipe,
-							    ring->queue) ?
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
 			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	return amdgpu_ring_init(adev, ring, 1024,
@@ -3391,9 +3390,7 @@ static void gfx_v9_0_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *m
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
-		if (amdgpu_gfx_is_high_priority_compute_queue(adev,
-							      ring->pipe,
-							      ring->queue)) {
+		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring)) {
 			mqd->cp_hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH;
 			mqd->cp_hqd_queue_priority =
 				AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM;
@@ -6671,6 +6668,65 @@ static void gfx_v9_0_emit_mem_sync(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */
 }
 
+static void gfx_v9_0_emit_wave_limit_cs(struct amdgpu_ring *ring,
+					uint32_t pipe, bool enable)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t val;
+	uint32_t wcl_cs_reg;
+
+	/* mmSPI_WCL_PIPE_PERCENT_CS[0-7]_DEFAULT values are same */
+	val = enable ? 0x1 : mmSPI_WCL_PIPE_PERCENT_CS0_DEFAULT;
+
+	switch (pipe) {
+	case 0:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, 0, mmSPI_WCL_PIPE_PERCENT_CS0);
+		break;
+	case 1:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, 0, mmSPI_WCL_PIPE_PERCENT_CS1);
+		break;
+	case 2:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, 0, mmSPI_WCL_PIPE_PERCENT_CS2);
+		break;
+	case 3:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, 0, mmSPI_WCL_PIPE_PERCENT_CS3);
+		break;
+	default:
+		DRM_DEBUG("invalid pipe %d\n", pipe);
+		return;
+	}
+
+	amdgpu_ring_emit_wreg(ring, wcl_cs_reg, val);
+
+}
+static void gfx_v9_0_emit_wave_limit(struct amdgpu_ring *ring, bool enable)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t val;
+	int i;
+
+
+	/* mmSPI_WCL_PIPE_PERCENT_GFX is 7 bit multiplier register to limit
+	 * number of gfx waves. Setting 5 bit will make sure gfx only gets
+	 * around 25% of gpu resources.
+	 */
+	val = enable ? 0x1f : mmSPI_WCL_PIPE_PERCENT_GFX_DEFAULT;
+	amdgpu_ring_emit_wreg(ring,
+			      SOC15_REG_OFFSET(GC, 0, mmSPI_WCL_PIPE_PERCENT_GFX),
+			      val);
+
+	/* Restrict waves for normal/low priority compute queues as well
+	 * to get best QoS for high priority compute jobs.
+	 *
+	 * amdgpu controls only 1st ME(0-3 CS pipes).
+	 */
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		if (i != ring->pipe)
+			gfx_v9_0_emit_wave_limit_cs(ring, i, enable);
+
+	}
+}
+
 static const struct amd_ip_funcs gfx_v9_0_ip_funcs = {
 	.name = "gfx_v9_0",
 	.early_init = gfx_v9_0_early_init,
@@ -6760,7 +6816,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
 		2 + /* gfx_v9_0_ring_emit_vm_flush */
 		8 + 8 + 8 + /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */
-		7, /* gfx_v9_0_emit_mem_sync */
+		7 + /* gfx_v9_0_emit_mem_sync */
+		5 + /* gfx_v9_0_emit_wave_limit for updating mmSPI_WCL_PIPE_PERCENT_GFX register */
+		15, /* for updating 3 mmSPI_WCL_PIPE_PERCENT_CS registers */
 	.emit_ib_size =	7, /* gfx_v9_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v9_0_ring_emit_ib_compute,
 	.emit_fence = gfx_v9_0_ring_emit_fence,
@@ -6776,6 +6834,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait,
 	.emit_mem_sync = gfx_v9_0_emit_mem_sync,
+	.emit_wave_limit = gfx_v9_0_emit_wave_limit,
 };
 
 static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 5648c48be77f..3b7c6c31fce1 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -27,8 +27,6 @@
 #include "gmc_v10_0.h"
 #include "umc_v8_7.h"
 
-#include "hdp/hdp_5_0_0_offset.h"
-#include "hdp/hdp_5_0_0_sh_mask.h"
 #include "athub/athub_2_0_0_sh_mask.h"
 #include "athub/athub_2_0_0_offset.h"
 #include "dcn/dcn_2_0_0_offset.h"
@@ -312,7 +310,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	int r;
 
 	/* flush hdp cache */
-	adev->nbio.funcs->hdp_flush(adev, NULL);
+	adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	/* For SRIOV run time, driver shouldn't access the register through MMIO
 	 * Directly use kiq to do the vm invalidation instead
@@ -995,7 +993,6 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev)
 {
 	int r;
 	bool value;
-	u32 tmp;
 
 	if (adev->gart.bo == NULL) {
 		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
@@ -1014,15 +1011,10 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	tmp = RREG32_SOC15(HDP, 0, mmHDP_MISC_CNTL);
-	tmp |= HDP_MISC_CNTL__FLUSH_INVALIDATE_CACHE_MASK;
-	WREG32_SOC15(HDP, 0, mmHDP_MISC_CNTL, tmp);
-
-	tmp = RREG32_SOC15(HDP, 0, mmHDP_HOST_PATH_CNTL);
-	WREG32_SOC15(HDP, 0, mmHDP_HOST_PATH_CNTL, tmp);
+	adev->hdp.funcs->init_registers(adev);
 
 	/* Flush HDP after it is initialized */
-	adev->nbio.funcs->hdp_flush(adev, NULL);
+	adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
 		false : true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index e22268f9dba7..3686e777c76c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -31,8 +31,6 @@
 #include "amdgpu_atomfirmware.h"
 #include "amdgpu_gem.h"
 
-#include "hdp/hdp_4_0_offset.h"
-#include "hdp/hdp_4_0_sh_mask.h"
 #include "gc/gc_9_0_sh_mask.h"
 #include "dce/dce_12_0_offset.h"
 #include "dce/dce_12_0_sh_mask.h"
@@ -241,60 +239,44 @@ static const char *mmhub_client_ids_vega20[][2] = {
 };
 
 static const char *mmhub_client_ids_arcturus[][2] = {
+	[0][0] = "DBGU1",
+	[1][0] = "XDP",
 	[2][0] = "MP1",
-	[3][0] = "MP0",
-	[10][0] = "UTCL2",
-	[13][0] = "OSS",
 	[14][0] = "HDP",
-	[15][0] = "SDMA0",
-	[32+15][0] = "SDMA1",
-	[64+15][0] = "SDMA2",
-	[96+15][0] = "SDMA3",
-	[128+15][0] = "SDMA4",
-	[160+11][0] = "JPEG",
-	[160+12][0] = "VCN",
-	[160+13][0] = "VCNU",
-	[160+15][0] = "SDMA5",
-	[192+10][0] = "UTCL2",
-	[192+11][0] = "JPEG1",
-	[192+12][0] = "VCN1",
-	[192+13][0] = "VCN1U",
-	[192+15][0] = "SDMA6",
-	[224+15][0] = "SDMA7",
+	[171][0] = "JPEG",
+	[172][0] = "VCN",
+	[173][0] = "VCNU",
+	[203][0] = "JPEG1",
+	[204][0] = "VCN1",
+	[205][0] = "VCN1U",
+	[256][0] = "SDMA0",
+	[257][0] = "SDMA1",
+	[258][0] = "SDMA2",
+	[259][0] = "SDMA3",
+	[260][0] = "SDMA4",
+	[261][0] = "SDMA5",
+	[262][0] = "SDMA6",
+	[263][0] = "SDMA7",
+	[384][0] = "OSS",
 	[0][1] = "DBGU1",
 	[1][1] = "XDP",
 	[2][1] = "MP1",
-	[3][1] = "MP0",
-	[13][1] = "OSS",
 	[14][1] = "HDP",
-	[15][1] = "SDMA0",
-	[32+15][1] = "SDMA1",
-	[64+15][1] = "SDMA2",
-	[96+15][1] = "SDMA3",
-	[128+15][1] = "SDMA4",
-	[160+11][1] = "JPEG",
-	[160+12][1] = "VCN",
-	[160+13][1] = "VCNU",
-	[160+15][1] = "SDMA5",
-	[192+11][1] = "JPEG1",
-	[192+12][1] = "VCN1",
-	[192+13][1] = "VCN1U",
-	[192+15][1] = "SDMA6",
-	[224+15][1] = "SDMA7",
-};
-
-static const u32 golden_settings_vega10_hdp[] =
-{
-	0xf64, 0x0fffffff, 0x00000000,
-	0xf65, 0x0fffffff, 0x00000000,
-	0xf66, 0x0fffffff, 0x00000000,
-	0xf67, 0x0fffffff, 0x00000000,
-	0xf68, 0x0fffffff, 0x00000000,
-	0xf6a, 0x0fffffff, 0x00000000,
-	0xf6b, 0x0fffffff, 0x00000000,
-	0xf6c, 0x0fffffff, 0x00000000,
-	0xf6d, 0x0fffffff, 0x00000000,
-	0xf6e, 0x0fffffff, 0x00000000,
+	[171][1] = "JPEG",
+	[172][1] = "VCN",
+	[173][1] = "VCNU",
+	[203][1] = "JPEG1",
+	[204][1] = "VCN1",
+	[205][1] = "VCN1U",
+	[256][1] = "SDMA0",
+	[257][1] = "SDMA1",
+	[258][1] = "SDMA2",
+	[259][1] = "SDMA3",
+	[260][1] = "SDMA4",
+	[261][1] = "SDMA5",
+	[262][1] = "SDMA6",
+	[263][1] = "SDMA7",
+	[384][1] = "OSS",
 };
 
 static const struct soc15_reg_golden golden_settings_mmhub_1_0_0[] =
@@ -1571,7 +1553,6 @@ static int gmc_v9_0_hw_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	bool value;
 	int r, i;
-	u32 tmp;
 
 	/* The sequence of these two function calls matters.*/
 	gmc_v9_0_init_golden_registers(adev);
@@ -1583,31 +1564,13 @@ static int gmc_v9_0_hw_init(void *handle)
 		WREG32_FIELD15(DCE, 0, VGA_RENDER_CONTROL, VGA_VSTATUS_CNTL, 0);
 	}
 
-	amdgpu_device_program_register_sequence(adev,
-						golden_settings_vega10_hdp,
-						ARRAY_SIZE(golden_settings_vega10_hdp));
-
 	if (adev->mmhub.funcs->update_power_gating)
 		adev->mmhub.funcs->update_power_gating(adev, true);
 
-	switch (adev->asic_type) {
-	case CHIP_ARCTURUS:
-		WREG32_FIELD15(HDP, 0, HDP_MMHUB_CNTL, HDP_MMHUB_GCC, 1);
-		break;
-	default:
-		break;
-	}
-
-	WREG32_FIELD15(HDP, 0, HDP_MISC_CNTL, FLUSH_INVALIDATE_CACHE, 1);
-
-	tmp = RREG32_SOC15(HDP, 0, mmHDP_HOST_PATH_CNTL);
-	WREG32_SOC15(HDP, 0, mmHDP_HOST_PATH_CNTL, tmp);
-
-	WREG32_SOC15(HDP, 0, mmHDP_NONSURFACE_BASE, (adev->gmc.vram_start >> 8));
-	WREG32_SOC15(HDP, 0, mmHDP_NONSURFACE_BASE_HI, (adev->gmc.vram_start >> 40));
+	adev->hdp.funcs->init_registers(adev);
 
 	/* After HDP is initialized, flush HDP.*/
-	adev->nbio.funcs->hdp_flush(adev, NULL);
+	adev->hdp.funcs->flush_hdp(adev, NULL);
 
 	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS)
 		value = false;
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
new file mode 100644
index 000000000000..e46621fed5b9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
@@ -0,0 +1,137 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_atombios.h"
+#include "hdp_v4_0.h"
+#include "amdgpu_ras.h"
+
+#include "hdp/hdp_4_0_offset.h"
+#include "hdp/hdp_4_0_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+/* for Vega20 register name change */
+#define mmHDP_MEM_POWER_CTRL    0x00d4
+#define HDP_MEM_POWER_CTRL__IPH_MEM_POWER_CTRL_EN_MASK  0x00000001L
+#define HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK    0x00000002L
+#define HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK   0x00010000L
+#define HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK     0x00020000L
+#define mmHDP_MEM_POWER_CTRL_BASE_IDX   0
+
+static void hdp_v4_0_flush_hdp(struct amdgpu_device *adev,
+				struct amdgpu_ring *ring)
+{
+	if (!ring || !ring->funcs->emit_wreg)
+		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+	else
+		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+}
+
+static void hdp_v4_0_invalidate_hdp(struct amdgpu_device *adev,
+				    struct amdgpu_ring *ring)
+{
+	if (!ring || !ring->funcs->emit_wreg)
+		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
+	else
+		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
+			HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
+}
+
+static void hdp_v4_0_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__HDP))
+		return;
+	/*read back hdp ras counter to reset it to 0 */
+	RREG32_SOC15(HDP, 0, mmHDP_EDC_CNT);
+}
+
+static void hdp_v4_0_update_clock_gating(struct amdgpu_device *adev,
+					 bool enable)
+{
+	uint32_t def, data;
+
+	if (adev->asic_type == CHIP_VEGA10 ||
+	    adev->asic_type == CHIP_VEGA12 ||
+	    adev->asic_type == CHIP_RAVEN) {
+		def = data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS));
+
+		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS))
+			data |= HDP_MEM_POWER_LS__LS_ENABLE_MASK;
+		else
+			data &= ~HDP_MEM_POWER_LS__LS_ENABLE_MASK;
+
+		if (def != data)
+			WREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS), data);
+	} else {
+		def = data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_CTRL));
+
+		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS))
+			data |= HDP_MEM_POWER_CTRL__IPH_MEM_POWER_CTRL_EN_MASK |
+				HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK |
+				HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK |
+				HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK;
+		else
+			data &= ~(HDP_MEM_POWER_CTRL__IPH_MEM_POWER_CTRL_EN_MASK |
+				  HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK |
+				  HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK |
+				  HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK);
+
+		if (def != data)
+			WREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_CTRL), data);
+	}
+}
+
+static void hdp_v4_0_get_clockgating_state(struct amdgpu_device *adev,
+					    u32 *flags)
+{
+	int data;
+
+	/* AMD_CG_SUPPORT_HDP_LS */
+	data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS));
+	if (data & HDP_MEM_POWER_LS__LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_LS;
+}
+
+static void hdp_v4_0_init_registers(struct amdgpu_device *adev)
+{
+	switch (adev->asic_type) {
+	case CHIP_ARCTURUS:
+		WREG32_FIELD15(HDP, 0, HDP_MMHUB_CNTL, HDP_MMHUB_GCC, 1);
+		break;
+	default:
+		break;
+	}
+
+	WREG32_FIELD15(HDP, 0, HDP_MISC_CNTL, FLUSH_INVALIDATE_CACHE, 1);
+
+	WREG32_SOC15(HDP, 0, mmHDP_NONSURFACE_BASE, (adev->gmc.vram_start >> 8));
+	WREG32_SOC15(HDP, 0, mmHDP_NONSURFACE_BASE_HI, (adev->gmc.vram_start >> 40));
+}
+
+const struct amdgpu_hdp_funcs hdp_v4_0_funcs = {
+	.flush_hdp = hdp_v4_0_flush_hdp,
+	.invalidate_hdp = hdp_v4_0_invalidate_hdp,
+	.reset_ras_error_count = hdp_v4_0_reset_ras_error_count,
+	.update_clock_gating = hdp_v4_0_update_clock_gating,
+	.get_clock_gating_state = hdp_v4_0_get_clockgating_state,
+	.init_registers = hdp_v4_0_init_registers,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h
new file mode 100644
index 000000000000..d1e6399e8c46
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __HDP_V4_0_H__
+#define __HDP_V4_0_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_hdp_funcs hdp_v4_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
new file mode 100644
index 000000000000..7a15e669b68d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
@@ -0,0 +1,212 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_atombios.h"
+#include "hdp_v5_0.h"
+
+#include "hdp/hdp_5_0_0_offset.h"
+#include "hdp/hdp_5_0_0_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void hdp_v5_0_flush_hdp(struct amdgpu_device *adev,
+				struct amdgpu_ring *ring)
+{
+	if (!ring || !ring->funcs->emit_wreg)
+		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+	else
+		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+}
+
+static void hdp_v5_0_invalidate_hdp(struct amdgpu_device *adev,
+				    struct amdgpu_ring *ring)
+{
+	if (!ring || !ring->funcs->emit_wreg) {
+		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
+	} else {
+		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
+					HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
+	}
+}
+
+static void hdp_v5_0_update_mem_power_gating(struct amdgpu_device *adev,
+					  bool enable)
+{
+	uint32_t hdp_clk_cntl, hdp_clk_cntl1;
+	uint32_t hdp_mem_pwr_cntl;
+
+	if (!(adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS |
+				AMD_CG_SUPPORT_HDP_DS |
+				AMD_CG_SUPPORT_HDP_SD)))
+		return;
+
+	hdp_clk_cntl = hdp_clk_cntl1 = RREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL);
+	hdp_mem_pwr_cntl = RREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL);
+
+	/* Before doing clock/power mode switch,
+	 * forced on IPH & RC clock */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     IPH_MEM_CLK_SOFT_OVERRIDE, 1);
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 1);
+	WREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL, hdp_clk_cntl);
+
+	/* HDP 5.0 doesn't support dynamic power mode switch,
+	 * disable clock and power gating before any changing */
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 IPH_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 IPH_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 IPH_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 IPH_MEM_POWER_SD_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_SD_EN, 0);
+	WREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+
+	/* only one clock gating mode (LS/DS/SD) can be enabled */
+	if (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS) {
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+						 HDP_MEM_POWER_CTRL,
+						 IPH_MEM_POWER_LS_EN, enable);
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+						 HDP_MEM_POWER_CTRL,
+						 RC_MEM_POWER_LS_EN, enable);
+	} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_DS) {
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+						 HDP_MEM_POWER_CTRL,
+						 IPH_MEM_POWER_DS_EN, enable);
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+						 HDP_MEM_POWER_CTRL,
+						 RC_MEM_POWER_DS_EN, enable);
+	} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_SD) {
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+						 HDP_MEM_POWER_CTRL,
+						 IPH_MEM_POWER_SD_EN, enable);
+		/* RC should not use shut down mode, fallback to ds */
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+						 HDP_MEM_POWER_CTRL,
+						 RC_MEM_POWER_DS_EN, enable);
+	}
+
+	/* confirmed that IPH_MEM_POWER_CTRL_EN and RC_MEM_POWER_CTRL_EN have to
+	 * be set for SRAM LS/DS/SD */
+	if (adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_HDP_DS |
+			      AMD_CG_SUPPORT_HDP_SD)) {
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+						 IPH_MEM_POWER_CTRL_EN, 1);
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+						 RC_MEM_POWER_CTRL_EN, 1);
+	}
+
+	WREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+
+	/* restore IPH & RC clock override after clock/power mode changing */
+	WREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL, hdp_clk_cntl1);
+}
+
+static void hdp_v5_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						      bool enable)
+{
+	uint32_t hdp_clk_cntl;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_HDP_MGCG))
+		return;
+
+	hdp_clk_cntl = RREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL);
+
+	if (enable) {
+		hdp_clk_cntl &=
+			~(uint32_t)
+			(HDP_CLK_CNTL__IPH_MEM_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK);
+	} else {
+		hdp_clk_cntl |= HDP_CLK_CNTL__IPH_MEM_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK;
+	}
+
+	WREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL, hdp_clk_cntl);
+}
+
+static void hdp_v5_0_update_clock_gating(struct amdgpu_device *adev,
+					      bool enable)
+{
+	hdp_v5_0_update_mem_power_gating(adev, enable);
+	hdp_v5_0_update_medium_grain_clock_gating(adev, enable);
+}
+
+static void hdp_v5_0_get_clockgating_state(struct amdgpu_device *adev,
+					    u32 *flags)
+{
+	uint32_t tmp;
+
+	/* AMD_CG_SUPPORT_HDP_MGCG */
+	tmp = RREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL);
+	if (!(tmp & (HDP_CLK_CNTL__IPH_MEM_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK)))
+		*flags |= AMD_CG_SUPPORT_HDP_MGCG;
+
+	/* AMD_CG_SUPPORT_HDP_LS/DS/SD */
+	tmp = RREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL);
+	if (tmp & HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_LS;
+	else if (tmp & HDP_MEM_POWER_CTRL__IPH_MEM_POWER_DS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_DS;
+	else if (tmp & HDP_MEM_POWER_CTRL__IPH_MEM_POWER_SD_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_SD;
+}
+
+static void hdp_v5_0_init_registers(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(HDP, 0, mmHDP_MISC_CNTL);
+	tmp |= HDP_MISC_CNTL__FLUSH_INVALIDATE_CACHE_MASK;
+	WREG32_SOC15(HDP, 0, mmHDP_MISC_CNTL, tmp);
+}
+
+const struct amdgpu_hdp_funcs hdp_v5_0_funcs = {
+	.flush_hdp = hdp_v5_0_flush_hdp,
+	.invalidate_hdp = hdp_v5_0_invalidate_hdp,
+	.update_clock_gating = hdp_v5_0_update_clock_gating,
+	.get_clock_gating_state = hdp_v5_0_get_clockgating_state,
+	.init_registers = hdp_v5_0_init_registers,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.h b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.h
new file mode 100644
index 000000000000..2d5ec2b419f3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __HDP_V5_0_H__
+#define __HDP_V5_0_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_hdp_funcs hdp_v5_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
index 37d8b6ca4dab..cc957471f31e 100644
--- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
@@ -194,19 +194,29 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
-	if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) {
-		wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
-		/* When a ring buffer overflow happen start parsing interrupt
-		 * from the last not overwritten vector (wptr + 16). Hopefully
-		 * this should allow us to catchup.
-		 */
-		dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
-			 wptr, ih->rptr, (wptr + 16) & ih->ptr_mask);
-		ih->rptr = (wptr + 16) & ih->ptr_mask;
-		tmp = RREG32(mmIH_RB_CNTL);
-		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
-		WREG32(mmIH_RB_CNTL, tmp);
-	}
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	/* Double check that the overflow wasn't already cleared. */
+	wptr = RREG32(mmIH_RB_WPTR);
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 16). Hopefully
+	 * this should allow us to catchup.
+	 */
+	dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+		wptr, ih->rptr, (wptr + 16) & ih->ptr_mask);
+	ih->rptr = (wptr + 16) & ih->ptr_mask;
+	tmp = RREG32(mmIH_RB_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32(mmIH_RB_CNTL, tmp);
+
+
+out:
 	return (wptr & ih->ptr_mask);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
index b72c8e4ca36b..ab9be5ad5a5f 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
@@ -310,7 +310,7 @@ static void mmhub_v2_3_setup_vmid_config(struct amdgpu_device *adev)
 		/* Send no-retry XNACK on fault to suppress VM fault storm. */
 		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
 				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
-				    !amdgpu_noretry);
+				    !adev->gmc.noretry);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL,
 				    i * hub->ctx_distance, tmp);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
@@ -491,12 +491,11 @@ mmhub_v2_3_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 {
 	uint32_t def, data, def1, data1;
 
-	def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
+	def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_CGTT_CLK_CTRL);
 	def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2);
 
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)) {
-		data |= MM_ATC_L2_MISC_CG__ENABLE_MASK;
-
+		data &= ~MM_ATC_L2_CGTT_CLK_CTRL__SOFT_OVERRIDE_MASK;
 		data1 &= ~(DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
 		           DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
 		           DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
@@ -505,8 +504,7 @@ mmhub_v2_3_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 		           DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
 
 	} else {
-		data &= ~MM_ATC_L2_MISC_CG__ENABLE_MASK;
-
+		data |= MM_ATC_L2_CGTT_CLK_CTRL__SOFT_OVERRIDE_MASK;
 		data1 |= (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
 			  DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
 			  DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
@@ -516,7 +514,7 @@ mmhub_v2_3_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 	}
 
 	if (def != data)
-		WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG, data);
+		WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_CGTT_CLK_CTRL, data);
 	if (def1 != data1)
 		WREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2, data1);
 }
@@ -525,17 +523,44 @@ static void
 mmhub_v2_3_update_medium_grain_light_sleep(struct amdgpu_device *adev,
 					   bool enable)
 {
-	uint32_t def, data;
-
-	def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
-
-	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_LS))
-		data |= MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
-	else
-		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+	uint32_t def, data, def1, data1, def2, data2;
+
+	def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_CGTT_CLK_CTRL);
+	def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_WR_CGTT_CLK_CTRL);
+	def2 = data2 = RREG32_SOC15(MMHUB, 0, mmDAGB0_RD_CGTT_CLK_CTRL);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_LS)) {
+		data &= ~MM_ATC_L2_CGTT_CLK_CTRL__MGLS_OVERRIDE_MASK;
+		data1 &= ~(DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_WRITE_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_READ_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_RETURN_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_REGISTER_MASK);
+		data2 &= ~(DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_WRITE_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_READ_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_RETURN_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_REGISTER_MASK);
+	} else {
+		data |= MM_ATC_L2_CGTT_CLK_CTRL__MGLS_OVERRIDE_MASK;
+		data1 |= (DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_WRITE_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_READ_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_RETURN_MASK |
+			DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_REGISTER_MASK);
+		data2 |= (DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_WRITE_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_READ_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_RETURN_MASK |
+			DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_REGISTER_MASK);
+	}
 
 	if (def != data)
-		WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG, data);
+		WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_CGTT_CLK_CTRL, data);
+	if (def1 != data1)
+		WREG32_SOC15(MMHUB, 0, mmDAGB0_WR_CGTT_CLK_CTRL, data1);
+	if (def2 != data2)
+		WREG32_SOC15(MMHUB, 0, mmDAGB0_RD_CGTT_CLK_CTRL, data2);
 }
 
 static int mmhub_v2_3_set_clockgating(struct amdgpu_device *adev,
@@ -554,26 +579,39 @@ static int mmhub_v2_3_set_clockgating(struct amdgpu_device *adev,
 
 static void mmhub_v2_3_get_clockgating(struct amdgpu_device *adev, u32 *flags)
 {
-	int data, data1;
+	int data, data1, data2, data3;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
 
-	data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
-	data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2);
+	data = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2);
+	data1  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_CGTT_CLK_CTRL);
+	data2 = RREG32_SOC15(MMHUB, 0, mmDAGB0_WR_CGTT_CLK_CTRL);
+	data3 = RREG32_SOC15(MMHUB, 0, mmDAGB0_RD_CGTT_CLK_CTRL);
 
 	/* AMD_CG_SUPPORT_MC_MGCG */
-	if ((data & MM_ATC_L2_MISC_CG__ENABLE_MASK) &&
-	    !(data1 & (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
+	if (!(data & (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
 		       DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
 		       DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
 		       DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
 		       DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
-		       DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK)))
-		*flags |= AMD_CG_SUPPORT_MC_MGCG;
+		       DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK))
+		&& !(data1 & MM_ATC_L2_CGTT_CLK_CTRL__SOFT_OVERRIDE_MASK)) {
+			*flags |= AMD_CG_SUPPORT_MC_MGCG;
+	}
 
 	/* AMD_CG_SUPPORT_MC_LS */
-	if (data & MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
+	if (!(data1 & MM_ATC_L2_CGTT_CLK_CTRL__MGLS_OVERRIDE_MASK)
+		&& !(data2 & (DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_MASK |
+				DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_WRITE_MASK |
+				DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_READ_MASK |
+				DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_RETURN_MASK |
+				DAGB0_WR_CGTT_CLK_CTRL__LS_OVERRIDE_REGISTER_MASK))
+		&& !(data3 & (DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_MASK |
+				DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_WRITE_MASK |
+				DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_READ_MASK |
+				DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_RETURN_MASK |
+				DAGB0_RD_CGTT_CLK_CTRL__LS_OVERRIDE_REGISTER_MASK)))
 		*flags |= AMD_CG_SUPPORT_MC_LS;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index 7767ccca526b..3ee481557fc9 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -255,6 +255,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
 	if (!down_read_trylock(&adev->reset_sem))
 		return;
 
+	amdgpu_virt_fini_data_exchange(adev);
 	atomic_set(&adev->in_gpu_reset, 1);
 
 	do {
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
index dd5c1e6ce009..48e588d3c409 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
@@ -276,6 +276,7 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
 	if (!down_read_trylock(&adev->reset_sem))
 		return;
 
+	amdgpu_virt_fini_data_exchange(adev);
 	atomic_set(&adev->in_gpu_reset, 1);
 
 	do {
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
index 7ba229e43799..f4e4040bbd25 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
@@ -40,6 +40,53 @@
 static void navi10_ih_set_interrupt_funcs(struct amdgpu_device *adev);
 
 /**
+ * navi10_ih_init_register_offset - Initialize register offset for ih rings
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize register offset ih rings (NAVI10).
+ */
+static void navi10_ih_init_register_offset(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (adev->irq.ih.ring_size) {
+		ih_regs = &adev->irq.ih.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR);
+		ih_regs->ih_rb_wptr_addr_lo = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO);
+		ih_regs->ih_rb_wptr_addr_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL;
+	}
+
+	if (adev->irq.ih1.ring_size) {
+		ih_regs = &adev->irq.ih1.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_RING1);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI_RING1);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING1);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING1);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING1);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING1);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING1;
+	}
+
+	if (adev->irq.ih2.ring_size) {
+		ih_regs = &adev->irq.ih2.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_RING2);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI_RING2);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING2);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING2);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING2);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING2);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING2;
+	}
+}
+
+/**
  * force_update_wptr_for_self_int - Force update the wptr for self interrupt
  *
  * @adev: amdgpu_device pointer
@@ -82,133 +129,66 @@ force_update_wptr_for_self_int(struct amdgpu_device *adev,
 }
 
 /**
- * navi10_ih_enable_interrupts - Enable the interrupt ring buffer
+ * navi10_ih_toggle_ring_interrupts - toggle the interrupt ring buffer
  *
  * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointet
+ * @enable: true - enable the interrupts, false - disable the interrupts
  *
- * Enable the interrupt ring buffer (NAVI10).
+ * Toggle the interrupt ring buffer (NAVI10)
  */
-static void navi10_ih_enable_interrupts(struct amdgpu_device *adev)
+static int navi10_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
+					    struct amdgpu_ih_ring *ih,
+					    bool enable)
 {
-	u32 ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL);
-
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_ENABLE, 1);
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 1);
-	if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL, ih_rb_cntl)) {
-			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
-			return;
-		}
-	} else {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
-	}
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
 
-	adev->irq.ih.enabled = true;
+	ih_regs = &ih->ih_regs;
 
-	if (adev->irq.ih1.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
-					   RB_ENABLE, 1);
-		if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING1 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl);
-		}
-		adev->irq.ih1.enabled = true;
-	}
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+	/* enable_intr field is only valid in ring0 */
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
+	WREG32(ih_regs->ih_rb_cntl, tmp);
 
-	if (adev->irq.ih2.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING2,
-					   RB_ENABLE, 1);
-		if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING2,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING2 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl);
-		}
-		adev->irq.ih2.enabled = true;
+	if (enable) {
+		ih->enabled = true;
+	} else {
+		/* set rptr, wptr to 0 */
+		WREG32(ih_regs->ih_rb_rptr, 0);
+		WREG32(ih_regs->ih_rb_wptr, 0);
+		ih->enabled = false;
+		ih->rptr = 0;
 	}
 
-	if (adev->irq.ih_soft.ring_size)
-		adev->irq.ih_soft.enabled = true;
+	return 0;
 }
 
 /**
- * navi10_ih_disable_interrupts - Disable the interrupt ring buffer
+ * navi10_ih_toggle_interrupts - Toggle all the available interrupt ring buffers
  *
  * @adev: amdgpu_device pointer
+ * @enable: enable or disable interrupt ring buffers
  *
- * Disable the interrupt ring buffer (NAVI10).
+ * Toggle all the available interrupt ring buffers (NAVI10).
  */
-static void navi10_ih_disable_interrupts(struct amdgpu_device *adev)
+static int navi10_ih_toggle_interrupts(struct amdgpu_device *adev, bool enable)
 {
-	u32 ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL);
-
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_ENABLE, 0);
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 0);
-	if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL, ih_rb_cntl)) {
-			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
-			return;
-		}
-	} else {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
-	}
-
-	/* set rptr, wptr to 0 */
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0);
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0);
-	adev->irq.ih.enabled = false;
-	adev->irq.ih.rptr = 0;
-
-	if (adev->irq.ih1.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
-					   RB_ENABLE, 0);
-		if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING1 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl);
-		}
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING1, 0);
-		adev->irq.ih1.enabled = false;
-		adev->irq.ih1.rptr = 0;
-	}
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
+	int i;
+	int r;
 
-	if (adev->irq.ih2.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING2,
-					   RB_ENABLE, 0);
-		if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING2,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING2 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl);
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			r = navi10_ih_toggle_ring_interrupts(adev, ih[i], enable);
+			if (r)
+				return r;
 		}
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING2, 0);
-		adev->irq.ih2.enabled = false;
-		adev->irq.ih2.rptr = 0;
 	}
 
+	return 0;
 }
 
 static uint32_t navi10_ih_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl)
@@ -253,22 +233,49 @@ static uint32_t navi10_ih_doorbell_rptr(struct amdgpu_ih_ring *ih)
 	return ih_doorbell_rtpr;
 }
 
-static void navi10_ih_reroute_ih(struct amdgpu_device *adev)
+/**
+ * navi10_ih_enable_ring - enable an ih ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Enable an ih ring buffer (NAVI10)
+ */
+static int navi10_ih_enable_ring(struct amdgpu_device *adev,
+				 struct amdgpu_ih_ring *ih)
 {
+	struct amdgpu_ih_regs *ih_regs;
 	uint32_t tmp;
 
-	/* Reroute to IH ring 1 for VMC */
-	WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_INDEX, 0x12);
-	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, CLIENT_TYPE, 1);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-	WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA, tmp);
-
-	/* Reroute IH ring 1 for UMC */
-	WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_INDEX, 0x1B);
-	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-	WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA, tmp);
+	ih_regs = &ih->ih_regs;
+
+	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
+	WREG32(ih_regs->ih_rb_base, ih->gpu_addr >> 8);
+	WREG32(ih_regs->ih_rb_base_hi, (ih->gpu_addr >> 40) & 0xff);
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = navi10_ih_rb_cntl(ih, tmp);
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
+	if (ih == &adev->irq.ih1) {
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
+	}
+	WREG32(ih_regs->ih_rb_cntl, tmp);
+
+	if (ih == &adev->irq.ih) {
+		/* set the ih ring 0 writeback address whether it's enabled or not */
+		WREG32(ih_regs->ih_rb_wptr_addr_lo, lower_32_bits(ih->wptr_addr));
+		WREG32(ih_regs->ih_rb_wptr_addr_hi, upper_32_bits(ih->wptr_addr) & 0xFFFF);
+	}
+
+	/* set rptr, wptr to 0 */
+	WREG32(ih_regs->ih_rb_wptr, 0);
+	WREG32(ih_regs->ih_rb_rptr, 0);
+
+	WREG32(ih_regs->ih_doorbell_rptr, navi10_ih_doorbell_rptr(ih));
+
+	return 0;
 }
 
 /**
@@ -284,36 +291,21 @@ static void navi10_ih_reroute_ih(struct amdgpu_device *adev)
  */
 static int navi10_ih_irq_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_ih_ring *ih = &adev->irq.ih;
-	u32 ih_rb_cntl, ih_chicken;
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
+	u32 ih_chicken;
 	u32 tmp;
+	int ret;
+	int i;
 
 	/* disable irqs */
-	navi10_ih_disable_interrupts(adev);
+	ret = navi10_ih_toggle_interrupts(adev, false);
+	if (ret)
+		return ret;
 
 	adev->nbio.funcs->ih_control(adev);
 
-	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, ih->gpu_addr >> 8);
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, (ih->gpu_addr >> 40) & 0xff);
-
-	ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL);
-	ih_rb_cntl = navi10_ih_rb_cntl(ih, ih_rb_cntl);
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RPTR_REARM,
-				   !!adev->irq.msi_enabled);
-	if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL, ih_rb_cntl)) {
-			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
-			return -ETIMEDOUT;
-		}
-	} else {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
-	}
-	if (adev->irq.ih1.ring_size)
-		navi10_ih_reroute_ih(adev);
-
 	if (unlikely(adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT)) {
-		if (ih->use_bus_addr) {
+		if (ih[0]->use_bus_addr) {
 			switch (adev->asic_type) {
 			case CHIP_SIENNA_CICHLID:
 			case CHIP_NAVY_FLOUNDER:
@@ -334,77 +326,17 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
 		}
 	}
 
-	/* set the writeback address whether it's enabled or not */
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO,
-		     lower_32_bits(ih->wptr_addr));
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI,
-		     upper_32_bits(ih->wptr_addr) & 0xFFFF);
-
-	/* set rptr, wptr to 0 */
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0);
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0);
-
-	WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR,
-			navi10_ih_doorbell_rptr(ih));
-
-	adev->nbio.funcs->ih_doorbell_range(adev, ih->use_doorbell,
-					    ih->doorbell_index);
-
-	ih = &adev->irq.ih1;
-	if (ih->ring_size) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_RING1, ih->gpu_addr >> 8);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI_RING1,
-			     (ih->gpu_addr >> 40) & 0xff);
-
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-		ih_rb_cntl = navi10_ih_rb_cntl(ih, ih_rb_cntl);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
-					   WPTR_OVERFLOW_ENABLE, 0);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
-					   RB_FULL_DRAIN_ENABLE, 1);
-		if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING1 failed!\n");
-				return -ETIMEDOUT;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl);
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			ret = navi10_ih_enable_ring(adev, ih[i]);
+			if (ret)
+				return ret;
 		}
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING1, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, 0);
-
-		WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING1,
-				navi10_ih_doorbell_rptr(ih));
-	}
-
-	ih = &adev->irq.ih2;
-	if (ih->ring_size) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_RING2, ih->gpu_addr >> 8);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI_RING2,
-			     (ih->gpu_addr >> 40) & 0xff);
-
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-		ih_rb_cntl = navi10_ih_rb_cntl(ih, ih_rb_cntl);
-
-		if (amdgpu_sriov_vf(adev) && adev->asic_type < CHIP_NAVI10) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING2,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING2 failed!\n");
-				return -ETIMEDOUT;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl);
-		}
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING2, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, 0);
-
-		WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING2,
-			     navi10_ih_doorbell_rptr(ih));
 	}
 
+	/* update doorbell range for ih ring 0*/
+	adev->nbio.funcs->ih_doorbell_range(adev, ih[0]->use_doorbell,
+					    ih[0]->doorbell_index);
 
 	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL);
 	tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL,
@@ -418,10 +350,15 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
 	pci_set_master(adev->pdev);
 
 	/* enable interrupts */
-	navi10_ih_enable_interrupts(adev);
+	ret = navi10_ih_toggle_interrupts(adev, true);
+	if (ret)
+		return ret;
 	/* enable wptr force update for self int */
 	force_update_wptr_for_self_int(adev, 0, 8, true);
 
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
 	return 0;
 }
 
@@ -435,7 +372,7 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
 static void navi10_ih_irq_disable(struct amdgpu_device *adev)
 {
 	force_update_wptr_for_self_int(adev, 0, 8, false);
-	navi10_ih_disable_interrupts(adev);
+	navi10_ih_toggle_interrupts(adev, false);
 
 	/* Wait and acknowledge irq */
 	mdelay(1);
@@ -455,23 +392,16 @@ static void navi10_ih_irq_disable(struct amdgpu_device *adev)
 static u32 navi10_ih_get_wptr(struct amdgpu_device *adev,
 			      struct amdgpu_ih_ring *ih)
 {
-	u32 wptr, reg, tmp;
+	u32 wptr, tmp;
+	struct amdgpu_ih_regs *ih_regs;
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
+	ih_regs = &ih->ih_regs;
 
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
-	if (ih == &adev->irq.ih)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR);
-	else if (ih == &adev->irq.ih1)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING1);
-	else if (ih == &adev->irq.ih2)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING2);
-	else
-		BUG();
-
-	wptr = RREG32_NO_KIQ(reg);
+	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
@@ -486,68 +416,14 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev,
 		 wptr, ih->rptr, tmp);
 	ih->rptr = tmp;
 
-	if (ih == &adev->irq.ih)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL);
-	else if (ih == &adev->irq.ih1)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-	else if (ih == &adev->irq.ih2)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-	else
-		BUG();
-
-	tmp = RREG32_NO_KIQ(reg);
+	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
-	WREG32_NO_KIQ(reg, tmp);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
 out:
 	return (wptr & ih->ptr_mask);
 }
 
 /**
- * navi10_ih_decode_iv - decode an interrupt vector
- *
- * @adev: amdgpu_device pointer
- * @ih: IH ring buffer to decode
- * @entry: IV entry to place decoded information into
- *
- * Decodes the interrupt vector at the current rptr
- * position and also advance the position.
- */
-static void navi10_ih_decode_iv(struct amdgpu_device *adev,
-				struct amdgpu_ih_ring *ih,
-				struct amdgpu_iv_entry *entry)
-{
-	/* wptr/rptr are in bytes! */
-	u32 ring_index = ih->rptr >> 2;
-	uint32_t dw[8];
-
-	dw[0] = le32_to_cpu(ih->ring[ring_index + 0]);
-	dw[1] = le32_to_cpu(ih->ring[ring_index + 1]);
-	dw[2] = le32_to_cpu(ih->ring[ring_index + 2]);
-	dw[3] = le32_to_cpu(ih->ring[ring_index + 3]);
-	dw[4] = le32_to_cpu(ih->ring[ring_index + 4]);
-	dw[5] = le32_to_cpu(ih->ring[ring_index + 5]);
-	dw[6] = le32_to_cpu(ih->ring[ring_index + 6]);
-	dw[7] = le32_to_cpu(ih->ring[ring_index + 7]);
-
-	entry->client_id = dw[0] & 0xff;
-	entry->src_id = (dw[0] >> 8) & 0xff;
-	entry->ring_id = (dw[0] >> 16) & 0xff;
-	entry->vmid = (dw[0] >> 24) & 0xf;
-	entry->vmid_src = (dw[0] >> 31);
-	entry->timestamp = dw[1] | ((u64)(dw[2] & 0xffff) << 32);
-	entry->timestamp_src = dw[2] >> 31;
-	entry->pasid = dw[3] & 0xffff;
-	entry->pasid_src = dw[3] >> 31;
-	entry->src_data[0] = dw[4];
-	entry->src_data[1] = dw[5];
-	entry->src_data[2] = dw[6];
-	entry->src_data[3] = dw[7];
-
-	/* wptr/rptr are in bytes! */
-	ih->rptr += 32;
-}
-
-/**
  * navi10_ih_irq_rearm - rearm IRQ if lost
  *
  * @adev: amdgpu_device pointer
@@ -557,22 +433,15 @@ static void navi10_ih_decode_iv(struct amdgpu_device *adev,
 static void navi10_ih_irq_rearm(struct amdgpu_device *adev,
 			       struct amdgpu_ih_ring *ih)
 {
-	uint32_t reg_rptr = 0;
 	uint32_t v = 0;
 	uint32_t i = 0;
+	struct amdgpu_ih_regs *ih_regs;
 
-	if (ih == &adev->irq.ih)
-		reg_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR);
-	else if (ih == &adev->irq.ih1)
-		reg_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING1);
-	else if (ih == &adev->irq.ih2)
-		reg_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING2);
-	else
-		return;
+	ih_regs = &ih->ih_regs;
 
 	/* Rearm IRQ / re-write doorbell if doorbell write is lost */
 	for (i = 0; i < MAX_REARM_RETRY; i++) {
-		v = RREG32_NO_KIQ(reg_rptr);
+		v = RREG32_NO_KIQ(ih_regs->ih_rb_rptr);
 		if ((v < ih->ring_size) && (v != ih->rptr))
 			WDOORBELL32(ih->doorbell_index, ih->rptr);
 		else
@@ -591,6 +460,8 @@ static void navi10_ih_irq_rearm(struct amdgpu_device *adev,
 static void navi10_ih_set_rptr(struct amdgpu_device *adev,
 			       struct amdgpu_ih_ring *ih)
 {
+	struct amdgpu_ih_regs *ih_regs;
+
 	if (ih->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
 		*ih->rptr_cpu = ih->rptr;
@@ -598,12 +469,9 @@ static void navi10_ih_set_rptr(struct amdgpu_device *adev,
 
 		if (amdgpu_sriov_vf(adev))
 			navi10_ih_irq_rearm(adev, ih);
-	} else if (ih == &adev->irq.ih) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, ih->rptr);
-	} else if (ih == &adev->irq.ih1) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, ih->rptr);
-	} else if (ih == &adev->irq.ih2) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, ih->rptr);
+	} else {
+		ih_regs = &ih->ih_regs;
+		WREG32(ih_regs->ih_rb_rptr, ih->rptr);
 	}
 }
 
@@ -685,23 +553,8 @@ static int navi10_ih_sw_init(void *handle)
 	adev->irq.ih1.ring_size = 0;
 	adev->irq.ih2.ring_size = 0;
 
-	if (adev->asic_type < CHIP_NAVI10) {
-		r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true);
-		if (r)
-			return r;
-
-		adev->irq.ih1.use_doorbell = true;
-		adev->irq.ih1.doorbell_index =
-					(adev->doorbell_index.ih + 1) << 1;
-
-		r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true);
-		if (r)
-			return r;
-
-		adev->irq.ih2.use_doorbell = true;
-		adev->irq.ih2.doorbell_index =
-					(adev->doorbell_index.ih + 2) << 1;
-	}
+	/* initialize ih control registers offset */
+	navi10_ih_init_register_offset(adev);
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
 	if (r)
@@ -717,6 +570,7 @@ static int navi10_ih_sw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	amdgpu_irq_fini(adev);
+	amdgpu_ih_ring_fini(adev, &adev->irq.ih_soft);
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih2);
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih1);
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih);
@@ -848,7 +702,7 @@ static const struct amd_ip_funcs navi10_ih_ip_funcs = {
 
 static const struct amdgpu_ih_funcs navi10_ih_funcs = {
 	.get_wptr = navi10_ih_get_wptr,
-	.decode_iv = navi10_ih_decode_iv,
+	.decode_iv = amdgpu_ih_decode_iv_helper,
 	.set_rptr = navi10_ih_set_rptr
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
index b5c3db16c2b0..05ddec7ba7e2 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
@@ -34,6 +34,14 @@
 #define smnCPM_CONTROL		0x11180460
 #define smnPCIE_CNTL2		0x11180070
 #define smnPCIE_LC_CNTL		0x11140280
+#define smnPCIE_LC_CNTL3	0x111402d4
+#define smnPCIE_LC_CNTL6	0x111402ec
+#define smnPCIE_LC_CNTL7	0x111402f0
+#define smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2	0x1014008c
+#define smnRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL	0x10123538
+#define smnBIF_CFG_DEV0_EPF0_PCIE_LTR_CAP	0x10140324
+#define smnPSWUSP0_PCIE_LC_CNTL2		0x111402c4
+#define smnNBIF_MGCG_CTRL_LCLK			0x1013a21c
 
 #define mmBIF_SDMA2_DOORBELL_RANGE		0x01d6
 #define mmBIF_SDMA2_DOORBELL_RANGE_BASE_IDX	2
@@ -80,15 +88,6 @@ static void nbio_v2_3_mc_access_enable(struct amdgpu_device *adev, bool enable)
 		WREG32_SOC15(NBIO, 0, mmBIF_FB_EN, 0);
 }
 
-static void nbio_v2_3_hdp_flush(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-	else
-		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-}
-
 static u32 nbio_v2_3_get_memsize(struct amdgpu_device *adev)
 {
 	return RREG32_SOC15(NBIO, 0, mmRCC_DEV0_EPF0_RCC_CONFIG_MEMSIZE);
@@ -359,6 +358,111 @@ static void nbio_v2_3_enable_aspm(struct amdgpu_device *adev,
 		WREG32_PCIE(smnPCIE_LC_CNTL, data);
 }
 
+static void nbio_v2_3_program_ltr(struct amdgpu_device *adev)
+{
+	uint32_t def, data;
+
+	WREG32_PCIE(smnRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL, 0x75EB);
+
+	def = data = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP2);
+	data &= ~RCC_BIF_STRAP2__STRAP_LTR_IN_ASPML1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP2, data);
+
+	def = data = RREG32_PCIE(smnRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL);
+	data &= ~EP_PCIE_TX_LTR_CNTL__LTR_PRIV_MSG_DIS_IN_PM_NON_D0_MASK;
+	if (def != data)
+		WREG32_PCIE(smnRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL, data);
+
+	def = data = RREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2);
+	data |= BIF_CFG_DEV0_EPF0_DEVICE_CNTL2__LTR_EN_MASK;
+	if (def != data)
+		WREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
+}
+
+static void nbio_v2_3_program_aspm(struct amdgpu_device *adev)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
+	data &= ~PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK;
+	data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
+	data |= PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
+	if (def != data)
+		WREG32_PCIE(smnPCIE_LC_CNTL, data);
+
+	def = data = RREG32_PCIE(smnPCIE_LC_CNTL7);
+	data |= PCIE_LC_CNTL7__LC_NBIF_ASPM_INPUT_EN_MASK;
+	if (def != data)
+		WREG32_PCIE(smnPCIE_LC_CNTL7, data);
+
+	def = data = RREG32_PCIE(smnNBIF_MGCG_CTRL_LCLK);
+	data |= NBIF_MGCG_CTRL_LCLK__NBIF_MGCG_REG_DIS_LCLK_MASK;
+	if (def != data)
+		WREG32_PCIE(smnNBIF_MGCG_CTRL_LCLK, data);
+
+	def = data = RREG32_PCIE(smnPCIE_LC_CNTL3);
+	data |= PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
+	if (def != data)
+		WREG32_PCIE(smnPCIE_LC_CNTL3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP3);
+	data &= ~RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER_MASK;
+	data &= ~RCC_BIF_STRAP3__STRAP_VLINK_PM_L1_ENTRY_TIMER_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP5);
+	data &= ~RCC_BIF_STRAP5__STRAP_VLINK_LDN_ENTRY_TIMER_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP5, data);
+
+	def = data = RREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2);
+	data &= ~BIF_CFG_DEV0_EPF0_DEVICE_CNTL2__LTR_EN_MASK;
+	if (def != data)
+		WREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
+
+	WREG32_PCIE(smnBIF_CFG_DEV0_EPF0_PCIE_LTR_CAP, 0x10011001);
+
+	def = data = RREG32_PCIE(smnPSWUSP0_PCIE_LC_CNTL2);
+	data |= PSWUSP0_PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L1_MASK |
+		PSWUSP0_PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L23_MASK;
+	data &= ~PSWUSP0_PCIE_LC_CNTL2__LC_RCV_L0_TO_RCV_L0S_DIS_MASK;
+	if (def != data)
+		WREG32_PCIE(smnPSWUSP0_PCIE_LC_CNTL2, data);
+
+	def = data = RREG32_PCIE(smnPCIE_LC_CNTL6);
+	data |= PCIE_LC_CNTL6__LC_L1_POWERDOWN_MASK |
+		PCIE_LC_CNTL6__LC_RX_L0S_STANDBY_EN_MASK;
+	if (def != data)
+		WREG32_PCIE(smnPCIE_LC_CNTL6, data);
+
+	nbio_v2_3_program_ltr(adev);
+
+	def = data = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP3);
+	data |= 0x5DE0 << RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER__SHIFT;
+	data |= 0x0010 << RCC_BIF_STRAP3__STRAP_VLINK_PM_L1_ENTRY_TIMER__SHIFT;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP5);
+	data |= 0x0010 << RCC_BIF_STRAP5__STRAP_VLINK_LDN_ENTRY_TIMER__SHIFT;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP5, data);
+
+	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
+	data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
+	data |= 0x9 << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
+	data |= 0x1 << PCIE_LC_CNTL__LC_PMI_TO_L1_DIS__SHIFT;
+	if (def != data)
+		WREG32_PCIE(smnPCIE_LC_CNTL, data);
+
+	def = data = RREG32_PCIE(smnPCIE_LC_CNTL3);
+	data &= ~PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
+	if (def != data)
+		WREG32_PCIE(smnPCIE_LC_CNTL3, data);
+}
+
 const struct amdgpu_nbio_funcs nbio_v2_3_funcs = {
 	.get_hdp_flush_req_offset = nbio_v2_3_get_hdp_flush_req_offset,
 	.get_hdp_flush_done_offset = nbio_v2_3_get_hdp_flush_done_offset,
@@ -366,7 +470,6 @@ const struct amdgpu_nbio_funcs nbio_v2_3_funcs = {
 	.get_pcie_data_offset = nbio_v2_3_get_pcie_data_offset,
 	.get_rev_id = nbio_v2_3_get_rev_id,
 	.mc_access_enable = nbio_v2_3_mc_access_enable,
-	.hdp_flush = nbio_v2_3_hdp_flush,
 	.get_memsize = nbio_v2_3_get_memsize,
 	.sdma_doorbell_range = nbio_v2_3_sdma_doorbell_range,
 	.vcn_doorbell_range = nbio_v2_3_vcn_doorbell_range,
@@ -380,4 +483,5 @@ const struct amdgpu_nbio_funcs nbio_v2_3_funcs = {
 	.init_registers = nbio_v2_3_init_registers,
 	.remap_hdp_registers = nbio_v2_3_remap_hdp_registers,
 	.enable_aspm = nbio_v2_3_enable_aspm,
+	.program_aspm =  nbio_v2_3_program_aspm,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
index d2f1fe55d388..83ea063388fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
@@ -29,6 +29,15 @@
 #include "nbio/nbio_6_1_sh_mask.h"
 #include "nbio/nbio_6_1_smn.h"
 #include "vega10_enum.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void nbio_v6_1_remap_hdp_registers(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, mmREMAP_HDP_MEM_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL);
+	WREG32_SOC15(NBIO, 0, mmREMAP_HDP_REG_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_REG_FLUSH_CNTL);
+}
 
 static u32 nbio_v6_1_get_rev_id(struct amdgpu_device *adev)
 {
@@ -50,18 +59,6 @@ static void nbio_v6_1_mc_access_enable(struct amdgpu_device *adev, bool enable)
 		WREG32_SOC15(NBIO, 0, mmBIF_FB_EN, 0);
 }
 
-static void nbio_v6_1_hdp_flush(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_SOC15_NO_KIQ(NBIO, 0,
-				    mmBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL,
-				    0);
-	else
-		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
-			NBIO, 0, mmBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL), 0);
-}
-
 static u32 nbio_v6_1_get_memsize(struct amdgpu_device *adev)
 {
 	return RREG32_SOC15(NBIO, 0, mmRCC_PF_0_0_RCC_CONFIG_MEMSIZE);
@@ -266,7 +263,6 @@ const struct amdgpu_nbio_funcs nbio_v6_1_funcs = {
 	.get_pcie_data_offset = nbio_v6_1_get_pcie_data_offset,
 	.get_rev_id = nbio_v6_1_get_rev_id,
 	.mc_access_enable = nbio_v6_1_mc_access_enable,
-	.hdp_flush = nbio_v6_1_hdp_flush,
 	.get_memsize = nbio_v6_1_get_memsize,
 	.sdma_doorbell_range = nbio_v6_1_sdma_doorbell_range,
 	.enable_doorbell_aperture = nbio_v6_1_enable_doorbell_aperture,
@@ -277,4 +273,5 @@ const struct amdgpu_nbio_funcs nbio_v6_1_funcs = {
 	.get_clockgating_state = nbio_v6_1_get_clockgating_state,
 	.ih_control = nbio_v6_1_ih_control,
 	.init_registers = nbio_v6_1_init_registers,
+	.remap_hdp_registers = nbio_v6_1_remap_hdp_registers,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
index ae685813c419..3c00666a13e1 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
@@ -60,15 +60,6 @@ static void nbio_v7_0_mc_access_enable(struct amdgpu_device *adev, bool enable)
 		WREG32_SOC15(NBIO, 0, mmBIF_FB_EN, 0);
 }
 
-static void nbio_v7_0_hdp_flush(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-	else
-		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-}
-
 static u32 nbio_v7_0_get_memsize(struct amdgpu_device *adev)
 {
 	return RREG32_SOC15(NBIO, 0, mmRCC_CONFIG_MEMSIZE);
@@ -292,7 +283,6 @@ const struct amdgpu_nbio_funcs nbio_v7_0_funcs = {
 	.get_pcie_data_offset = nbio_v7_0_get_pcie_data_offset,
 	.get_rev_id = nbio_v7_0_get_rev_id,
 	.mc_access_enable = nbio_v7_0_mc_access_enable,
-	.hdp_flush = nbio_v7_0_hdp_flush,
 	.get_memsize = nbio_v7_0_get_memsize,
 	.sdma_doorbell_range = nbio_v7_0_sdma_doorbell_range,
 	.vcn_doorbell_range = nbio_v7_0_vcn_doorbell_range,
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
index aa36022670f9..598ce0e93627 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
@@ -56,15 +56,6 @@ static void nbio_v7_2_mc_access_enable(struct amdgpu_device *adev, bool enable)
 		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN, 0);
 }
 
-static void nbio_v7_2_hdp_flush(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-	else
-		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-}
-
 static u32 nbio_v7_2_get_memsize(struct amdgpu_device *adev)
 {
 	return RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_CONFIG_MEMSIZE);
@@ -325,7 +316,6 @@ const struct amdgpu_nbio_funcs nbio_v7_2_funcs = {
 	.get_pcie_port_data_offset = nbio_v7_2_get_pcie_port_data_offset,
 	.get_rev_id = nbio_v7_2_get_rev_id,
 	.mc_access_enable = nbio_v7_2_mc_access_enable,
-	.hdp_flush = nbio_v7_2_hdp_flush,
 	.get_memsize = nbio_v7_2_get_memsize,
 	.sdma_doorbell_range = nbio_v7_2_sdma_doorbell_range,
 	.vcn_doorbell_range = nbio_v7_2_vcn_doorbell_range,
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
index eadc9526d33f..4bc1d1434065 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
@@ -82,15 +82,6 @@ static void nbio_v7_4_mc_access_enable(struct amdgpu_device *adev, bool enable)
 		WREG32_SOC15(NBIO, 0, mmBIF_FB_EN, 0);
 }
 
-static void nbio_v7_4_hdp_flush(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-	else
-		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-}
-
 static u32 nbio_v7_4_get_memsize(struct amdgpu_device *adev)
 {
 	return RREG32_SOC15(NBIO, 0, mmRCC_CONFIG_MEMSIZE);
@@ -541,7 +532,6 @@ const struct amdgpu_nbio_funcs nbio_v7_4_funcs = {
 	.get_pcie_data_offset = nbio_v7_4_get_pcie_data_offset,
 	.get_rev_id = nbio_v7_4_get_rev_id,
 	.mc_access_enable = nbio_v7_4_mc_access_enable,
-	.hdp_flush = nbio_v7_4_hdp_flush,
 	.get_memsize = nbio_v7_4_get_memsize,
 	.sdma_doorbell_range = nbio_v7_4_sdma_doorbell_range,
 	.vcn_doorbell_range = nbio_v7_4_vcn_doorbell_range,
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 6bee3677394a..c625c5d8ed89 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -38,9 +38,6 @@
 
 #include "gc/gc_10_1_0_offset.h"
 #include "gc/gc_10_1_0_sh_mask.h"
-#include "hdp/hdp_5_0_0_offset.h"
-#include "hdp/hdp_5_0_0_sh_mask.h"
-#include "smuio/smuio_11_0_0_offset.h"
 #include "mp/mp_11_0_offset.h"
 
 #include "soc15.h"
@@ -50,6 +47,7 @@
 #include "mmhub_v2_0.h"
 #include "nbio_v2_3.h"
 #include "nbio_v7_2.h"
+#include "hdp_v5_0.h"
 #include "nv.h"
 #include "navi10_ih.h"
 #include "gfx_v10_0.h"
@@ -62,6 +60,8 @@
 #include "dce_virtual.h"
 #include "mes_v10_1.h"
 #include "mxgpu_nv.h"
+#include "smuio_v11_0.h"
+#include "smuio_v11_0_6.h"
 
 static const struct amd_ip_funcs nv_common_ip_funcs;
 
@@ -203,6 +203,7 @@ static bool nv_read_bios_from_rom(struct amdgpu_device *adev,
 {
 	u32 *dw_ptr;
 	u32 i, length_dw;
+	u32 rom_index_offset, rom_data_offset;
 
 	if (bios == NULL)
 		return false;
@@ -215,11 +216,16 @@ static bool nv_read_bios_from_rom(struct amdgpu_device *adev,
 	dw_ptr = (u32 *)bios;
 	length_dw = ALIGN(length_bytes, 4) / 4;
 
+	rom_index_offset =
+		adev->smuio.funcs->get_rom_index_offset(adev);
+	rom_data_offset =
+		adev->smuio.funcs->get_rom_data_offset(adev);
+
 	/* set rom index to 0 */
-	WREG32(SOC15_REG_OFFSET(SMUIO, 0, mmROM_INDEX), 0);
+	WREG32(rom_index_offset, 0);
 	/* read out the rom data */
 	for (i = 0; i < length_dw; i++)
-		dw_ptr[i] = RREG32(SOC15_REG_OFFSET(SMUIO, 0, mmROM_DATA));
+		dw_ptr[i] = RREG32(rom_data_offset);
 
 	return true;
 }
@@ -336,6 +342,38 @@ static int nv_asic_mode1_reset(struct amdgpu_device *adev)
 	return ret;
 }
 
+static int nv_asic_mode2_reset(struct amdgpu_device *adev)
+{
+	u32 i;
+	int ret = 0;
+
+	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
+
+	/* disable BM */
+	pci_clear_master(adev->pdev);
+
+	amdgpu_device_cache_pci_state(adev->pdev);
+
+	ret = amdgpu_dpm_mode2_reset(adev);
+	if (ret)
+		dev_err(adev->dev, "GPU mode2 reset failed\n");
+
+	amdgpu_device_load_pci_state(adev->pdev);
+
+	/* wait for asic to come out of reset */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		u32 memsize = adev->nbio.funcs->get_memsize(adev);
+
+		if (memsize != 0xffffffff)
+			break;
+		udelay(1);
+	}
+
+	amdgpu_atombios_scratch_regs_engine_hung(adev, false);
+
+	return ret;
+}
+
 static bool nv_asic_supports_baco(struct amdgpu_device *adev)
 {
 	struct smu_context *smu = &adev->smu;
@@ -352,7 +390,9 @@ nv_asic_reset_method(struct amdgpu_device *adev)
 	struct smu_context *smu = &adev->smu;
 
 	if (amdgpu_reset_method == AMD_RESET_METHOD_MODE1 ||
-	    amdgpu_reset_method == AMD_RESET_METHOD_BACO)
+	    amdgpu_reset_method == AMD_RESET_METHOD_MODE2 ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_BACO ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_PCI)
 		return amdgpu_reset_method;
 
 	if (amdgpu_reset_method != -1)
@@ -360,6 +400,8 @@ nv_asic_reset_method(struct amdgpu_device *adev)
 				  amdgpu_reset_method);
 
 	switch (adev->asic_type) {
+	case CHIP_VANGOGH:
+		return AMD_RESET_METHOD_MODE2;
 	case CHIP_SIENNA_CICHLID:
 	case CHIP_NAVY_FLOUNDER:
 	case CHIP_DIMGREY_CAVEFISH:
@@ -377,7 +419,16 @@ static int nv_asic_reset(struct amdgpu_device *adev)
 	int ret = 0;
 	struct smu_context *smu = &adev->smu;
 
-	if (nv_asic_reset_method(adev) == AMD_RESET_METHOD_BACO) {
+	/* skip reset on vangogh for now */
+	if (adev->asic_type == CHIP_VANGOGH)
+		return 0;
+
+	switch (nv_asic_reset_method(adev)) {
+	case AMD_RESET_METHOD_PCI:
+		dev_info(adev->dev, "PCI reset\n");
+		ret = amdgpu_device_pci_reset(adev);
+		break;
+	case AMD_RESET_METHOD_BACO:
 		dev_info(adev->dev, "BACO reset\n");
 
 		ret = smu_baco_enter(smu);
@@ -386,9 +437,15 @@ static int nv_asic_reset(struct amdgpu_device *adev)
 		ret = smu_baco_exit(smu);
 		if (ret)
 			return ret;
-	} else {
+		break;
+	case AMD_RESET_METHOD_MODE2:
+		dev_info(adev->dev, "MODE2 reset\n");
+		ret = nv_asic_mode2_reset(adev);
+		break;
+	default:
 		dev_info(adev->dev, "MODE1 reset\n");
 		ret = nv_asic_mode1_reset(adev);
+		break;
 	}
 
 	return ret;
@@ -423,11 +480,14 @@ static void nv_pcie_gen3_enable(struct amdgpu_device *adev)
 
 static void nv_program_aspm(struct amdgpu_device *adev)
 {
-
-	if (amdgpu_aspm == 0)
+	if (amdgpu_aspm != 1)
 		return;
 
-	/* todo */
+	if ((adev->asic_type >= CHIP_SIENNA_CICHLID) &&
+	    !(adev->flags & AMD_IS_APU) &&
+	    (adev->nbio.funcs->program_aspm))
+		adev->nbio.funcs->program_aspm(adev);
+
 }
 
 static void nv_enable_doorbell_aperture(struct amdgpu_device *adev,
@@ -498,7 +558,8 @@ static bool nv_is_headless_sku(struct pci_dev *pdev)
 {
 	if ((pdev->device == 0x731E &&
 	    (pdev->revision == 0xC6 || pdev->revision == 0xC7)) ||
-	    (pdev->device == 0x7340 && pdev->revision == 0xC9))
+	    (pdev->device == 0x7340 && pdev->revision == 0xC9)  ||
+	    (pdev->device == 0x7360 && pdev->revision == 0xC7))
 		return true;
 	return false;
 }
@@ -514,6 +575,12 @@ int nv_set_ip_blocks(struct amdgpu_device *adev)
 		adev->nbio.funcs = &nbio_v2_3_funcs;
 		adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg;
 	}
+	adev->hdp.funcs = &hdp_v5_0_funcs;
+
+	if (adev->asic_type >= CHIP_SIENNA_CICHLID)
+		adev->smuio.funcs = &smuio_v11_0_6_funcs;
+	else
+		adev->smuio.funcs = &smuio_v11_0_funcs;
 
 	if (adev->asic_type == CHIP_SIENNA_CICHLID)
 		adev->gmc.xgmi.supported = true;
@@ -568,7 +635,8 @@ int nv_set_ip_blocks(struct amdgpu_device *adev)
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT &&
 		    !amdgpu_sriov_vf(adev))
 			amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
-		amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block);
+		if (!nv_is_headless_sku(adev->pdev))
+		        amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block);
 		if (!amdgpu_sriov_vf(adev))
 			amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block);
 		break;
@@ -669,22 +737,6 @@ static uint32_t nv_get_rev_id(struct amdgpu_device *adev)
 	return adev->nbio.funcs->get_rev_id(adev);
 }
 
-static void nv_flush_hdp(struct amdgpu_device *adev, struct amdgpu_ring *ring)
-{
-	adev->nbio.funcs->hdp_flush(adev, ring);
-}
-
-static void nv_invalidate_hdp(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg) {
-		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
-	} else {
-		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
-					HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
-	}
-}
-
 static bool nv_need_full_reset(struct amdgpu_device *adev)
 {
 	return true;
@@ -768,10 +820,10 @@ static int nv_update_umd_stable_pstate(struct amdgpu_device *adev,
 	 * The ASPM function is not fully enabled and verified on
 	 * Navi yet. Temporarily skip this until ASPM enabled.
 	 */
-#if 0
-	if (adev->nbio.funcs->enable_aspm)
+	if ((adev->asic_type >= CHIP_SIENNA_CICHLID) &&
+	    !(adev->flags & AMD_IS_APU) &&
+	    (adev->nbio.funcs->enable_aspm))
 		adev->nbio.funcs->enable_aspm(adev, !enter);
-#endif
 
 	return 0;
 }
@@ -788,8 +840,6 @@ static const struct amdgpu_asic_funcs nv_asic_funcs =
 	.set_uvd_clocks = &nv_set_uvd_clocks,
 	.set_vce_clocks = &nv_set_vce_clocks,
 	.get_config_memsize = &nv_get_config_memsize,
-	.flush_hdp = &nv_flush_hdp,
-	.invalidate_hdp = &nv_invalidate_hdp,
 	.init_doorbell_index = &nv_init_doorbell_index,
 	.need_full_reset = &nv_need_full_reset,
 	.need_reset_on_init = &nv_need_reset_on_init,
@@ -1080,120 +1130,6 @@ static int nv_common_soft_reset(void *handle)
 	return 0;
 }
 
-static void nv_update_hdp_mem_power_gating(struct amdgpu_device *adev,
-					   bool enable)
-{
-	uint32_t hdp_clk_cntl, hdp_clk_cntl1;
-	uint32_t hdp_mem_pwr_cntl;
-
-	if (!(adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS |
-				AMD_CG_SUPPORT_HDP_DS |
-				AMD_CG_SUPPORT_HDP_SD)))
-		return;
-
-	hdp_clk_cntl = hdp_clk_cntl1 = RREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL);
-	hdp_mem_pwr_cntl = RREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL);
-
-	/* Before doing clock/power mode switch,
-	 * forced on IPH & RC clock */
-	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
-				     IPH_MEM_CLK_SOFT_OVERRIDE, 1);
-	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
-				     RC_MEM_CLK_SOFT_OVERRIDE, 1);
-	WREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL, hdp_clk_cntl);
-
-	/* HDP 5.0 doesn't support dynamic power mode switch,
-	 * disable clock and power gating before any changing */
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 IPH_MEM_POWER_CTRL_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 IPH_MEM_POWER_LS_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 IPH_MEM_POWER_DS_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 IPH_MEM_POWER_SD_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 RC_MEM_POWER_CTRL_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 RC_MEM_POWER_LS_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 RC_MEM_POWER_DS_EN, 0);
-	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-					 RC_MEM_POWER_SD_EN, 0);
-	WREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
-
-	/* only one clock gating mode (LS/DS/SD) can be enabled */
-	if (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS) {
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
-						 HDP_MEM_POWER_CTRL,
-						 IPH_MEM_POWER_LS_EN, enable);
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
-						 HDP_MEM_POWER_CTRL,
-						 RC_MEM_POWER_LS_EN, enable);
-	} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_DS) {
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
-						 HDP_MEM_POWER_CTRL,
-						 IPH_MEM_POWER_DS_EN, enable);
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
-						 HDP_MEM_POWER_CTRL,
-						 RC_MEM_POWER_DS_EN, enable);
-	} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_SD) {
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
-						 HDP_MEM_POWER_CTRL,
-						 IPH_MEM_POWER_SD_EN, enable);
-		/* RC should not use shut down mode, fallback to ds */
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
-						 HDP_MEM_POWER_CTRL,
-						 RC_MEM_POWER_DS_EN, enable);
-	}
-
-	/* confirmed that IPH_MEM_POWER_CTRL_EN and RC_MEM_POWER_CTRL_EN have to
-	 * be set for SRAM LS/DS/SD */
-	if (adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_HDP_DS |
-							AMD_CG_SUPPORT_HDP_SD)) {
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-						IPH_MEM_POWER_CTRL_EN, 1);
-		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
-						RC_MEM_POWER_CTRL_EN, 1);
-	}
-
-	WREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
-
-	/* restore IPH & RC clock override after clock/power mode changing */
-	WREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL, hdp_clk_cntl1);
-}
-
-static void nv_update_hdp_clock_gating(struct amdgpu_device *adev,
-				       bool enable)
-{
-	uint32_t hdp_clk_cntl;
-
-	if (!(adev->cg_flags & AMD_CG_SUPPORT_HDP_MGCG))
-		return;
-
-	hdp_clk_cntl = RREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL);
-
-	if (enable) {
-		hdp_clk_cntl &=
-			~(uint32_t)
-			  (HDP_CLK_CNTL__IPH_MEM_CLK_SOFT_OVERRIDE_MASK |
-			   HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
-			   HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
-			   HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
-			   HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
-			   HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK);
-	} else {
-		hdp_clk_cntl |= HDP_CLK_CNTL__IPH_MEM_CLK_SOFT_OVERRIDE_MASK |
-			HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
-			HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
-			HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
-			HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
-			HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK;
-	}
-
-	WREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL, hdp_clk_cntl);
-}
-
 static int nv_common_set_clockgating_state(void *handle,
 					   enum amd_clockgating_state state)
 {
@@ -1213,9 +1149,9 @@ static int nv_common_set_clockgating_state(void *handle,
 				state == AMD_CG_STATE_GATE);
 		adev->nbio.funcs->update_medium_grain_light_sleep(adev,
 				state == AMD_CG_STATE_GATE);
-		nv_update_hdp_mem_power_gating(adev,
-				   state == AMD_CG_STATE_GATE);
-		nv_update_hdp_clock_gating(adev,
+		adev->hdp.funcs->update_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		adev->smuio.funcs->update_rom_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
 		break;
 	default:
@@ -1234,31 +1170,15 @@ static int nv_common_set_powergating_state(void *handle,
 static void nv_common_get_clockgating_state(void *handle, u32 *flags)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	uint32_t tmp;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
 
 	adev->nbio.funcs->get_clockgating_state(adev, flags);
 
-	/* AMD_CG_SUPPORT_HDP_MGCG */
-	tmp = RREG32_SOC15(HDP, 0, mmHDP_CLK_CNTL);
-	if (!(tmp & (HDP_CLK_CNTL__IPH_MEM_CLK_SOFT_OVERRIDE_MASK |
-		     HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
-		     HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
-		     HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
-		     HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
-		     HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK)))
-		*flags |= AMD_CG_SUPPORT_HDP_MGCG;
-
-	/* AMD_CG_SUPPORT_HDP_LS/DS/SD */
-	tmp = RREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL);
-	if (tmp & HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK)
-		*flags |= AMD_CG_SUPPORT_HDP_LS;
-	else if (tmp & HDP_MEM_POWER_CTRL__IPH_MEM_POWER_DS_EN_MASK)
-		*flags |= AMD_CG_SUPPORT_HDP_DS;
-	else if (tmp & HDP_MEM_POWER_CTRL__IPH_MEM_POWER_SD_EN_MASK)
-		*flags |= AMD_CG_SUPPORT_HDP_SD;
+	adev->hdp.funcs->get_clock_gating_state(adev, flags);
+
+	adev->smuio.funcs->get_clock_gating_state(adev, flags);
 
 	return;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h
index d65a5339d354..3ba7bdfde65d 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h
@@ -47,7 +47,7 @@ enum psp_gfx_crtl_cmd_id
     GFX_CTRL_CMD_ID_DISABLE_INT     = 0x00060000,   /* disable PSP-to-Gfx interrupt */
     GFX_CTRL_CMD_ID_MODE1_RST       = 0x00070000,   /* trigger the Mode 1 reset */
     GFX_CTRL_CMD_ID_GBR_IH_SET      = 0x00080000,   /* set Gbr IH_RB_CNTL registers */
-    GFX_CTRL_CMD_ID_CONSUME_CMD     = 0x000A0000,   /* send interrupt to psp for updating write pointer of vf */
+    GFX_CTRL_CMD_ID_CONSUME_CMD     = 0x00090000,   /* send interrupt to psp for updating write pointer of vf */
     GFX_CTRL_CMD_ID_DESTROY_GPCOM_RING = 0x000C0000, /* destroy GPCOM ring */
 
     GFX_CTRL_CMD_ID_MAX             = 0x000F0000,   /* max command ID */
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
index d7f92634eba2..4b1cc5e9ee92 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
@@ -92,8 +92,6 @@ static int psp_v10_0_init_microcode(struct psp_context *psp)
 			(uint8_t *)ta_hdr +
 			le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
 
-		adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
-
 		adev->psp.ta_dtm_ucode_version =
 			le32_to_cpu(ta_hdr->ta_dtm_ucode_version);
 		adev->psp.ta_dtm_ucode_size =
@@ -101,6 +99,16 @@ static int psp_v10_0_init_microcode(struct psp_context *psp)
 		adev->psp.ta_dtm_start_addr =
 			(uint8_t *)adev->psp.ta_hdcp_start_addr +
 			le32_to_cpu(ta_hdr->ta_dtm_offset_bytes);
+
+		adev->psp.ta_securedisplay_ucode_version =
+			le32_to_cpu(ta_hdr->ta_securedisplay_ucode_version);
+		adev->psp.ta_securedisplay_ucode_size =
+			le32_to_cpu(ta_hdr->ta_securedisplay_size_bytes);
+		adev->psp.ta_securedisplay_start_addr =
+			(uint8_t *)adev->psp.ta_hdcp_start_addr +
+			le32_to_cpu(ta_hdr->ta_securedisplay_offset_bytes);
+
+		adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
index bd4248c93c49..c325d6f53a71 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
@@ -392,37 +392,6 @@ static int psp_v11_0_bootloader_load_sos(struct psp_context *psp)
 	return ret;
 }
 
-static void psp_v11_0_reroute_ih(struct psp_context *psp)
-{
-	struct amdgpu_device *adev = psp->adev;
-	uint32_t tmp;
-
-	/* Change IH ring for VMC */
-	tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1244b);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, CLIENT_TYPE, 1);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, 3);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, tmp);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_GBR_IH_SET);
-
-	mdelay(20);
-	psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-		     0x80000000, 0x8000FFFF, false);
-
-	/* Change IH ring for UMC */
-	tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1216b);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, 4);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, tmp);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_GBR_IH_SET);
-
-	mdelay(20);
-	psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-		     0x80000000, 0x8000FFFF, false);
-}
-
 static int psp_v11_0_ring_init(struct psp_context *psp,
 			      enum psp_ring_type ring_type)
 {
@@ -430,11 +399,6 @@ static int psp_v11_0_ring_init(struct psp_context *psp,
 	struct psp_ring *ring;
 	struct amdgpu_device *adev = psp->adev;
 
-	if ((!amdgpu_sriov_vf(adev)) &&
-	    !(adev->asic_type >= CHIP_SIENNA_CICHLID &&
-	    adev->asic_type <= CHIP_DIMGREY_CAVEFISH))
-		psp_v11_0_reroute_ih(psp);
-
 	ring = &psp->km_ring;
 
 	ring->ring_type = ring_type;
@@ -726,7 +690,7 @@ static int psp_v11_0_memory_training(struct psp_context *psp, uint32_t ops)
 		}
 
 		memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
-		adev->nbio.funcs->hdp_flush(adev, NULL);
+		adev->hdp.funcs->flush_hdp(adev, NULL);
 		vfree(buf);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index ce56e93c6886..c8c22c1d1e65 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -46,7 +46,6 @@
 #include "sdma6/sdma6_4_2_2_sh_mask.h"
 #include "sdma7/sdma7_4_2_2_offset.h"
 #include "sdma7/sdma7_4_2_2_sh_mask.h"
-#include "hdp/hdp_4_0_offset.h"
 #include "sdma0/sdma0_4_1_default.h"
 
 #include "soc15_common.h"
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
index b208b81005bb..d345e324837d 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
@@ -32,7 +32,6 @@
 
 #include "gc/gc_10_1_0_offset.h"
 #include "gc/gc_10_1_0_sh_mask.h"
-#include "hdp/hdp_5_0_0_offset.h"
 #include "ivsrcid/sdma0/irqsrcs_sdma0_5_0.h"
 #include "ivsrcid/sdma1/irqsrcs_sdma1_5_0.h"
 
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index f1ba36a094da..690a5090475a 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -119,15 +119,7 @@ static int sdma_v5_2_init_inst_ctx(struct amdgpu_sdma_instance *sdma_inst)
 
 static void sdma_v5_2_destroy_inst_ctx(struct amdgpu_device *adev)
 {
-	int i;
-
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		release_firmware(adev->sdma.instance[i].fw);
-		adev->sdma.instance[i].fw = NULL;
-
-		if (adev->asic_type == CHIP_SIENNA_CICHLID)
-			break;
-	}
+	release_firmware(adev->sdma.instance[0].fw);
 
 	memset((void *)adev->sdma.instance, 0,
 	       sizeof(struct amdgpu_sdma_instance) * AMDGPU_MAX_SDMA_INSTANCES);
@@ -185,23 +177,10 @@ static int sdma_v5_2_init_microcode(struct amdgpu_device *adev)
 	if (err)
 		goto out;
 
-	for (i = 1; i < adev->sdma.num_instances; i++) {
-		if (adev->asic_type >= CHIP_SIENNA_CICHLID &&
-		    adev->asic_type <= CHIP_DIMGREY_CAVEFISH) {
-			memcpy((void *)&adev->sdma.instance[i],
-			       (void *)&adev->sdma.instance[0],
-			       sizeof(struct amdgpu_sdma_instance));
-		} else {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma%d.bin", chip_name, i);
-			err = request_firmware(&adev->sdma.instance[i].fw, fw_name, adev->dev);
-			if (err)
-				goto out;
-
-			err = sdma_v5_2_init_inst_ctx(&adev->sdma.instance[i]);
-			if (err)
-				goto out;
-		}
-	}
+	for (i = 1; i < adev->sdma.num_instances; i++)
+		memcpy((void *)&adev->sdma.instance[i],
+		       (void *)&adev->sdma.instance[0],
+		       sizeof(struct amdgpu_sdma_instance));
 
 	DRM_DEBUG("psp_load == '%s'\n",
 		  adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ? "true" : "false");
diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c
index 3cf0589bfea5..6b5cf7882a12 100644
--- a/drivers/gpu/drm/amd/amdgpu/si.c
+++ b/drivers/gpu/drm/amd/amdgpu/si.c
@@ -1270,7 +1270,7 @@ static int si_gpu_pci_config_reset(struct amdgpu_device *adev)
 	u32 i;
 	int r = -EINVAL;
 
-	dev_info(adev->dev, "GPU pci config reset\n");
+	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
 
 	/* set mclk/sclk to bypass */
 	si_set_clk_bypass_mode(adev);
@@ -1294,20 +1294,6 @@ static int si_gpu_pci_config_reset(struct amdgpu_device *adev)
 		}
 		udelay(1);
 	}
-
-	return r;
-}
-
-static int si_asic_reset(struct amdgpu_device *adev)
-{
-	int r;
-
-	dev_info(adev->dev, "PCI CONFIG reset\n");
-
-	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
-
-	r = si_gpu_pci_config_reset(adev);
-
 	amdgpu_atombios_scratch_regs_engine_hung(adev, false);
 
 	return r;
@@ -1321,14 +1307,34 @@ static bool si_asic_supports_baco(struct amdgpu_device *adev)
 static enum amd_reset_method
 si_asic_reset_method(struct amdgpu_device *adev)
 {
-	if (amdgpu_reset_method != AMD_RESET_METHOD_LEGACY &&
-	    amdgpu_reset_method != -1)
+	if (amdgpu_reset_method == AMD_RESET_METHOD_PCI)
+		return amdgpu_reset_method;
+	else if (amdgpu_reset_method != AMD_RESET_METHOD_LEGACY &&
+		 amdgpu_reset_method != -1)
 		dev_warn(adev->dev, "Specified reset method:%d isn't supported, using AUTO instead.\n",
-				  amdgpu_reset_method);
+			 amdgpu_reset_method);
 
 	return AMD_RESET_METHOD_LEGACY;
 }
 
+static int si_asic_reset(struct amdgpu_device *adev)
+{
+	int r;
+
+	switch (si_asic_reset_method(adev)) {
+	case AMD_RESET_METHOD_PCI:
+		dev_info(adev->dev, "PCI reset\n");
+		r = amdgpu_device_pci_reset(adev);
+		break;
+	default:
+		dev_info(adev->dev, "PCI CONFIG reset\n");
+		r = si_gpu_pci_config_reset(adev);
+		break;
+	}
+
+	return r;
+}
+
 static u32 si_get_config_memsize(struct amdgpu_device *adev)
 {
 	return RREG32(mmCONFIG_MEMSIZE);
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c
new file mode 100644
index 000000000000..3a18dbb55c32
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c
@@ -0,0 +1,77 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "smuio_v11_0_6.h"
+#include "smuio/smuio_11_0_6_offset.h"
+#include "smuio/smuio_11_0_6_sh_mask.h"
+
+static u32 smuio_v11_0_6_get_rom_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(SMUIO, 0, mmROM_INDEX);
+}
+
+static u32 smuio_v11_0_6_get_rom_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(SMUIO, 0, mmROM_DATA);
+}
+
+static void smuio_v11_0_6_update_rom_clock_gating(struct amdgpu_device *adev, bool enable)
+{
+	u32 def, data;
+
+	/* enable/disable ROM CG is not supported on APU */
+	if (adev->flags & AMD_IS_APU)
+		return;
+
+	def = data = RREG32_SOC15(SMUIO, 0, mmCGTT_ROM_CLK_CTRL0);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_ROM_MGCG))
+		data &= ~(CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE0_MASK |
+			CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE1_MASK);
+	else
+		data |= CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE0_MASK |
+			CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE1_MASK;
+
+	if (def != data)
+		WREG32_SOC15(SMUIO, 0, mmCGTT_ROM_CLK_CTRL0, data);
+}
+
+static void smuio_v11_0_6_get_clock_gating_state(struct amdgpu_device *adev, u32 *flags)
+{
+	u32 data;
+
+	/* CGTT_ROM_CLK_CTRL0 is not available for APU */
+	if (adev->flags & AMD_IS_APU)
+		return;
+
+	data = RREG32_SOC15(SMUIO, 0, mmCGTT_ROM_CLK_CTRL0);
+	if (!(data & CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE0_MASK))
+		*flags |= AMD_CG_SUPPORT_ROM_MGCG;
+}
+
+const struct amdgpu_smuio_funcs smuio_v11_0_6_funcs = {
+	.get_rom_index_offset = smuio_v11_0_6_get_rom_index_offset,
+	.get_rom_data_offset = smuio_v11_0_6_get_rom_data_offset,
+	.update_rom_clock_gating = smuio_v11_0_6_update_rom_clock_gating,
+	.get_clock_gating_state = smuio_v11_0_6_get_clock_gating_state,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.h b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.h
new file mode 100644
index 000000000000..3c3f4ab0bc9b
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SMUIO_V11_0_6_H__
+#define __SMUIO_V11_0_6_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_smuio_funcs smuio_v11_0_6_funcs;
+
+#endif /* __SMUIO_V11_0_6_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index 8a23636ecc27..1221aa6b40a9 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -40,8 +40,6 @@
 #include "gc/gc_9_0_sh_mask.h"
 #include "sdma0/sdma0_4_0_offset.h"
 #include "sdma1/sdma1_4_0_offset.h"
-#include "hdp/hdp_4_0_offset.h"
-#include "hdp/hdp_4_0_sh_mask.h"
 #include "nbio/nbio_7_0_default.h"
 #include "nbio/nbio_7_0_offset.h"
 #include "nbio/nbio_7_0_sh_mask.h"
@@ -59,7 +57,9 @@
 #include "nbio_v6_1.h"
 #include "nbio_v7_0.h"
 #include "nbio_v7_4.h"
+#include "hdp_v4_0.h"
 #include "vega10_ih.h"
+#include "vega20_ih.h"
 #include "navi10_ih.h"
 #include "sdma_v4_0.h"
 #include "uvd_v7_0.h"
@@ -83,14 +83,6 @@
 #define mmMP0_MISC_LIGHT_SLEEP_CTRL                                                             0x01ba
 #define mmMP0_MISC_LIGHT_SLEEP_CTRL_BASE_IDX                                                    0
 
-/* for Vega20 register name change */
-#define mmHDP_MEM_POWER_CTRL	0x00d4
-#define HDP_MEM_POWER_CTRL__IPH_MEM_POWER_CTRL_EN_MASK	0x00000001L
-#define HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK	0x00000002L
-#define HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK	0x00010000L
-#define HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK		0x00020000L
-#define mmHDP_MEM_POWER_CTRL_BASE_IDX	0
-
 /*
  * Indirect registers accessor
  */
@@ -241,6 +233,8 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev)
 {
 	u32 reference_clock = adev->clock.spll.reference_freq;
 
+	if (adev->asic_type == CHIP_RENOIR)
+		return 10000;
 	if (adev->asic_type == CHIP_RAVEN)
 		return reference_clock / 4;
 
@@ -487,7 +481,8 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 
 	if (amdgpu_reset_method == AMD_RESET_METHOD_MODE1 ||
 	    amdgpu_reset_method == AMD_RESET_METHOD_MODE2 ||
-		amdgpu_reset_method == AMD_RESET_METHOD_BACO)
+	    amdgpu_reset_method == AMD_RESET_METHOD_BACO ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_PCI)
 		return amdgpu_reset_method;
 
 	if (amdgpu_reset_method != -1)
@@ -532,15 +527,18 @@ static int soc15_asic_reset(struct amdgpu_device *adev)
 		return 0;
 
 	switch (soc15_asic_reset_method(adev)) {
-		case AMD_RESET_METHOD_BACO:
-			dev_info(adev->dev, "BACO reset\n");
-			return soc15_asic_baco_reset(adev);
-		case AMD_RESET_METHOD_MODE2:
-			dev_info(adev->dev, "MODE2 reset\n");
-			return amdgpu_dpm_mode2_reset(adev);
-		default:
-			dev_info(adev->dev, "MODE1 reset\n");
-			return soc15_asic_mode1_reset(adev);
+	case AMD_RESET_METHOD_PCI:
+		dev_info(adev->dev, "PCI reset\n");
+		return amdgpu_device_pci_reset(adev);
+	case AMD_RESET_METHOD_BACO:
+		dev_info(adev->dev, "BACO reset\n");
+		return soc15_asic_baco_reset(adev);
+	case AMD_RESET_METHOD_MODE2:
+		dev_info(adev->dev, "MODE2 reset\n");
+		return amdgpu_dpm_mode2_reset(adev);
+	default:
+		dev_info(adev->dev, "MODE1 reset\n");
+		return soc15_asic_mode1_reset(adev);
 	}
 }
 
@@ -699,6 +697,7 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 		adev->nbio.funcs = &nbio_v6_1_funcs;
 		adev->nbio.hdp_flush_reg = &nbio_v6_1_hdp_flush_reg;
 	}
+	adev->hdp.funcs = &hdp_v4_0_funcs;
 
 	if (adev->asic_type == CHIP_VEGA20 || adev->asic_type == CHIP_ARCTURUS)
 		adev->df.funcs = &df_v3_6_funcs;
@@ -729,12 +728,12 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 					amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block);
 			}
 			if (adev->asic_type == CHIP_VEGA20)
-				amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+				amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block);
 			else
 				amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block);
 		} else {
 			if (adev->asic_type == CHIP_VEGA20)
-				amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+				amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block);
 			else
 				amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block);
 			if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) {
@@ -787,9 +786,9 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 		if (amdgpu_sriov_vf(adev)) {
 			if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
 				amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
-			amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+			amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block);
 		} else {
-			amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+			amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block);
 			if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
 				amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
 		}
@@ -834,35 +833,12 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 	return 0;
 }
 
-static void soc15_flush_hdp(struct amdgpu_device *adev, struct amdgpu_ring *ring)
-{
-	adev->nbio.funcs->hdp_flush(adev, ring);
-}
-
-static void soc15_invalidate_hdp(struct amdgpu_device *adev,
-				 struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
-	else
-		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
-			HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
-}
-
 static bool soc15_need_full_reset(struct amdgpu_device *adev)
 {
 	/* change this when we implement soft reset */
 	return true;
 }
 
-static void vega20_reset_hdp_ras_error_count(struct amdgpu_device *adev)
-{
-	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__HDP))
-		return;
-	/*read back hdp ras counter to reset it to 0 */
-	RREG32_SOC15(HDP, 0, mmHDP_EDC_CNT);
-}
-
 static void soc15_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0,
 				 uint64_t *count1)
 {
@@ -1011,8 +987,6 @@ static const struct amdgpu_asic_funcs soc15_asic_funcs =
 	.set_uvd_clocks = &soc15_set_uvd_clocks,
 	.set_vce_clocks = &soc15_set_vce_clocks,
 	.get_config_memsize = &soc15_get_config_memsize,
-	.flush_hdp = &soc15_flush_hdp,
-	.invalidate_hdp = &soc15_invalidate_hdp,
 	.need_full_reset = &soc15_need_full_reset,
 	.init_doorbell_index = &vega10_doorbell_index_init,
 	.get_pcie_usage = &soc15_get_pcie_usage,
@@ -1034,9 +1008,6 @@ static const struct amdgpu_asic_funcs vega20_asic_funcs =
 	.set_uvd_clocks = &soc15_set_uvd_clocks,
 	.set_vce_clocks = &soc15_set_vce_clocks,
 	.get_config_memsize = &soc15_get_config_memsize,
-	.flush_hdp = &soc15_flush_hdp,
-	.invalidate_hdp = &soc15_invalidate_hdp,
-	.reset_hdp_ras_error_count = &vega20_reset_hdp_ras_error_count,
 	.need_full_reset = &soc15_need_full_reset,
 	.init_doorbell_index = &vega20_doorbell_index_init,
 	.get_pcie_usage = &vega20_get_pcie_usage,
@@ -1239,7 +1210,8 @@ static int soc15_common_early_init(void *handle)
 		break;
 	case CHIP_RENOIR:
 		adev->asic_funcs = &soc15_asic_funcs;
-		if (adev->pdev->device == 0x1636)
+		if ((adev->pdev->device == 0x1636) ||
+		    (adev->pdev->device == 0x164c))
 			adev->apu_flags |= AMD_APU_IS_RENOIR;
 		else
 			adev->apu_flags |= AMD_APU_IS_GREEN_SARDINE;
@@ -1293,9 +1265,8 @@ static int soc15_common_late_init(void *handle)
 	if (amdgpu_sriov_vf(adev))
 		xgpu_ai_mailbox_get_irq(adev);
 
-	if (adev->asic_funcs &&
-	    adev->asic_funcs->reset_hdp_ras_error_count)
-		adev->asic_funcs->reset_hdp_ras_error_count(adev);
+	if (adev->hdp.funcs->reset_ras_error_count)
+		adev->hdp.funcs->reset_ras_error_count(adev);
 
 	if (adev->nbio.funcs->ras_late_init)
 		r = adev->nbio.funcs->ras_late_init(adev);
@@ -1421,41 +1392,6 @@ static int soc15_common_soft_reset(void *handle)
 	return 0;
 }
 
-static void soc15_update_hdp_light_sleep(struct amdgpu_device *adev, bool enable)
-{
-	uint32_t def, data;
-
-	if (adev->asic_type == CHIP_VEGA20 ||
-		adev->asic_type == CHIP_ARCTURUS ||
-		adev->asic_type == CHIP_RENOIR) {
-		def = data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_CTRL));
-
-		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS))
-			data |= HDP_MEM_POWER_CTRL__IPH_MEM_POWER_CTRL_EN_MASK |
-				HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK |
-				HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK |
-				HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK;
-		else
-			data &= ~(HDP_MEM_POWER_CTRL__IPH_MEM_POWER_CTRL_EN_MASK |
-				HDP_MEM_POWER_CTRL__IPH_MEM_POWER_LS_EN_MASK |
-				HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK |
-				HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK);
-
-		if (def != data)
-			WREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_CTRL), data);
-	} else {
-		def = data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS));
-
-		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS))
-			data |= HDP_MEM_POWER_LS__LS_ENABLE_MASK;
-		else
-			data &= ~HDP_MEM_POWER_LS__LS_ENABLE_MASK;
-
-		if (def != data)
-			WREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS), data);
-	}
-}
-
 static void soc15_update_drm_clock_gating(struct amdgpu_device *adev, bool enable)
 {
 	uint32_t def, data;
@@ -1516,7 +1452,7 @@ static int soc15_common_set_clockgating_state(void *handle,
 				state == AMD_CG_STATE_GATE);
 		adev->nbio.funcs->update_medium_grain_light_sleep(adev,
 				state == AMD_CG_STATE_GATE);
-		soc15_update_hdp_light_sleep(adev,
+		adev->hdp.funcs->update_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
 		soc15_update_drm_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
@@ -1533,7 +1469,7 @@ static int soc15_common_set_clockgating_state(void *handle,
 				state == AMD_CG_STATE_GATE);
 		adev->nbio.funcs->update_medium_grain_light_sleep(adev,
 				state == AMD_CG_STATE_GATE);
-		soc15_update_hdp_light_sleep(adev,
+		adev->hdp.funcs->update_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
 		soc15_update_drm_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
@@ -1541,7 +1477,7 @@ static int soc15_common_set_clockgating_state(void *handle,
 				state == AMD_CG_STATE_GATE);
 		break;
 	case CHIP_ARCTURUS:
-		soc15_update_hdp_light_sleep(adev,
+		adev->hdp.funcs->update_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
 		break;
 	default:
@@ -1560,10 +1496,7 @@ static void soc15_common_get_clockgating_state(void *handle, u32 *flags)
 
 	adev->nbio.funcs->get_clockgating_state(adev, flags);
 
-	/* AMD_CG_SUPPORT_HDP_LS */
-	data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS));
-	if (data & HDP_MEM_POWER_LS__LS_ENABLE_MASK)
-		*flags |= AMD_CG_SUPPORT_HDP_LS;
+	adev->hdp.funcs->get_clock_gating_state(adev, flags);
 
 	/* AMD_CG_SUPPORT_DRM_MGCG */
 	data = RREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_CGTT_CTRL0));
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h b/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
new file mode 100644
index 000000000000..5039375bb1d4
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
@@ -0,0 +1,154 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _TA_SECUREDISPLAY_IF_H
+#define _TA_SECUREDISPLAY_IF_H
+
+/** Secure Display related enumerations */
+/**********************************************************/
+
+/** @enum ta_securedisplay_command
+ *    Secure Display Command ID
+ */
+enum ta_securedisplay_command {
+	/* Query whether TA is responding used only for validation purpose */
+	TA_SECUREDISPLAY_COMMAND__QUERY_TA              = 1,
+	/* Send region of Interest and CRC value to I2C */
+	TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC          = 2,
+	/* Maximum Command ID */
+	TA_SECUREDISPLAY_COMMAND__MAX_ID                = 0x7FFFFFFF,
+};
+
+/** @enum ta_securedisplay_status
+ *    Secure Display status returns in shared buffer status
+ */
+enum ta_securedisplay_status {
+	TA_SECUREDISPLAY_STATUS__SUCCESS                 = 0x00,         /* Success */
+	TA_SECUREDISPLAY_STATUS__GENERIC_FAILURE         = 0x01,         /* Generic Failure */
+	TA_SECUREDISPLAY_STATUS__INVALID_PARAMETER       = 0x02,         /* Invalid Parameter */
+	TA_SECUREDISPLAY_STATUS__NULL_POINTER            = 0x03,         /* Null Pointer*/
+	TA_SECUREDISPLAY_STATUS__I2C_WRITE_ERROR         = 0x04,         /* Fail to Write to I2C */
+	TA_SECUREDISPLAY_STATUS__READ_DIO_SCRATCH_ERROR  = 0x05, /*Fail Read DIO Scratch Register*/
+	TA_SECUREDISPLAY_STATUS__READ_CRC_ERROR          = 0x06,         /* Fail to Read CRC*/
+
+	TA_SECUREDISPLAY_STATUS__MAX                     = 0x7FFFFFFF,/* Maximum Value for status*/
+};
+
+/** @enum ta_securedisplay_max_phy
+ *    Physical ID number to use for reading corresponding DIO Scratch register for ROI
+ */
+enum  ta_securedisplay_max_phy {
+	TA_SECUREDISPLAY_PHY0                           = 0,
+	TA_SECUREDISPLAY_PHY1                           = 1,
+	TA_SECUREDISPLAY_PHY2                           = 2,
+	TA_SECUREDISPLAY_PHY3                           = 3,
+	TA_SECUREDISPLAY_MAX_PHY                        = 4,
+};
+
+/** @enum ta_securedisplay_ta_query_cmd_ret
+ *    A predefined specific reteurn value which is 0xAB only used to validate
+ *    communication to Secure Display TA is functional.
+ *    This value is used to validate whether TA is responding successfully
+ */
+enum ta_securedisplay_ta_query_cmd_ret {
+	/* This is a value to validate if TA is loaded successfully */
+	TA_SECUREDISPLAY_QUERY_CMD_RET                 = 0xAB,
+};
+
+/** @enum ta_securedisplay_buffer_size
+ *    I2C Buffer size which contains 8 bytes of ROI  (X start, X end, Y start, Y end)
+ *    and 6 bytes of CRC( R,G,B) and 1  byte for physical ID
+ */
+enum ta_securedisplay_buffer_size {
+	/* 15 bytes = 8 byte (ROI) + 6 byte(CRC) + 1 byte(phy_id) */
+	TA_SECUREDISPLAY_I2C_BUFFER_SIZE                = 15,
+};
+
+/** Input/output structures for Secure Display commands */
+/**********************************************************/
+/**
+ * Input structures
+ */
+
+/** @struct ta_securedisplay_send_roi_crc_input
+ *    Physical ID to determine which DIO scratch register should be used to get ROI
+ */
+struct ta_securedisplay_send_roi_crc_input {
+	uint32_t  phy_id;  /* Physical ID */
+};
+
+/** @union ta_securedisplay_cmd_input
+ *    Input buffer
+ */
+union ta_securedisplay_cmd_input {
+	/* send ROI and CRC input buffer format */
+	struct ta_securedisplay_send_roi_crc_input        send_roi_crc;
+	uint32_t                                          reserved[4];
+};
+
+/**
+ * Output structures
+ */
+
+/** @struct ta_securedisplay_query_ta_output
+ *  Output buffer format for query TA whether TA is responding used only for validation purpose
+ */
+struct ta_securedisplay_query_ta_output {
+	/* return value from TA when it is queried for validation purpose only */
+	uint32_t  query_cmd_ret;
+};
+
+/** @struct ta_securedisplay_send_roi_crc_output
+ *  Output buffer format for send ROI CRC command which will pass I2c buffer created inside TA
+ *  and used to write to I2C used only for validation purpose
+ */
+struct ta_securedisplay_send_roi_crc_output {
+	uint8_t  i2c_buf[TA_SECUREDISPLAY_I2C_BUFFER_SIZE];  /* I2C buffer */
+	uint8_t  reserved;
+};
+
+/** @union ta_securedisplay_cmd_output
+ *    Output buffer
+ */
+union ta_securedisplay_cmd_output {
+	/* Query TA output buffer format used only for validation purpose*/
+	struct ta_securedisplay_query_ta_output            query_ta;
+	/* Send ROI CRC output buffer format used only for validation purpose */
+	struct ta_securedisplay_send_roi_crc_output        send_roi_crc;
+	uint32_t                                           reserved[4];
+};
+
+/** @struct securedisplay_cmd
+ *    Secure Display Command which is shared buffer memory
+ */
+struct securedisplay_cmd {
+	uint32_t                             cmd_id;                    /* +0  Bytes Command ID */
+	enum ta_securedisplay_status         status;     /* +4  Bytes Status of Secure Display TA */
+	uint32_t                             reserved[2];               /* +8  Bytes Reserved */
+	union ta_securedisplay_cmd_input     securedisplay_in_message;  /* +16 Bytes Input Buffer */
+	union ta_securedisplay_cmd_output    securedisplay_out_message;/* +32 Bytes Output Buffer */
+	/**@note Total 48 Bytes */
+};
+
+#endif   //_TA_SECUREDISPLAY_IF_H
+
diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
index ce3319993b4b..249fcbee7871 100644
--- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
@@ -196,19 +196,30 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
-	if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) {
-		wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
-		/* When a ring buffer overflow happen start parsing interrupt
-		 * from the last not overwritten vector (wptr + 16). Hopefully
-		 * this should allow us to catchup.
-		 */
-		dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
-			 wptr, ih->rptr, (wptr + 16) & ih->ptr_mask);
-		ih->rptr = (wptr + 16) & ih->ptr_mask;
-		tmp = RREG32(mmIH_RB_CNTL);
-		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
-		WREG32(mmIH_RB_CNTL, tmp);
-	}
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	/* Double check that the overflow wasn't already cleared. */
+	wptr = RREG32(mmIH_RB_WPTR);
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 16). Hopefully
+	 * this should allow us to catchup.
+	 */
+
+	dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+		wptr, ih->rptr, (wptr + 16) & ih->ptr_mask);
+	ih->rptr = (wptr + 16) & ih->ptr_mask;
+	tmp = RREG32(mmIH_RB_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32(mmIH_RB_CNTL, tmp);
+
+out:
 	return (wptr & ih->ptr_mask);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index 312ecf6d24a0..7cd67cb2ac5f 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -36,7 +36,6 @@
 #include "vce/vce_4_0_default.h"
 #include "vce/vce_4_0_sh_mask.h"
 #include "nbif/nbif_6_1_offset.h"
-#include "hdp/hdp_4_0_offset.h"
 #include "mmhub/mmhub_1_0_offset.h"
 #include "mmhub/mmhub_1_0_sh_mask.h"
 #include "ivsrcid/uvd/irqsrcs_uvd_7_0.h"
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index c734e31a9e65..6117931fa8d7 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -32,7 +32,6 @@
 
 #include "vcn/vcn_1_0_offset.h"
 #include "vcn/vcn_1_0_sh_mask.h"
-#include "hdp/hdp_4_0_offset.h"
 #include "mmhub/mmhub_9_1_offset.h"
 #include "mmhub/mmhub_9_1_sh_mask.h"
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
index e5ae31eb744e..88626d83e07b 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
@@ -38,132 +38,120 @@
 static void vega10_ih_set_interrupt_funcs(struct amdgpu_device *adev);
 
 /**
- * vega10_ih_enable_interrupts - Enable the interrupt ring buffer
+ * vega10_ih_init_register_offset - Initialize register offset for ih rings
  *
  * @adev: amdgpu_device pointer
  *
- * Enable the interrupt ring buffer (VEGA10).
+ * Initialize register offset ih rings (VEGA10).
  */
-static void vega10_ih_enable_interrupts(struct amdgpu_device *adev)
+static void vega10_ih_init_register_offset(struct amdgpu_device *adev)
 {
-	u32 ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL);
-
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_ENABLE, 1);
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 1);
-	if (amdgpu_sriov_vf(adev)) {
-		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL, ih_rb_cntl)) {
-			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
-			return;
-		}
-	} else {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (adev->irq.ih.ring_size) {
+		ih_regs = &adev->irq.ih.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR);
+		ih_regs->ih_rb_wptr_addr_lo = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO);
+		ih_regs->ih_rb_wptr_addr_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL;
 	}
-	adev->irq.ih.enabled = true;
 
 	if (adev->irq.ih1.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
-					   RB_ENABLE, 1);
-		if (amdgpu_sriov_vf(adev)) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING1 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl);
-		}
-		adev->irq.ih1.enabled = true;
+		ih_regs = &adev->irq.ih1.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_RING1);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI_RING1);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING1);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING1);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING1);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING1);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING1;
 	}
 
 	if (adev->irq.ih2.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING2,
-					   RB_ENABLE, 1);
-		if (amdgpu_sriov_vf(adev)) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING2,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING2 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl);
-		}
-		adev->irq.ih2.enabled = true;
+		ih_regs = &adev->irq.ih2.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_RING2);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI_RING2);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING2);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING2);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING2);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING2);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING2;
 	}
-
-	if (adev->irq.ih_soft.ring_size)
-		adev->irq.ih_soft.enabled = true;
 }
 
 /**
- * vega10_ih_disable_interrupts - Disable the interrupt ring buffer
+ * vega10_ih_toggle_ring_interrupts - toggle the interrupt ring buffer
  *
  * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointet
+ * @enable: true - enable the interrupts, false - disable the interrupts
  *
- * Disable the interrupt ring buffer (VEGA10).
+ * Toggle the interrupt ring buffer (VEGA10)
  */
-static void vega10_ih_disable_interrupts(struct amdgpu_device *adev)
+static int vega10_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
+					    struct amdgpu_ih_ring *ih,
+					    bool enable)
 {
-	u32 ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL);
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
 
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_ENABLE, 0);
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 0);
+	ih_regs = &ih->ih_regs;
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+	/* enable_intr field is only valid in ring0 */
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
 	if (amdgpu_sriov_vf(adev)) {
-		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL, ih_rb_cntl)) {
-			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
-			return;
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			dev_err(adev->dev, "PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
 		}
 	} else {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
+		WREG32(ih_regs->ih_rb_cntl, tmp);
 	}
 
-	/* set rptr, wptr to 0 */
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0);
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0);
-	adev->irq.ih.enabled = false;
-	adev->irq.ih.rptr = 0;
-
-	if (adev->irq.ih1.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
-					   RB_ENABLE, 0);
-		if (amdgpu_sriov_vf(adev)) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING1 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl);
-		}
+	if (enable) {
+		ih->enabled = true;
+	} else {
 		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING1, 0);
-		adev->irq.ih1.enabled = false;
-		adev->irq.ih1.rptr = 0;
+		WREG32(ih_regs->ih_rb_rptr, 0);
+		WREG32(ih_regs->ih_rb_wptr, 0);
+		ih->enabled = false;
+		ih->rptr = 0;
 	}
 
-	if (adev->irq.ih2.ring_size) {
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING2,
-					   RB_ENABLE, 0);
-		if (amdgpu_sriov_vf(adev)) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING2,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING2 failed!\n");
-				return;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl);
-		}
+	return 0;
+}
 
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING2, 0);
-		adev->irq.ih2.enabled = false;
-		adev->irq.ih2.rptr = 0;
+/**
+ * vega10_ih_toggle_interrupts - Toggle all the available interrupt ring buffers
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable or disable interrupt ring buffers
+ *
+ * Toggle all the available interrupt ring buffers (VEGA10).
+ */
+static int vega10_ih_toggle_interrupts(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
+	int i;
+	int r;
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			r = vega10_ih_toggle_ring_interrupts(adev, ih[i], enable);
+			if (r)
+				return r;
+		}
 	}
+
+	return 0;
 }
 
 static uint32_t vega10_ih_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl)
@@ -209,6 +197,58 @@ static uint32_t vega10_ih_doorbell_rptr(struct amdgpu_ih_ring *ih)
 }
 
 /**
+ * vega10_ih_enable_ring - enable an ih ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Enable an ih ring buffer (VEGA10)
+ */
+static int vega10_ih_enable_ring(struct amdgpu_device *adev,
+				 struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
+	WREG32(ih_regs->ih_rb_base, ih->gpu_addr >> 8);
+	WREG32(ih_regs->ih_rb_base_hi, (ih->gpu_addr >> 40) & 0xff);
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = vega10_ih_rb_cntl(ih, tmp);
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
+	if (ih == &adev->irq.ih1) {
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
+	}
+	if (amdgpu_sriov_vf(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			dev_err(adev->dev, "PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (ih == &adev->irq.ih) {
+		/* set the ih ring 0 writeback address whether it's enabled or not */
+		WREG32(ih_regs->ih_rb_wptr_addr_lo, lower_32_bits(ih->wptr_addr));
+		WREG32(ih_regs->ih_rb_wptr_addr_hi, upper_32_bits(ih->wptr_addr) & 0xFFFF);
+	}
+
+	/* set rptr, wptr to 0 */
+	WREG32(ih_regs->ih_rb_wptr, 0);
+	WREG32(ih_regs->ih_rb_rptr, 0);
+
+	WREG32(ih_regs->ih_doorbell_rptr, vega10_ih_doorbell_rptr(ih));
+
+	return 0;
+}
+
+/**
  * vega10_ih_irq_init - init and enable the interrupt ring
  *
  * @adev: amdgpu_device pointer
@@ -221,116 +261,34 @@ static uint32_t vega10_ih_doorbell_rptr(struct amdgpu_ih_ring *ih)
  */
 static int vega10_ih_irq_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_ih_ring *ih;
-	u32 ih_rb_cntl, ih_chicken;
-	int ret = 0;
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
+	u32 ih_chicken;
+	int ret;
+	int i;
 	u32 tmp;
 
 	/* disable irqs */
-	vega10_ih_disable_interrupts(adev);
+	ret = vega10_ih_toggle_interrupts(adev, false);
+	if (ret)
+		return ret;
 
 	adev->nbio.funcs->ih_control(adev);
 
-	ih = &adev->irq.ih;
-	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, ih->gpu_addr >> 8);
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, (ih->gpu_addr >> 40) & 0xff);
-
-	ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL);
-	ih_rb_cntl = vega10_ih_rb_cntl(ih, ih_rb_cntl);
-	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RPTR_REARM,
-				   !!adev->irq.msi_enabled);
-	if (amdgpu_sriov_vf(adev)) {
-		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL, ih_rb_cntl)) {
-			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
-			return -ETIMEDOUT;
-		}
-	} else {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
-	}
-
-	if ((adev->asic_type == CHIP_ARCTURUS &&
-	     adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) ||
-	    adev->asic_type == CHIP_RENOIR) {
+	if (adev->asic_type == CHIP_RENOIR) {
 		ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN);
 		if (adev->irq.ih.use_bus_addr) {
 			ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
 						   MC_SPACE_GPA_ENABLE, 1);
-		} else {
-			ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
-						   MC_SPACE_FBPA_ENABLE, 1);
 		}
 		WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN, ih_chicken);
 	}
 
-	/* set the writeback address whether it's enabled or not */
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO,
-		     lower_32_bits(ih->wptr_addr));
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI,
-		     upper_32_bits(ih->wptr_addr) & 0xFFFF);
-
-	/* set rptr, wptr to 0 */
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0);
-	WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0);
-
-	WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR,
-		     vega10_ih_doorbell_rptr(ih));
-
-	ih = &adev->irq.ih1;
-	if (ih->ring_size) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_RING1, ih->gpu_addr >> 8);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI_RING1,
-			     (ih->gpu_addr >> 40) & 0xff);
-
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-		ih_rb_cntl = vega10_ih_rb_cntl(ih, ih_rb_cntl);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
-					   WPTR_OVERFLOW_ENABLE, 0);
-		ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
-					   RB_FULL_DRAIN_ENABLE, 1);
-		if (amdgpu_sriov_vf(adev)) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING1 failed!\n");
-				return -ETIMEDOUT;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl);
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			ret = vega10_ih_enable_ring(adev, ih[i]);
+			if (ret)
+				return ret;
 		}
-
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING1, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, 0);
-
-		WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING1,
-			     vega10_ih_doorbell_rptr(ih));
-	}
-
-	ih = &adev->irq.ih2;
-	if (ih->ring_size) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_RING2, ih->gpu_addr >> 8);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI_RING2,
-			     (ih->gpu_addr >> 40) & 0xff);
-
-		ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-		ih_rb_cntl = vega10_ih_rb_cntl(ih, ih_rb_cntl);
-
-		if (amdgpu_sriov_vf(adev)) {
-			if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING2,
-						ih_rb_cntl)) {
-				DRM_ERROR("program IH_RB_CNTL_RING2 failed!\n");
-				return -ETIMEDOUT;
-			}
-		} else {
-			WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl);
-		}
-
-		/* set rptr, wptr to 0 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING2, 0);
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, 0);
-
-		WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING2,
-			     vega10_ih_doorbell_rptr(ih));
 	}
 
 	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL);
@@ -345,9 +303,14 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev)
 	pci_set_master(adev->pdev);
 
 	/* enable interrupts */
-	vega10_ih_enable_interrupts(adev);
+	ret = vega10_ih_toggle_interrupts(adev, true);
+	if (ret)
+		return ret;
 
-	return ret;
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
+	return 0;
 }
 
 /**
@@ -359,7 +322,7 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev)
  */
 static void vega10_ih_irq_disable(struct amdgpu_device *adev)
 {
-	vega10_ih_disable_interrupts(adev);
+	vega10_ih_toggle_interrupts(adev, false);
 
 	/* Wait and acknowledge irq */
 	mdelay(1);
@@ -379,25 +342,17 @@ static void vega10_ih_irq_disable(struct amdgpu_device *adev)
 static u32 vega10_ih_get_wptr(struct amdgpu_device *adev,
 			      struct amdgpu_ih_ring *ih)
 {
-	u32 wptr, reg, tmp;
+	u32 wptr, tmp;
+	struct amdgpu_ih_regs *ih_regs;
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
+	ih_regs = &ih->ih_regs;
 
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
 	/* Double check that the overflow wasn't already cleared. */
-
-	if (ih == &adev->irq.ih)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR);
-	else if (ih == &adev->irq.ih1)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING1);
-	else if (ih == &adev->irq.ih2)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING2);
-	else
-		BUG();
-
-	wptr = RREG32_NO_KIQ(reg);
+	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
@@ -413,69 +368,15 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev,
 		 wptr, ih->rptr, tmp);
 	ih->rptr = tmp;
 
-	if (ih == &adev->irq.ih)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL);
-	else if (ih == &adev->irq.ih1)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING1);
-	else if (ih == &adev->irq.ih2)
-		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING2);
-	else
-		BUG();
-
-	tmp = RREG32_NO_KIQ(reg);
+	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
-	WREG32_NO_KIQ(reg, tmp);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
 
 out:
 	return (wptr & ih->ptr_mask);
 }
 
 /**
- * vega10_ih_decode_iv - decode an interrupt vector
- *
- * @adev: amdgpu_device pointer
- * @ih: IH ring buffer to decode
- * @entry: IV entry to place decoded information into
- *
- * Decodes the interrupt vector at the current rptr
- * position and also advance the position.
- */
-static void vega10_ih_decode_iv(struct amdgpu_device *adev,
-				struct amdgpu_ih_ring *ih,
-				struct amdgpu_iv_entry *entry)
-{
-	/* wptr/rptr are in bytes! */
-	u32 ring_index = ih->rptr >> 2;
-	uint32_t dw[8];
-
-	dw[0] = le32_to_cpu(ih->ring[ring_index + 0]);
-	dw[1] = le32_to_cpu(ih->ring[ring_index + 1]);
-	dw[2] = le32_to_cpu(ih->ring[ring_index + 2]);
-	dw[3] = le32_to_cpu(ih->ring[ring_index + 3]);
-	dw[4] = le32_to_cpu(ih->ring[ring_index + 4]);
-	dw[5] = le32_to_cpu(ih->ring[ring_index + 5]);
-	dw[6] = le32_to_cpu(ih->ring[ring_index + 6]);
-	dw[7] = le32_to_cpu(ih->ring[ring_index + 7]);
-
-	entry->client_id = dw[0] & 0xff;
-	entry->src_id = (dw[0] >> 8) & 0xff;
-	entry->ring_id = (dw[0] >> 16) & 0xff;
-	entry->vmid = (dw[0] >> 24) & 0xf;
-	entry->vmid_src = (dw[0] >> 31);
-	entry->timestamp = dw[1] | ((u64)(dw[2] & 0xffff) << 32);
-	entry->timestamp_src = dw[2] >> 31;
-	entry->pasid = dw[3] & 0xffff;
-	entry->pasid_src = dw[3] >> 31;
-	entry->src_data[0] = dw[4];
-	entry->src_data[1] = dw[5];
-	entry->src_data[2] = dw[6];
-	entry->src_data[3] = dw[7];
-
-	/* wptr/rptr are in bytes! */
-	ih->rptr += 32;
-}
-
-/**
  * vega10_ih_irq_rearm - rearm IRQ if lost
  *
  * @adev: amdgpu_device pointer
@@ -485,22 +386,14 @@ static void vega10_ih_decode_iv(struct amdgpu_device *adev,
 static void vega10_ih_irq_rearm(struct amdgpu_device *adev,
 			       struct amdgpu_ih_ring *ih)
 {
-	uint32_t reg_rptr = 0;
 	uint32_t v = 0;
 	uint32_t i = 0;
+	struct amdgpu_ih_regs *ih_regs;
 
-	if (ih == &adev->irq.ih)
-		reg_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR);
-	else if (ih == &adev->irq.ih1)
-		reg_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING1);
-	else if (ih == &adev->irq.ih2)
-		reg_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING2);
-	else
-		return;
-
+	ih_regs = &ih->ih_regs;
 	/* Rearm IRQ / re-wwrite doorbell if doorbell write is lost */
 	for (i = 0; i < MAX_REARM_RETRY; i++) {
-		v = RREG32_NO_KIQ(reg_rptr);
+		v = RREG32_NO_KIQ(ih_regs->ih_rb_rptr);
 		if ((v < ih->ring_size) && (v != ih->rptr))
 			WDOORBELL32(ih->doorbell_index, ih->rptr);
 		else
@@ -519,6 +412,8 @@ static void vega10_ih_irq_rearm(struct amdgpu_device *adev,
 static void vega10_ih_set_rptr(struct amdgpu_device *adev,
 			       struct amdgpu_ih_ring *ih)
 {
+	struct amdgpu_ih_regs *ih_regs;
+
 	if (ih->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
 		*ih->rptr_cpu = ih->rptr;
@@ -526,12 +421,9 @@ static void vega10_ih_set_rptr(struct amdgpu_device *adev,
 
 		if (amdgpu_sriov_vf(adev))
 			vega10_ih_irq_rearm(adev, ih);
-	} else if (ih == &adev->irq.ih) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, ih->rptr);
-	} else if (ih == &adev->irq.ih1) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, ih->rptr);
-	} else if (ih == &adev->irq.ih2) {
-		WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, ih->rptr);
+	} else {
+		ih_regs = &ih->ih_regs;
+		WREG32(ih_regs->ih_rb_rptr, ih->rptr);
 	}
 }
 
@@ -600,19 +492,23 @@ static int vega10_ih_sw_init(void *handle)
 	adev->irq.ih.use_doorbell = true;
 	adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1;
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true);
-	if (r)
-		return r;
+	if (!(adev->flags & AMD_IS_APU)) {
+		r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true);
+		if (r)
+			return r;
 
-	adev->irq.ih1.use_doorbell = true;
-	adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
+		adev->irq.ih1.use_doorbell = true;
+		adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true);
-	if (r)
-		return r;
+		r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true);
+		if (r)
+			return r;
 
-	adev->irq.ih2.use_doorbell = true;
-	adev->irq.ih2.doorbell_index = (adev->doorbell_index.ih + 2) << 1;
+		adev->irq.ih2.use_doorbell = true;
+		adev->irq.ih2.doorbell_index = (adev->doorbell_index.ih + 2) << 1;
+	}
+	/* initialize ih control registers offset */
+	vega10_ih_init_register_offset(adev);
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
 	if (r)
@@ -628,6 +524,7 @@ static int vega10_ih_sw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	amdgpu_irq_fini(adev);
+	amdgpu_ih_ring_fini(adev, &adev->irq.ih_soft);
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih2);
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih1);
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih);
@@ -698,15 +595,11 @@ static void vega10_ih_update_clockgating_state(struct amdgpu_device *adev,
 		def = data = RREG32_SOC15(OSSSYS, 0, mmIH_CLK_CTRL);
 		field_val = enable ? 0 : 1;
 		/**
-		 * Vega10 does not have IH_RETRY_INT_CAM_MEM_CLK_SOFT_OVERRIDE
-		 * and IH_BUFFER_MEM_CLK_SOFT_OVERRIDE field.
+		 * Vega10/12 and RAVEN don't have IH_BUFFER_MEM_CLK_SOFT_OVERRIDE field.
 		 */
-		if (adev->asic_type > CHIP_VEGA10) {
-			data = REG_SET_FIELD(data, IH_CLK_CTRL,
-				     IH_RETRY_INT_CAM_MEM_CLK_SOFT_OVERRIDE, field_val);
+		if (adev->asic_type == CHIP_RENOIR)
 			data = REG_SET_FIELD(data, IH_CLK_CTRL,
 				     IH_BUFFER_MEM_CLK_SOFT_OVERRIDE, field_val);
-		}
 
 		data = REG_SET_FIELD(data, IH_CLK_CTRL,
 				     DBUS_MUX_CLK_SOFT_OVERRIDE, field_val);
@@ -759,7 +652,7 @@ const struct amd_ip_funcs vega10_ih_ip_funcs = {
 
 static const struct amdgpu_ih_funcs vega10_ih_funcs = {
 	.get_wptr = vega10_ih_get_wptr,
-	.decode_iv = vega10_ih_decode_iv,
+	.decode_iv = amdgpu_ih_decode_iv_helper,
 	.set_rptr = vega10_ih_set_rptr
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
new file mode 100644
index 000000000000..5a3c867d5881
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
@@ -0,0 +1,703 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ih.h"
+#include "soc15.h"
+
+#include "oss/osssys_4_2_0_offset.h"
+#include "oss/osssys_4_2_0_sh_mask.h"
+
+#include "soc15_common.h"
+#include "vega20_ih.h"
+
+#define MAX_REARM_RETRY 10
+
+static void vega20_ih_set_interrupt_funcs(struct amdgpu_device *adev);
+
+/**
+ * vega20_ih_init_register_offset - Initialize register offset for ih rings
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize register offset ih rings (VEGA20).
+ */
+static void vega20_ih_init_register_offset(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (adev->irq.ih.ring_size) {
+		ih_regs = &adev->irq.ih.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR);
+		ih_regs->ih_rb_wptr_addr_lo = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO);
+		ih_regs->ih_rb_wptr_addr_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL;
+	}
+
+	if (adev->irq.ih1.ring_size) {
+		ih_regs = &adev->irq.ih1.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_RING1);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI_RING1);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING1);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING1);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING1);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING1);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING1;
+	}
+
+	if (adev->irq.ih2.ring_size) {
+		ih_regs = &adev->irq.ih2.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_RING2);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI_RING2);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING2);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING2);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR_RING2);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR_RING2);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING2;
+	}
+}
+
+/**
+ * vega20_ih_toggle_ring_interrupts - toggle the interrupt ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ * @enable: true - enable the interrupts, false - disable the interrupts
+ *
+ * Toggle the interrupt ring buffer (VEGA20)
+ */
+static int vega20_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
+					    struct amdgpu_ih_ring *ih,
+					    bool enable)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+	/* enable_intr field is only valid in ring0 */
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
+	if (amdgpu_sriov_vf(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			dev_err(adev->dev, "PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (enable) {
+		ih->enabled = true;
+	} else {
+		/* set rptr, wptr to 0 */
+		WREG32(ih_regs->ih_rb_rptr, 0);
+		WREG32(ih_regs->ih_rb_wptr, 0);
+		ih->enabled = false;
+		ih->rptr = 0;
+	}
+
+	return 0;
+}
+
+/**
+ * vega20_ih_toggle_interrupts - Toggle all the available interrupt ring buffers
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable or disable interrupt ring buffers
+ *
+ * Toggle all the available interrupt ring buffers (VEGA20).
+ */
+static int vega20_ih_toggle_interrupts(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
+	int i;
+	int r;
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			r = vega20_ih_toggle_ring_interrupts(adev, ih[i], enable);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static uint32_t vega20_ih_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl)
+{
+	int rb_bufsz = order_base_2(ih->ring_size / 4);
+
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   MC_SPACE, ih->use_bus_addr ? 1 : 4);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_CLEAR, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_SIZE, rb_bufsz);
+	/* Ring Buffer write pointer writeback. If enabled, IH_RB_WPTR register
+	 * value is written to memory
+	 */
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_WRITEBACK_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SNOOP, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_RO, 0);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_VMID, 0);
+
+	return ih_rb_cntl;
+}
+
+static uint32_t vega20_ih_doorbell_rptr(struct amdgpu_ih_ring *ih)
+{
+	u32 ih_doorbell_rtpr = 0;
+
+	if (ih->use_doorbell) {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR, OFFSET,
+						 ih->doorbell_index);
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 1);
+	} else {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 0);
+	}
+	return ih_doorbell_rtpr;
+}
+
+/**
+ * vega20_ih_enable_ring - enable an ih ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Enable an ih ring buffer (VEGA20)
+ */
+static int vega20_ih_enable_ring(struct amdgpu_device *adev,
+				 struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
+	WREG32(ih_regs->ih_rb_base, ih->gpu_addr >> 8);
+	WREG32(ih_regs->ih_rb_base_hi, (ih->gpu_addr >> 40) & 0xff);
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = vega20_ih_rb_cntl(ih, tmp);
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
+	if (ih == &adev->irq.ih1) {
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
+	}
+	if (amdgpu_sriov_vf(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			dev_err(adev->dev, "PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (ih == &adev->irq.ih) {
+		/* set the ih ring 0 writeback address whether it's enabled or not */
+		WREG32(ih_regs->ih_rb_wptr_addr_lo, lower_32_bits(ih->wptr_addr));
+		WREG32(ih_regs->ih_rb_wptr_addr_hi, upper_32_bits(ih->wptr_addr) & 0xFFFF);
+	}
+
+	/* set rptr, wptr to 0 */
+	WREG32(ih_regs->ih_rb_wptr, 0);
+	WREG32(ih_regs->ih_rb_rptr, 0);
+
+	WREG32(ih_regs->ih_doorbell_rptr, vega20_ih_doorbell_rptr(ih));
+
+	return 0;
+}
+
+/**
+ * vega20_ih_reroute_ih - reroute VMC/UTCL2 ih to an ih ring
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Reroute VMC and UMC interrupts on primary ih ring to
+ * ih ring 1 so they won't lose when bunches of page faults
+ * interrupts overwhelms the interrupt handler(VEGA20)
+ */
+static void vega20_ih_reroute_ih(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* vega20 ih reroute will go through psp
+	 * this function is only used for arcturus
+	 */
+	if (adev->asic_type == CHIP_ARCTURUS) {
+		/* Reroute to IH ring 1 for VMC */
+		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_INDEX, 0x12);
+		tmp = RREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA);
+		tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, CLIENT_TYPE, 1);
+		tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
+		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA, tmp);
+
+		/* Reroute IH ring 1 for UTCL2 */
+		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_INDEX, 0x1B);
+		tmp = RREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA);
+		tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
+		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA, tmp);
+	}
+}
+
+/**
+ * vega20_ih_irq_init - init and enable the interrupt ring
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Allocate a ring buffer for the interrupt controller,
+ * enable the RLC, disable interrupts, enable the IH
+ * ring buffer and enable it (VI).
+ * Called at device load and reume.
+ * Returns 0 for success, errors for failure.
+ */
+static int vega20_ih_irq_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
+	u32 ih_chicken;
+	int ret;
+	int i;
+	u32 tmp;
+
+	/* disable irqs */
+	ret = vega20_ih_toggle_interrupts(adev, false);
+	if (ret)
+		return ret;
+
+	adev->nbio.funcs->ih_control(adev);
+
+	if (adev->asic_type == CHIP_ARCTURUS &&
+	    adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN);
+		if (adev->irq.ih.use_bus_addr) {
+			ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
+						   MC_SPACE_GPA_ENABLE, 1);
+		}
+		WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN, ih_chicken);
+	}
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			if (i == 1)
+				vega20_ih_reroute_ih(adev);
+			ret = vega20_ih_enable_ring(adev, ih[i]);
+			if (ret)
+				return ret;
+		}
+	}
+
+	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL,
+			    CLIENT18_IS_STORM_CLIENT, 1);
+	WREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL, tmp);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1);
+	WREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL, tmp);
+
+	pci_set_master(adev->pdev);
+
+	/* enable interrupts */
+	ret = vega20_ih_toggle_interrupts(adev, true);
+	if (ret)
+		return ret;
+
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
+	return 0;
+}
+
+/**
+ * vega20_ih_irq_disable - disable interrupts
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Disable interrupts on the hw (VEGA20).
+ */
+static void vega20_ih_irq_disable(struct amdgpu_device *adev)
+{
+	vega20_ih_toggle_interrupts(adev, false);
+
+	/* Wait and acknowledge irq */
+	mdelay(1);
+}
+
+/**
+ * vega20_ih_get_wptr - get the IH ring buffer wptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Get the IH ring buffer wptr from either the register
+ * or the writeback memory buffer (VEGA20).  Also check for
+ * ring buffer overflow and deal with it.
+ * Returns the value of the wptr.
+ */
+static u32 vega20_ih_get_wptr(struct amdgpu_device *adev,
+			      struct amdgpu_ih_ring *ih)
+{
+	u32 wptr, tmp;
+	struct amdgpu_ih_regs *ih_regs;
+
+	wptr = le32_to_cpu(*ih->wptr_cpu);
+	ih_regs = &ih->ih_regs;
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	/* Double check that the overflow wasn't already cleared. */
+	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 32). Hopefully
+	 * this should allow us to catchup.
+	 */
+	tmp = (wptr + 32) & ih->ptr_mask;
+	dev_warn(adev->dev, "IH ring buffer overflow "
+		 "(0x%08X, 0x%08X, 0x%08X)\n",
+		 wptr, ih->rptr, tmp);
+	ih->rptr = tmp;
+
+	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
+out:
+	return (wptr & ih->ptr_mask);
+}
+
+/**
+ * vega20_ih_irq_rearm - rearm IRQ if lost
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ */
+static void vega20_ih_irq_rearm(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	uint32_t v = 0;
+	uint32_t i = 0;
+	struct amdgpu_ih_regs *ih_regs;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Rearm IRQ / re-wwrite doorbell if doorbell write is lost */
+	for (i = 0; i < MAX_REARM_RETRY; i++) {
+		v = RREG32_NO_KIQ(ih_regs->ih_rb_rptr);
+		if ((v < ih->ring_size) && (v != ih->rptr))
+			WDOORBELL32(ih->doorbell_index, ih->rptr);
+		else
+			break;
+	}
+}
+
+/**
+ * vega20_ih_set_rptr - set the IH ring buffer rptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Set the IH ring buffer rptr.
+ */
+static void vega20_ih_set_rptr(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (ih->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		*ih->rptr_cpu = ih->rptr;
+		WDOORBELL32(ih->doorbell_index, ih->rptr);
+
+		if (amdgpu_sriov_vf(adev))
+			vega20_ih_irq_rearm(adev, ih);
+	} else {
+		ih_regs = &ih->ih_regs;
+		WREG32(ih_regs->ih_rb_rptr, ih->rptr);
+	}
+}
+
+/**
+ * vega20_ih_self_irq - dispatch work for ring 1 and 2
+ *
+ * @adev: amdgpu_device pointer
+ * @source: irq source
+ * @entry: IV with WPTR update
+ *
+ * Update the WPTR from the IV and schedule work to handle the entries.
+ */
+static int vega20_ih_self_irq(struct amdgpu_device *adev,
+			      struct amdgpu_irq_src *source,
+			      struct amdgpu_iv_entry *entry)
+{
+	uint32_t wptr = cpu_to_le32(entry->src_data[0]);
+
+	switch (entry->ring_id) {
+	case 1:
+		*adev->irq.ih1.wptr_cpu = wptr;
+		schedule_work(&adev->irq.ih1_work);
+		break;
+	case 2:
+		*adev->irq.ih2.wptr_cpu = wptr;
+		schedule_work(&adev->irq.ih2_work);
+		break;
+	default: break;
+	}
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vega20_ih_self_irq_funcs = {
+	.process = vega20_ih_self_irq,
+};
+
+static void vega20_ih_set_self_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.self_irq.num_types = 0;
+	adev->irq.self_irq.funcs = &vega20_ih_self_irq_funcs;
+}
+
+static int vega20_ih_early_init(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	vega20_ih_set_interrupt_funcs(adev);
+	vega20_ih_set_self_irq_funcs(adev);
+	return 0;
+}
+
+static int vega20_ih_sw_init(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int r;
+
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_IH, 0,
+			      &adev->irq.self_irq);
+	if (r)
+		return r;
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, true);
+	if (r)
+		return r;
+
+	adev->irq.ih.use_doorbell = true;
+	adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1;
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true);
+	if (r)
+		return r;
+
+	adev->irq.ih1.use_doorbell = true;
+	adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true);
+	if (r)
+		return r;
+
+	adev->irq.ih2.use_doorbell = true;
+	adev->irq.ih2.doorbell_index = (adev->doorbell_index.ih + 2) << 1;
+
+	/* initialize ih control registers offset */
+	vega20_ih_init_register_offset(adev);
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_init(adev);
+
+	return r;
+}
+
+static int vega20_ih_sw_fini(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	amdgpu_irq_fini(adev);
+	amdgpu_ih_ring_fini(adev, &adev->irq.ih_soft);
+	amdgpu_ih_ring_fini(adev, &adev->irq.ih2);
+	amdgpu_ih_ring_fini(adev, &adev->irq.ih1);
+	amdgpu_ih_ring_fini(adev, &adev->irq.ih);
+
+	return 0;
+}
+
+static int vega20_ih_hw_init(void *handle)
+{
+	int r;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	r = vega20_ih_irq_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int vega20_ih_hw_fini(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	vega20_ih_irq_disable(adev);
+
+	return 0;
+}
+
+static int vega20_ih_suspend(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	return vega20_ih_hw_fini(adev);
+}
+
+static int vega20_ih_resume(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	return vega20_ih_hw_init(adev);
+}
+
+static bool vega20_ih_is_idle(void *handle)
+{
+	/* todo */
+	return true;
+}
+
+static int vega20_ih_wait_for_idle(void *handle)
+{
+	/* todo */
+	return -ETIMEDOUT;
+}
+
+static int vega20_ih_soft_reset(void *handle)
+{
+	/* todo */
+
+	return 0;
+}
+
+static void vega20_ih_update_clockgating_state(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t data, def, field_val;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_IH_CG) {
+		def = data = RREG32_SOC15(OSSSYS, 0, mmIH_CLK_CTRL);
+		field_val = enable ? 0 : 1;
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     IH_RETRY_INT_CAM_MEM_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     IH_BUFFER_MEM_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DBUS_MUX_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     OSSSYS_SHARE_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     LIMIT_SMN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DYN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     REG_CLK_SOFT_OVERRIDE, field_val);
+		if (def != data)
+			WREG32_SOC15(OSSSYS, 0, mmIH_CLK_CTRL, data);
+	}
+}
+
+static int vega20_ih_set_clockgating_state(void *handle,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	vega20_ih_update_clockgating_state(adev,
+				state == AMD_CG_STATE_GATE);
+	return 0;
+
+}
+
+static int vega20_ih_set_powergating_state(void *handle,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+const struct amd_ip_funcs vega20_ih_ip_funcs = {
+	.name = "vega20_ih",
+	.early_init = vega20_ih_early_init,
+	.late_init = NULL,
+	.sw_init = vega20_ih_sw_init,
+	.sw_fini = vega20_ih_sw_fini,
+	.hw_init = vega20_ih_hw_init,
+	.hw_fini = vega20_ih_hw_fini,
+	.suspend = vega20_ih_suspend,
+	.resume = vega20_ih_resume,
+	.is_idle = vega20_ih_is_idle,
+	.wait_for_idle = vega20_ih_wait_for_idle,
+	.soft_reset = vega20_ih_soft_reset,
+	.set_clockgating_state = vega20_ih_set_clockgating_state,
+	.set_powergating_state = vega20_ih_set_powergating_state,
+};
+
+static const struct amdgpu_ih_funcs vega20_ih_funcs = {
+	.get_wptr = vega20_ih_get_wptr,
+	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.set_rptr = vega20_ih_set_rptr
+};
+
+static void vega20_ih_set_interrupt_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.ih_funcs = &vega20_ih_funcs;
+}
+
+const struct amdgpu_ip_block_version vega20_ih_ip_block =
+{
+	.type = AMD_IP_BLOCK_TYPE_IH,
+	.major = 4,
+	.minor = 2,
+	.rev = 0,
+	.funcs = &vega20_ih_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.h b/drivers/gpu/drm/amd/amdgpu/vega20_ih.h
new file mode 100644
index 000000000000..7af6d8758ee3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VEGA20_IH_H__
+#define __VEGA20_IH_H__
+
+extern const struct amd_ip_funcs vega20_ih_ip_funcs;
+extern const struct amdgpu_ip_block_version vega20_ih_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c
index d56b474b3a21..eafb76aebd00 100644
--- a/drivers/gpu/drm/amd/amdgpu/vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/vi.c
@@ -642,11 +642,21 @@ static int vi_read_register(struct amdgpu_device *adev, u32 se_num,
 	return -EINVAL;
 }
 
-static int vi_gpu_pci_config_reset(struct amdgpu_device *adev)
+/**
+ * vi_asic_pci_config_reset - soft reset GPU
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Use PCI Config method to reset the GPU.
+ *
+ * Returns 0 for success.
+ */
+static int vi_asic_pci_config_reset(struct amdgpu_device *adev)
 {
 	u32 i;
+	int r = -EINVAL;
 
-	dev_info(adev->dev, "GPU pci config reset\n");
+	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
 
 	/* disable BM */
 	pci_clear_master(adev->pdev);
@@ -661,29 +671,11 @@ static int vi_gpu_pci_config_reset(struct amdgpu_device *adev)
 			/* enable BM */
 			pci_set_master(adev->pdev);
 			adev->has_hw_reset = true;
-			return 0;
+			r = 0;
+			break;
 		}
 		udelay(1);
 	}
-	return -EINVAL;
-}
-
-/**
- * vi_asic_pci_config_reset - soft reset GPU
- *
- * @adev: amdgpu_device pointer
- *
- * Use PCI Config method to reset the GPU.
- *
- * Returns 0 for success.
- */
-static int vi_asic_pci_config_reset(struct amdgpu_device *adev)
-{
-	int r;
-
-	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
-
-	r = vi_gpu_pci_config_reset(adev);
 
 	amdgpu_atombios_scratch_regs_engine_hung(adev, false);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/Kconfig b/drivers/gpu/drm/amd/amdkfd/Kconfig
index e8fb10c41f16..f02c938f75da 100644
--- a/drivers/gpu/drm/amd/amdkfd/Kconfig
+++ b/drivers/gpu/drm/amd/amdkfd/Kconfig
@@ -7,6 +7,8 @@ config HSA_AMD
 	bool "HSA kernel driver for AMD GPU devices"
 	depends on DRM_AMDGPU && (X86_64 || ARM64 || PPC64)
 	imply AMD_IOMMU_V2 if X86_64
+	select HMM_MIRROR
 	select MMU_NOTIFIER
+	select DRM_AMDGPU_USERPTR
 	help
 	  Enable this if you want to use HSA features on AMD GPU devices.
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
index 8cac497c2c45..a5640a6138cf 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
@@ -1040,11 +1040,14 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size)
 				(struct crat_subtype_iolink *)sub_type_hdr);
 		if (ret < 0)
 			return ret;
-		crat_table->length += (sub_type_hdr->length * entries);
-		crat_table->total_entries += entries;
 
-		sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr +
-				sub_type_hdr->length * entries);
+		if (entries) {
+			crat_table->length += (sub_type_hdr->length * entries);
+			crat_table->total_entries += entries;
+
+			sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr +
+					sub_type_hdr->length * entries);
+		}
 #else
 		pr_info("IO link not available for non x86 platforms\n");
 #endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
index 16262e5d93f5..7351dd195274 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -243,11 +243,11 @@ get_sh_mem_bases_nybble_64(struct kfd_process_device *pdd)
 static inline void dqm_lock(struct device_queue_manager *dqm)
 {
 	mutex_lock(&dqm->lock_hidden);
-	dqm->saved_flags = memalloc_nofs_save();
+	dqm->saved_flags = memalloc_noreclaim_save();
 }
 static inline void dqm_unlock(struct device_queue_manager *dqm)
 {
-	memalloc_nofs_restore(dqm->saved_flags);
+	memalloc_noreclaim_restore(dqm->saved_flags);
 	mutex_unlock(&dqm->lock_hidden);
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
index 241bd6ff79f4..74a460be077b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
@@ -44,6 +44,25 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev,
 	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
 	pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry);
 
+	/* Only handle clients we care about */
+	if (client_id != SOC15_IH_CLIENTID_GRBM_CP &&
+	    client_id != SOC15_IH_CLIENTID_SDMA0 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA1 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA2 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA3 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA4 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA5 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA6 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA7 &&
+	    client_id != SOC15_IH_CLIENTID_VMC &&
+	    client_id != SOC15_IH_CLIENTID_VMC1 &&
+	    client_id != SOC15_IH_CLIENTID_UTCL2 &&
+	    client_id != SOC15_IH_CLIENTID_SE0SH &&
+	    client_id != SOC15_IH_CLIENTID_SE1SH &&
+	    client_id != SOC15_IH_CLIENTID_SE2SH &&
+	    client_id != SOC15_IH_CLIENTID_SE3SH)
+		return false;
+
 	/* This is a known issue for gfx9. Under non HWS, pasid is not set
 	 * in the interrupt payload, so we need to find out the pasid on our
 	 * own.
@@ -96,17 +115,30 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev,
 	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
 	context_id = SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry);
 
-	if (source_id == SOC15_INTSRC_CP_END_OF_PIPE)
-		kfd_signal_event_interrupt(pasid, context_id, 32);
-	else if (source_id == SOC15_INTSRC_SDMA_TRAP)
-		kfd_signal_event_interrupt(pasid, context_id & 0xfffffff, 28);
-	else if (source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG)
-		kfd_signal_event_interrupt(pasid, context_id & 0xffffff, 24);
-	else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE)
-		kfd_signal_hw_exception_event(pasid);
-	else if (client_id == SOC15_IH_CLIENTID_VMC ||
-		client_id == SOC15_IH_CLIENTID_VMC1 ||
-		 client_id == SOC15_IH_CLIENTID_UTCL2) {
+	if (client_id == SOC15_IH_CLIENTID_GRBM_CP ||
+	    client_id == SOC15_IH_CLIENTID_SE0SH ||
+	    client_id == SOC15_IH_CLIENTID_SE1SH ||
+	    client_id == SOC15_IH_CLIENTID_SE2SH ||
+	    client_id == SOC15_IH_CLIENTID_SE3SH) {
+		if (source_id == SOC15_INTSRC_CP_END_OF_PIPE)
+			kfd_signal_event_interrupt(pasid, context_id, 32);
+		else if (source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG)
+			kfd_signal_event_interrupt(pasid, context_id & 0xffffff, 24);
+		else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE)
+			kfd_signal_hw_exception_event(pasid);
+	} else if (client_id == SOC15_IH_CLIENTID_SDMA0 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA1 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA2 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA3 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA4 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA5 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA6 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA7) {
+		if (source_id == SOC15_INTSRC_SDMA_TRAP)
+			kfd_signal_event_interrupt(pasid, context_id & 0xfffffff, 28);
+	} else if (client_id == SOC15_IH_CLIENTID_VMC ||
+		   client_id == SOC15_IH_CLIENTID_VMC1 ||
+		   client_id == SOC15_IH_CLIENTID_UTCL2) {
 		struct kfd_vm_fault_info info = {0};
 		uint16_t ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index a3fc23873819..0be72789ccbc 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -497,8 +497,6 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 			      dev->node_props.num_sdma_queues_per_engine);
 	sysfs_show_32bit_prop(buffer, offs, "num_cp_queues",
 			      dev->node_props.num_cp_queues);
-	sysfs_show_64bit_prop(buffer, offs, "unique_id",
-			      dev->node_props.unique_id);
 
 	if (dev->gpu) {
 		log_max_watch_addr =
@@ -529,6 +527,9 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 				      dev->node_props.capability);
 		sysfs_show_32bit_prop(buffer, offs, "sdma_fw_version",
 				      dev->gpu->sdma_fw_version);
+		sysfs_show_64bit_prop(buffer, offs, "unique_id",
+				      amdgpu_amdkfd_get_unique_id(dev->gpu->kgd));
+
 	}
 
 	return sysfs_show_32bit_prop(buffer, offs, "max_engine_clk_ccompute",
@@ -1340,7 +1341,6 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 		dev->gpu->dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) ?
 		amdgpu_amdkfd_get_num_gws(dev->gpu->kgd) : 0;
 	dev->node_props.num_cp_queues = get_cp_queues_num(dev->gpu->dqm);
-	dev->node_props.unique_id = amdgpu_amdkfd_get_unique_id(dev->gpu->kgd);
 
 	kfd_fill_mem_clk_max_info(dev);
 	kfd_fill_iolink_non_crat_info(dev);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
index 326d9b26b7aa..416fd910e12e 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
@@ -57,7 +57,6 @@
 
 struct kfd_node_properties {
 	uint64_t hive_id;
-	uint64_t unique_id;
 	uint32_t cpu_cores_count;
 	uint32_t simd_count;
 	uint32_t mem_banks_count;
diff --git a/drivers/gpu/drm/amd/display/Kconfig b/drivers/gpu/drm/amd/display/Kconfig
index 797b5d4b43e5..e509a175ed17 100644
--- a/drivers/gpu/drm/amd/display/Kconfig
+++ b/drivers/gpu/drm/amd/display/Kconfig
@@ -6,7 +6,7 @@ config DRM_AMD_DC
 	bool "AMD DC - Enable new display engine"
 	default y
 	select SND_HDA_COMPONENT if SND_HDA_CORE
-	select DRM_AMD_DC_DCN if (X86 || PPC64 || (ARM64 && KERNEL_MODE_NEON)) && !(KCOV_INSTRUMENT_ALL && KCOV_ENABLE_COMPARISONS)
+	select DRM_AMD_DC_DCN if (X86 || PPC64) && !(KCOV_INSTRUMENT_ALL && KCOV_ENABLE_COMPARISONS)
 	help
 	  Choose this option if you want to use the new display engine
 	  support for AMDGPU. This adds required support for Vega and
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 22124f76d0b5..55e39b462a5e 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -60,7 +60,6 @@
 
 #include <linux/module.h>
 #include <linux/moduleparam.h>
-#include <linux/version.h>
 #include <linux/types.h>
 #include <linux/pm_runtime.h>
 #include <linux/pci.h>
@@ -938,42 +937,49 @@ static void mmhub_read_system_context(struct amdgpu_device *adev, struct dc_phy_
 
 }
 #endif
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+static void event_mall_stutter(struct work_struct *work)
+{
 
-#ifdef CONFIG_DEBUG_FS
-static int create_crtc_crc_properties(struct amdgpu_display_manager *dm)
+	struct vblank_workqueue *vblank_work = container_of(work, struct vblank_workqueue, mall_work);
+	struct amdgpu_display_manager *dm = vblank_work->dm;
+
+	mutex_lock(&dm->dc_lock);
+
+	if (vblank_work->enable)
+		dm->active_vblank_irq_count++;
+	else
+		dm->active_vblank_irq_count--;
+
+
+	dc_allow_idle_optimizations(
+		dm->dc, dm->active_vblank_irq_count == 0 ? true : false);
+
+	DRM_DEBUG_DRIVER("Allow idle optimizations (MALL): %d\n", dm->active_vblank_irq_count == 0);
+
+
+	mutex_unlock(&dm->dc_lock);
+}
+
+static struct vblank_workqueue *vblank_create_workqueue(struct amdgpu_device *adev, struct dc *dc)
 {
-	dm->crc_win_x_start_property =
-		drm_property_create_range(adev_to_drm(dm->adev),
-					  DRM_MODE_PROP_ATOMIC,
-					  "AMD_CRC_WIN_X_START", 0, U16_MAX);
-	if (!dm->crc_win_x_start_property)
-		return -ENOMEM;
 
-	dm->crc_win_y_start_property =
-		drm_property_create_range(adev_to_drm(dm->adev),
-					  DRM_MODE_PROP_ATOMIC,
-					  "AMD_CRC_WIN_Y_START", 0, U16_MAX);
-	if (!dm->crc_win_y_start_property)
-		return -ENOMEM;
+	int max_caps = dc->caps.max_links;
+	struct vblank_workqueue *vblank_work;
+	int i = 0;
 
-	dm->crc_win_x_end_property =
-		drm_property_create_range(adev_to_drm(dm->adev),
-					  DRM_MODE_PROP_ATOMIC,
-					  "AMD_CRC_WIN_X_END", 0, U16_MAX);
-	if (!dm->crc_win_x_end_property)
-		return -ENOMEM;
+	vblank_work = kcalloc(max_caps, sizeof(*vblank_work), GFP_KERNEL);
+	if (ZERO_OR_NULL_PTR(vblank_work)) {
+		kfree(vblank_work);
+		return NULL;
+	}
 
-	dm->crc_win_y_end_property =
-		drm_property_create_range(adev_to_drm(dm->adev),
-					  DRM_MODE_PROP_ATOMIC,
-					  "AMD_CRC_WIN_Y_END", 0, U16_MAX);
-	if (!dm->crc_win_y_end_property)
-		return -ENOMEM;
+	for (i = 0; i < max_caps; i++)
+		INIT_WORK(&vblank_work[i].mall_work, event_mall_stutter);
 
-	return 0;
+	return vblank_work;
 }
 #endif
-
 static int amdgpu_dm_init(struct amdgpu_device *adev)
 {
 	struct dc_init_data init_data;
@@ -993,6 +999,9 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 
 	mutex_init(&adev->dm.dc_lock);
 	mutex_init(&adev->dm.audio_lock);
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	spin_lock_init(&adev->dm.vblank_lock);
+#endif
 
 	if(amdgpu_dm_irq_init(adev)) {
 		DRM_ERROR("amdgpu: failed to initialize DM IRQ support.\n");
@@ -1051,8 +1060,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 
 	init_data.flags.power_down_display_on_boot = true;
 
-	init_data.soc_bounding_box = adev->dm.soc_bounding_box;
-
 	/* Display Core create. */
 	adev->dm.dc = dc_create(&init_data);
 
@@ -1109,6 +1116,17 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 
 	amdgpu_dm_init_color_mod();
 
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	if (adev->dm.dc->caps.max_links > 0) {
+		adev->dm.vblank_workqueue = vblank_create_workqueue(adev, adev->dm.dc);
+
+		if (!adev->dm.vblank_workqueue)
+			DRM_ERROR("amdgpu: failed to initialize vblank_workqueue.\n");
+		else
+			DRM_DEBUG_DRIVER("amdgpu: vblank_workqueue init done %p.\n", adev->dm.vblank_workqueue);
+	}
+#endif
+
 #ifdef CONFIG_DRM_AMD_DC_HDCP
 	if (adev->dm.dc->caps.max_links > 0 && adev->asic_type >= CHIP_RAVEN) {
 		adev->dm.hdcp_workqueue = hdcp_create_workqueue(adev, &init_params.cp_psp, adev->dm.dc);
@@ -1121,10 +1139,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 		dc_init_callbacks(adev->dm.dc, &init_params);
 	}
 #endif
-#ifdef CONFIG_DEBUG_FS
-	if (create_crtc_crc_properties(&adev->dm))
-		DRM_ERROR("amdgpu: failed to create crc property.\n");
-#endif
 	if (amdgpu_dm_initialize_drm_device(adev)) {
 		DRM_ERROR(
 		"amdgpu: failed to initialize sw for display support.\n");
@@ -1170,7 +1184,7 @@ static void amdgpu_dm_fini(struct amdgpu_device *adev)
 
 #ifdef CONFIG_DRM_AMD_DC_HDCP
 	if (adev->dm.hdcp_workqueue) {
-		hdcp_destroy(adev->dm.hdcp_workqueue);
+		hdcp_destroy(&adev->dev->kobj, adev->dm.hdcp_workqueue);
 		adev->dm.hdcp_workqueue = NULL;
 	}
 
@@ -1817,6 +1831,11 @@ static int dm_suspend(void *handle)
 
 	if (amdgpu_in_reset(adev)) {
 		mutex_lock(&dm->dc_lock);
+
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+		dc_allow_idle_optimizations(adev->dm.dc, false);
+#endif
+
 		dm->cached_dc_state = dc_copy_state(dm->dc->current_state);
 
 		dm_gpureset_toggle_interrupts(adev, dm->cached_dc_state, false);
@@ -1872,8 +1891,8 @@ static void emulated_link_detect(struct dc_link *link)
 	link->type = dc_connection_none;
 	prev_sink = link->local_sink;
 
-	if (prev_sink != NULL)
-		dc_sink_retain(prev_sink);
+	if (prev_sink)
+		dc_sink_release(prev_sink);
 
 	switch (link->connector_signal) {
 	case SIGNAL_TYPE_HDMI_TYPE_A: {
@@ -2248,6 +2267,11 @@ static void update_connector_ext_caps(struct amdgpu_dm_connector *aconnector)
 	    caps->ext_caps->bits.hdr_aux_backlight_control == 1)
 		caps->aux_support = true;
 
+	if (amdgpu_backlight == 0)
+		caps->aux_support = false;
+	else if (amdgpu_backlight == 1)
+		caps->aux_support = true;
+
 	/* From the specification (CTA-861-G), for calculating the maximum
 	 * luminance we need to use:
 	 *	Luminance = 50*2**(CV/32)
@@ -2369,8 +2393,10 @@ void amdgpu_dm_update_connector_after_detect(
 		 * TODO: check if we still need the S3 mode update workaround.
 		 * If yes, put it here.
 		 */
-		if (aconnector->dc_sink)
+		if (aconnector->dc_sink) {
 			amdgpu_dm_update_freesync_caps(connector, NULL);
+			dc_sink_release(aconnector->dc_sink);
+		}
 
 		aconnector->dc_sink = sink;
 		dc_sink_retain(aconnector->dc_sink);
@@ -2386,9 +2412,6 @@ void amdgpu_dm_update_connector_after_detect(
 
 			drm_connector_update_edid_property(connector,
 							   aconnector->edid);
-			aconnector->num_modes = drm_add_edid_modes(connector, aconnector->edid);
-			drm_connector_list_update(connector);
-
 			if (aconnector->dc_link->aux_mode)
 				drm_dp_cec_set_edid(&aconnector->dm_dp_aux.aux,
 						    aconnector->edid);
@@ -3167,19 +3190,6 @@ static void amdgpu_dm_update_backlight_caps(struct amdgpu_display_manager *dm)
 #endif
 }
 
-static int set_backlight_via_aux(struct dc_link *link, uint32_t brightness)
-{
-	bool rc;
-
-	if (!link)
-		return 1;
-
-	rc = dc_link_set_backlight_level_nits(link, true, brightness,
-					      AUX_BL_DEFAULT_TRANSITION_TIME_MS);
-
-	return rc ? 0 : 1;
-}
-
 static int get_brightness_range(const struct amdgpu_dm_backlight_caps *caps,
 				unsigned *min, unsigned *max)
 {
@@ -3242,9 +3252,10 @@ static int amdgpu_dm_backlight_update_status(struct backlight_device *bd)
 	brightness = convert_brightness_from_user(&caps, bd->props.brightness);
 	// Change brightness based on AUX property
 	if (caps.aux_support)
-		return set_backlight_via_aux(link, brightness);
-
-	rc = dc_link_set_backlight_level(dm->backlight_link, brightness, 0);
+		rc = dc_link_set_backlight_level_nits(link, true, brightness,
+						      AUX_BL_DEFAULT_TRANSITION_TIME_MS);
+	else
+		rc = dc_link_set_backlight_level(dm->backlight_link, brightness, 0);
 
 	return rc ? 0 : 1;
 }
@@ -3252,11 +3263,27 @@ static int amdgpu_dm_backlight_update_status(struct backlight_device *bd)
 static int amdgpu_dm_backlight_get_brightness(struct backlight_device *bd)
 {
 	struct amdgpu_display_manager *dm = bl_get_data(bd);
-	int ret = dc_link_get_backlight_level(dm->backlight_link);
+	struct amdgpu_dm_backlight_caps caps;
+
+	amdgpu_dm_update_backlight_caps(dm);
+	caps = dm->backlight_caps;
+
+	if (caps.aux_support) {
+		struct dc_link *link = (struct dc_link *)dm->backlight_link;
+		u32 avg, peak;
+		bool rc;
+
+		rc = dc_link_get_backlight_level_nits(link, &avg, &peak);
+		if (!rc)
+			return bd->props.brightness;
+		return convert_brightness_to_user(&caps, avg);
+	} else {
+		int ret = dc_link_get_backlight_level(dm->backlight_link);
 
-	if (ret == DC_ERROR_UNEXPECTED)
-		return bd->props.brightness;
-	return convert_brightness_to_user(&dm->backlight_caps, ret);
+		if (ret == DC_ERROR_UNEXPECTED)
+			return bd->props.brightness;
+		return convert_brightness_to_user(&caps, ret);
+	}
 }
 
 static const struct backlight_ops amdgpu_dm_backlight_ops = {
@@ -3760,10 +3787,53 @@ static const struct drm_encoder_funcs amdgpu_dm_encoder_funcs = {
 };
 
 
+static void get_min_max_dc_plane_scaling(struct drm_device *dev,
+					 struct drm_framebuffer *fb,
+					 int *min_downscale, int *max_upscale)
+{
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct dc *dc = adev->dm.dc;
+	/* Caps for all supported planes are the same on DCE and DCN 1 - 3 */
+	struct dc_plane_cap *plane_cap = &dc->caps.planes[0];
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_NV21:
+		*max_upscale = plane_cap->max_upscale_factor.nv12;
+		*min_downscale = plane_cap->max_downscale_factor.nv12;
+		break;
+
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+		*max_upscale = plane_cap->max_upscale_factor.fp16;
+		*min_downscale = plane_cap->max_downscale_factor.fp16;
+		break;
+
+	default:
+		*max_upscale = plane_cap->max_upscale_factor.argb8888;
+		*min_downscale = plane_cap->max_downscale_factor.argb8888;
+		break;
+	}
+
+	/*
+	 * A factor of 1 in the plane_cap means to not allow scaling, ie. use a
+	 * scaling factor of 1.0 == 1000 units.
+	 */
+	if (*max_upscale == 1)
+		*max_upscale = 1000;
+
+	if (*min_downscale == 1)
+		*min_downscale = 1000;
+}
+
+
 static int fill_dc_scaling_info(const struct drm_plane_state *state,
 				struct dc_scaling_info *scaling_info)
 {
-	int scale_w, scale_h;
+	int scale_w, scale_h, min_downscale, max_upscale;
 
 	memset(scaling_info, 0, sizeof(*scaling_info));
 
@@ -3795,17 +3865,25 @@ static int fill_dc_scaling_info(const struct drm_plane_state *state,
 	/* DRM doesn't specify clipping on destination output. */
 	scaling_info->clip_rect = scaling_info->dst_rect;
 
-	/* TODO: Validate scaling per-format with DC plane caps */
+	/* Validate scaling per-format with DC plane caps */
+	if (state->plane && state->plane->dev && state->fb) {
+		get_min_max_dc_plane_scaling(state->plane->dev, state->fb,
+					     &min_downscale, &max_upscale);
+	} else {
+		min_downscale = 250;
+		max_upscale = 16000;
+	}
+
 	scale_w = scaling_info->dst_rect.width * 1000 /
 		  scaling_info->src_rect.width;
 
-	if (scale_w < 250 || scale_w > 16000)
+	if (scale_w < min_downscale || scale_w > max_upscale)
 		return -EINVAL;
 
 	scale_h = scaling_info->dst_rect.height * 1000 /
 		  scaling_info->src_rect.height;
 
-	if (scale_h < 250 || scale_h > 16000)
+	if (scale_h < min_downscale || scale_h > max_upscale)
 		return -EINVAL;
 
 	/*
@@ -4646,6 +4724,7 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev,
 	dc_plane_state->global_alpha_value = plane_info.global_alpha_value;
 	dc_plane_state->dcc = plane_info.dcc;
 	dc_plane_state->layer_index = plane_info.layer_index; // Always returns 0
+	dc_plane_state->flip_int_enabled = true;
 
 	/*
 	 * Always set input transfer function, since plane state is refreshed
@@ -5333,64 +5412,12 @@ dm_crtc_duplicate_state(struct drm_crtc *crtc)
 	state->crc_src = cur->crc_src;
 	state->cm_has_degamma = cur->cm_has_degamma;
 	state->cm_is_degamma_srgb = cur->cm_is_degamma_srgb;
-#ifdef CONFIG_DEBUG_FS
-	state->crc_window = cur->crc_window;
-#endif
+
 	/* TODO Duplicate dc_stream after objects are stream object is flattened */
 
 	return &state->base;
 }
 
-#ifdef CONFIG_DEBUG_FS
-static int amdgpu_dm_crtc_atomic_set_property(struct drm_crtc *crtc,
-					    struct drm_crtc_state *crtc_state,
-					    struct drm_property *property,
-					    uint64_t val)
-{
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct dm_crtc_state *dm_new_state =
-		to_dm_crtc_state(crtc_state);
-
-	if (property == adev->dm.crc_win_x_start_property)
-		dm_new_state->crc_window.x_start = val;
-	else if (property == adev->dm.crc_win_y_start_property)
-		dm_new_state->crc_window.y_start = val;
-	else if (property == adev->dm.crc_win_x_end_property)
-		dm_new_state->crc_window.x_end = val;
-	else if (property == adev->dm.crc_win_y_end_property)
-		dm_new_state->crc_window.y_end = val;
-	else
-		return -EINVAL;
-
-	return 0;
-}
-
-static int amdgpu_dm_crtc_atomic_get_property(struct drm_crtc *crtc,
-					    const struct drm_crtc_state *state,
-					    struct drm_property *property,
-					    uint64_t *val)
-{
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct dm_crtc_state *dm_state =
-		to_dm_crtc_state(state);
-
-	if (property == adev->dm.crc_win_x_start_property)
-		*val = dm_state->crc_window.x_start;
-	else if (property == adev->dm.crc_win_y_start_property)
-		*val = dm_state->crc_window.y_start;
-	else if (property == adev->dm.crc_win_x_end_property)
-		*val = dm_state->crc_window.x_end;
-	else if (property == adev->dm.crc_win_y_end_property)
-		*val = dm_state->crc_window.y_end;
-	else
-		return -EINVAL;
-
-	return 0;
-}
-#endif
-
 static inline int dm_set_vupdate_irq(struct drm_crtc *crtc, bool enable)
 {
 	enum dc_irq_source irq_source;
@@ -5413,6 +5440,10 @@ static inline int dm_set_vblank(struct drm_crtc *crtc, bool enable)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
 	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state);
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	struct amdgpu_display_manager *dm = &adev->dm;
+	unsigned long flags;
+#endif
 	int rc = 0;
 
 	if (enable) {
@@ -5428,7 +5459,23 @@ static inline int dm_set_vblank(struct drm_crtc *crtc, bool enable)
 		return rc;
 
 	irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst;
-	return dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY;
+
+	if (!dc_interrupt_set(adev->dm.dc, irq_source, enable))
+		return -EBUSY;
+
+	if (amdgpu_in_reset(adev))
+		return 0;
+
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	spin_lock_irqsave(&dm->vblank_lock, flags);
+	dm->vblank_workqueue->dm = dm;
+	dm->vblank_workqueue->otg_inst = acrtc->otg_inst;
+	dm->vblank_workqueue->enable = enable;
+	spin_unlock_irqrestore(&dm->vblank_lock, flags);
+	schedule_work(&dm->vblank_workqueue->mall_work);
+#endif
+
+	return 0;
 }
 
 static int dm_enable_vblank(struct drm_crtc *crtc)
@@ -5456,10 +5503,6 @@ static const struct drm_crtc_funcs amdgpu_dm_crtc_funcs = {
 	.enable_vblank = dm_enable_vblank,
 	.disable_vblank = dm_disable_vblank,
 	.get_vblank_timestamp = drm_crtc_vblank_helper_get_vblank_timestamp,
-#ifdef CONFIG_DEBUG_FS
-	.atomic_set_property = amdgpu_dm_crtc_atomic_set_property,
-	.atomic_get_property = amdgpu_dm_crtc_atomic_get_property,
-#endif
 };
 
 static enum drm_connector_status
@@ -6423,12 +6466,51 @@ static void dm_plane_helper_cleanup_fb(struct drm_plane *plane,
 static int dm_plane_helper_check_state(struct drm_plane_state *state,
 				       struct drm_crtc_state *new_crtc_state)
 {
-	int max_downscale = 0;
-	int max_upscale = INT_MAX;
+	struct drm_framebuffer *fb = state->fb;
+	int min_downscale, max_upscale;
+	int min_scale = 0;
+	int max_scale = INT_MAX;
+
+	/* Plane enabled? Validate viewport and get scaling factors from plane caps. */
+	if (fb && state->crtc) {
+		/* Validate viewport to cover the case when only the position changes */
+		if (state->plane->type != DRM_PLANE_TYPE_CURSOR) {
+			int viewport_width = state->crtc_w;
+			int viewport_height = state->crtc_h;
+
+			if (state->crtc_x < 0)
+				viewport_width += state->crtc_x;
+			else if (state->crtc_x + state->crtc_w > new_crtc_state->mode.crtc_hdisplay)
+				viewport_width = new_crtc_state->mode.crtc_hdisplay - state->crtc_x;
+
+			if (state->crtc_y < 0)
+				viewport_height += state->crtc_y;
+			else if (state->crtc_y + state->crtc_h > new_crtc_state->mode.crtc_vdisplay)
+				viewport_height = new_crtc_state->mode.crtc_vdisplay - state->crtc_y;
+
+			/* If completely outside of screen, viewport_width and/or viewport_height will be negative,
+			 * which is still OK to satisfy the condition below, thereby also covering these cases
+			 * (when plane is completely outside of screen).
+			 * x2 for width is because of pipe-split.
+			 */
+			if (viewport_width < MIN_VIEWPORT_SIZE*2 || viewport_height < MIN_VIEWPORT_SIZE)
+				return -EINVAL;
+		}
+
+		/* Get min/max allowed scaling factors from plane caps. */
+		get_min_max_dc_plane_scaling(state->crtc->dev, fb,
+					     &min_downscale, &max_upscale);
+		/*
+		 * Convert to drm convention: 16.16 fixed point, instead of dc's
+		 * 1.0 == 1000. Also drm scaling is src/dst instead of dc's
+		 * dst/src, so min_scale = 1.0 / max_upscale, etc.
+		 */
+		min_scale = (1000 << 16) / max_upscale;
+		max_scale = (1000 << 16) / min_downscale;
+	}
 
-	/* TODO: These should be checked against DC plane caps */
 	return drm_atomic_helper_check_plane_state(
-		state, new_crtc_state, max_downscale, max_upscale, true, true);
+		state, new_crtc_state, min_scale, max_scale, true, true);
 }
 
 static int dm_plane_atomic_check(struct drm_plane *plane,
@@ -6666,25 +6748,6 @@ static int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
 	return 0;
 }
 
-#ifdef CONFIG_DEBUG_FS
-static void attach_crtc_crc_properties(struct amdgpu_display_manager *dm,
-				struct amdgpu_crtc *acrtc)
-{
-	drm_object_attach_property(&acrtc->base.base,
-				   dm->crc_win_x_start_property,
-				   0);
-	drm_object_attach_property(&acrtc->base.base,
-				   dm->crc_win_y_start_property,
-				   0);
-	drm_object_attach_property(&acrtc->base.base,
-				   dm->crc_win_x_end_property,
-				   0);
-	drm_object_attach_property(&acrtc->base.base,
-				   dm->crc_win_y_end_property,
-				   0);
-}
-#endif
-
 static int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
 			       struct drm_plane *plane,
 			       uint32_t crtc_index)
@@ -6732,9 +6795,7 @@ static int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
 	drm_crtc_enable_color_mgmt(&acrtc->base, MAX_COLOR_LUT_ENTRIES,
 				   true, MAX_COLOR_LUT_ENTRIES);
 	drm_mode_crtc_set_gamma_size(&acrtc->base, MAX_COLOR_LEGACY_LUT_ENTRIES);
-#ifdef CONFIG_DEBUG_FS
-	attach_crtc_crc_properties(dm, acrtc);
-#endif
+
 	return 0;
 
 fail:
@@ -8371,7 +8432,6 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 	 */
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
 		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
-		bool configure_crc = false;
 
 		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 
@@ -8381,30 +8441,21 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 			dc_stream_retain(dm_new_crtc_state->stream);
 			acrtc->dm_irq_params.stream = dm_new_crtc_state->stream;
 			manage_dm_interrupts(adev, acrtc, true);
-		}
+
 #ifdef CONFIG_DEBUG_FS
-		if (new_crtc_state->active &&
-			amdgpu_dm_is_valid_crc_source(dm_new_crtc_state->crc_src)) {
 			/**
 			 * Frontend may have changed so reapply the CRC capture
 			 * settings for the stream.
 			 */
 			dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
-			dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
 
-			if (amdgpu_dm_crc_window_is_default(dm_new_crtc_state)) {
-				if (!old_crtc_state->active || drm_atomic_crtc_needs_modeset(new_crtc_state))
-					configure_crc = true;
-			} else {
-				if (amdgpu_dm_crc_window_changed(dm_new_crtc_state, dm_old_crtc_state))
-					configure_crc = true;
-			}
-
-			if (configure_crc)
+			if (amdgpu_dm_is_valid_crc_source(dm_new_crtc_state->crc_src)) {
 				amdgpu_dm_crtc_configure_crc_source(
-					crtc, dm_new_crtc_state, dm_new_crtc_state->crc_src);
-		}
+					crtc, dm_new_crtc_state,
+					dm_new_crtc_state->crc_src);
+			}
 #endif
+		}
 	}
 
 	for_each_new_crtc_in_state(state, crtc, new_crtc_state, j)
@@ -8489,14 +8540,14 @@ static int dm_force_atomic_commit(struct drm_connector *connector)
 
 	ret = PTR_ERR_OR_ZERO(conn_state);
 	if (ret)
-		goto err;
+		goto out;
 
 	/* Attach crtc to drm_atomic_state*/
 	crtc_state = drm_atomic_get_crtc_state(state, &disconnected_acrtc->base);
 
 	ret = PTR_ERR_OR_ZERO(crtc_state);
 	if (ret)
-		goto err;
+		goto out;
 
 	/* force a restore */
 	crtc_state->mode_changed = true;
@@ -8506,17 +8557,15 @@ static int dm_force_atomic_commit(struct drm_connector *connector)
 
 	ret = PTR_ERR_OR_ZERO(plane_state);
 	if (ret)
-		goto err;
-
+		goto out;
 
 	/* Call commit internally with the state we just constructed */
 	ret = drm_atomic_commit(state);
-	if (!ret)
-		return 0;
 
-err:
-	DRM_ERROR("Restoring old state failed with %i\n", ret);
+out:
 	drm_atomic_state_put(state);
+	if (ret)
+		DRM_ERROR("Restoring old state failed with %i\n", ret);
 
 	return ret;
 }
@@ -9723,6 +9772,10 @@ void amdgpu_dm_update_freesync_caps(struct drm_connector *connector,
 			amdgpu_dm_connector->max_vfreq = range->max_vfreq;
 			amdgpu_dm_connector->pixel_clock_mhz =
 				range->pixel_clock_mhz * 10;
+
+			connector->display_info.monitor_range.min_vfreq = range->min_vfreq;
+			connector->display_info.monitor_range.max_vfreq = range->max_vfreq;
+
 			break;
 		}
 
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
index 2ee6edb3df93..8bfe901cf237 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
@@ -58,10 +58,10 @@
 /* Forward declarations */
 struct amdgpu_device;
 struct drm_device;
-struct amdgpu_dm_irq_handler_data;
 struct dc;
 struct amdgpu_bo;
 struct dmub_srv;
+struct dc_plane_state;
 
 struct common_irq_params {
 	struct amdgpu_device *adev;
@@ -93,6 +93,20 @@ struct dm_compressor_info {
 };
 
 /**
+ * struct vblank_workqueue - Works to be executed in a separate thread during vblank
+ * @mall_work: work for mall stutter
+ * @dm: amdgpu display manager device
+ * @otg_inst: otg instance of which vblank is being set
+ * @enable: true if enable vblank
+ */
+struct vblank_workqueue {
+	struct work_struct mall_work;
+	struct amdgpu_display_manager *dm;
+	int otg_inst;
+	bool enable;
+};
+
+/**
  * struct amdgpu_dm_backlight_caps - Information about backlight
  *
  * Describe the backlight support for ACPI or eDP AUX.
@@ -244,6 +258,15 @@ struct amdgpu_display_manager {
 	struct mutex audio_lock;
 
 	/**
+	 * @vblank_work_lock:
+	 *
+	 * Guards access to deferred vblank work state.
+	 */
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	spinlock_t vblank_lock;
+#endif
+
+	/**
 	 * @audio_component:
 	 *
 	 * Used to notify ELD changes to sound driver.
@@ -321,6 +344,10 @@ struct amdgpu_display_manager {
 	struct hdcp_workqueue *hdcp_workqueue;
 #endif
 
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	struct vblank_workqueue *vblank_workqueue;
+#endif
+
 	struct drm_atomic_state *cached_state;
 	struct dc_state *cached_dc_state;
 
@@ -336,32 +363,13 @@ struct amdgpu_display_manager {
 	 */
 	const struct gpu_info_soc_bounding_box_v1_0 *soc_bounding_box;
 
-#ifdef CONFIG_DEBUG_FS
-	/**
-	 * @crc_win_x_start_property:
-	 *
-	 * X start of the crc calculation window
-	 */
-	struct drm_property *crc_win_x_start_property;
 	/**
-	 * @crc_win_y_start_property:
+	 * @active_vblank_irq_count:
 	 *
-	 * Y start of the crc calculation window
+	 * number of currently active vblank irqs
 	 */
-	struct drm_property *crc_win_y_start_property;
-	/**
-	 * @crc_win_x_end_property:
-	 *
-	 * X end of the crc calculation window
-	 */
-	struct drm_property *crc_win_x_end_property;
-	/**
-	 * @crc_win_y_end_property:
-	 *
-	 * Y end of the crc calculation window
-	 */
-	struct drm_property *crc_win_y_end_property;
-#endif
+	uint32_t active_vblank_irq_count;
+
 	/**
 	 * @mst_encoders:
 	 *
@@ -438,25 +446,11 @@ struct amdgpu_dm_connector {
 
 extern const struct amdgpu_ip_block_version dm_ip_block;
 
-struct amdgpu_framebuffer;
-struct amdgpu_display_manager;
-struct dc_validation_set;
-struct dc_plane_state;
-
 struct dm_plane_state {
 	struct drm_plane_state base;
 	struct dc_plane_state *dc_state;
 };
 
-#ifdef CONFIG_DEBUG_FS
-struct crc_rec {
-	uint16_t x_start;
-	uint16_t y_start;
-	uint16_t x_end;
-	uint16_t y_end;
-	};
-#endif
-
 struct dm_crtc_state {
 	struct drm_crtc_state base;
 	struct dc_stream_state *stream;
@@ -479,9 +473,6 @@ struct dm_crtc_state {
 	struct dc_info_packet vrr_infopacket;
 
 	int abm_level;
-#ifdef CONFIG_DEBUG_FS
-	struct crc_rec crc_window;
-#endif
 };
 
 #define to_dm_crtc_state(x) container_of(x, struct dm_crtc_state, base)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c
index 7b886a779a8c..66cb8730586b 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c
@@ -81,41 +81,6 @@ const char *const *amdgpu_dm_crtc_get_crc_sources(struct drm_crtc *crtc,
 	return pipe_crc_sources;
 }
 
-static void amdgpu_dm_set_crc_window_default(struct dm_crtc_state *dm_crtc_state)
-{
-	dm_crtc_state->crc_window.x_start = 0;
-	dm_crtc_state->crc_window.y_start = 0;
-	dm_crtc_state->crc_window.x_end = 0;
-	dm_crtc_state->crc_window.y_end = 0;
-}
-
-bool amdgpu_dm_crc_window_is_default(struct dm_crtc_state *dm_crtc_state)
-{
-	bool ret = true;
-
-	if ((dm_crtc_state->crc_window.x_start != 0) ||
-		(dm_crtc_state->crc_window.y_start != 0) ||
-		(dm_crtc_state->crc_window.x_end != 0) ||
-		(dm_crtc_state->crc_window.y_end != 0))
-		ret = false;
-
-	return ret;
-}
-
-bool amdgpu_dm_crc_window_changed(struct dm_crtc_state *dm_new_crtc_state,
-					struct dm_crtc_state *dm_old_crtc_state)
-{
-	bool ret = false;
-
-	if ((dm_new_crtc_state->crc_window.x_start != dm_old_crtc_state->crc_window.x_start) ||
-		(dm_new_crtc_state->crc_window.y_start != dm_old_crtc_state->crc_window.y_start) ||
-		(dm_new_crtc_state->crc_window.x_end != dm_old_crtc_state->crc_window.x_end) ||
-		(dm_new_crtc_state->crc_window.y_end != dm_old_crtc_state->crc_window.y_end))
-		ret = true;
-
-	return ret;
-}
-
 int
 amdgpu_dm_crtc_verify_crc_source(struct drm_crtc *crtc, const char *src_name,
 				 size_t *values_cnt)
@@ -140,7 +105,6 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc,
 	struct dc_stream_state *stream_state = dm_crtc_state->stream;
 	bool enable = amdgpu_dm_is_valid_crc_source(source);
 	int ret = 0;
-	struct crc_params *crc_window = NULL, tmp_window;
 
 	/* Configuration will be deferred to stream enable. */
 	if (!stream_state)
@@ -150,24 +114,8 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc,
 
 	/* Enable CRTC CRC generation if necessary. */
 	if (dm_is_crc_source_crtc(source) || source == AMDGPU_DM_PIPE_CRC_SOURCE_NONE) {
-		if (!enable)
-			amdgpu_dm_set_crc_window_default(dm_crtc_state);
-
-		if (!amdgpu_dm_crc_window_is_default(dm_crtc_state)) {
-			crc_window = &tmp_window;
-
-			tmp_window.windowa_x_start = dm_crtc_state->crc_window.x_start;
-			tmp_window.windowa_y_start = dm_crtc_state->crc_window.y_start;
-			tmp_window.windowa_x_end = dm_crtc_state->crc_window.x_end;
-			tmp_window.windowa_y_end = dm_crtc_state->crc_window.y_end;
-			tmp_window.windowb_x_start = dm_crtc_state->crc_window.x_start;
-			tmp_window.windowb_y_start = dm_crtc_state->crc_window.y_start;
-			tmp_window.windowb_x_end = dm_crtc_state->crc_window.x_end;
-			tmp_window.windowb_y_end = dm_crtc_state->crc_window.y_end;
-		}
-
 		if (!dc_stream_configure_crc(stream_state->ctx->dc,
-					     stream_state, crc_window, enable, enable)) {
+					     stream_state, NULL, enable, enable)) {
 			ret = -EINVAL;
 			goto unlock;
 		}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
index 0235bfb246e5..f7d731797d3f 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
@@ -47,9 +47,6 @@ static inline bool amdgpu_dm_is_valid_crc_source(enum amdgpu_dm_pipe_crc_source
 
 /* amdgpu_dm_crc.c */
 #ifdef CONFIG_DEBUG_FS
-bool amdgpu_dm_crc_window_is_default(struct dm_crtc_state *dm_crtc_state);
-bool amdgpu_dm_crc_window_changed(struct dm_crtc_state *dm_new_crtc_state,
-					struct dm_crtc_state *dm_old_crtc_state);
 int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc,
 					struct dm_crtc_state *dm_crtc_state,
 					enum amdgpu_dm_pipe_crc_source source);
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
index 11459fb09a37..360952129b6d 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
@@ -691,7 +691,7 @@ static ssize_t dp_phy_test_pattern_debugfs_write(struct file *f, const char __us
 	return size;
 }
 
-/**
+/*
  * Returns the DMCUB tracebuffer contents.
  * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_dmub_tracebuffer
  */
@@ -735,7 +735,7 @@ static int dmub_tracebuffer_show(struct seq_file *m, void *data)
 	return 0;
 }
 
-/**
+/*
  * Returns the DMCUB firmware state contents.
  * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_dmub_fw_state
  */
@@ -1063,7 +1063,7 @@ static int dp_dsc_fec_support_show(struct seq_file *m, void *data)
  *	echo 0 > /sys/kernel/debug/dri/0/DP-X/trigger_hotplug
  *
  */
-static ssize_t dp_trigger_hotplug(struct file *f, const char __user *buf,
+static ssize_t trigger_hotplug(struct file *f, const char __user *buf,
 							size_t size, loff_t *pos)
 {
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
@@ -2214,9 +2214,9 @@ static const struct file_operations dp_dsc_slice_bpg_offset_debugfs_fops = {
 	.llseek = default_llseek
 };
 
-static const struct file_operations dp_trigger_hotplug_debugfs_fops = {
+static const struct file_operations trigger_hotplug_debugfs_fops = {
 	.owner = THIS_MODULE,
-	.write = dp_trigger_hotplug,
+	.write = trigger_hotplug,
 	.llseek = default_llseek
 };
 
@@ -2270,7 +2270,6 @@ static const struct {
 	const struct file_operations *fops;
 } dp_debugfs_entries[] = {
 		{"link_settings", &dp_link_settings_debugfs_fops},
-		{"trigger_hotplug", &dp_trigger_hotplug_debugfs_fops},
 		{"phy_settings", &dp_phy_settings_debugfs_fop},
 		{"test_pattern", &dp_phy_test_pattern_fops},
 #ifdef CONFIG_DRM_AMD_DC_HDCP
@@ -2367,6 +2366,9 @@ void connector_debugfs_init(struct amdgpu_dm_connector *connector)
 	debugfs_create_file("output_bpc", 0644, dir, connector,
 			    &output_bpc_fops);
 
+	debugfs_create_file("trigger_hotplug", 0644, dir, connector,
+			    &trigger_hotplug_debugfs_fops);
+
 	connector->debugfs_dpcd_address = 0;
 	connector->debugfs_dpcd_size = 0;
 
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
index c2cd184f0bbd..0cdbfcd475ec 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
@@ -376,7 +376,7 @@ static void event_cpirq(struct work_struct *work)
 }
 
 
-void hdcp_destroy(struct hdcp_workqueue *hdcp_work)
+void hdcp_destroy(struct kobject *kobj, struct hdcp_workqueue *hdcp_work)
 {
 	int i = 0;
 
@@ -385,6 +385,7 @@ void hdcp_destroy(struct hdcp_workqueue *hdcp_work)
 		cancel_delayed_work_sync(&hdcp_work[i].watchdog_timer_dwork);
 	}
 
+	sysfs_remove_bin_file(kobj, &hdcp_work[0].attr);
 	kfree(hdcp_work->srm);
 	kfree(hdcp_work->srm_temp);
 	kfree(hdcp_work);
@@ -449,11 +450,12 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
 		link->mode = mod_hdcp_signal_type_to_operation_mode(aconnector->dc_sink->sink_signal);
 
 	display->controller = CONTROLLER_ID_D0 + config->otg_inst;
-	display->dig_fe = config->stream_enc_inst;
-	link->dig_be = config->link_enc_inst;
+	display->dig_fe = config->dig_fe;
+	link->dig_be = config->dig_be;
 	link->ddc_line = aconnector->dc_link->ddc_hw_inst + 1;
 	link->dp.rev = aconnector->dc_link->dpcd_caps.dpcd_rev.raw;
-	link->dp.mst_supported = config->mst_supported;
+	link->dp.assr_enabled = config->assr_enabled;
+	link->dp.mst_enabled = config->mst_enabled;
 	display->adjust.disable = 1;
 	link->adjust.auth_delay = 3;
 	link->adjust.hdcp1.disable = 0;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h
index 5159b3a5e5b0..09294ff122fe 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h
@@ -69,7 +69,7 @@ void hdcp_update_display(struct hdcp_workqueue *hdcp_work,
 
 void hdcp_reset_display(struct hdcp_workqueue *work, unsigned int link_index);
 void hdcp_handle_cpirq(struct hdcp_workqueue *work, unsigned int link_index);
-void hdcp_destroy(struct hdcp_workqueue *work);
+void hdcp_destroy(struct kobject *kobj, struct hdcp_workqueue *work);
 
 struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev, struct cp_psp *cp_psp, struct dc *dc);
 
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
index f6f487e9fe2d..5750818db8f6 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
@@ -25,7 +25,6 @@
 
 #include <linux/string.h>
 #include <linux/acpi.h>
-#include <linux/version.h>
 #include <linux/i2c.h>
 
 #include <drm/drm_probe_helper.h>
@@ -527,11 +526,11 @@ bool dm_helpers_submit_i2c(
 bool dm_helpers_dp_write_dsc_enable(
 		struct dc_context *ctx,
 		const struct dc_stream_state *stream,
-		bool enable
-)
+		bool enable)
 {
 	uint8_t enable_dsc = enable ? 1 : 0;
 	struct amdgpu_dm_connector *aconnector;
+	uint8_t ret = 0;
 
 	if (!stream)
 		return false;
@@ -542,13 +541,13 @@ bool dm_helpers_dp_write_dsc_enable(
 		if (!aconnector->dsc_aux)
 			return false;
 
-		return (drm_dp_dpcd_write(aconnector->dsc_aux, DP_DSC_ENABLE, &enable_dsc, 1) >= 0);
+		ret = drm_dp_dpcd_write(aconnector->dsc_aux, DP_DSC_ENABLE, &enable_dsc, 1);
 	}
 
 	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT)
 		return dm_helpers_dp_write_dpcd(ctx, stream->link, DP_DSC_ENABLE, &enable_dsc, 1);
 
-	return false;
+	return (ret > 0);
 }
 
 bool dm_helpers_is_dp_sink_present(struct dc_link *link)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c
index 26ed70e5538a..e0000c180ed1 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c
@@ -662,6 +662,20 @@ static int amdgpu_dm_set_crtc_irq_state(struct amdgpu_device *adev,
 		__func__);
 }
 
+static int amdgpu_dm_set_vline0_irq_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int crtc_id,
+					enum amdgpu_interrupt_state state)
+{
+	return dm_irq_state(
+		adev,
+		source,
+		crtc_id,
+		state,
+		IRQ_TYPE_VLINE0,
+		__func__);
+}
+
 static int amdgpu_dm_set_vupdate_irq_state(struct amdgpu_device *adev,
 					   struct amdgpu_irq_src *source,
 					   unsigned int crtc_id,
@@ -681,6 +695,11 @@ static const struct amdgpu_irq_src_funcs dm_crtc_irq_funcs = {
 	.process = amdgpu_dm_irq_handler,
 };
 
+static const struct amdgpu_irq_src_funcs dm_vline0_irq_funcs = {
+	.set = amdgpu_dm_set_vline0_irq_state,
+	.process = amdgpu_dm_irq_handler,
+};
+
 static const struct amdgpu_irq_src_funcs dm_vupdate_irq_funcs = {
 	.set = amdgpu_dm_set_vupdate_irq_state,
 	.process = amdgpu_dm_irq_handler,
@@ -702,6 +721,9 @@ void amdgpu_dm_set_irq_funcs(struct amdgpu_device *adev)
 	adev->crtc_irq.num_types = adev->mode_info.num_crtc;
 	adev->crtc_irq.funcs = &dm_crtc_irq_funcs;
 
+	adev->vline0_irq.num_types = adev->mode_info.num_crtc;
+	adev->vline0_irq.funcs = &dm_vline0_irq_funcs;
+
 	adev->vupdate_irq.num_types = adev->mode_info.num_crtc;
 	adev->vupdate_irq.funcs = &dm_vupdate_irq_funcs;
 
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 8ab0b9060d2b..41b09ab22233 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -23,7 +23,6 @@
  *
  */
 
-#include <linux/version.h>
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_dp_mst_helper.h>
@@ -833,6 +832,9 @@ bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 		if (computed_streams[i])
 			continue;
 
+		if (dcn20_remove_stream_from_ctx(stream->ctx->dc, dc_state, stream) != DC_OK)
+			return false;
+
 		mutex_lock(&aconnector->mst_mgr.lock);
 		if (!compute_mst_dsc_configs_for_link(state, dc_state, stream->link)) {
 			mutex_unlock(&aconnector->mst_mgr.lock);
@@ -850,7 +852,8 @@ bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 		stream = dc_state->streams[i];
 
 		if (stream->timing.flags.DSC == 1)
-			dc_stream_add_dsc_to_resource(stream->ctx->dc, dc_state, stream);
+			if (dc_stream_add_dsc_to_resource(stream->ctx->dc, dc_state, stream) != DC_OK)
+				return false;
 	}
 
 	return true;
diff --git a/drivers/gpu/drm/amd/display/dc/Makefile b/drivers/gpu/drm/amd/display/dc/Makefile
index bf8fe0471b8f..5bf2f2375b40 100644
--- a/drivers/gpu/drm/amd/display/dc/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/Makefile
@@ -69,5 +69,7 @@ AMD_DISPLAY_FILES += $(AMD_DISPLAY_CORE)
 AMD_DISPLAY_FILES += $(AMD_DM_REG_UPDATE)
 
 DC_DMUB += dc_dmub_srv.o
+DC_EDID += dc_edid_parser.o
 AMD_DISPLAY_DMUB = $(addprefix $(AMDDALPATH)/dc/,$(DC_DMUB))
-AMD_DISPLAY_FILES += $(AMD_DISPLAY_DMUB)
+AMD_DISPLAY_EDID = $(addprefix $(AMDDALPATH)/dc/,$(DC_EDID))
+AMD_DISPLAY_FILES += $(AMD_DISPLAY_DMUB) $(AMD_DISPLAY_EDID)
diff --git a/drivers/gpu/drm/amd/display/dc/basics/conversion.c b/drivers/gpu/drm/amd/display/dc/basics/conversion.c
index 24ed03d8cda7..6767fab55c26 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/conversion.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/conversion.c
@@ -73,12 +73,9 @@ uint16_t fixed_point_to_int_frac(
 
 	return result;
 }
-/**
-* convert_float_matrix
-* This converts a double into HW register spec defined format S2D13.
-* @param :
-* @return None
-*/
+/*
+ * convert_float_matrix - This converts a double into HW register spec defined format S2D13.
+ */
 void convert_float_matrix(
 	uint16_t *matrix,
 	struct fixed31_32 *flt,
diff --git a/drivers/gpu/drm/amd/display/dc/basics/dc_common.c b/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
index b2fc4f8e6482..ad04ef98e652 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
@@ -49,20 +49,24 @@ bool is_rgb_cspace(enum dc_color_space output_color_space)
 	}
 }
 
-bool is_lower_pipe_tree_visible(struct pipe_ctx *pipe_ctx)
+bool is_child_pipe_tree_visible(struct pipe_ctx *pipe_ctx)
 {
 	if (pipe_ctx->plane_state && pipe_ctx->plane_state->visible)
 		return true;
-	if (pipe_ctx->bottom_pipe && is_lower_pipe_tree_visible(pipe_ctx->bottom_pipe))
+	if (pipe_ctx->bottom_pipe && is_child_pipe_tree_visible(pipe_ctx->bottom_pipe))
+		return true;
+	if (pipe_ctx->next_odm_pipe && is_child_pipe_tree_visible(pipe_ctx->next_odm_pipe))
 		return true;
 	return false;
 }
 
-bool is_upper_pipe_tree_visible(struct pipe_ctx *pipe_ctx)
+bool is_parent_pipe_tree_visible(struct pipe_ctx *pipe_ctx)
 {
 	if (pipe_ctx->plane_state && pipe_ctx->plane_state->visible)
 		return true;
-	if (pipe_ctx->top_pipe && is_upper_pipe_tree_visible(pipe_ctx->top_pipe))
+	if (pipe_ctx->top_pipe && is_parent_pipe_tree_visible(pipe_ctx->top_pipe))
+		return true;
+	if (pipe_ctx->prev_odm_pipe && is_parent_pipe_tree_visible(pipe_ctx->prev_odm_pipe))
 		return true;
 	return false;
 }
@@ -71,9 +75,13 @@ bool is_pipe_tree_visible(struct pipe_ctx *pipe_ctx)
 {
 	if (pipe_ctx->plane_state && pipe_ctx->plane_state->visible)
 		return true;
-	if (pipe_ctx->top_pipe && is_upper_pipe_tree_visible(pipe_ctx->top_pipe))
+	if (pipe_ctx->top_pipe && is_parent_pipe_tree_visible(pipe_ctx->top_pipe))
+		return true;
+	if (pipe_ctx->bottom_pipe && is_child_pipe_tree_visible(pipe_ctx->bottom_pipe))
+		return true;
+	if (pipe_ctx->prev_odm_pipe && is_parent_pipe_tree_visible(pipe_ctx->prev_odm_pipe))
 		return true;
-	if (pipe_ctx->bottom_pipe && is_lower_pipe_tree_visible(pipe_ctx->bottom_pipe))
+	if (pipe_ctx->next_odm_pipe && is_child_pipe_tree_visible(pipe_ctx->next_odm_pipe))
 		return true;
 	return false;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/basics/dc_common.h b/drivers/gpu/drm/amd/display/dc/basics/dc_common.h
index 7c0cbf47e8ce..b061497480b8 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/dc_common.h
+++ b/drivers/gpu/drm/amd/display/dc/basics/dc_common.h
@@ -30,9 +30,9 @@
 
 bool is_rgb_cspace(enum dc_color_space output_color_space);
 
-bool is_lower_pipe_tree_visible(struct pipe_ctx *pipe_ctx);
+bool is_child_pipe_tree_visible(struct pipe_ctx *pipe_ctx);
 
-bool is_upper_pipe_tree_visible(struct pipe_ctx *pipe_ctx);
+bool is_parent_pipe_tree_visible(struct pipe_ctx *pipe_ctx);
 
 bool is_pipe_tree_visible(struct pipe_ctx *pipe_ctx);
 
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c
index 23a373ca94b5..c67d21a5ee52 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c
@@ -911,11 +911,11 @@ static enum bp_result get_ss_info_from_tbl(
  * ver 2.1 can co-exist with SS_Info table. Expect ASIC_InternalSS_Info ver 3.1,
  * there is only one entry for each signal /ss id.  However, there is
  * no planning of supporting multiple spread Sprectum entry for EverGreen
- * @param [in] this
- * @param [in] signal, ASSignalType to be converted to info index
- * @param [in] index, number of entries that match the converted info index
- * @param [out] ss_info, sprectrum information structure,
- * @return Bios parser result code
+ * @dcb:     pointer to the DC BIOS
+ * @signal:  ASSignalType to be converted to info index
+ * @index:   number of entries that match the converted info index
+ * @ss_info: sprectrum information structure,
+ * return:   Bios parser result code
  */
 static enum bp_result bios_parser_get_spread_spectrum_info(
 	struct dc_bios *dcb,
@@ -985,10 +985,10 @@ static enum bp_result get_ss_info_from_internal_ss_info_tbl_V2_1(
  * There can not be more than 1 entry for  ASIC_InternalSS_Info Ver 2.1 or
  * SS_Info.
  *
- * @param this
- * @param id, spread sprectrum info index
- * @param pSSinfo, sprectrum information structure,
- * @return Bios parser result code
+ * @bp:      pointer to the BIOS parser
+ * @id:      spread sprectrum info index
+ * @ss_info: sprectrum information structure,
+ * return:   BIOS parser result code
  */
 static enum bp_result get_ss_info_from_tbl(
 	struct bios_parser *bp,
@@ -1011,9 +1011,10 @@ static enum bp_result get_ss_info_from_tbl(
  * from the VBIOS
  * There will not be multiple entry for Ver 2.1
  *
- * @param id, spread sprectrum info index
- * @param pSSinfo, sprectrum information structure,
- * @return Bios parser result code
+ * @bp:    pointer to the Bios parser
+ * @id:    spread sprectrum info index
+ * @info:  sprectrum information structure,
+ * return: Bios parser result code
  */
 static enum bp_result get_ss_info_from_internal_ss_info_tbl_V2_1(
 	struct bios_parser *bp,
@@ -1076,9 +1077,10 @@ static enum bp_result get_ss_info_from_internal_ss_info_tbl_V2_1(
  * of entries that matches the id
  * for, the SS_Info table, there should not be more than 1 entry match.
  *
- * @param [in] id, spread sprectrum id
- * @param [out] pSSinfo, sprectrum information structure,
- * @return Bios parser result code
+ * @bp:      pointer to the Bios parser
+ * @id:      spread sprectrum id
+ * @ss_info: sprectrum information structure,
+ * return:   Bios parser result code
  */
 static enum bp_result get_ss_info_from_ss_info_table(
 	struct bios_parser *bp,
@@ -1451,16 +1453,14 @@ static enum bp_result get_embedded_panel_info_v1_3(
 }
 
 /**
- * bios_parser_get_encoder_cap_info
+ * bios_parser_get_encoder_cap_info - get encoder capability
+ *                                    information of input object id
  *
- * @brief
- *  Get encoder capability information of input object id
- *
- * @param object_id, Object id
- * @param object_id, encoder cap information structure
- *
- * @return Bios parser result code
+ * @dcb:       pointer to the DC BIOS
+ * @object_id: object id
+ * @info:      encoder cap information structure
  *
+ * return: Bios parser result code
  */
 static enum bp_result bios_parser_get_encoder_cap_info(
 	struct dc_bios *dcb,
@@ -1490,17 +1490,12 @@ static enum bp_result bios_parser_get_encoder_cap_info(
 }
 
 /**
- * get_encoder_cap_record
- *
- * @brief
- *  Get encoder cap record for the object
- *
- * @param object, ATOM object
+ * get_encoder_cap_record - Get encoder cap record for the object
  *
- * @return atom encoder cap record
- *
- * @note
- *  search all records to find the ATOM_ENCODER_CAP_RECORD_V2 record
+ * @bp:      pointer to the BIOS parser
+ * @object:  ATOM object
+ * return:   atom encoder cap record
+ * note:     search all records to find the ATOM_ENCODER_CAP_RECORD_V2 record
  */
 static ATOM_ENCODER_CAP_RECORD_V2 *get_encoder_cap_record(
 	struct bios_parser *bp,
@@ -1557,8 +1552,9 @@ static uint32_t get_ss_entry_number_from_ss_info_tbl(
  * Get Number of SpreadSpectrum Entry from the ASIC_InternalSS_Info table from
  * the VBIOS that match the SSid (to be converted from signal)
  *
- * @param[in] signal, ASSignalType to be converted to SSid
- * @return number of SS Entry that match the signal
+ * @dcb:    pointer to the DC BIOS
+ * @signal: ASSignalType to be converted to SSid
+ * return: number of SS Entry that match the signal
  */
 static uint32_t bios_parser_get_ss_entry_number(
 	struct dc_bios *dcb,
@@ -1608,10 +1604,10 @@ static uint32_t bios_parser_get_ss_entry_number(
  * get_ss_entry_number_from_ss_info_tbl
  * Get Number of spread spectrum entry from the SS_Info table from the VBIOS.
  *
- * @note There can only be one entry for each id for SS_Info Table
- *
- * @param [in] id, spread spectrum id
- * @return number of SS Entry that match the id
+ * @bp:  pointer to the BIOS parser
+ * @id:  spread spectrum id
+ * return: number of SS Entry that match the id
+ * note: There can only be one entry for each id for SS_Info Table
  */
 static uint32_t get_ss_entry_number_from_ss_info_tbl(
 	struct bios_parser *bp,
@@ -1679,8 +1675,9 @@ static uint32_t get_ss_entry_number_from_ss_info_tbl(
  * There can not be more than 1 entry for  ASIC_InternalSS_Info Ver 2.1 or
  * SS_Info.
  *
- * @param id, spread sprectrum info index
- * @return Bios parser result code
+ * @bp:    pointer to the BIOS parser
+ * @id:    spread sprectrum info index
+ * return: Bios parser result code
  */
 static uint32_t get_ss_entry_number(struct bios_parser *bp, uint32_t id)
 {
@@ -1696,8 +1693,9 @@ static uint32_t get_ss_entry_number(struct bios_parser *bp, uint32_t id)
  * Ver 2.1 from the VBIOS
  * There will not be multiple entry for Ver 2.1
  *
- * @param id, spread sprectrum info index
- * @return number of SS Entry that match the id
+ * @bp:    pointer to the BIOS parser
+ * @id:    spread sprectrum info index
+ * return: number of SS Entry that match the id
  */
 static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_v2_1(
 	struct bios_parser *bp,
@@ -1731,8 +1729,9 @@ static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_v2_1(
  * Get Number of SpreadSpectrum Entry from the ASIC_InternalSS_Info table of
  * the VBIOS that matches id
  *
- * @param[in]  id, spread sprectrum id
- * @return number of SS Entry that match the id
+ * @bp:    pointer to the BIOS parser
+ * @id:    spread sprectrum id
+ * return: number of SS Entry that match the id
  */
 static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_V3_1(
 	struct bios_parser *bp,
@@ -1767,10 +1766,11 @@ static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_V3_1(
  * bios_parser_get_gpio_pin_info
  * Get GpioPin information of input gpio id
  *
- * @param gpio_id, GPIO ID
- * @param info, GpioPin information structure
- * @return Bios parser result code
- * @note
+ * @dcb:     pointer to the DC BIOS
+ * @gpio_id: GPIO ID
+ * @info:    GpioPin information structure
+ * return:   Bios parser result code
+ * note:
  *  to get the GPIO PIN INFO, we need:
  *  1. get the GPIO_ID from other object table, see GetHPDInfo()
  *  2. in DATA_TABLE.GPIO_Pin_LUT, search all records, to get the registerA
@@ -2197,13 +2197,10 @@ static uint32_t get_support_mask_for_device_id(struct device_id device_id)
 }
 
 /**
- * bios_parser_set_scratch_critical_state
- *
- * @brief
- *  update critical state bit in VBIOS scratch register
- *
- * @param
- *  bool - to set or reset state
+ * bios_parser_set_scratch_critical_state - update critical state
+ *                                          bit in VBIOS scratch register
+ * @dcb:    pointer to the DC BIOS
+ * @state:  set or reset state
  */
 static void bios_parser_set_scratch_critical_state(
 	struct dc_bios *dcb,
@@ -2222,7 +2219,7 @@ static void bios_parser_set_scratch_critical_state(
  * bios_parser *bp - [in]BIOS parser handler to get master data table
  * integrated_info *info - [out] store and output integrated info
  *
- * @return
+ * return:
  * enum bp_result - BP_RESULT_OK if information is available,
  *                  BP_RESULT_BADBIOSTABLE otherwise.
  */
@@ -2372,7 +2369,7 @@ static enum bp_result get_integrated_info_v8(
  * bios_parser *bp - [in]BIOS parser handler to get master data table
  * integrated_info *info - [out] store and output integrated info
  *
- * @return
+ * return:
  * enum bp_result - BP_RESULT_OK if information is available,
  *                  BP_RESULT_BADBIOSTABLE otherwise.
  */
@@ -2509,7 +2506,7 @@ static enum bp_result get_integrated_info_v9(
  * bios_parser *bp - [in]BIOS parser handler to get master data table
  * integrated_info *info - [out] store and output integrated info
  *
- * @return
+ * return:
  * enum bp_result - BP_RESULT_OK if information is available,
  *                  BP_RESULT_BADBIOSTABLE otherwise.
  */
@@ -2585,7 +2582,7 @@ static struct integrated_info *bios_parser_create_integrated_info(
 	return NULL;
 }
 
-enum bp_result update_slot_layout_info(
+static enum bp_result update_slot_layout_info(
 	struct dc_bios *dcb,
 	unsigned int i,
 	struct slot_layout_info *slot_layout_info,
@@ -2689,7 +2686,7 @@ enum bp_result update_slot_layout_info(
 }
 
 
-enum bp_result get_bracket_layout_record(
+static enum bp_result get_bracket_layout_record(
 	struct dc_bios *dcb,
 	unsigned int bracket_layout_id,
 	struct slot_layout_info *slot_layout_info)
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
index 670c26583817..9f9fda3118d1 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
@@ -485,10 +485,11 @@ static struct atom_hpd_int_record *get_hpd_record(
  * bios_parser_get_gpio_pin_info
  * Get GpioPin information of input gpio id
  *
- * @param gpio_id, GPIO ID
- * @param info, GpioPin information structure
- * @return Bios parser result code
- * @note
+ * @dcb:     pointer to the DC BIOS
+ * @gpio_id: GPIO ID
+ * @info:    GpioPin information structure
+ * return: Bios parser result code
+ * note:
  *  to get the GPIO PIN INFO, we need:
  *  1. get the GPIO_ID from other object table, see GetHPDInfo()
  *  2. in DATA_TABLE.GPIO_Pin_LUT, search all records,
@@ -801,11 +802,11 @@ static enum bp_result get_ss_info_v4_2(
  * ver 3.1,
  * there is only one entry for each signal /ss id.  However, there is
  * no planning of supporting multiple spread Sprectum entry for EverGreen
- * @param [in] this
- * @param [in] signal, ASSignalType to be converted to info index
- * @param [in] index, number of entries that match the converted info index
- * @param [out] ss_info, sprectrum information structure,
- * @return Bios parser result code
+ * @dcb:     pointer to the DC BIOS
+ * @signal:  ASSignalType to be converted to info index
+ * @index:   number of entries that match the converted info index
+ * @ss_info: sprectrum information structure,
+ * return: Bios parser result code
  */
 static enum bp_result bios_parser_get_spread_spectrum_info(
 	struct dc_bios *dcb,
@@ -1196,13 +1197,11 @@ static bool bios_parser_is_accelerated_mode(
 }
 
 /**
- * bios_parser_set_scratch_critical_state
+ * bios_parser_set_scratch_critical_state - update critical state bit
+ *                                          in VBIOS scratch register
  *
- * @brief
- *  update critical state bit in VBIOS scratch register
- *
- * @param
- *  bool - to set or reset state
+ * @dcb:   pointer to the DC BIO
+ * @state: set or reset state
  */
 static void bios_parser_set_scratch_critical_state(
 	struct dc_bios *dcb,
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table.c b/drivers/gpu/drm/amd/display/dc/bios/command_table.c
index 070459e3e407..afc10b954ffa 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table.c
@@ -245,6 +245,23 @@ static enum bp_result encoder_control_digx_v3(
 					cntl->enable_dp_audio);
 	params.ucLaneNum = (uint8_t)(cntl->lanes_number);
 
+	switch (cntl->color_depth) {
+	case COLOR_DEPTH_888:
+		params.ucBitPerColor = PANEL_8BIT_PER_COLOR;
+		break;
+	case COLOR_DEPTH_101010:
+		params.ucBitPerColor = PANEL_10BIT_PER_COLOR;
+		break;
+	case COLOR_DEPTH_121212:
+		params.ucBitPerColor = PANEL_12BIT_PER_COLOR;
+		break;
+	case COLOR_DEPTH_161616:
+		params.ucBitPerColor = PANEL_16BIT_PER_COLOR;
+		break;
+	default:
+		break;
+	}
+
 	if (EXEC_BIOS_CMD_TABLE(DIGxEncoderControl, params))
 		result = BP_RESULT_OK;
 
@@ -274,6 +291,23 @@ static enum bp_result encoder_control_digx_v4(
 					cntl->enable_dp_audio));
 	params.ucLaneNum = (uint8_t)(cntl->lanes_number);
 
+	switch (cntl->color_depth) {
+	case COLOR_DEPTH_888:
+		params.ucBitPerColor = PANEL_8BIT_PER_COLOR;
+		break;
+	case COLOR_DEPTH_101010:
+		params.ucBitPerColor = PANEL_10BIT_PER_COLOR;
+		break;
+	case COLOR_DEPTH_121212:
+		params.ucBitPerColor = PANEL_12BIT_PER_COLOR;
+		break;
+	case COLOR_DEPTH_161616:
+		params.ucBitPerColor = PANEL_16BIT_PER_COLOR;
+		break;
+	default:
+		break;
+	}
+
 	if (EXEC_BIOS_CMD_TABLE(DIGxEncoderControl, params))
 		result = BP_RESULT_OK;
 
@@ -1057,6 +1091,19 @@ static enum bp_result set_pixel_clock_v5(
 		 * driver choose program it itself, i.e. here we program it
 		 * to 888 by default.
 		 */
+		if (bp_params->signal_type == SIGNAL_TYPE_HDMI_TYPE_A)
+			switch (bp_params->color_depth) {
+			case TRANSMITTER_COLOR_DEPTH_30:
+				/* yes this is correct, the atom define is wrong */
+				clk.sPCLKInput.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_32BPP;
+				break;
+			case TRANSMITTER_COLOR_DEPTH_36:
+				/* yes this is correct, the atom define is wrong */
+				clk.sPCLKInput.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_30BPP;
+				break;
+			default:
+				break;
+			}
 
 		if (EXEC_BIOS_CMD_TABLE(SetPixelClock, clk))
 			result = BP_RESULT_OK;
@@ -1135,6 +1182,20 @@ static enum bp_result set_pixel_clock_v6(
 		 * driver choose program it itself, i.e. here we pass required
 		 * target rate that includes deep color.
 		 */
+		if (bp_params->signal_type == SIGNAL_TYPE_HDMI_TYPE_A)
+			switch (bp_params->color_depth) {
+			case TRANSMITTER_COLOR_DEPTH_30:
+				clk.sPCLKInput.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_30BPP_V6;
+				break;
+			case TRANSMITTER_COLOR_DEPTH_36:
+				clk.sPCLKInput.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_36BPP_V6;
+				break;
+			case TRANSMITTER_COLOR_DEPTH_48:
+				clk.sPCLKInput.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_48BPP;
+				break;
+			default:
+				break;
+			}
 
 		if (EXEC_BIOS_CMD_TABLE(SetPixelClock, clk))
 			result = BP_RESULT_OK;
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c
index 48b4ef03fc8f..5b77251e0590 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c
@@ -114,18 +114,14 @@ bool dal_cmd_table_helper_controller_id_to_atom(
 }
 
 /**
-* translate_transmitter_bp_to_atom
-*
-* @brief
-*  Translate the Transmitter to the corresponding ATOM BIOS value
-*
-* @param
-*   input transmitter
-*   output digitalTransmitter
-*    // =00: Digital Transmitter1 ( UNIPHY linkAB )
-*    // =01: Digital Transmitter2 ( UNIPHY linkCD )
-*    // =02: Digital Transmitter3 ( UNIPHY linkEF )
-*/
+ * translate_transmitter_bp_to_atom - Translate the Transmitter to the
+ *                                    corresponding ATOM BIOS value
+ * @t: transmitter
+ * returns: output digitalTransmitter
+ *    // =00: Digital Transmitter1 ( UNIPHY linkAB )
+ *    // =01: Digital Transmitter2 ( UNIPHY linkCD )
+ *    // =02: Digital Transmitter3 ( UNIPHY linkEF )
+ */
 uint8_t dal_cmd_table_helper_transmitter_bp_to_atom(
 	enum transmitter t)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
index 7736c92d55c4..455ee2be15a3 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
@@ -128,18 +128,14 @@ bool dal_cmd_table_helper_controller_id_to_atom2(
 }
 
 /**
-* translate_transmitter_bp_to_atom
-*
-* @brief
-*  Translate the Transmitter to the corresponding ATOM BIOS value
-*
-* @param
-*   input transmitter
-*   output digitalTransmitter
-*    // =00: Digital Transmitter1 ( UNIPHY linkAB )
-*    // =01: Digital Transmitter2 ( UNIPHY linkCD )
-*    // =02: Digital Transmitter3 ( UNIPHY linkEF )
-*/
+ * translate_transmitter_bp_to_atom2 - Translate the Transmitter to the
+ *                                     corresponding ATOM BIOS value
+ *  @t: transmitter
+ *  returns: digitalTransmitter
+ *    // =00: Digital Transmitter1 ( UNIPHY linkAB )
+ *    // =01: Digital Transmitter2 ( UNIPHY linkCD )
+ *    // =02: Digital Transmitter3 ( UNIPHY linkEF )
+ */
 uint8_t dal_cmd_table_helper_transmitter_bp_to_atom2(
 	enum transmitter t)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/Makefile b/drivers/gpu/drm/amd/display/dc/calcs/Makefile
index 64f515d74410..f3c00f479e1c 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/calcs/Makefile
@@ -33,10 +33,6 @@ ifdef CONFIG_PPC64
 calcs_ccflags := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-calcs_rcflags := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c b/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c
index ef41b287cbe2..e633f8a51edb 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c
@@ -106,7 +106,6 @@ static void calculate_bandwidth(
 	bool lpt_enabled;
 	enum bw_defines sclk_message;
 	enum bw_defines yclk_message;
-	enum bw_defines v_filter_init_mode[maximum_number_of_surfaces];
 	enum bw_defines tiling_mode[maximum_number_of_surfaces];
 	enum bw_defines surface_type[maximum_number_of_surfaces];
 	enum bw_defines voltage;
@@ -792,12 +791,8 @@ static void calculate_bandwidth(
 				data->v_filter_init[i] = bw_add(data->v_filter_init[i], bw_int_to_fixed(1));
 			}
 			if (data->stereo_mode[i] == bw_def_top_bottom) {
-				v_filter_init_mode[i] = bw_def_manual;
 				data->v_filter_init[i] = bw_min2(data->v_filter_init[i], bw_int_to_fixed(4));
 			}
-			else {
-				v_filter_init_mode[i] = bw_def_auto;
-			}
 			if (data->stereo_mode[i] == bw_def_top_bottom) {
 				data->num_lines_at_frame_start = bw_int_to_fixed(1);
 			}
@@ -2730,7 +2725,7 @@ void bw_calcs_init(struct bw_calcs_dceip *bw_dceip,
 
 }
 
-/**
+/*
  * Compare calculated (required) clocks against the clocks available at
  * maximum voltage (max Performance Level).
  */
@@ -3001,13 +2996,12 @@ static bool all_displays_in_sync(const struct pipe_ctx pipe[],
 	return true;
 }
 
-/**
+/*
  * Return:
  *	true -	Display(s) configuration supported.
  *		In this case 'calcs_output' contains data for HW programming
  *	false - Display(s) configuration not supported (not enough bandwidth).
  */
-
 bool bw_calcs(struct dc_context *ctx,
 	const struct bw_calcs_dceip *dceip,
 	const struct bw_calcs_vbios *vbios,
@@ -3028,7 +3022,7 @@ bool bw_calcs(struct dc_context *ctx,
 		calcs_output->all_displays_in_sync = false;
 
 	if (data->number_of_displays != 0) {
-		uint8_t yclk_lvl, sclk_lvl;
+		uint8_t yclk_lvl;
 		struct bw_fixed high_sclk = vbios->high_sclk;
 		struct bw_fixed mid1_sclk = vbios->mid1_sclk;
 		struct bw_fixed mid2_sclk = vbios->mid2_sclk;
@@ -3049,7 +3043,6 @@ bool bw_calcs(struct dc_context *ctx,
 		calculate_bandwidth(dceip, vbios, data);
 
 		yclk_lvl = data->y_clk_level;
-		sclk_lvl = data->sclk_level;
 
 		calcs_output->nbp_state_change_enable =
 			data->nbp_state_change_enable;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile b/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
index d59b380e7b7f..ff96bee57bfc 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
@@ -104,13 +104,6 @@ ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/clk_mgr/dcn21/rn_clk_mgr.o := $(call cc-option,-mno-gnu-attribute)
 endif
 
-# prevent build errors:
-# ...: '-mgeneral-regs-only' is incompatible with the use of floating-point types
-# this file is unused on arm64, just like on ppc64
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/clk_mgr/dcn21/rn_clk_mgr.o := -mgeneral-regs-only
-endif
-
 AMD_DAL_CLK_MGR_DCN21 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn21/,$(CLK_MGR_DCN21))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN21)
@@ -125,13 +118,6 @@ ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/clk_mgr/dcn30/dcn30_clk_mgr.o := $(call cc-option,-mno-gnu-attribute)
 endif
 
-# prevent build errors:
-# ...: '-mgeneral-regs-only' is incompatible with the use of floating-point types
-# this file is unused on arm64, just like on ppc64
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/clk_mgr/dcn30/dcn30_clk_mgr.o := -mgeneral-regs-only
-endif
-
 AMD_DAL_CLK_MGR_DCN30 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn30/,$(CLK_MGR_DCN30))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN30)
@@ -146,13 +132,6 @@ ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/clk_mgr/dcn301/vg_clk_mgr.o := $(call cc-option,-mno-gnu-attribute)
 endif
 
-# prevent build errors:
-# ...: '-mgeneral-regs-only' is incompatible with the use of floating-point types
-# this file is unused on arm64, just like on ppc64
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/clk_mgr/dcn301/vg_clk_mgr.o := -mgeneral-regs-only
-endif
-
 AMD_DAL_CLK_MGR_DCN301 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn301/,$(CLK_MGR_DCN301))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN301)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
index 75b8240ed059..e133edc587d3 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
@@ -187,17 +187,6 @@ static void ramp_up_dispclk_with_dpp(
 	clk_mgr->base.clks.max_supported_dppclk_khz = new_clocks->max_supported_dppclk_khz;
 }
 
-static bool is_mpo_enabled(struct dc_state *context)
-{
-	int i;
-
-	for (i = 0; i < context->stream_count; i++) {
-		if (context->stream_status[i].plane_count > 1)
-			return true;
-	}
-	return false;
-}
-
 static void rv1_update_clocks(struct clk_mgr *clk_mgr_base,
 			struct dc_state *context,
 			bool safe_to_lower)
@@ -295,22 +284,9 @@ static void rv1_update_clocks(struct clk_mgr *clk_mgr_base,
 		if (pp_smu->set_hard_min_fclk_by_freq &&
 				pp_smu->set_hard_min_dcfclk_by_freq &&
 				pp_smu->set_min_deep_sleep_dcfclk) {
-			// Only increase clocks when display is active and MPO is enabled
-			if (display_count && is_mpo_enabled(context)) {
-				pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu,
-						((new_clocks->fclk_khz / 1000) *  101) / 100);
-				pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu,
-						((new_clocks->dcfclk_khz / 1000) * 101) / 100);
-				pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu,
-						(new_clocks->dcfclk_deep_sleep_khz + 999) / 1000);
-			} else {
-				pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu,
-						new_clocks->fclk_khz / 1000);
-				pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu,
-						new_clocks->dcfclk_khz / 1000);
-				pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu,
-						(new_clocks->dcfclk_deep_sleep_khz + 999) / 1000);
-			}
+			pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu, new_clocks->fclk_khz / 1000);
+			pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu, new_clocks->dcfclk_khz / 1000);
+			pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu, (new_clocks->dcfclk_deep_sleep_khz + 999) / 1000);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c
index f2114bc910bf..ec9dc265cde0 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c
@@ -257,8 +257,7 @@ void dcn2_update_clocks(struct clk_mgr *clk_mgr_base,
 			if (update_dppclk || update_dispclk)
 				dcn20_update_clocks_update_dentist(clk_mgr);
 			// always update dtos unless clock is lowered and not safe to lower
-			if (new_clocks->dppclk_khz >= dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz)
-				dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
 		}
 	}
 
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c
index 5b466f440d67..c7e5a64e06af 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c
@@ -146,15 +146,15 @@ static noinline void dcn3_build_wm_range_table(struct clk_mgr_internal *clk_mgr)
 	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_uclk = 0xFFFF;
 
 	/* Set D - MALL - SR enter and exit times adjusted for MALL */
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].valid = true;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us = pstate_latency_us;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us = 2;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us = 4;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.wm_type = WATERMARKS_MALL;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_dcfclk = 0;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_dcfclk = 0xFFFF;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_uclk = min_uclk_mhz;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_uclk = 0xFFFF;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].valid = true;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us = pstate_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us = 2;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us = 4;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.wm_type = WATERMARKS_MALL;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_dcfclk = 0;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_uclk = 0xFFFF;
 }
 
 void dcn3_init_clocks(struct clk_mgr *clk_mgr_base)
@@ -251,6 +251,7 @@ static void dcn3_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu;
 	bool force_reset = false;
 	bool update_uclk = false;
+	bool p_state_change_support;
 
 	if (dc->work_arounds.skip_clock_update || !clk_mgr->smu_present)
 		return;
@@ -291,8 +292,9 @@ static void dcn3_update_clocks(struct clk_mgr *clk_mgr_base,
 		clk_mgr_base->clks.socclk_khz = new_clocks->socclk_khz;
 
 	clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support;
-	if (should_update_pstate_support(safe_to_lower, new_clocks->p_state_change_support, clk_mgr_base->clks.p_state_change_support)) {
-		clk_mgr_base->clks.p_state_change_support = new_clocks->p_state_change_support;
+	p_state_change_support = new_clocks->p_state_change_support || (display_count == 0);
+	if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.p_state_change_support)) {
+		clk_mgr_base->clks.p_state_change_support = p_state_change_support;
 
 		/* to disable P-State switching, set UCLK min = max */
 		if (!clk_mgr_base->clks.p_state_change_support)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c
index cfa8e02cf103..68942bbc7472 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c
@@ -103,7 +103,7 @@ int dcn301_smu_send_msg_with_param(
 	/* Trigger the message transaction by writing the message ID */
 	REG_WRITE(MP1_SMN_C2PMSG_67, msg_id);
 
-	result = dcn301_smu_wait_for_response(clk_mgr, 10, 1000);
+	result = dcn301_smu_wait_for_response(clk_mgr, 10, 200000);
 
 	ASSERT(result == VBIOSSMC_Result_OK);
 
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c
index 991b9c5beaa3..aadb801447a7 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c
@@ -75,7 +75,8 @@ int vg_get_active_display_cnt_wa(
 		const struct dc_link *link = dc->links[i];
 
 		/* abusing the fact that the dig and phy are coupled to see if the phy is enabled */
-		if (link->link_enc->funcs->is_dig_enabled(link->link_enc))
+		if (link->link_enc->funcs->is_dig_enabled &&
+				link->link_enc->funcs->is_dig_enabled(link->link_enc))
 			display_count++;
 	}
 
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 58eb0d69873a..8f8a13c7cf73 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -175,6 +175,8 @@ static bool create_links(
 
 	connectors_num = bios->funcs->get_connectors_number(bios);
 
+	DC_LOG_DC("BIOS object table - number of connectors: %d", connectors_num);
+
 	if (connectors_num > ENUM_ID_COUNT) {
 		dm_error(
 			"DC: Number of connectors %d exceeds maximum of %d!\n",
@@ -193,6 +195,8 @@ static bool create_links(
 		struct link_init_data link_init_params = {0};
 		struct dc_link *link;
 
+		DC_LOG_DC("BIOS object table - printing link object info for connector number: %d, link_index: %d", i, dc->link_count);
+
 		link_init_params.ctx = dc->ctx;
 		/* next BIOS object table connector */
 		link_init_params.connector_index = i;
@@ -201,30 +205,14 @@ static bool create_links(
 		link = link_create(&link_init_params);
 
 		if (link) {
-			bool should_destory_link = false;
-
-			if (link->connector_signal == SIGNAL_TYPE_EDP) {
-				if (dc->config.edp_not_connected) {
-					if (!IS_DIAG_DC(dc->ctx->dce_environment))
-						should_destory_link = true;
-				} else {
-					enum dc_connection_type type;
-					dc_link_detect_sink(link, &type);
-					if (type == dc_connection_none)
-						should_destory_link = true;
-				}
-			}
-
-			if (dc->config.force_enum_edp || !should_destory_link) {
 				dc->links[dc->link_count] = link;
 				link->dc = dc;
 				++dc->link_count;
-			} else {
-				link_destroy(&link);
-			}
 		}
 	}
 
+	DC_LOG_DC("BIOS object table - end");
+
 	for (i = 0; i < num_virtual_links; i++) {
 		struct dc_link *link = kzalloc(sizeof(*link), GFP_KERNEL);
 		struct encoder_init_data enc_init = {0};
@@ -284,20 +272,16 @@ static void dc_perf_trace_destroy(struct dc_perf_trace **perf_trace)
 }
 
 /**
- *****************************************************************************
- *  Function: dc_stream_adjust_vmin_vmax
+ *  dc_stream_adjust_vmin_vmax:
  *
- *  @brief
- *     Looks up the pipe context of dc_stream_state and updates the
- *     vertical_total_min and vertical_total_max of the DRR, Dynamic Refresh
- *     Rate, which is a power-saving feature that targets reducing panel
- *     refresh rate while the screen is static
+ *  Looks up the pipe context of dc_stream_state and updates the
+ *  vertical_total_min and vertical_total_max of the DRR, Dynamic Refresh
+ *  Rate, which is a power-saving feature that targets reducing panel
+ *  refresh rate while the screen is static
  *
- *  @param [in] dc: dc reference
- *  @param [in] stream: Initial dc stream state
- *  @param [in] adjust: Updated parameters for vertical_total_min and
- *  vertical_total_max
- *****************************************************************************
+ *  @dc:     dc reference
+ *  @stream: Initial dc stream state
+ *  @adjust: Updated parameters for vertical_total_min and vertical_total_max
  */
 bool dc_stream_adjust_vmin_vmax(struct dc *dc,
 		struct dc_stream_state *stream,
@@ -355,6 +339,7 @@ bool dc_stream_get_crtc_position(struct dc *dc,
  * @dc: DC Object
  * @stream: The stream to configure CRC on.
  * @enable: Enable CRC if true, disable otherwise.
+ * @crc_window: CRC window (x/y start/end) information
  * @continuous: Capture CRC on every frame if true. Otherwise, only capture
  *              once.
  *
@@ -420,7 +405,9 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
  * dc_stream_get_crc() - Get CRC values for the given stream.
  * @dc: DC object
  * @stream: The DC stream state of the stream to get CRCs from.
- * @r_cr, g_y, b_cb: CRC values for the three channels are stored here.
+ * @r_cr: CRC value for the first of the 3 channels stored here.
+ * @g_y:  CRC value for the second of the 3 channels stored here.
+ * @b_cb: CRC value for the third of the 3 channels stored here.
  *
  * dc_stream_configure_crc needs to be called beforehand to enable CRCs.
  * Return false if stream is not found, or if CRCs are not enabled.
@@ -707,7 +694,6 @@ static bool dc_construct(struct dc *dc,
 	}
 
 	dc->dcn_ip = dcn_ip;
-	dc->soc_bounding_box = init_params->soc_bounding_box;
 #endif
 
 	if (!dc_construct_ctx(dc, init_params)) {
@@ -757,6 +743,10 @@ static bool dc_construct(struct dc *dc,
 	if (!dc->res_pool)
 		goto fail;
 
+	/* set i2c speed if not done by the respective dcnxxx__resource.c */
+	if (dc->caps.i2c_speed_in_khz_hdcp == 0)
+		dc->caps.i2c_speed_in_khz_hdcp = dc->caps.i2c_speed_in_khz;
+
 	dc->clk_mgr = dc_clk_mgr_create(dc->ctx, dc->res_pool->pp_smu, dc->res_pool->dccg);
 	if (!dc->clk_mgr)
 		goto fail;
@@ -764,8 +754,6 @@ static bool dc_construct(struct dc *dc,
 	dc->clk_mgr->force_smu_not_present = init_params->force_smu_not_present;
 #endif
 
-	dc->debug.force_ignore_link_settings = init_params->force_ignore_link_settings;
-
 	if (dc->res_pool->funcs->update_bw_bounding_box)
 		dc->res_pool->funcs->update_bw_bounding_box(dc, dc->clk_mgr->bw_params);
 
@@ -803,7 +791,8 @@ static void disable_all_writeback_pipes_for_stream(
 		stream->writeback_info[i].wb_enabled = false;
 }
 
-void apply_ctx_interdependent_lock(struct dc *dc, struct dc_state *context, struct dc_stream_state *stream, bool lock)
+static void apply_ctx_interdependent_lock(struct dc *dc, struct dc_state *context,
+					  struct dc_stream_state *stream, bool lock)
 {
 	int i = 0;
 
@@ -964,19 +953,15 @@ struct dc *dc_create(const struct dc_init_data *init_params)
 	struct dc *dc = kzalloc(sizeof(*dc), GFP_KERNEL);
 	unsigned int full_pipe_count;
 
-	if (NULL == dc)
-		goto alloc_fail;
+	if (!dc)
+		return NULL;
 
 	if (init_params->dce_environment == DCE_ENV_VIRTUAL_HW) {
-		if (false == dc_construct_ctx(dc, init_params)) {
-			dc_destruct(dc);
-			goto construct_fail;
-		}
+		if (!dc_construct_ctx(dc, init_params))
+			goto destruct_dc;
 	} else {
-		if (false == dc_construct(dc, init_params)) {
-			dc_destruct(dc);
-			goto construct_fail;
-		}
+		if (!dc_construct(dc, init_params))
+			goto destruct_dc;
 
 		full_pipe_count = dc->res_pool->pipe_count;
 		if (dc->res_pool->underlay_pipe_index != NO_UNDERLAY_PIPE)
@@ -1007,15 +992,36 @@ struct dc *dc_create(const struct dc_init_data *init_params)
 
 	return dc;
 
-construct_fail:
+destruct_dc:
+	dc_destruct(dc);
 	kfree(dc);
-
-alloc_fail:
 	return NULL;
 }
 
+static void detect_edp_presence(struct dc *dc)
+{
+	struct dc_link *edp_link = get_edp_link(dc);
+	bool edp_sink_present = true;
+
+	if (!edp_link)
+		return;
+
+	if (dc->config.edp_not_connected) {
+			edp_sink_present = false;
+	} else {
+		enum dc_connection_type type;
+		dc_link_detect_sink(edp_link, &type);
+		if (type == dc_connection_none)
+			edp_sink_present = false;
+	}
+
+	edp_link->edp_sink_present = edp_sink_present;
+}
+
 void dc_hardware_init(struct dc *dc)
 {
+
+	detect_edp_presence(dc);
 	if (dc->ctx->dce_environment != DCE_ENV_VIRTUAL_HW)
 		dc->hwss.init_hw(dc);
 }
@@ -1493,7 +1499,7 @@ bool dc_commit_state(struct dc *dc, struct dc_state *context)
 	enum dc_status result = DC_ERROR_UNEXPECTED;
 	int i;
 
-	if (false == context_changed(dc, context))
+	if (!context_changed(dc, context))
 		return DC_OK;
 
 	DC_LOG_DC("%s: %d streams\n",
@@ -1540,7 +1546,7 @@ bool dc_acquire_release_mpc_3dlut(
 		if (found_pipe_idx) {
 			if (acquire && pool->funcs->acquire_post_bldn_3dlut)
 				ret = pool->funcs->acquire_post_bldn_3dlut(res_ctx, pool, mpcc_id, lut, shaper);
-			else if (acquire == false && pool->funcs->release_post_bldn_3dlut)
+			else if (!acquire && pool->funcs->release_post_bldn_3dlut)
 				ret = pool->funcs->release_post_bldn_3dlut(res_ctx, pool, lut, shaper);
 		}
 	}
@@ -2016,7 +2022,7 @@ static enum surface_update_type check_update_surfaces_for_stream(
 	return overall_type;
 }
 
-/**
+/*
  * dc_check_update_surfaces_for_stream() - Determine update type (fast, med, or full)
  *
  * See :c:type:`enum surface_update_type <surface_update_type>` for explanation of update types
@@ -2270,6 +2276,9 @@ static void copy_stream_update_to_stream(struct dc *dc,
 
 	if (update->dither_option)
 		stream->dither_option = *update->dither_option;
+
+	if (update->pending_test_pattern)
+		stream->test_pattern = *update->pending_test_pattern;
 	/* update current stream with writeback info */
 	if (update->wb_update) {
 		int i;
@@ -2366,6 +2375,15 @@ static void commit_planes_do_stream_update(struct dc *dc,
 				}
 			}
 
+			if (stream_update->pending_test_pattern) {
+				dc_link_dp_set_test_pattern(stream->link,
+					stream->test_pattern.type,
+					stream->test_pattern.color_space,
+					stream->test_pattern.p_link_settings,
+					stream->test_pattern.p_custom_pattern,
+					stream->test_pattern.cust_pattern_size);
+			}
+
 			/* Full fe update*/
 			if (update_type == UPDATE_TYPE_FAST)
 				continue;
@@ -2819,7 +2837,7 @@ enum dc_irq_source dc_interrupt_to_irq_source(
 	return dal_irq_service_to_irq_source(dc->res_pool->irqs, src_id, ext_id);
 }
 
-/**
+/*
  * dc_interrupt_set() - Enable/disable an AMD hw interrupt source
  */
 bool dc_interrupt_set(struct dc *dc, enum dc_irq_source src, bool enable)
@@ -2953,7 +2971,7 @@ static bool link_add_remote_sink_helper(struct dc_link *dc_link, struct dc_sink
 	return true;
 }
 
-/**
+/*
  * dc_link_add_remote_sink() - Create a sink and attach it to an existing link
  *
  * EDID length is in bytes
@@ -3016,7 +3034,7 @@ fail_add_sink:
 	return NULL;
 }
 
-/**
+/*
  * dc_link_remove_remote_sink() - Remove a remote sink from a dc_link
  *
  * Note that this just removes the struct dc_sink - it doesn't
@@ -3143,9 +3161,11 @@ void dc_lock_memory_clock_frequency(struct dc *dc)
 			core_link_enable_stream(dc->current_state, &dc->current_state->res_ctx.pipe_ctx[i]);
 }
 
-bool dc_is_plane_eligible_for_idle_optimizaitons(struct dc *dc,
-						 struct dc_plane_state *plane)
+bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc, struct dc_plane_state *plane,
+		struct dc_cursor_attributes *cursor_attr)
 {
+	if (dc->hwss.does_plane_fit_in_mall && dc->hwss.does_plane_fit_in_mall(dc, plane, cursor_attr))
+		return true;
 	return false;
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
index 9e1071b2181f..bd0101013ec8 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
@@ -203,9 +203,21 @@ static bool program_hpd_filter(const struct dc_link *link)
 	return result;
 }
 
+bool dc_link_wait_for_t12(struct dc_link *link)
+{
+	if (link->connector_signal == SIGNAL_TYPE_EDP && link->dc->hwss.edp_wait_for_T12) {
+		link->dc->hwss.edp_wait_for_T12(link);
+
+		return true;
+	}
+
+	return false;
+}
+
 /**
  * dc_link_detect_sink() - Determine if there is a sink connected
  *
+ * @link: pointer to the dc link
  * @type: Returned connection type
  * Does not detect downstream devices, such as MST sinks
  * or display connected through active dongles
@@ -342,7 +354,7 @@ static enum signal_type get_basic_signal_type(struct graphics_object_id encoder,
 	return SIGNAL_TYPE_NONE;
 }
 
-/**
+/*
  * dc_link_is_dp_sink_present() - Check if there is a native DP
  * or passive DP-HDMI dongle connected
  */
@@ -596,8 +608,6 @@ static void query_hdcp_capability(enum signal_type signal, struct dc_link *link)
 	dc_process_hdcp_msg(signal, link, &msg22);
 
 	if (signal == SIGNAL_TYPE_DISPLAY_PORT || signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		enum hdcp_message_status status = HDCP_MESSAGE_UNSUPPORTED;
-
 		msg14.data = &link->hdcp_caps.bcaps.raw;
 		msg14.length = sizeof(link->hdcp_caps.bcaps.raw);
 		msg14.msg_id = HDCP_MESSAGE_ID_READ_BCAPS;
@@ -605,7 +615,7 @@ static void query_hdcp_capability(enum signal_type signal, struct dc_link *link)
 		msg14.link = HDCP_LINK_PRIMARY;
 		msg14.max_retries = 5;
 
-		status = dc_process_hdcp_msg(signal, link, &msg14);
+		dc_process_hdcp_msg(signal, link, &msg14);
 	}
 
 }
@@ -830,7 +840,7 @@ static bool wait_for_entering_dp_alt_mode(struct dc_link *link)
 	return false;
 }
 
-/**
+/*
  * dc_link_detect() - Detect if a sink is attached to a given link
  *
  * link->local_sink is created or destroyed as needed.
@@ -1065,9 +1075,6 @@ static bool dc_link_detect_helper(struct dc_link *link,
 			break;
 		}
 
-		if (link->local_sink->edid_caps.panel_patch.disable_fec)
-			link->ctx->dc->debug.disable_fec = true;
-
 		// Check if edid is the same
 		if ((prev_sink) &&
 		    (edid_status == EDID_THE_SAME || edid_status == EDID_OK))
@@ -1366,13 +1373,17 @@ static bool dc_link_construct(struct dc_link *link,
 	struct dc_context *dc_ctx = init_params->ctx;
 	struct encoder_init_data enc_init_data = { 0 };
 	struct panel_cntl_init_data panel_cntl_init_data = { 0 };
-	struct integrated_info info = {{{ 0 }}};
+	struct integrated_info *info;
 	struct dc_bios *bios = init_params->dc->ctx->dc_bios;
 	const struct dc_vbios_funcs *bp_funcs = bios->funcs;
 	struct bp_disp_connector_caps_info disp_connect_caps_info = { 0 };
 
 	DC_LOGGER_INIT(dc_ctx->logger);
 
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (!info)
+		goto create_fail;
+
 	link->irq_source_hpd = DC_IRQ_SOURCE_INVALID;
 	link->irq_source_hpd_rx = DC_IRQ_SOURCE_INVALID;
 
@@ -1390,10 +1401,12 @@ static bool dc_link_construct(struct dc_link *link,
 	link->link_id =
 		bios->funcs->get_connector_id(bios, init_params->connector_index);
 
+	DC_LOG_DC("BIOS object table - link_id: %d", link->link_id.id);
 
 	if (bios->funcs->get_disp_connector_caps_info) {
 		bios->funcs->get_disp_connector_caps_info(bios, link->link_id, &disp_connect_caps_info);
 		link->is_internal_display = disp_connect_caps_info.INTERNAL_DISPLAY;
+		DC_LOG_DC("BIOS object table - is_internal_display: %d", link->is_internal_display);
 	}
 
 	if (link->link_id.type != OBJECT_TYPE_CONNECTOR) {
@@ -1408,10 +1421,14 @@ static bool dc_link_construct(struct dc_link *link,
 
 	link->hpd_gpio = get_hpd_gpio(link->ctx->dc_bios, link->link_id,
 				      link->ctx->gpio_service);
+
 	if (link->hpd_gpio) {
 		dal_gpio_open(link->hpd_gpio, GPIO_MODE_INTERRUPT);
 		dal_gpio_unlock_pin(link->hpd_gpio);
 		link->irq_source_hpd = dal_irq_get_source(link->hpd_gpio);
+
+		DC_LOG_DC("BIOS object table - hpd_gpio id: %d", link->hpd_gpio->id);
+		DC_LOG_DC("BIOS object table - hpd_gpio en: %d", link->hpd_gpio->en);
 	}
 
 	switch (link->link_id.id) {
@@ -1470,6 +1487,11 @@ static bool dc_link_construct(struct dc_link *link,
 		goto ddc_create_fail;
 	}
 
+	if (!link->ddc->ddc_pin) {
+		DC_ERROR("Failed to get I2C info for connector!\n");
+		goto ddc_create_fail;
+	}
+
 	link->ddc_hw_inst =
 		dal_ddc_get_line(dal_ddc_service_get_ddc_pin(link->ddc));
 
@@ -1508,6 +1530,8 @@ static bool dc_link_construct(struct dc_link *link,
 		goto link_enc_create_fail;
 	}
 
+	DC_LOG_DC("BIOS object table - DP_IS_USB_C: %d", link->link_enc->features.flags.bits.DP_IS_USB_C);
+
 	link->link_enc_hw_inst = link->link_enc->transmitter;
 
 	for (i = 0; i < 4; i++) {
@@ -1530,16 +1554,20 @@ static bool dc_link_construct(struct dc_link *link,
 		if (link->device_tag.dev_id.device_type == DEVICE_TYPE_LCD &&
 		    link->connector_signal == SIGNAL_TYPE_RGB)
 			continue;
+
+		DC_LOG_DC("BIOS object table - device_tag.acpi_device: %d", link->device_tag.acpi_device);
+		DC_LOG_DC("BIOS object table - device_tag.dev_id.device_type: %d", link->device_tag.dev_id.device_type);
+		DC_LOG_DC("BIOS object table - device_tag.dev_id.enum_id: %d", link->device_tag.dev_id.enum_id);
 		break;
 	}
 
 	if (bios->integrated_info)
-		info = *bios->integrated_info;
+		memcpy(info, bios->integrated_info, sizeof(*info));
 
 	/* Look for channel mapping corresponding to connector and device tag */
 	for (i = 0; i < MAX_NUMBER_OF_EXT_DISPLAY_PATH; i++) {
 		struct external_display_path *path =
-			&info.ext_disp_conn_info.path[i];
+			&info->ext_disp_conn_info.path[i];
 
 		if (path->device_connector_id.enum_id == link->link_id.enum_id &&
 		    path->device_connector_id.id == link->link_id.id &&
@@ -1548,10 +1576,14 @@ static bool dc_link_construct(struct dc_link *link,
 			    path->device_acpi_enum == link->device_tag.acpi_device) {
 				link->ddi_channel_mapping = path->channel_mapping;
 				link->chip_caps = path->caps;
+				DC_LOG_DC("BIOS object table - ddi_channel_mapping: 0x%04X", link->ddi_channel_mapping.raw);
+				DC_LOG_DC("BIOS object table - chip_caps: %d", link->chip_caps);
 			} else if (path->device_tag ==
 				   link->device_tag.dev_id.raw_device_tag) {
 				link->ddi_channel_mapping = path->channel_mapping;
 				link->chip_caps = path->caps;
+				DC_LOG_DC("BIOS object table - ddi_channel_mapping: 0x%04X", link->ddi_channel_mapping.raw);
+				DC_LOG_DC("BIOS object table - chip_caps: %d", link->chip_caps);
 			}
 			break;
 		}
@@ -1570,6 +1602,7 @@ static bool dc_link_construct(struct dc_link *link,
 
 	link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
 
+	DC_LOG_DC("BIOS object table - %s finished successfully.\n", __func__);
 	return true;
 device_tag_fail:
 	link->link_enc->funcs->destroy(&link->link_enc);
@@ -1586,6 +1619,9 @@ create_fail:
 		link->hpd_gpio = NULL;
 	}
 
+	DC_LOG_DC("BIOS object table - %s failed.\n", __func__);
+	kfree(info);
+
 	return false;
 }
 
@@ -2487,9 +2523,14 @@ enum dc_status dc_link_validate_mode_timing(
 static struct abm *get_abm_from_stream_res(const struct dc_link *link)
 {
 	int i;
-	struct dc *dc = link->ctx->dc;
+	struct dc *dc = NULL;
 	struct abm *abm = NULL;
 
+	if (!link || !link->ctx)
+		return NULL;
+
+	dc = link->ctx->dc;
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx pipe_ctx = dc->current_state->res_ctx.pipe_ctx[i];
 		struct dc_stream_state *stream = pipe_ctx.stream;
@@ -2561,7 +2602,6 @@ bool dc_link_set_backlight_level(const struct dc_link *link,
 			if (pipe_ctx->plane_state == NULL)
 				frame_ramp = 0;
 		} else {
-			ASSERT(false);
 			return false;
 		}
 
@@ -3128,17 +3168,17 @@ static void update_psp_stream_config(struct pipe_ctx *pipe_ctx, bool dpms_off)
 {
 	struct cp_psp *cp_psp = &pipe_ctx->stream->ctx->cp_psp;
 	if (cp_psp && cp_psp->funcs.update_stream_config) {
-		struct cp_psp_stream_config config;
-
-		memset(&config, 0, sizeof(config));
+		struct cp_psp_stream_config config = {0};
+		enum dp_panel_mode panel_mode =
+				dp_get_panel_mode(pipe_ctx->stream->link);
 
 		config.otg_inst = (uint8_t) pipe_ctx->stream_res.tg->inst;
-		/*stream_enc_inst*/
-		config.stream_enc_inst = (uint8_t) pipe_ctx->stream_res.stream_enc->stream_enc_inst;
-		config.link_enc_inst = pipe_ctx->stream->link->link_enc_hw_inst;
+		config.dig_fe = (uint8_t) pipe_ctx->stream_res.stream_enc->stream_enc_inst;
+		config.dig_be = pipe_ctx->stream->link->link_enc_hw_inst;
 		config.dpms_off = dpms_off;
 		config.dm_stream_ctx = pipe_ctx->stream->dm_stream_context;
-		config.mst_supported = (pipe_ctx->stream->signal ==
+		config.assr_enabled = (panel_mode == DP_PANEL_MODE_EDP);
+		config.mst_enabled = (pipe_ctx->stream->signal ==
 				SIGNAL_TYPE_DISPLAY_PORT_MST);
 		cp_psp->funcs.update_stream_config(cp_psp->handle, &config);
 	}
@@ -3391,10 +3431,7 @@ void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable)
 }
 
 /**
- *****************************************************************************
- *  Function: dc_link_enable_hpd_filter
- *
- *  @brief
+ *  dc_link_enable_hpd_filter:
  *     If enable is true, programs HPD filter on associated HPD line using
  *     delay_on_disconnect/delay_on_connect values dependent on
  *     link->connector_signal
@@ -3402,9 +3439,8 @@ void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable)
  *     If enable is false, programs HPD filter on associated HPD line with no
  *     delays on connect or disconnect
  *
- *  @param [in] link: pointer to the dc link
- *  @param [in] enable: boolean specifying whether to enable hbd
- *****************************************************************************
+ *  @link:   pointer to the dc link
+ *  @enable: boolean specifying whether to enable hbd
  */
 void dc_link_enable_hpd_filter(struct dc_link *link, bool enable)
 {
@@ -3630,7 +3666,7 @@ uint32_t dc_link_bandwidth_kbps(
 	link_bw_kbps *= 8;   /* 8 bits per byte*/
 	link_bw_kbps *= link_setting->lane_count;
 
-	if (dc_link_is_fec_supported(link) && !link->dc->debug.disable_fec) {
+	if (dc_link_should_enable_fec(link)) {
 		/* Account for FEC overhead.
 		 * We have to do it based on caps,
 		 * and not based on FEC being set ready,
@@ -3651,8 +3687,8 @@ uint32_t dc_link_bandwidth_kbps(
 		 * but the difference is minimal and is in a safe direction,
 		 * which all works well around potential ambiguity of DP 1.4a spec.
 		 */
-		link_bw_kbps = mul_u64_u32_shr(BIT_ULL(32) * 970LL / 1000,
-					       link_bw_kbps, 32);
+		long long fec_link_bw_kbps = link_bw_kbps * 970LL;
+		link_bw_kbps = (uint32_t)(div64_s64(fec_link_bw_kbps, 1000LL));
 	}
 
 	return link_bw_kbps;
@@ -3682,3 +3718,19 @@ bool dc_link_is_fec_supported(const struct dc_link *link)
 			!IS_FPGA_MAXIMUS_DC(link->ctx->dce_environment));
 }
 
+bool dc_link_should_enable_fec(const struct dc_link *link)
+{
+	bool is_fec_disable = false;
+	bool ret = false;
+
+	if ((link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT_MST &&
+			link->local_sink &&
+			link->local_sink->edid_caps.panel_patch.disable_fec) ||
+			link->connector_signal == SIGNAL_TYPE_EDP) // Disable FEC for eDP
+		is_fec_disable = true;
+
+	if (dc_link_is_fec_supported(link) && !link->dc->debug.disable_fec && !is_fec_disable)
+		ret = true;
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c
index c5936e064360..ae6484ab567b 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c
@@ -37,12 +37,16 @@
 #include "dc_link_ddc.h"
 #include "dce/dce_aux.h"
 
-/*DP to Dual link DVI converter*/
+#define DC_LOGGER_INIT(logger)
+
+static const uint8_t DP_VGA_DONGLE_BRANCH_DEV_NAME[] = "DpVga";
+/* DP to Dual link DVI converter */
 static const uint8_t DP_DVI_CONVERTER_ID_4[] = "m2DVIa";
 static const uint8_t DP_DVI_CONVERTER_ID_5[] = "3393N2";
 
 #define AUX_POWER_UP_WA_DELAY 500
 #define I2C_OVER_AUX_DEFER_WA_DELAY 70
+#define DPVGA_DONGLE_AUX_DEFER_WA_DELAY 40
 #define I2C_OVER_AUX_DEFER_WA_DELAY_1MS 1
 
 /* CV smart dongle slave address for retrieving supported HDTV modes*/
@@ -194,6 +198,10 @@ static void ddc_service_construct(
 	if (BP_RESULT_OK != dcb->funcs->get_i2c_info(dcb, init_data->id, &i2c_info)) {
 		ddc_service->ddc_pin = NULL;
 	} else {
+		DC_LOGGER_INIT(ddc_service->ctx->logger);
+		DC_LOG_DC("BIOS object table - i2c_line: %d", i2c_info.i2c_line);
+		DC_LOG_DC("BIOS object table - i2c_engine_id: %d", i2c_info.i2c_engine_id);
+
 		hw_info.ddc_channel = i2c_info.i2c_line;
 		if (ddc_service->link != NULL)
 			hw_info.hw_supported = i2c_info.i2c_hw_assist;
@@ -286,6 +294,15 @@ static uint32_t defer_delay_converter_wa(
 {
 	struct dc_link *link = ddc->link;
 
+	if (link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_VGA_CONVERTER &&
+		link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_0080E1 &&
+		!memcmp(link->dpcd_caps.branch_dev_name,
+		    DP_VGA_DONGLE_BRANCH_DEV_NAME,
+			sizeof(link->dpcd_caps.branch_dev_name)))
+
+		return defer_delay > DPVGA_DONGLE_AUX_DEFER_WA_DELAY ?
+			defer_delay : DPVGA_DONGLE_AUX_DEFER_WA_DELAY;
+
 	if (link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_0080E1 &&
 	    !memcmp(link->dpcd_caps.branch_dev_name,
 		    DP_DVI_CONVERTER_ID_4,
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
index 2fc12239b22c..c1391bfb7a9b 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
@@ -126,9 +126,7 @@ static void dpcd_set_training_pattern(
 static enum dc_dp_training_pattern decide_cr_training_pattern(
 		const struct dc_link_settings *link_settings)
 {
-	enum dc_dp_training_pattern pattern = DP_TRAINING_PATTERN_SEQUENCE_1;
-
-	return pattern;
+	return DP_TRAINING_PATTERN_SEQUENCE_1;
 }
 
 static enum dc_dp_training_pattern decide_eq_training_pattern(struct dc_link *link,
@@ -892,13 +890,13 @@ static uint32_t translate_training_aux_read_interval(uint32_t dpcd_aux_read_inte
 
 	switch (dpcd_aux_read_interval) {
 	case 0x01:
-		aux_rd_interval_us = 400;
+		aux_rd_interval_us = 4000;
 		break;
 	case 0x02:
-		aux_rd_interval_us = 4000;
+		aux_rd_interval_us = 8000;
 		break;
 	case 0x03:
-		aux_rd_interval_us = 8000;
+		aux_rd_interval_us = 12000;
 		break;
 	case 0x04:
 		aux_rd_interval_us = 16000;
@@ -2399,6 +2397,9 @@ static bool decide_dp_link_settings(struct dc_link *link, struct dc_link_setting
 			initial_link_setting;
 	uint32_t link_bw;
 
+	if (req_bw > dc_link_bandwidth_kbps(link, &link->verified_link_cap))
+		return false;
+
 	/* search for the minimum link setting that:
 	 * 1. is supported according to the link training result
 	 * 2. could support the b/w requested by the timing
@@ -3045,14 +3046,14 @@ bool dc_link_handle_hpd_rx_irq(struct dc_link *link, union hpd_irq_data *out_hpd
 		for (i = 0; i < MAX_PIPES; i++) {
 			pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
 			if (pipe_ctx && pipe_ctx->stream && !pipe_ctx->stream->dpms_off &&
-					pipe_ctx->stream->link == link)
+					pipe_ctx->stream->link == link && !pipe_ctx->prev_odm_pipe)
 				core_link_disable_stream(pipe_ctx);
 		}
 
 		for (i = 0; i < MAX_PIPES; i++) {
 			pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
 			if (pipe_ctx && pipe_ctx->stream && !pipe_ctx->stream->dpms_off &&
-					pipe_ctx->stream->link == link)
+					pipe_ctx->stream->link == link && !pipe_ctx->prev_odm_pipe)
 				core_link_enable_stream(link->dc->current_state, pipe_ctx);
 		}
 
@@ -3707,7 +3708,7 @@ bool detect_dp_sink_caps(struct dc_link *link)
 	/* TODO save sink caps in link->sink */
 }
 
-enum dc_link_rate linkRateInKHzToLinkRateMultiplier(uint32_t link_rate_in_khz)
+static enum dc_link_rate linkRateInKHzToLinkRateMultiplier(uint32_t link_rate_in_khz)
 {
 	enum dc_link_rate link_rate;
 	// LinkRate is normally stored as a multiplier of 0.27 Gbps per lane. Do the translation.
@@ -3992,7 +3993,7 @@ bool dc_link_dp_set_test_pattern(
 	unsigned int cust_pattern_size)
 {
 	struct pipe_ctx *pipes = link->dc->current_state->res_ctx.pipe_ctx;
-	struct pipe_ctx *pipe_ctx = &pipes[0];
+	struct pipe_ctx *pipe_ctx = NULL;
 	unsigned int lane;
 	unsigned int i;
 	unsigned char link_qual_pattern[LANE_COUNT_DP_MAX] = {0};
@@ -4002,12 +4003,18 @@ bool dc_link_dp_set_test_pattern(
 	memset(&training_pattern, 0, sizeof(training_pattern));
 
 	for (i = 0; i < MAX_PIPES; i++) {
+		if (pipes[i].stream == NULL)
+			continue;
+
 		if (pipes[i].stream->link == link && !pipes[i].top_pipe && !pipes[i].prev_odm_pipe) {
 			pipe_ctx = &pipes[i];
 			break;
 		}
 	}
 
+	if (pipe_ctx == NULL)
+		return false;
+
 	/* Reset CRTC Test Pattern if it is currently running and request is VideoMode */
 	if (link->test_pattern_enabled && test_pattern ==
 			DP_TEST_PATTERN_VIDEO_MODE) {
@@ -4339,7 +4346,7 @@ void dp_set_fec_ready(struct dc_link *link, bool ready)
 	struct link_encoder *link_enc = link->link_enc;
 	uint8_t fec_config = 0;
 
-	if (!dc_link_is_fec_supported(link) || link->dc->debug.disable_fec)
+	if (!dc_link_should_enable_fec(link))
 		return;
 
 	if (link_enc->funcs->fec_set_ready &&
@@ -4374,7 +4381,7 @@ void dp_set_fec_enable(struct dc_link *link, bool enable)
 {
 	struct link_encoder *link_enc = link->link_enc;
 
-	if (!dc_link_is_fec_supported(link) || link->dc->debug.disable_fec)
+	if (!dc_link_should_enable_fec(link))
 		return;
 
 	if (link_enc->funcs->fec_set_enable &&
@@ -4400,24 +4407,39 @@ void dp_set_fec_enable(struct dc_link *link, bool enable)
 void dpcd_set_source_specific_data(struct dc_link *link)
 {
 	if (!link->dc->vendor_signature.is_valid) {
-		enum dc_status result_write_min_hblank = DC_NOT_SUPPORTED;
-		struct dpcd_amd_signature amd_signature;
-		amd_signature.AMD_IEEE_TxSignature_byte1 = 0x0;
-		amd_signature.AMD_IEEE_TxSignature_byte2 = 0x0;
-		amd_signature.AMD_IEEE_TxSignature_byte3 = 0x1A;
-		amd_signature.device_id_byte1 =
+		enum dc_status __maybe_unused result_write_min_hblank = DC_NOT_SUPPORTED;
+		struct dpcd_amd_signature amd_signature = {0};
+		struct dpcd_amd_device_id amd_device_id = {0};
+
+		amd_device_id.device_id_byte1 =
 				(uint8_t)(link->ctx->asic_id.chip_id);
-		amd_signature.device_id_byte2 =
+		amd_device_id.device_id_byte2 =
 				(uint8_t)(link->ctx->asic_id.chip_id >> 8);
-		memset(&amd_signature.zero, 0, 4);
-		amd_signature.dce_version =
+		amd_device_id.dce_version =
 				(uint8_t)(link->ctx->dce_version);
-		amd_signature.dal_version_byte1 = 0x0; // needed? where to get?
-		amd_signature.dal_version_byte2 = 0x0; // needed? where to get?
+		amd_device_id.dal_version_byte1 = 0x0; // needed? where to get?
+		amd_device_id.dal_version_byte2 = 0x0; // needed? where to get?
 
-		core_link_write_dpcd(link, DP_SOURCE_OUI,
+		core_link_read_dpcd(link, DP_SOURCE_OUI,
+				(uint8_t *)(&amd_signature),
+				sizeof(amd_signature));
+
+		if (!((amd_signature.AMD_IEEE_TxSignature_byte1 == 0x0) &&
+			(amd_signature.AMD_IEEE_TxSignature_byte2 == 0x0) &&
+			(amd_signature.AMD_IEEE_TxSignature_byte3 == 0x1A))) {
+
+			amd_signature.AMD_IEEE_TxSignature_byte1 = 0x0;
+			amd_signature.AMD_IEEE_TxSignature_byte2 = 0x0;
+			amd_signature.AMD_IEEE_TxSignature_byte3 = 0x1A;
+
+			core_link_write_dpcd(link, DP_SOURCE_OUI,
 				(uint8_t *)(&amd_signature),
 				sizeof(amd_signature));
+		}
+
+		core_link_write_dpcd(link, DP_SOURCE_OUI+0x03,
+				(uint8_t *)(&amd_device_id),
+				sizeof(amd_device_id));
 
 		if (link->ctx->dce_version >= DCN_VERSION_2_0 &&
 			link->dc->caps.min_horizontal_blanking_period != 0) {
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index 07c22556480b..0c26c2ade782 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -1117,7 +1117,7 @@ static void calculate_inits_and_adj_vp(struct pipe_ctx *pipe_ctx)
  * We also need to make sure pipe_ctx->plane_res.scl_data.h_active uses the
  * original h_border_left value in its calculation.
  */
-int shift_border_left_to_dst(struct pipe_ctx *pipe_ctx)
+static int shift_border_left_to_dst(struct pipe_ctx *pipe_ctx)
 {
 	int store_h_border_left = pipe_ctx->stream->timing.h_border_left;
 
@@ -1128,8 +1128,8 @@ int shift_border_left_to_dst(struct pipe_ctx *pipe_ctx)
 	return store_h_border_left;
 }
 
-void restore_border_left_from_dst(struct pipe_ctx *pipe_ctx,
-                                  int store_h_border_left)
+static void restore_border_left_from_dst(struct pipe_ctx *pipe_ctx,
+					 int store_h_border_left)
 {
 	pipe_ctx->stream->dst.x -= store_h_border_left;
 	pipe_ctx->stream->timing.h_border_left = store_h_border_left;
@@ -1153,8 +1153,8 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 
 	calculate_viewport(pipe_ctx);
 
-	if (pipe_ctx->plane_res.scl_data.viewport.height < 12 ||
-		pipe_ctx->plane_res.scl_data.viewport.width < 12) {
+	if (pipe_ctx->plane_res.scl_data.viewport.height < MIN_VIEWPORT_SIZE ||
+		pipe_ctx->plane_res.scl_data.viewport.width < MIN_VIEWPORT_SIZE) {
 		if (store_h_border_left) {
 			restore_border_left_from_dst(pipe_ctx,
 				store_h_border_left);
@@ -1697,7 +1697,7 @@ static bool are_stream_backends_same(
 	return true;
 }
 
-/**
+/*
  * dc_is_stream_unchanged() - Compare two stream states for equivalence.
  *
  * Checks if there a difference between the two states
@@ -1718,7 +1718,7 @@ bool dc_is_stream_unchanged(
 	return true;
 }
 
-/**
+/*
  * dc_is_stream_scaling_unchanged() - Compare scaling rectangles of two streams.
  */
 bool dc_is_stream_scaling_unchanged(
@@ -1833,7 +1833,7 @@ static struct audio *find_first_free_audio(
 	return 0;
 }
 
-/**
+/*
  * dc_add_stream_to_ctx() - Add a new dc_stream_state to a dc_state.
  */
 enum dc_status dc_add_stream_to_ctx(
@@ -1860,7 +1860,7 @@ enum dc_status dc_add_stream_to_ctx(
 	return res;
 }
 
-/**
+/*
  * dc_remove_stream_from_ctx() - Remove a stream from a dc_state.
  */
 enum dc_status dc_remove_stream_from_ctx(
@@ -2075,6 +2075,20 @@ static int acquire_resource_from_hw_enabled_state(
 	return -1;
 }
 
+static void mark_seamless_boot_stream(
+		const struct dc  *dc,
+		struct dc_stream_state *stream)
+{
+	struct dc_bios *dcb = dc->ctx->dc_bios;
+
+	/* TODO: Check Linux */
+	if (dc->config.allow_seamless_boot_optimization &&
+			!dcb->funcs->is_accelerated_mode(dcb)) {
+		if (dc_validate_seamless_boot_timing(dc, stream->sink, &stream->timing))
+			stream->apply_seamless_boot_optimization = true;
+	}
+}
+
 enum dc_status resource_map_pool_resources(
 		const struct dc  *dc,
 		struct dc_state *context,
@@ -2085,22 +2099,20 @@ enum dc_status resource_map_pool_resources(
 	struct dc_context *dc_ctx = dc->ctx;
 	struct pipe_ctx *pipe_ctx = NULL;
 	int pipe_idx = -1;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
 
 	calculate_phy_pix_clks(stream);
 
-	/* TODO: Check Linux */
-	if (dc->config.allow_seamless_boot_optimization &&
-			!dcb->funcs->is_accelerated_mode(dcb)) {
-		if (dc_validate_seamless_boot_timing(dc, stream->sink, &stream->timing))
-			stream->apply_seamless_boot_optimization = true;
-	}
+	mark_seamless_boot_stream(dc, stream);
 
-	if (stream->apply_seamless_boot_optimization)
+	if (stream->apply_seamless_boot_optimization) {
 		pipe_idx = acquire_resource_from_hw_enabled_state(
 				&context->res_ctx,
 				pool,
 				stream);
+		if (pipe_idx < 0)
+			/* hw resource was assigned to other stream */
+			stream->apply_seamless_boot_optimization = false;
+	}
 
 	if (pipe_idx < 0)
 		/* acquire new resources */
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
index c103f858375d..25fa712a7847 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
@@ -244,7 +244,7 @@ struct dc_stream_status *dc_stream_get_status(
 }
 
 #ifndef TRIM_FSFT
-/**
+/*
  * dc_optimize_timing_for_fsft() - dc to optimize timing
  */
 bool dc_optimize_timing_for_fsft(
@@ -260,8 +260,7 @@ bool dc_optimize_timing_for_fsft(
 }
 #endif
 
-
-/**
+/*
  * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address
  */
 bool dc_stream_set_cursor_attributes(
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
index 3d7d27435f15..e6b9c6a71841 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
@@ -115,7 +115,7 @@ struct dc_plane_state *dc_create_plane_state(struct dc *dc)
 	return plane_state;
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: dc_plane_get_status
  *
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 3aedadb34548..18ed0d3f247e 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -42,12 +42,13 @@
 #include "inc/hw/dmcu.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.2.116"
+#define DC_VER "3.2.122"
 
 #define MAX_SURFACES 3
 #define MAX_PLANES 6
 #define MAX_STREAMS 6
 #define MAX_SINKS_PER_LINK 4
+#define MIN_VIEWPORT_SIZE 12
 
 /*******************************************************************************
  * Display Core Interfaces
@@ -171,6 +172,9 @@ struct dc_caps {
 	bool dmcub_support;
 	uint32_t num_of_internal_disp;
 	enum dp_protocol_version max_dp_protocol_version;
+	unsigned int mall_size_per_mem_channel;
+	unsigned int mall_size_total;
+	unsigned int cursor_cache_size;
 	struct dc_plane_cap planes[MAX_PLANES];
 	struct dc_color_caps color;
 };
@@ -481,7 +485,6 @@ struct dc_debug_options {
 	bool performance_trace;
 	bool az_endpoint_mute_only;
 	bool always_use_regamma;
-	bool p010_mpo_support;
 	bool recovery_enabled;
 	bool avoid_vbios_exec_table;
 	bool scl_reset_length10;
@@ -499,6 +502,9 @@ struct dc_debug_options {
 	bool dmcub_emulation;
 #if defined(CONFIG_DRM_AMD_DC_DCN)
 	bool disable_idle_power_optimizations;
+	unsigned int mall_size_override;
+	unsigned int mall_additional_timer_percent;
+	bool mall_error_as_fatal;
 #endif
 	bool dmub_command_table; /* for testing only */
 	struct dc_bw_validation_profile bw_val_profile;
@@ -521,7 +527,6 @@ struct dc_debug_options {
 	bool usbc_combo_phy_reset_wa;
 	bool disable_dsc;
 	bool enable_dram_clock_change_one_display_vactive;
-	bool force_ignore_link_settings;
 	union mem_low_power_enable_options enable_mem_low_power;
 };
 
@@ -633,7 +638,6 @@ struct dc {
 
 	const char *build_id;
 	struct vm_helper *vm_helper;
-	const struct gpu_info_soc_bounding_box_v1_0 *soc_bounding_box;
 };
 
 enum frame_buffer_mode {
@@ -671,16 +675,10 @@ struct dc_init_data {
 	struct dc_config flags;
 	uint64_t log_mask;
 
-	/**
-	 * gpu_info FW provided soc bounding box struct or 0 if not
-	 * available in FW
-	 */
-	const struct gpu_info_soc_bounding_box_v1_0 *soc_bounding_box;
 	struct dpcd_vendor_signature vendor_signature;
 #if defined(CONFIG_DRM_AMD_DC_DCN)
 	bool force_smu_not_present;
 #endif
-	bool force_ignore_link_settings;
 };
 
 struct dc_callback_init {
@@ -889,6 +887,7 @@ struct dc_plane_state {
 	int layer_index;
 
 	union surface_update_flags update_flags;
+	bool flip_int_enabled;
 	/* private to DC core */
 	struct dc_plane_status status;
 	struct dc_context *ctx;
@@ -1269,8 +1268,8 @@ enum dc_status dc_set_clock(struct dc *dc, enum dc_clock_type clock_type, uint32
 void dc_get_clock(struct dc *dc, enum dc_clock_type clock_type, struct dc_clock_config *clock_cfg);
 #if defined(CONFIG_DRM_AMD_DC_DCN)
 
-bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc,
-						 struct dc_plane_state *plane);
+bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc, struct dc_plane_state *plane,
+				struct dc_cursor_attributes *cursor_attr);
 
 void dc_allow_idle_optimizations(struct dc *dc, bool allow);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
index 80a2191a3115..cc6fb838420e 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
@@ -451,6 +451,9 @@ struct dpcd_amd_signature {
 	uint8_t AMD_IEEE_TxSignature_byte1;
 	uint8_t AMD_IEEE_TxSignature_byte2;
 	uint8_t AMD_IEEE_TxSignature_byte3;
+};
+
+struct dpcd_amd_device_id {
 	uint8_t device_id_byte1;
 	uint8_t device_id_byte2;
 	uint8_t zero[4];
diff --git a/drivers/gpu/drm/amd/display/dc/dc_edid_parser.c b/drivers/gpu/drm/amd/display/dc/dc_edid_parser.c
new file mode 100644
index 000000000000..0db5b49e9d5e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_edid_parser.c
@@ -0,0 +1,80 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dce/dce_dmcu.h"
+#include "dc_edid_parser.h"
+
+bool dc_edid_parser_send_cea(struct dc *dc,
+		int offset,
+		int total_length,
+		uint8_t *data,
+		int length)
+{
+	struct dmcu *dmcu = dc->res_pool->dmcu;
+
+	if (dmcu &&
+	    dmcu->funcs->is_dmcu_initialized(dmcu) &&
+	    dmcu->funcs->send_edid_cea) {
+		return dmcu->funcs->send_edid_cea(dmcu,
+				offset,
+				total_length,
+				data,
+				length);
+	}
+
+	return false;
+}
+
+bool dc_edid_parser_recv_cea_ack(struct dc *dc, int *offset)
+{
+	struct dmcu *dmcu = dc->res_pool->dmcu;
+
+	if (dmcu &&
+	    dmcu->funcs->is_dmcu_initialized(dmcu) &&
+	    dmcu->funcs->recv_edid_cea_ack) {
+		return dmcu->funcs->recv_edid_cea_ack(dmcu, offset);
+	}
+
+	return false;
+}
+
+bool dc_edid_parser_recv_amd_vsdb(struct dc *dc,
+		int *version,
+		int *min_frame_rate,
+		int *max_frame_rate)
+{
+	struct dmcu *dmcu = dc->res_pool->dmcu;
+
+	if (dmcu &&
+	    dmcu->funcs->is_dmcu_initialized(dmcu) &&
+	    dmcu->funcs->recv_amd_vsdb) {
+		return dmcu->funcs->recv_amd_vsdb(dmcu,
+				version,
+				min_frame_rate,
+				max_frame_rate);
+	}
+
+	return false;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dc_edid_parser.h b/drivers/gpu/drm/amd/display/dc/dc_edid_parser.h
new file mode 100644
index 000000000000..da67ec06f0a2
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_edid_parser.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef _DC_EDID_PARSER_H_
+#define _DC_EDID_PARSER_H_
+
+#include "core_types.h"
+
+bool dc_edid_parser_send_cea(struct dc *dc,
+		int offset,
+		int total_length,
+		uint8_t *data,
+		int length);
+
+bool dc_edid_parser_recv_cea_ack(struct dc *dc, int *offset);
+
+bool dc_edid_parser_recv_amd_vsdb(struct dc *dc,
+		int *version,
+		int *min_frame_rate,
+		int *max_frame_rate);
+
+#endif /* _DC_EDID_PARSER_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c
index 57edb25fc381..a612ba6dc389 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c
@@ -34,6 +34,7 @@
 
 #include "dc.h"
 #include "dc_dmub_srv.h"
+#include "reg_helper.h"
 
 static inline void submit_dmub_read_modify_write(
 	struct dc_reg_helper_state *offload,
diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
index 701aa7178a89..b41e6367b15e 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
@@ -71,6 +71,7 @@ struct dc_plane_address {
 	union {
 		struct{
 			PHYSICAL_ADDRESS_LOC addr;
+			PHYSICAL_ADDRESS_LOC cursor_cache_addr;
 			PHYSICAL_ADDRESS_LOC meta_addr;
 			union large_integer dcc_const_color;
 		} grph;
diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h
index 6d9a60c9dcc0..e189f16bc026 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_link.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_link.h
@@ -103,6 +103,8 @@ struct dc_link {
 	bool lttpr_non_transparent_mode;
 	bool is_internal_display;
 
+	bool edp_sink_present;
+
 	/* caps is the same as reported_link_cap. link_traing use
 	 * reported_link_cap. Will clean up.  TODO
 	 */
@@ -259,6 +261,13 @@ enum dc_status dc_link_reallocate_mst_payload(struct dc_link *link);
 bool dc_link_handle_hpd_rx_irq(struct dc_link *dc_link,
 		union hpd_irq_data *hpd_irq_dpcd_data, bool *out_link_loss);
 
+/*
+ * On eDP links this function call will stall until T12 has elapsed.
+ * If the panel is not in power off state, this function will return
+ * immediately.
+ */
+bool dc_link_wait_for_t12(struct dc_link *link);
+
 enum dc_status read_hpd_rx_irq_data(
 	struct dc_link *link,
 	union hpd_irq_data *irq_data);
@@ -369,5 +378,6 @@ uint32_t dc_bandwidth_in_kbps_from_timing(
 	const struct dc_crtc_timing *timing);
 
 bool dc_link_is_fec_supported(const struct dc_link *link);
+bool dc_link_should_enable_fec(const struct dc_link *link);
 
 #endif /* DC_LINK_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h
index b7910976b81a..80b67b860091 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_stream.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h
@@ -130,6 +130,14 @@ union stream_update_flags {
 	uint32_t raw;
 };
 
+struct test_pattern {
+	enum dp_test_pattern type;
+	enum dp_test_pattern_color_space color_space;
+	struct link_training_settings const *p_link_settings;
+	unsigned char const *p_custom_pattern;
+	unsigned int cust_pattern_size;
+};
+
 struct dc_stream_state {
 	// sink is deprecated, new code should not reference
 	// this pointer
@@ -227,6 +235,8 @@ struct dc_stream_state {
 
 	uint32_t stream_id;
 	bool is_dsc_enabled;
+
+	struct test_pattern test_pattern;
 	union stream_update_flags update_flags;
 };
 
@@ -261,6 +271,7 @@ struct dc_stream_update {
 	struct dc_dsc_config *dsc_config;
 	struct dc_transfer_func *func_shaper;
 	struct dc_3dlut *lut3d_func;
+	struct test_pattern *pending_test_pattern;
 };
 
 bool dc_is_stream_unchanged(
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c b/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c
index 2a2a0fdb9253..7866cf2a668f 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c
@@ -868,7 +868,7 @@ void dce_aud_wall_dto_setup(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-void dce60_aud_wall_dto_setup(
+static void dce60_aud_wall_dto_setup(
 	struct audio *audio,
 	enum signal_type signal,
 	const struct audio_crtc_info *crtc_info,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
index cda5fd0464bc..d51b5fe91287 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
@@ -388,12 +388,6 @@ static enum aux_channel_operation_result get_channel_status(
 	}
 }
 
-enum i2caux_engine_type get_engine_type(
-		const struct dce_aux *engine)
-{
-	return I2CAUX_ENGINE_TYPE_AUX;
-}
-
 static bool acquire(
 	struct dce_aux *engine,
 	struct ddc *ddc)
@@ -582,7 +576,7 @@ int dce_aux_transfer_raw(struct ddc_service *ddc,
 	*operation_result = get_channel_status(aux_engine, &returned_bytes);
 
 	if (*operation_result == AUX_CHANNEL_OPERATION_SUCCEEDED) {
-		int bytes_replied = 0;
+		int __maybe_unused bytes_replied = 0;
 		bytes_replied = read_channel_reply(aux_engine, payload->length,
 					 payload->data, payload->reply,
 					 &status);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h
index 382465862f29..277484cf853e 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h
@@ -124,7 +124,6 @@ struct dce110_aux_registers {
 	AUX_SF(AUX_SW_CONTROL, AUX_SW_GO, mask_sh),\
 	AUX_SF(AUX_SW_DATA, AUX_SW_AUTOINCREMENT_DISABLE, mask_sh),\
 	AUX_SF(AUX_SW_DATA, AUX_SW_DATA_RW, mask_sh),\
-	AUX_SF(AUX_SW_DATA, AUX_SW_AUTOINCREMENT_DISABLE, mask_sh),\
 	AUX_SF(AUX_SW_DATA, AUX_SW_INDEX, mask_sh),\
 	AUX_SF(AUX_SW_DATA, AUX_SW_DATA, mask_sh),\
 	AUX_SF(AUX_SW_STATUS, AUX_SW_REPLY_BYTE_COUNT, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
index fb733f573715..dec58b3c42e4 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
@@ -113,20 +113,19 @@ static const struct spread_spectrum_data *get_ss_data_entry(
 }
 
 /**
- * Function: calculate_fb_and_fractional_fb_divider
+ * calculate_fb_and_fractional_fb_divider - Calculates feedback and fractional
+ *                                          feedback dividers values
  *
- * * DESCRIPTION: Calculates feedback and fractional feedback dividers values
+ * @calc_pll_cs:	    Pointer to clock source information
+ * @target_pix_clk_100hz:   Desired frequency in 100 Hz
+ * @ref_divider:            Reference divider (already known)
+ * @post_divider:           Post Divider (already known)
+ * @feedback_divider_param: Pointer where to store
+ *			    calculated feedback divider value
+ * @fract_feedback_divider_param: Pointer where to store
+ *			    calculated fract feedback divider value
  *
- *PARAMETERS:
- * targetPixelClock             Desired frequency in 100 Hz
- * ref_divider                  Reference divider (already known)
- * postDivider                  Post Divider (already known)
- * feedback_divider_param       Pointer where to store
- *					calculated feedback divider value
- * fract_feedback_divider_param Pointer where to store
- *					calculated fract feedback divider value
- *
- *RETURNS:
+ * return:
  * It fills the locations pointed by feedback_divider_param
  *					and fract_feedback_divider_param
  * It returns	- true if feedback divider not 0
@@ -175,22 +174,22 @@ static bool calculate_fb_and_fractional_fb_divider(
 }
 
 /**
-*calc_fb_divider_checking_tolerance
-*
-*DESCRIPTION: Calculates Feedback and Fractional Feedback divider values
-*		for passed Reference and Post divider, checking for tolerance.
-*PARAMETERS:
-* pll_settings		Pointer to structure
-* ref_divider		Reference divider (already known)
-* postDivider		Post Divider (already known)
-* tolerance		Tolerance for Calculated Pixel Clock to be within
-*
-*RETURNS:
-* It fills the PLLSettings structure with PLL Dividers values
-* if calculated values are within required tolerance
-* It returns	- true if error is within tolerance
-*		- false if error is not within tolerance
-*/
+ * calc_fb_divider_checking_tolerance - Calculates Feedback and
+ *                                      Fractional Feedback divider values
+ *		                        for passed Reference and Post divider,
+ *                                      checking for tolerance.
+ * @calc_pll_cs:	Pointer to clock source information
+ * @pll_settings:	Pointer to PLL settings
+ * @ref_divider:	Reference divider (already known)
+ * @post_divider:	Post Divider (already known)
+ * @tolerance:		Tolerance for Calculated Pixel Clock to be within
+ *
+ * return:
+ *  It fills the PLLSettings structure with PLL Dividers values
+ *  if calculated values are within required tolerance
+ *  It returns	- true if error is within tolerance
+ *		- false if error is not within tolerance
+ */
 static bool calc_fb_divider_checking_tolerance(
 		struct calc_pll_clock_source *calc_pll_cs,
 		struct pll_settings *pll_settings,
@@ -241,7 +240,7 @@ static bool calc_fb_divider_checking_tolerance(
 		pll_settings->calculated_pix_clk_100hz =
 			actual_calculated_clock_100hz;
 		pll_settings->vco_freq =
-			actual_calculated_clock_100hz * post_divider / 10;
+			div_u64((u64)actual_calculated_clock_100hz * post_divider, 10);
 		return true;
 	}
 	return false;
@@ -460,7 +459,7 @@ static bool pll_adjust_pix_clk(
 	return false;
 }
 
-/**
+/*
  * Calculate PLL Dividers for given Clock Value.
  * First will call VBIOS Adjust Exec table to check if requested Pixel clock
  * will be Adjusted based on usage.
@@ -871,6 +870,20 @@ static bool dce110_program_pix_clk(
 	bp_pc_params.flags.SET_EXTERNAL_REF_DIV_SRC =
 					pll_settings->use_external_clk;
 
+	switch (pix_clk_params->color_depth) {
+	case COLOR_DEPTH_101010:
+		bp_pc_params.color_depth = TRANSMITTER_COLOR_DEPTH_30;
+		break;
+	case COLOR_DEPTH_121212:
+		bp_pc_params.color_depth = TRANSMITTER_COLOR_DEPTH_36;
+		break;
+	case COLOR_DEPTH_161616:
+		bp_pc_params.color_depth = TRANSMITTER_COLOR_DEPTH_48;
+		break;
+	default:
+		break;
+	}
+
 	if (clk_src->bios->funcs->set_pixel_clock(
 			clk_src->bios, &bp_pc_params) != BP_RESULT_OK)
 		return false;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c
index f3ed8b619caf..ddc789daf3b1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c
@@ -57,6 +57,9 @@
 #define MCP_SYNC_PHY_LOCK 0x90
 #define MCP_SYNC_PHY_UNLOCK 0x91
 #define MCP_BL_SET_PWM_FRAC 0x6A  /* Enable or disable Fractional PWM */
+#define MCP_SEND_EDID_CEA 0xA0
+#define EDID_CEA_CMD_ACK 1
+#define EDID_CEA_CMD_NACK 2
 #define MASTER_COMM_CNTL_REG__MASTER_COMM_INTERRUPT_MASK   0x00000001L
 
 // PSP FW version
@@ -65,13 +68,17 @@
 //Register access policy version
 #define mmMP0_SMN_C2PMSG_91				0x1609B
 
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+static const uint32_t abm_gain_stepsize = 0x0060;
+#endif
+
 static bool dce_dmcu_init(struct dmcu *dmcu)
 {
 	// Do nothing
 	return true;
 }
 
-bool dce_dmcu_load_iram(struct dmcu *dmcu,
+static bool dce_dmcu_load_iram(struct dmcu *dmcu,
 		unsigned int start_offset,
 		const char *src,
 		unsigned int bytes)
@@ -807,6 +814,120 @@ static bool dcn20_unlock_phy(struct dmcu *dmcu)
 	return true;
 }
 
+static bool dcn10_send_edid_cea(struct dmcu *dmcu,
+		int offset,
+		int total_length,
+		uint8_t *data,
+		int length)
+{
+	struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu);
+	uint32_t header, data1, data2;
+
+	/* If microcontroller is not running, do nothing */
+	if (dmcu->dmcu_state != DMCU_RUNNING)
+		return false;
+
+	if (length > 8 || length <= 0)
+		return false;
+
+	header = ((uint32_t)offset & 0xFFFF) << 16 | (total_length & 0xFFFF);
+	data1 = (((uint32_t)data[0]) << 24) | (((uint32_t)data[1]) << 16) |
+		(((uint32_t)data[2]) << 8) | ((uint32_t)data[3]);
+	data2 = (((uint32_t)data[4]) << 24) | (((uint32_t)data[5]) << 16) |
+		(((uint32_t)data[6]) << 8) | ((uint32_t)data[7]);
+
+	/* waitDMCUReadyForCmd */
+	REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 1, 10000);
+
+	/* setDMCUParam_Cmd */
+	REG_UPDATE(MASTER_COMM_CMD_REG, MASTER_COMM_CMD_REG_BYTE0, MCP_SEND_EDID_CEA);
+
+	REG_WRITE(MASTER_COMM_DATA_REG1, header);
+	REG_WRITE(MASTER_COMM_DATA_REG2, data1);
+	REG_WRITE(MASTER_COMM_DATA_REG3, data2);
+
+	/* notifyDMCUMsg */
+	REG_UPDATE(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 1);
+
+	/* waitDMCUReadyForCmd */
+	REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 1, 10000);
+
+	return true;
+}
+
+static bool dcn10_get_scp_results(struct dmcu *dmcu,
+		uint32_t *cmd,
+		uint32_t *data1,
+		uint32_t *data2,
+		uint32_t *data3)
+{
+	struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu);
+
+	/* If microcontroller is not running, do nothing */
+	if (dmcu->dmcu_state != DMCU_RUNNING)
+		return false;
+
+	*cmd = REG_READ(SLAVE_COMM_CMD_REG);
+	*data1 =  REG_READ(SLAVE_COMM_DATA_REG1);
+	*data2 =  REG_READ(SLAVE_COMM_DATA_REG2);
+	*data3 =  REG_READ(SLAVE_COMM_DATA_REG3);
+
+	/* clear SCP interrupt */
+	REG_UPDATE(SLAVE_COMM_CNTL_REG, SLAVE_COMM_INTERRUPT, 0);
+
+	return true;
+}
+
+static bool dcn10_recv_amd_vsdb(struct dmcu *dmcu,
+		int *version,
+		int *min_frame_rate,
+		int *max_frame_rate)
+{
+	uint32_t data[4];
+	int cmd, ack, len;
+
+	if (!dcn10_get_scp_results(dmcu, &data[0], &data[1], &data[2], &data[3]))
+		return false;
+
+	cmd = data[0] & 0x3FF;
+	len = (data[0] >> 10) & 0x3F;
+	ack = data[1];
+
+	if (cmd != MCP_SEND_EDID_CEA || ack != EDID_CEA_CMD_ACK || len != 12)
+		return false;
+
+	if ((data[2] & 0xFF)) {
+		*version = (data[2] >> 8) & 0xFF;
+		*min_frame_rate = (data[3] >> 16) & 0xFFFF;
+		*max_frame_rate = data[3] & 0xFFFF;
+		return true;
+	}
+
+	return false;
+}
+
+static bool dcn10_recv_edid_cea_ack(struct dmcu *dmcu, int *offset)
+{
+	uint32_t data[4];
+	int cmd, ack;
+
+	if (!dcn10_get_scp_results(dmcu,
+				&data[0], &data[1], &data[2], &data[3]))
+		return false;
+
+	cmd = data[0] & 0x3FF;
+	ack = data[1];
+
+	if (cmd != MCP_SEND_EDID_CEA)
+		return false;
+
+	if (ack == EDID_CEA_CMD_ACK)
+		return true;
+
+	*offset = data[2]; /* nack */
+	return false;
+}
+
 #endif //(CONFIG_DRM_AMD_DC_DCN)
 
 static const struct dmcu_funcs dce_funcs = {
@@ -829,6 +950,9 @@ static const struct dmcu_funcs dcn10_funcs = {
 	.get_psr_state = dcn10_get_dmcu_psr_state,
 	.set_psr_wait_loop = dcn10_psr_wait_loop,
 	.get_psr_wait_loop = dcn10_get_psr_wait_loop,
+	.send_edid_cea = dcn10_send_edid_cea,
+	.recv_amd_vsdb = dcn10_recv_amd_vsdb,
+	.recv_edid_cea_ack = dcn10_recv_edid_cea_ack,
 	.is_dmcu_initialized = dcn10_is_dmcu_initialized
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.h b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.h
index 93e7f34d4775..ff726b35ef6a 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.h
@@ -40,6 +40,10 @@
 	SR(MASTER_COMM_DATA_REG3), \
 	SR(MASTER_COMM_CMD_REG), \
 	SR(MASTER_COMM_CNTL_REG), \
+	SR(SLAVE_COMM_DATA_REG1), \
+	SR(SLAVE_COMM_DATA_REG2), \
+	SR(SLAVE_COMM_DATA_REG3), \
+	SR(SLAVE_COMM_CMD_REG), \
 	SR(DMCU_IRAM_RD_CTRL), \
 	SR(DMCU_IRAM_RD_DATA), \
 	SR(DMCU_INTERRUPT_TO_UC_EN_MASK), \
@@ -112,6 +116,7 @@
 	DMCU_SF(MASTER_COMM_CMD_REG, \
 			MASTER_COMM_CMD_REG_BYTE0, mask_sh), \
 	DMCU_SF(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, mask_sh), \
+	DMCU_SF(SLAVE_COMM_CNTL_REG, SLAVE_COMM_INTERRUPT, mask_sh), \
 	DMCU_SF(DMCU_INTERRUPT_TO_UC_EN_MASK, \
 			STATIC_SCREEN1_INT_TO_UC_EN, mask_sh), \
 	DMCU_SF(DMCU_INTERRUPT_TO_UC_EN_MASK, \
@@ -179,6 +184,7 @@
 	type UC_IN_RESET; \
 	type MASTER_COMM_CMD_REG_BYTE0; \
 	type MASTER_COMM_INTERRUPT; \
+	type SLAVE_COMM_INTERRUPT; \
 	type DPHY_RX_FAST_TRAINING_CAPABLE; \
 	type DPHY_LOAD_BS_COUNT; \
 	type STATIC_SCREEN1_INT_TO_UC_EN; \
@@ -211,6 +217,11 @@ struct dce_dmcu_registers {
 	uint32_t MASTER_COMM_DATA_REG3;
 	uint32_t MASTER_COMM_CMD_REG;
 	uint32_t MASTER_COMM_CNTL_REG;
+	uint32_t SLAVE_COMM_DATA_REG1;
+	uint32_t SLAVE_COMM_DATA_REG2;
+	uint32_t SLAVE_COMM_DATA_REG3;
+	uint32_t SLAVE_COMM_CMD_REG;
+	uint32_t SLAVE_COMM_CNTL_REG;
 	uint32_t DMCU_IRAM_RD_CTRL;
 	uint32_t DMCU_IRAM_RD_DATA;
 	uint32_t DMCU_INTERRUPT_TO_UC_EN_MASK;
@@ -317,6 +328,4 @@ struct dmcu *dcn21_dmcu_create(
 
 void dce_dmcu_destroy(struct dmcu **dmcu);
 
-static const uint32_t abm_gain_stepsize = 0x0060;
-
 #endif /* _DCE_ABM_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c
index 7fbd92fbc63a..a524f471e0d7 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c
@@ -435,7 +435,7 @@ struct dce_i2c_hw *acquire_i2c_hw_engine(
 	return dce_i2c_hw;
 }
 
-enum i2c_channel_operation_result dce_i2c_hw_engine_wait_on_operation_result(
+static enum i2c_channel_operation_result dce_i2c_hw_engine_wait_on_operation_result(
 	struct dce_i2c_hw *dce_i2c_hw,
 	uint32_t timeout,
 	enum i2c_channel_operation_result expected_result)
@@ -502,7 +502,7 @@ static uint32_t get_transaction_timeout_hw(
 	return period_timeout * num_of_clock_stretches;
 }
 
-bool dce_i2c_hw_engine_submit_payload(
+static bool dce_i2c_hw_engine_submit_payload(
 	struct dce_i2c_hw *dce_i2c_hw,
 	struct i2c_payload *payload,
 	bool middle_of_transaction,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c
index 87d8428df6c4..6846afd83701 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c
@@ -339,7 +339,7 @@ static bool start_sync_sw(
 	return false;
 }
 
-void dce_i2c_sw_engine_set_speed(
+static void dce_i2c_sw_engine_set_speed(
 	struct dce_i2c_sw *engine,
 	uint32_t speed)
 {
@@ -353,7 +353,7 @@ void dce_i2c_sw_engine_set_speed(
 		engine->clock_delay = 12;
 }
 
-bool dce_i2c_sw_engine_acquire_engine(
+static bool dce_i2c_sw_engine_acquire_engine(
 	struct dce_i2c_sw *engine,
 	struct ddc *ddc)
 {
@@ -397,7 +397,7 @@ bool dce_i2c_engine_acquire_sw(
 
 
 
-void dce_i2c_sw_engine_submit_channel_request(
+static void dce_i2c_sw_engine_submit_channel_request(
 	struct dce_i2c_sw *engine,
 	struct i2c_request_transaction_data *req)
 {
@@ -440,7 +440,8 @@ void dce_i2c_sw_engine_submit_channel_request(
 		I2C_CHANNEL_OPERATION_SUCCEEDED :
 		I2C_CHANNEL_OPERATION_FAILED;
 }
-bool dce_i2c_sw_engine_submit_payload(
+
+static bool dce_i2c_sw_engine_submit_payload(
 	struct dce_i2c_sw *engine,
 	struct i2c_payload *payload,
 	bool middle_of_transaction)
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c
index 210466b2d863..1e77ffee71b3 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c
@@ -1197,7 +1197,7 @@ void dce110_link_encoder_enable_dp_mst_output(
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
 /* enables DP PHY output */
-void dce60_link_encoder_enable_dp_output(
+static void dce60_link_encoder_enable_dp_output(
 	struct link_encoder *enc,
 	const struct dc_link_settings *link_settings,
 	enum clock_source_id clock_source)
@@ -1236,7 +1236,7 @@ void dce60_link_encoder_enable_dp_output(
 }
 
 /* enables DP PHY output in MST mode */
-void dce60_link_encoder_enable_dp_mst_output(
+static void dce60_link_encoder_enable_dp_mst_output(
 	struct link_encoder *enc,
 	const struct dc_link_settings *link_settings,
 	enum clock_source_id clock_source)
@@ -1426,7 +1426,7 @@ void dce110_link_encoder_dp_set_phy_pattern(
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
 /* set DP PHY test and training patterns */
-void dce60_link_encoder_dp_set_phy_pattern(
+static void dce60_link_encoder_dp_set_phy_pattern(
 	struct link_encoder *enc,
 	const struct encoder_set_dp_phy_pattern_param *param)
 {
@@ -1503,7 +1503,6 @@ void dce110_link_encoder_update_mst_stream_allocation_table(
 	const struct link_mst_stream_allocation_table *table)
 {
 	struct dce110_link_encoder *enc110 = TO_DCE110_LINK_ENC(enc);
-	uint32_t value0 = 0;
 	uint32_t value1 = 0;
 	uint32_t value2 = 0;
 	uint32_t slots = 0;
@@ -1604,7 +1603,7 @@ void dce110_link_encoder_update_mst_stream_allocation_table(
 	do {
 		udelay(10);
 
-		value0 = REG_READ(DP_MSE_SAT_UPDATE);
+		REG_READ(DP_MSE_SAT_UPDATE);
 
 		REG_GET(DP_MSE_SAT_UPDATE,
 				DP_MSE_SAT_UPDATE, &value1);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c
index e459ae65aaf7..4600231da6cb 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c
@@ -97,7 +97,7 @@ enum {
 
 
 
-/**
+/*
  *	set_truncation
  *	1) set truncation depth: 0 for 18 bpp or 1 for 24 bpp
  *	2) enable truncation
@@ -142,7 +142,7 @@ static void set_truncation(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-/**
+/*
  *	dce60_set_truncation
  *	1) set truncation depth: 0 for 18 bpp or 1 for 24 bpp
  *	2) enable truncation
@@ -183,7 +183,7 @@ static void dce60_set_truncation(
 }
 #endif
 
-/**
+/*
  *	set_spatial_dither
  *	1) set spatial dithering mode: pattern of seed
  *	2) set spatial dithering depth: 0 for 18bpp or 1 for 24bpp
@@ -291,7 +291,7 @@ static void set_spatial_dither(
 		FMT_SPATIAL_DITHER_EN, 1);
 }
 
-/**
+/*
  *	SetTemporalDither (Frame Modulation)
  *	1) set temporal dither depth
  *	2) select pattern: from hard-coded pattern or programmable pattern
@@ -355,7 +355,7 @@ static void set_temporal_dither(
 		FMT_TEMPORAL_DITHER_EN, 1);
 }
 
-/**
+/*
  *	Set Clamping
  *	1) Set clamping format based on bpc - 0 for 6bpc (No clamping)
  *		1 for 8 bpc
@@ -415,7 +415,7 @@ void dce110_opp_set_clamping(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-/**
+/*
  *	Set Clamping for DCE6 parts
  *	1) Set clamping format based on bpc - 0 for 6bpc (No clamping)
  *		1 for 8 bpc
@@ -424,7 +424,7 @@ void dce110_opp_set_clamping(
  *		7 for programable
  *	2) Enable clamp if Limited range requested
  */
-void dce60_opp_set_clamping(
+static void dce60_opp_set_clamping(
 	struct dce110_opp *opp110,
 	const struct clamping_and_pixel_encoding_params *params)
 {
@@ -465,7 +465,7 @@ void dce60_opp_set_clamping(
 }
 #endif
 
-/**
+/*
  *	set_pixel_encoding
  *
  *	Set Pixel Encoding
@@ -501,7 +501,7 @@ static void set_pixel_encoding(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-/**
+/*
  *	dce60_set_pixel_encoding
  *	DCE6 has no FMT_SUBSAMPLING_{MODE,ORDER} bits in FMT_CONTROL reg
  *	Set Pixel Encoding
@@ -545,7 +545,7 @@ void dce110_opp_program_bit_depth_reduction(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-void dce60_opp_program_bit_depth_reduction(
+static void dce60_opp_program_bit_depth_reduction(
 	struct output_pixel_processor *opp,
 	const struct bit_depth_reduction_params *params)
 {
@@ -568,7 +568,7 @@ void dce110_opp_program_clamping_and_pixel_encoding(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-void dce60_opp_program_clamping_and_pixel_encoding(
+static void dce60_opp_program_clamping_and_pixel_encoding(
 	struct output_pixel_processor *opp,
 	const struct clamping_and_pixel_encoding_params *params)
 {
@@ -678,7 +678,7 @@ void dce110_opp_program_fmt(
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
-void dce60_opp_program_fmt(
+static void dce60_opp_program_fmt(
 	struct output_pixel_processor *opp,
 	struct bit_depth_reduction_params *fmt_bit_depth,
 	struct clamping_and_pixel_encoding_params *clamping)
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h
index 4d484ef60f35..bf1ffc3629c7 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h
@@ -111,7 +111,6 @@ enum dce110_opp_reg_type {
 	OPP_SF(FMT_DITHER_RAND_R_SEED, FMT_RAND_R_SEED, mask_sh),\
 	OPP_SF(FMT_DITHER_RAND_G_SEED, FMT_RAND_G_SEED, mask_sh),\
 	OPP_SF(FMT_DITHER_RAND_B_SEED, FMT_RAND_B_SEED, mask_sh),\
-	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_EN, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_RESET, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_OFFSET, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_DEPTH, mask_sh),\
@@ -219,7 +218,6 @@ enum dce110_opp_reg_type {
 	OPP_SF(FMT_DITHER_RAND_R_SEED, FMT_RAND_R_SEED, mask_sh),\
 	OPP_SF(FMT_DITHER_RAND_G_SEED, FMT_RAND_G_SEED, mask_sh),\
 	OPP_SF(FMT_DITHER_RAND_B_SEED, FMT_RAND_B_SEED, mask_sh),\
-	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_EN, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_RESET, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_OFFSET, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_DEPTH, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c b/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c
index 761fdfc1f5bd..e92339235863 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c
@@ -50,16 +50,16 @@ static unsigned int dce_get_16_bit_backlight_from_pwm(struct panel_cntl *panel_c
 {
 	uint64_t current_backlight;
 	uint32_t round_result;
-	uint32_t pwm_period_cntl, bl_period, bl_int_count;
-	uint32_t bl_pwm_cntl, bl_pwm, fractional_duty_cycle_en;
+	uint32_t bl_period, bl_int_count;
+	uint32_t bl_pwm, fractional_duty_cycle_en;
 	uint32_t bl_period_mask, bl_pwm_mask;
 	struct dce_panel_cntl *dce_panel_cntl = TO_DCE_PANEL_CNTL(panel_cntl);
 
-	pwm_period_cntl = REG_READ(BL_PWM_PERIOD_CNTL);
+	REG_READ(BL_PWM_PERIOD_CNTL);
 	REG_GET(BL_PWM_PERIOD_CNTL, BL_PWM_PERIOD, &bl_period);
 	REG_GET(BL_PWM_PERIOD_CNTL, BL_PWM_PERIOD_BITCNT, &bl_int_count);
 
-	bl_pwm_cntl = REG_READ(BL_PWM_CNTL);
+	REG_READ(BL_PWM_CNTL);
 	REG_GET(BL_PWM_CNTL, BL_ACTIVE_INT_FRAC_CNT, (uint32_t *)(&bl_pwm));
 	REG_GET(BL_PWM_CNTL, BL_PWM_FRACTIONAL_EN, &fractional_duty_cycle_en);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
index ada57f745fd7..8d4263da59f2 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
@@ -67,7 +67,6 @@ static void dce110_update_generic_info_packet(
 	uint32_t packet_index,
 	const struct dc_info_packet *info_packet)
 {
-	uint32_t regval;
 	/* TODOFPGA Figure out a proper number for max_retries polling for lock
 	 * use 50 for now.
 	 */
@@ -99,7 +98,7 @@ static void dce110_update_generic_info_packet(
 	}
 	/* choose which generic packet to use */
 	{
-		regval = REG_READ(AFMT_VBI_PACKET_CONTROL);
+		REG_READ(AFMT_VBI_PACKET_CONTROL);
 		REG_UPDATE(AFMT_VBI_PACKET_CONTROL,
 				AFMT_GENERIC_INDEX, packet_index);
 	}
@@ -564,6 +563,7 @@ static void dce110_stream_encoder_hdmi_set_stream_attribute(
 	cntl.enable_dp_audio = enable_audio;
 	cntl.pixel_clock = actual_pix_clk_khz;
 	cntl.lanes_number = LANE_COUNT_FOUR;
+	cntl.color_depth = crtc_timing->display_color_depth;
 
 	if (enc110->base.bp->funcs->encoder_control(
 			enc110->base.bp, &cntl) != BP_RESULT_OK)
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c b/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c
index 130a0a0c8332..151dc7bf6d23 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c
@@ -493,7 +493,6 @@ static void dce60_transform_set_scaler(
 {
 	struct dce_transform *xfm_dce = TO_DCE_TRANSFORM(xfm);
 	bool is_scaling_required;
-	bool filter_updated = false;
 	const uint16_t *coeffs_v, *coeffs_h;
 
 	/*Use whole line buffer memory always*/
@@ -558,7 +557,6 @@ static void dce60_transform_set_scaler(
 
 			xfm_dce->filter_v = coeffs_v;
 			xfm_dce->filter_h = coeffs_h;
-			filter_updated = true;
 		}
 	}
 
@@ -601,12 +599,12 @@ static void set_clamp(
 		clamp_max = 0x3FC0;
 		break;
 	case COLOR_DEPTH_101010:
-		/* 10bit MSB aligned on 14 bit bus '11 1111 1111 1100' */
-		clamp_max = 0x3FFC;
+		/* 10bit MSB aligned on 14 bit bus '11 1111 1111 0000' */
+		clamp_max = 0x3FF0;
 		break;
 	case COLOR_DEPTH_121212:
-		/* 12bit MSB aligned on 14 bit bus '11 1111 1111 1111' */
-		clamp_max = 0x3FFF;
+		/* 12bit MSB aligned on 14 bit bus '11 1111 1111 1100' */
+		clamp_max = 0x3FFC;
 		break;
 	default:
 		clamp_max = 0x3FC0;
@@ -1037,34 +1035,23 @@ static void dce60_transform_set_pixel_storage_depth(
 	const struct bit_depth_reduction_params *bit_depth_params)
 {
 	struct dce_transform *xfm_dce = TO_DCE_TRANSFORM(xfm);
-	int pixel_depth, expan_mode;
 	enum dc_color_depth color_depth;
 
 	switch (depth) {
 	case LB_PIXEL_DEPTH_18BPP:
 		color_depth = COLOR_DEPTH_666;
-		pixel_depth = 2;
-		expan_mode  = 1;
 		break;
 	case LB_PIXEL_DEPTH_24BPP:
 		color_depth = COLOR_DEPTH_888;
-		pixel_depth = 1;
-		expan_mode  = 1;
 		break;
 	case LB_PIXEL_DEPTH_30BPP:
 		color_depth = COLOR_DEPTH_101010;
-		pixel_depth = 0;
-		expan_mode  = 1;
 		break;
 	case LB_PIXEL_DEPTH_36BPP:
 		color_depth = COLOR_DEPTH_121212;
-		pixel_depth = 3;
-		expan_mode  = 0;
 		break;
 	default:
 		color_depth = COLOR_DEPTH_101010;
-		pixel_depth = 0;
-		expan_mode  = 1;
 		BREAK_TO_DEBUGGER();
 		break;
 	}
@@ -1113,7 +1100,7 @@ static void program_gamut_remap(
 
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: dal_transform_wide_gamut_set_gamut_remap
  *
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c
index 0cf130dc4e52..453aaa5757bd 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c
@@ -57,6 +57,7 @@ static void dmub_abm_enable_fractional_pwm(struct dc_context *dc)
 	union dmub_rb_cmd cmd;
 	uint32_t fractional_pwm = (dc->dc->config.disable_fractional_pwm == false) ? 1 : 0;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.abm_set_pwm_frac.header.type = DMUB_CMD__ABM;
 	cmd.abm_set_pwm_frac.header.sub_type = DMUB_CMD__ABM_SET_PWM_FRAC;
 	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.fractional_pwm = fractional_pwm;
@@ -135,6 +136,7 @@ static bool dmub_abm_set_level(struct abm *abm, uint32_t level)
 	union dmub_rb_cmd cmd;
 	struct dc_context *dc = abm->ctx;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.abm_set_level.header.type = DMUB_CMD__ABM;
 	cmd.abm_set_level.header.sub_type = DMUB_CMD__ABM_SET_LEVEL;
 	cmd.abm_set_level.abm_set_level_data.level = level;
@@ -160,6 +162,7 @@ static bool dmub_abm_init_config(struct abm *abm,
 	// Copy iramtable into cw7
 	memcpy(dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, (void *)src, bytes);
 
+	memset(&cmd, 0, sizeof(cmd));
 	// Fw will copy from cw7 to fw_state
 	cmd.abm_init_config.header.type = DMUB_CMD__ABM;
 	cmd.abm_init_config.header.sub_type = DMUB_CMD__ABM_INIT_CONFIG;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
index d399270fd17e..c97ee5abc0ef 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
@@ -33,8 +33,9 @@ void dmub_hw_lock_mgr_cmd(struct dc_dmub_srv *dmub_srv,
 				union dmub_hw_lock_flags *hw_locks,
 				struct dmub_hw_lock_inst_flags *inst_flags)
 {
-	union dmub_rb_cmd cmd = { 0 };
+	union dmub_rb_cmd cmd;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.lock_hw.header.type = DMUB_CMD__HW_LOCK;
 	cmd.lock_hw.header.sub_type = 0;
 	cmd.lock_hw.header.payload_bytes = sizeof(struct dmub_cmd_lock_hw_data);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
index 17e84f34ceba..69e34bef274c 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
@@ -31,7 +31,7 @@
 
 #define MAX_PIPES 6
 
-/**
+/*
  * Convert dmcub psr state to dmcu psr state.
  */
 static enum dc_psr_state convert_psr_state(uint32_t raw_state)
@@ -74,7 +74,7 @@ static enum dc_psr_state convert_psr_state(uint32_t raw_state)
 	return state;
 }
 
-/**
+/*
  * Get PSR state from firmware.
  */
 static void dmub_psr_get_state(struct dmub_psr *dmub, enum dc_psr_state *state)
@@ -90,7 +90,7 @@ static void dmub_psr_get_state(struct dmub_psr *dmub, enum dc_psr_state *state)
 	*state = convert_psr_state(raw_state);
 }
 
-/**
+/*
  * Set PSR version.
  */
 static bool dmub_psr_set_version(struct dmub_psr *dmub, struct dc_stream_state *stream)
@@ -101,6 +101,7 @@ static bool dmub_psr_set_version(struct dmub_psr *dmub, struct dc_stream_state *
 	if (stream->link->psr_settings.psr_version == DC_PSR_VERSION_UNSUPPORTED)
 		return false;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.psr_set_version.header.type = DMUB_CMD__PSR;
 	cmd.psr_set_version.header.sub_type = DMUB_CMD__PSR_SET_VERSION;
 	switch (stream->link->psr_settings.psr_version) {
@@ -121,7 +122,7 @@ static bool dmub_psr_set_version(struct dmub_psr *dmub, struct dc_stream_state *
 	return true;
 }
 
-/**
+/*
  * Enable/Disable PSR.
  */
 static void dmub_psr_enable(struct dmub_psr *dmub, bool enable, bool wait)
@@ -131,7 +132,7 @@ static void dmub_psr_enable(struct dmub_psr *dmub, bool enable, bool wait)
 	uint32_t retry_count;
 	enum dc_psr_state state = PSR_STATE0;
 
-
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.psr_enable.header.type = DMUB_CMD__PSR;
 
 	if (enable)
@@ -170,7 +171,7 @@ static void dmub_psr_enable(struct dmub_psr *dmub, bool enable, bool wait)
 	}
 }
 
-/**
+/*
  * Set PSR level.
  */
 static void dmub_psr_set_level(struct dmub_psr *dmub, uint16_t psr_level)
@@ -184,6 +185,7 @@ static void dmub_psr_set_level(struct dmub_psr *dmub, uint16_t psr_level)
 	if (state == PSR_STATE0)
 		return;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.psr_set_level.header.type = DMUB_CMD__PSR;
 	cmd.psr_set_level.header.sub_type = DMUB_CMD__PSR_SET_LEVEL;
 	cmd.psr_set_level.header.payload_bytes = sizeof(struct dmub_cmd_psr_set_level_data);
@@ -194,7 +196,7 @@ static void dmub_psr_set_level(struct dmub_psr *dmub, uint16_t psr_level)
 	dc_dmub_srv_wait_idle(dc->dmub_srv);
 }
 
-/**
+/*
  * Setup PSR by programming phy registers and sending psr hw context values to firmware.
  */
 static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
@@ -233,6 +235,7 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
 	link->link_enc->funcs->psr_program_secondary_packet(link->link_enc,
 			psr_context->sdpTransmitLineNumDeadline);
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.psr_copy_settings.header.type = DMUB_CMD__PSR;
 	cmd.psr_copy_settings.header.sub_type = DMUB_CMD__PSR_COPY_SETTINGS;
 	cmd.psr_copy_settings.header.payload_bytes = sizeof(struct dmub_cmd_psr_copy_settings_data);
@@ -277,7 +280,7 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
 	return true;
 }
 
-/**
+/*
  * Send command to PSR to force static ENTER and ignore all state changes until exit
  */
 static void dmub_psr_force_static(struct dmub_psr *dmub)
@@ -285,6 +288,7 @@ static void dmub_psr_force_static(struct dmub_psr *dmub)
 	union dmub_rb_cmd cmd;
 	struct dc_context *dc = dmub->ctx;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.psr_force_static.header.type = DMUB_CMD__PSR;
 	cmd.psr_force_static.header.sub_type = DMUB_CMD__PSR_FORCE_STATIC;
 	cmd.psr_enable.header.payload_bytes = 0;
@@ -294,7 +298,7 @@ static void dmub_psr_force_static(struct dmub_psr *dmub)
 	dc_dmub_srv_wait_idle(dc->dmub_srv);
 }
 
-/**
+/*
  * Get PSR residency from firmware.
  */
 static void dmub_psr_get_residency(struct dmub_psr *dmub, uint32_t *residency)
@@ -316,7 +320,7 @@ static const struct dmub_psr_funcs psr_funcs = {
 	.psr_get_residency		= dmub_psr_get_residency,
 };
 
-/**
+/*
  * Construct PSR object.
  */
 static void dmub_psr_construct(struct dmub_psr *psr, struct dc_context *ctx)
@@ -325,7 +329,7 @@ static void dmub_psr_construct(struct dmub_psr *psr, struct dc_context *ctx)
 	psr->funcs = &psr_funcs;
 }
 
-/**
+/*
  * Allocate and initialize PSR object.
  */
 struct dmub_psr *dmub_psr_create(struct dc_context *ctx)
@@ -342,7 +346,7 @@ struct dmub_psr *dmub_psr_create(struct dc_context *ctx)
 	return psr;
 }
 
-/**
+/*
  * Deallocate PSR object.
  */
 void dmub_psr_destroy(struct dmub_psr **dmub)
diff --git a/drivers/gpu/drm/amd/display/dc/dce100/Makefile b/drivers/gpu/drm/amd/display/dc/dce100/Makefile
index a822d4e2a169..ff20c47f559e 100644
--- a/drivers/gpu/drm/amd/display/dc/dce100/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce100/Makefile
@@ -23,6 +23,8 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
+CFLAGS_$(AMDDALPATH)/dc/dce100/dce100_resource.o = $(call cc-disable-warning, override-init)
+
 DCE100 = dce100_resource.o dce100_hw_sequencer.o
 
 AMD_DAL_DCE100 = $(addprefix $(AMDDALPATH)/dc/dce100/,$(DCE100))
diff --git a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c
index 8ab9d6c79808..635ef0e7c782 100644
--- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c
@@ -58,6 +58,8 @@
 #include "dce/dce_abm.h"
 #include "dce/dce_i2c.h"
 
+#include "dce100_resource.h"
+
 #ifndef mmMC_HUB_RDREQ_DMIF_LIMIT
 #include "gmc/gmc_8_2_d.h"
 #include "gmc/gmc_8_2_sh_mask.h"
@@ -385,7 +387,7 @@ static const struct dc_plane_cap plane_cap = {
 	.pixel_format_support = {
 			.argb8888 = true,
 			.nv12 = false,
-			.fp16 = false
+			.fp16 = true
 	},
 
 	.max_upscale_factor = {
@@ -611,7 +613,7 @@ static const struct encoder_feature_support link_enc_feature = {
 		.flags.bits.IS_TPS3_CAPABLE = true
 };
 
-struct link_encoder *dce100_link_encoder_create(
+static struct link_encoder *dce100_link_encoder_create(
 	const struct encoder_init_data *enc_init_data)
 {
 	struct dce110_link_encoder *enc110 =
@@ -650,7 +652,7 @@ static struct panel_cntl *dce100_panel_cntl_create(const struct panel_cntl_init_
 	return &panel_cntl->base;
 }
 
-struct output_pixel_processor *dce100_opp_create(
+static struct output_pixel_processor *dce100_opp_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -665,7 +667,7 @@ struct output_pixel_processor *dce100_opp_create(
 	return &opp->base;
 }
 
-struct dce_aux *dce100_aux_engine_create(
+static struct dce_aux *dce100_aux_engine_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -703,7 +705,7 @@ static const struct dce_i2c_mask i2c_masks = {
 		I2C_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(_MASK)
 };
 
-struct dce_i2c_hw *dce100_i2c_hw_create(
+static struct dce_i2c_hw *dce100_i2c_hw_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -718,7 +720,7 @@ struct dce_i2c_hw *dce100_i2c_hw_create(
 
 	return dce_i2c_hw;
 }
-struct clock_source *dce100_clock_source_create(
+static struct clock_source *dce100_clock_source_create(
 	struct dc_context *ctx,
 	struct dc_bios *bios,
 	enum clock_source_id id,
@@ -742,7 +744,7 @@ struct clock_source *dce100_clock_source_create(
 	return NULL;
 }
 
-void dce100_clock_source_destroy(struct clock_source **clk_src)
+static void dce100_clock_source_destroy(struct clock_source **clk_src)
 {
 	kfree(TO_DCE110_CLK_SRC(*clk_src));
 	*clk_src = NULL;
@@ -831,7 +833,7 @@ static enum dc_status build_mapped_resource(
 	return DC_OK;
 }
 
-bool dce100_validate_bandwidth(
+static bool dce100_validate_bandwidth(
 	struct dc  *dc,
 	struct dc_state *context,
 	bool fast_validate)
@@ -876,7 +878,7 @@ static bool dce100_validate_surface_sets(
 	return true;
 }
 
-enum dc_status dce100_validate_global(
+static enum dc_status dce100_validate_global(
 		struct dc  *dc,
 		struct dc_state *context)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/Makefile b/drivers/gpu/drm/amd/display/dc/dce110/Makefile
index d564c0eb8b04..84ab48df0c26 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce110/Makefile
@@ -23,6 +23,8 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
+CFLAGS_$(AMDDALPATH)/dc/dce110/dce110_resource.o = $(call cc-disable-warning, override-init)
+
 DCE110 = dce110_timing_generator.o \
 dce110_compressor.o dce110_hw_sequencer.o dce110_resource.o \
 dce110_opp_regamma_v.o dce110_opp_csc_v.o dce110_timing_generator_v.o \
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c
index 72b580a4eb85..44564a4742b5 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c
@@ -412,36 +412,6 @@ void dce110_compressor_destroy(struct compressor **compressor)
 	*compressor = NULL;
 }
 
-bool dce110_get_required_compressed_surfacesize(struct fbc_input_info fbc_input_info,
-						struct fbc_requested_compressed_size size)
-{
-	bool result = false;
-
-	unsigned int max_x = FBC_MAX_X, max_y = FBC_MAX_Y;
-
-	get_max_support_fbc_buffersize(&max_x, &max_y);
-
-	if (fbc_input_info.dynamic_fbc_buffer_alloc == 0) {
-		/*
-		 * For DCE11 here use Max HW supported size:  HW Support up to 3840x2400 resolution
-		 * or 18000 chunks.
-		 */
-		size.preferred_size = size.min_size = align_to_chunks_number_per_line(max_x) * max_y * 4;  /* (For FBC when LPT not supported). */
-		size.preferred_size_alignment = size.min_size_alignment = 0x100;       /* For FBC when LPT not supported */
-		size.bits.preferred_must_be_framebuffer_pool = 1;
-		size.bits.min_must_be_framebuffer_pool = 1;
-
-		result = true;
-	}
-	/*
-	 * Maybe to add registry key support with optional size here to override above
-	 * for debugging purposes
-	 */
-
-	return result;
-}
-
-
 void get_max_support_fbc_buffersize(unsigned int *max_x, unsigned int *max_y)
 {
 	*max_x = FBC_MAX_X;
@@ -455,31 +425,6 @@ void get_max_support_fbc_buffersize(unsigned int *max_x, unsigned int *max_y)
 	 */
 }
 
-
-unsigned int controller_id_to_index(enum controller_id controller_id)
-{
-	unsigned int index = 0;
-
-	switch (controller_id) {
-	case CONTROLLER_ID_D0:
-		index = 0;
-		break;
-	case CONTROLLER_ID_D1:
-		index = 1;
-		break;
-	case CONTROLLER_ID_D2:
-		index = 2;
-		break;
-	case CONTROLLER_ID_D3:
-		index = 3;
-		break;
-	default:
-		break;
-	}
-	return index;
-}
-
-
 static const struct compressor_funcs dce110_compressor_funcs = {
 	.power_up_fbc = dce110_compressor_power_up_fbc,
 	.enable_fbc = dce110_compressor_enable_fbc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index 4c230f1de9a3..caee1c9f54bd 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -921,6 +921,37 @@ void dce110_edp_power_control(
 	}
 }
 
+void dce110_edp_wait_for_T12(
+		struct dc_link *link)
+{
+	struct dc_context *ctx = link->ctx;
+
+	if (dal_graphics_object_id_get_connector_id(link->link_enc->connector)
+			!= CONNECTOR_ID_EDP) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	if (!link->panel_cntl)
+		return;
+
+	if (!link->panel_cntl->funcs->is_panel_powered_on(link->panel_cntl) &&
+			link->link_trace.time_stamp.edp_poweroff != 0) {
+		unsigned int t12_duration = 500; // Default T12 as per spec
+		unsigned long long current_ts = dm_get_timestamp(ctx);
+		unsigned long long time_since_edp_poweroff_ms =
+				div64_u64(dm_get_elapse_time_in_ns(
+						ctx,
+						current_ts,
+						link->link_trace.time_stamp.edp_poweroff), 1000000);
+
+		t12_duration += link->local_sink->edid_caps.panel_patch.extra_t12_ms; // Add extra T12
+
+		if (time_since_edp_poweroff_ms < t12_duration)
+			msleep(t12_duration - time_since_edp_poweroff_ms);
+	}
+}
+
 /*todo: cloned in stream enc, fix*/
 /*
  * @brief
@@ -1628,7 +1659,7 @@ static struct dc_link *get_edp_link_with_sink(
 	return link;
 }
 
-/**
+/*
  * When ASIC goes from VBIOS/VGA mode to driver/accelerated mode we need:
  *  1. Power down all DC HW blocks
  *  2. Disable VGA engine on all controllers
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
index d54172d88f5f..8bbb499067f7 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
@@ -34,6 +34,7 @@
 #include "inc/dce_calcs.h"
 
 #include "dce/dce_mem_input.h"
+#include "dce110_mem_input_v.h"
 
 static void set_flip_control(
 	struct dce_mem_input *mem_input110,
@@ -468,7 +469,7 @@ static void program_pixel_format(
 	}
 }
 
-bool dce_mem_input_v_is_surface_pending(struct mem_input *mem_input)
+static bool dce_mem_input_v_is_surface_pending(struct mem_input *mem_input)
 {
 	struct dce_mem_input *mem_input110 = TO_DCE_MEM_INPUT(mem_input);
 	uint32_t value;
@@ -483,7 +484,7 @@ bool dce_mem_input_v_is_surface_pending(struct mem_input *mem_input)
 	return false;
 }
 
-bool dce_mem_input_v_program_surface_flip_and_addr(
+static bool dce_mem_input_v_program_surface_flip_and_addr(
 	struct mem_input *mem_input,
 	const struct dc_plane_address *address,
 	bool flip_immediate)
@@ -560,7 +561,7 @@ static const unsigned int *get_dvmm_hw_setting(
 	}
 }
 
-void dce_mem_input_v_program_pte_vm(
+static void dce_mem_input_v_program_pte_vm(
 		struct mem_input *mem_input,
 		enum surface_pixel_format format,
 		union dc_tiling_info *tiling_info,
@@ -633,7 +634,7 @@ void dce_mem_input_v_program_pte_vm(
 	dm_write_reg(mem_input110->base.ctx, mmUNP_DVMM_PTE_ARB_CONTROL_C, value);
 }
 
-void dce_mem_input_v_program_surface_config(
+static void dce_mem_input_v_program_surface_config(
 	struct mem_input *mem_input,
 	enum surface_pixel_format format,
 	union dc_tiling_info *tiling_info,
@@ -919,7 +920,7 @@ static void program_nbp_watermark_c(
 			marks);
 }
 
-void dce_mem_input_v_program_display_marks(
+static void dce_mem_input_v_program_display_marks(
 	struct mem_input *mem_input,
 	struct dce_watermarks nbp,
 	struct dce_watermarks stutter,
@@ -942,7 +943,7 @@ void dce_mem_input_v_program_display_marks(
 
 }
 
-void dce_mem_input_program_chroma_display_marks(
+static void dce_mem_input_program_chroma_display_marks(
 	struct mem_input *mem_input,
 	struct dce_watermarks nbp,
 	struct dce_watermarks stutter,
@@ -963,7 +964,7 @@ void dce_mem_input_program_chroma_display_marks(
 		stutter);
 }
 
-void dce110_allocate_mem_input_v(
+static void dce110_allocate_mem_input_v(
 	struct mem_input *mi,
 	uint32_t h_total,/* for current stream */
 	uint32_t v_total,/* for current stream */
@@ -1005,7 +1006,7 @@ void dce110_allocate_mem_input_v(
 
 }
 
-void dce110_free_mem_input_v(
+static void dce110_free_mem_input_v(
 	struct mem_input *mi,
 	uint32_t total_stream_num)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c
index 3f63822b8e28..d7fcc5cccdce 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c
@@ -410,7 +410,7 @@ static const struct dc_plane_cap plane_cap = {
 		.pixel_format_support = {
 				.argb8888 = true,
 				.nv12 = false,
-				.fp16 = false
+				.fp16 = true
 		},
 
 		.max_upscale_factor = {
@@ -715,7 +715,7 @@ static struct output_pixel_processor *dce110_opp_create(
 	return &opp->base;
 }
 
-struct dce_aux *dce110_aux_engine_create(
+static struct dce_aux *dce110_aux_engine_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -753,7 +753,7 @@ static const struct dce_i2c_mask i2c_masks = {
 		I2C_COMMON_MASK_SH_LIST_DCE110(_MASK)
 };
 
-struct dce_i2c_hw *dce110_i2c_hw_create(
+static struct dce_i2c_hw *dce110_i2c_hw_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -768,7 +768,7 @@ struct dce_i2c_hw *dce110_i2c_hw_create(
 
 	return dce_i2c_hw;
 }
-struct clock_source *dce110_clock_source_create(
+static struct clock_source *dce110_clock_source_create(
 	struct dc_context *ctx,
 	struct dc_bios *bios,
 	enum clock_source_id id,
@@ -792,7 +792,7 @@ struct clock_source *dce110_clock_source_create(
 	return NULL;
 }
 
-void dce110_clock_source_destroy(struct clock_source **clk_src)
+static void dce110_clock_source_destroy(struct clock_source **clk_src)
 {
 	struct dce110_clk_src *dce110_clk_src;
 
@@ -1034,8 +1034,8 @@ static bool dce110_validate_bandwidth(
 	return result;
 }
 
-enum dc_status dce110_validate_plane(const struct dc_plane_state *plane_state,
-				     struct dc_caps *caps)
+static enum dc_status dce110_validate_plane(const struct dc_plane_state *plane_state,
+					    struct dc_caps *caps)
 {
 	if (((plane_state->dst_rect.width * 2) < plane_state->src_rect.width) ||
 	    ((plane_state->dst_rect.height * 2) < plane_state->src_rect.height))
@@ -1089,7 +1089,7 @@ static bool dce110_validate_surface_sets(
 	return true;
 }
 
-enum dc_status dce110_validate_global(
+static enum dc_status dce110_validate_global(
 		struct dc *dc,
 		struct dc_state *context)
 {
@@ -1272,7 +1272,6 @@ static bool underlay_create(struct dc_context *ctx, struct resource_pool *pool)
 
 	/* update the public caps to indicate an underlay is available */
 	ctx->dc->caps.max_slave_planes = 1;
-	ctx->dc->caps.max_slave_planes = 1;
 
 	return true;
 }
@@ -1333,7 +1332,7 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc)
 		1000);
 }
 
-const struct resource_caps *dce110_resource_cap(
+static const struct resource_caps *dce110_resource_cap(
 	struct hw_asic_id *asic_id)
 {
 	if (ASIC_REV_IS_STONEY(asic_id->hw_internal_rev))
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
index 1ea7db8eeb98..d88a74559edd 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
@@ -75,7 +75,7 @@ static void dce110_timing_generator_apply_front_porch_workaround(
 	}
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: is_in_vertical_blank
  *
@@ -116,7 +116,7 @@ void dce110_timing_generator_set_early_control(
 	dm_write_reg(tg->ctx, address, regval);
 }
 
-/**
+/*
  * Enable CRTC
  * Enable CRTC - call ASIC Control Object to enable Timing generator.
  */
@@ -175,7 +175,7 @@ void dce110_timing_generator_program_blank_color(
 	dm_write_reg(tg->ctx, addr, value);
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: disable_stereo
  *
@@ -226,7 +226,7 @@ static void disable_stereo(struct timing_generator *tg)
 }
 #endif
 
-/**
+/*
  * disable_crtc - call ASIC Control Object to disable Timing generator.
  */
 bool dce110_timing_generator_disable_crtc(struct timing_generator *tg)
@@ -247,11 +247,10 @@ bool dce110_timing_generator_disable_crtc(struct timing_generator *tg)
 	return result == BP_RESULT_OK;
 }
 
-/**
-* program_horz_count_by_2
-* Programs DxCRTC_HORZ_COUNT_BY2_EN - 1 for DVI 30bpp mode, 0 otherwise
-*
-*/
+/*
+ * program_horz_count_by_2
+ * Programs DxCRTC_HORZ_COUNT_BY2_EN - 1 for DVI 30bpp mode, 0 otherwise
+ */
 static void program_horz_count_by_2(
 	struct timing_generator *tg,
 	const struct dc_crtc_timing *timing)
@@ -273,7 +272,7 @@ static void program_horz_count_by_2(
 			CRTC_REG(mmCRTC_COUNT_CONTROL), regval);
 }
 
-/**
+/*
  * program_timing_generator
  * Program CRTC Timing Registers - DxCRTC_H_*, DxCRTC_V_*, Pixel repetition.
  * Call ASIC Control Object to program Timings.
@@ -352,7 +351,7 @@ bool dce110_timing_generator_program_timing_generator(
 	return result == BP_RESULT_OK;
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: set_drr
  *
@@ -521,7 +520,7 @@ uint32_t dce110_timing_generator_get_vblank_counter(struct timing_generator *tg)
 	return field;
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: dce110_timing_generator_get_position
  *
@@ -557,7 +556,7 @@ void dce110_timing_generator_get_position(struct timing_generator *tg,
 			CRTC_VERT_COUNT_NOM);
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: get_crtc_scanoutpos
  *
@@ -1106,11 +1105,11 @@ void dce110_timing_generator_set_test_pattern(
 	}
 }
 
-/**
-* dce110_timing_generator_validate_timing
-* The timing generators support a maximum display size of is 8192 x 8192 pixels,
-* including both active display and blanking periods. Check H Total and V Total.
-*/
+/*
+ * dce110_timing_generator_validate_timing
+ * The timing generators support a maximum display size of is 8192 x 8192 pixels,
+ * including both active display and blanking periods. Check H Total and V Total.
+ */
 bool dce110_timing_generator_validate_timing(
 	struct timing_generator *tg,
 	const struct dc_crtc_timing *timing,
@@ -1167,9 +1166,9 @@ bool dce110_timing_generator_validate_timing(
 	return true;
 }
 
-/**
-* Wait till we are at the beginning of VBlank.
-*/
+/*
+ * Wait till we are at the beginning of VBlank.
+ */
 void dce110_timing_generator_wait_for_vblank(struct timing_generator *tg)
 {
 	/* We want to catch beginning of VBlank here, so if the first try are
@@ -1191,9 +1190,9 @@ void dce110_timing_generator_wait_for_vblank(struct timing_generator *tg)
 	}
 }
 
-/**
-* Wait till we are in VActive (anywhere in VActive)
-*/
+/*
+ * Wait till we are in VActive (anywhere in VActive)
+ */
 void dce110_timing_generator_wait_for_vactive(struct timing_generator *tg)
 {
 	while (dce110_timing_generator_is_in_vertical_blank(tg)) {
@@ -1204,7 +1203,7 @@ void dce110_timing_generator_wait_for_vactive(struct timing_generator *tg)
 	}
 }
 
-/**
+/*
  *****************************************************************************
  *  Function: dce110_timing_generator_setup_global_swap_lock
  *
@@ -1215,7 +1214,6 @@ void dce110_timing_generator_wait_for_vactive(struct timing_generator *tg)
  *  @param [in] gsl_params: setup data
  *****************************************************************************
  */
-
 void dce110_timing_generator_setup_global_swap_lock(
 	struct timing_generator *tg,
 	const struct dcp_gsl_params *gsl_params)
@@ -1351,10 +1349,7 @@ void dce110_timing_generator_tear_down_global_swap_lock(
 
 	/* Restore DCP_GSL_PURPOSE_SURFACE_FLIP */
 	{
-		uint32_t value_crtc_vtotal;
-
-		value_crtc_vtotal = dm_read_reg(tg->ctx,
-				CRTC_REG(mmCRTC_V_TOTAL));
+		dm_read_reg(tg->ctx, CRTC_REG(mmCRTC_V_TOTAL));
 
 		set_reg_field_value(value,
 				0,
@@ -1385,7 +1380,7 @@ void dce110_timing_generator_tear_down_global_swap_lock(
 
 	dm_write_reg(tg->ctx, address, value);
 }
-/**
+/*
  *****************************************************************************
  *  Function: is_counter_moving
  *
@@ -1767,7 +1762,7 @@ void dce110_timing_generator_disable_reset_trigger(
 	dm_write_reg(tg->ctx, CRTC_REG(mmCRTC_TRIGB_CNTL), value);
 }
 
-/**
+/*
  *****************************************************************************
  *  @brief
  *     Checks whether CRTC triggered reset occurred
@@ -1794,7 +1789,7 @@ bool dce110_timing_generator_did_triggered_reset_occur(
 	return (force || vert_sync);
 }
 
-/**
+/*
  * dce110_timing_generator_disable_vga
  * Turn OFF VGA Mode and Timing  - DxVGA_CONTROL
  * VGA Mode and VGA Timing is used by VBIOS on CRT Monitors;
@@ -1840,14 +1835,13 @@ void dce110_timing_generator_disable_vga(
 	dm_write_reg(tg->ctx, addr, value);
 }
 
-/**
-* set_overscan_color_black
-*
-* @param :black_color is one of the color space
-*    :this routine will set overscan black color according to the color space.
-* @return none
-*/
-
+/*
+ * set_overscan_color_black
+ *
+ * @param :black_color is one of the color space
+ *    :this routine will set overscan black color according to the color space.
+ * @return none
+ */
 void dce110_timing_generator_set_overscan_color_black(
 	struct timing_generator *tg,
 	const struct tg_color *color)
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c
index a13a2f58944e..c509384fff54 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c
@@ -46,17 +46,16 @@
  *
  **********************************************************************************/
 
-/**
-* Enable CRTCV
-*/
+/*
+ * Enable CRTCV
+ */
 
 static bool dce110_timing_generator_v_enable_crtc(struct timing_generator *tg)
 {
 /*
-* Set MASTER_UPDATE_MODE to 0
-* This is needed for DRR, and also suggested to be default value by Syed.
-*/
-
+ * Set MASTER_UPDATE_MODE to 0
+ * This is needed for DRR, and also suggested to be default value by Syed.
+ */
 	uint32_t value;
 
 	value = 0;
@@ -209,9 +208,9 @@ static void dce110_timing_generator_v_wait_for_vblank(struct timing_generator *t
 	}
 }
 
-/**
-* Wait till we are in VActive (anywhere in VActive)
-*/
+/*
+ * Wait till we are in VActive (anywhere in VActive)
+ */
 static void dce110_timing_generator_v_wait_for_vactive(struct timing_generator *tg)
 {
 	while (dce110_timing_generator_v_is_in_vertical_blank(tg)) {
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c
index b1aaab5590cc..29438c6050db 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c
@@ -217,16 +217,15 @@ static bool setup_scaling_configuration(
 	return is_scaling_needed;
 }
 
-/**
-* Function:
-* void program_overscan
-*
-* Purpose: Programs overscan border
-* Input:   overscan
-*
-* Output:
-   void
-*/
+/*
+ * Function:
+ * void program_overscan
+ *
+ * Purpose: Programs overscan border
+ * Input:   overscan
+ *
+ * Output: void
+ */
 static void program_overscan(
 		struct dce_transform *xfm_dce,
 		const struct scaler_data *data)
diff --git a/drivers/gpu/drm/amd/display/dc/dce112/Makefile b/drivers/gpu/drm/amd/display/dc/dce112/Makefile
index 8e090446d511..9de6501702d2 100644
--- a/drivers/gpu/drm/amd/display/dc/dce112/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce112/Makefile
@@ -23,6 +23,8 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
+CFLAGS_$(AMDDALPATH)/dc/dce112/dce112_resource.o = $(call cc-disable-warning, override-init)
+
 DCE112 = dce112_compressor.o dce112_hw_sequencer.o \
 dce112_resource.o
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c
index f99b1c084590..ee55cda854bf 100644
--- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c
@@ -59,7 +59,9 @@
 #include "dce/dce_11_2_sh_mask.h"
 
 #include "dce100/dce100_resource.h"
-#define DC_LOGGER \
+#include "dce112_resource.h"
+
+#define DC_LOGGER				\
 		dc->ctx->logger
 
 #ifndef mmDP_DPHY_INTERNAL_CTRL
@@ -617,7 +619,7 @@ static const struct encoder_feature_support link_enc_feature = {
 		.flags.bits.IS_TPS4_CAPABLE = true
 };
 
-struct link_encoder *dce112_link_encoder_create(
+static struct link_encoder *dce112_link_encoder_create(
 	const struct encoder_init_data *enc_init_data)
 {
 	struct dce110_link_encoder *enc110 =
@@ -671,7 +673,7 @@ static struct input_pixel_processor *dce112_ipp_create(
 	return &ipp->base;
 }
 
-struct output_pixel_processor *dce112_opp_create(
+static struct output_pixel_processor *dce112_opp_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -686,7 +688,7 @@ struct output_pixel_processor *dce112_opp_create(
 	return &opp->base;
 }
 
-struct dce_aux *dce112_aux_engine_create(
+static struct dce_aux *dce112_aux_engine_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -724,7 +726,7 @@ static const struct dce_i2c_mask i2c_masks = {
 		I2C_COMMON_MASK_SH_LIST_DCE110(_MASK)
 };
 
-struct dce_i2c_hw *dce112_i2c_hw_create(
+static struct dce_i2c_hw *dce112_i2c_hw_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -739,7 +741,7 @@ struct dce_i2c_hw *dce112_i2c_hw_create(
 
 	return dce_i2c_hw;
 }
-struct clock_source *dce112_clock_source_create(
+static struct clock_source *dce112_clock_source_create(
 	struct dc_context *ctx,
 	struct dc_bios *bios,
 	enum clock_source_id id,
@@ -763,7 +765,7 @@ struct clock_source *dce112_clock_source_create(
 	return NULL;
 }
 
-void dce112_clock_source_destroy(struct clock_source **clk_src)
+static void dce112_clock_source_destroy(struct clock_source **clk_src)
 {
 	kfree(TO_DCE110_CLK_SRC(*clk_src));
 	*clk_src = NULL;
@@ -1024,7 +1026,7 @@ enum dc_status dce112_add_stream_to_ctx(
 	return result;
 }
 
-enum dc_status dce112_validate_global(
+static enum dc_status dce112_validate_global(
 		struct dc *dc,
 		struct dc_state *context)
 {
@@ -1202,7 +1204,7 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc)
 	dm_pp_notify_wm_clock_changes(dc->ctx, &clk_ranges);
 }
 
-const struct resource_caps *dce112_resource_cap(
+static const struct resource_caps *dce112_resource_cap(
 	struct hw_asic_id *asic_id)
 {
 	if (ASIC_REV_IS_POLARIS11_M(asic_id->hw_internal_rev) ||
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/Makefile b/drivers/gpu/drm/amd/display/dc/dce120/Makefile
index 37db1f8d45ea..a9cc4b73270b 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce120/Makefile
@@ -24,6 +24,8 @@
 # It provides the control and status of HW CRTC block.
 
 
+CFLAGS_$(AMDDALPATH)/dc/dce120/dce120_resource.o = $(call cc-disable-warning, override-init)
+
 DCE120 = dce120_resource.o dce120_timing_generator.o \
 dce120_hw_sequencer.o
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c
index 66a13aa39c95..d4afe6c824d2 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c
@@ -50,6 +50,7 @@ struct dce120_hw_seq_reg_offsets {
 	uint32_t crtc;
 };
 
+#if 0
 static const struct dce120_hw_seq_reg_offsets reg_offsets[] = {
 {
 	.crtc = (mmCRTC0_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
@@ -79,7 +80,6 @@ static const struct dce120_hw_seq_reg_offsets reg_offsets[] = {
 /*******************************************************************************
  * Private definitions
  ******************************************************************************/
-#if 0
 static void dce120_init_pte(struct dc_context *ctx, uint8_t controller_id)
 {
 	uint32_t addr;
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c
index f1e3d2888eac..c65e4d125c8e 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c
@@ -423,7 +423,7 @@ static const struct dce110_clk_src_mask cs_mask = {
 		CS_COMMON_MASK_SH_LIST_DCE_112(_MASK)
 };
 
-struct output_pixel_processor *dce120_opp_create(
+static struct output_pixel_processor *dce120_opp_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -437,7 +437,7 @@ struct output_pixel_processor *dce120_opp_create(
 			     ctx, inst, &opp_regs[inst], &opp_shift, &opp_mask);
 	return &opp->base;
 }
-struct dce_aux *dce120_aux_engine_create(
+static struct dce_aux *dce120_aux_engine_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -475,7 +475,7 @@ static const struct dce_i2c_mask i2c_masks = {
 		I2C_COMMON_MASK_SH_LIST_DCE110(_MASK)
 };
 
-struct dce_i2c_hw *dce120_i2c_hw_create(
+static struct dce_i2c_hw *dce120_i2c_hw_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
index 915fbb8e8168..b57c466124e7 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
@@ -69,7 +69,7 @@
 #define CRTC_REG_SET_3(reg, field1, val1, field2, val2, field3, val3)	\
 		CRTC_REG_SET_N(reg, 3, FD(reg##__##field1), val1, FD(reg##__##field2), val2, FD(reg##__##field3), val3)
 
-/**
+/*
  *****************************************************************************
  *  Function: is_in_vertical_blank
  *
@@ -98,7 +98,7 @@ static bool dce120_timing_generator_is_in_vertical_blank(
 
 
 /* determine if given timing can be supported by TG */
-bool dce120_timing_generator_validate_timing(
+static bool dce120_timing_generator_validate_timing(
 	struct timing_generator *tg,
 	const struct dc_crtc_timing *timing,
 	enum signal_type signal)
@@ -125,7 +125,7 @@ bool dce120_timing_generator_validate_timing(
 	return true;
 }
 
-bool dce120_tg_validate_timing(struct timing_generator *tg,
+static bool dce120_tg_validate_timing(struct timing_generator *tg,
 	const struct dc_crtc_timing *timing)
 {
 	return dce120_timing_generator_validate_timing(tg, timing, SIGNAL_TYPE_NONE);
@@ -133,7 +133,7 @@ bool dce120_tg_validate_timing(struct timing_generator *tg,
 
 /******** HW programming ************/
 /* Disable/Enable Timing Generator */
-bool dce120_timing_generator_enable_crtc(struct timing_generator *tg)
+static bool dce120_timing_generator_enable_crtc(struct timing_generator *tg)
 {
 	enum bp_result result;
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -153,7 +153,7 @@ bool dce120_timing_generator_enable_crtc(struct timing_generator *tg)
 	return result == BP_RESULT_OK;
 }
 
-void dce120_timing_generator_set_early_control(
+static void dce120_timing_generator_set_early_control(
 		struct timing_generator *tg,
 		uint32_t early_cntl)
 {
@@ -166,7 +166,7 @@ void dce120_timing_generator_set_early_control(
 /**************** TG current status ******************/
 
 /* return the current frame counter. Used by Linux kernel DRM */
-uint32_t dce120_timing_generator_get_vblank_counter(
+static uint32_t dce120_timing_generator_get_vblank_counter(
 		struct timing_generator *tg)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -181,7 +181,7 @@ uint32_t dce120_timing_generator_get_vblank_counter(
 }
 
 /* Get current H and V position */
-void dce120_timing_generator_get_crtc_position(
+static void dce120_timing_generator_get_crtc_position(
 	struct timing_generator *tg,
 	struct crtc_position *position)
 {
@@ -207,7 +207,7 @@ void dce120_timing_generator_get_crtc_position(
 }
 
 /* wait until TG is in beginning of vertical blank region */
-void dce120_timing_generator_wait_for_vblank(struct timing_generator *tg)
+static void dce120_timing_generator_wait_for_vblank(struct timing_generator *tg)
 {
 	/* We want to catch beginning of VBlank here, so if the first try are
 	 * in VBlank, we might be very close to Active, in this case wait for
@@ -229,7 +229,7 @@ void dce120_timing_generator_wait_for_vblank(struct timing_generator *tg)
 }
 
 /* wait until TG is in beginning of active region */
-void dce120_timing_generator_wait_for_vactive(struct timing_generator *tg)
+static void dce120_timing_generator_wait_for_vactive(struct timing_generator *tg)
 {
 	while (dce120_timing_generator_is_in_vertical_blank(tg)) {
 		if (!tg->funcs->is_counter_moving(tg)) {
@@ -242,7 +242,7 @@ void dce120_timing_generator_wait_for_vactive(struct timing_generator *tg)
 /*********** Timing Generator Synchronization routines ****/
 
 /* Setups Global Swap Lock group, TimingServer or TimingClient*/
-void dce120_timing_generator_setup_global_swap_lock(
+static void dce120_timing_generator_setup_global_swap_lock(
 	struct timing_generator *tg,
 	const struct dcp_gsl_params *gsl_params)
 {
@@ -279,7 +279,7 @@ void dce120_timing_generator_setup_global_swap_lock(
 }
 
 /* Clear all the register writes done by setup_global_swap_lock */
-void dce120_timing_generator_tear_down_global_swap_lock(
+static void dce120_timing_generator_tear_down_global_swap_lock(
 	struct timing_generator *tg)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -300,7 +300,7 @@ void dce120_timing_generator_tear_down_global_swap_lock(
 }
 
 /* Reset slave controllers on master VSync */
-void dce120_timing_generator_enable_reset_trigger(
+static void dce120_timing_generator_enable_reset_trigger(
 	struct timing_generator *tg,
 	int source)
 {
@@ -347,7 +347,7 @@ void dce120_timing_generator_enable_reset_trigger(
 }
 
 /* disabling trigger-reset */
-void dce120_timing_generator_disable_reset_trigger(
+static void dce120_timing_generator_disable_reset_trigger(
 	struct timing_generator *tg)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -367,7 +367,7 @@ void dce120_timing_generator_disable_reset_trigger(
 }
 
 /* Checks whether CRTC triggered reset occurred */
-bool dce120_timing_generator_did_triggered_reset_occur(
+static bool dce120_timing_generator_did_triggered_reset_occur(
 	struct timing_generator *tg)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -384,7 +384,7 @@ bool dce120_timing_generator_did_triggered_reset_occur(
 
 /******** Stuff to move to other virtual HW objects *****************/
 /* Move to enable accelerated mode */
-void dce120_timing_generator_disable_vga(struct timing_generator *tg)
+static void dce120_timing_generator_disable_vga(struct timing_generator *tg)
 {
 	uint32_t offset = 0;
 	uint32_t value = 0;
@@ -425,7 +425,7 @@ void dce120_timing_generator_disable_vga(struct timing_generator *tg)
 }
 /* TODO: Should we move it to transform */
 /* Fully program CRTC timing in timing generator */
-void dce120_timing_generator_program_blanking(
+static void dce120_timing_generator_program_blanking(
 	struct timing_generator *tg,
 	const struct dc_crtc_timing *timing)
 {
@@ -485,7 +485,7 @@ void dce120_timing_generator_program_blanking(
 
 /* TODO: Should we move it to opp? */
 /* Combine with below and move YUV/RGB color conversion to SW layer */
-void dce120_timing_generator_program_blank_color(
+static void dce120_timing_generator_program_blank_color(
 	struct timing_generator *tg,
 	const struct tg_color *black_color)
 {
@@ -498,7 +498,7 @@ void dce120_timing_generator_program_blank_color(
 		CRTC_BLACK_COLOR_R_CR, black_color->color_r_cr);
 }
 /* Combine with above and move YUV/RGB color conversion to SW layer */
-void dce120_timing_generator_set_overscan_color_black(
+static void dce120_timing_generator_set_overscan_color_black(
 	struct timing_generator *tg,
 	const struct tg_color *color)
 {
@@ -540,7 +540,7 @@ void dce120_timing_generator_set_overscan_color_black(
 	 */
 }
 
-void dce120_timing_generator_set_drr(
+static void dce120_timing_generator_set_drr(
 	struct timing_generator *tg,
 	const struct drr_params *params)
 {
@@ -589,50 +589,7 @@ void dce120_timing_generator_set_drr(
 	}
 }
 
-/**
- *****************************************************************************
- *  Function: dce120_timing_generator_get_position
- *
- *  @brief
- *     Returns CRTC vertical/horizontal counters
- *
- *  @param [out] position
- *****************************************************************************
- */
-void dce120_timing_generator_get_position(struct timing_generator *tg,
-	struct crtc_position *position)
-{
-	uint32_t value;
-	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
-
-	value = dm_read_reg_soc15(
-			tg->ctx,
-			mmCRTC0_CRTC_STATUS_POSITION,
-			tg110->offsets.crtc);
-
-	position->horizontal_count = get_reg_field_value(
-			value,
-			CRTC0_CRTC_STATUS_POSITION,
-			CRTC_HORZ_COUNT);
-
-	position->vertical_count = get_reg_field_value(
-			value,
-			CRTC0_CRTC_STATUS_POSITION,
-			CRTC_VERT_COUNT);
-
-	value = dm_read_reg_soc15(
-			tg->ctx,
-			mmCRTC0_CRTC_NOM_VERT_POSITION,
-			tg110->offsets.crtc);
-
-	position->nominal_vcount = get_reg_field_value(
-			value,
-			CRTC0_CRTC_NOM_VERT_POSITION,
-			CRTC_VERT_COUNT_NOM);
-}
-
-
-void dce120_timing_generator_get_crtc_scanoutpos(
+static void dce120_timing_generator_get_crtc_scanoutpos(
 	struct timing_generator *tg,
 	uint32_t *v_blank_start,
 	uint32_t *v_blank_end,
@@ -661,7 +618,7 @@ void dce120_timing_generator_get_crtc_scanoutpos(
 	*v_position = position.vertical_count;
 }
 
-void dce120_timing_generator_enable_advanced_request(
+static void dce120_timing_generator_enable_advanced_request(
 	struct timing_generator *tg,
 	bool enable,
 	const struct dc_crtc_timing *timing)
@@ -699,7 +656,7 @@ void dce120_timing_generator_enable_advanced_request(
 			value);
 }
 
-void dce120_tg_program_blank_color(struct timing_generator *tg,
+static void dce120_tg_program_blank_color(struct timing_generator *tg,
 	const struct tg_color *black_color)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -722,7 +679,7 @@ void dce120_tg_program_blank_color(struct timing_generator *tg,
 		value);
 }
 
-void dce120_tg_set_overscan_color(struct timing_generator *tg,
+static void dce120_tg_set_overscan_color(struct timing_generator *tg,
 	const struct tg_color *overscan_color)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -749,7 +706,7 @@ static void dce120_tg_program_timing(struct timing_generator *tg,
 		dce120_timing_generator_program_blanking(tg, timing);
 }
 
-bool dce120_tg_is_blanked(struct timing_generator *tg)
+static bool dce120_tg_is_blanked(struct timing_generator *tg)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
 	uint32_t value = dm_read_reg_soc15(
@@ -770,7 +727,7 @@ bool dce120_tg_is_blanked(struct timing_generator *tg)
 	return false;
 }
 
-void dce120_tg_set_blank(struct timing_generator *tg,
+static void dce120_tg_set_blank(struct timing_generator *tg,
 		bool enable_blanking)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
@@ -789,7 +746,7 @@ void dce120_tg_set_blank(struct timing_generator *tg,
 bool dce120_tg_validate_timing(struct timing_generator *tg,
 	const struct dc_crtc_timing *timing);
 
-void dce120_tg_wait_for_state(struct timing_generator *tg,
+static void dce120_tg_wait_for_state(struct timing_generator *tg,
 	enum crtc_state state)
 {
 	switch (state) {
@@ -806,7 +763,7 @@ void dce120_tg_wait_for_state(struct timing_generator *tg,
 	}
 }
 
-void dce120_tg_set_colors(struct timing_generator *tg,
+static void dce120_tg_set_colors(struct timing_generator *tg,
 	const struct tg_color *blank_color,
 	const struct tg_color *overscan_color)
 {
@@ -833,7 +790,7 @@ static void dce120_timing_generator_set_static_screen_control(
 			CRTC_STATIC_SCREEN_FRAME_COUNT, num_frames);
 }
 
-void dce120_timing_generator_set_test_pattern(
+static void dce120_timing_generator_set_test_pattern(
 	struct timing_generator *tg,
 	/* TODO: replace 'controller_dp_test_pattern' by 'test_pattern_mode'
 	 * because this is not DP-specific (which is probably somewhere in DP
diff --git a/drivers/gpu/drm/amd/display/dc/dce60/Makefile b/drivers/gpu/drm/amd/display/dc/dce60/Makefile
index 7036c3bd0f87..dda596fa1cd7 100644
--- a/drivers/gpu/drm/amd/display/dc/dce60/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce60/Makefile
@@ -23,6 +23,8 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
+CFLAGS_AMDDALPATH)/dc/dce60/dce60_resource.o = $(call cc-disable-warning, override-init)
+
 DCE60 = dce60_timing_generator.o dce60_hw_sequencer.o \
 	dce60_resource.o
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.c b/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.c
index e9dd78c484d6..dcfa0a3efa00 100644
--- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.c
@@ -60,6 +60,8 @@
 #include "dce/dce_i2c.h"
 /* TODO remove this include */
 
+#include "dce60_resource.h"
+
 #ifndef mmMC_HUB_RDREQ_DMIF_LIMIT
 #include "gmc/gmc_6_0_d.h"
 #include "gmc/gmc_6_0_sh_mask.h"
@@ -519,7 +521,7 @@ static struct output_pixel_processor *dce60_opp_create(
 	return &opp->base;
 }
 
-struct dce_aux *dce60_aux_engine_create(
+static struct dce_aux *dce60_aux_engine_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -557,7 +559,7 @@ static const struct dce_i2c_mask i2c_masks = {
 		I2C_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(_MASK)
 };
 
-struct dce_i2c_hw *dce60_i2c_hw_create(
+static struct dce_i2c_hw *dce60_i2c_hw_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -573,7 +575,7 @@ struct dce_i2c_hw *dce60_i2c_hw_create(
 	return dce_i2c_hw;
 }
 
-struct dce_i2c_sw *dce60_i2c_sw_create(
+static struct dce_i2c_sw *dce60_i2c_sw_create(
 	struct dc_context *ctx)
 {
 	struct dce_i2c_sw *dce_i2c_sw =
@@ -707,7 +709,7 @@ static const struct encoder_feature_support link_enc_feature = {
 		.flags.bits.IS_TPS3_CAPABLE = true
 };
 
-struct link_encoder *dce60_link_encoder_create(
+static struct link_encoder *dce60_link_encoder_create(
 	const struct encoder_init_data *enc_init_data)
 {
 	struct dce110_link_encoder *enc110 =
@@ -746,7 +748,7 @@ static struct panel_cntl *dce60_panel_cntl_create(const struct panel_cntl_init_d
 	return &panel_cntl->base;
 }
 
-struct clock_source *dce60_clock_source_create(
+static struct clock_source *dce60_clock_source_create(
 	struct dc_context *ctx,
 	struct dc_bios *bios,
 	enum clock_source_id id,
@@ -770,7 +772,7 @@ struct clock_source *dce60_clock_source_create(
 	return NULL;
 }
 
-void dce60_clock_source_destroy(struct clock_source **clk_src)
+static void dce60_clock_source_destroy(struct clock_source **clk_src)
 {
 	kfree(TO_DCE110_CLK_SRC(*clk_src));
 	*clk_src = NULL;
@@ -860,7 +862,7 @@ static void dce60_resource_destruct(struct dce110_resource_pool *pool)
 	}
 }
 
-bool dce60_validate_bandwidth(
+static bool dce60_validate_bandwidth(
 	struct dc *dc,
 	struct dc_state *context,
 	bool fast_validate)
@@ -905,7 +907,7 @@ static bool dce60_validate_surface_sets(
 	return true;
 }
 
-enum dc_status dce60_validate_global(
+static enum dc_status dce60_validate_global(
 		struct dc *dc,
 		struct dc_state *context)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c
index fc1af0ff0ca4..c1a85ee374d9 100644
--- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c
@@ -189,8 +189,8 @@ static bool dce60_is_tg_enabled(struct timing_generator *tg)
 	return field == 1;
 }
 
-bool dce60_configure_crc(struct timing_generator *tg,
-			  const struct crc_params *params)
+static bool dce60_configure_crc(struct timing_generator *tg,
+				const struct crc_params *params)
 {
 	/* Cannot configure crc on a CRTC that is disabled */
 	if (!dce60_is_tg_enabled(tg))
diff --git a/drivers/gpu/drm/amd/display/dc/dce80/Makefile b/drivers/gpu/drm/amd/display/dc/dce80/Makefile
index 666fcb2bdbba..0a9d1a350d8b 100644
--- a/drivers/gpu/drm/amd/display/dc/dce80/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce80/Makefile
@@ -23,6 +23,8 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
+CFLAGS_$(AMDDALPATH)/dc/dce80/dce80_resource.o = $(call cc-disable-warning, override-init)
+
 DCE80 = dce80_timing_generator.o dce80_hw_sequencer.o \
 	dce80_resource.o
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c
index 390a0fa37239..612450f99278 100644
--- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c
@@ -60,6 +60,8 @@
 #include "dce/dce_i2c.h"
 /* TODO remove this include */
 
+#include "dce80_resource.h"
+
 #ifndef mmMC_HUB_RDREQ_DMIF_LIMIT
 #include "gmc/gmc_7_1_d.h"
 #include "gmc/gmc_7_1_sh_mask.h"
@@ -402,7 +404,7 @@ static const struct dc_plane_cap plane_cap = {
 	.pixel_format_support = {
 			.argb8888 = true,
 			.nv12 = false,
-			.fp16 = false
+			.fp16 = true
 	},
 
 	.max_upscale_factor = {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/Makefile b/drivers/gpu/drm/amd/display/dc/dcn10/Makefile
index 733e6e6e43bd..62ad1a11bff9 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/Makefile
@@ -31,11 +31,4 @@ DCN10 = dcn10_init.o dcn10_resource.o dcn10_ipp.o dcn10_hw_sequencer.o \
 
 AMD_DAL_DCN10 = $(addprefix $(AMDDALPATH)/dc/dcn10/,$(DCN10))
 
-# fix:
-# ...: '-mgeneral-regs-only' is incompatible with the use of floating-point types
-# aarch64 does not support soft-float, so use hard-float and handle this in code
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn10/dcn10_resource.o := -mgeneral-regs-only
-endif
-
 AMD_DISPLAY_FILES += $(AMD_DAL_DCN10)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c
index 4d3f7d5e1473..904c2d278998 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c
@@ -577,7 +577,7 @@ void dpp1_power_on_degamma_lut(
 	struct dcn10_dpp *dpp = TO_DCN10_DPP(dpp_base);
 
 	REG_SET(CM_MEM_PWR_CTRL, 0,
-			SHARED_MEM_PWR_DIS, power_on == true ? 0:1);
+			SHARED_MEM_PWR_DIS, power_on ? 0:1);
 
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
index 9e796dfeac20..714c71a5fbde 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
@@ -1257,6 +1257,16 @@ void hubp1_soft_reset(struct hubp *hubp, bool reset)
 	REG_UPDATE(DCHUBP_CNTL, HUBP_DISABLE, reset ? 1 : 0);
 }
 
+void hubp1_set_flip_int(struct hubp *hubp)
+{
+	struct dcn10_hubp *hubp1 = TO_DCN10_HUBP(hubp);
+
+	REG_UPDATE(DCSURF_SURFACE_FLIP_INTERRUPT,
+		SURFACE_FLIP_INT_MASK, 1);
+
+	return;
+}
+
 void hubp1_init(struct hubp *hubp)
 {
 	//do nothing
@@ -1290,6 +1300,7 @@ static const struct hubp_funcs dcn10_hubp_funcs = {
 	.dmdata_load = NULL,
 	.hubp_soft_reset = hubp1_soft_reset,
 	.hubp_in_blank = hubp1_in_blank,
+	.hubp_set_flip_int = hubp1_set_flip_int,
 };
 
 /*****************************************/
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h
index a9a6ed7f4f99..e2f2f6995935 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h
@@ -74,6 +74,7 @@
 	SRI(DCSURF_SURFACE_EARLIEST_INUSE_C, HUBPREQ, id),\
 	SRI(DCSURF_SURFACE_EARLIEST_INUSE_HIGH_C, HUBPREQ, id),\
 	SRI(DCSURF_SURFACE_CONTROL, HUBPREQ, id),\
+	SRI(DCSURF_SURFACE_FLIP_INTERRUPT, HUBPREQ, id),\
 	SRI(HUBPRET_CONTROL, HUBPRET, id),\
 	SRI(DCN_EXPANSION_MODE, HUBPREQ, id),\
 	SRI(DCHUBP_REQ_SIZE_CONFIG, HUBP, id),\
@@ -183,6 +184,7 @@
 	uint32_t DCSURF_SURFACE_EARLIEST_INUSE_C; \
 	uint32_t DCSURF_SURFACE_EARLIEST_INUSE_HIGH_C; \
 	uint32_t DCSURF_SURFACE_CONTROL; \
+	uint32_t DCSURF_SURFACE_FLIP_INTERRUPT; \
 	uint32_t HUBPRET_CONTROL; \
 	uint32_t DCN_EXPANSION_MODE; \
 	uint32_t DCHUBP_REQ_SIZE_CONFIG; \
@@ -332,6 +334,7 @@
 	HUBP_SF(HUBPREQ0_DCSURF_SURFACE_CONTROL, SECONDARY_META_SURFACE_TMZ_C, mask_sh),\
 	HUBP_SF(HUBPREQ0_DCSURF_SURFACE_CONTROL, SECONDARY_SURFACE_DCC_EN, mask_sh),\
 	HUBP_SF(HUBPREQ0_DCSURF_SURFACE_CONTROL, SECONDARY_SURFACE_DCC_IND_64B_BLK, mask_sh),\
+	HUBP_SF(HUBPREQ0_DCSURF_SURFACE_FLIP_INTERRUPT, SURFACE_FLIP_INT_MASK, mask_sh),\
 	HUBP_SF(HUBPRET0_HUBPRET_CONTROL, DET_BUF_PLANE1_BASE_ADDRESS, mask_sh),\
 	HUBP_SF(HUBPRET0_HUBPRET_CONTROL, CROSSBAR_SRC_CB_B, mask_sh),\
 	HUBP_SF(HUBPRET0_HUBPRET_CONTROL, CROSSBAR_SRC_CR_R, mask_sh),\
@@ -531,6 +534,7 @@
 	type PRIMARY_SURFACE_DCC_IND_64B_BLK;\
 	type SECONDARY_SURFACE_DCC_EN;\
 	type SECONDARY_SURFACE_DCC_IND_64B_BLK;\
+	type SURFACE_FLIP_INT_MASK;\
 	type DET_BUF_PLANE1_BASE_ADDRESS;\
 	type CROSSBAR_SRC_CB_B;\
 	type CROSSBAR_SRC_CR_R;\
@@ -777,4 +781,6 @@ void hubp1_read_state_common(struct hubp *hubp);
 bool hubp1_in_blank(struct hubp *hubp);
 void hubp1_soft_reset(struct hubp *hubp, bool reset);
 
+void hubp1_set_flip_int(struct hubp *hubp);
+
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index cfc130e2d6fd..9ba5c624770d 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -647,8 +647,13 @@ static void power_on_plane(
 	if (REG(DC_IP_REQUEST_CNTL)) {
 		REG_SET(DC_IP_REQUEST_CNTL, 0,
 				IP_REQUEST_EN, 1);
-		hws->funcs.dpp_pg_control(hws, plane_id, true);
-		hws->funcs.hubp_pg_control(hws, plane_id, true);
+
+		if (hws->funcs.dpp_pg_control)
+			hws->funcs.dpp_pg_control(hws, plane_id, true);
+
+		if (hws->funcs.hubp_pg_control)
+			hws->funcs.hubp_pg_control(hws, plane_id, true);
+
 		REG_SET(DC_IP_REQUEST_CNTL, 0,
 				IP_REQUEST_EN, 0);
 		DC_LOG_DEBUG(
@@ -1082,8 +1087,13 @@ void dcn10_plane_atomic_power_down(struct dc *dc,
 	if (REG(DC_IP_REQUEST_CNTL)) {
 		REG_SET(DC_IP_REQUEST_CNTL, 0,
 				IP_REQUEST_EN, 1);
-		hws->funcs.dpp_pg_control(hws, dpp->inst, false);
-		hws->funcs.hubp_pg_control(hws, hubp->inst, false);
+
+		if (hws->funcs.dpp_pg_control)
+			hws->funcs.dpp_pg_control(hws, dpp->inst, false);
+
+		if (hws->funcs.hubp_pg_control)
+			hws->funcs.hubp_pg_control(hws, hubp->inst, false);
+
 		dpp->funcs->dpp_reset(dpp);
 		REG_SET(DC_IP_REQUEST_CNTL, 0,
 				IP_REQUEST_EN, 0);
@@ -1214,6 +1224,7 @@ void dcn10_init_pipes(struct dc *dc, struct dc_state *context)
 			// signals when OTG blanked. This is to prevent pipe from
 			// requesting data while in PSR.
 			tg->funcs->tg_init(tg);
+			hubp->power_gated = true;
 			continue;
 		}
 
@@ -2185,6 +2196,13 @@ static void dcn10_enable_plane(
 	if (dc->debug.sanity_checks) {
 		hws->funcs.verify_allow_pstate_change_high(dc);
 	}
+
+	if (!pipe_ctx->top_pipe
+		&& pipe_ctx->plane_state
+		&& pipe_ctx->plane_state->flip_int_enabled
+		&& pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int)
+			pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int(pipe_ctx->plane_res.hubp);
+
 }
 
 void dcn10_program_gamut_remap(struct pipe_ctx *pipe_ctx)
@@ -2624,7 +2642,7 @@ static void dcn10_update_dchubp_dpp(
 	hws->funcs.update_plane_addr(dc, pipe_ctx);
 
 	if (is_pipe_tree_visible(pipe_ctx))
-		hubp->funcs->set_blank(hubp, false);
+		dc->hwss.set_hubp_blank(dc, pipe_ctx, false);
 }
 
 void dcn10_blank_pixel_data(
@@ -3124,7 +3142,7 @@ void dcn10_setup_stereo(struct pipe_ctx *pipe_ctx, struct dc *dc)
 
 	pipe_ctx->stream_res.opp->funcs->opp_program_stereo(
 		pipe_ctx->stream_res.opp,
-		flags.PROGRAM_STEREO == 1 ? true:false,
+		flags.PROGRAM_STEREO == 1,
 		&stream->timing);
 
 	pipe_ctx->stream_res.tg->funcs->program_stereo(
@@ -3135,13 +3153,16 @@ void dcn10_setup_stereo(struct pipe_ctx *pipe_ctx, struct dc *dc)
 	return;
 }
 
-static struct hubp *get_hubp_by_inst(struct resource_pool *res_pool, int mpcc_inst)
+static struct pipe_ctx *get_pipe_ctx_by_hubp_inst(struct dc_state *context, int mpcc_inst)
 {
 	int i;
 
-	for (i = 0; i < res_pool->pipe_count; i++) {
-		if (res_pool->hubps[i]->inst == mpcc_inst)
-			return res_pool->hubps[i];
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (context->res_ctx.pipe_ctx[i].plane_res.hubp
+				&& context->res_ctx.pipe_ctx[i].plane_res.hubp->inst == mpcc_inst) {
+			return &context->res_ctx.pipe_ctx[i];
+		}
+
 	}
 	ASSERT(false);
 	return NULL;
@@ -3164,11 +3185,23 @@ void dcn10_wait_for_mpcc_disconnect(
 
 	for (mpcc_inst = 0; mpcc_inst < MAX_PIPES; mpcc_inst++) {
 		if (pipe_ctx->stream_res.opp->mpcc_disconnect_pending[mpcc_inst]) {
-			struct hubp *hubp = get_hubp_by_inst(res_pool, mpcc_inst);
+			struct pipe_ctx *restore_bottom_pipe;
+			struct pipe_ctx *restore_top_pipe;
+			struct pipe_ctx *inst_pipe_ctx = get_pipe_ctx_by_hubp_inst(dc->current_state, mpcc_inst);
 
+			ASSERT(inst_pipe_ctx);
 			res_pool->mpc->funcs->wait_for_idle(res_pool->mpc, mpcc_inst);
 			pipe_ctx->stream_res.opp->mpcc_disconnect_pending[mpcc_inst] = false;
-			hubp->funcs->set_blank(hubp, true);
+			/*
+			 * Set top and bottom pipes NULL, as we don't want
+			 * to blank those pipes when disconnecting from MPCC
+			 */
+			restore_bottom_pipe = inst_pipe_ctx->bottom_pipe;
+			restore_top_pipe = inst_pipe_ctx->top_pipe;
+			inst_pipe_ctx->top_pipe = inst_pipe_ctx->bottom_pipe = NULL;
+			dc->hwss.set_hubp_blank(dc, inst_pipe_ctx, true);
+			inst_pipe_ctx->top_pipe = restore_top_pipe;
+			inst_pipe_ctx->bottom_pipe = restore_bottom_pipe;
 		}
 	}
 
@@ -3721,3 +3754,10 @@ void dcn10_get_clock(struct dc *dc,
 				dc->clk_mgr->funcs->get_clock(dc->clk_mgr, context, clock_type, clock_cfg);
 
 }
+
+void dcn10_set_hubp_blank(const struct dc *dc,
+				struct pipe_ctx *pipe_ctx,
+				bool blank_enable)
+{
+	pipe_ctx->plane_res.hubp->funcs->set_blank(pipe_ctx->plane_res.hubp, blank_enable);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h
index e5691e499023..89e6dfb63da0 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h
@@ -163,6 +163,8 @@ void dcn10_wait_for_mpcc_disconnect(
 void dce110_edp_backlight_control(
 		struct dc_link *link,
 		bool enable);
+void dce110_edp_wait_for_T12(
+		struct dc_link *link);
 void dce110_edp_power_control(
 		struct dc_link *link,
 		bool power_up);
@@ -202,5 +204,8 @@ void dcn10_wait_for_pending_cleared(struct dc *dc,
 		struct dc_state *context);
 void dcn10_set_hdr_multiplier(struct pipe_ctx *pipe_ctx);
 void dcn10_verify_allow_pstate_change_high(struct dc *dc);
+void dcn10_set_hubp_blank(const struct dc *dc,
+				struct pipe_ctx *pipe_ctx,
+				bool blank_enable);
 
 #endif /* __DC_HWSS_DCN10_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c
index 7f4766e45dff..e8b6065fffad 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c
@@ -47,7 +47,7 @@
 
 unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...)
 {
-	unsigned int ret_vsnprintf;
+	int ret_vsnprintf;
 	unsigned int chars_printed;
 
 	va_list args;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c
index 254300b06b43..2f1b802e66a1 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c
@@ -79,6 +79,7 @@ static const struct hw_sequencer_funcs dcn10_funcs = {
 	.set_backlight_level = dce110_set_backlight_level,
 	.set_abm_immediate_disable = dce110_set_abm_immediate_disable,
 	.set_pipe = dce110_set_pipe,
+	.set_hubp_blank = dcn10_set_hubp_blank,
 };
 
 static const struct hwseq_private_funcs dcn10_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
index 81db0179f7ea..e4701825b5a0 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
@@ -480,7 +480,6 @@ unsigned int dcn10_get_dig_frontend(struct link_encoder *enc)
 		break;
 	default:
 		// invalid source select DIG
-		ASSERT(false);
 		result = ENGINE_ID_UNKNOWN;
 	}
 
@@ -956,6 +955,21 @@ void dcn10_link_encoder_enable_tmds_output(
 	}
 }
 
+void dcn10_link_encoder_enable_tmds_output_with_clk_pattern_wa(
+	struct link_encoder *enc,
+	enum clock_source_id clock_source,
+	enum dc_color_depth color_depth,
+	enum signal_type signal,
+	uint32_t pixel_clock)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	dcn10_link_encoder_enable_tmds_output(
+		enc, clock_source, color_depth, signal, pixel_clock);
+
+	REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+}
+
 /* enables DP PHY output */
 void dcn10_link_encoder_enable_dp_output(
 	struct link_encoder *enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
index d4caad670855..3e1a582e4b88 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
@@ -42,6 +42,7 @@
 #define LE_DCN_COMMON_REG_LIST(id) \
 	SRI(DIG_BE_CNTL, DIG, id), \
 	SRI(DIG_BE_EN_CNTL, DIG, id), \
+	SRI(DIG_CLOCK_PATTERN, DIG, id), \
 	SRI(TMDS_CTL_BITS, DIG, id), \
 	SRI(DP_CONFIG, DP, id), \
 	SRI(DP_DPHY_CNTL, DP, id), \
@@ -83,6 +84,7 @@ struct dcn10_link_enc_hpd_registers {
 struct dcn10_link_enc_registers {
 	uint32_t DIG_BE_CNTL;
 	uint32_t DIG_BE_EN_CNTL;
+	uint32_t DIG_CLOCK_PATTERN;
 	uint32_t DP_CONFIG;
 	uint32_t DP_DPHY_CNTL;
 	uint32_t DP_DPHY_INTERNAL_CTRL;
@@ -168,6 +170,7 @@ struct dcn10_link_enc_registers {
 	LE_SF(DIG0_DIG_BE_CNTL, DIG_HPD_SELECT, mask_sh),\
 	LE_SF(DIG0_DIG_BE_CNTL, DIG_MODE, mask_sh),\
 	LE_SF(DIG0_DIG_BE_CNTL, DIG_FE_SOURCE_SELECT, mask_sh),\
+	LE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
 	LE_SF(DIG0_TMDS_CTL_BITS, TMDS_CTL0, mask_sh), \
 	LE_SF(DP0_DP_DPHY_CNTL, DPHY_BYPASS, mask_sh),\
 	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE0, mask_sh),\
@@ -218,6 +221,7 @@ struct dcn10_link_enc_registers {
 	type DIG_HPD_SELECT;\
 	type DIG_MODE;\
 	type DIG_FE_SOURCE_SELECT;\
+	type DIG_CLOCK_PATTERN;\
 	type DPHY_BYPASS;\
 	type DPHY_ATEST_SEL_LANE0;\
 	type DPHY_ATEST_SEL_LANE1;\
@@ -536,6 +540,13 @@ void dcn10_link_encoder_enable_tmds_output(
 	enum signal_type signal,
 	uint32_t pixel_clock);
 
+void dcn10_link_encoder_enable_tmds_output_with_clk_pattern_wa(
+	struct link_encoder *enc,
+	enum clock_source_id clock_source,
+	enum dc_color_depth color_depth,
+	enum signal_type signal,
+	uint32_t pixel_clock);
+
 /* enables DP PHY output */
 void dcn10_link_encoder_enable_dp_output(
 	struct link_encoder *enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c
index 100ce0e28fd5..b096011acb49 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c
@@ -470,7 +470,7 @@ void mpc1_cursor_lock(struct mpc *mpc, int opp_id, bool lock)
 unsigned int mpc1_get_mpc_out_mux(struct mpc *mpc, int opp_id)
 {
 	struct dcn10_mpc *mpc10 = TO_DCN10_MPC(mpc);
-	uint32_t val = 0;
+	uint32_t val = 0xf;
 
 	if (opp_id < MAX_OPP && REG(MUX[opp_id]))
 		REG_GET(MUX[opp_id], MPC_OUT_MUX, &val);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index f033397a84e9..6138f4887de7 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -659,6 +659,16 @@ void optc1_unlock(struct timing_generator *optc)
 			OTG_MASTER_UPDATE_LOCK, 0);
 }
 
+bool optc1_is_locked(struct timing_generator *optc)
+{
+	struct optc *optc1 = DCN10TG_FROM_TG(optc);
+	uint32_t locked;
+
+	REG_GET(OTG_MASTER_UPDATE_LOCK, UPDATE_LOCK_STATUS, &locked);
+
+	return (locked == 1);
+}
+
 void optc1_get_position(struct timing_generator *optc,
 		struct crtc_position *position)
 {
@@ -1513,6 +1523,7 @@ static const struct timing_generator_funcs dcn10_tg_funcs = {
 		.enable_crtc_reset = optc1_enable_crtc_reset,
 		.disable_reset_trigger = optc1_disable_reset_trigger,
 		.lock = optc1_lock,
+		.is_locked = optc1_is_locked,
 		.unlock = optc1_unlock,
 		.enable_optc_clock = optc1_enable_optc_clock,
 		.set_drr = optc1_set_drr,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
index b12bd9aae52f..b222c67973d4 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
@@ -638,6 +638,7 @@ void optc1_set_blank(struct timing_generator *optc,
 		bool enable_blanking);
 
 bool optc1_is_blanked(struct timing_generator *optc);
+bool optc1_is_locked(struct timing_generator *optc);
 
 void optc1_program_blank_color(
 		struct timing_generator *optc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c
index bdc37831535e..90e912fef2b3 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c
@@ -608,8 +608,8 @@ static const struct dc_debug_options debug_defaults_drv = {
 		.disable_pplib_clock_request = false,
 		.disable_pplib_wm_range = false,
 		.pplib_wm_report_mode = WM_REPORT_DEFAULT,
-		.pipe_split_policy = MPC_SPLIT_DYNAMIC,
-		.force_single_disp_pipe_split = true,
+		.pipe_split_policy = MPC_SPLIT_AVOID,
+		.force_single_disp_pipe_split = false,
 		.disable_dcc = DCC_ENABLE,
 		.voltage_align_fclk = true,
 		.disable_stereo_support = true,
@@ -1534,15 +1534,8 @@ static bool dcn10_resource_construct(
 	memcpy(dc->dcn_ip, &dcn10_ip_defaults, sizeof(dcn10_ip_defaults));
 	memcpy(dc->dcn_soc, &dcn10_soc_defaults, sizeof(dcn10_soc_defaults));
 
-#if defined(CONFIG_ARM64)
-	/* Aarch64 does not support -msoft-float/-mfloat-abi=soft */
-	DC_FP_START();
-	dcn10_resource_construct_fp(dc);
-	DC_FP_END();
-#else
 	/* Other architectures we build for build this with soft-float */
 	dcn10_resource_construct_fp(dc);
-#endif
 
 	pool->base.pp_smu = dcn10_pp_smu_create(ctx);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/Makefile b/drivers/gpu/drm/amd/display/dc/dcn20/Makefile
index 624cb1341ef1..5fcaf78334ff 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/Makefile
@@ -17,10 +17,6 @@ ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c
index 0df0da2e6a4d..bec7059f6d5d 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c
@@ -1597,6 +1597,7 @@ static struct hubp_funcs dcn20_hubp_funcs = {
 	.validate_dml_output = hubp2_validate_dml_output,
 	.hubp_in_blank = hubp1_in_blank,
 	.hubp_soft_reset = hubp1_soft_reset,
+	.hubp_set_flip_int = hubp1_set_flip_int,
 };
 
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
index cb822df21b7c..5342c309b78c 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
@@ -1062,8 +1062,13 @@ static void dcn20_power_on_plane(
 	if (REG(DC_IP_REQUEST_CNTL)) {
 		REG_SET(DC_IP_REQUEST_CNTL, 0,
 				IP_REQUEST_EN, 1);
-		dcn20_dpp_pg_control(hws, pipe_ctx->plane_res.dpp->inst, true);
-		dcn20_hubp_pg_control(hws, pipe_ctx->plane_res.hubp->inst, true);
+
+		if (hws->funcs.dpp_pg_control)
+			hws->funcs.dpp_pg_control(hws, pipe_ctx->plane_res.dpp->inst, true);
+
+		if (hws->funcs.hubp_pg_control)
+			hws->funcs.hubp_pg_control(hws, pipe_ctx->plane_res.hubp->inst, true);
+
 		REG_SET(DC_IP_REQUEST_CNTL, 0,
 				IP_REQUEST_EN, 0);
 		DC_LOG_DEBUG(
@@ -1141,6 +1146,12 @@ void dcn20_enable_plane(
 		pipe_ctx->plane_res.hubp->funcs->hubp_set_vm_system_aperture_settings(pipe_ctx->plane_res.hubp, &apt);
 	}
 
+	if (!pipe_ctx->top_pipe
+		&& pipe_ctx->plane_state
+		&& pipe_ctx->plane_state->flip_int_enabled
+		&& pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int)
+			pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int(pipe_ctx->plane_res.hubp);
+
 //	if (dc->debug.sanity_checks) {
 //		dcn10_verify_allow_pstate_change_high(dc);
 //	}
@@ -1570,8 +1581,8 @@ static void dcn20_update_dchubp_dpp(
 
 
 
-	if (pipe_ctx->update_flags.bits.enable)
-		hubp->funcs->set_blank(hubp, false);
+	if (is_pipe_tree_visible(pipe_ctx))
+		dc->hwss.set_hubp_blank(dc, pipe_ctx, false);
 }
 
 
@@ -1765,6 +1776,14 @@ void dcn20_post_unlock_program_front_end(
 		}
 	}
 
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->vtp_locked) {
+			dc->hwss.set_hubp_blank(dc, pipe, true);
+			pipe->vtp_locked = false;
+		}
+	}
 	/* WA to apply WM setting*/
 	if (hwseq->wa.DEGVIDCN21)
 		dc->res_pool->hubbub->funcs->apply_DEDCN21_147_wa(dc->res_pool->hubbub);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c
index de9dcbeea150..51a4166e9750 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c
@@ -94,6 +94,7 @@ static const struct hw_sequencer_funcs dcn20_funcs = {
 	.optimize_timing_for_fsft = dcn20_optimize_timing_for_fsft,
 #endif
 	.set_disp_pattern_generator = dcn20_set_disp_pattern_generator,
+	.set_hubp_blank = dcn10_set_hubp_blank,
 };
 
 static const struct hwseq_private_funcs dcn20_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.c
index 15c2ff264ff6..fa013496e26b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.c
@@ -363,7 +363,7 @@ static const struct link_encoder_funcs dcn20_link_enc_funcs = {
 		dcn10_link_encoder_validate_output_with_stream,
 	.hw_init = enc2_hw_init,
 	.setup = dcn10_link_encoder_setup,
-	.enable_tmds_output = dcn10_link_encoder_enable_tmds_output,
+	.enable_tmds_output = dcn10_link_encoder_enable_tmds_output_with_clk_pattern_wa,
 	.enable_dp_output = dcn20_link_encoder_enable_dp_output,
 	.enable_dp_mst_output = dcn10_link_encoder_enable_dp_mst_output,
 	.disable_output = dcn10_link_encoder_disable_output,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
index e04ecf0fc0db..2c2dbfcd8957 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
@@ -297,8 +297,8 @@ static struct _vcs_dpi_soc_bounding_box_st dcn2_0_soc = {
 			},
 		},
 	.num_states = 5,
-	.sr_exit_time_us = 11.6,
-	.sr_enter_plus_exit_time_us = 13.9,
+	.sr_exit_time_us = 8.6,
+	.sr_enter_plus_exit_time_us = 10.9,
 	.urgent_latency_us = 4.0,
 	.urgent_latency_pixel_data_only_us = 4.0,
 	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
@@ -2097,6 +2097,7 @@ int dcn20_populate_dml_pipes_from_context(
 			pipes[pipe_cnt].pipe.dest.pixel_rate_mhz *= 2;
 		pipes[pipe_cnt].pipe.dest.otg_inst = res_ctx->pipe_ctx[i].stream_res.tg->inst;
 		pipes[pipe_cnt].dout.dp_lanes = 4;
+		pipes[pipe_cnt].dout.is_virtual = 0;
 		pipes[pipe_cnt].pipe.dest.vtotal_min = res_ctx->pipe_ctx[i].stream->adjust.v_total_min;
 		pipes[pipe_cnt].pipe.dest.vtotal_max = res_ctx->pipe_ctx[i].stream->adjust.v_total_max;
 		switch (get_num_odm_splits(&res_ctx->pipe_ctx[i])) {
@@ -2150,6 +2151,7 @@ int dcn20_populate_dml_pipes_from_context(
 			break;
 		default:
 			/* In case there is no signal, set dp with 4 lanes to allow max config */
+			pipes[pipe_cnt].dout.is_virtual = 1;
 			pipes[pipe_cnt].dout.output_type = dm_dp;
 			pipes[pipe_cnt].dout.dp_lanes = 4;
 		}
@@ -2517,8 +2519,7 @@ struct pipe_ctx *dcn20_find_secondary_pipe(struct dc *dc,
 		 * if this primary pipe has a bottom pipe in prev. state
 		 * and if the bottom pipe is still available (which it should be),
 		 * pick that pipe as secondary
-		 * Same logic applies for ODM pipes. Since mpo is not allowed with odm
-		 * check in else case.
+		 * Same logic applies for ODM pipes
 		 */
 		if (dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].bottom_pipe) {
 			preferred_pipe_idx = dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].bottom_pipe->pipe_idx;
@@ -2526,7 +2527,9 @@ struct pipe_ctx *dcn20_find_secondary_pipe(struct dc *dc,
 				secondary_pipe = &res_ctx->pipe_ctx[preferred_pipe_idx];
 				secondary_pipe->pipe_idx = preferred_pipe_idx;
 			}
-		} else if (dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].next_odm_pipe) {
+		}
+		if (secondary_pipe == NULL &&
+				dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].next_odm_pipe) {
 			preferred_pipe_idx = dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].next_odm_pipe->pipe_idx;
 			if (res_ctx->pipe_ctx[preferred_pipe_idx].stream == NULL) {
 				secondary_pipe = &res_ctx->pipe_ctx[preferred_pipe_idx];
@@ -3244,7 +3247,7 @@ restore_dml_state:
 bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context,
 		bool fast_validate)
 {
-	bool voltage_supported = false;
+	bool voltage_supported;
 	DC_FP_START();
 	voltage_supported = dcn20_validate_bandwidth_fp(dc, context, fast_validate);
 	DC_FP_END();
@@ -3505,7 +3508,8 @@ void dcn20_update_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_s
 		calculated_states[i].dram_speed_mts = uclk_states[i] * 16 / 1000;
 
 		// FCLK:UCLK ratio is 1.08
-		min_fclk_required_by_uclk = mul_u64_u32_shr(BIT_ULL(32) * 1080 / 1000000, uclk_states[i], 32);
+		min_fclk_required_by_uclk = div_u64(((unsigned long long)uclk_states[i]) * 1080,
+			1000000);
 
 		calculated_states[i].fabricclk_mhz = (min_fclk_required_by_uclk < min_dcfclk) ?
 				min_dcfclk : min_fclk_required_by_uclk;
@@ -3605,7 +3609,6 @@ static enum dml_project get_dml_project_version(uint32_t hw_internal_rev)
 static bool init_soc_bounding_box(struct dc *dc,
 				  struct dcn20_resource_pool *pool)
 {
-	const struct gpu_info_soc_bounding_box_v1_0 *bb = dc->soc_bounding_box;
 	struct _vcs_dpi_soc_bounding_box_st *loaded_bb =
 			get_asic_rev_soc_bb(dc->ctx->asic_id.hw_internal_rev);
 	struct _vcs_dpi_ip_params_st *loaded_ip =
@@ -3613,116 +3616,6 @@ static bool init_soc_bounding_box(struct dc *dc,
 
 	DC_LOGGER_INIT(dc->ctx->logger);
 
-	/* TODO: upstream NV12 bounding box when its launched */
-	if (!bb && ASICREV_IS_NAVI12_P(dc->ctx->asic_id.hw_internal_rev)) {
-		DC_LOG_ERROR("%s: not valid soc bounding box/n", __func__);
-		return false;
-	}
-
-	if (bb && ASICREV_IS_NAVI12_P(dc->ctx->asic_id.hw_internal_rev)) {
-		int i;
-
-		dcn2_0_nv12_soc.sr_exit_time_us =
-				fixed16_to_double_to_cpu(bb->sr_exit_time_us);
-		dcn2_0_nv12_soc.sr_enter_plus_exit_time_us =
-				fixed16_to_double_to_cpu(bb->sr_enter_plus_exit_time_us);
-		dcn2_0_nv12_soc.urgent_latency_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_us);
-		dcn2_0_nv12_soc.urgent_latency_pixel_data_only_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_pixel_data_only_us);
-		dcn2_0_nv12_soc.urgent_latency_pixel_mixed_with_vm_data_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_pixel_mixed_with_vm_data_us);
-		dcn2_0_nv12_soc.urgent_latency_vm_data_only_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_vm_data_only_us);
-		dcn2_0_nv12_soc.urgent_out_of_order_return_per_channel_pixel_only_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_pixel_only_bytes);
-		dcn2_0_nv12_soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes);
-		dcn2_0_nv12_soc.urgent_out_of_order_return_per_channel_vm_only_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_vm_only_bytes);
-		dcn2_0_nv12_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_pixel_only);
-		dcn2_0_nv12_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm);
-		dcn2_0_nv12_soc.pct_ideal_dram_sdp_bw_after_urgent_vm_only =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_vm_only);
-		dcn2_0_nv12_soc.max_avg_sdp_bw_use_normal_percent =
-				fixed16_to_double_to_cpu(bb->max_avg_sdp_bw_use_normal_percent);
-		dcn2_0_nv12_soc.max_avg_dram_bw_use_normal_percent =
-				fixed16_to_double_to_cpu(bb->max_avg_dram_bw_use_normal_percent);
-		dcn2_0_nv12_soc.writeback_latency_us =
-				fixed16_to_double_to_cpu(bb->writeback_latency_us);
-		dcn2_0_nv12_soc.ideal_dram_bw_after_urgent_percent =
-				fixed16_to_double_to_cpu(bb->ideal_dram_bw_after_urgent_percent);
-		dcn2_0_nv12_soc.max_request_size_bytes =
-				le32_to_cpu(bb->max_request_size_bytes);
-		dcn2_0_nv12_soc.dram_channel_width_bytes =
-				le32_to_cpu(bb->dram_channel_width_bytes);
-		dcn2_0_nv12_soc.fabric_datapath_to_dcn_data_return_bytes =
-				le32_to_cpu(bb->fabric_datapath_to_dcn_data_return_bytes);
-		dcn2_0_nv12_soc.dcn_downspread_percent =
-				fixed16_to_double_to_cpu(bb->dcn_downspread_percent);
-		dcn2_0_nv12_soc.downspread_percent =
-				fixed16_to_double_to_cpu(bb->downspread_percent);
-		dcn2_0_nv12_soc.dram_page_open_time_ns =
-				fixed16_to_double_to_cpu(bb->dram_page_open_time_ns);
-		dcn2_0_nv12_soc.dram_rw_turnaround_time_ns =
-				fixed16_to_double_to_cpu(bb->dram_rw_turnaround_time_ns);
-		dcn2_0_nv12_soc.dram_return_buffer_per_channel_bytes =
-				le32_to_cpu(bb->dram_return_buffer_per_channel_bytes);
-		dcn2_0_nv12_soc.round_trip_ping_latency_dcfclk_cycles =
-				le32_to_cpu(bb->round_trip_ping_latency_dcfclk_cycles);
-		dcn2_0_nv12_soc.urgent_out_of_order_return_per_channel_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_bytes);
-		dcn2_0_nv12_soc.channel_interleave_bytes =
-				le32_to_cpu(bb->channel_interleave_bytes);
-		dcn2_0_nv12_soc.num_banks =
-				le32_to_cpu(bb->num_banks);
-		dcn2_0_nv12_soc.num_chans =
-				le32_to_cpu(bb->num_chans);
-		dcn2_0_nv12_soc.vmm_page_size_bytes =
-				le32_to_cpu(bb->vmm_page_size_bytes);
-		dcn2_0_nv12_soc.dram_clock_change_latency_us =
-				fixed16_to_double_to_cpu(bb->dram_clock_change_latency_us);
-		// HACK!! Lower uclock latency switch time so we don't switch
-		dcn2_0_nv12_soc.dram_clock_change_latency_us = 10;
-		dcn2_0_nv12_soc.writeback_dram_clock_change_latency_us =
-				fixed16_to_double_to_cpu(bb->writeback_dram_clock_change_latency_us);
-		dcn2_0_nv12_soc.return_bus_width_bytes =
-				le32_to_cpu(bb->return_bus_width_bytes);
-		dcn2_0_nv12_soc.dispclk_dppclk_vco_speed_mhz =
-				le32_to_cpu(bb->dispclk_dppclk_vco_speed_mhz);
-		dcn2_0_nv12_soc.xfc_bus_transport_time_us =
-				le32_to_cpu(bb->xfc_bus_transport_time_us);
-		dcn2_0_nv12_soc.xfc_xbuf_latency_tolerance_us =
-				le32_to_cpu(bb->xfc_xbuf_latency_tolerance_us);
-		dcn2_0_nv12_soc.use_urgent_burst_bw =
-				le32_to_cpu(bb->use_urgent_burst_bw);
-		dcn2_0_nv12_soc.num_states =
-				le32_to_cpu(bb->num_states);
-
-		for (i = 0; i < dcn2_0_nv12_soc.num_states; i++) {
-			dcn2_0_nv12_soc.clock_limits[i].state =
-					le32_to_cpu(bb->clock_limits[i].state);
-			dcn2_0_nv12_soc.clock_limits[i].dcfclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dcfclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].fabricclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].fabricclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].dispclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dispclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].dppclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dppclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].phyclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].phyclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].socclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].socclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].dscclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dscclk_mhz);
-			dcn2_0_nv12_soc.clock_limits[i].dram_speed_mts =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dram_speed_mts);
-		}
-	}
-
 	if (pool->base.pp_smu) {
 		struct pp_smu_nv_clock_table max_clocks = {0};
 		unsigned int uclk_states[8] = {0};
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/Makefile b/drivers/gpu/drm/amd/display/dc/dcn21/Makefile
index 1ee5fc03b7b3..bb8c95141082 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/Makefile
@@ -13,10 +13,6 @@ ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c
index f9045852728f..b0c9180b808f 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c
@@ -838,6 +838,7 @@ static struct hubp_funcs dcn21_hubp_funcs = {
 	.hubp_set_flip_control_surface_gsl = hubp2_set_flip_control_surface_gsl,
 	.hubp_init = hubp21_init,
 	.validate_dml_output = hubp21_validate_dml_output,
+	.hubp_set_flip_int = hubp1_set_flip_int,
 };
 
 bool hubp21_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c
index 96ee0b82f458..d3b643089603 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c
@@ -123,7 +123,7 @@ void dcn21_optimize_pwr_state(
  * PHY will hang on the next mode set attempt.
  * if enable PLL follow by disable PLL (without executing lane enable/disable),
  * RDPCS_PHY_DP_MPLLB_STATE remains 1,
- * which indicate that PLL disable attempt actually didn’t go through.
+ * which indicate that PLL disable attempt actually didn't go through.
  * As a workaround, insert PHY lane enable/disable before PLL disable.
  */
 void dcn21_PLAT_58856_wa(struct dc_state *context, struct pipe_ctx *pipe_ctx)
@@ -143,6 +143,7 @@ static bool dmub_abm_set_pipe(struct abm *abm, uint32_t otg_inst, uint32_t optio
 	struct dc_context *dc = abm->ctx;
 	uint32_t ramping_boundary = 0xFFFF;
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.abm_set_pipe.header.type = DMUB_CMD__ABM;
 	cmd.abm_set_pipe.header.sub_type = DMUB_CMD__ABM_SET_PIPE;
 	cmd.abm_set_pipe.abm_set_pipe_data.otg_inst = otg_inst;
@@ -212,6 +213,7 @@ bool dcn21_set_backlight_level(struct pipe_ctx *pipe_ctx,
 	if (abm && panel_cntl)
 		dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst);
 
+	memset(&cmd, 0, sizeof(cmd));
 	cmd.abm_set_backlight.header.type = DMUB_CMD__ABM;
 	cmd.abm_set_backlight.header.sub_type = DMUB_CMD__ABM_SET_BACKLIGHT;
 	cmd.abm_set_backlight.abm_set_backlight_data.frame_ramp = frame_ramp;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c
index 074e2713257f..0597391b2171 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c
@@ -99,6 +99,7 @@ static const struct hw_sequencer_funcs dcn21_funcs = {
 #endif
 	.is_abm_supported = dcn21_is_abm_supported,
 	.set_disp_pattern_generator = dcn20_set_disp_pattern_generator,
+	.set_hubp_blank = dcn10_set_hubp_blank,
 };
 
 static const struct hwseq_private_funcs dcn21_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
index 1c88d2edd381..173488ab787a 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
@@ -906,6 +906,8 @@ enum dcn20_clk_src_array_id {
 	DCN20_CLK_SRC_PLL0,
 	DCN20_CLK_SRC_PLL1,
 	DCN20_CLK_SRC_PLL2,
+	DCN20_CLK_SRC_PLL3,
+	DCN20_CLK_SRC_PLL4,
 	DCN20_CLK_SRC_TOTAL_DCN21
 };
 
@@ -1060,8 +1062,6 @@ static void patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_s
 {
 	int i;
 
-	DC_FP_START();
-
 	if (dc->bb_overrides.sr_exit_time_ns) {
 		for (i = 0; i < WM_SET_COUNT; i++) {
 			  dc->clk_mgr->bw_params->wm_table.entries[i].sr_exit_time_us =
@@ -1086,8 +1086,6 @@ static void patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_s
 				dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
 		}
 	}
-
-	DC_FP_END();
 }
 
 void dcn21_calculate_wm(
@@ -1327,8 +1325,8 @@ validate_out:
 	return out;
 }
 
-bool dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context,
-		bool fast_validate)
+static noinline bool dcn21_validate_bandwidth_fp(struct dc *dc,
+		struct dc_state *context, bool fast_validate)
 {
 	bool out = false;
 
@@ -1337,7 +1335,7 @@ bool dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context,
 	int vlevel = 0;
 	int pipe_split_from[MAX_PIPES];
 	int pipe_cnt = 0;
-	display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_KERNEL);
+	display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_ATOMIC);
 	DC_LOGGER_INIT(dc->ctx->logger);
 
 	BW_VAL_TRACE_COUNT();
@@ -1381,6 +1379,22 @@ validate_out:
 
 	return out;
 }
+
+/*
+ * Some of the functions further below use the FPU, so we need to wrap this
+ * with DC_FP_START()/DC_FP_END(). Use the same approach as for
+ * dcn20_validate_bandwidth in dcn20_resource.c.
+ */
+bool dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context,
+		bool fast_validate)
+{
+	bool voltage_supported;
+	DC_FP_START();
+	voltage_supported = dcn21_validate_bandwidth_fp(dc, context, fast_validate);
+	DC_FP_END();
+	return voltage_supported;
+}
+
 static void dcn21_destroy_resource_pool(struct resource_pool **pool)
 {
 	struct dcn21_resource_pool *dcn21_pool = TO_DCN21_RES_POOL(*pool);
@@ -2030,6 +2044,14 @@ static bool dcn21_resource_construct(
 			dcn21_clock_source_create(ctx, ctx->dc_bios,
 				CLOCK_SOURCE_COMBO_PHY_PLL2,
 				&clk_src_regs[2], false);
+	pool->base.clock_sources[DCN20_CLK_SRC_PLL3] =
+			dcn21_clock_source_create(ctx, ctx->dc_bios,
+				CLOCK_SOURCE_COMBO_PHY_PLL3,
+				&clk_src_regs[3], false);
+	pool->base.clock_sources[DCN20_CLK_SRC_PLL4] =
+			dcn21_clock_source_create(ctx, ctx->dc_bios,
+				CLOCK_SOURCE_COMBO_PHY_PLL4,
+				&clk_src_regs[4], false);
 
 	pool->base.clk_src_count = DCN20_CLK_SRC_TOTAL_DCN21;
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/Makefile b/drivers/gpu/drm/amd/display/dc/dcn30/Makefile
index 248c2711aace..dfd77b3cc84d 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/Makefile
@@ -32,8 +32,8 @@ DCN30 = dcn30_init.o dcn30_hubbub.o dcn30_hubp.o dcn30_dpp.o dcn30_optc.o \
 
 
 ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o := -mhard-float -msse
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o := -mhard-float -msse
+CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o := -msse
+CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o := -msse
 endif
 
 ifdef CONFIG_PPC64
@@ -41,15 +41,12 @@ CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o := -mhard-float -maltivec
 CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o := -mgeneral-regs-only
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
 endif
+CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o += -mhard-float
+CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o += -mhard-float
 endif
 
 ifdef CONFIG_X86
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp_cm.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp_cm.c
index 9da66e491116..33985401f25c 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp_cm.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp_cm.c
@@ -133,7 +133,6 @@ static void dpp3_power_on_gamcor_lut(
 		struct dpp *dpp_base,
 	bool power_on)
 {
-	uint32_t power_status;
 	struct dcn3_dpp *dpp = TO_DCN30_DPP(dpp_base);
 
 	if (dpp_base->ctx->dc->debug.enable_mem_low_power.bits.cm) {
@@ -143,12 +142,6 @@ static void dpp3_power_on_gamcor_lut(
 	} else
 		REG_SET(CM_MEM_PWR_CTRL, 0,
 				GAMCOR_MEM_PWR_DIS, power_on == true ? 0:1);
-
-	REG_GET(CM_MEM_PWR_STATUS, GAMCOR_MEM_PWR_STATE, &power_status);
-	if (power_status != 0)
-		BREAK_TO_DEBUGGER();
-
-
 }
 
 void dpp3_program_cm_dealpha(
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c
index 88ffa9ff1ed1..f24612523248 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c
@@ -511,6 +511,7 @@ static struct hubp_funcs dcn30_hubp_funcs = {
 	.hubp_init = hubp3_init,
 	.hubp_in_blank = hubp1_in_blank,
 	.hubp_soft_reset = hubp1_soft_reset,
+	.hubp_set_flip_int = hubp1_set_flip_int,
 };
 
 bool hubp3_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.h b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.h
index 5fa150f34c60..705fbfc37502 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.h
@@ -62,6 +62,7 @@
 	HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_NO_OUTSTANDING_REQ, mask_sh),\
 	HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_VTG_SEL, mask_sh),\
 	HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_DISABLE, mask_sh),\
+	HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_IN_BLANK, mask_sh),\
 	HUBP_SF(HUBP0_DCSURF_ADDR_CONFIG, NUM_PIPES, mask_sh),\
 	HUBP_SF(HUBP0_DCSURF_ADDR_CONFIG, PIPE_INTERLEAVE, mask_sh),\
 	HUBP_SF(HUBP0_DCSURF_ADDR_CONFIG, MAX_COMPRESSED_FRAGS, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c
index 3deb3fb1724d..06dc1e2e8383 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c
@@ -539,6 +539,8 @@ void dcn30_init_hw(struct dc *dc)
 
 					fe = dc->links[i]->link_enc->funcs->get_dig_frontend(
 										dc->links[i]->link_enc);
+					if (fe == ENGINE_ID_UNKNOWN)
+						continue;
 
 					for (j = 0; j < dc->res_pool->stream_enc_count; j++) {
 						if (fe == dc->res_pool->stream_enc[j]->id) {
@@ -710,8 +712,11 @@ void dcn30_program_dmdata_engine(struct pipe_ctx *pipe_ctx)
 bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable)
 {
 	union dmub_rb_cmd cmd;
-	unsigned int surface_size, refresh_hz, denom;
 	uint32_t tmr_delay = 0, tmr_scale = 0;
+	struct dc_cursor_attributes cursor_attr;
+	bool cursor_cache_enable = false;
+	struct dc_stream_state *stream = NULL;
+	struct dc_plane_state *plane = NULL;
 
 	if (!dc->ctx->dmub_srv)
 		return false;
@@ -722,72 +727,150 @@ bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable)
 
 			/* First, check no-memory-requests case */
 			for (i = 0; i < dc->current_state->stream_count; i++) {
-				if (dc->current_state->stream_status[i]
-					    .plane_count)
+				if (dc->current_state->stream_status[i].plane_count)
 					/* Fail eligibility on a visible stream */
 					break;
 			}
 
-			if (dc->current_state->stream_count == 1 // single display only
-			    && dc->current_state->stream_status[0].plane_count == 1 // single surface only
-			    && dc->current_state->stream_status[0].plane_states[0]->address.page_table_base.quad_part == 0 // no VM
-			    // Only 8 and 16 bit formats
-			    && dc->current_state->stream_status[0].plane_states[0]->format <= SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F
-			    && dc->current_state->stream_status[0].plane_states[0]->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB8888) {
-				surface_size = dc->current_state->stream_status[0].plane_states[0]->plane_size.surface_pitch *
-					dc->current_state->stream_status[0].plane_states[0]->plane_size.surface_size.height *
-					(dc->current_state->stream_status[0].plane_states[0]->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616 ?
-					 8 : 4);
-			} else {
-				// TODO: remove hard code size
-				surface_size = 128 * 1024 * 1024;
+			if (i == dc->current_state->stream_count) {
+				/* Enable no-memory-requests case */
+				memset(&cmd, 0, sizeof(cmd));
+				cmd.mall.header.type = DMUB_CMD__MALL;
+				cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_NO_DF_REQ;
+				cmd.mall.header.payload_bytes = sizeof(cmd.mall) - sizeof(cmd.mall.header);
+
+				dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
+				dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
+
+				return true;
+			}
+
+			stream = dc->current_state->streams[0];
+			plane = (stream ? dc->current_state->stream_status[0].plane_states[0] : NULL);
+
+			if (stream && plane) {
+				cursor_cache_enable = stream->cursor_position.enable &&
+						plane->address.grph.cursor_cache_addr.quad_part;
+				cursor_attr = stream->cursor_attributes;
 			}
 
-			// TODO: remove hard code size
-			if (surface_size < 128 * 1024 * 1024) {
-				refresh_hz = div_u64((unsigned long long) dc->current_state->streams[0]->timing.pix_clk_100hz *
-						     100LL,
-						     (dc->current_state->streams[0]->timing.v_total *
-						      dc->current_state->streams[0]->timing.h_total));
+			/*
+			 * Second, check MALL eligibility
+			 *
+			 * single display only, single surface only, 8 and 16 bit formats only, no VM,
+			 * do not use MALL for displays that support PSR as they use D0i3.2 in DMCUB FW
+			 *
+			 * TODO: When we implement multi-display, PSR displays will be allowed if there is
+			 * a non-PSR display present, since in that case we can't do D0i3.2
+			 */
+			if (dc->current_state->stream_count == 1 &&
+					stream->link->psr_settings.psr_version == DC_PSR_VERSION_UNSUPPORTED &&
+					dc->current_state->stream_status[0].plane_count == 1 &&
+					plane->format <= SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F &&
+					plane->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB8888 &&
+					plane->address.page_table_base.quad_part == 0 &&
+					dc->hwss.does_plane_fit_in_mall &&
+					dc->hwss.does_plane_fit_in_mall(dc, plane,
+							cursor_cache_enable ? &cursor_attr : NULL)) {
+				unsigned int v_total = stream->adjust.v_total_max ?
+						stream->adjust.v_total_max : stream->timing.v_total;
+				unsigned int refresh_hz = div_u64((unsigned long long) stream->timing.pix_clk_100hz *
+						100LL, (v_total * stream->timing.h_total));
 
 				/*
-				 * Delay_Us = 65.28 * (64 + MallFrameCacheTmrDly) * 2^MallFrameCacheTmrScale
-				 * Delay_Us / 65.28 = (64 + MallFrameCacheTmrDly) * 2^MallFrameCacheTmrScale
-				 * (Delay_Us / 65.28) / 2^MallFrameCacheTmrScale = 64 + MallFrameCacheTmrDly
-				 * MallFrameCacheTmrDly = ((Delay_Us / 65.28) / 2^MallFrameCacheTmrScale) - 64
-				 *                      = (1000000 / refresh) / 65.28 / 2^MallFrameCacheTmrScale - 64
-				 *                      = 1000000 / (refresh * 65.28 * 2^MallFrameCacheTmrScale) - 64
-				 *                      = (1000000 * 100) / (refresh * 6528 * 2^MallFrameCacheTmrScale) - 64
+				 * one frame time in microsec:
+				 * Delay_Us = 1000000 / refresh
+				 * dynamic_delay_us = 1000000 / refresh + 2 * stutter_period
+				 *
+				 * one frame time modified by 'additional timer percent' (p):
+				 * Delay_Us_modified = dynamic_delay_us + dynamic_delay_us * p / 100
+				 *                   = dynamic_delay_us * (1 + p / 100)
+				 *                   = (1000000 / refresh + 2 * stutter_period) * (100 + p) / 100
+				 *                   = (1000000 + 2 * stutter_period * refresh) * (100 + p) / (100 * refresh)
+				 *
+				 * formula for timer duration based on parameters, from regspec:
+				 * dynamic_delay_us = 65.28 * (64 + MallFrameCacheTmrDly) * 2^MallFrameCacheTmrScale
+				 *
+				 * dynamic_delay_us / 65.28 = (64 + MallFrameCacheTmrDly) * 2^MallFrameCacheTmrScale
+				 * (dynamic_delay_us / 65.28) / 2^MallFrameCacheTmrScale = 64 + MallFrameCacheTmrDly
+				 * MallFrameCacheTmrDly = ((dynamic_delay_us / 65.28) / 2^MallFrameCacheTmrScale) - 64
+				 *                      = (1000000 + 2 * stutter_period * refresh) * (100 + p) / (100 * refresh) / 65.28 / 2^MallFrameCacheTmrScale - 64
+				 *                      = (1000000 + 2 * stutter_period * refresh) * (100 + p) / (refresh * 6528 * 2^MallFrameCacheTmrScale) - 64
 				 *
 				 * need to round up the result of the division before the subtraction
 				 */
-				denom = refresh_hz * 6528;
-				tmr_delay = div_u64((100000000LL + denom - 1), denom) - 64LL;
+				unsigned int denom = refresh_hz * 6528;
+				unsigned int stutter_period = dc->current_state->perf_params.stutter_period_us;
+
+				tmr_delay = div_u64(((1000000LL + 2 * stutter_period * refresh_hz) *
+						(100LL + dc->debug.mall_additional_timer_percent) + denom - 1),
+						denom) - 64LL;
 
 				/* scale should be increased until it fits into 6 bits */
 				while (tmr_delay & ~0x3F) {
 					tmr_scale++;
 
 					if (tmr_scale > 3) {
-						/* The delay exceeds the range of the hystersis timer */
+						/* Delay exceeds range of hysteresis timer */
 						ASSERT(false);
 						return false;
 					}
 
 					denom *= 2;
-					tmr_delay = div_u64((100000000LL + denom - 1), denom) - 64LL;
+					tmr_delay = div_u64(((1000000LL + 2 * stutter_period * refresh_hz) *
+							(100LL + dc->debug.mall_additional_timer_percent) + denom - 1),
+							denom) - 64LL;
+				}
+
+				/* Copy HW cursor */
+				if (cursor_cache_enable) {
+					memset(&cmd, 0, sizeof(cmd));
+					cmd.mall.header.type = DMUB_CMD__MALL;
+					cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_COPY_CURSOR;
+					cmd.mall.header.payload_bytes =
+							sizeof(cmd.mall) - sizeof(cmd.mall.header);
+
+					switch (cursor_attr.color_format) {
+					case CURSOR_MODE_MONO:
+						cmd.mall.cursor_bpp = 2;
+						break;
+					case CURSOR_MODE_COLOR_1BIT_AND:
+					case CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA:
+					case CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA:
+						cmd.mall.cursor_bpp = 32;
+						break;
+
+					case CURSOR_MODE_COLOR_64BIT_FP_PRE_MULTIPLIED:
+					case CURSOR_MODE_COLOR_64BIT_FP_UN_PRE_MULTIPLIED:
+						cmd.mall.cursor_bpp = 64;
+						break;
+					}
+
+					cmd.mall.cursor_copy_src.quad_part = cursor_attr.address.quad_part;
+					cmd.mall.cursor_copy_dst.quad_part =
+							plane->address.grph.cursor_cache_addr.quad_part;
+					cmd.mall.cursor_width = cursor_attr.width;
+					cmd.mall.cursor_height = cursor_attr.height;
+					cmd.mall.cursor_pitch = cursor_attr.pitch;
+
+					dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
+					dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
+					dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
+
+					/* Use copied cursor, and it's okay to not switch back */
+					cursor_attr.address.quad_part =
+							plane->address.grph.cursor_cache_addr.quad_part;
+					dc_stream_set_cursor_attributes(stream, &cursor_attr);
 				}
 
 				/* Enable MALL */
 				memset(&cmd, 0, sizeof(cmd));
 				cmd.mall.header.type = DMUB_CMD__MALL;
-				cmd.mall.header.sub_type =
-					DMUB_CMD__MALL_ACTION_ALLOW;
-				cmd.mall.header.payload_bytes =
-					sizeof(cmd.mall) -
-					sizeof(cmd.mall.header);
+				cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_ALLOW;
+				cmd.mall.header.payload_bytes = sizeof(cmd.mall) - sizeof(cmd.mall.header);
 				cmd.mall.tmr_delay = tmr_delay;
 				cmd.mall.tmr_scale = tmr_scale;
+				cmd.mall.debug_bits = dc->debug.mall_error_as_fatal;
 
 				dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
 				dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
@@ -814,6 +897,40 @@ bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable)
 	return true;
 }
 
+bool dcn30_does_plane_fit_in_mall(struct dc *dc, struct dc_plane_state *plane, struct dc_cursor_attributes *cursor_attr)
+{
+	// add meta size?
+	unsigned int surface_size = plane->plane_size.surface_pitch * plane->plane_size.surface_size.height *
+			(plane->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616 ? 8 : 4);
+	unsigned int mall_size = dc->caps.mall_size_total;
+	unsigned int cursor_size = 0;
+
+	if (dc->debug.mall_size_override)
+		mall_size = 1024 * 1024 * dc->debug.mall_size_override;
+
+	if (cursor_attr) {
+		cursor_size = dc->caps.max_cursor_size * dc->caps.max_cursor_size;
+
+		switch (cursor_attr->color_format) {
+		case CURSOR_MODE_MONO:
+			cursor_size /= 2;
+			break;
+		case CURSOR_MODE_COLOR_1BIT_AND:
+		case CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA:
+		case CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA:
+			cursor_size *= 4;
+			break;
+
+		case CURSOR_MODE_COLOR_64BIT_FP_PRE_MULTIPLIED:
+		case CURSOR_MODE_COLOR_64BIT_FP_UN_PRE_MULTIPLIED:
+			cursor_size *= 8;
+			break;
+		}
+	}
+
+	return (surface_size + cursor_size) < mall_size;
+}
+
 void dcn30_hardware_release(struct dc *dc)
 {
 	/* if pstate unsupported, force it supported */
@@ -823,6 +940,53 @@ void dcn30_hardware_release(struct dc *dc)
 				dc->res_pool->hubbub, true, true);
 }
 
+void dcn30_set_hubp_blank(const struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		bool blank_enable)
+{
+	struct pipe_ctx *mpcc_pipe;
+	struct pipe_ctx *odm_pipe;
+
+	if (blank_enable) {
+		struct plane_resource *plane_res = &pipe_ctx->plane_res;
+		struct stream_resource *stream_res = &pipe_ctx->stream_res;
+
+		/* Wait for enter vblank */
+		stream_res->tg->funcs->wait_for_state(stream_res->tg, CRTC_STATE_VBLANK);
+
+		/* Blank HUBP to allow p-state during blank on all timings */
+		pipe_ctx->plane_res.hubp->funcs->set_blank(pipe_ctx->plane_res.hubp, true);
+		/* Confirm hubp in blank */
+		ASSERT(plane_res->hubp->funcs->hubp_in_blank(plane_res->hubp));
+		/* Toggle HUBP_DISABLE */
+		plane_res->hubp->funcs->hubp_soft_reset(plane_res->hubp, true);
+		plane_res->hubp->funcs->hubp_soft_reset(plane_res->hubp, false);
+		for (mpcc_pipe = pipe_ctx->bottom_pipe; mpcc_pipe; mpcc_pipe = mpcc_pipe->bottom_pipe) {
+			mpcc_pipe->plane_res.hubp->funcs->set_blank(mpcc_pipe->plane_res.hubp, true);
+			/* Confirm hubp in blank */
+			ASSERT(mpcc_pipe->plane_res.hubp->funcs->hubp_in_blank(mpcc_pipe->plane_res.hubp));
+			/* Toggle HUBP_DISABLE */
+			mpcc_pipe->plane_res.hubp->funcs->hubp_soft_reset(mpcc_pipe->plane_res.hubp, true);
+			mpcc_pipe->plane_res.hubp->funcs->hubp_soft_reset(mpcc_pipe->plane_res.hubp, false);
+
+		}
+		for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
+			odm_pipe->plane_res.hubp->funcs->set_blank(odm_pipe->plane_res.hubp, true);
+			/* Confirm hubp in blank */
+			ASSERT(odm_pipe->plane_res.hubp->funcs->hubp_in_blank(odm_pipe->plane_res.hubp));
+			/* Toggle HUBP_DISABLE */
+			odm_pipe->plane_res.hubp->funcs->hubp_soft_reset(odm_pipe->plane_res.hubp, true);
+			odm_pipe->plane_res.hubp->funcs->hubp_soft_reset(odm_pipe->plane_res.hubp, false);
+		}
+	} else {
+		pipe_ctx->plane_res.hubp->funcs->set_blank(pipe_ctx->plane_res.hubp, false);
+		for (mpcc_pipe = pipe_ctx->bottom_pipe; mpcc_pipe; mpcc_pipe = mpcc_pipe->bottom_pipe)
+			mpcc_pipe->plane_res.hubp->funcs->set_blank(mpcc_pipe->plane_res.hubp, false);
+		for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
+			odm_pipe->plane_res.hubp->funcs->set_blank(odm_pipe->plane_res.hubp, false);
+	}
+}
+
 void dcn30_set_disp_pattern_generator(const struct dc *dc,
 		struct pipe_ctx *pipe_ctx,
 		enum controller_dp_test_pattern test_pattern,
@@ -831,6 +995,25 @@ void dcn30_set_disp_pattern_generator(const struct dc *dc,
 		const struct tg_color *solid_color,
 		int width, int height, int offset)
 {
-	pipe_ctx->stream_res.opp->funcs->opp_set_disp_pattern_generator(pipe_ctx->stream_res.opp, test_pattern,
-			color_space, color_depth, solid_color, width, height, offset);
+	struct stream_resource *stream_res = &pipe_ctx->stream_res;
+
+	if (test_pattern != CONTROLLER_DP_TEST_PATTERN_VIDEOMODE) {
+		pipe_ctx->vtp_locked = false;
+		/* turning on DPG */
+		stream_res->opp->funcs->opp_set_disp_pattern_generator(stream_res->opp, test_pattern, color_space,
+				color_depth, solid_color, width, height, offset);
+
+		/* Defer hubp blank if tg is locked */
+		if (stream_res->tg->funcs->is_tg_enabled(stream_res->tg)) {
+			if (stream_res->tg->funcs->is_locked(stream_res->tg))
+				pipe_ctx->vtp_locked = true;
+			else
+				dc->hwss.set_hubp_blank(dc, pipe_ctx, true);
+		}
+	} else {
+		dc->hwss.set_hubp_blank(dc, pipe_ctx, false);
+		/* turning off DPG */
+		stream_res->opp->funcs->opp_set_disp_pattern_generator(stream_res->opp, test_pattern, color_space,
+				color_depth, solid_color, width, height, offset);
+	}
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.h
index 7d32c43aafe0..3b7d4812e311 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.h
@@ -65,6 +65,9 @@ void dcn30_set_avmute(struct pipe_ctx *pipe_ctx, bool enable);
 void dcn30_update_info_frame(struct pipe_ctx *pipe_ctx);
 void dcn30_program_dmdata_engine(struct pipe_ctx *pipe_ctx);
 
+bool dcn30_does_plane_fit_in_mall(struct dc *dc, struct dc_plane_state *plane,
+		struct dc_cursor_attributes *cursor_attr);
+
 bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable);
 
 void dcn30_hardware_release(struct dc *dc);
@@ -77,4 +80,8 @@ void dcn30_set_disp_pattern_generator(const struct dc *dc,
 		const struct tg_color *solid_color,
 		int width, int height, int offset);
 
+void dcn30_set_hubp_blank(const struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		bool blank_enable);
+
 #endif /* __DC_HWSS_DCN30_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c
index 6125fe440ad0..204444fead97 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c
@@ -71,6 +71,7 @@ static const struct hw_sequencer_funcs dcn30_funcs = {
 	.edp_backlight_control = dce110_edp_backlight_control,
 	.edp_power_control = dce110_edp_power_control,
 	.edp_wait_for_hpd_ready = dce110_edp_wait_for_hpd_ready,
+	.edp_wait_for_T12 = dce110_edp_wait_for_T12,
 	.set_cursor_position = dcn10_set_cursor_position,
 	.set_cursor_attribute = dcn10_set_cursor_attribute,
 	.set_cursor_sdr_white_level = dcn10_set_cursor_sdr_white_level,
@@ -91,11 +92,13 @@ static const struct hw_sequencer_funcs dcn30_funcs = {
 	.get_vupdate_offset_from_vsync = dcn10_get_vupdate_offset_from_vsync,
 	.calc_vupdate_position = dcn10_calc_vupdate_position,
 	.apply_idle_power_optimizations = dcn30_apply_idle_power_optimizations,
+	.does_plane_fit_in_mall = dcn30_does_plane_fit_in_mall,
 	.set_backlight_level = dcn21_set_backlight_level,
 	.set_abm_immediate_disable = dcn21_set_abm_immediate_disable,
 	.hardware_release = dcn30_hardware_release,
 	.set_pipe = dcn21_set_pipe,
 	.set_disp_pattern_generator = dcn30_set_disp_pattern_generator,
+	.set_hubp_blank = dcn30_set_hubp_blank,
 };
 
 static const struct hwseq_private_funcs dcn30_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c
index 3ba3991ee612..8980c90b2277 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c
@@ -309,6 +309,7 @@ static struct timing_generator_funcs dcn30_tg_funcs = {
 		.enable_crtc_reset = optc1_enable_crtc_reset,
 		.disable_reset_trigger = optc1_disable_reset_trigger,
 		.lock = optc3_lock,
+		.is_locked = optc1_is_locked,
 		.unlock = optc1_unlock,
 		.lock_doublebuffer_enable = optc3_lock_doublebuffer_enable,
 		.lock_doublebuffer_disable = optc3_lock_doublebuffer_disable,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
index 5e126fdf6ec1..fb7f1dea3c46 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
@@ -1716,125 +1716,22 @@ static bool is_soc_bounding_box_valid(struct dc *dc)
 static bool init_soc_bounding_box(struct dc *dc,
 				  struct dcn30_resource_pool *pool)
 {
-	const struct gpu_info_soc_bounding_box_v1_0 *bb = dc->soc_bounding_box;
 	struct _vcs_dpi_soc_bounding_box_st *loaded_bb = &dcn3_0_soc;
 	struct _vcs_dpi_ip_params_st *loaded_ip = &dcn3_0_ip;
 
 	DC_LOGGER_INIT(dc->ctx->logger);
 
-	if (!bb && !is_soc_bounding_box_valid(dc)) {
+	if (!is_soc_bounding_box_valid(dc)) {
 		DC_LOG_ERROR("%s: not valid soc bounding box/n", __func__);
 		return false;
 	}
 
-	if (bb && !is_soc_bounding_box_valid(dc)) {
-		int i;
-
-		dcn3_0_soc.sr_exit_time_us =
-				fixed16_to_double_to_cpu(bb->sr_exit_time_us);
-		dcn3_0_soc.sr_enter_plus_exit_time_us =
-				fixed16_to_double_to_cpu(bb->sr_enter_plus_exit_time_us);
-		dcn3_0_soc.urgent_latency_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_us);
-		dcn3_0_soc.urgent_latency_pixel_data_only_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_pixel_data_only_us);
-		dcn3_0_soc.urgent_latency_pixel_mixed_with_vm_data_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_pixel_mixed_with_vm_data_us);
-		dcn3_0_soc.urgent_latency_vm_data_only_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_vm_data_only_us);
-		dcn3_0_soc.urgent_out_of_order_return_per_channel_pixel_only_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_pixel_only_bytes);
-		dcn3_0_soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes);
-		dcn3_0_soc.urgent_out_of_order_return_per_channel_vm_only_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_vm_only_bytes);
-		dcn3_0_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_pixel_only);
-		dcn3_0_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm);
-		dcn3_0_soc.pct_ideal_dram_sdp_bw_after_urgent_vm_only =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_vm_only);
-		dcn3_0_soc.max_avg_sdp_bw_use_normal_percent =
-				fixed16_to_double_to_cpu(bb->max_avg_sdp_bw_use_normal_percent);
-		dcn3_0_soc.max_avg_dram_bw_use_normal_percent =
-				fixed16_to_double_to_cpu(bb->max_avg_dram_bw_use_normal_percent);
-		dcn3_0_soc.writeback_latency_us =
-				fixed16_to_double_to_cpu(bb->writeback_latency_us);
-		dcn3_0_soc.ideal_dram_bw_after_urgent_percent =
-				fixed16_to_double_to_cpu(bb->ideal_dram_bw_after_urgent_percent);
-		dcn3_0_soc.max_request_size_bytes =
-				le32_to_cpu(bb->max_request_size_bytes);
-		dcn3_0_soc.dram_channel_width_bytes =
-				le32_to_cpu(bb->dram_channel_width_bytes);
-		dcn3_0_soc.fabric_datapath_to_dcn_data_return_bytes =
-				le32_to_cpu(bb->fabric_datapath_to_dcn_data_return_bytes);
-		dcn3_0_soc.dcn_downspread_percent =
-				fixed16_to_double_to_cpu(bb->dcn_downspread_percent);
-		dcn3_0_soc.downspread_percent =
-				fixed16_to_double_to_cpu(bb->downspread_percent);
-		dcn3_0_soc.dram_page_open_time_ns =
-				fixed16_to_double_to_cpu(bb->dram_page_open_time_ns);
-		dcn3_0_soc.dram_rw_turnaround_time_ns =
-				fixed16_to_double_to_cpu(bb->dram_rw_turnaround_time_ns);
-		dcn3_0_soc.dram_return_buffer_per_channel_bytes =
-				le32_to_cpu(bb->dram_return_buffer_per_channel_bytes);
-		dcn3_0_soc.round_trip_ping_latency_dcfclk_cycles =
-				le32_to_cpu(bb->round_trip_ping_latency_dcfclk_cycles);
-		dcn3_0_soc.urgent_out_of_order_return_per_channel_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_bytes);
-		dcn3_0_soc.channel_interleave_bytes =
-				le32_to_cpu(bb->channel_interleave_bytes);
-		dcn3_0_soc.num_banks =
-				le32_to_cpu(bb->num_banks);
-		dcn3_0_soc.num_chans =
-				le32_to_cpu(bb->num_chans);
-		dcn3_0_soc.gpuvm_min_page_size_bytes =
-				le32_to_cpu(bb->vmm_page_size_bytes);
-		dcn3_0_soc.dram_clock_change_latency_us =
-				fixed16_to_double_to_cpu(bb->dram_clock_change_latency_us);
-		dcn3_0_soc.writeback_dram_clock_change_latency_us =
-				fixed16_to_double_to_cpu(bb->writeback_dram_clock_change_latency_us);
-		dcn3_0_soc.return_bus_width_bytes =
-				le32_to_cpu(bb->return_bus_width_bytes);
-		dcn3_0_soc.dispclk_dppclk_vco_speed_mhz =
-				le32_to_cpu(bb->dispclk_dppclk_vco_speed_mhz);
-		dcn3_0_soc.xfc_bus_transport_time_us =
-				le32_to_cpu(bb->xfc_bus_transport_time_us);
-		dcn3_0_soc.xfc_xbuf_latency_tolerance_us =
-				le32_to_cpu(bb->xfc_xbuf_latency_tolerance_us);
-		dcn3_0_soc.use_urgent_burst_bw =
-				le32_to_cpu(bb->use_urgent_burst_bw);
-		dcn3_0_soc.num_states =
-				le32_to_cpu(bb->num_states);
-
-		for (i = 0; i < dcn3_0_soc.num_states; i++) {
-			dcn3_0_soc.clock_limits[i].state =
-					le32_to_cpu(bb->clock_limits[i].state);
-			dcn3_0_soc.clock_limits[i].dcfclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dcfclk_mhz);
-			dcn3_0_soc.clock_limits[i].fabricclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].fabricclk_mhz);
-			dcn3_0_soc.clock_limits[i].dispclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dispclk_mhz);
-			dcn3_0_soc.clock_limits[i].dppclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dppclk_mhz);
-			dcn3_0_soc.clock_limits[i].phyclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].phyclk_mhz);
-			dcn3_0_soc.clock_limits[i].socclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].socclk_mhz);
-			dcn3_0_soc.clock_limits[i].dscclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dscclk_mhz);
-			dcn3_0_soc.clock_limits[i].dram_speed_mts =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dram_speed_mts);
-		}
-	}
-
 	loaded_ip->max_num_otg = pool->base.res_cap->num_timing_generator;
 	loaded_ip->max_num_dpp = pool->base.pipe_count;
 	loaded_ip->clamp_min_dcfclk = dc->config.clamp_min_dcfclk;
 	dcn20_patch_bounding_box(dc, loaded_bb);
 
-	if (!bb && dc->ctx->dc_bios->funcs->get_soc_bb_info) {
+	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
 		struct bp_soc_bb_info bb_info = {0};
 
 		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
@@ -2292,17 +2189,15 @@ static noinline void dcn30_calculate_wm_and_dlg_fp(
 		unsigned int min_dram_speed_mts = context->bw_ctx.dml.vba.DRAMSpeed;
 		unsigned int min_dram_speed_mts_margin = 160;
 
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->dummy_pstate_table[0].dummy_pstate_latency_us;
-
 		if (context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_dram_clock_change_unsupported)
 			min_dram_speed_mts = dc->clk_mgr->bw_params->clk_table.entries[dc->clk_mgr->bw_params->clk_table.num_entries - 1].memclk_mhz * 16;
 
-		for (i = 3; i > 0; i--) {
-			if ((min_dram_speed_mts + min_dram_speed_mts_margin > dc->clk_mgr->bw_params->dummy_pstate_table[i].dram_speed_mts) &&
-					(min_dram_speed_mts - min_dram_speed_mts_margin < dc->clk_mgr->bw_params->dummy_pstate_table[i].dram_speed_mts))
-				context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->dummy_pstate_table[i].dummy_pstate_latency_us;
-		}
+		/* find largest table entry that is lower than dram speed, but lower than DPM0 still uses DPM0 */
+		for (i = 3; i > 0; i--)
+			if (min_dram_speed_mts + min_dram_speed_mts_margin > dc->clk_mgr->bw_params->dummy_pstate_table[i].dram_speed_mts)
+				break;
 
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->dummy_pstate_table[i].dummy_pstate_latency_us;
 		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us;
 		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us;
 	}
@@ -2437,16 +2332,28 @@ validate_out:
 	return out;
 }
 
-static noinline void get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
-                                                       unsigned int *optimal_dcfclk,
-                                                       unsigned int *optimal_fclk)
+/*
+ * This must be noinline to ensure anything that deals with FP registers
+ * is contained within this call; previously our compiling with hard-float
+ * would result in fp instructions being emitted outside of the boundaries
+ * of the DC_FP_START/END macros, which makes sense as the compiler has no
+ * idea about what is wrapped and what is not
+ *
+ * This is largely just a workaround to avoid breakage introduced with 5.6,
+ * ideally all fp-using code should be moved into its own file, only that
+ * should be compiled with hard-float, and all code exported from there
+ * should be strictly wrapped with DC_FP_START/END
+ */
+static noinline void dcn30_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk)
 {
        double bw_from_dram, bw_from_dram1, bw_from_dram2;
 
        bw_from_dram1 = uclk_mts * dcn3_0_soc.num_chans *
-                       dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_dram_bw_use_normal_percent / 100);
+		dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_dram_bw_use_normal_percent / 100);
        bw_from_dram2 = uclk_mts * dcn3_0_soc.num_chans *
-                       dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100);
+		dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100);
 
        bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
 
@@ -2505,7 +2412,7 @@ void dcn30_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params
 		// Calculate optimal dcfclk for each uclk
 		for (i = 0; i < num_uclk_states; i++) {
 			DC_FP_START();
-			get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
+			dcn30_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
 					&optimal_dcfclk_for_uclk[i], NULL);
 			DC_FP_END();
 			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) {
@@ -2601,6 +2508,19 @@ static const struct resource_funcs dcn30_res_pool_funcs = {
 	.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
 };
 
+#define CTX ctx
+
+#define REG(reg_name) \
+	(DCN_BASE.instance[0].segment[mm ## reg_name ## _BASE_IDX] + mm ## reg_name)
+
+static uint32_t read_pipe_fuses(struct dc_context *ctx)
+{
+	uint32_t value = REG_READ(CC_DC_PIPE_DIS);
+	/* Support for max 6 pipes */
+	value = value & 0x3f;
+	return value;
+}
+
 static bool dcn30_resource_construct(
 	uint8_t num_virtual_links,
 	struct dc *dc,
@@ -2610,6 +2530,15 @@ static bool dcn30_resource_construct(
 	struct dc_context *ctx = dc->ctx;
 	struct irq_service_init_data init_data;
 	struct ddc_service_init_data ddc_init_data;
+	uint32_t pipe_fuses = read_pipe_fuses(ctx);
+	uint32_t num_pipes = 0;
+
+	if (!(pipe_fuses == 0 || pipe_fuses == 0x3e)) {
+		BREAK_TO_DEBUGGER();
+		dm_error("DC: Unexpected fuse recipe for navi2x !\n");
+		/* fault to single pipe */
+		pipe_fuses = 0x3e;
+	}
 
 	DC_FP_START();
 
@@ -2631,6 +2560,10 @@ static bool dcn30_resource_construct(
 	dc->caps.max_cursor_size = 256;
 	dc->caps.min_horizontal_blanking_period = 80;
 	dc->caps.dmdata_alloc_size = 2048;
+	dc->caps.mall_size_per_mem_channel = 8;
+	/* total size = mall per channel * num channels * 1024 * 1024 */
+	dc->caps.mall_size_total = dc->caps.mall_size_per_mem_channel * dc->ctx->dc_bios->vram_info.num_chans * 1048576;
+	dc->caps.cursor_cache_size = dc->caps.max_cursor_size * dc->caps.max_cursor_size * 8;
 
 	dc->caps.max_slave_planes = 1;
 	dc->caps.post_blend_color_processing = true;
@@ -2739,6 +2672,15 @@ static bool dcn30_resource_construct(
 	/* PP Lib and SMU interfaces */
 	init_soc_bounding_box(dc, pool);
 
+	num_pipes = dcn3_0_ip.max_num_dpp;
+
+	for (i = 0; i < dcn3_0_ip.max_num_dpp; i++)
+		if (pipe_fuses & 1 << i)
+			num_pipes--;
+
+	dcn3_0_ip.max_num_dpp = num_pipes;
+	dcn3_0_ip.max_num_otg = num_pipes;
+
 	dml_init_instance(&dc->dml, &dcn3_0_soc, &dcn3_0_ip, DML_PROJECT_DCN30);
 
 	/* IRQ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/Makefile b/drivers/gpu/drm/amd/display/dc/dcn301/Makefile
index 2fd5d34e4ba6..09264716d1dc 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn301/Makefile
@@ -14,21 +14,18 @@ DCN301 = dcn301_init.o dcn301_resource.o dcn301_dccg.o \
 		dcn301_dio_link_encoder.o dcn301_hwseq.o dcn301_panel_cntl.o dcn301_hubbub.o
 
 ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn301/dcn301_resource.o := -mhard-float -msse
+CFLAGS_$(AMDDALPATH)/dc/dcn301/dcn301_resource.o := -msse
 endif
 
 ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/dcn301/dcn301_resource.o := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn301/dcn301_resource.o := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
 endif
+CFLAGS_$(AMDDALPATH)/dc/dcn301/dcn301_resource.o += -mhard-float
 endif
 
 ifdef CONFIG_X86
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c
index bdad72140cbc..b8bf6d61005b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c
@@ -98,6 +98,7 @@ static const struct hw_sequencer_funcs dcn301_funcs = {
 	.set_abm_immediate_disable = dcn21_set_abm_immediate_disable,
 	.set_pipe = dcn21_set_pipe,
 	.set_disp_pattern_generator = dcn30_set_disp_pattern_generator,
+	.set_hubp_blank = dcn30_set_hubp_blank,
 };
 
 static const struct hwseq_private_funcs dcn301_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c
index 4825c5c1c6ed..c494235016e0 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c
@@ -1489,124 +1489,21 @@ static bool is_soc_bounding_box_valid(struct dc *dc)
 static bool init_soc_bounding_box(struct dc *dc,
 				  struct dcn301_resource_pool *pool)
 {
-	const struct gpu_info_soc_bounding_box_v1_0 *bb = dc->soc_bounding_box;
 	struct _vcs_dpi_soc_bounding_box_st *loaded_bb = &dcn3_01_soc;
 	struct _vcs_dpi_ip_params_st *loaded_ip = &dcn3_01_ip;
 
 	DC_LOGGER_INIT(dc->ctx->logger);
 
-	if (!bb && !is_soc_bounding_box_valid(dc)) {
+	if (!is_soc_bounding_box_valid(dc)) {
 		DC_LOG_ERROR("%s: not valid soc bounding box/n", __func__);
 		return false;
 	}
 
-	if (bb && !is_soc_bounding_box_valid(dc)) {
-		int i;
-
-		dcn3_01_soc.sr_exit_time_us =
-				fixed16_to_double_to_cpu(bb->sr_exit_time_us);
-		dcn3_01_soc.sr_enter_plus_exit_time_us =
-				fixed16_to_double_to_cpu(bb->sr_enter_plus_exit_time_us);
-		dcn3_01_soc.urgent_latency_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_us);
-		dcn3_01_soc.urgent_latency_pixel_data_only_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_pixel_data_only_us);
-		dcn3_01_soc.urgent_latency_pixel_mixed_with_vm_data_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_pixel_mixed_with_vm_data_us);
-		dcn3_01_soc.urgent_latency_vm_data_only_us =
-				fixed16_to_double_to_cpu(bb->urgent_latency_vm_data_only_us);
-		dcn3_01_soc.urgent_out_of_order_return_per_channel_pixel_only_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_pixel_only_bytes);
-		dcn3_01_soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes);
-		dcn3_01_soc.urgent_out_of_order_return_per_channel_vm_only_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_vm_only_bytes);
-		dcn3_01_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_pixel_only);
-		dcn3_01_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm);
-		dcn3_01_soc.pct_ideal_dram_sdp_bw_after_urgent_vm_only =
-				fixed16_to_double_to_cpu(bb->pct_ideal_dram_sdp_bw_after_urgent_vm_only);
-		dcn3_01_soc.max_avg_sdp_bw_use_normal_percent =
-				fixed16_to_double_to_cpu(bb->max_avg_sdp_bw_use_normal_percent);
-		dcn3_01_soc.max_avg_dram_bw_use_normal_percent =
-				fixed16_to_double_to_cpu(bb->max_avg_dram_bw_use_normal_percent);
-		dcn3_01_soc.writeback_latency_us =
-				fixed16_to_double_to_cpu(bb->writeback_latency_us);
-		dcn3_01_soc.ideal_dram_bw_after_urgent_percent =
-				fixed16_to_double_to_cpu(bb->ideal_dram_bw_after_urgent_percent);
-		dcn3_01_soc.max_request_size_bytes =
-				le32_to_cpu(bb->max_request_size_bytes);
-		dcn3_01_soc.dram_channel_width_bytes =
-				le32_to_cpu(bb->dram_channel_width_bytes);
-		dcn3_01_soc.fabric_datapath_to_dcn_data_return_bytes =
-				le32_to_cpu(bb->fabric_datapath_to_dcn_data_return_bytes);
-		dcn3_01_soc.dcn_downspread_percent =
-				fixed16_to_double_to_cpu(bb->dcn_downspread_percent);
-		dcn3_01_soc.downspread_percent =
-				fixed16_to_double_to_cpu(bb->downspread_percent);
-		dcn3_01_soc.dram_page_open_time_ns =
-				fixed16_to_double_to_cpu(bb->dram_page_open_time_ns);
-		dcn3_01_soc.dram_rw_turnaround_time_ns =
-				fixed16_to_double_to_cpu(bb->dram_rw_turnaround_time_ns);
-		dcn3_01_soc.dram_return_buffer_per_channel_bytes =
-				le32_to_cpu(bb->dram_return_buffer_per_channel_bytes);
-		dcn3_01_soc.round_trip_ping_latency_dcfclk_cycles =
-				le32_to_cpu(bb->round_trip_ping_latency_dcfclk_cycles);
-		dcn3_01_soc.urgent_out_of_order_return_per_channel_bytes =
-				le32_to_cpu(bb->urgent_out_of_order_return_per_channel_bytes);
-		dcn3_01_soc.channel_interleave_bytes =
-				le32_to_cpu(bb->channel_interleave_bytes);
-		dcn3_01_soc.num_banks =
-				le32_to_cpu(bb->num_banks);
-		dcn3_01_soc.num_chans =
-				le32_to_cpu(bb->num_chans);
-		dcn3_01_soc.gpuvm_min_page_size_bytes =
-				le32_to_cpu(bb->vmm_page_size_bytes);
-		dcn3_01_soc.dram_clock_change_latency_us =
-				fixed16_to_double_to_cpu(bb->dram_clock_change_latency_us);
-		dcn3_01_soc.writeback_dram_clock_change_latency_us =
-				fixed16_to_double_to_cpu(bb->writeback_dram_clock_change_latency_us);
-		dcn3_01_soc.return_bus_width_bytes =
-				le32_to_cpu(bb->return_bus_width_bytes);
-		dcn3_01_soc.dispclk_dppclk_vco_speed_mhz =
-				le32_to_cpu(bb->dispclk_dppclk_vco_speed_mhz);
-		dcn3_01_soc.xfc_bus_transport_time_us =
-				le32_to_cpu(bb->xfc_bus_transport_time_us);
-		dcn3_01_soc.xfc_xbuf_latency_tolerance_us =
-				le32_to_cpu(bb->xfc_xbuf_latency_tolerance_us);
-		dcn3_01_soc.use_urgent_burst_bw =
-				le32_to_cpu(bb->use_urgent_burst_bw);
-		dcn3_01_soc.num_states =
-				le32_to_cpu(bb->num_states);
-
-		for (i = 0; i < dcn3_01_soc.num_states; i++) {
-			dcn3_01_soc.clock_limits[i].state =
-					le32_to_cpu(bb->clock_limits[i].state);
-			dcn3_01_soc.clock_limits[i].dcfclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dcfclk_mhz);
-			dcn3_01_soc.clock_limits[i].fabricclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].fabricclk_mhz);
-			dcn3_01_soc.clock_limits[i].dispclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dispclk_mhz);
-			dcn3_01_soc.clock_limits[i].dppclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dppclk_mhz);
-			dcn3_01_soc.clock_limits[i].phyclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].phyclk_mhz);
-			dcn3_01_soc.clock_limits[i].socclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].socclk_mhz);
-			dcn3_01_soc.clock_limits[i].dscclk_mhz =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dscclk_mhz);
-			dcn3_01_soc.clock_limits[i].dram_speed_mts =
-					fixed16_to_double_to_cpu(bb->clock_limits[i].dram_speed_mts);
-		}
-	}
-
 	loaded_ip->max_num_otg = pool->base.res_cap->num_timing_generator;
 	loaded_ip->max_num_dpp = pool->base.pipe_count;
 	dcn20_patch_bounding_box(dc, loaded_bb);
 
-	if (!bb && dc->ctx->dc_bios->funcs->get_soc_bb_info) {
+	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
 		struct bp_soc_bb_info bb_info = {0};
 
 		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
@@ -1722,15 +1619,110 @@ static void dcn301_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *b
 	dml_init_instance(&dc->dml, &dcn3_01_soc, &dcn3_01_ip, DML_PROJECT_DCN30);
 }
 
+static void calculate_wm_set_for_vlevel(
+		int vlevel,
+		struct wm_range_table_entry *table_entry,
+		struct dcn_watermarks *wm_set,
+		struct display_mode_lib *dml,
+		display_e2e_pipe_params_st *pipes,
+		int pipe_cnt)
+{
+	double dram_clock_change_latency_cached = dml->soc.dram_clock_change_latency_us;
+
+	ASSERT(vlevel < dml->soc.num_states);
+	/* only pipe 0 is read for voltage and dcf/soc clocks */
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dml->soc.clock_limits[vlevel].dcfclk_mhz;
+	pipes[0].clks_cfg.socclk_mhz = dml->soc.clock_limits[vlevel].socclk_mhz;
+
+	dml->soc.dram_clock_change_latency_us = table_entry->pstate_latency_us;
+	dml->soc.sr_exit_time_us = table_entry->sr_exit_time_us;
+	dml->soc.sr_enter_plus_exit_time_us = table_entry->sr_enter_plus_exit_time_us;
+
+	wm_set->urgent_ns = get_wm_urgent(dml, pipes, pipe_cnt) * 1000;
+	wm_set->cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(dml, pipes, pipe_cnt) * 1000;
+	wm_set->cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(dml, pipes, pipe_cnt) * 1000;
+	wm_set->cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(dml, pipes, pipe_cnt) * 1000;
+	wm_set->pte_meta_urgent_ns = get_wm_memory_trip(dml, pipes, pipe_cnt) * 1000;
+	wm_set->frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(dml, pipes, pipe_cnt) * 1000;
+	wm_set->frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(dml, pipes, pipe_cnt) * 1000;
+	wm_set->urgent_latency_ns = get_urgent_latency(dml, pipes, pipe_cnt) * 1000;
+	dml->soc.dram_clock_change_latency_us = dram_clock_change_latency_cached;
+
+}
+
+static void dcn301_calculate_wm_and_dlg(
+		struct dc *dc, struct dc_state *context,
+		display_e2e_pipe_params_st *pipes,
+		int pipe_cnt,
+		int vlevel_req)
+{
+	int i, pipe_idx;
+	int vlevel, vlevel_max;
+	struct wm_range_table_entry *table_entry;
+	struct clk_bw_params *bw_params = dc->clk_mgr->bw_params;
+
+	ASSERT(bw_params);
+
+	vlevel_max = bw_params->clk_table.num_entries - 1;
+
+	/* WM Set D */
+	table_entry = &bw_params->wm_table.entries[WM_D];
+	if (table_entry->wm_type == WM_TYPE_RETRAINING)
+		vlevel = 0;
+	else
+		vlevel = vlevel_max;
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.d,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+	/* WM Set C */
+	table_entry = &bw_params->wm_table.entries[WM_C];
+	vlevel = min(max(vlevel_req, 2), vlevel_max);
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.c,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+	/* WM Set B */
+	table_entry = &bw_params->wm_table.entries[WM_B];
+	vlevel = min(max(vlevel_req, 1), vlevel_max);
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.b,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+
+	/* WM Set A */
+	table_entry = &bw_params->wm_table.entries[WM_A];
+	vlevel = min(vlevel_req, vlevel_max);
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.a,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt);
+		pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+		if (dc->config.forced_clocks) {
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
+		}
+		if (dc->debug.min_disp_clk_khz > pipes[pipe_idx].clks_cfg.dispclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
+		if (dc->debug.min_dpp_clk_khz > pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
+
+		pipe_idx++;
+	}
+
+	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+}
+
 static struct resource_funcs dcn301_res_pool_funcs = {
 	.destroy = dcn301_destroy_resource_pool,
 	.link_enc_create = dcn301_link_encoder_create,
 	.panel_cntl_create = dcn301_panel_cntl_create,
 	.validate_bandwidth = dcn30_validate_bandwidth,
-	.calculate_wm_and_dlg = dcn30_calculate_wm_and_dlg,
+	.calculate_wm_and_dlg = dcn301_calculate_wm_and_dlg,
 	.populate_dml_pipes = dcn30_populate_dml_pipes_from_context,
 	.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
 	.add_stream_to_ctx = dcn30_add_stream_to_ctx,
+	.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
 	.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
 	.populate_dml_writeback_from_context = dcn30_populate_dml_writeback_from_context,
 	.set_mcif_arb_params = dcn30_set_mcif_arb_params,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/Makefile b/drivers/gpu/drm/amd/display/dc/dcn302/Makefile
index 36e44e1b07fa..101620a8867a 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn302/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn302/Makefile
@@ -13,21 +13,18 @@
 DCN3_02 = dcn302_init.o dcn302_hwseq.o dcn302_resource.o
 
 ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o := -mhard-float -msse
+CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o := -msse
 endif
 
 ifdef CONFIG_PPC64
 CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
 endif
+CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o += -mhard-float
 endif
 
 ifdef CONFIG_X86
diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c
index 808c4dcdb3ac..4b659b63f75b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c
@@ -53,6 +53,8 @@
 #include "dce/dce_i2c_hw.h"
 #include "dce/dce_panel_cntl.h"
 #include "dce/dmub_abm.h"
+#include "dce/dmub_psr.h"
+#include "clk_mgr.h"
 
 #include "hw_sequencer_private.h"
 #include "reg_helper.h"
@@ -162,8 +164,8 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_02_soc = {
 
 		.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
 		.num_states = 1,
-		.sr_exit_time_us = 5.20,
-		.sr_enter_plus_exit_time_us = 9.60,
+		.sr_exit_time_us = 12,
+		.sr_enter_plus_exit_time_us = 20,
 		.urgent_latency_us = 4.0,
 		.urgent_latency_pixel_data_only_us = 4.0,
 		.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
@@ -190,7 +192,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_02_soc = {
 		.num_banks = 8,
 		.gpuvm_min_page_size_bytes = 4096,
 		.hostvm_min_page_size_bytes = 4096,
-		.dram_clock_change_latency_us = 350,
+		.dram_clock_change_latency_us = 404,
 		.dummy_pstate_latency_us = 5,
 		.writeback_dram_clock_change_latency_us = 23.0,
 		.return_bus_width_bytes = 64,
@@ -238,6 +240,7 @@ static const struct dc_debug_options debug_defaults_diags = {
 		.dwb_fi_phase = -1, // -1 = disable
 		.dmub_command_table = true,
 		.enable_tri_buf = true,
+		.disable_psr = true,
 };
 
 enum dcn302_clk_src_array_id {
@@ -1213,6 +1216,9 @@ static void dcn302_resource_destruct(struct resource_pool *pool)
 			dce_abm_destroy(&pool->multiple_abms[i]);
 	}
 
+	if (pool->psr != NULL)
+		dmub_psr_destroy(&pool->psr);
+
 	if (pool->dccg != NULL)
 		dcn_dccg_destroy(&pool->dccg);
 }
@@ -1224,6 +1230,165 @@ static void dcn302_destroy_resource_pool(struct resource_pool **pool)
 	*pool = NULL;
 }
 
+static void dcn302_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk)
+{
+	double bw_from_dram, bw_from_dram1, bw_from_dram2;
+
+	bw_from_dram1 = uclk_mts * dcn3_02_soc.num_chans *
+		dcn3_02_soc.dram_channel_width_bytes * (dcn3_02_soc.max_avg_dram_bw_use_normal_percent / 100);
+	bw_from_dram2 = uclk_mts * dcn3_02_soc.num_chans *
+		dcn3_02_soc.dram_channel_width_bytes * (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100);
+
+	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
+
+	if (optimal_fclk)
+		*optimal_fclk = bw_from_dram /
+		(dcn3_02_soc.fabric_datapath_to_dcn_data_return_bytes * (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100));
+
+	if (optimal_dcfclk)
+		*optimal_dcfclk =  bw_from_dram /
+		(dcn3_02_soc.return_bus_width_bytes * (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100));
+}
+
+void dcn302_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	unsigned int i, j;
+	unsigned int num_states = 0;
+
+	unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
+	unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
+	unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
+	unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
+
+	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {694, 875, 1000, 1200};
+	unsigned int num_dcfclk_sta_targets = 4;
+	unsigned int num_uclk_states;
+
+
+	if (dc->ctx->dc_bios->vram_info.num_chans)
+		dcn3_02_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dcn3_02_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+
+	dcn3_02_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+
+	if (bw_params->clk_table.entries[0].memclk_mhz) {
+		int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
+
+		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
+				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
+				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
+				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
+				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		}
+		if (!max_dcfclk_mhz)
+			max_dcfclk_mhz = dcn3_02_soc.clock_limits[0].dcfclk_mhz;
+		if (!max_dispclk_mhz)
+			max_dispclk_mhz = dcn3_02_soc.clock_limits[0].dispclk_mhz;
+		if (!max_dppclk_mhz)
+			max_dppclk_mhz = dcn3_02_soc.clock_limits[0].dppclk_mhz;
+		if (!max_phyclk_mhz)
+			max_phyclk_mhz = dcn3_02_soc.clock_limits[0].phyclk_mhz;
+
+		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			/* If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array */
+			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
+			num_dcfclk_sta_targets++;
+		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			/* If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates */
+			for (i = 0; i < num_dcfclk_sta_targets; i++) {
+				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
+					dcfclk_sta_targets[i] = max_dcfclk_mhz;
+					break;
+				}
+			}
+			/* Update size of array since we "removed" duplicates */
+			num_dcfclk_sta_targets = i + 1;
+		}
+
+		num_uclk_states = bw_params->clk_table.num_entries;
+
+		/* Calculate optimal dcfclk for each uclk */
+		for (i = 0; i < num_uclk_states; i++) {
+			dcn302_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
+					&optimal_dcfclk_for_uclk[i], NULL);
+			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) {
+				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
+			}
+		}
+
+		/* Calculate optimal uclk for each dcfclk sta target */
+		for (i = 0; i < num_dcfclk_sta_targets; i++) {
+			for (j = 0; j < num_uclk_states; j++) {
+				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+					optimal_uclk_for_dcfclk_sta_targets[i] =
+							bw_params->clk_table.entries[j].memclk_mhz * 16;
+					break;
+				}
+			}
+		}
+
+		i = 0;
+		j = 0;
+		/* create the final dcfclk and uclk table */
+		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
+			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j] && i < num_dcfclk_sta_targets) {
+				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+			} else {
+				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+					dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+				} else {
+					j = num_uclk_states;
+				}
+			}
+		}
+
+		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
+			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+		}
+
+		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
+				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+		}
+
+		dcn3_02_soc.num_states = num_states;
+		for (i = 0; i < dcn3_02_soc.num_states; i++) {
+			dcn3_02_soc.clock_limits[i].state = i;
+			dcn3_02_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
+			dcn3_02_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
+			dcn3_02_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
+
+			/* Fill all states with max values of all other clocks */
+			dcn3_02_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
+			dcn3_02_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
+			dcn3_02_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
+			dcn3_02_soc.clock_limits[i].dtbclk_mhz = dcn3_02_soc.clock_limits[0].dtbclk_mhz;
+			/* These clocks cannot come from bw_params, always fill from dcn3_02_soc[1] */
+			/* FCLK, PHYCLK_D18, SOCCLK, DSCCLK */
+			dcn3_02_soc.clock_limits[i].phyclk_d18_mhz = dcn3_02_soc.clock_limits[0].phyclk_d18_mhz;
+			dcn3_02_soc.clock_limits[i].socclk_mhz = dcn3_02_soc.clock_limits[0].socclk_mhz;
+			dcn3_02_soc.clock_limits[i].dscclk_mhz = dcn3_02_soc.clock_limits[0].dscclk_mhz;
+		}
+		/* re-init DML with updated bb */
+		dml_init_instance(&dc->dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
+		if (dc->current_state)
+			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
+	}
+}
+
 static struct resource_funcs dcn302_res_pool_funcs = {
 		.destroy = dcn302_destroy_resource_pool,
 		.link_enc_create = dcn302_link_encoder_create,
@@ -1240,7 +1405,7 @@ static struct resource_funcs dcn302_res_pool_funcs = {
 		.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
 		.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
 		.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
-		.update_bw_bounding_box = dcn30_update_bw_bounding_box,
+		.update_bw_bounding_box = dcn302_update_bw_bounding_box,
 		.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
 };
 
@@ -1311,7 +1476,10 @@ static bool dcn302_resource_construct(
 	dc->caps.max_cursor_size = 256;
 	dc->caps.min_horizontal_blanking_period = 80;
 	dc->caps.dmdata_alloc_size = 2048;
-
+	dc->caps.mall_size_per_mem_channel = 4;
+	/* total size = mall per channel * num channels * 1024 * 1024 */
+	dc->caps.mall_size_total = dc->caps.mall_size_per_mem_channel * dc->ctx->dc_bios->vram_info.num_chans * 1048576;
+	dc->caps.cursor_cache_size = dc->caps.max_cursor_size * dc->caps.max_cursor_size * 8;
 	dc->caps.max_slave_planes = 1;
 	dc->caps.post_blend_color_processing = true;
 	dc->caps.force_dp_tps4_for_cp2520 = true;
@@ -1354,8 +1522,6 @@ static bool dcn302_resource_construct(
 
 	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV)
 		dc->debug = debug_defaults_drv;
-	else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS)
-		dc->debug = debug_defaults_diags;
 	else
 		dc->debug = debug_defaults_diags;
 
@@ -1469,6 +1635,14 @@ static bool dcn302_resource_construct(
 	}
 	pool->timing_generator_count = i;
 
+	/* PSR */
+	pool->psr = dmub_psr_create(ctx);
+	if (pool->psr == NULL) {
+		dm_error("DC: failed to create psr!\n");
+		BREAK_TO_DEBUGGER();
+		goto create_fail;
+	}
+
 	/* ABMs */
 	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
 		pool->multiple_abms[i] = dmub_abm_create(ctx, &abm_regs[i], &abm_shift, &abm_mask);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.h b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.h
index 71f7deed18e3..42d2c73e30bc 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.h
@@ -30,4 +30,6 @@
 
 struct resource_pool *dcn302_create_resource_pool(const struct dc_init_data *init_data, struct dc *dc);
 
+void dcn302_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+
 #endif /* _DCN302_RESOURCE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h b/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h
index 5da7677627a1..cac0b2c0d31b 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h
@@ -30,9 +30,10 @@ struct dc_link;
 
 struct cp_psp_stream_config {
 	uint8_t otg_inst;
-	uint8_t link_enc_inst;
-	uint8_t stream_enc_inst;
-	uint8_t mst_supported;
+	uint8_t dig_be;
+	uint8_t dig_fe;
+	uint8_t assr_enabled;
+	uint8_t mst_enabled;
 	void *dm_stream_ctx;
 	bool dpms_off;
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dml/Makefile b/drivers/gpu/drm/amd/display/dc/dml/Makefile
index a02a33dcd70b..6bb7f2905821 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dml/Makefile
@@ -33,10 +33,6 @@ ifdef CONFIG_PPC64
 dml_ccflags := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-dml_rcflags := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c
index 45f028986a8d..0f3f510fd83b 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c
@@ -3138,7 +3138,7 @@ static void CalculateFlipSchedule(
 				4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime + 0.125),
 				1) / 4.0;
 
-		if ((GPUVMEnable == true || DCCEnable == true)) {
+		if ((GPUVMEnable || DCCEnable)) {
 			mode_lib->vba.ImmediateFlipBW[0] = BandwidthAvailableForImmediateFlip
 					* ImmediateFlipBytes / TotImmediateFlipBytes;
 			TimeForFetchingRowInVBlankImmediateFlip = dml_max(
@@ -4168,10 +4168,11 @@ void dml20_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	for (i = 0; i <= mode_lib->vba.soc.num_states; i++) {
 		locals->DIOSupport[i] = true;
 		for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
-			if (locals->OutputBppPerState[i][k] == BPP_INVALID
-					|| (mode_lib->vba.OutputFormat[k] == dm_420
+			if (!mode_lib->vba.skip_dio_check[k]
+					&& (locals->OutputBppPerState[i][k] == BPP_INVALID
+						|| (mode_lib->vba.OutputFormat[k] == dm_420
 							&& mode_lib->vba.Interlace[k] == true
-							&& mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true)) {
+							&& mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true))) {
 				locals->DIOSupport[i] = false;
 			}
 		}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c
index 860e72a51534..210c96cd5b03 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c
@@ -2635,14 +2635,15 @@ static void dml20v2_DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndP
 	}
 
 	if (mode_lib->vba.DRAMClockChangeSupportsVActive &&
-			mode_lib->vba.MinActiveDRAMClockChangeMargin > 60 &&
-			mode_lib->vba.PrefetchMode[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb] == 0) {
+			mode_lib->vba.MinActiveDRAMClockChangeMargin > 60) {
 		mode_lib->vba.DRAMClockChangeWatermark += 25;
 
 		for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) {
-			if (mode_lib->vba.DRAMClockChangeWatermark >
-			dml_max(mode_lib->vba.StutterEnterPlusExitWatermark, mode_lib->vba.UrgentWatermark))
-				mode_lib->vba.MinTTUVBlank[k] += 25;
+			if (mode_lib->vba.PrefetchMode[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb] == 0) {
+				if (mode_lib->vba.DRAMClockChangeWatermark >
+				dml_max(mode_lib->vba.StutterEnterPlusExitWatermark, mode_lib->vba.UrgentWatermark))
+					mode_lib->vba.MinTTUVBlank[k] += 25;
+			}
 		}
 
 		mode_lib->vba.DRAMClockChangeSupport[0][0] = dm_dram_clock_change_vactive;
@@ -3262,6 +3263,7 @@ static void CalculateFlipSchedule(
 
 static unsigned int TruncToValidBPP(
 		double DecimalBPP,
+		double DesiredBPP,
 		bool DSCEnabled,
 		enum output_encoder_class Output,
 		enum output_format_class Format,
@@ -3269,31 +3271,31 @@ static unsigned int TruncToValidBPP(
 {
 	if (Output == dm_hdmi) {
 		if (Format == dm_420) {
-			if (DecimalBPP >= 18)
+			if (DecimalBPP >= 18 && (DesiredBPP == 0 || DesiredBPP == 18))
 				return 18;
-			else if (DecimalBPP >= 15)
+			else if (DecimalBPP >= 15 && (DesiredBPP == 0 || DesiredBPP == 15))
 				return 15;
-			else if (DecimalBPP >= 12)
+			else if (DecimalBPP >= 12 && (DesiredBPP == 0 || DesiredBPP == 12))
 				return 12;
 			else
 				return BPP_INVALID;
 		} else if (Format == dm_444) {
-			if (DecimalBPP >= 36)
+			if (DecimalBPP >= 36 && (DesiredBPP == 0 || DesiredBPP == 36))
 				return 36;
-			else if (DecimalBPP >= 30)
+			else if (DecimalBPP >= 30 && (DesiredBPP == 0 || DesiredBPP == 30))
 				return 30;
-			else if (DecimalBPP >= 24)
+			else if (DecimalBPP >= 24 && (DesiredBPP == 0 || DesiredBPP == 24))
 				return 24;
-			else if (DecimalBPP >= 18)
+			else if (DecimalBPP >= 18 && (DesiredBPP == 0 || DesiredBPP == 18))
 				return 18;
 			else
 				return BPP_INVALID;
 		} else {
-			if (DecimalBPP / 1.5 >= 24)
+			if (DecimalBPP / 1.5 >= 24 && (DesiredBPP == 0 || DesiredBPP == 24))
 				return 24;
-			else if (DecimalBPP / 1.5 >= 20)
+			else if (DecimalBPP / 1.5 >= 20 && (DesiredBPP == 0 || DesiredBPP == 20))
 				return 20;
-			else if (DecimalBPP / 1.5 >= 16)
+			else if (DecimalBPP / 1.5 >= 16 && (DesiredBPP == 0 || DesiredBPP == 16))
 				return 16;
 			else
 				return BPP_INVALID;
@@ -3301,53 +3303,86 @@ static unsigned int TruncToValidBPP(
 	} else {
 		if (DSCEnabled) {
 			if (Format == dm_420) {
-				if (DecimalBPP < 6)
-					return BPP_INVALID;
-				else if (DecimalBPP >= 1.5 * DSCInputBitPerComponent - 1 / 16)
-					return 1.5 * DSCInputBitPerComponent - 1 / 16;
-				else
-					return dml_floor(16 * DecimalBPP, 1) / 16;
+				if (DesiredBPP == 0) {
+					if (DecimalBPP < 6)
+						return BPP_INVALID;
+					else if (DecimalBPP >= 1.5 * DSCInputBitPerComponent - 1.0 / 16.0)
+						return 1.5 * DSCInputBitPerComponent - 1.0 / 16.0;
+					else
+						return dml_floor(16 * DecimalBPP, 1) / 16.0;
+				} else {
+					if (DecimalBPP < 6
+							|| DesiredBPP < 6
+							|| DesiredBPP > 1.5 * DSCInputBitPerComponent - 1.0 / 16.0
+							|| DecimalBPP < DesiredBPP) {
+						return BPP_INVALID;
+					} else {
+						return DesiredBPP;
+					}
+				}
 			} else if (Format == dm_n422) {
-				if (DecimalBPP < 7)
-					return BPP_INVALID;
-				else if (DecimalBPP >= 2 * DSCInputBitPerComponent - 1 / 16)
-					return 2 * DSCInputBitPerComponent - 1 / 16;
-				else
-					return dml_floor(16 * DecimalBPP, 1) / 16;
+				if (DesiredBPP == 0) {
+					if (DecimalBPP < 7)
+						return BPP_INVALID;
+					else if (DecimalBPP >= 2 * DSCInputBitPerComponent - 1.0 / 16.0)
+						return 2 * DSCInputBitPerComponent - 1.0 / 16.0;
+					else
+						return dml_floor(16 * DecimalBPP, 1) / 16.0;
+				} else {
+					if (DecimalBPP < 7
+							|| DesiredBPP < 7
+							|| DesiredBPP > 2 * DSCInputBitPerComponent - 1.0 / 16.0
+							|| DecimalBPP < DesiredBPP) {
+						return BPP_INVALID;
+					} else {
+						return DesiredBPP;
+					}
+				}
 			} else {
-				if (DecimalBPP < 8)
-					return BPP_INVALID;
-				else if (DecimalBPP >= 3 * DSCInputBitPerComponent - 1 / 16)
-					return 3 * DSCInputBitPerComponent - 1 / 16;
-				else
-					return dml_floor(16 * DecimalBPP, 1) / 16;
+				if (DesiredBPP == 0) {
+					if (DecimalBPP < 8)
+						return BPP_INVALID;
+					else if (DecimalBPP >= 3 * DSCInputBitPerComponent - 1.0 / 16.0)
+						return 3 * DSCInputBitPerComponent - 1.0 / 16.0;
+					else
+						return dml_floor(16 * DecimalBPP, 1) / 16.0;
+				} else {
+					if (DecimalBPP < 8
+							|| DesiredBPP < 8
+							|| DesiredBPP > 3 * DSCInputBitPerComponent - 1.0 / 16.0
+							|| DecimalBPP < DesiredBPP) {
+						return BPP_INVALID;
+					} else {
+						return DesiredBPP;
+					}
+				}
 			}
 		} else if (Format == dm_420) {
-			if (DecimalBPP >= 18)
+			if (DecimalBPP >= 18 && (DesiredBPP == 0 || DesiredBPP == 18))
 				return 18;
-			else if (DecimalBPP >= 15)
+			else if (DecimalBPP >= 15 && (DesiredBPP == 0 || DesiredBPP == 15))
 				return 15;
-			else if (DecimalBPP >= 12)
+			else if (DecimalBPP >= 12 && (DesiredBPP == 0 || DesiredBPP == 12))
 				return 12;
 			else
 				return BPP_INVALID;
 		} else if (Format == dm_s422 || Format == dm_n422) {
-			if (DecimalBPP >= 24)
+			if (DecimalBPP >= 24 && (DesiredBPP == 0 || DesiredBPP == 24))
 				return 24;
-			else if (DecimalBPP >= 20)
+			else if (DecimalBPP >= 20 && (DesiredBPP == 0 || DesiredBPP == 20))
 				return 20;
-			else if (DecimalBPP >= 16)
+			else if (DecimalBPP >= 16 && (DesiredBPP == 0 || DesiredBPP == 16))
 				return 16;
 			else
 				return BPP_INVALID;
 		} else {
-			if (DecimalBPP >= 36)
+			if (DecimalBPP >= 36 && (DesiredBPP == 0 || DesiredBPP == 36))
 				return 36;
-			else if (DecimalBPP >= 30)
+			else if (DecimalBPP >= 30 && (DesiredBPP == 0 || DesiredBPP == 30))
 				return 30;
-			else if (DecimalBPP >= 24)
+			else if (DecimalBPP >= 24 && (DesiredBPP == 0 || DesiredBPP == 24))
 				return 24;
-			else if (DecimalBPP >= 18)
+			else if (DecimalBPP >= 18 && (DesiredBPP == 0 || DesiredBPP == 18))
 				return 18;
 			else
 				return BPP_INVALID;
@@ -4136,6 +4171,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 					locals->RequiresFEC[i][k] = 0;
 					locals->OutputBppPerState[i][k] = TruncToValidBPP(
 							dml_min(600.0, mode_lib->vba.PHYCLKPerState[i]) / mode_lib->vba.PixelClockBackEnd[k] * 24,
+							mode_lib->vba.ForcedOutputLinkBPP[k],
 							false,
 							mode_lib->vba.Output[k],
 							mode_lib->vba.OutputFormat[k],
@@ -4152,6 +4188,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						mode_lib->vba.Outbpp = TruncToValidBPP(
 								(1.0 - mode_lib->vba.Downspreading / 100.0) * 270.0
 								* mode_lib->vba.OutputLinkDPLanes[k] / mode_lib->vba.PixelClockBackEnd[k] * 8.0,
+								mode_lib->vba.ForcedOutputLinkBPP[k],
 								false,
 								mode_lib->vba.Output[k],
 								mode_lib->vba.OutputFormat[k],
@@ -4159,6 +4196,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						mode_lib->vba.OutbppDSC = TruncToValidBPP(
 								(1.0 - mode_lib->vba.Downspreading / 100.0) * (1.0 - mode_lib->vba.EffectiveFECOverhead / 100.0) * 270.0
 								* mode_lib->vba.OutputLinkDPLanes[k] / mode_lib->vba.PixelClockBackEnd[k] * 8.0,
+								mode_lib->vba.ForcedOutputLinkBPP[k],
 								true,
 								mode_lib->vba.Output[k],
 								mode_lib->vba.OutputFormat[k],
@@ -4181,6 +4219,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						mode_lib->vba.Outbpp = TruncToValidBPP(
 								(1.0 - mode_lib->vba.Downspreading / 100.0) * 540.0
 								* mode_lib->vba.OutputLinkDPLanes[k] / mode_lib->vba.PixelClockBackEnd[k] * 8.0,
+								mode_lib->vba.ForcedOutputLinkBPP[k],
 									false,
 									mode_lib->vba.Output[k],
 									mode_lib->vba.OutputFormat[k],
@@ -4188,6 +4227,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						mode_lib->vba.OutbppDSC = TruncToValidBPP(
 								(1.0 - mode_lib->vba.Downspreading / 100.0) * (1.0 - mode_lib->vba.EffectiveFECOverhead / 100.0) * 540.0
 								* mode_lib->vba.OutputLinkDPLanes[k] / mode_lib->vba.PixelClockBackEnd[k] * 8.0,
+								mode_lib->vba.ForcedOutputLinkBPP[k],
 								true,
 								mode_lib->vba.Output[k],
 								mode_lib->vba.OutputFormat[k],
@@ -4212,6 +4252,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						mode_lib->vba.Outbpp = TruncToValidBPP(
 								(1.0 - mode_lib->vba.Downspreading / 100.0) * 810.0
 								* mode_lib->vba.OutputLinkDPLanes[k] / mode_lib->vba.PixelClockBackEnd[k] * 8.0,
+								mode_lib->vba.ForcedOutputLinkBPP[k],
 								false,
 								mode_lib->vba.Output[k],
 								mode_lib->vba.OutputFormat[k],
@@ -4219,6 +4260,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						mode_lib->vba.OutbppDSC = TruncToValidBPP(
 								(1.0 - mode_lib->vba.Downspreading / 100.0) * (1.0 - mode_lib->vba.EffectiveFECOverhead / 100.0) * 810.0
 								* mode_lib->vba.OutputLinkDPLanes[k] / mode_lib->vba.PixelClockBackEnd[k] * 8.0,
+								mode_lib->vba.ForcedOutputLinkBPP[k],
 								true,
 								mode_lib->vba.Output[k],
 								mode_lib->vba.OutputFormat[k],
@@ -4247,10 +4289,11 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 	for (i = 0; i <= mode_lib->vba.soc.num_states; i++) {
 		locals->DIOSupport[i] = true;
 		for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
-			if (locals->OutputBppPerState[i][k] == BPP_INVALID
-					|| (mode_lib->vba.OutputFormat[k] == dm_420
+			if (!mode_lib->vba.skip_dio_check[k]
+					&& (locals->OutputBppPerState[i][k] == BPP_INVALID
+						|| (mode_lib->vba.OutputFormat[k] == dm_420
 							&& mode_lib->vba.Interlace[k] == true
-							&& mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true)) {
+							&& mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true))) {
 				locals->DIOSupport[i] = false;
 			}
 		}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c
index 86ff24dffc3e..398210d1af34 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c
@@ -4257,10 +4257,11 @@ void dml21_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	for (i = 0; i <= mode_lib->vba.soc.num_states; i++) {
 		locals->DIOSupport[i] = true;
 		for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
-			if (locals->OutputBppPerState[i][k] == BPP_INVALID
-					|| (mode_lib->vba.OutputFormat[k] == dm_420
+			if (!mode_lib->vba.skip_dio_check[k]
+					&& (locals->OutputBppPerState[i][k] == BPP_INVALID
+						|| (mode_lib->vba.OutputFormat[k] == dm_420
 							&& mode_lib->vba.Interlace[k] == true
-							&& mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true)) {
+							&& mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true))) {
 				locals->DIOSupport[i] = false;
 			}
 		}
@@ -5121,48 +5122,48 @@ void dml21_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		for (j = 0; j < 2; j++) {
 			enum dm_validation_status status = DML_VALIDATION_OK;
 
-			if (mode_lib->vba.ScaleRatioAndTapsSupport != true) {
+			if (!mode_lib->vba.ScaleRatioAndTapsSupport) {
 				status = DML_FAIL_SCALE_RATIO_TAP;
-			} else if (mode_lib->vba.SourceFormatPixelAndScanSupport != true) {
+			} else if (!mode_lib->vba.SourceFormatPixelAndScanSupport) {
 				status = DML_FAIL_SOURCE_PIXEL_FORMAT;
-			} else if (locals->ViewportSizeSupport[i][0] != true) {
+			} else if (!locals->ViewportSizeSupport[i][0]) {
 				status = DML_FAIL_VIEWPORT_SIZE;
-			} else if (locals->DIOSupport[i] != true) {
+			} else if (!locals->DIOSupport[i]) {
 				status = DML_FAIL_DIO_SUPPORT;
-			} else if (locals->NotEnoughDSCUnits[i] != false) {
+			} else if (locals->NotEnoughDSCUnits[i]) {
 				status = DML_FAIL_NOT_ENOUGH_DSC;
-			} else if (locals->DSCCLKRequiredMoreThanSupported[i] != false) {
+			} else if (locals->DSCCLKRequiredMoreThanSupported[i]) {
 				status = DML_FAIL_DSC_CLK_REQUIRED;
-			} else if (locals->ROBSupport[i][0] != true) {
+			} else if (!locals->ROBSupport[i][0]) {
 				status = DML_FAIL_REORDERING_BUFFER;
-			} else if (locals->DISPCLK_DPPCLK_Support[i][j] != true) {
+			} else if (!locals->DISPCLK_DPPCLK_Support[i][j]) {
 				status = DML_FAIL_DISPCLK_DPPCLK;
-			} else if (locals->TotalAvailablePipesSupport[i][j] != true) {
+			} else if (!locals->TotalAvailablePipesSupport[i][j]) {
 				status = DML_FAIL_TOTAL_AVAILABLE_PIPES;
-			} else if (mode_lib->vba.NumberOfOTGSupport != true) {
+			} else if (!mode_lib->vba.NumberOfOTGSupport) {
 				status = DML_FAIL_NUM_OTG;
-			} else if (mode_lib->vba.WritebackModeSupport != true) {
+			} else if (!mode_lib->vba.WritebackModeSupport) {
 				status = DML_FAIL_WRITEBACK_MODE;
-			} else if (mode_lib->vba.WritebackLatencySupport != true) {
+			} else if (!mode_lib->vba.WritebackLatencySupport) {
 				status = DML_FAIL_WRITEBACK_LATENCY;
-			} else if (mode_lib->vba.WritebackScaleRatioAndTapsSupport != true) {
+			} else if (!mode_lib->vba.WritebackScaleRatioAndTapsSupport) {
 				status = DML_FAIL_WRITEBACK_SCALE_RATIO_TAP;
-			} else if (mode_lib->vba.CursorSupport != true) {
+			} else if (!mode_lib->vba.CursorSupport) {
 				status = DML_FAIL_CURSOR_SUPPORT;
-			} else if (mode_lib->vba.PitchSupport != true) {
+			} else if (!mode_lib->vba.PitchSupport) {
 				status = DML_FAIL_PITCH_SUPPORT;
-			} else if (locals->TotalVerticalActiveBandwidthSupport[i][0] != true) {
+			} else if (!locals->TotalVerticalActiveBandwidthSupport[i][0]) {
 				status = DML_FAIL_TOTAL_V_ACTIVE_BW;
-			} else if (locals->PTEBufferSizeNotExceeded[i][j] != true) {
+			} else if (!locals->PTEBufferSizeNotExceeded[i][j]) {
 				status = DML_FAIL_PTE_BUFFER_SIZE;
-			} else if (mode_lib->vba.NonsupportedDSCInputBPC != false) {
+			} else if (mode_lib->vba.NonsupportedDSCInputBPC) {
 				status = DML_FAIL_DSC_INPUT_BPC;
-			} else if ((mode_lib->vba.HostVMEnable != false
-					&& locals->ImmediateFlipSupportedForState[i][j] != true)) {
+			} else if ((mode_lib->vba.HostVMEnable
+					&& !locals->ImmediateFlipSupportedForState[i][j])) {
 				status = DML_FAIL_HOST_VM_IMMEDIATE_FLIP;
-			} else if (locals->PrefetchSupported[i][j] != true) {
+			} else if (!locals->PrefetchSupported[i][j]) {
 				status = DML_FAIL_PREFETCH_SUPPORT;
-			} else if (locals->VRatioInPrefetchSupported[i][j] != true) {
+			} else if (!locals->VRatioInPrefetchSupported[i][j]) {
 				status = DML_FAIL_V_RATIO_PREFETCH;
 			}
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
index 319dec59bcd1..bc07082c1357 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
@@ -1219,13 +1219,13 @@ static bool CalculatePrefetchSchedule(
 			dml_print("DML: prefetch_bw_equ: %f\n", prefetch_bw_equ);
 
 			if (prefetch_bw_equ > 0) {
-				if (GPUVMEnable == true) {
+				if (GPUVMEnable) {
 					Tvm_equ = dml_max3(*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / prefetch_bw_equ, Tvm_trips, LineTime / 4);
 				} else {
 					Tvm_equ = LineTime / 4;
 				}
 
-				if ((GPUVMEnable == true || myPipe->DCCEnable == true)) {
+				if ((GPUVMEnable || myPipe->DCCEnable)) {
 					Tr0_equ = dml_max4(
 							(MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / prefetch_bw_equ,
 							Tr0_trips,
@@ -4263,7 +4263,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	for (i = 0; i < v->soc.num_states; i++) {
 		v->DIOSupport[i] = true;
 		for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
-			if (v->BlendingAndTiming[k] == k && (v->Output[k] == dm_dp || v->Output[k] == dm_edp || v->Output[k] == dm_hdmi)
+			if (!v->skip_dio_check[k] && v->BlendingAndTiming[k] == k && (v->Output[k] == dm_dp || v->Output[k] == dm_edp || v->Output[k] == dm_hdmi)
 					&& (v->OutputBppPerState[i][k] == 0
 							|| (v->OutputFormat[k] == dm_420 && v->Interlace[k] == true && v->ProgressiveToInterlaceUnitInOPP == true))) {
 				v->DIOSupport[i] = false;
@@ -5558,7 +5558,7 @@ static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 		}
 	}
 
-	if (mode_lib->vba.MinActiveDRAMClockChangeMargin > 0 && PrefetchMode == 0) {
+	if (mode_lib->vba.MinActiveDRAMClockChangeMargin > 0) {
 		*DRAMClockChangeSupport = dm_dram_clock_change_vactive;
 	} else if (((mode_lib->vba.SynchronizedVBlank == true || mode_lib->vba.TotalNumberOfActiveOTG == 1 || SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank > 0) && PrefetchMode == 0)) {
 		*DRAMClockChangeSupport = dm_dram_clock_change_vblank;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
index 5b5916b5bc71..0f14f205ebe5 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
@@ -165,8 +165,8 @@ static void handle_det_buf_split(struct display_mode_lib *mode_lib,
 	unsigned int swath_bytes_c = 0;
 	unsigned int full_swath_bytes_packed_l = 0;
 	unsigned int full_swath_bytes_packed_c = 0;
-	bool req128_l = 0;
-	bool req128_c = 0;
+	bool req128_l = false;
+	bool req128_c = false;
 	bool surf_linear = (pipe_src_param.sw_mode == dm_sw_linear);
 	bool surf_vert = (pipe_src_param.source_scan == dm_vert);
 	unsigned int log2_swath_height_l = 0;
@@ -191,37 +191,37 @@ static void handle_det_buf_split(struct display_mode_lib *mode_lib,
 		total_swath_bytes = 2 * full_swath_bytes_packed_l;
 
 	if (total_swath_bytes <= detile_buf_size_in_bytes) { //full 256b request
-		req128_l = 0;
-		req128_c = 0;
+		req128_l = false;
+		req128_c = false;
 		swath_bytes_l = full_swath_bytes_packed_l;
 		swath_bytes_c = full_swath_bytes_packed_c;
 	} else if (!rq_param->yuv420) {
-		req128_l = 1;
-		req128_c = 0;
+		req128_l = true;
+		req128_c = false;
 		swath_bytes_c = full_swath_bytes_packed_c;
 		swath_bytes_l = full_swath_bytes_packed_l / 2;
 	} else if ((double)full_swath_bytes_packed_l / (double)full_swath_bytes_packed_c < 1.5) {
-		req128_l = 0;
-		req128_c = 1;
+		req128_l = false;
+		req128_c = true;
 		swath_bytes_l = full_swath_bytes_packed_l;
 		swath_bytes_c = full_swath_bytes_packed_c / 2;
 
 		total_swath_bytes = 2 * swath_bytes_l + 2 * swath_bytes_c;
 
 		if (total_swath_bytes > detile_buf_size_in_bytes) {
-			req128_l = 1;
+			req128_l = true;
 			swath_bytes_l = full_swath_bytes_packed_l / 2;
 		}
 	} else {
-		req128_l = 1;
-		req128_c = 0;
+		req128_l = true;
+		req128_c = false;
 		swath_bytes_l = full_swath_bytes_packed_l/2;
 		swath_bytes_c = full_swath_bytes_packed_c;
 
 		total_swath_bytes = 2 * swath_bytes_l + 2 * swath_bytes_c;
 
 		if (total_swath_bytes > detile_buf_size_in_bytes) {
-			req128_c = 1;
+			req128_c = true;
 			swath_bytes_c = full_swath_bytes_packed_c/2;
 		}
 	}
@@ -1006,8 +1006,8 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 
 	double min_dst_y_ttu_vblank = 0;
 	unsigned int dlg_vblank_start = 0;
-	bool dual_plane = 0;
-	bool mode_422 = 0;
+	bool dual_plane = false;
+	bool mode_422 = false;
 	unsigned int access_dir = 0;
 	unsigned int vp_height_l = 0;
 	unsigned int vp_width_l = 0;
@@ -1021,7 +1021,7 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	double hratio_c = 0;
 	double vratio_l = 0;
 	double vratio_c = 0;
-	bool scl_enable = 0;
+	bool scl_enable = false;
 
 	double line_time_in_us = 0;
 	//	double vinit_l;
@@ -1156,7 +1156,7 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	// Source
 	//			 dcc_en			  = src.dcc;
 	dual_plane = is_dual_plane((enum source_format_class)(src->source_format));
-	mode_422 = 0; // TODO
+	mode_422 = false; // TODO
 	access_dir = (src->source_scan == dm_vert); // vp access direction: horizontal or vertical accessed
 	vp_height_l = src->viewport_height;
 	vp_width_l = src->viewport_width;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
index dd0c3b1780d7..0c5128187e08 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
@@ -297,6 +297,7 @@ struct _vcs_dpi_display_output_params_st {
 	int num_active_wb;
 	int output_bpc;
 	int output_type;
+	int is_virtual;
 	int output_format;
 	int dsc_slices;
 	int max_audio_sample_rate;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
index c9fbb33f05a3..bc0485a59018 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
@@ -451,6 +451,8 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 				dout->output_bpp;
 		mode_lib->vba.Output[mode_lib->vba.NumberOfActivePlanes] =
 				(enum output_encoder_class) (dout->output_type);
+		mode_lib->vba.skip_dio_check[mode_lib->vba.NumberOfActivePlanes] =
+				dout->is_virtual;
 
 		if (!dout->dsc_enable)
 			mode_lib->vba.ForcedOutputLinkBPP[mode_lib->vba.NumberOfActivePlanes] = dout->output_bpp;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
index 3529fedc4c52..025aa5bd8ea0 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
@@ -340,6 +340,7 @@ struct vba_vars_st {
 	unsigned int DSCInputBitPerComponent[DC__NUM_DPP__MAX];
 	enum output_format_class OutputFormat[DC__NUM_DPP__MAX];
 	enum output_encoder_class Output[DC__NUM_DPP__MAX];
+	bool skip_dio_check[DC__NUM_DPP__MAX];
 	unsigned int BlendingAndTiming[DC__NUM_DPP__MAX];
 	bool SynchronizedVBlank;
 	unsigned int NumberOfCursors[DC__NUM_DPP__MAX];
diff --git a/drivers/gpu/drm/amd/display/dc/dsc/Makefile b/drivers/gpu/drm/amd/display/dc/dsc/Makefile
index f2624a1156e5..8d31eb75c6a6 100644
--- a/drivers/gpu/drm/amd/display/dc/dsc/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dsc/Makefile
@@ -10,10 +10,6 @@ ifdef CONFIG_PPC64
 dsc_ccflags := -mhard-float -maltivec
 endif
 
-ifdef CONFIG_ARM64
-dsc_rcflags := -mgeneral-regs-only
-endif
-
 ifdef CONFIG_CC_IS_GCC
 ifeq ($(call cc-ifversion, -lt, 0701, y), y)
 IS_OLD_GCC = 1
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.c b/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.c
index df68430aeb0c..c6e28f6bf1a2 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.c
+++ b/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.c
@@ -28,6 +28,7 @@
  */
 
 #include "dm_services.h"
+#include "hw_factory_diag.h"
 #include "include/gpio_types.h"
 #include "../hw_factory.h"
 
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.h b/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.h
index 8a74f6adb8ee..bf68eb1d9a1d 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.h
+++ b/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.h
@@ -26,6 +26,8 @@
 #ifndef __DAL_HW_FACTORY_DIAG_FPGA_H__
 #define __DAL_HW_FACTORY_DIAG_FPGA_H__
 
+struct hw_factory;
+
 /* Initialize HW factory function pointers and pin info */
 void dal_hw_factory_diag_fpga_init(struct hw_factory *factory);
 
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_translate_diag.c b/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_translate_diag.c
index bf9068846927..e5138a5a8eb5 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_translate_diag.c
+++ b/drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_translate_diag.c
@@ -24,6 +24,7 @@
  */
 
 #include "dm_services.h"
+#include "hw_translate_diag.h"
 #include "include/gpio_types.h"
 
 #include "../hw_translate.h"
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/hw_ddc.c b/drivers/gpu/drm/amd/display/dc/gpio/hw_ddc.c
index 1ae153eab31d..7a8cec2d7a90 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/hw_ddc.c
+++ b/drivers/gpu/drm/amd/display/dc/gpio/hw_ddc.c
@@ -107,13 +107,12 @@ static enum gpio_result set_config(
 					msleep(3);
 			}
 		} else {
-			uint32_t reg2;
 			uint32_t sda_pd_dis = 0;
 			uint32_t scl_pd_dis = 0;
 
-			reg2 = REG_GET_2(gpio.MASK_reg,
-					DC_GPIO_SDA_PD_DIS, &sda_pd_dis,
-					DC_GPIO_SCL_PD_DIS, &scl_pd_dis);
+			REG_GET_2(gpio.MASK_reg,
+				  DC_GPIO_SDA_PD_DIS, &sda_pd_dis,
+				  DC_GPIO_SCL_PD_DIS, &scl_pd_dis);
 
 			if (sda_pd_dis) {
 				REG_SET(gpio.MASK_reg, regval,
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c b/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c
index da73bfb3cacd..92c65d2fa7d7 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c
+++ b/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c
@@ -119,17 +119,3 @@ bool dal_hw_factory_init(
 		return false;
 	}
 }
-
-void dal_hw_factory_destroy(
-	struct dc_context *ctx,
-	struct hw_factory **factory)
-{
-	if (!factory || !*factory) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	kfree(*factory);
-
-	*factory = NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
index 2d77eac66cb0..8efa1b80546d 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
@@ -333,6 +333,7 @@ struct pipe_ctx {
 	union pipe_update_flags update_flags;
 	struct dwbc *dwbc;
 	struct mcif_wb *mcif_wb;
+	bool vtp_locked;
 };
 
 struct resource_context {
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
index ffd37696b6b9..316301fc1e30 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
@@ -309,9 +309,9 @@ static inline bool should_set_clock(bool safe_to_lower, int calc_clk, int cur_cl
 static inline bool should_update_pstate_support(bool safe_to_lower, bool calc_support, bool cur_support)
 {
 	if (cur_support != calc_support) {
-		if (calc_support == true && safe_to_lower)
+		if (calc_support && safe_to_lower)
 			return true;
-		else if (calc_support == false && !safe_to_lower)
+		else if (!calc_support && !safe_to_lower)
 			return true;
 	}
 
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h
index 69d9fbfb4bec..cd1c0dc32bf8 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h
@@ -74,6 +74,16 @@ struct dmcu_funcs {
 	bool (*is_dmcu_initialized)(struct dmcu *dmcu);
 	bool (*lock_phy)(struct dmcu *dmcu);
 	bool (*unlock_phy)(struct dmcu *dmcu);
+	bool (*send_edid_cea)(struct dmcu *dmcu,
+			int offset,
+			int total_length,
+			uint8_t *data,
+			int length);
+	bool (*recv_amd_vsdb)(struct dmcu *dmcu,
+			int *version,
+			int *min_frame_rate,
+			int *max_frame_rate);
+	bool (*recv_edid_cea_ack)(struct dmcu *dmcu, int *offset);
 };
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h b/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
index 22f3f643ed1b..346dcd87dc10 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
@@ -191,6 +191,8 @@ struct hubp_funcs {
 	bool (*hubp_in_blank)(struct hubp *hubp);
 	void (*hubp_soft_reset)(struct hubp *hubp, bool reset);
 
+	void (*hubp_set_flip_int)(struct hubp *hubp);
+
 };
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index f7632fe25976..754832d216fd 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -190,6 +190,7 @@ struct timing_generator_funcs {
 	void (*set_blank)(struct timing_generator *tg,
 					bool enable_blanking);
 	bool (*is_blanked)(struct timing_generator *tg);
+	bool (*is_locked)(struct timing_generator *tg);
 	void (*set_overscan_blank_color) (struct timing_generator *tg, const struct tg_color *color);
 	void (*set_blank_color)(struct timing_generator *tg, const struct tg_color *color);
 	void (*set_colors)(struct timing_generator *tg,
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h
index 62804dc7b698..0586ab2ffd6a 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h
@@ -54,6 +54,7 @@ struct hw_sequencer_funcs {
 	/* Embedded Display Related */
 	void (*edp_power_control)(struct dc_link *link, bool enable);
 	void (*edp_wait_for_hpd_ready)(struct dc_link *link, bool power_up);
+	void (*edp_wait_for_T12)(struct dc_link *link);
 
 	/* Pipe Programming Related */
 	void (*init_hw)(struct dc *dc);
@@ -217,6 +218,9 @@ struct hw_sequencer_funcs {
 	/* Idle Optimization Related */
 	bool (*apply_idle_power_optimizations)(struct dc *dc, bool enable);
 
+	bool (*does_plane_fit_in_mall)(struct dc *dc, struct dc_plane_state *plane,
+			struct dc_cursor_attributes *cursor_attr);
+
 	bool (*is_abm_supported)(struct dc *dc,
 			struct dc_state *context, struct dc_stream_state *stream);
 
@@ -227,6 +231,10 @@ struct hw_sequencer_funcs {
 			enum dc_color_depth color_depth,
 			const struct tg_color *solid_color,
 			int width, int height, int offset);
+
+	void (*set_hubp_blank)(const struct dc *dc,
+			struct pipe_ctx *pipe_ctx,
+			bool blank_enable);
 };
 
 void color_space_to_black_color(
diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c b/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c
index f956b3bde680..34f43cb650f8 100644
--- a/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c
+++ b/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c
@@ -58,6 +58,18 @@ enum dc_irq_source to_dal_irq_source_dcn10(
 		return DC_IRQ_SOURCE_VBLANK5;
 	case DCN_1_0__SRCID__DC_D6_OTG_VSTARTUP:
 		return DC_IRQ_SOURCE_VBLANK6;
+	case DCN_1_0__SRCID__OTG1_VERTICAL_INTERRUPT0_CONTROL:
+		return DC_IRQ_SOURCE_DC1_VLINE0;
+	case DCN_1_0__SRCID__OTG2_VERTICAL_INTERRUPT0_CONTROL:
+		return DC_IRQ_SOURCE_DC2_VLINE0;
+	case DCN_1_0__SRCID__OTG3_VERTICAL_INTERRUPT0_CONTROL:
+		return DC_IRQ_SOURCE_DC3_VLINE0;
+	case DCN_1_0__SRCID__OTG4_VERTICAL_INTERRUPT0_CONTROL:
+		return DC_IRQ_SOURCE_DC4_VLINE0;
+	case DCN_1_0__SRCID__OTG5_VERTICAL_INTERRUPT0_CONTROL:
+		return DC_IRQ_SOURCE_DC5_VLINE0;
+	case DCN_1_0__SRCID__OTG6_VERTICAL_INTERRUPT0_CONTROL:
+		return DC_IRQ_SOURCE_DC6_VLINE0;
 	case DCN_1_0__SRCID__OTG0_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
 		return DC_IRQ_SOURCE_VUPDATE1;
 	case DCN_1_0__SRCID__OTG1_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
@@ -167,6 +179,11 @@ static const struct irq_source_info_funcs vblank_irq_info_funcs = {
 	.ack = NULL
 };
 
+static const struct irq_source_info_funcs vline0_irq_info_funcs = {
+	.set = NULL,
+	.ack = NULL
+};
+
 static const struct irq_source_info_funcs vupdate_no_lock_irq_info_funcs = {
 	.set = NULL,
 	.ack = NULL
@@ -241,6 +258,14 @@ static const struct irq_source_info_funcs vupdate_no_lock_irq_info_funcs = {
 		.funcs = &vblank_irq_info_funcs\
 	}
 
+#define vline0_int_entry(reg_num)\
+	[DC_IRQ_SOURCE_DC1_VLINE0 + reg_num] = {\
+		IRQ_REG_ENTRY(OTG, reg_num,\
+			OTG_VERTICAL_INTERRUPT0_CONTROL, OTG_VERTICAL_INTERRUPT0_INT_ENABLE,\
+			OTG_VERTICAL_INTERRUPT0_CONTROL, OTG_VERTICAL_INTERRUPT0_CLEAR),\
+		.funcs = &vline0_irq_info_funcs\
+	}
+
 #define dummy_irq_entry() \
 	{\
 		.funcs = &dummy_irq_info_funcs\
@@ -349,6 +374,12 @@ irq_source_info_dcn10[DAL_IRQ_SOURCES_NUMBER] = {
 	vblank_int_entry(3),
 	vblank_int_entry(4),
 	vblank_int_entry(5),
+	vline0_int_entry(0),
+	vline0_int_entry(1),
+	vline0_int_entry(2),
+	vline0_int_entry(3),
+	vline0_int_entry(4),
+	vline0_int_entry(5),
 };
 
 static const struct irq_service_funcs irq_service_funcs_dcn10 = {
diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c b/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c
index 1b971265418b..0e0f494fbb5e 100644
--- a/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c
+++ b/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c
@@ -168,6 +168,11 @@ static const struct irq_source_info_funcs vblank_irq_info_funcs = {
 	.ack = NULL
 };
 
+static const struct irq_source_info_funcs vupdate_no_lock_irq_info_funcs = {
+	.set = NULL,
+	.ack = NULL
+};
+
 #undef BASE_INNER
 #define BASE_INNER(seg) DMU_BASE__INST0_SEG ## seg
 
@@ -230,6 +235,17 @@ static const struct irq_source_info_funcs vblank_irq_info_funcs = {
 		.funcs = &vblank_irq_info_funcs\
 	}
 
+/* vupdate_no_lock_int_entry maps to DC_IRQ_SOURCE_VUPDATEx, to match semantic
+ * of DCE's DC_IRQ_SOURCE_VUPDATEx.
+ */
+#define vupdate_no_lock_int_entry(reg_num)\
+	[DC_IRQ_SOURCE_VUPDATE1 + reg_num] = {\
+		IRQ_REG_ENTRY(OTG, reg_num,\
+			OTG_GLOBAL_SYNC_STATUS, VUPDATE_NO_LOCK_INT_EN,\
+			OTG_GLOBAL_SYNC_STATUS, VUPDATE_NO_LOCK_EVENT_CLEAR),\
+		.funcs = &vupdate_no_lock_irq_info_funcs\
+	}
+
 #define vblank_int_entry(reg_num)\
 	[DC_IRQ_SOURCE_VBLANK1 + reg_num] = {\
 		IRQ_REG_ENTRY(OTG, reg_num,\
@@ -338,6 +354,12 @@ irq_source_info_dcn21[DAL_IRQ_SOURCES_NUMBER] = {
 	vupdate_int_entry(3),
 	vupdate_int_entry(4),
 	vupdate_int_entry(5),
+	vupdate_no_lock_int_entry(0),
+	vupdate_no_lock_int_entry(1),
+	vupdate_no_lock_int_entry(2),
+	vupdate_no_lock_int_entry(3),
+	vupdate_no_lock_int_entry(4),
+	vupdate_no_lock_int_entry(5),
 	vblank_int_entry(0),
 	vblank_int_entry(1),
 	vblank_int_entry(2),
diff --git a/drivers/gpu/drm/amd/display/dc/irq/irq_service.c b/drivers/gpu/drm/amd/display/dc/irq/irq_service.c
index 6bf27bde8724..5f245bde54ff 100644
--- a/drivers/gpu/drm/amd/display/dc/irq/irq_service.c
+++ b/drivers/gpu/drm/amd/display/dc/irq/irq_service.c
@@ -79,7 +79,7 @@ void dal_irq_service_destroy(struct irq_service **irq_service)
 	*irq_service = NULL;
 }
 
-const struct irq_source_info *find_irq_source_info(
+static const struct irq_source_info *find_irq_source_info(
 	struct irq_service *irq_service,
 	enum dc_irq_source source)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/irq_types.h b/drivers/gpu/drm/amd/display/dc/irq_types.h
index d0ccd81ad5b4..87812d81fed3 100644
--- a/drivers/gpu/drm/amd/display/dc/irq_types.h
+++ b/drivers/gpu/drm/amd/display/dc/irq_types.h
@@ -160,6 +160,7 @@ enum irq_type
 	IRQ_TYPE_PFLIP = DC_IRQ_SOURCE_PFLIP1,
 	IRQ_TYPE_VUPDATE = DC_IRQ_SOURCE_VUPDATE1,
 	IRQ_TYPE_VBLANK = DC_IRQ_SOURCE_VBLANK1,
+	IRQ_TYPE_VLINE0 = DC_IRQ_SOURCE_DC1_VLINE0,
 };
 
 #define DAL_VALID_IRQ_SRC_NUM(src) \
diff --git a/drivers/gpu/drm/amd/display/dc/os_types.h b/drivers/gpu/drm/amd/display/dc/os_types.h
index 95cb56929e79..126c2f3a4dd3 100644
--- a/drivers/gpu/drm/amd/display/dc/os_types.h
+++ b/drivers/gpu/drm/amd/display/dc/os_types.h
@@ -55,10 +55,6 @@
 #include <asm/fpu/api.h>
 #define DC_FP_START() kernel_fpu_begin()
 #define DC_FP_END() kernel_fpu_end()
-#elif defined(CONFIG_ARM64)
-#include <asm/neon.h>
-#define DC_FP_START() kernel_neon_begin()
-#define DC_FP_END() kernel_neon_end()
 #elif defined(CONFIG_PPC64)
 #include <asm/switch_to.h>
 #include <asm/cputable.h>
diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
index 249a076d6f69..072b4e7e624b 100644
--- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
+++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
@@ -47,10 +47,10 @@
 
 /* Firmware versioning. */
 #ifdef DMUB_EXPOSE_VERSION
-#define DMUB_FW_VERSION_GIT_HASH 0xf51b86a
+#define DMUB_FW_VERSION_GIT_HASH 0x6444c02e7
 #define DMUB_FW_VERSION_MAJOR 0
 #define DMUB_FW_VERSION_MINOR 0
-#define DMUB_FW_VERSION_REVISION 47
+#define DMUB_FW_VERSION_REVISION 51
 #define DMUB_FW_VERSION_TEST 0
 #define DMUB_FW_VERSION_VBIOS 0
 #define DMUB_FW_VERSION_HOTFIX 0
@@ -458,6 +458,10 @@ struct dmub_rb_cmd_mall {
 	uint16_t cursor_pitch;
 	uint16_t cursor_height;
 	uint8_t cursor_bpp;
+	uint8_t debug_bits;
+
+	uint8_t reserved1;
+	uint8_t reserved2;
 };
 
 struct dmub_cmd_digx_encoder_control_data {
@@ -487,13 +491,34 @@ struct dmub_rb_cmd_enable_disp_power_gating {
 	struct dmub_cmd_enable_disp_power_gating_data power_gating;
 };
 
-struct dmub_cmd_dig1_transmitter_control_data {
+struct dmub_dig_transmitter_control_data_v1_7 {
+	uint8_t phyid; /**< 0=UNIPHYA, 1=UNIPHYB, 2=UNIPHYC, 3=UNIPHYD, 4=UNIPHYE, 5=UNIPHYF */
+	uint8_t action; /**< Defined as ATOM_TRANSMITER_ACTION_xxx */
+	union {
+		uint8_t digmode; /**< enum atom_encode_mode_def */
+		uint8_t dplaneset; /**< DP voltage swing and pre-emphasis value, "DP_LANE_SET__xDB_y_zV" */
+	} mode_laneset;
+	uint8_t lanenum; /**< Number of lanes */
+	union {
+		uint32_t symclk_10khz; /**< Symbol Clock in 10Khz */
+	} symclk_units;
+	uint8_t hpdsel; /**< =1: HPD1, =2: HPD2, ..., =6: HPD6, =0: HPD is not assigned */
+	uint8_t digfe_sel; /**< DIG front-end selection, bit0 means DIG0 FE is enabled */
+	uint8_t connobj_id; /**< Connector Object Id defined in ObjectId.h */
+	uint8_t reserved0; /**< For future use */
+	uint8_t reserved1; /**< For future use */
+	uint8_t reserved2[3]; /**< For future use */
+	uint32_t reserved3[11]; /**< For future use */
+};
+
+union dmub_cmd_dig1_transmitter_control_data {
 	struct dig_transmitter_control_parameters_v1_6 dig;
+	struct dmub_dig_transmitter_control_data_v1_7 dig_v1_7;
 };
 
 struct dmub_rb_cmd_dig1_transmitter_control {
 	struct dmub_cmd_header header;
-	struct dmub_cmd_dig1_transmitter_control_data transmitter_control;
+	union dmub_cmd_dig1_transmitter_control_data transmitter_control;
 };
 
 struct dmub_rb_cmd_dpphy_init {
@@ -624,6 +649,7 @@ enum dmub_cmd_mall_type {
 	DMUB_CMD__MALL_ACTION_ALLOW = 0,
 	DMUB_CMD__MALL_ACTION_DISALLOW = 1,
 	DMUB_CMD__MALL_ACTION_COPY_CURSOR = 2,
+	DMUB_CMD__MALL_ACTION_NO_DF_REQ = 3,
 };
 
 struct dmub_cmd_psr_copy_settings_data {
@@ -648,6 +674,7 @@ struct dmub_cmd_psr_copy_settings_data {
 	uint8_t multi_disp_optimizations_en;
 	uint16_t init_sdp_deadline;
 	uint16_t pad2;
+	uint32_t line_time_in_us;
 };
 
 struct dmub_rb_cmd_psr_copy_settings {
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c
index cafba1d23c6a..8e8e65fa83c0 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c
@@ -81,6 +81,13 @@ static inline void dmub_dcn20_translate_addr(const union dmub_addr *addr_in,
 	addr_out->quad_part = addr_in->quad_part - fb_base + fb_offset;
 }
 
+bool dmub_dcn20_use_cached_inbox(struct dmub_srv *dmub)
+{
+	/* Cached inbox is not supported in this fw version range */
+	return !(dmub->fw_version >= DMUB_FW_VERSION(1, 0, 0) &&
+		 dmub->fw_version <= DMUB_FW_VERSION(1, 10, 0));
+}
+
 void dmub_dcn20_reset(struct dmub_srv *dmub)
 {
 	union dmub_gpint_data_register cmd;
@@ -216,7 +223,7 @@ void dmub_dcn20_setup_windows(struct dmub_srv *dmub,
 	dmub_dcn20_translate_addr(&cw4->offset, fb_base, fb_offset, &offset);
 
 	/* New firmware can support CW4. */
-	if (dmub->fw_version > DMUB_FW_VERSION(1, 0, 10)) {
+	if (dmub_dcn20_use_cached_inbox(dmub)) {
 		REG_WRITE(DMCUB_REGION3_CW4_OFFSET, offset.u.low_part);
 		REG_WRITE(DMCUB_REGION3_CW4_OFFSET_HIGH, offset.u.high_part);
 		REG_WRITE(DMCUB_REGION3_CW4_BASE_ADDRESS, cw4->region.base);
@@ -255,7 +262,7 @@ void dmub_dcn20_setup_mailbox(struct dmub_srv *dmub,
 			      const struct dmub_region *inbox1)
 {
 	/* New firmware can support CW4 for the inbox. */
-	if (dmub->fw_version > DMUB_FW_VERSION(1, 0, 10))
+	if (dmub_dcn20_use_cached_inbox(dmub))
 		REG_WRITE(DMCUB_INBOX1_BASE_ADDRESS, inbox1->base);
 	else
 		REG_WRITE(DMCUB_INBOX1_BASE_ADDRESS, 0x80000000);
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h
index d438f365cbb0..a62be9c0652e 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h
@@ -198,4 +198,6 @@ void dmub_dcn20_skip_dmub_panel_power_sequence(struct dmub_srv *dmub, bool skip)
 
 union dmub_fw_boot_status dmub_dcn20_get_fw_boot_status(struct dmub_srv *dmub);
 
+bool dmub_dcn20_use_cached_inbox(struct dmub_srv *dmub);
+
 #endif /* _DMUB_DCN20_H_ */
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn30.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn30.c
index f00df02ded81..b4bc0df2f14a 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn30.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn30.c
@@ -26,6 +26,7 @@
 #include "../dmub_srv.h"
 #include "dmub_reg.h"
 #include "dmub_dcn20.h"
+#include "dmub_dcn30.h"
 
 #include "sienna_cichlid_ip_offset.h"
 #include "dcn/dcn_3_0_0_offset.h"
@@ -154,7 +155,7 @@ void dmub_dcn30_setup_windows(struct dmub_srv *dmub,
 	offset = cw4->offset;
 
 	/* New firmware can support CW4. */
-	if (dmub->fw_version > DMUB_FW_VERSION(1, 0, 10)) {
+	if (dmub_dcn20_use_cached_inbox(dmub)) {
 		REG_WRITE(DMCUB_REGION3_CW4_OFFSET, offset.u.low_part);
 		REG_WRITE(DMCUB_REGION3_CW4_OFFSET_HIGH, offset.u.high_part);
 		REG_WRITE(DMCUB_REGION3_CW4_BASE_ADDRESS, cw4->region.base);
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c
index f388d36af0b6..61f64a295f06 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c
@@ -406,6 +406,9 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 	dmub->fb_offset = params->fb_offset;
 	dmub->psp_version = params->psp_version;
 
+	if (dmub->hw_funcs.reset)
+		dmub->hw_funcs.reset(dmub);
+
 	if (inst_fb && data_fb) {
 		cw0.offset.quad_part = inst_fb->gpu_addr;
 		cw0.region.base = DMUB_CW0_BASE;
@@ -427,9 +430,6 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 
 	}
 
-	if (dmub->hw_funcs.reset)
-		dmub->hw_funcs.reset(dmub);
-
 	if (inst_fb && data_fb && bios_fb && mail_fb && tracebuff_fb &&
 	    fw_state_fb && scratch_mem_fb) {
 		cw2.offset.quad_part = data_fb->gpu_addr;
@@ -489,9 +489,6 @@ enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub)
 	if (!dmub->sw_init)
 		return DMUB_STATUS_INVALID;
 
-	if (dmub->hw_init == false)
-		return DMUB_STATUS_OK;
-
 	if (dmub->hw_funcs.reset)
 		dmub->hw_funcs.reset(dmub);
 
diff --git a/drivers/gpu/drm/amd/display/modules/color/color_table.c b/drivers/gpu/drm/amd/display/modules/color/color_table.c
index 692e536e7d05..410f2a82b9a2 100644
--- a/drivers/gpu/drm/amd/display/modules/color/color_table.c
+++ b/drivers/gpu/drm/amd/display/modules/color/color_table.c
@@ -1,10 +1,26 @@
 /*
- * Copyright (c) 2019 Advanced Micro Devices, Inc. (unpublished)
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
  *
- * All rights reserved.  This notice is intended as a precaution against
- * inadvertent publication and does not imply publication or any waiver
- * of confidentiality.  The year included in the foregoing notice is the
- * year of creation of the work.
  */
 
 #include "color_table.h"
diff --git a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h
index 6c678cfb82e3..5c22cf7e6118 100644
--- a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h
+++ b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h
@@ -397,7 +397,7 @@ static inline uint8_t is_dp_hdcp(struct mod_hdcp *hdcp)
 static inline uint8_t is_dp_mst_hdcp(struct mod_hdcp *hdcp)
 {
 	return (hdcp->connection.link.mode == MOD_HDCP_MODE_DP &&
-			hdcp->connection.link.dp.mst_supported);
+			hdcp->connection.link.dp.mst_enabled);
 }
 
 static inline uint8_t is_hdmi_dvi_sl_hdcp(struct mod_hdcp *hdcp)
diff --git a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_psp.c b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_psp.c
index 3a367a5968ae..904ce9b88088 100644
--- a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_psp.c
+++ b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_psp.c
@@ -106,7 +106,7 @@ enum mod_hdcp_status mod_hdcp_add_display_to_topology(struct mod_hdcp *hdcp,
 	dtm_cmd->dtm_in_message.topology_update_v2.dig_be = link->dig_be;
 	dtm_cmd->dtm_in_message.topology_update_v2.dig_fe = display->dig_fe;
 	if (is_dp_hdcp(hdcp))
-		dtm_cmd->dtm_in_message.topology_update_v2.is_assr = link->dp.assr_supported;
+		dtm_cmd->dtm_in_message.topology_update_v2.is_assr = link->dp.assr_enabled;
 
 	dtm_cmd->dtm_in_message.topology_update_v2.dp_mst_vcid = display->vc_id;
 	dtm_cmd->dtm_in_message.topology_update_v2.max_hdcp_supported_version =
@@ -548,6 +548,8 @@ enum mod_hdcp_status mod_hdcp_hdcp2_validate_ake_cert(struct mod_hdcp *hdcp)
 			   TA_HDCP2_MSG_AUTHENTICATION_STATUS__RECEIVERID_REVOKED) {
 			hdcp->connection.is_hdcp2_revoked = 1;
 			status = MOD_HDCP_STATUS_HDCP2_AKE_CERT_REVOKED;
+		}  else {
+			status = MOD_HDCP_STATUS_HDCP2_VALIDATE_AKE_CERT_FAILURE;
 		}
 	}
 	mutex_unlock(&psp->hdcp_context.mutex);
diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h b/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h
index eed560eecbab..d223ed3be5d3 100644
--- a/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h
+++ b/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h
@@ -101,8 +101,8 @@ enum mod_hdcp_status {
 
 struct mod_hdcp_displayport {
 	uint8_t rev;
-	uint8_t assr_supported;
-	uint8_t mst_supported;
+	uint8_t assr_enabled;
+	uint8_t mst_enabled;
 };
 
 struct mod_hdcp_hdmi {
diff --git a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c
index 0fdf7a3e96de..57f198de5e2c 100644
--- a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c
+++ b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c
@@ -409,16 +409,11 @@ void mod_build_vsc_infopacket(const struct dc_stream_state *stream,
 }
 
 /**
- *****************************************************************************
- *  Function: mod_build_hf_vsif_infopacket
+ *  mod_build_hf_vsif_infopacket - Prepare HDMI Vendor Specific info frame.
+ *                                 Follows HDMI Spec to build up Vendor Specific info frame
  *
- *  @brief
- *     Prepare HDMI Vendor Specific info frame.
- *     Follows HDMI Spec to build up Vendor Specific info frame
- *
- *  @param [in] stream: contains data we may need to construct VSIF (i.e. timing_3d_format, etc.)
- *  @param [out] info_packet:   output structure where to store VSIF
- *****************************************************************************
+ *  @stream:      contains data we may need to construct VSIF (i.e. timing_3d_format, etc.)
+ *  @info_packet: output structure where to store VSIF
  */
 void mod_build_hf_vsif_infopacket(const struct dc_stream_state *stream,
 		struct dc_info_packet *info_packet)
diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
index 4fd8bce95d84..6270ecbd2438 100644
--- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
+++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
@@ -266,7 +266,7 @@ static void fill_backlight_transform_table_v_2_2(struct dmcu_iram_parameters par
 	 * format U4.10.
 	 */
 	for (i = 1; i+1 < num_entries; i++) {
-		lut_index = (params.backlight_lut_array_size - 1) * i / (num_entries - 1);
+		lut_index = DIV_ROUNDUP((i * params.backlight_lut_array_size), num_entries);
 		ASSERT(lut_index < params.backlight_lut_array_size);
 
 		table->backlight_thresholds[i] = (big_endian) ?
@@ -278,7 +278,7 @@ static void fill_backlight_transform_table_v_2_2(struct dmcu_iram_parameters par
 	}
 }
 
-void fill_iram_v_2(struct iram_table_v_2 *ram_table, struct dmcu_iram_parameters params)
+static void fill_iram_v_2(struct iram_table_v_2 *ram_table, struct dmcu_iram_parameters params)
 {
 	unsigned int set = params.set;
 
@@ -452,7 +452,7 @@ void fill_iram_v_2(struct iram_table_v_2 *ram_table, struct dmcu_iram_parameters
 			params, ram_table);
 }
 
-void fill_iram_v_2_2(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parameters params)
+static void fill_iram_v_2_2(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parameters params)
 {
 	unsigned int set = params.set;
 
@@ -598,7 +598,7 @@ void fill_iram_v_2_2(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parame
 			params, ram_table, true);
 }
 
-void fill_iram_v_2_3(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parameters params, bool big_endian)
+static void fill_iram_v_2_3(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parameters params, bool big_endian)
 {
 	unsigned int i, j;
 	unsigned int set = params.set;
diff --git a/drivers/gpu/drm/amd/include/amd_pcie.h b/drivers/gpu/drm/amd/include/amd_pcie.h
index 9cb9ceb4d74d..a1ece3eecdf5 100644
--- a/drivers/gpu/drm/amd/include/amd_pcie.h
+++ b/drivers/gpu/drm/amd/include/amd_pcie.h
@@ -28,6 +28,7 @@
 #define CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2        0x00020000
 #define CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3        0x00040000
 #define CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4        0x00080000
+#define CAIL_PCIE_LINK_SPEED_SUPPORT_GEN5        0x00100000
 #define CAIL_PCIE_LINK_SPEED_SUPPORT_MASK        0xFFFF0000
 #define CAIL_PCIE_LINK_SPEED_SUPPORT_SHIFT       16
 
@@ -36,6 +37,7 @@
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2   0x00000002
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3   0x00000004
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN4   0x00000008
+#define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN5   0x00000010
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_MASK   0x0000FFFF
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_SHIFT  0
 
diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h
index 9676016a37ce..43ed6291b2b8 100644
--- a/drivers/gpu/drm/amd/include/amd_shared.h
+++ b/drivers/gpu/drm/amd/include/amd_shared.h
@@ -213,6 +213,7 @@ enum PP_FEATURE_MASK {
 	PP_ACG_MASK = 0x10000,
 	PP_STUTTER_MODE = 0x20000,
 	PP_AVFS_MASK = 0x40000,
+	PP_GFX_DCS_MASK = 0x80000,
 };
 
 enum DC_FEATURE_MASK {
diff --git a/drivers/gpu/drm/amd/include/asic_reg/oss/osssys_4_2_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/oss/osssys_4_2_0_offset.h
new file mode 100644
index 000000000000..bd129266ebfd
--- /dev/null
+++ b/drivers/gpu/drm/amd/include/asic_reg/oss/osssys_4_2_0_offset.h
@@ -0,0 +1,345 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef _osssys_4_2_0_OFFSET_HEADER
+#define _osssys_4_2_0_OFFSET_HEADER
+
+
+
+// addressBlock: osssys_osssysdec
+// base address: 0x4280
+#define mmIH_VMID_0_LUT                                                                                0x0000
+#define mmIH_VMID_0_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_1_LUT                                                                                0x0001
+#define mmIH_VMID_1_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_2_LUT                                                                                0x0002
+#define mmIH_VMID_2_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_3_LUT                                                                                0x0003
+#define mmIH_VMID_3_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_4_LUT                                                                                0x0004
+#define mmIH_VMID_4_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_5_LUT                                                                                0x0005
+#define mmIH_VMID_5_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_6_LUT                                                                                0x0006
+#define mmIH_VMID_6_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_7_LUT                                                                                0x0007
+#define mmIH_VMID_7_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_8_LUT                                                                                0x0008
+#define mmIH_VMID_8_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_9_LUT                                                                                0x0009
+#define mmIH_VMID_9_LUT_BASE_IDX                                                                       0
+#define mmIH_VMID_10_LUT                                                                               0x000a
+#define mmIH_VMID_10_LUT_BASE_IDX                                                                      0
+#define mmIH_VMID_11_LUT                                                                               0x000b
+#define mmIH_VMID_11_LUT_BASE_IDX                                                                      0
+#define mmIH_VMID_12_LUT                                                                               0x000c
+#define mmIH_VMID_12_LUT_BASE_IDX                                                                      0
+#define mmIH_VMID_13_LUT                                                                               0x000d
+#define mmIH_VMID_13_LUT_BASE_IDX                                                                      0
+#define mmIH_VMID_14_LUT                                                                               0x000e
+#define mmIH_VMID_14_LUT_BASE_IDX                                                                      0
+#define mmIH_VMID_15_LUT                                                                               0x000f
+#define mmIH_VMID_15_LUT_BASE_IDX                                                                      0
+#define mmIH_VMID_0_LUT_MM                                                                             0x0010
+#define mmIH_VMID_0_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_1_LUT_MM                                                                             0x0011
+#define mmIH_VMID_1_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_2_LUT_MM                                                                             0x0012
+#define mmIH_VMID_2_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_3_LUT_MM                                                                             0x0013
+#define mmIH_VMID_3_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_4_LUT_MM                                                                             0x0014
+#define mmIH_VMID_4_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_5_LUT_MM                                                                             0x0015
+#define mmIH_VMID_5_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_6_LUT_MM                                                                             0x0016
+#define mmIH_VMID_6_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_7_LUT_MM                                                                             0x0017
+#define mmIH_VMID_7_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_8_LUT_MM                                                                             0x0018
+#define mmIH_VMID_8_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_9_LUT_MM                                                                             0x0019
+#define mmIH_VMID_9_LUT_MM_BASE_IDX                                                                    0
+#define mmIH_VMID_10_LUT_MM                                                                            0x001a
+#define mmIH_VMID_10_LUT_MM_BASE_IDX                                                                   0
+#define mmIH_VMID_11_LUT_MM                                                                            0x001b
+#define mmIH_VMID_11_LUT_MM_BASE_IDX                                                                   0
+#define mmIH_VMID_12_LUT_MM                                                                            0x001c
+#define mmIH_VMID_12_LUT_MM_BASE_IDX                                                                   0
+#define mmIH_VMID_13_LUT_MM                                                                            0x001d
+#define mmIH_VMID_13_LUT_MM_BASE_IDX                                                                   0
+#define mmIH_VMID_14_LUT_MM                                                                            0x001e
+#define mmIH_VMID_14_LUT_MM_BASE_IDX                                                                   0
+#define mmIH_VMID_15_LUT_MM                                                                            0x001f
+#define mmIH_VMID_15_LUT_MM_BASE_IDX                                                                   0
+#define mmIH_COOKIE_0                                                                                  0x0020
+#define mmIH_COOKIE_0_BASE_IDX                                                                         0
+#define mmIH_COOKIE_1                                                                                  0x0021
+#define mmIH_COOKIE_1_BASE_IDX                                                                         0
+#define mmIH_COOKIE_2                                                                                  0x0022
+#define mmIH_COOKIE_2_BASE_IDX                                                                         0
+#define mmIH_COOKIE_3                                                                                  0x0023
+#define mmIH_COOKIE_3_BASE_IDX                                                                         0
+#define mmIH_COOKIE_4                                                                                  0x0024
+#define mmIH_COOKIE_4_BASE_IDX                                                                         0
+#define mmIH_COOKIE_5                                                                                  0x0025
+#define mmIH_COOKIE_5_BASE_IDX                                                                         0
+#define mmIH_COOKIE_6                                                                                  0x0026
+#define mmIH_COOKIE_6_BASE_IDX                                                                         0
+#define mmIH_COOKIE_7                                                                                  0x0027
+#define mmIH_COOKIE_7_BASE_IDX                                                                         0
+#define mmIH_REGISTER_LAST_PART0                                                                       0x003f
+#define mmIH_REGISTER_LAST_PART0_BASE_IDX                                                              0
+#define mmSEM_REQ_INPUT_0                                                                              0x0040
+#define mmSEM_REQ_INPUT_0_BASE_IDX                                                                     0
+#define mmSEM_REQ_INPUT_1                                                                              0x0041
+#define mmSEM_REQ_INPUT_1_BASE_IDX                                                                     0
+#define mmSEM_REQ_INPUT_2                                                                              0x0042
+#define mmSEM_REQ_INPUT_2_BASE_IDX                                                                     0
+#define mmSEM_REQ_INPUT_3                                                                              0x0043
+#define mmSEM_REQ_INPUT_3_BASE_IDX                                                                     0
+#define mmSEM_REGISTER_LAST_PART0                                                                      0x007f
+#define mmSEM_REGISTER_LAST_PART0_BASE_IDX                                                             0
+#define mmIH_RB_CNTL                                                                                   0x0080
+#define mmIH_RB_CNTL_BASE_IDX                                                                          0
+#define mmIH_RB_BASE                                                                                   0x0081
+#define mmIH_RB_BASE_BASE_IDX                                                                          0
+#define mmIH_RB_BASE_HI                                                                                0x0082
+#define mmIH_RB_BASE_HI_BASE_IDX                                                                       0
+#define mmIH_RB_RPTR                                                                                   0x0083
+#define mmIH_RB_RPTR_BASE_IDX                                                                          0
+#define mmIH_RB_WPTR                                                                                   0x0084
+#define mmIH_RB_WPTR_BASE_IDX                                                                          0
+#define mmIH_RB_WPTR_ADDR_HI                                                                           0x0085
+#define mmIH_RB_WPTR_ADDR_HI_BASE_IDX                                                                  0
+#define mmIH_RB_WPTR_ADDR_LO                                                                           0x0086
+#define mmIH_RB_WPTR_ADDR_LO_BASE_IDX                                                                  0
+#define mmIH_DOORBELL_RPTR                                                                             0x0087
+#define mmIH_DOORBELL_RPTR_BASE_IDX                                                                    0
+#define mmIH_RB_CNTL_RING1                                                                             0x008c
+#define mmIH_RB_CNTL_RING1_BASE_IDX                                                                    0
+#define mmIH_RB_BASE_RING1                                                                             0x008d
+#define mmIH_RB_BASE_RING1_BASE_IDX                                                                    0
+#define mmIH_RB_BASE_HI_RING1                                                                          0x008e
+#define mmIH_RB_BASE_HI_RING1_BASE_IDX                                                                 0
+#define mmIH_RB_RPTR_RING1                                                                             0x008f
+#define mmIH_RB_RPTR_RING1_BASE_IDX                                                                    0
+#define mmIH_RB_WPTR_RING1                                                                             0x0090
+#define mmIH_RB_WPTR_RING1_BASE_IDX                                                                    0
+#define mmIH_DOORBELL_RPTR_RING1                                                                       0x0093
+#define mmIH_DOORBELL_RPTR_RING1_BASE_IDX                                                              0
+#define mmIH_RB_CNTL_RING2                                                                             0x0098
+#define mmIH_RB_CNTL_RING2_BASE_IDX                                                                    0
+#define mmIH_RB_BASE_RING2                                                                             0x0099
+#define mmIH_RB_BASE_RING2_BASE_IDX                                                                    0
+#define mmIH_RB_BASE_HI_RING2                                                                          0x009a
+#define mmIH_RB_BASE_HI_RING2_BASE_IDX                                                                 0
+#define mmIH_RB_RPTR_RING2                                                                             0x009b
+#define mmIH_RB_RPTR_RING2_BASE_IDX                                                                    0
+#define mmIH_RB_WPTR_RING2                                                                             0x009c
+#define mmIH_RB_WPTR_RING2_BASE_IDX                                                                    0
+#define mmIH_DOORBELL_RPTR_RING2                                                                       0x009f
+#define mmIH_DOORBELL_RPTR_RING2_BASE_IDX                                                              0
+#define mmIH_VERSION                                                                                   0x00a5
+#define mmIH_VERSION_BASE_IDX                                                                          0
+#define mmIH_CNTL                                                                                      0x00c0
+#define mmIH_CNTL_BASE_IDX                                                                             0
+#define mmIH_CNTL2                                                                                     0x00c1
+#define mmIH_CNTL2_BASE_IDX                                                                            0
+#define mmIH_STATUS                                                                                    0x00c2
+#define mmIH_STATUS_BASE_IDX                                                                           0
+#define mmIH_PERFMON_CNTL                                                                              0x00c3
+#define mmIH_PERFMON_CNTL_BASE_IDX                                                                     0
+#define mmIH_PERFCOUNTER0_RESULT                                                                       0x00c4
+#define mmIH_PERFCOUNTER0_RESULT_BASE_IDX                                                              0
+#define mmIH_PERFCOUNTER1_RESULT                                                                       0x00c5
+#define mmIH_PERFCOUNTER1_RESULT_BASE_IDX                                                              0
+#define mmIH_DSM_MATCH_VALUE_BIT_31_0                                                                  0x00c7
+#define mmIH_DSM_MATCH_VALUE_BIT_31_0_BASE_IDX                                                         0
+#define mmIH_DSM_MATCH_VALUE_BIT_63_32                                                                 0x00c8
+#define mmIH_DSM_MATCH_VALUE_BIT_63_32_BASE_IDX                                                        0
+#define mmIH_DSM_MATCH_VALUE_BIT_95_64                                                                 0x00c9
+#define mmIH_DSM_MATCH_VALUE_BIT_95_64_BASE_IDX                                                        0
+#define mmIH_DSM_MATCH_FIELD_CONTROL                                                                   0x00ca
+#define mmIH_DSM_MATCH_FIELD_CONTROL_BASE_IDX                                                          0
+#define mmIH_DSM_MATCH_DATA_CONTROL                                                                    0x00cb
+#define mmIH_DSM_MATCH_DATA_CONTROL_BASE_IDX                                                           0
+#define mmIH_DSM_MATCH_FCN_ID                                                                          0x00cc
+#define mmIH_DSM_MATCH_FCN_ID_BASE_IDX                                                                 0
+#define mmIH_LIMIT_INT_RATE_CNTL                                                                       0x00cd
+#define mmIH_LIMIT_INT_RATE_CNTL_BASE_IDX                                                              0
+#define mmIH_VF_RB_STATUS                                                                              0x00ce
+#define mmIH_VF_RB_STATUS_BASE_IDX                                                                     0
+#define mmIH_VF_RB_STATUS2                                                                             0x00cf
+#define mmIH_VF_RB_STATUS2_BASE_IDX                                                                    0
+#define mmIH_VF_RB1_STATUS                                                                             0x00d0
+#define mmIH_VF_RB1_STATUS_BASE_IDX                                                                    0
+#define mmIH_VF_RB1_STATUS2                                                                            0x00d1
+#define mmIH_VF_RB1_STATUS2_BASE_IDX                                                                   0
+#define mmIH_VF_RB2_STATUS                                                                             0x00d2
+#define mmIH_VF_RB2_STATUS_BASE_IDX                                                                    0
+#define mmIH_VF_RB2_STATUS2                                                                            0x00d3
+#define mmIH_VF_RB2_STATUS2_BASE_IDX                                                                   0
+#define mmIH_INT_FLOOD_CNTL                                                                            0x00d5
+#define mmIH_INT_FLOOD_CNTL_BASE_IDX                                                                   0
+#define mmIH_RB0_INT_FLOOD_STATUS                                                                      0x00d6
+#define mmIH_RB0_INT_FLOOD_STATUS_BASE_IDX                                                             0
+#define mmIH_RB1_INT_FLOOD_STATUS                                                                      0x00d7
+#define mmIH_RB1_INT_FLOOD_STATUS_BASE_IDX                                                             0
+#define mmIH_RB2_INT_FLOOD_STATUS                                                                      0x00d8
+#define mmIH_RB2_INT_FLOOD_STATUS_BASE_IDX                                                             0
+#define mmIH_INT_FLOOD_STATUS                                                                          0x00d9
+#define mmIH_INT_FLOOD_STATUS_BASE_IDX                                                                 0
+#define mmIH_STORM_CLIENT_LIST_CNTL                                                                    0x00da
+#define mmIH_STORM_CLIENT_LIST_CNTL_BASE_IDX                                                           0
+#define mmIH_CLK_CTRL                                                                                  0x00db
+#define mmIH_CLK_CTRL_BASE_IDX                                                                         0
+#define mmIH_INT_FLAGS                                                                                 0x00dc
+#define mmIH_INT_FLAGS_BASE_IDX                                                                        0
+#define mmIH_LAST_INT_INFO0                                                                            0x00dd
+#define mmIH_LAST_INT_INFO0_BASE_IDX                                                                   0
+#define mmIH_LAST_INT_INFO1                                                                            0x00de
+#define mmIH_LAST_INT_INFO1_BASE_IDX                                                                   0
+#define mmIH_LAST_INT_INFO2                                                                            0x00df
+#define mmIH_LAST_INT_INFO2_BASE_IDX                                                                   0
+#define mmIH_SCRATCH                                                                                   0x00e0
+#define mmIH_SCRATCH_BASE_IDX                                                                          0
+#define mmIH_CLIENT_CREDIT_ERROR                                                                       0x00e1
+#define mmIH_CLIENT_CREDIT_ERROR_BASE_IDX                                                              0
+#define mmIH_GPU_IOV_VIOLATION_LOG                                                                     0x00e2
+#define mmIH_GPU_IOV_VIOLATION_LOG_BASE_IDX                                                            0
+#define mmIH_COOKIE_REC_VIOLATION_LOG                                                                  0x00e3
+#define mmIH_COOKIE_REC_VIOLATION_LOG_BASE_IDX                                                         0
+#define mmIH_CREDIT_STATUS                                                                             0x00e4
+#define mmIH_CREDIT_STATUS_BASE_IDX                                                                    0
+#define mmIH_MMHUB_ERROR                                                                               0x00e5
+#define mmIH_MMHUB_ERROR_BASE_IDX                                                                      0
+#define mmIH_MEM_POWER_CTRL                                                                            0x00e8
+#define mmIH_MEM_POWER_CTRL_BASE_IDX                                                                   0
+#define mmIH_REGISTER_LAST_PART2                                                                       0x00ff
+#define mmIH_REGISTER_LAST_PART2_BASE_IDX                                                              0
+#define mmSEM_CLK_CTRL                                                                                 0x0100
+#define mmSEM_CLK_CTRL_BASE_IDX                                                                        0
+#define mmSEM_UTC_CREDIT                                                                               0x0101
+#define mmSEM_UTC_CREDIT_BASE_IDX                                                                      0
+#define mmSEM_UTC_CONFIG                                                                               0x0102
+#define mmSEM_UTC_CONFIG_BASE_IDX                                                                      0
+#define mmSEM_UTCL2_TRAN_EN_LUT                                                                        0x0103
+#define mmSEM_UTCL2_TRAN_EN_LUT_BASE_IDX                                                               0
+#define mmSEM_MCIF_CONFIG                                                                              0x0104
+#define mmSEM_MCIF_CONFIG_BASE_IDX                                                                     0
+#define mmSEM_PERFMON_CNTL                                                                             0x0105
+#define mmSEM_PERFMON_CNTL_BASE_IDX                                                                    0
+#define mmSEM_PERFCOUNTER0_RESULT                                                                      0x0106
+#define mmSEM_PERFCOUNTER0_RESULT_BASE_IDX                                                             0
+#define mmSEM_PERFCOUNTER1_RESULT                                                                      0x0107
+#define mmSEM_PERFCOUNTER1_RESULT_BASE_IDX                                                             0
+#define mmSEM_STATUS                                                                                   0x0108
+#define mmSEM_STATUS_BASE_IDX                                                                          0
+#define mmSEM_MAILBOX_CLIENTCONFIG                                                                     0x0109
+#define mmSEM_MAILBOX_CLIENTCONFIG_BASE_IDX                                                            0
+#define mmSEM_MAILBOX                                                                                  0x010a
+#define mmSEM_MAILBOX_BASE_IDX                                                                         0
+#define mmSEM_MAILBOX_CONTROL                                                                          0x010b
+#define mmSEM_MAILBOX_CONTROL_BASE_IDX                                                                 0
+#define mmSEM_CHICKEN_BITS                                                                             0x010c
+#define mmSEM_CHICKEN_BITS_BASE_IDX                                                                    0
+#define mmSEM_MAILBOX_CLIENTCONFIG_EXTRA                                                               0x010d
+#define mmSEM_MAILBOX_CLIENTCONFIG_EXTRA_BASE_IDX                                                      0
+#define mmSEM_GPU_IOV_VIOLATION_LOG                                                                    0x010e
+#define mmSEM_GPU_IOV_VIOLATION_LOG_BASE_IDX                                                           0
+#define mmSEM_OUTSTANDING_THRESHOLD                                                                    0x010f
+#define mmSEM_OUTSTANDING_THRESHOLD_BASE_IDX                                                           0
+#define mmSEM_MEM_POWER_CTRL                                                                           0x0110
+#define mmSEM_MEM_POWER_CTRL_BASE_IDX                                                                  0
+#define mmSEM_REGISTER_LAST_PART2                                                                      0x017f
+#define mmSEM_REGISTER_LAST_PART2_BASE_IDX                                                             0
+#define mmIH_ACTIVE_FCN_ID                                                                             0x0180
+#define mmIH_ACTIVE_FCN_ID_BASE_IDX                                                                    0
+#define mmIH_VIRT_RESET_REQ                                                                            0x0181
+#define mmIH_VIRT_RESET_REQ_BASE_IDX                                                                   0
+#define mmIH_CLIENT_CFG                                                                                0x0184
+#define mmIH_CLIENT_CFG_BASE_IDX                                                                       0
+#define mmIH_CLIENT_CFG_INDEX                                                                          0x0188
+#define mmIH_CLIENT_CFG_INDEX_BASE_IDX                                                                 0
+#define mmIH_CLIENT_CFG_DATA                                                                           0x0189
+#define mmIH_CLIENT_CFG_DATA_BASE_IDX                                                                  0
+#define mmIH_CID_REMAP_INDEX                                                                           0x018a
+#define mmIH_CID_REMAP_INDEX_BASE_IDX                                                                  0
+#define mmIH_CID_REMAP_DATA                                                                            0x018b
+#define mmIH_CID_REMAP_DATA_BASE_IDX                                                                   0
+#define mmIH_CHICKEN                                                                                   0x018c
+#define mmIH_CHICKEN_BASE_IDX                                                                          0
+#define mmIH_MMHUB_CNTL                                                                                0x018d
+#define mmIH_MMHUB_CNTL_BASE_IDX                                                                       0
+#define mmIH_INT_DROP_CNTL                                                                             0x018e
+#define mmIH_INT_DROP_CNTL_BASE_IDX                                                                    0
+#define mmIH_INT_DROP_MATCH_VALUE0                                                                     0x018f
+#define mmIH_INT_DROP_MATCH_VALUE0_BASE_IDX                                                            0
+#define mmIH_INT_DROP_MATCH_VALUE1                                                                     0x0190
+#define mmIH_INT_DROP_MATCH_VALUE1_BASE_IDX                                                            0
+#define mmIH_INT_DROP_MATCH_MASK0                                                                      0x0191
+#define mmIH_INT_DROP_MATCH_MASK0_BASE_IDX                                                             0
+#define mmIH_INT_DROP_MATCH_MASK1                                                                      0x0192
+#define mmIH_INT_DROP_MATCH_MASK1_BASE_IDX                                                             0
+#define mmIH_REGISTER_LAST_PART1                                                                       0x019f
+#define mmIH_REGISTER_LAST_PART1_BASE_IDX                                                              0
+#define mmSEM_ACTIVE_FCN_ID                                                                            0x01a0
+#define mmSEM_ACTIVE_FCN_ID_BASE_IDX                                                                   0
+#define mmSEM_VIRT_RESET_REQ                                                                           0x01a1
+#define mmSEM_VIRT_RESET_REQ_BASE_IDX                                                                  0
+#define mmSEM_RESP_SDMA0                                                                               0x01a4
+#define mmSEM_RESP_SDMA0_BASE_IDX                                                                      0
+#define mmSEM_RESP_SDMA1                                                                               0x01a5
+#define mmSEM_RESP_SDMA1_BASE_IDX                                                                      0
+#define mmSEM_RESP_UVD                                                                                 0x01a6
+#define mmSEM_RESP_UVD_BASE_IDX                                                                        0
+#define mmSEM_RESP_VCE_0                                                                               0x01a7
+#define mmSEM_RESP_VCE_0_BASE_IDX                                                                      0
+#define mmSEM_RESP_ACP                                                                                 0x01a8
+#define mmSEM_RESP_ACP_BASE_IDX                                                                        0
+#define mmSEM_RESP_ISP                                                                                 0x01a9
+#define mmSEM_RESP_ISP_BASE_IDX                                                                        0
+#define mmSEM_RESP_VCE_1                                                                               0x01aa
+#define mmSEM_RESP_VCE_1_BASE_IDX                                                                      0
+#define mmSEM_RESP_VP8                                                                                 0x01ab
+#define mmSEM_RESP_VP8_BASE_IDX                                                                        0
+#define mmSEM_RESP_GC                                                                                  0x01ac
+#define mmSEM_RESP_GC_BASE_IDX                                                                         0
+#define mmSEM_RESP_UVD_1                                                                               0x01ad
+#define mmSEM_RESP_UVD_1_BASE_IDX                                                                      0
+#define mmSEM_CID_REMAP_INDEX                                                                          0x01b0
+#define mmSEM_CID_REMAP_INDEX_BASE_IDX                                                                 0
+#define mmSEM_CID_REMAP_DATA                                                                           0x01b1
+#define mmSEM_CID_REMAP_DATA_BASE_IDX                                                                  0
+#define mmSEM_ATOMIC_OP_LUT                                                                            0x01b2
+#define mmSEM_ATOMIC_OP_LUT_BASE_IDX                                                                   0
+#define mmSEM_EDC_CONFIG                                                                               0x01b3
+#define mmSEM_EDC_CONFIG_BASE_IDX                                                                      0
+#define mmSEM_CHICKEN_BITS2                                                                            0x01b4
+#define mmSEM_CHICKEN_BITS2_BASE_IDX                                                                   0
+#define mmSEM_MMHUB_CNTL                                                                               0x01b5
+#define mmSEM_MMHUB_CNTL_BASE_IDX                                                                      0
+#define mmSEM_REGISTER_LAST_PART1                                                                      0x01bf
+#define mmSEM_REGISTER_LAST_PART1_BASE_IDX                                                             0
+
+#endif
diff --git a/drivers/gpu/drm/amd/include/asic_reg/oss/osssys_4_2_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/oss/osssys_4_2_0_sh_mask.h
new file mode 100644
index 000000000000..3ea83ea9ce3a
--- /dev/null
+++ b/drivers/gpu/drm/amd/include/asic_reg/oss/osssys_4_2_0_sh_mask.h
@@ -0,0 +1,1300 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef _osssys_4_2_0_SH_MASK_HEADER
+#define _osssys_4_2_0_SH_MASK_HEADER
+
+
+// addressBlock: osssys_osssysdec
+//IH_VMID_0_LUT
+#define IH_VMID_0_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_0_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_1_LUT
+#define IH_VMID_1_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_1_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_2_LUT
+#define IH_VMID_2_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_2_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_3_LUT
+#define IH_VMID_3_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_3_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_4_LUT
+#define IH_VMID_4_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_4_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_5_LUT
+#define IH_VMID_5_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_5_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_6_LUT
+#define IH_VMID_6_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_6_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_7_LUT
+#define IH_VMID_7_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_7_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_8_LUT
+#define IH_VMID_8_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_8_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_9_LUT
+#define IH_VMID_9_LUT__PASID__SHIFT                                                                           0x0
+#define IH_VMID_9_LUT__PASID_MASK                                                                             0x0000FFFFL
+//IH_VMID_10_LUT
+#define IH_VMID_10_LUT__PASID__SHIFT                                                                          0x0
+#define IH_VMID_10_LUT__PASID_MASK                                                                            0x0000FFFFL
+//IH_VMID_11_LUT
+#define IH_VMID_11_LUT__PASID__SHIFT                                                                          0x0
+#define IH_VMID_11_LUT__PASID_MASK                                                                            0x0000FFFFL
+//IH_VMID_12_LUT
+#define IH_VMID_12_LUT__PASID__SHIFT                                                                          0x0
+#define IH_VMID_12_LUT__PASID_MASK                                                                            0x0000FFFFL
+//IH_VMID_13_LUT
+#define IH_VMID_13_LUT__PASID__SHIFT                                                                          0x0
+#define IH_VMID_13_LUT__PASID_MASK                                                                            0x0000FFFFL
+//IH_VMID_14_LUT
+#define IH_VMID_14_LUT__PASID__SHIFT                                                                          0x0
+#define IH_VMID_14_LUT__PASID_MASK                                                                            0x0000FFFFL
+//IH_VMID_15_LUT
+#define IH_VMID_15_LUT__PASID__SHIFT                                                                          0x0
+#define IH_VMID_15_LUT__PASID_MASK                                                                            0x0000FFFFL
+//IH_VMID_0_LUT_MM
+#define IH_VMID_0_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_0_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_1_LUT_MM
+#define IH_VMID_1_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_1_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_2_LUT_MM
+#define IH_VMID_2_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_2_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_3_LUT_MM
+#define IH_VMID_3_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_3_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_4_LUT_MM
+#define IH_VMID_4_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_4_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_5_LUT_MM
+#define IH_VMID_5_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_5_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_6_LUT_MM
+#define IH_VMID_6_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_6_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_7_LUT_MM
+#define IH_VMID_7_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_7_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_8_LUT_MM
+#define IH_VMID_8_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_8_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_9_LUT_MM
+#define IH_VMID_9_LUT_MM__PASID__SHIFT                                                                        0x0
+#define IH_VMID_9_LUT_MM__PASID_MASK                                                                          0x0000FFFFL
+//IH_VMID_10_LUT_MM
+#define IH_VMID_10_LUT_MM__PASID__SHIFT                                                                       0x0
+#define IH_VMID_10_LUT_MM__PASID_MASK                                                                         0x0000FFFFL
+//IH_VMID_11_LUT_MM
+#define IH_VMID_11_LUT_MM__PASID__SHIFT                                                                       0x0
+#define IH_VMID_11_LUT_MM__PASID_MASK                                                                         0x0000FFFFL
+//IH_VMID_12_LUT_MM
+#define IH_VMID_12_LUT_MM__PASID__SHIFT                                                                       0x0
+#define IH_VMID_12_LUT_MM__PASID_MASK                                                                         0x0000FFFFL
+//IH_VMID_13_LUT_MM
+#define IH_VMID_13_LUT_MM__PASID__SHIFT                                                                       0x0
+#define IH_VMID_13_LUT_MM__PASID_MASK                                                                         0x0000FFFFL
+//IH_VMID_14_LUT_MM
+#define IH_VMID_14_LUT_MM__PASID__SHIFT                                                                       0x0
+#define IH_VMID_14_LUT_MM__PASID_MASK                                                                         0x0000FFFFL
+//IH_VMID_15_LUT_MM
+#define IH_VMID_15_LUT_MM__PASID__SHIFT                                                                       0x0
+#define IH_VMID_15_LUT_MM__PASID_MASK                                                                         0x0000FFFFL
+//IH_COOKIE_0
+#define IH_COOKIE_0__CLIENT_ID__SHIFT                                                                         0x0
+#define IH_COOKIE_0__SOURCE_ID__SHIFT                                                                         0x8
+#define IH_COOKIE_0__RING_ID__SHIFT                                                                           0x10
+#define IH_COOKIE_0__VM_ID__SHIFT                                                                             0x18
+#define IH_COOKIE_0__RESERVED__SHIFT                                                                          0x1c
+#define IH_COOKIE_0__VMID_TYPE__SHIFT                                                                         0x1f
+#define IH_COOKIE_0__CLIENT_ID_MASK                                                                           0x000000FFL
+#define IH_COOKIE_0__SOURCE_ID_MASK                                                                           0x0000FF00L
+#define IH_COOKIE_0__RING_ID_MASK                                                                             0x00FF0000L
+#define IH_COOKIE_0__VM_ID_MASK                                                                               0x0F000000L
+#define IH_COOKIE_0__RESERVED_MASK                                                                            0x70000000L
+#define IH_COOKIE_0__VMID_TYPE_MASK                                                                           0x80000000L
+//IH_COOKIE_1
+#define IH_COOKIE_1__TIMESTAMP_31_0__SHIFT                                                                    0x0
+#define IH_COOKIE_1__TIMESTAMP_31_0_MASK                                                                      0xFFFFFFFFL
+//IH_COOKIE_2
+#define IH_COOKIE_2__TIMESTAMP_47_32__SHIFT                                                                   0x0
+#define IH_COOKIE_2__RESERVED__SHIFT                                                                          0x10
+#define IH_COOKIE_2__TIMESTAMP_SRC__SHIFT                                                                     0x1f
+#define IH_COOKIE_2__TIMESTAMP_47_32_MASK                                                                     0x0000FFFFL
+#define IH_COOKIE_2__RESERVED_MASK                                                                            0x7FFF0000L
+#define IH_COOKIE_2__TIMESTAMP_SRC_MASK                                                                       0x80000000L
+//IH_COOKIE_3
+#define IH_COOKIE_3__PAS_ID__SHIFT                                                                            0x0
+#define IH_COOKIE_3__RESERVED__SHIFT                                                                          0x10
+#define IH_COOKIE_3__PASID_SRC__SHIFT                                                                         0x1f
+#define IH_COOKIE_3__PAS_ID_MASK                                                                              0x0000FFFFL
+#define IH_COOKIE_3__RESERVED_MASK                                                                            0x7FFF0000L
+#define IH_COOKIE_3__PASID_SRC_MASK                                                                           0x80000000L
+//IH_COOKIE_4
+#define IH_COOKIE_4__CONTEXT_ID_31_0__SHIFT                                                                   0x0
+#define IH_COOKIE_4__CONTEXT_ID_31_0_MASK                                                                     0xFFFFFFFFL
+//IH_COOKIE_5
+#define IH_COOKIE_5__CONTEXT_ID_63_32__SHIFT                                                                  0x0
+#define IH_COOKIE_5__CONTEXT_ID_63_32_MASK                                                                    0xFFFFFFFFL
+//IH_COOKIE_6
+#define IH_COOKIE_6__CONTEXT_ID_95_64__SHIFT                                                                  0x0
+#define IH_COOKIE_6__CONTEXT_ID_95_64_MASK                                                                    0xFFFFFFFFL
+//IH_COOKIE_7
+#define IH_COOKIE_7__CONTEXT_ID_128_96__SHIFT                                                                 0x0
+#define IH_COOKIE_7__CONTEXT_ID_128_96_MASK                                                                   0xFFFFFFFFL
+//IH_REGISTER_LAST_PART0
+#define IH_REGISTER_LAST_PART0__RESERVED__SHIFT                                                               0x0
+#define IH_REGISTER_LAST_PART0__RESERVED_MASK                                                                 0xFFFFFFFFL
+//SEM_REQ_INPUT_0
+#define SEM_REQ_INPUT_0__DATA__SHIFT                                                                          0x0
+#define SEM_REQ_INPUT_0__DATA_MASK                                                                            0xFFFFFFFFL
+//SEM_REQ_INPUT_1
+#define SEM_REQ_INPUT_1__DATA__SHIFT                                                                          0x0
+#define SEM_REQ_INPUT_1__DATA_MASK                                                                            0xFFFFFFFFL
+//SEM_REQ_INPUT_2
+#define SEM_REQ_INPUT_2__DATA__SHIFT                                                                          0x0
+#define SEM_REQ_INPUT_2__DATA_MASK                                                                            0xFFFFFFFFL
+//SEM_REQ_INPUT_3
+#define SEM_REQ_INPUT_3__DATA__SHIFT                                                                          0x0
+#define SEM_REQ_INPUT_3__DATA_MASK                                                                            0xFFFFFFFFL
+//SEM_REGISTER_LAST_PART0
+#define SEM_REGISTER_LAST_PART0__RESERVED__SHIFT                                                              0x0
+#define SEM_REGISTER_LAST_PART0__RESERVED_MASK                                                                0xFFFFFFFFL
+//IH_RB_CNTL
+#define IH_RB_CNTL__RB_ENABLE__SHIFT                                                                          0x0
+#define IH_RB_CNTL__RB_SIZE__SHIFT                                                                            0x1
+#define IH_RB_CNTL__RB_GPU_TS_ENABLE__SHIFT                                                                   0x7
+#define IH_RB_CNTL__WPTR_WRITEBACK_ENABLE__SHIFT                                                              0x8
+#define IH_RB_CNTL__RB_FULL_DRAIN_ENABLE__SHIFT                                                               0x9
+#define IH_RB_CNTL__FULL_DRAIN_CLEAR__SHIFT                                                                   0xa
+#define IH_RB_CNTL__PAGE_RB_CLEAR__SHIFT                                                                      0xb
+#define IH_RB_CNTL__RB_USED_INT_THRESHOLD__SHIFT                                                              0xc
+#define IH_RB_CNTL__WPTR_OVERFLOW_ENABLE__SHIFT                                                               0x10
+#define IH_RB_CNTL__ENABLE_INTR__SHIFT                                                                        0x11
+#define IH_RB_CNTL__MC_SWAP__SHIFT                                                                            0x12
+#define IH_RB_CNTL__MC_SNOOP__SHIFT                                                                           0x14
+#define IH_RB_CNTL__RPTR_REARM__SHIFT                                                                         0x15
+#define IH_RB_CNTL__MC_RO__SHIFT                                                                              0x16
+#define IH_RB_CNTL__MC_VMID__SHIFT                                                                            0x18
+#define IH_RB_CNTL__MC_SPACE__SHIFT                                                                           0x1c
+#define IH_RB_CNTL__WPTR_OVERFLOW_CLEAR__SHIFT                                                                0x1f
+#define IH_RB_CNTL__RB_ENABLE_MASK                                                                            0x00000001L
+#define IH_RB_CNTL__RB_SIZE_MASK                                                                              0x0000003EL
+#define IH_RB_CNTL__RB_GPU_TS_ENABLE_MASK                                                                     0x00000080L
+#define IH_RB_CNTL__WPTR_WRITEBACK_ENABLE_MASK                                                                0x00000100L
+#define IH_RB_CNTL__RB_FULL_DRAIN_ENABLE_MASK                                                                 0x00000200L
+#define IH_RB_CNTL__FULL_DRAIN_CLEAR_MASK                                                                     0x00000400L
+#define IH_RB_CNTL__PAGE_RB_CLEAR_MASK                                                                        0x00000800L
+#define IH_RB_CNTL__RB_USED_INT_THRESHOLD_MASK                                                                0x0000F000L
+#define IH_RB_CNTL__WPTR_OVERFLOW_ENABLE_MASK                                                                 0x00010000L
+#define IH_RB_CNTL__ENABLE_INTR_MASK                                                                          0x00020000L
+#define IH_RB_CNTL__MC_SWAP_MASK                                                                              0x000C0000L
+#define IH_RB_CNTL__MC_SNOOP_MASK                                                                             0x00100000L
+#define IH_RB_CNTL__RPTR_REARM_MASK                                                                           0x00200000L
+#define IH_RB_CNTL__MC_RO_MASK                                                                                0x00400000L
+#define IH_RB_CNTL__MC_VMID_MASK                                                                              0x0F000000L
+#define IH_RB_CNTL__MC_SPACE_MASK                                                                             0x70000000L
+#define IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK                                                                  0x80000000L
+//IH_RB_BASE
+#define IH_RB_BASE__ADDR__SHIFT                                                                               0x0
+#define IH_RB_BASE__ADDR_MASK                                                                                 0xFFFFFFFFL
+//IH_RB_BASE_HI
+#define IH_RB_BASE_HI__ADDR__SHIFT                                                                            0x0
+#define IH_RB_BASE_HI__ADDR_MASK                                                                              0x000000FFL
+//IH_RB_RPTR
+#define IH_RB_RPTR__OFFSET__SHIFT                                                                             0x2
+#define IH_RB_RPTR__OFFSET_MASK                                                                               0x0003FFFCL
+//IH_RB_WPTR
+#define IH_RB_WPTR__RB_OVERFLOW__SHIFT                                                                        0x0
+#define IH_RB_WPTR__OFFSET__SHIFT                                                                             0x2
+#define IH_RB_WPTR__RB_LEFT_NONE__SHIFT                                                                       0x12
+#define IH_RB_WPTR__RB_MAY_OVERFLOW__SHIFT                                                                    0x13
+#define IH_RB_WPTR__RB_OVERFLOW_MASK                                                                          0x00000001L
+#define IH_RB_WPTR__OFFSET_MASK                                                                               0x0003FFFCL
+#define IH_RB_WPTR__RB_LEFT_NONE_MASK                                                                         0x00040000L
+#define IH_RB_WPTR__RB_MAY_OVERFLOW_MASK                                                                      0x00080000L
+//IH_RB_WPTR_ADDR_HI
+#define IH_RB_WPTR_ADDR_HI__ADDR__SHIFT                                                                       0x0
+#define IH_RB_WPTR_ADDR_HI__ADDR_MASK                                                                         0x0000FFFFL
+//IH_RB_WPTR_ADDR_LO
+#define IH_RB_WPTR_ADDR_LO__ADDR__SHIFT                                                                       0x2
+#define IH_RB_WPTR_ADDR_LO__ADDR_MASK                                                                         0xFFFFFFFCL
+//IH_DOORBELL_RPTR
+#define IH_DOORBELL_RPTR__OFFSET__SHIFT                                                                       0x0
+#define IH_DOORBELL_RPTR__ENABLE__SHIFT                                                                       0x1c
+#define IH_DOORBELL_RPTR__OFFSET_MASK                                                                         0x03FFFFFFL
+#define IH_DOORBELL_RPTR__ENABLE_MASK                                                                         0x10000000L
+//IH_RB_CNTL_RING1
+#define IH_RB_CNTL_RING1__RB_ENABLE__SHIFT                                                                    0x0
+#define IH_RB_CNTL_RING1__RB_SIZE__SHIFT                                                                      0x1
+#define IH_RB_CNTL_RING1__RB_GPU_TS_ENABLE__SHIFT                                                             0x7
+#define IH_RB_CNTL_RING1__RB_FULL_DRAIN_ENABLE__SHIFT                                                         0x9
+#define IH_RB_CNTL_RING1__FULL_DRAIN_CLEAR__SHIFT                                                             0xa
+#define IH_RB_CNTL_RING1__PAGE_RB_CLEAR__SHIFT                                                                0xb
+#define IH_RB_CNTL_RING1__RB_USED_INT_THRESHOLD__SHIFT                                                        0xc
+#define IH_RB_CNTL_RING1__WPTR_OVERFLOW_ENABLE__SHIFT                                                         0x10
+#define IH_RB_CNTL_RING1__MC_SWAP__SHIFT                                                                      0x12
+#define IH_RB_CNTL_RING1__MC_SNOOP__SHIFT                                                                     0x14
+#define IH_RB_CNTL_RING1__MC_RO__SHIFT                                                                        0x16
+#define IH_RB_CNTL_RING1__MC_VMID__SHIFT                                                                      0x18
+#define IH_RB_CNTL_RING1__MC_SPACE__SHIFT                                                                     0x1c
+#define IH_RB_CNTL_RING1__WPTR_OVERFLOW_CLEAR__SHIFT                                                          0x1f
+#define IH_RB_CNTL_RING1__RB_ENABLE_MASK                                                                      0x00000001L
+#define IH_RB_CNTL_RING1__RB_SIZE_MASK                                                                        0x0000003EL
+#define IH_RB_CNTL_RING1__RB_GPU_TS_ENABLE_MASK                                                               0x00000080L
+#define IH_RB_CNTL_RING1__RB_FULL_DRAIN_ENABLE_MASK                                                           0x00000200L
+#define IH_RB_CNTL_RING1__FULL_DRAIN_CLEAR_MASK                                                               0x00000400L
+#define IH_RB_CNTL_RING1__PAGE_RB_CLEAR_MASK                                                                  0x00000800L
+#define IH_RB_CNTL_RING1__RB_USED_INT_THRESHOLD_MASK                                                          0x0000F000L
+#define IH_RB_CNTL_RING1__WPTR_OVERFLOW_ENABLE_MASK                                                           0x00010000L
+#define IH_RB_CNTL_RING1__MC_SWAP_MASK                                                                        0x000C0000L
+#define IH_RB_CNTL_RING1__MC_SNOOP_MASK                                                                       0x00100000L
+#define IH_RB_CNTL_RING1__MC_RO_MASK                                                                          0x00400000L
+#define IH_RB_CNTL_RING1__MC_VMID_MASK                                                                        0x0F000000L
+#define IH_RB_CNTL_RING1__MC_SPACE_MASK                                                                       0x70000000L
+#define IH_RB_CNTL_RING1__WPTR_OVERFLOW_CLEAR_MASK                                                            0x80000000L
+//IH_RB_BASE_RING1
+#define IH_RB_BASE_RING1__ADDR__SHIFT                                                                         0x0
+#define IH_RB_BASE_RING1__ADDR_MASK                                                                           0xFFFFFFFFL
+//IH_RB_BASE_HI_RING1
+#define IH_RB_BASE_HI_RING1__ADDR__SHIFT                                                                      0x0
+#define IH_RB_BASE_HI_RING1__ADDR_MASK                                                                        0x000000FFL
+//IH_RB_RPTR_RING1
+#define IH_RB_RPTR_RING1__OFFSET__SHIFT                                                                       0x2
+#define IH_RB_RPTR_RING1__OFFSET_MASK                                                                         0x0003FFFCL
+//IH_RB_WPTR_RING1
+#define IH_RB_WPTR_RING1__RB_OVERFLOW__SHIFT                                                                  0x0
+#define IH_RB_WPTR_RING1__OFFSET__SHIFT                                                                       0x2
+#define IH_RB_WPTR_RING1__RB_LEFT_NONE__SHIFT                                                                 0x12
+#define IH_RB_WPTR_RING1__RB_MAY_OVERFLOW__SHIFT                                                              0x13
+#define IH_RB_WPTR_RING1__RB_OVERFLOW_MASK                                                                    0x00000001L
+#define IH_RB_WPTR_RING1__OFFSET_MASK                                                                         0x0003FFFCL
+#define IH_RB_WPTR_RING1__RB_LEFT_NONE_MASK                                                                   0x00040000L
+#define IH_RB_WPTR_RING1__RB_MAY_OVERFLOW_MASK                                                                0x00080000L
+//IH_DOORBELL_RPTR_RING1
+#define IH_DOORBELL_RPTR_RING1__OFFSET__SHIFT                                                                 0x0
+#define IH_DOORBELL_RPTR_RING1__ENABLE__SHIFT                                                                 0x1c
+#define IH_DOORBELL_RPTR_RING1__OFFSET_MASK                                                                   0x03FFFFFFL
+#define IH_DOORBELL_RPTR_RING1__ENABLE_MASK                                                                   0x10000000L
+//IH_RB_CNTL_RING2
+#define IH_RB_CNTL_RING2__RB_ENABLE__SHIFT                                                                    0x0
+#define IH_RB_CNTL_RING2__RB_SIZE__SHIFT                                                                      0x1
+#define IH_RB_CNTL_RING2__RB_GPU_TS_ENABLE__SHIFT                                                             0x7
+#define IH_RB_CNTL_RING2__RB_FULL_DRAIN_ENABLE__SHIFT                                                         0x9
+#define IH_RB_CNTL_RING2__FULL_DRAIN_CLEAR__SHIFT                                                             0xa
+#define IH_RB_CNTL_RING2__PAGE_RB_CLEAR__SHIFT                                                                0xb
+#define IH_RB_CNTL_RING2__RB_USED_INT_THRESHOLD__SHIFT                                                        0xc
+#define IH_RB_CNTL_RING2__WPTR_OVERFLOW_ENABLE__SHIFT                                                         0x10
+#define IH_RB_CNTL_RING2__MC_SWAP__SHIFT                                                                      0x12
+#define IH_RB_CNTL_RING2__MC_SNOOP__SHIFT                                                                     0x14
+#define IH_RB_CNTL_RING2__MC_RO__SHIFT                                                                        0x16
+#define IH_RB_CNTL_RING2__MC_VMID__SHIFT                                                                      0x18
+#define IH_RB_CNTL_RING2__MC_SPACE__SHIFT                                                                     0x1c
+#define IH_RB_CNTL_RING2__WPTR_OVERFLOW_CLEAR__SHIFT                                                          0x1f
+#define IH_RB_CNTL_RING2__RB_ENABLE_MASK                                                                      0x00000001L
+#define IH_RB_CNTL_RING2__RB_SIZE_MASK                                                                        0x0000003EL
+#define IH_RB_CNTL_RING2__RB_GPU_TS_ENABLE_MASK                                                               0x00000080L
+#define IH_RB_CNTL_RING2__RB_FULL_DRAIN_ENABLE_MASK                                                           0x00000200L
+#define IH_RB_CNTL_RING2__FULL_DRAIN_CLEAR_MASK                                                               0x00000400L
+#define IH_RB_CNTL_RING2__PAGE_RB_CLEAR_MASK                                                                  0x00000800L
+#define IH_RB_CNTL_RING2__RB_USED_INT_THRESHOLD_MASK                                                          0x0000F000L
+#define IH_RB_CNTL_RING2__WPTR_OVERFLOW_ENABLE_MASK                                                           0x00010000L
+#define IH_RB_CNTL_RING2__MC_SWAP_MASK                                                                        0x000C0000L
+#define IH_RB_CNTL_RING2__MC_SNOOP_MASK                                                                       0x00100000L
+#define IH_RB_CNTL_RING2__MC_RO_MASK                                                                          0x00400000L
+#define IH_RB_CNTL_RING2__MC_VMID_MASK                                                                        0x0F000000L
+#define IH_RB_CNTL_RING2__MC_SPACE_MASK                                                                       0x70000000L
+#define IH_RB_CNTL_RING2__WPTR_OVERFLOW_CLEAR_MASK                                                            0x80000000L
+//IH_RB_BASE_RING2
+#define IH_RB_BASE_RING2__ADDR__SHIFT                                                                         0x0
+#define IH_RB_BASE_RING2__ADDR_MASK                                                                           0xFFFFFFFFL
+//IH_RB_BASE_HI_RING2
+#define IH_RB_BASE_HI_RING2__ADDR__SHIFT                                                                      0x0
+#define IH_RB_BASE_HI_RING2__ADDR_MASK                                                                        0x000000FFL
+//IH_RB_RPTR_RING2
+#define IH_RB_RPTR_RING2__OFFSET__SHIFT                                                                       0x2
+#define IH_RB_RPTR_RING2__OFFSET_MASK                                                                         0x0003FFFCL
+//IH_RB_WPTR_RING2
+#define IH_RB_WPTR_RING2__RB_OVERFLOW__SHIFT                                                                  0x0
+#define IH_RB_WPTR_RING2__OFFSET__SHIFT                                                                       0x2
+#define IH_RB_WPTR_RING2__RB_LEFT_NONE__SHIFT                                                                 0x12
+#define IH_RB_WPTR_RING2__RB_MAY_OVERFLOW__SHIFT                                                              0x13
+#define IH_RB_WPTR_RING2__RB_OVERFLOW_MASK                                                                    0x00000001L
+#define IH_RB_WPTR_RING2__OFFSET_MASK                                                                         0x0003FFFCL
+#define IH_RB_WPTR_RING2__RB_LEFT_NONE_MASK                                                                   0x00040000L
+#define IH_RB_WPTR_RING2__RB_MAY_OVERFLOW_MASK                                                                0x00080000L
+//IH_DOORBELL_RPTR_RING2
+#define IH_DOORBELL_RPTR_RING2__OFFSET__SHIFT                                                                 0x0
+#define IH_DOORBELL_RPTR_RING2__ENABLE__SHIFT                                                                 0x1c
+#define IH_DOORBELL_RPTR_RING2__OFFSET_MASK                                                                   0x03FFFFFFL
+#define IH_DOORBELL_RPTR_RING2__ENABLE_MASK                                                                   0x10000000L
+//IH_VERSION
+#define IH_VERSION__MINVER__SHIFT                                                                             0x0
+#define IH_VERSION__MAJVER__SHIFT                                                                             0x8
+#define IH_VERSION__REV__SHIFT                                                                                0x10
+#define IH_VERSION__MINVER_MASK                                                                               0x0000007FL
+#define IH_VERSION__MAJVER_MASK                                                                               0x00007F00L
+#define IH_VERSION__REV_MASK                                                                                  0x003F0000L
+//IH_CNTL
+#define IH_CNTL__WPTR_WRITEBACK_TIMER__SHIFT                                                                  0x0
+#define IH_CNTL__IH_IDLE_HYSTERESIS_CNTL__SHIFT                                                               0x6
+#define IH_CNTL__IH_FIFO_HIGHWATER__SHIFT                                                                     0x8
+#define IH_CNTL__MC_WR_CLEAN_CNT__SHIFT                                                                       0x14
+#define IH_CNTL__WPTR_WRITEBACK_TIMER_MASK                                                                    0x0000001FL
+#define IH_CNTL__IH_IDLE_HYSTERESIS_CNTL_MASK                                                                 0x000000C0L
+#define IH_CNTL__IH_FIFO_HIGHWATER_MASK                                                                       0x00007F00L
+#define IH_CNTL__MC_WR_CLEAN_CNT_MASK                                                                         0x01F00000L
+//IH_CNTL2
+#define IH_CNTL2__SELF_IV_FORCE_WPTR_UPDATE_TIMEOUT__SHIFT                                                    0x0
+#define IH_CNTL2__SELF_IV_FORCE_WPTR_UPDATE_ENABLE__SHIFT                                                     0x8
+#define IH_CNTL2__SELF_IV_FORCE_WPTR_UPDATE_TIMEOUT_MASK                                                      0x0000001FL
+#define IH_CNTL2__SELF_IV_FORCE_WPTR_UPDATE_ENABLE_MASK                                                       0x00000100L
+//IH_STATUS
+#define IH_STATUS__IDLE__SHIFT                                                                                0x0
+#define IH_STATUS__INPUT_IDLE__SHIFT                                                                          0x1
+#define IH_STATUS__BUFFER_IDLE__SHIFT                                                                         0x2
+#define IH_STATUS__RB_FULL__SHIFT                                                                             0x3
+#define IH_STATUS__RB_FULL_DRAIN__SHIFT                                                                       0x4
+#define IH_STATUS__RB_OVERFLOW__SHIFT                                                                         0x5
+#define IH_STATUS__MC_WR_IDLE__SHIFT                                                                          0x6
+#define IH_STATUS__MC_WR_STALL__SHIFT                                                                         0x7
+#define IH_STATUS__MC_WR_CLEAN_PENDING__SHIFT                                                                 0x8
+#define IH_STATUS__MC_WR_CLEAN_STALL__SHIFT                                                                   0x9
+#define IH_STATUS__BIF_INTERRUPT_LINE__SHIFT                                                                  0xa
+#define IH_STATUS__SWITCH_READY__SHIFT                                                                        0xb
+#define IH_STATUS__RB1_FULL__SHIFT                                                                            0xc
+#define IH_STATUS__RB1_FULL_DRAIN__SHIFT                                                                      0xd
+#define IH_STATUS__RB1_OVERFLOW__SHIFT                                                                        0xe
+#define IH_STATUS__RB2_FULL__SHIFT                                                                            0xf
+#define IH_STATUS__RB2_FULL_DRAIN__SHIFT                                                                      0x10
+#define IH_STATUS__RB2_OVERFLOW__SHIFT                                                                        0x11
+#define IH_STATUS__SELF_INT_GEN_IDLE__SHIFT                                                                   0x12
+#define IH_STATUS__IDLE_MASK                                                                                  0x00000001L
+#define IH_STATUS__INPUT_IDLE_MASK                                                                            0x00000002L
+#define IH_STATUS__BUFFER_IDLE_MASK                                                                           0x00000004L
+#define IH_STATUS__RB_FULL_MASK                                                                               0x00000008L
+#define IH_STATUS__RB_FULL_DRAIN_MASK                                                                         0x00000010L
+#define IH_STATUS__RB_OVERFLOW_MASK                                                                           0x00000020L
+#define IH_STATUS__MC_WR_IDLE_MASK                                                                            0x00000040L
+#define IH_STATUS__MC_WR_STALL_MASK                                                                           0x00000080L
+#define IH_STATUS__MC_WR_CLEAN_PENDING_MASK                                                                   0x00000100L
+#define IH_STATUS__MC_WR_CLEAN_STALL_MASK                                                                     0x00000200L
+#define IH_STATUS__BIF_INTERRUPT_LINE_MASK                                                                    0x00000400L
+#define IH_STATUS__SWITCH_READY_MASK                                                                          0x00000800L
+#define IH_STATUS__RB1_FULL_MASK                                                                              0x00001000L
+#define IH_STATUS__RB1_FULL_DRAIN_MASK                                                                        0x00002000L
+#define IH_STATUS__RB1_OVERFLOW_MASK                                                                          0x00004000L
+#define IH_STATUS__RB2_FULL_MASK                                                                              0x00008000L
+#define IH_STATUS__RB2_FULL_DRAIN_MASK                                                                        0x00010000L
+#define IH_STATUS__RB2_OVERFLOW_MASK                                                                          0x00020000L
+#define IH_STATUS__SELF_INT_GEN_IDLE_MASK                                                                     0x00040000L
+//IH_PERFMON_CNTL
+#define IH_PERFMON_CNTL__ENABLE0__SHIFT                                                                       0x0
+#define IH_PERFMON_CNTL__CLEAR0__SHIFT                                                                        0x1
+#define IH_PERFMON_CNTL__PERF_SEL0__SHIFT                                                                     0x2
+#define IH_PERFMON_CNTL__ENABLE1__SHIFT                                                                       0x10
+#define IH_PERFMON_CNTL__CLEAR1__SHIFT                                                                        0x11
+#define IH_PERFMON_CNTL__PERF_SEL1__SHIFT                                                                     0x12
+#define IH_PERFMON_CNTL__ENABLE0_MASK                                                                         0x00000001L
+#define IH_PERFMON_CNTL__CLEAR0_MASK                                                                          0x00000002L
+#define IH_PERFMON_CNTL__PERF_SEL0_MASK                                                                       0x000007FCL
+#define IH_PERFMON_CNTL__ENABLE1_MASK                                                                         0x00010000L
+#define IH_PERFMON_CNTL__CLEAR1_MASK                                                                          0x00020000L
+#define IH_PERFMON_CNTL__PERF_SEL1_MASK                                                                       0x07FC0000L
+//IH_PERFCOUNTER0_RESULT
+#define IH_PERFCOUNTER0_RESULT__PERF_COUNT__SHIFT                                                             0x0
+#define IH_PERFCOUNTER0_RESULT__PERF_COUNT_MASK                                                               0xFFFFFFFFL
+//IH_PERFCOUNTER1_RESULT
+#define IH_PERFCOUNTER1_RESULT__PERF_COUNT__SHIFT                                                             0x0
+#define IH_PERFCOUNTER1_RESULT__PERF_COUNT_MASK                                                               0xFFFFFFFFL
+//IH_DSM_MATCH_VALUE_BIT_31_0
+#define IH_DSM_MATCH_VALUE_BIT_31_0__VALUE__SHIFT                                                             0x0
+#define IH_DSM_MATCH_VALUE_BIT_31_0__VALUE_MASK                                                               0xFFFFFFFFL
+//IH_DSM_MATCH_VALUE_BIT_63_32
+#define IH_DSM_MATCH_VALUE_BIT_63_32__VALUE__SHIFT                                                            0x0
+#define IH_DSM_MATCH_VALUE_BIT_63_32__VALUE_MASK                                                              0xFFFFFFFFL
+//IH_DSM_MATCH_VALUE_BIT_95_64
+#define IH_DSM_MATCH_VALUE_BIT_95_64__VALUE__SHIFT                                                            0x0
+#define IH_DSM_MATCH_VALUE_BIT_95_64__VALUE_MASK                                                              0xFFFFFFFFL
+//IH_DSM_MATCH_FIELD_CONTROL
+#define IH_DSM_MATCH_FIELD_CONTROL__SRC_EN__SHIFT                                                             0x0
+#define IH_DSM_MATCH_FIELD_CONTROL__FCNID_EN__SHIFT                                                           0x1
+#define IH_DSM_MATCH_FIELD_CONTROL__TIMESTAMP_EN__SHIFT                                                       0x2
+#define IH_DSM_MATCH_FIELD_CONTROL__RINGID_EN__SHIFT                                                          0x3
+#define IH_DSM_MATCH_FIELD_CONTROL__VMID_EN__SHIFT                                                            0x4
+#define IH_DSM_MATCH_FIELD_CONTROL__PASID_EN__SHIFT                                                           0x5
+#define IH_DSM_MATCH_FIELD_CONTROL__CLIENT_ID_EN__SHIFT                                                       0x6
+#define IH_DSM_MATCH_FIELD_CONTROL__SRC_EN_MASK                                                               0x00000001L
+#define IH_DSM_MATCH_FIELD_CONTROL__FCNID_EN_MASK                                                             0x00000002L
+#define IH_DSM_MATCH_FIELD_CONTROL__TIMESTAMP_EN_MASK                                                         0x00000004L
+#define IH_DSM_MATCH_FIELD_CONTROL__RINGID_EN_MASK                                                            0x00000008L
+#define IH_DSM_MATCH_FIELD_CONTROL__VMID_EN_MASK                                                              0x00000010L
+#define IH_DSM_MATCH_FIELD_CONTROL__PASID_EN_MASK                                                             0x00000020L
+#define IH_DSM_MATCH_FIELD_CONTROL__CLIENT_ID_EN_MASK                                                         0x00000040L
+//IH_DSM_MATCH_DATA_CONTROL
+#define IH_DSM_MATCH_DATA_CONTROL__VALUE__SHIFT                                                               0x0
+#define IH_DSM_MATCH_DATA_CONTROL__VALUE_MASK                                                                 0x0FFFFFFFL
+//IH_DSM_MATCH_FCN_ID
+#define IH_DSM_MATCH_FCN_ID__PF_VF__SHIFT                                                                     0x0
+#define IH_DSM_MATCH_FCN_ID__VF_ID__SHIFT                                                                     0x1
+#define IH_DSM_MATCH_FCN_ID__PF_VF_MASK                                                                       0x00000001L
+#define IH_DSM_MATCH_FCN_ID__VF_ID_MASK                                                                       0x0000001EL
+//IH_LIMIT_INT_RATE_CNTL
+#define IH_LIMIT_INT_RATE_CNTL__LIMIT_ENABLE__SHIFT                                                           0x0
+#define IH_LIMIT_INT_RATE_CNTL__PERF_INTERVAL__SHIFT                                                          0x1
+#define IH_LIMIT_INT_RATE_CNTL__PERF_THRESHOLD__SHIFT                                                         0x5
+#define IH_LIMIT_INT_RATE_CNTL__RETURN_DELAY__SHIFT                                                           0x11
+#define IH_LIMIT_INT_RATE_CNTL__PERF_RESULT__SHIFT                                                            0x15
+#define IH_LIMIT_INT_RATE_CNTL__LIMIT_ENABLE_MASK                                                             0x00000001L
+#define IH_LIMIT_INT_RATE_CNTL__PERF_INTERVAL_MASK                                                            0x0000001EL
+#define IH_LIMIT_INT_RATE_CNTL__PERF_THRESHOLD_MASK                                                           0x0000FFE0L
+#define IH_LIMIT_INT_RATE_CNTL__RETURN_DELAY_MASK                                                             0x001E0000L
+#define IH_LIMIT_INT_RATE_CNTL__PERF_RESULT_MASK                                                              0xFFE00000L
+//IH_VF_RB_STATUS
+#define IH_VF_RB_STATUS__RB_FULL_DRAIN_VF__SHIFT                                                              0x0
+#define IH_VF_RB_STATUS__RB_OVERFLOW_VF__SHIFT                                                                0x10
+#define IH_VF_RB_STATUS__RB_FULL_DRAIN_VF_MASK                                                                0x0000FFFFL
+#define IH_VF_RB_STATUS__RB_OVERFLOW_VF_MASK                                                                  0xFFFF0000L
+//IH_VF_RB_STATUS2
+#define IH_VF_RB_STATUS2__RB_FULL_VF__SHIFT                                                                   0x0
+#define IH_VF_RB_STATUS2__BIF_INTERRUPT_LINE_VF__SHIFT                                                        0x10
+#define IH_VF_RB_STATUS2__RB_FULL_VF_MASK                                                                     0x0000FFFFL
+#define IH_VF_RB_STATUS2__BIF_INTERRUPT_LINE_VF_MASK                                                          0xFFFF0000L
+//IH_VF_RB1_STATUS
+#define IH_VF_RB1_STATUS__RB_FULL_DRAIN_VF__SHIFT                                                             0x0
+#define IH_VF_RB1_STATUS__RB_OVERFLOW_VF__SHIFT                                                               0x10
+#define IH_VF_RB1_STATUS__RB_FULL_DRAIN_VF_MASK                                                               0x0000FFFFL
+#define IH_VF_RB1_STATUS__RB_OVERFLOW_VF_MASK                                                                 0xFFFF0000L
+//IH_VF_RB1_STATUS2
+#define IH_VF_RB1_STATUS2__RB_FULL_VF__SHIFT                                                                  0x0
+#define IH_VF_RB1_STATUS2__RB_FULL_VF_MASK                                                                    0x0000FFFFL
+//IH_VF_RB2_STATUS
+#define IH_VF_RB2_STATUS__RB_FULL_DRAIN_VF__SHIFT                                                             0x0
+#define IH_VF_RB2_STATUS__RB_OVERFLOW_VF__SHIFT                                                               0x10
+#define IH_VF_RB2_STATUS__RB_FULL_DRAIN_VF_MASK                                                               0x0000FFFFL
+#define IH_VF_RB2_STATUS__RB_OVERFLOW_VF_MASK                                                                 0xFFFF0000L
+//IH_VF_RB2_STATUS2
+#define IH_VF_RB2_STATUS2__RB_FULL_VF__SHIFT                                                                  0x0
+#define IH_VF_RB2_STATUS2__RB_FULL_VF_MASK                                                                    0x0000FFFFL
+//IH_INT_FLOOD_CNTL
+#define IH_INT_FLOOD_CNTL__HIGHWATER__SHIFT                                                                   0x0
+#define IH_INT_FLOOD_CNTL__FLOOD_CNTL_ENABLE__SHIFT                                                           0x3
+#define IH_INT_FLOOD_CNTL__CLEAR_INT_FLOOD_STATUS__SHIFT                                                      0x4
+#define IH_INT_FLOOD_CNTL__HIGHWATER_MASK                                                                     0x00000007L
+#define IH_INT_FLOOD_CNTL__FLOOD_CNTL_ENABLE_MASK                                                             0x00000008L
+#define IH_INT_FLOOD_CNTL__CLEAR_INT_FLOOD_STATUS_MASK                                                        0x00000010L
+//IH_RB0_INT_FLOOD_STATUS
+#define IH_RB0_INT_FLOOD_STATUS__RB_INT_DROPPED_VF__SHIFT                                                     0x0
+#define IH_RB0_INT_FLOOD_STATUS__RB_INT_DROPPED__SHIFT                                                        0x1f
+#define IH_RB0_INT_FLOOD_STATUS__RB_INT_DROPPED_VF_MASK                                                       0x0000FFFFL
+#define IH_RB0_INT_FLOOD_STATUS__RB_INT_DROPPED_MASK                                                          0x80000000L
+//IH_RB1_INT_FLOOD_STATUS
+#define IH_RB1_INT_FLOOD_STATUS__RB_INT_DROPPED_VF__SHIFT                                                     0x0
+#define IH_RB1_INT_FLOOD_STATUS__RB_INT_DROPPED__SHIFT                                                        0x1f
+#define IH_RB1_INT_FLOOD_STATUS__RB_INT_DROPPED_VF_MASK                                                       0x0000FFFFL
+#define IH_RB1_INT_FLOOD_STATUS__RB_INT_DROPPED_MASK                                                          0x80000000L
+//IH_RB2_INT_FLOOD_STATUS
+#define IH_RB2_INT_FLOOD_STATUS__RB_INT_DROPPED_VF__SHIFT                                                     0x0
+#define IH_RB2_INT_FLOOD_STATUS__RB_INT_DROPPED__SHIFT                                                        0x1f
+#define IH_RB2_INT_FLOOD_STATUS__RB_INT_DROPPED_VF_MASK                                                       0x0000FFFFL
+#define IH_RB2_INT_FLOOD_STATUS__RB_INT_DROPPED_MASK                                                          0x80000000L
+//IH_INT_FLOOD_STATUS
+#define IH_INT_FLOOD_STATUS__INT_DROP_CNT__SHIFT                                                              0x0
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_CLIENT_ID__SHIFT                                                  0x8
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_SOURCE_ID__SHIFT                                                  0x10
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_VF_ID__SHIFT                                                      0x18
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_VF__SHIFT                                                         0x1c
+#define IH_INT_FLOOD_STATUS__INT_DROPPED__SHIFT                                                               0x1e
+#define IH_INT_FLOOD_STATUS__INT_DROP_CNT_MASK                                                                0x000000FFL
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_CLIENT_ID_MASK                                                    0x0000FF00L
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_SOURCE_ID_MASK                                                    0x00FF0000L
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_VF_ID_MASK                                                        0x0F000000L
+#define IH_INT_FLOOD_STATUS__FIRST_DROP_INT_VF_MASK                                                           0x10000000L
+#define IH_INT_FLOOD_STATUS__INT_DROPPED_MASK                                                                 0x40000000L
+//IH_STORM_CLIENT_LIST_CNTL
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT1_IS_STORM_CLIENT__SHIFT                                             0x1
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT2_IS_STORM_CLIENT__SHIFT                                             0x2
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT3_IS_STORM_CLIENT__SHIFT                                             0x3
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT4_IS_STORM_CLIENT__SHIFT                                             0x4
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT5_IS_STORM_CLIENT__SHIFT                                             0x5
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT6_IS_STORM_CLIENT__SHIFT                                             0x6
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT7_IS_STORM_CLIENT__SHIFT                                             0x7
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT8_IS_STORM_CLIENT__SHIFT                                             0x8
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT9_IS_STORM_CLIENT__SHIFT                                             0x9
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT10_IS_STORM_CLIENT__SHIFT                                            0xa
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT11_IS_STORM_CLIENT__SHIFT                                            0xb
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT12_IS_STORM_CLIENT__SHIFT                                            0xc
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT13_IS_STORM_CLIENT__SHIFT                                            0xd
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT14_IS_STORM_CLIENT__SHIFT                                            0xe
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT15_IS_STORM_CLIENT__SHIFT                                            0xf
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT16_IS_STORM_CLIENT__SHIFT                                            0x10
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT17_IS_STORM_CLIENT__SHIFT                                            0x11
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT18_IS_STORM_CLIENT__SHIFT                                            0x12
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT19_IS_STORM_CLIENT__SHIFT                                            0x13
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT20_IS_STORM_CLIENT__SHIFT                                            0x14
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT21_IS_STORM_CLIENT__SHIFT                                            0x15
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT22_IS_STORM_CLIENT__SHIFT                                            0x16
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT23_IS_STORM_CLIENT__SHIFT                                            0x17
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT24_IS_STORM_CLIENT__SHIFT                                            0x18
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT25_IS_STORM_CLIENT__SHIFT                                            0x19
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT26_IS_STORM_CLIENT__SHIFT                                            0x1a
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT27_IS_STORM_CLIENT__SHIFT                                            0x1b
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT28_IS_STORM_CLIENT__SHIFT                                            0x1c
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT29_IS_STORM_CLIENT__SHIFT                                            0x1d
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT30_IS_STORM_CLIENT__SHIFT                                            0x1e
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT31_IS_STORM_CLIENT__SHIFT                                            0x1f
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT1_IS_STORM_CLIENT_MASK                                               0x00000002L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT2_IS_STORM_CLIENT_MASK                                               0x00000004L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT3_IS_STORM_CLIENT_MASK                                               0x00000008L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT4_IS_STORM_CLIENT_MASK                                               0x00000010L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT5_IS_STORM_CLIENT_MASK                                               0x00000020L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT6_IS_STORM_CLIENT_MASK                                               0x00000040L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT7_IS_STORM_CLIENT_MASK                                               0x00000080L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT8_IS_STORM_CLIENT_MASK                                               0x00000100L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT9_IS_STORM_CLIENT_MASK                                               0x00000200L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT10_IS_STORM_CLIENT_MASK                                              0x00000400L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT11_IS_STORM_CLIENT_MASK                                              0x00000800L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT12_IS_STORM_CLIENT_MASK                                              0x00001000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT13_IS_STORM_CLIENT_MASK                                              0x00002000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT14_IS_STORM_CLIENT_MASK                                              0x00004000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT15_IS_STORM_CLIENT_MASK                                              0x00008000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT16_IS_STORM_CLIENT_MASK                                              0x00010000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT17_IS_STORM_CLIENT_MASK                                              0x00020000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT18_IS_STORM_CLIENT_MASK                                              0x00040000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT19_IS_STORM_CLIENT_MASK                                              0x00080000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT20_IS_STORM_CLIENT_MASK                                              0x00100000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT21_IS_STORM_CLIENT_MASK                                              0x00200000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT22_IS_STORM_CLIENT_MASK                                              0x00400000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT23_IS_STORM_CLIENT_MASK                                              0x00800000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT24_IS_STORM_CLIENT_MASK                                              0x01000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT25_IS_STORM_CLIENT_MASK                                              0x02000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT26_IS_STORM_CLIENT_MASK                                              0x04000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT27_IS_STORM_CLIENT_MASK                                              0x08000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT28_IS_STORM_CLIENT_MASK                                              0x10000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT29_IS_STORM_CLIENT_MASK                                              0x20000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT30_IS_STORM_CLIENT_MASK                                              0x40000000L
+#define IH_STORM_CLIENT_LIST_CNTL__CLIENT31_IS_STORM_CLIENT_MASK                                              0x80000000L
+//IH_CLK_CTRL
+#define IH_CLK_CTRL__IH_RETRY_INT_CAM_MEM_CLK_SOFT_OVERRIDE__SHIFT                                            0x19
+#define IH_CLK_CTRL__IH_BUFFER_MEM_CLK_SOFT_OVERRIDE__SHIFT                                                   0x1a
+#define IH_CLK_CTRL__DBUS_MUX_CLK_SOFT_OVERRIDE__SHIFT                                                        0x1b
+#define IH_CLK_CTRL__OSSSYS_SHARE_CLK_SOFT_OVERRIDE__SHIFT                                                    0x1c
+#define IH_CLK_CTRL__LIMIT_SMN_CLK_SOFT_OVERRIDE__SHIFT                                                       0x1d
+#define IH_CLK_CTRL__DYN_CLK_SOFT_OVERRIDE__SHIFT                                                             0x1e
+#define IH_CLK_CTRL__REG_CLK_SOFT_OVERRIDE__SHIFT                                                             0x1f
+#define IH_CLK_CTRL__IH_RETRY_INT_CAM_MEM_CLK_SOFT_OVERRIDE_MASK                                              0x02000000L
+#define IH_CLK_CTRL__IH_BUFFER_MEM_CLK_SOFT_OVERRIDE_MASK                                                     0x04000000L
+#define IH_CLK_CTRL__DBUS_MUX_CLK_SOFT_OVERRIDE_MASK                                                          0x08000000L
+#define IH_CLK_CTRL__OSSSYS_SHARE_CLK_SOFT_OVERRIDE_MASK                                                      0x10000000L
+#define IH_CLK_CTRL__LIMIT_SMN_CLK_SOFT_OVERRIDE_MASK                                                         0x20000000L
+#define IH_CLK_CTRL__DYN_CLK_SOFT_OVERRIDE_MASK                                                               0x40000000L
+#define IH_CLK_CTRL__REG_CLK_SOFT_OVERRIDE_MASK                                                               0x80000000L
+//IH_INT_FLAGS
+#define IH_INT_FLAGS__CLIENT_0_FLAG__SHIFT                                                                    0x0
+#define IH_INT_FLAGS__CLIENT_1_FLAG__SHIFT                                                                    0x1
+#define IH_INT_FLAGS__CLIENT_2_FLAG__SHIFT                                                                    0x2
+#define IH_INT_FLAGS__CLIENT_3_FLAG__SHIFT                                                                    0x3
+#define IH_INT_FLAGS__CLIENT_4_FLAG__SHIFT                                                                    0x4
+#define IH_INT_FLAGS__CLIENT_5_FLAG__SHIFT                                                                    0x5
+#define IH_INT_FLAGS__CLIENT_6_FLAG__SHIFT                                                                    0x6
+#define IH_INT_FLAGS__CLIENT_7_FLAG__SHIFT                                                                    0x7
+#define IH_INT_FLAGS__CLIENT_8_FLAG__SHIFT                                                                    0x8
+#define IH_INT_FLAGS__CLIENT_9_FLAG__SHIFT                                                                    0x9
+#define IH_INT_FLAGS__CLIENT_10_FLAG__SHIFT                                                                   0xa
+#define IH_INT_FLAGS__CLIENT_11_FLAG__SHIFT                                                                   0xb
+#define IH_INT_FLAGS__CLIENT_12_FLAG__SHIFT                                                                   0xc
+#define IH_INT_FLAGS__CLIENT_13_FLAG__SHIFT                                                                   0xd
+#define IH_INT_FLAGS__CLIENT_14_FLAG__SHIFT                                                                   0xe
+#define IH_INT_FLAGS__CLIENT_15_FLAG__SHIFT                                                                   0xf
+#define IH_INT_FLAGS__CLIENT_16_FLAG__SHIFT                                                                   0x10
+#define IH_INT_FLAGS__CLIENT_17_FLAG__SHIFT                                                                   0x11
+#define IH_INT_FLAGS__CLIENT_18_FLAG__SHIFT                                                                   0x12
+#define IH_INT_FLAGS__CLIENT_19_FLAG__SHIFT                                                                   0x13
+#define IH_INT_FLAGS__CLIENT_20_FLAG__SHIFT                                                                   0x14
+#define IH_INT_FLAGS__CLIENT_21_FLAG__SHIFT                                                                   0x15
+#define IH_INT_FLAGS__CLIENT_22_FLAG__SHIFT                                                                   0x16
+#define IH_INT_FLAGS__CLIENT_23_FLAG__SHIFT                                                                   0x17
+#define IH_INT_FLAGS__CLIENT_24_FLAG__SHIFT                                                                   0x18
+#define IH_INT_FLAGS__CLIENT_25_FLAG__SHIFT                                                                   0x19
+#define IH_INT_FLAGS__CLIENT_26_FLAG__SHIFT                                                                   0x1a
+#define IH_INT_FLAGS__CLIENT_27_FLAG__SHIFT                                                                   0x1b
+#define IH_INT_FLAGS__CLIENT_28_FLAG__SHIFT                                                                   0x1c
+#define IH_INT_FLAGS__CLIENT_29_FLAG__SHIFT                                                                   0x1d
+#define IH_INT_FLAGS__CLIENT_30_FLAG__SHIFT                                                                   0x1e
+#define IH_INT_FLAGS__CLIENT_31_FLAG__SHIFT                                                                   0x1f
+#define IH_INT_FLAGS__CLIENT_0_FLAG_MASK                                                                      0x00000001L
+#define IH_INT_FLAGS__CLIENT_1_FLAG_MASK                                                                      0x00000002L
+#define IH_INT_FLAGS__CLIENT_2_FLAG_MASK                                                                      0x00000004L
+#define IH_INT_FLAGS__CLIENT_3_FLAG_MASK                                                                      0x00000008L
+#define IH_INT_FLAGS__CLIENT_4_FLAG_MASK                                                                      0x00000010L
+#define IH_INT_FLAGS__CLIENT_5_FLAG_MASK                                                                      0x00000020L
+#define IH_INT_FLAGS__CLIENT_6_FLAG_MASK                                                                      0x00000040L
+#define IH_INT_FLAGS__CLIENT_7_FLAG_MASK                                                                      0x00000080L
+#define IH_INT_FLAGS__CLIENT_8_FLAG_MASK                                                                      0x00000100L
+#define IH_INT_FLAGS__CLIENT_9_FLAG_MASK                                                                      0x00000200L
+#define IH_INT_FLAGS__CLIENT_10_FLAG_MASK                                                                     0x00000400L
+#define IH_INT_FLAGS__CLIENT_11_FLAG_MASK                                                                     0x00000800L
+#define IH_INT_FLAGS__CLIENT_12_FLAG_MASK                                                                     0x00001000L
+#define IH_INT_FLAGS__CLIENT_13_FLAG_MASK                                                                     0x00002000L
+#define IH_INT_FLAGS__CLIENT_14_FLAG_MASK                                                                     0x00004000L
+#define IH_INT_FLAGS__CLIENT_15_FLAG_MASK                                                                     0x00008000L
+#define IH_INT_FLAGS__CLIENT_16_FLAG_MASK                                                                     0x00010000L
+#define IH_INT_FLAGS__CLIENT_17_FLAG_MASK                                                                     0x00020000L
+#define IH_INT_FLAGS__CLIENT_18_FLAG_MASK                                                                     0x00040000L
+#define IH_INT_FLAGS__CLIENT_19_FLAG_MASK                                                                     0x00080000L
+#define IH_INT_FLAGS__CLIENT_20_FLAG_MASK                                                                     0x00100000L
+#define IH_INT_FLAGS__CLIENT_21_FLAG_MASK                                                                     0x00200000L
+#define IH_INT_FLAGS__CLIENT_22_FLAG_MASK                                                                     0x00400000L
+#define IH_INT_FLAGS__CLIENT_23_FLAG_MASK                                                                     0x00800000L
+#define IH_INT_FLAGS__CLIENT_24_FLAG_MASK                                                                     0x01000000L
+#define IH_INT_FLAGS__CLIENT_25_FLAG_MASK                                                                     0x02000000L
+#define IH_INT_FLAGS__CLIENT_26_FLAG_MASK                                                                     0x04000000L
+#define IH_INT_FLAGS__CLIENT_27_FLAG_MASK                                                                     0x08000000L
+#define IH_INT_FLAGS__CLIENT_28_FLAG_MASK                                                                     0x10000000L
+#define IH_INT_FLAGS__CLIENT_29_FLAG_MASK                                                                     0x20000000L
+#define IH_INT_FLAGS__CLIENT_30_FLAG_MASK                                                                     0x40000000L
+#define IH_INT_FLAGS__CLIENT_31_FLAG_MASK                                                                     0x80000000L
+//IH_LAST_INT_INFO0
+#define IH_LAST_INT_INFO0__CLIENT_ID__SHIFT                                                                   0x0
+#define IH_LAST_INT_INFO0__SOURCE_ID__SHIFT                                                                   0x8
+#define IH_LAST_INT_INFO0__RING_ID__SHIFT                                                                     0x10
+#define IH_LAST_INT_INFO0__VM_ID__SHIFT                                                                       0x18
+#define IH_LAST_INT_INFO0__VMID_TYPE__SHIFT                                                                   0x1f
+#define IH_LAST_INT_INFO0__CLIENT_ID_MASK                                                                     0x000000FFL
+#define IH_LAST_INT_INFO0__SOURCE_ID_MASK                                                                     0x0000FF00L
+#define IH_LAST_INT_INFO0__RING_ID_MASK                                                                       0x00FF0000L
+#define IH_LAST_INT_INFO0__VM_ID_MASK                                                                         0x0F000000L
+#define IH_LAST_INT_INFO0__VMID_TYPE_MASK                                                                     0x80000000L
+//IH_LAST_INT_INFO1
+#define IH_LAST_INT_INFO1__CONTEXT_ID__SHIFT                                                                  0x0
+#define IH_LAST_INT_INFO1__CONTEXT_ID_MASK                                                                    0xFFFFFFFFL
+//IH_LAST_INT_INFO2
+#define IH_LAST_INT_INFO2__PAS_ID__SHIFT                                                                      0x0
+#define IH_LAST_INT_INFO2__VF_ID__SHIFT                                                                       0x10
+#define IH_LAST_INT_INFO2__VF__SHIFT                                                                          0x14
+#define IH_LAST_INT_INFO2__PAS_ID_MASK                                                                        0x0000FFFFL
+#define IH_LAST_INT_INFO2__VF_ID_MASK                                                                         0x000F0000L
+#define IH_LAST_INT_INFO2__VF_MASK                                                                            0x00100000L
+//IH_SCRATCH
+#define IH_SCRATCH__DATA__SHIFT                                                                               0x0
+#define IH_SCRATCH__DATA_MASK                                                                                 0xFFFFFFFFL
+//IH_CLIENT_CREDIT_ERROR
+#define IH_CLIENT_CREDIT_ERROR__CLEAR__SHIFT                                                                  0x0
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_1_ERROR__SHIFT                                                         0x1
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_2_ERROR__SHIFT                                                         0x2
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_3_ERROR__SHIFT                                                         0x3
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_4_ERROR__SHIFT                                                         0x4
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_5_ERROR__SHIFT                                                         0x5
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_6_ERROR__SHIFT                                                         0x6
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_7_ERROR__SHIFT                                                         0x7
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_8_ERROR__SHIFT                                                         0x8
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_9_ERROR__SHIFT                                                         0x9
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_10_ERROR__SHIFT                                                        0xa
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_11_ERROR__SHIFT                                                        0xb
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_12_ERROR__SHIFT                                                        0xc
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_13_ERROR__SHIFT                                                        0xd
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_14_ERROR__SHIFT                                                        0xe
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_15_ERROR__SHIFT                                                        0xf
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_16_ERROR__SHIFT                                                        0x10
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_17_ERROR__SHIFT                                                        0x11
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_18_ERROR__SHIFT                                                        0x12
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_19_ERROR__SHIFT                                                        0x13
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_20_ERROR__SHIFT                                                        0x14
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_21_ERROR__SHIFT                                                        0x15
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_22_ERROR__SHIFT                                                        0x16
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_23_ERROR__SHIFT                                                        0x17
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_24_ERROR__SHIFT                                                        0x18
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_25_ERROR__SHIFT                                                        0x19
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_26_ERROR__SHIFT                                                        0x1a
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_27_ERROR__SHIFT                                                        0x1b
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_28_ERROR__SHIFT                                                        0x1c
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_29_ERROR__SHIFT                                                        0x1d
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_30_ERROR__SHIFT                                                        0x1e
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_31_ERROR__SHIFT                                                        0x1f
+#define IH_CLIENT_CREDIT_ERROR__CLEAR_MASK                                                                    0x00000001L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_1_ERROR_MASK                                                           0x00000002L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_2_ERROR_MASK                                                           0x00000004L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_3_ERROR_MASK                                                           0x00000008L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_4_ERROR_MASK                                                           0x00000010L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_5_ERROR_MASK                                                           0x00000020L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_6_ERROR_MASK                                                           0x00000040L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_7_ERROR_MASK                                                           0x00000080L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_8_ERROR_MASK                                                           0x00000100L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_9_ERROR_MASK                                                           0x00000200L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_10_ERROR_MASK                                                          0x00000400L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_11_ERROR_MASK                                                          0x00000800L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_12_ERROR_MASK                                                          0x00001000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_13_ERROR_MASK                                                          0x00002000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_14_ERROR_MASK                                                          0x00004000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_15_ERROR_MASK                                                          0x00008000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_16_ERROR_MASK                                                          0x00010000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_17_ERROR_MASK                                                          0x00020000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_18_ERROR_MASK                                                          0x00040000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_19_ERROR_MASK                                                          0x00080000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_20_ERROR_MASK                                                          0x00100000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_21_ERROR_MASK                                                          0x00200000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_22_ERROR_MASK                                                          0x00400000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_23_ERROR_MASK                                                          0x00800000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_24_ERROR_MASK                                                          0x01000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_25_ERROR_MASK                                                          0x02000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_26_ERROR_MASK                                                          0x04000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_27_ERROR_MASK                                                          0x08000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_28_ERROR_MASK                                                          0x10000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_29_ERROR_MASK                                                          0x20000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_30_ERROR_MASK                                                          0x40000000L
+#define IH_CLIENT_CREDIT_ERROR__CLIENT_31_ERROR_MASK                                                          0x80000000L
+//IH_GPU_IOV_VIOLATION_LOG
+#define IH_GPU_IOV_VIOLATION_LOG__VIOLATION_STATUS__SHIFT                                                     0x0
+#define IH_GPU_IOV_VIOLATION_LOG__MULTIPLE_VIOLATION_STATUS__SHIFT                                            0x1
+#define IH_GPU_IOV_VIOLATION_LOG__ADDRESS__SHIFT                                                              0x2
+#define IH_GPU_IOV_VIOLATION_LOG__OPCODE__SHIFT                                                               0x12
+#define IH_GPU_IOV_VIOLATION_LOG__VF__SHIFT                                                                   0x13
+#define IH_GPU_IOV_VIOLATION_LOG__VF_ID__SHIFT                                                                0x14
+#define IH_GPU_IOV_VIOLATION_LOG__INITIATOR_ID__SHIFT                                                         0x18
+#define IH_GPU_IOV_VIOLATION_LOG__VIOLATION_STATUS_MASK                                                       0x00000001L
+#define IH_GPU_IOV_VIOLATION_LOG__MULTIPLE_VIOLATION_STATUS_MASK                                              0x00000002L
+#define IH_GPU_IOV_VIOLATION_LOG__ADDRESS_MASK                                                                0x0003FFFCL
+#define IH_GPU_IOV_VIOLATION_LOG__OPCODE_MASK                                                                 0x00040000L
+#define IH_GPU_IOV_VIOLATION_LOG__VF_MASK                                                                     0x00080000L
+#define IH_GPU_IOV_VIOLATION_LOG__VF_ID_MASK                                                                  0x00F00000L
+#define IH_GPU_IOV_VIOLATION_LOG__INITIATOR_ID_MASK                                                           0xFF000000L
+//IH_COOKIE_REC_VIOLATION_LOG
+#define IH_COOKIE_REC_VIOLATION_LOG__VIOLATION_STATUS__SHIFT                                                  0x0
+#define IH_COOKIE_REC_VIOLATION_LOG__CLIENT_ID__SHIFT                                                         0x10
+#define IH_COOKIE_REC_VIOLATION_LOG__INITIATOR_ID__SHIFT                                                      0x18
+#define IH_COOKIE_REC_VIOLATION_LOG__VIOLATION_STATUS_MASK                                                    0x00000001L
+#define IH_COOKIE_REC_VIOLATION_LOG__CLIENT_ID_MASK                                                           0x00FF0000L
+#define IH_COOKIE_REC_VIOLATION_LOG__INITIATOR_ID_MASK                                                        0xFF000000L
+//IH_CREDIT_STATUS
+#define IH_CREDIT_STATUS__CLIENT_1_CREDIT_RETURNED__SHIFT                                                     0x1
+#define IH_CREDIT_STATUS__CLIENT_2_CREDIT_RETURNED__SHIFT                                                     0x2
+#define IH_CREDIT_STATUS__CLIENT_3_CREDIT_RETURNED__SHIFT                                                     0x3
+#define IH_CREDIT_STATUS__CLIENT_4_CREDIT_RETURNED__SHIFT                                                     0x4
+#define IH_CREDIT_STATUS__CLIENT_5_CREDIT_RETURNED__SHIFT                                                     0x5
+#define IH_CREDIT_STATUS__CLIENT_6_CREDIT_RETURNED__SHIFT                                                     0x6
+#define IH_CREDIT_STATUS__CLIENT_7_CREDIT_RETURNED__SHIFT                                                     0x7
+#define IH_CREDIT_STATUS__CLIENT_8_CREDIT_RETURNED__SHIFT                                                     0x8
+#define IH_CREDIT_STATUS__CLIENT_9_CREDIT_RETURNED__SHIFT                                                     0x9
+#define IH_CREDIT_STATUS__CLIENT_10_CREDIT_RETURNED__SHIFT                                                    0xa
+#define IH_CREDIT_STATUS__CLIENT_11_CREDIT_RETURNED__SHIFT                                                    0xb
+#define IH_CREDIT_STATUS__CLIENT_12_CREDIT_RETURNED__SHIFT                                                    0xc
+#define IH_CREDIT_STATUS__CLIENT_13_CREDIT_RETURNED__SHIFT                                                    0xd
+#define IH_CREDIT_STATUS__CLIENT_14_CREDIT_RETURNED__SHIFT                                                    0xe
+#define IH_CREDIT_STATUS__CLIENT_15_CREDIT_RETURNED__SHIFT                                                    0xf
+#define IH_CREDIT_STATUS__CLIENT_16_CREDIT_RETURNED__SHIFT                                                    0x10
+#define IH_CREDIT_STATUS__CLIENT_17_CREDIT_RETURNED__SHIFT                                                    0x11
+#define IH_CREDIT_STATUS__CLIENT_18_CREDIT_RETURNED__SHIFT                                                    0x12
+#define IH_CREDIT_STATUS__CLIENT_19_CREDIT_RETURNED__SHIFT                                                    0x13
+#define IH_CREDIT_STATUS__CLIENT_20_CREDIT_RETURNED__SHIFT                                                    0x14
+#define IH_CREDIT_STATUS__CLIENT_21_CREDIT_RETURNED__SHIFT                                                    0x15
+#define IH_CREDIT_STATUS__CLIENT_22_CREDIT_RETURNED__SHIFT                                                    0x16
+#define IH_CREDIT_STATUS__CLIENT_23_CREDIT_RETURNED__SHIFT                                                    0x17
+#define IH_CREDIT_STATUS__CLIENT_24_CREDIT_RETURNED__SHIFT                                                    0x18
+#define IH_CREDIT_STATUS__CLIENT_25_CREDIT_RETURNED__SHIFT                                                    0x19
+#define IH_CREDIT_STATUS__CLIENT_26_CREDIT_RETURNED__SHIFT                                                    0x1a
+#define IH_CREDIT_STATUS__CLIENT_27_CREDIT_RETURNED__SHIFT                                                    0x1b
+#define IH_CREDIT_STATUS__CLIENT_28_CREDIT_RETURNED__SHIFT                                                    0x1c
+#define IH_CREDIT_STATUS__CLIENT_29_CREDIT_RETURNED__SHIFT                                                    0x1d
+#define IH_CREDIT_STATUS__CLIENT_30_CREDIT_RETURNED__SHIFT                                                    0x1e
+#define IH_CREDIT_STATUS__CLIENT_31_CREDIT_RETURNED__SHIFT                                                    0x1f
+#define IH_CREDIT_STATUS__CLIENT_1_CREDIT_RETURNED_MASK                                                       0x00000002L
+#define IH_CREDIT_STATUS__CLIENT_2_CREDIT_RETURNED_MASK                                                       0x00000004L
+#define IH_CREDIT_STATUS__CLIENT_3_CREDIT_RETURNED_MASK                                                       0x00000008L
+#define IH_CREDIT_STATUS__CLIENT_4_CREDIT_RETURNED_MASK                                                       0x00000010L
+#define IH_CREDIT_STATUS__CLIENT_5_CREDIT_RETURNED_MASK                                                       0x00000020L
+#define IH_CREDIT_STATUS__CLIENT_6_CREDIT_RETURNED_MASK                                                       0x00000040L
+#define IH_CREDIT_STATUS__CLIENT_7_CREDIT_RETURNED_MASK                                                       0x00000080L
+#define IH_CREDIT_STATUS__CLIENT_8_CREDIT_RETURNED_MASK                                                       0x00000100L
+#define IH_CREDIT_STATUS__CLIENT_9_CREDIT_RETURNED_MASK                                                       0x00000200L
+#define IH_CREDIT_STATUS__CLIENT_10_CREDIT_RETURNED_MASK                                                      0x00000400L
+#define IH_CREDIT_STATUS__CLIENT_11_CREDIT_RETURNED_MASK                                                      0x00000800L
+#define IH_CREDIT_STATUS__CLIENT_12_CREDIT_RETURNED_MASK                                                      0x00001000L
+#define IH_CREDIT_STATUS__CLIENT_13_CREDIT_RETURNED_MASK                                                      0x00002000L
+#define IH_CREDIT_STATUS__CLIENT_14_CREDIT_RETURNED_MASK                                                      0x00004000L
+#define IH_CREDIT_STATUS__CLIENT_15_CREDIT_RETURNED_MASK                                                      0x00008000L
+#define IH_CREDIT_STATUS__CLIENT_16_CREDIT_RETURNED_MASK                                                      0x00010000L
+#define IH_CREDIT_STATUS__CLIENT_17_CREDIT_RETURNED_MASK                                                      0x00020000L
+#define IH_CREDIT_STATUS__CLIENT_18_CREDIT_RETURNED_MASK                                                      0x00040000L
+#define IH_CREDIT_STATUS__CLIENT_19_CREDIT_RETURNED_MASK                                                      0x00080000L
+#define IH_CREDIT_STATUS__CLIENT_20_CREDIT_RETURNED_MASK                                                      0x00100000L
+#define IH_CREDIT_STATUS__CLIENT_21_CREDIT_RETURNED_MASK                                                      0x00200000L
+#define IH_CREDIT_STATUS__CLIENT_22_CREDIT_RETURNED_MASK                                                      0x00400000L
+#define IH_CREDIT_STATUS__CLIENT_23_CREDIT_RETURNED_MASK                                                      0x00800000L
+#define IH_CREDIT_STATUS__CLIENT_24_CREDIT_RETURNED_MASK                                                      0x01000000L
+#define IH_CREDIT_STATUS__CLIENT_25_CREDIT_RETURNED_MASK                                                      0x02000000L
+#define IH_CREDIT_STATUS__CLIENT_26_CREDIT_RETURNED_MASK                                                      0x04000000L
+#define IH_CREDIT_STATUS__CLIENT_27_CREDIT_RETURNED_MASK                                                      0x08000000L
+#define IH_CREDIT_STATUS__CLIENT_28_CREDIT_RETURNED_MASK                                                      0x10000000L
+#define IH_CREDIT_STATUS__CLIENT_29_CREDIT_RETURNED_MASK                                                      0x20000000L
+#define IH_CREDIT_STATUS__CLIENT_30_CREDIT_RETURNED_MASK                                                      0x40000000L
+#define IH_CREDIT_STATUS__CLIENT_31_CREDIT_RETURNED_MASK                                                      0x80000000L
+//IH_MMHUB_ERROR
+#define IH_MMHUB_ERROR__IH_BRESP_01__SHIFT                                                                    0x1
+#define IH_MMHUB_ERROR__IH_BRESP_10__SHIFT                                                                    0x2
+#define IH_MMHUB_ERROR__IH_BRESP_11__SHIFT                                                                    0x3
+#define IH_MMHUB_ERROR__IH_BUSER_NACK_01__SHIFT                                                               0x5
+#define IH_MMHUB_ERROR__IH_BUSER_NACK_10__SHIFT                                                               0x6
+#define IH_MMHUB_ERROR__IH_BUSER_NACK_11__SHIFT                                                               0x7
+#define IH_MMHUB_ERROR__IH_BRESP_01_MASK                                                                      0x00000002L
+#define IH_MMHUB_ERROR__IH_BRESP_10_MASK                                                                      0x00000004L
+#define IH_MMHUB_ERROR__IH_BRESP_11_MASK                                                                      0x00000008L
+#define IH_MMHUB_ERROR__IH_BUSER_NACK_01_MASK                                                                 0x00000020L
+#define IH_MMHUB_ERROR__IH_BUSER_NACK_10_MASK                                                                 0x00000040L
+#define IH_MMHUB_ERROR__IH_BUSER_NACK_11_MASK                                                                 0x00000080L
+//IH_MEM_POWER_CTRL
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_CTRL_EN__SHIFT                                                 0x0
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_LS_EN__SHIFT                                                   0x1
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_DS_EN__SHIFT                                                   0x2
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_SD_EN__SHIFT                                                   0x3
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_IDLE_HYSTERESIS__SHIFT                                               0x4
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_UP_RECOVER_DELAY__SHIFT                                        0x8
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_DOWN_LS_ENTER_DELAY__SHIFT                                     0xe
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_CTRL_EN_MASK                                                   0x00000001L
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_LS_EN_MASK                                                     0x00000002L
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_DS_EN_MASK                                                     0x00000004L
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_SD_EN_MASK                                                     0x00000008L
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_IDLE_HYSTERESIS_MASK                                                 0x00000070L
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_UP_RECOVER_DELAY_MASK                                          0x00003F00L
+#define IH_MEM_POWER_CTRL__IH_BUFFER_MEM_POWER_DOWN_LS_ENTER_DELAY_MASK                                       0x0000C000L
+//IH_REGISTER_LAST_PART2
+#define IH_REGISTER_LAST_PART2__RESERVED__SHIFT                                                               0x0
+#define IH_REGISTER_LAST_PART2__RESERVED_MASK                                                                 0xFFFFFFFFL
+//SEM_CLK_CTRL
+#define SEM_CLK_CTRL__ON_DELAY__SHIFT                                                                         0x0
+#define SEM_CLK_CTRL__OFF_HYSTERESIS__SHIFT                                                                   0x4
+#define SEM_CLK_CTRL__RESERVED__SHIFT                                                                         0xc
+#define SEM_CLK_CTRL__SOFT_OVERRIDE7__SHIFT                                                                   0x18
+#define SEM_CLK_CTRL__SOFT_OVERRIDE6__SHIFT                                                                   0x19
+#define SEM_CLK_CTRL__MEM_CLK_SOFT_OVERRIDE__SHIFT                                                            0x1a
+#define SEM_CLK_CTRL__SOFT_OVERRIDE4__SHIFT                                                                   0x1b
+#define SEM_CLK_CTRL__SOFT_OVERRIDE3__SHIFT                                                                   0x1c
+#define SEM_CLK_CTRL__SOFT_OVERRIDE2__SHIFT                                                                   0x1d
+#define SEM_CLK_CTRL__DYN_CLK_SOFT_OVERRIDE__SHIFT                                                            0x1e
+#define SEM_CLK_CTRL__REG_CLK_SOFT_OVERRIDE__SHIFT                                                            0x1f
+#define SEM_CLK_CTRL__ON_DELAY_MASK                                                                           0x0000000FL
+#define SEM_CLK_CTRL__OFF_HYSTERESIS_MASK                                                                     0x00000FF0L
+#define SEM_CLK_CTRL__RESERVED_MASK                                                                           0x00FFF000L
+#define SEM_CLK_CTRL__SOFT_OVERRIDE7_MASK                                                                     0x01000000L
+#define SEM_CLK_CTRL__SOFT_OVERRIDE6_MASK                                                                     0x02000000L
+#define SEM_CLK_CTRL__MEM_CLK_SOFT_OVERRIDE_MASK                                                              0x04000000L
+#define SEM_CLK_CTRL__SOFT_OVERRIDE4_MASK                                                                     0x08000000L
+#define SEM_CLK_CTRL__SOFT_OVERRIDE3_MASK                                                                     0x10000000L
+#define SEM_CLK_CTRL__SOFT_OVERRIDE2_MASK                                                                     0x20000000L
+#define SEM_CLK_CTRL__DYN_CLK_SOFT_OVERRIDE_MASK                                                              0x40000000L
+#define SEM_CLK_CTRL__REG_CLK_SOFT_OVERRIDE_MASK                                                              0x80000000L
+//SEM_UTC_CREDIT
+#define SEM_UTC_CREDIT__UTCL2_CREDIT__SHIFT                                                                   0x0
+#define SEM_UTC_CREDIT__WATERMARK__SHIFT                                                                      0x8
+#define SEM_UTC_CREDIT__UTCL2_CREDIT_MASK                                                                     0x0000001FL
+#define SEM_UTC_CREDIT__WATERMARK_MASK                                                                        0x00000F00L
+//SEM_UTC_CONFIG
+#define SEM_UTC_CONFIG__USE_MTYPE__SHIFT                                                                      0x0
+#define SEM_UTC_CONFIG__FORCE_SNOOP__SHIFT                                                                    0x3
+#define SEM_UTC_CONFIG__FORCE_GCC__SHIFT                                                                      0x4
+#define SEM_UTC_CONFIG__USE_PT_SNOOP__SHIFT                                                                   0x5
+#define SEM_UTC_CONFIG__USE_MTYPE_MASK                                                                        0x00000007L
+#define SEM_UTC_CONFIG__FORCE_SNOOP_MASK                                                                      0x00000008L
+#define SEM_UTC_CONFIG__FORCE_GCC_MASK                                                                        0x00000010L
+#define SEM_UTC_CONFIG__USE_PT_SNOOP_MASK                                                                     0x00000020L
+//SEM_UTCL2_TRAN_EN_LUT
+#define SEM_UTCL2_TRAN_EN_LUT__SDMA0_UTCL2_EN__SHIFT                                                          0x0
+#define SEM_UTCL2_TRAN_EN_LUT__SDMA1_UTCL2_EN__SHIFT                                                          0x1
+#define SEM_UTCL2_TRAN_EN_LUT__UVD_UTCL2_EN__SHIFT                                                            0x2
+#define SEM_UTCL2_TRAN_EN_LUT__VCE0_UTCL2_EN__SHIFT                                                           0x3
+#define SEM_UTCL2_TRAN_EN_LUT__ACP_UTCL2_EN__SHIFT                                                            0x4
+#define SEM_UTCL2_TRAN_EN_LUT__ISP_UTCL2_EN__SHIFT                                                            0x5
+#define SEM_UTCL2_TRAN_EN_LUT__VCE1_UTCL2_EN__SHIFT                                                           0x6
+#define SEM_UTCL2_TRAN_EN_LUT__VP8_UTCL2_EN__SHIFT                                                            0x7
+#define SEM_UTCL2_TRAN_EN_LUT__UVD1_UTCL2_EN__SHIFT                                                           0x8
+#define SEM_UTCL2_TRAN_EN_LUT__RESERVED__SHIFT                                                                0x9
+#define SEM_UTCL2_TRAN_EN_LUT__CP_UTCL2_EN__SHIFT                                                             0x1f
+#define SEM_UTCL2_TRAN_EN_LUT__SDMA0_UTCL2_EN_MASK                                                            0x00000001L
+#define SEM_UTCL2_TRAN_EN_LUT__SDMA1_UTCL2_EN_MASK                                                            0x00000002L
+#define SEM_UTCL2_TRAN_EN_LUT__UVD_UTCL2_EN_MASK                                                              0x00000004L
+#define SEM_UTCL2_TRAN_EN_LUT__VCE0_UTCL2_EN_MASK                                                             0x00000008L
+#define SEM_UTCL2_TRAN_EN_LUT__ACP_UTCL2_EN_MASK                                                              0x00000010L
+#define SEM_UTCL2_TRAN_EN_LUT__ISP_UTCL2_EN_MASK                                                              0x00000020L
+#define SEM_UTCL2_TRAN_EN_LUT__VCE1_UTCL2_EN_MASK                                                             0x00000040L
+#define SEM_UTCL2_TRAN_EN_LUT__VP8_UTCL2_EN_MASK                                                              0x00000080L
+#define SEM_UTCL2_TRAN_EN_LUT__UVD1_UTCL2_EN_MASK                                                             0x00000100L
+#define SEM_UTCL2_TRAN_EN_LUT__RESERVED_MASK                                                                  0x7FFFFE00L
+#define SEM_UTCL2_TRAN_EN_LUT__CP_UTCL2_EN_MASK                                                               0x80000000L
+//SEM_MCIF_CONFIG
+#define SEM_MCIF_CONFIG__MC_REQ_SWAP__SHIFT                                                                   0x0
+#define SEM_MCIF_CONFIG__MC_WRREQ_CREDIT__SHIFT                                                               0x2
+#define SEM_MCIF_CONFIG__MC_RDREQ_CREDIT__SHIFT                                                               0x8
+#define SEM_MCIF_CONFIG__MC_REQ_SWAP_MASK                                                                     0x00000003L
+#define SEM_MCIF_CONFIG__MC_WRREQ_CREDIT_MASK                                                                 0x000000FCL
+#define SEM_MCIF_CONFIG__MC_RDREQ_CREDIT_MASK                                                                 0x00003F00L
+//SEM_PERFMON_CNTL
+#define SEM_PERFMON_CNTL__PERF_ENABLE0__SHIFT                                                                 0x0
+#define SEM_PERFMON_CNTL__PERF_CLEAR0__SHIFT                                                                  0x1
+#define SEM_PERFMON_CNTL__PERF_SEL0__SHIFT                                                                    0x2
+#define SEM_PERFMON_CNTL__PERF_ENABLE1__SHIFT                                                                 0xa
+#define SEM_PERFMON_CNTL__PERF_CLEAR1__SHIFT                                                                  0xb
+#define SEM_PERFMON_CNTL__PERF_SEL1__SHIFT                                                                    0xc
+#define SEM_PERFMON_CNTL__PERF_ENABLE0_MASK                                                                   0x00000001L
+#define SEM_PERFMON_CNTL__PERF_CLEAR0_MASK                                                                    0x00000002L
+#define SEM_PERFMON_CNTL__PERF_SEL0_MASK                                                                      0x000003FCL
+#define SEM_PERFMON_CNTL__PERF_ENABLE1_MASK                                                                   0x00000400L
+#define SEM_PERFMON_CNTL__PERF_CLEAR1_MASK                                                                    0x00000800L
+#define SEM_PERFMON_CNTL__PERF_SEL1_MASK                                                                      0x000FF000L
+//SEM_PERFCOUNTER0_RESULT
+#define SEM_PERFCOUNTER0_RESULT__PERF_COUNT__SHIFT                                                            0x0
+#define SEM_PERFCOUNTER0_RESULT__PERF_COUNT_MASK                                                              0xFFFFFFFFL
+//SEM_PERFCOUNTER1_RESULT
+#define SEM_PERFCOUNTER1_RESULT__PERF_COUNT__SHIFT                                                            0x0
+#define SEM_PERFCOUNTER1_RESULT__PERF_COUNT_MASK                                                              0xFFFFFFFFL
+//SEM_STATUS
+#define SEM_STATUS__SEM_IDLE__SHIFT                                                                           0x0
+#define SEM_STATUS__SEM_INTERNAL_IDLE__SHIFT                                                                  0x1
+#define SEM_STATUS__MC_RDREQ_FIFO_FULL__SHIFT                                                                 0x2
+#define SEM_STATUS__MC_WRREQ_FIFO_FULL__SHIFT                                                                 0x3
+#define SEM_STATUS__WRITE1_FIFO_FULL__SHIFT                                                                   0x4
+#define SEM_STATUS__CHECK0_FIFO_FULL__SHIFT                                                                   0x5
+#define SEM_STATUS__MC_RDREQ_PENDING__SHIFT                                                                   0x6
+#define SEM_STATUS__MC_WRREQ_PENDING__SHIFT                                                                   0x7
+#define SEM_STATUS__SDMA0_MAILBOX_PENDING__SHIFT                                                              0x8
+#define SEM_STATUS__SDMA1_MAILBOX_PENDING__SHIFT                                                              0x9
+#define SEM_STATUS__UVD_MAILBOX_PENDING__SHIFT                                                                0xa
+#define SEM_STATUS__VCE_MAILBOX_PENDING__SHIFT                                                                0xb
+#define SEM_STATUS__CPG1_MAILBOX_PENDING__SHIFT                                                               0xc
+#define SEM_STATUS__CPG2_MAILBOX_PENDING__SHIFT                                                               0xd
+#define SEM_STATUS__VCE1_MAILBOX_PENDING__SHIFT                                                               0xe
+#define SEM_STATUS__ATC_REQ_PENDING__SHIFT                                                                    0xf
+#define SEM_STATUS__OUTSTANDING_CLEAN__SHIFT                                                                  0x10
+#define SEM_STATUS__INVREQ_FLUSH_VF_MISMATCH__SHIFT                                                           0x11
+#define SEM_STATUS__INVREQ_NONFLUSH_VF_MISMATCH__SHIFT                                                        0x12
+#define SEM_STATUS__INVREQ_CNT_IDLE__SHIFT                                                                    0x13
+#define SEM_STATUS__ENTRYLIST_IDLE__SHIFT                                                                     0x14
+#define SEM_STATUS__MIF_IDLE__SHIFT                                                                           0x15
+#define SEM_STATUS__REGISTER_IDLE__SHIFT                                                                      0x16
+#define SEM_STATUS__ATCL2_INVREQ_IDLE__SHIFT                                                                  0x17
+#define SEM_STATUS__UVD1_MAILBOX_PENDING__SHIFT                                                               0x18
+#define SEM_STATUS__SWITCH_READY__SHIFT                                                                       0x1f
+#define SEM_STATUS__SEM_IDLE_MASK                                                                             0x00000001L
+#define SEM_STATUS__SEM_INTERNAL_IDLE_MASK                                                                    0x00000002L
+#define SEM_STATUS__MC_RDREQ_FIFO_FULL_MASK                                                                   0x00000004L
+#define SEM_STATUS__MC_WRREQ_FIFO_FULL_MASK                                                                   0x00000008L
+#define SEM_STATUS__WRITE1_FIFO_FULL_MASK                                                                     0x00000010L
+#define SEM_STATUS__CHECK0_FIFO_FULL_MASK                                                                     0x00000020L
+#define SEM_STATUS__MC_RDREQ_PENDING_MASK                                                                     0x00000040L
+#define SEM_STATUS__MC_WRREQ_PENDING_MASK                                                                     0x00000080L
+#define SEM_STATUS__SDMA0_MAILBOX_PENDING_MASK                                                                0x00000100L
+#define SEM_STATUS__SDMA1_MAILBOX_PENDING_MASK                                                                0x00000200L
+#define SEM_STATUS__UVD_MAILBOX_PENDING_MASK                                                                  0x00000400L
+#define SEM_STATUS__VCE_MAILBOX_PENDING_MASK                                                                  0x00000800L
+#define SEM_STATUS__CPG1_MAILBOX_PENDING_MASK                                                                 0x00001000L
+#define SEM_STATUS__CPG2_MAILBOX_PENDING_MASK                                                                 0x00002000L
+#define SEM_STATUS__VCE1_MAILBOX_PENDING_MASK                                                                 0x00004000L
+#define SEM_STATUS__ATC_REQ_PENDING_MASK                                                                      0x00008000L
+#define SEM_STATUS__OUTSTANDING_CLEAN_MASK                                                                    0x00010000L
+#define SEM_STATUS__INVREQ_FLUSH_VF_MISMATCH_MASK                                                             0x00020000L
+#define SEM_STATUS__INVREQ_NONFLUSH_VF_MISMATCH_MASK                                                          0x00040000L
+#define SEM_STATUS__INVREQ_CNT_IDLE_MASK                                                                      0x00080000L
+#define SEM_STATUS__ENTRYLIST_IDLE_MASK                                                                       0x00100000L
+#define SEM_STATUS__MIF_IDLE_MASK                                                                             0x00200000L
+#define SEM_STATUS__REGISTER_IDLE_MASK                                                                        0x00400000L
+#define SEM_STATUS__ATCL2_INVREQ_IDLE_MASK                                                                    0x00800000L
+#define SEM_STATUS__UVD1_MAILBOX_PENDING_MASK                                                                 0x01000000L
+#define SEM_STATUS__SWITCH_READY_MASK                                                                         0x80000000L
+//SEM_MAILBOX_CLIENTCONFIG
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT0__SHIFT                                                           0x0
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT1__SHIFT                                                           0x3
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT2__SHIFT                                                           0x6
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT3__SHIFT                                                           0x9
+#define SEM_MAILBOX_CLIENTCONFIG__SDMA_CLIENT0__SHIFT                                                         0xc
+#define SEM_MAILBOX_CLIENTCONFIG__UVD_CLIENT0__SHIFT                                                          0xf
+#define SEM_MAILBOX_CLIENTCONFIG__SDMA1_CLIENT0__SHIFT                                                        0x12
+#define SEM_MAILBOX_CLIENTCONFIG__VCE_CLIENT0__SHIFT                                                          0x15
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT0_MASK                                                             0x00000007L
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT1_MASK                                                             0x00000038L
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT2_MASK                                                             0x000001C0L
+#define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT3_MASK                                                             0x00000E00L
+#define SEM_MAILBOX_CLIENTCONFIG__SDMA_CLIENT0_MASK                                                           0x00007000L
+#define SEM_MAILBOX_CLIENTCONFIG__UVD_CLIENT0_MASK                                                            0x00038000L
+#define SEM_MAILBOX_CLIENTCONFIG__SDMA1_CLIENT0_MASK                                                          0x001C0000L
+#define SEM_MAILBOX_CLIENTCONFIG__VCE_CLIENT0_MASK                                                            0x00E00000L
+//SEM_MAILBOX
+#define SEM_MAILBOX__HOSTPORT__SHIFT                                                                          0x0
+#define SEM_MAILBOX__RESERVED__SHIFT                                                                          0x10
+#define SEM_MAILBOX__HOSTPORT_MASK                                                                            0x0000FFFFL
+#define SEM_MAILBOX__RESERVED_MASK                                                                            0xFFFF0000L
+//SEM_MAILBOX_CONTROL
+#define SEM_MAILBOX_CONTROL__HOSTPORT_ENABLE__SHIFT                                                           0x0
+#define SEM_MAILBOX_CONTROL__RESERVED__SHIFT                                                                  0x10
+#define SEM_MAILBOX_CONTROL__HOSTPORT_ENABLE_MASK                                                             0x0000FFFFL
+#define SEM_MAILBOX_CONTROL__RESERVED_MASK                                                                    0xFFFF0000L
+//SEM_CHICKEN_BITS
+#define SEM_CHICKEN_BITS__VMID_PIPELINE_EN__SHIFT                                                             0x0
+#define SEM_CHICKEN_BITS__ENTRY_PIPELINE_EN__SHIFT                                                            0x1
+#define SEM_CHICKEN_BITS__CHECK_COUNTER_EN__SHIFT                                                             0x2
+#define SEM_CHICKEN_BITS__ECC_BEHAVIOR__SHIFT                                                                 0x3
+#define SEM_CHICKEN_BITS__PHY_TRAN_EN__SHIFT                                                                  0x6
+#define SEM_CHICKEN_BITS__ADDR_CMP_UNTRAN_EN__SHIFT                                                           0x7
+#define SEM_CHICKEN_BITS__IDLE_COUNTER_INDEX__SHIFT                                                           0x8
+#define SEM_CHICKEN_BITS__OUTSTANDING_CLEAN_COUNTER_INDEX__SHIFT                                              0xa
+#define SEM_CHICKEN_BITS__ATCL2_BUS_ID__SHIFT                                                                 0xc
+#define SEM_CHICKEN_BITS__ATOMIC_EN__SHIFT                                                                    0xe
+#define SEM_CHICKEN_BITS__EXTERNAL_ATOMIC_CHECK__SHIFT                                                        0xf
+#define SEM_CHICKEN_BITS__CLEAR_MAILBOX__SHIFT                                                                0x10
+#define SEM_CHICKEN_BITS__INVACK_AFTER_OUTSTANDING_CLEAN__SHIFT                                               0x12
+#define SEM_CHICKEN_BITS__UTC_TAG_CONFLICT_CHECK__SHIFT                                                       0x13
+#define SEM_CHICKEN_BITS__VMID_PIPELINE_EN_MASK                                                               0x00000001L
+#define SEM_CHICKEN_BITS__ENTRY_PIPELINE_EN_MASK                                                              0x00000002L
+#define SEM_CHICKEN_BITS__CHECK_COUNTER_EN_MASK                                                               0x00000004L
+#define SEM_CHICKEN_BITS__ECC_BEHAVIOR_MASK                                                                   0x00000018L
+#define SEM_CHICKEN_BITS__PHY_TRAN_EN_MASK                                                                    0x00000040L
+#define SEM_CHICKEN_BITS__ADDR_CMP_UNTRAN_EN_MASK                                                             0x00000080L
+#define SEM_CHICKEN_BITS__IDLE_COUNTER_INDEX_MASK                                                             0x00000300L
+#define SEM_CHICKEN_BITS__OUTSTANDING_CLEAN_COUNTER_INDEX_MASK                                                0x00000C00L
+#define SEM_CHICKEN_BITS__ATCL2_BUS_ID_MASK                                                                   0x00003000L
+#define SEM_CHICKEN_BITS__ATOMIC_EN_MASK                                                                      0x00004000L
+#define SEM_CHICKEN_BITS__EXTERNAL_ATOMIC_CHECK_MASK                                                          0x00008000L
+#define SEM_CHICKEN_BITS__CLEAR_MAILBOX_MASK                                                                  0x00030000L
+#define SEM_CHICKEN_BITS__INVACK_AFTER_OUTSTANDING_CLEAN_MASK                                                 0x00040000L
+#define SEM_CHICKEN_BITS__UTC_TAG_CONFLICT_CHECK_MASK                                                         0x00080000L
+//SEM_MAILBOX_CLIENTCONFIG_EXTRA
+#define SEM_MAILBOX_CLIENTCONFIG_EXTRA__VCE1_CLIENT0__SHIFT                                                   0x0
+#define SEM_MAILBOX_CLIENTCONFIG_EXTRA__UVD1_CLIENT0__SHIFT                                                   0x4
+#define SEM_MAILBOX_CLIENTCONFIG_EXTRA__VCE1_CLIENT0_MASK                                                     0x0000000FL
+#define SEM_MAILBOX_CLIENTCONFIG_EXTRA__UVD1_CLIENT0_MASK                                                     0x000000F0L
+//SEM_GPU_IOV_VIOLATION_LOG
+#define SEM_GPU_IOV_VIOLATION_LOG__VIOLATION_STATUS__SHIFT                                                    0x0
+#define SEM_GPU_IOV_VIOLATION_LOG__MULTIPLE_VIOLATION_STATUS__SHIFT                                           0x1
+#define SEM_GPU_IOV_VIOLATION_LOG__ADDRESS__SHIFT                                                             0x2
+#define SEM_GPU_IOV_VIOLATION_LOG__OPCODE__SHIFT                                                              0x12
+#define SEM_GPU_IOV_VIOLATION_LOG__VF__SHIFT                                                                  0x13
+#define SEM_GPU_IOV_VIOLATION_LOG__VF_ID__SHIFT                                                               0x14
+#define SEM_GPU_IOV_VIOLATION_LOG__INITIATOR_ID__SHIFT                                                        0x18
+#define SEM_GPU_IOV_VIOLATION_LOG__VIOLATION_STATUS_MASK                                                      0x00000001L
+#define SEM_GPU_IOV_VIOLATION_LOG__MULTIPLE_VIOLATION_STATUS_MASK                                             0x00000002L
+#define SEM_GPU_IOV_VIOLATION_LOG__ADDRESS_MASK                                                               0x0003FFFCL
+#define SEM_GPU_IOV_VIOLATION_LOG__OPCODE_MASK                                                                0x00040000L
+#define SEM_GPU_IOV_VIOLATION_LOG__VF_MASK                                                                    0x00080000L
+#define SEM_GPU_IOV_VIOLATION_LOG__VF_ID_MASK                                                                 0x00F00000L
+#define SEM_GPU_IOV_VIOLATION_LOG__INITIATOR_ID_MASK                                                          0xFF000000L
+//SEM_OUTSTANDING_THRESHOLD
+#define SEM_OUTSTANDING_THRESHOLD__VALUE__SHIFT                                                               0x0
+#define SEM_OUTSTANDING_THRESHOLD__VALUE_MASK                                                                 0x000000FFL
+//SEM_MEM_POWER_CTRL
+#define SEM_MEM_POWER_CTRL__MEM_POWER_CTRL_EN__SHIFT                                                          0x0
+#define SEM_MEM_POWER_CTRL__MEM_POWER_LS_EN__SHIFT                                                            0x1
+#define SEM_MEM_POWER_CTRL__MEM_POWER_DS_EN__SHIFT                                                            0x2
+#define SEM_MEM_POWER_CTRL__MEM_POWER_SD_EN__SHIFT                                                            0x3
+#define SEM_MEM_POWER_CTRL__MEM_IDLE_HYSTERESIS__SHIFT                                                        0x4
+#define SEM_MEM_POWER_CTRL__MEM_POWER_UP_RECOVER_DELAY__SHIFT                                                 0x8
+#define SEM_MEM_POWER_CTRL__MEM_POWER_DOWN_LS_ENTER_DELAY__SHIFT                                              0xe
+#define SEM_MEM_POWER_CTRL__MEM_POWER_CTRL_EN_MASK                                                            0x00000001L
+#define SEM_MEM_POWER_CTRL__MEM_POWER_LS_EN_MASK                                                              0x00000002L
+#define SEM_MEM_POWER_CTRL__MEM_POWER_DS_EN_MASK                                                              0x00000004L
+#define SEM_MEM_POWER_CTRL__MEM_POWER_SD_EN_MASK                                                              0x00000008L
+#define SEM_MEM_POWER_CTRL__MEM_IDLE_HYSTERESIS_MASK                                                          0x00000070L
+#define SEM_MEM_POWER_CTRL__MEM_POWER_UP_RECOVER_DELAY_MASK                                                   0x00003F00L
+#define SEM_MEM_POWER_CTRL__MEM_POWER_DOWN_LS_ENTER_DELAY_MASK                                                0x0000C000L
+//SEM_REGISTER_LAST_PART2
+#define SEM_REGISTER_LAST_PART2__RESERVED__SHIFT                                                              0x0
+#define SEM_REGISTER_LAST_PART2__RESERVED_MASK                                                                0xFFFFFFFFL
+//IH_ACTIVE_FCN_ID
+#define IH_ACTIVE_FCN_ID__VF_ID__SHIFT                                                                        0x0
+#define IH_ACTIVE_FCN_ID__RESERVED__SHIFT                                                                     0x4
+#define IH_ACTIVE_FCN_ID__PF_VF__SHIFT                                                                        0x1f
+#define IH_ACTIVE_FCN_ID__VF_ID_MASK                                                                          0x0000000FL
+#define IH_ACTIVE_FCN_ID__RESERVED_MASK                                                                       0x7FFFFFF0L
+#define IH_ACTIVE_FCN_ID__PF_VF_MASK                                                                          0x80000000L
+//IH_VIRT_RESET_REQ
+#define IH_VIRT_RESET_REQ__VF__SHIFT                                                                          0x0
+#define IH_VIRT_RESET_REQ__PF__SHIFT                                                                          0x1f
+#define IH_VIRT_RESET_REQ__VF_MASK                                                                            0x0000FFFFL
+#define IH_VIRT_RESET_REQ__PF_MASK                                                                            0x80000000L
+//IH_CLIENT_CFG
+#define IH_CLIENT_CFG__TOTAL_CLIENT_NUM__SHIFT                                                                0x0
+#define IH_CLIENT_CFG__TOTAL_CLIENT_NUM_MASK                                                                  0x0000001FL
+//IH_CLIENT_CFG_INDEX
+#define IH_CLIENT_CFG_INDEX__INDEX__SHIFT                                                                     0x0
+#define IH_CLIENT_CFG_INDEX__INDEX_MASK                                                                       0x0000001FL
+//IH_CLIENT_CFG_DATA
+#define IH_CLIENT_CFG_DATA__CREDIT_RETURN_ADDR__SHIFT                                                         0x0
+#define IH_CLIENT_CFG_DATA__CLIENT_TYPE__SHIFT                                                                0x12
+#define IH_CLIENT_CFG_DATA__RING_ID__SHIFT                                                                    0x14
+#define IH_CLIENT_CFG_DATA__VF_RB_SELECT__SHIFT                                                               0x16
+#define IH_CLIENT_CFG_DATA__OVERWRITE_RING_ID_WITH_ACTIVE_FCN_ID__SHIFT                                       0x18
+#define IH_CLIENT_CFG_DATA__CREDIT_RETURN_ADDR_MASK                                                           0x0003FFFFL
+#define IH_CLIENT_CFG_DATA__CLIENT_TYPE_MASK                                                                  0x000C0000L
+#define IH_CLIENT_CFG_DATA__RING_ID_MASK                                                                      0x00300000L
+#define IH_CLIENT_CFG_DATA__VF_RB_SELECT_MASK                                                                 0x00C00000L
+#define IH_CLIENT_CFG_DATA__OVERWRITE_RING_ID_WITH_ACTIVE_FCN_ID_MASK                                         0x01000000L
+//IH_CID_REMAP_INDEX
+#define IH_CID_REMAP_INDEX__INDEX__SHIFT                                                                      0x0
+#define IH_CID_REMAP_INDEX__INDEX_MASK                                                                        0x00000003L
+//IH_CID_REMAP_DATA
+#define IH_CID_REMAP_DATA__CLIENT_ID__SHIFT                                                                   0x0
+#define IH_CID_REMAP_DATA__INITIATOR_ID__SHIFT                                                                0x8
+#define IH_CID_REMAP_DATA__CLIENT_ID_REMAP__SHIFT                                                             0x10
+#define IH_CID_REMAP_DATA__CLIENT_ID_MASK                                                                     0x000000FFL
+#define IH_CID_REMAP_DATA__INITIATOR_ID_MASK                                                                  0x0000FF00L
+#define IH_CID_REMAP_DATA__CLIENT_ID_REMAP_MASK                                                               0x00FF0000L
+//IH_CHICKEN
+#define IH_CHICKEN__ACTIVE_FCN_ID_PROT_ENABLE__SHIFT                                                          0x0
+#define IH_CHICKEN__MC_SPACE_FBPA_ENABLE__SHIFT                                                               0x3
+#define IH_CHICKEN__MC_SPACE_GPA_ENABLE__SHIFT                                                                0x4
+#define IH_CHICKEN__ACTIVE_FCN_ID_PROT_ENABLE_MASK                                                            0x00000001L
+#define IH_CHICKEN__MC_SPACE_FBPA_ENABLE_MASK                                                                 0x00000008L
+#define IH_CHICKEN__MC_SPACE_GPA_ENABLE_MASK                                                                  0x00000010L
+//IH_MMHUB_CNTL
+#define IH_MMHUB_CNTL__UNITID__SHIFT                                                                          0x0
+#define IH_MMHUB_CNTL__IV_TLVL__SHIFT                                                                         0x8
+#define IH_MMHUB_CNTL__WPTR_WB_TLVL__SHIFT                                                                    0xc
+#define IH_MMHUB_CNTL__UNITID_MASK                                                                            0x0000003FL
+#define IH_MMHUB_CNTL__IV_TLVL_MASK                                                                           0x00000700L
+#define IH_MMHUB_CNTL__WPTR_WB_TLVL_MASK                                                                      0x00007000L
+//IH_INT_DROP_CNTL
+#define IH_INT_DROP_CNTL__INT_DROP_EN__SHIFT                                                                  0x0
+#define IH_INT_DROP_CNTL__CLIENT_ID_MATCH_EN__SHIFT                                                           0x1
+#define IH_INT_DROP_CNTL__SOURCE_ID_MATCH_EN__SHIFT                                                           0x2
+#define IH_INT_DROP_CNTL__VF_ID_MATCH_EN__SHIFT                                                               0x3
+#define IH_INT_DROP_CNTL__VF_MATCH_EN__SHIFT                                                                  0x4
+#define IH_INT_DROP_CNTL__CONTEXT_ID_MATCH_EN__SHIFT                                                          0x5
+#define IH_INT_DROP_CNTL__INT_DROP_MODE__SHIFT                                                                0x6
+#define IH_INT_DROP_CNTL__UTCL2_RETRY_INT_DROP_EN__SHIFT                                                      0x8
+#define IH_INT_DROP_CNTL__INT_DROPPED__SHIFT                                                                  0x10
+#define IH_INT_DROP_CNTL__INT_DROP_EN_MASK                                                                    0x00000001L
+#define IH_INT_DROP_CNTL__CLIENT_ID_MATCH_EN_MASK                                                             0x00000002L
+#define IH_INT_DROP_CNTL__SOURCE_ID_MATCH_EN_MASK                                                             0x00000004L
+#define IH_INT_DROP_CNTL__VF_ID_MATCH_EN_MASK                                                                 0x00000008L
+#define IH_INT_DROP_CNTL__VF_MATCH_EN_MASK                                                                    0x00000010L
+#define IH_INT_DROP_CNTL__CONTEXT_ID_MATCH_EN_MASK                                                            0x00000020L
+#define IH_INT_DROP_CNTL__INT_DROP_MODE_MASK                                                                  0x000000C0L
+#define IH_INT_DROP_CNTL__UTCL2_RETRY_INT_DROP_EN_MASK                                                        0x00000100L
+#define IH_INT_DROP_CNTL__INT_DROPPED_MASK                                                                    0x00010000L
+//IH_INT_DROP_MATCH_VALUE0
+#define IH_INT_DROP_MATCH_VALUE0__CLIENT_ID_MATCH_VALUE__SHIFT                                                0x0
+#define IH_INT_DROP_MATCH_VALUE0__SOURCE_ID_MATCH_VALUE__SHIFT                                                0x8
+#define IH_INT_DROP_MATCH_VALUE0__VF_ID_MATCH_VALUE__SHIFT                                                    0x10
+#define IH_INT_DROP_MATCH_VALUE0__VF_MATCH_VALUE__SHIFT                                                       0x17
+#define IH_INT_DROP_MATCH_VALUE0__CONTEXT_ID_39_32_MATCH_VALUE__SHIFT                                         0x18
+#define IH_INT_DROP_MATCH_VALUE0__CLIENT_ID_MATCH_VALUE_MASK                                                  0x000000FFL
+#define IH_INT_DROP_MATCH_VALUE0__SOURCE_ID_MATCH_VALUE_MASK                                                  0x0000FF00L
+#define IH_INT_DROP_MATCH_VALUE0__VF_ID_MATCH_VALUE_MASK                                                      0x000F0000L
+#define IH_INT_DROP_MATCH_VALUE0__VF_MATCH_VALUE_MASK                                                         0x00800000L
+#define IH_INT_DROP_MATCH_VALUE0__CONTEXT_ID_39_32_MATCH_VALUE_MASK                                           0xFF000000L
+//IH_INT_DROP_MATCH_VALUE1
+#define IH_INT_DROP_MATCH_VALUE1__CONTEXT_ID_31_0_MATCH_VALUE__SHIFT                                          0x0
+#define IH_INT_DROP_MATCH_VALUE1__CONTEXT_ID_31_0_MATCH_VALUE_MASK                                            0xFFFFFFFFL
+//IH_INT_DROP_MATCH_MASK0
+#define IH_INT_DROP_MATCH_MASK0__CLIENT_ID_MATCH_MASK__SHIFT                                                  0x0
+#define IH_INT_DROP_MATCH_MASK0__SOURCE_ID_MATCH_MASK__SHIFT                                                  0x8
+#define IH_INT_DROP_MATCH_MASK0__VF_ID_MATCH_MASK__SHIFT                                                      0x10
+#define IH_INT_DROP_MATCH_MASK0__VF_MATCH_MASK__SHIFT                                                         0x17
+#define IH_INT_DROP_MATCH_MASK0__CONTEXT_ID_39_32_MATCH_MASK__SHIFT                                           0x18
+#define IH_INT_DROP_MATCH_MASK0__CLIENT_ID_MATCH_MASK_MASK                                                    0x000000FFL
+#define IH_INT_DROP_MATCH_MASK0__SOURCE_ID_MATCH_MASK_MASK                                                    0x0000FF00L
+#define IH_INT_DROP_MATCH_MASK0__VF_ID_MATCH_MASK_MASK                                                        0x000F0000L
+#define IH_INT_DROP_MATCH_MASK0__VF_MATCH_MASK_MASK                                                           0x00800000L
+#define IH_INT_DROP_MATCH_MASK0__CONTEXT_ID_39_32_MATCH_MASK_MASK                                             0xFF000000L
+//IH_INT_DROP_MATCH_MASK1
+#define IH_INT_DROP_MATCH_MASK1__CONTEXT_ID_31_0_MATCH_MASK__SHIFT                                            0x0
+#define IH_INT_DROP_MATCH_MASK1__CONTEXT_ID_31_0_MATCH_MASK_MASK                                              0xFFFFFFFFL
+//IH_REGISTER_LAST_PART1
+#define IH_REGISTER_LAST_PART1__RESERVED__SHIFT                                                               0x0
+#define IH_REGISTER_LAST_PART1__RESERVED_MASK                                                                 0xFFFFFFFFL
+//SEM_ACTIVE_FCN_ID
+#define SEM_ACTIVE_FCN_ID__VFID__SHIFT                                                                        0x0
+#define SEM_ACTIVE_FCN_ID__VF__SHIFT                                                                          0x1f
+#define SEM_ACTIVE_FCN_ID__VFID_MASK                                                                          0x0000000FL
+#define SEM_ACTIVE_FCN_ID__VF_MASK                                                                            0x80000000L
+//SEM_VIRT_RESET_REQ
+#define SEM_VIRT_RESET_REQ__VF__SHIFT                                                                         0x0
+#define SEM_VIRT_RESET_REQ__PF__SHIFT                                                                         0x1f
+#define SEM_VIRT_RESET_REQ__VF_MASK                                                                           0x0000FFFFL
+#define SEM_VIRT_RESET_REQ__PF_MASK                                                                           0x80000000L
+//SEM_RESP_SDMA0
+#define SEM_RESP_SDMA0__ADDR__SHIFT                                                                           0x2
+#define SEM_RESP_SDMA0__ADDR_MASK                                                                             0x000FFFFCL
+//SEM_RESP_SDMA1
+#define SEM_RESP_SDMA1__ADDR__SHIFT                                                                           0x2
+#define SEM_RESP_SDMA1__ADDR_MASK                                                                             0x000FFFFCL
+//SEM_RESP_UVD
+#define SEM_RESP_UVD__ADDR__SHIFT                                                                             0x2
+#define SEM_RESP_UVD__ADDR_MASK                                                                               0x000FFFFCL
+//SEM_RESP_VCE_0
+#define SEM_RESP_VCE_0__ADDR__SHIFT                                                                           0x2
+#define SEM_RESP_VCE_0__ADDR_MASK                                                                             0x000FFFFCL
+//SEM_RESP_ACP
+#define SEM_RESP_ACP__ADDR__SHIFT                                                                             0x2
+#define SEM_RESP_ACP__ADDR_MASK                                                                               0x000FFFFCL
+//SEM_RESP_ISP
+#define SEM_RESP_ISP__ADDR__SHIFT                                                                             0x2
+#define SEM_RESP_ISP__ADDR_MASK                                                                               0x000FFFFCL
+//SEM_RESP_VCE_1
+#define SEM_RESP_VCE_1__ADDR__SHIFT                                                                           0x2
+#define SEM_RESP_VCE_1__ADDR_MASK                                                                             0x000FFFFCL
+//SEM_RESP_VP8
+#define SEM_RESP_VP8__ADDR__SHIFT                                                                             0x2
+#define SEM_RESP_VP8__ADDR_MASK                                                                               0x000FFFFCL
+//SEM_RESP_GC
+#define SEM_RESP_GC__ADDR__SHIFT                                                                              0x2
+#define SEM_RESP_GC__ADDR_MASK                                                                                0x000FFFFCL
+//SEM_RESP_UVD_1
+#define SEM_RESP_UVD_1__ADDR__SHIFT                                                                           0x2
+#define SEM_RESP_UVD_1__ADDR_MASK                                                                             0x000FFFFCL
+//SEM_CID_REMAP_INDEX
+#define SEM_CID_REMAP_INDEX__INDEX__SHIFT                                                                     0x0
+#define SEM_CID_REMAP_INDEX__INDEX_MASK                                                                       0x00000003L
+//SEM_CID_REMAP_DATA
+#define SEM_CID_REMAP_DATA__CLIENT_ID__SHIFT                                                                  0x0
+#define SEM_CID_REMAP_DATA__INITIATOR_ID__SHIFT                                                               0x8
+#define SEM_CID_REMAP_DATA__CLIENT_ID_REMAP__SHIFT                                                            0x10
+#define SEM_CID_REMAP_DATA__CLIENT_ID_MASK                                                                    0x000000FFL
+#define SEM_CID_REMAP_DATA__INITIATOR_ID_MASK                                                                 0x0000FF00L
+#define SEM_CID_REMAP_DATA__CLIENT_ID_REMAP_MASK                                                              0x00FF0000L
+//SEM_ATOMIC_OP_LUT
+#define SEM_ATOMIC_OP_LUT__SIGNAL_NORMAL__SHIFT                                                               0x0
+#define SEM_ATOMIC_OP_LUT__SIGNAL_WRITE1__SHIFT                                                               0x7
+#define SEM_ATOMIC_OP_LUT__WAIT_NORMAL__SHIFT                                                                 0xe
+#define SEM_ATOMIC_OP_LUT__WAIT_CHECK0__SHIFT                                                                 0x15
+#define SEM_ATOMIC_OP_LUT__SIGNAL_NORMAL_MASK                                                                 0x0000007FL
+#define SEM_ATOMIC_OP_LUT__SIGNAL_WRITE1_MASK                                                                 0x00003F80L
+#define SEM_ATOMIC_OP_LUT__WAIT_NORMAL_MASK                                                                   0x001FC000L
+#define SEM_ATOMIC_OP_LUT__WAIT_CHECK0_MASK                                                                   0x0FE00000L
+//SEM_EDC_CONFIG
+#define SEM_EDC_CONFIG__WRITE_DIS__SHIFT                                                                      0x0
+#define SEM_EDC_CONFIG__DIS_EDC__SHIFT                                                                        0x1
+#define SEM_EDC_CONFIG__WRITE_DIS_MASK                                                                        0x00000001L
+#define SEM_EDC_CONFIG__DIS_EDC_MASK                                                                          0x00000002L
+//SEM_CHICKEN_BITS2
+#define SEM_CHICKEN_BITS2__ACTIVE_FCN_ID_PROT_ENABLE__SHIFT                                                   0x0
+#define SEM_CHICKEN_BITS2__MM_CLIENT_USE_CONFIG_VFID__SHIFT                                                   0x1
+#define SEM_CHICKEN_BITS2__ACTIVE_FCN_ID_PROT_ENABLE_MASK                                                     0x00000001L
+#define SEM_CHICKEN_BITS2__MM_CLIENT_USE_CONFIG_VFID_MASK                                                     0x00000002L
+//SEM_MMHUB_CNTL
+#define SEM_MMHUB_CNTL__UNIT_ID__SHIFT                                                                        0x0
+#define SEM_MMHUB_CNTL__TLVL_VALUE__SHIFT                                                                     0x8
+#define SEM_MMHUB_CNTL__UNIT_ID_MASK                                                                          0x0000003FL
+#define SEM_MMHUB_CNTL__TLVL_VALUE_MASK                                                                       0x00000700L
+//SEM_REGISTER_LAST_PART1
+#define SEM_REGISTER_LAST_PART1__RESERVED__SHIFT                                                              0x0
+#define SEM_REGISTER_LAST_PART1__RESERVED_MASK                                                                0xFFFFFFFFL
+
+#endif
diff --git a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_11_0_6_offset.h b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_11_0_6_offset.h
new file mode 100644
index 000000000000..55facadea54b
--- /dev/null
+++ b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_11_0_6_offset.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright (C) 2020  Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#ifndef _smuio_11_0_6_OFFSET_HEADER
+#define _smuio_11_0_6_OFFSET_HEADER
+
+
+
+// addressBlock: smuio_smuio_SmuSmuioDec
+// base address: 0x5a000
+#define mmCGTT_ROM_CLK_CTRL0                                                                           0x00e4
+#define mmCGTT_ROM_CLK_CTRL0_BASE_IDX                                                                  0
+#define mmROM_INDEX                                                                                    0x00e5
+#define mmROM_INDEX_BASE_IDX                                                                           0
+#define mmROM_DATA                                                                                     0x00e6
+#define mmROM_DATA_BASE_IDX                                                                            0
+
+#endif
diff --git a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_11_0_6_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_11_0_6_sh_mask.h
new file mode 100644
index 000000000000..7d6a2fac2839
--- /dev/null
+++ b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_11_0_6_sh_mask.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (C) 2020  Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#ifndef _smuio_11_0_6_SH_MASK_HEADER
+#define _smuio_11_0_6_SH_MASK_HEADER
+
+
+//CGTT_ROM_CLK_CTRL0
+#define CGTT_ROM_CLK_CTRL0__ON_DELAY__SHIFT                                                                   0x0
+#define CGTT_ROM_CLK_CTRL0__OFF_HYSTERESIS__SHIFT                                                             0x4
+#define CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE1__SHIFT                                                             0x1e
+#define CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE0__SHIFT                                                             0x1f
+#define CGTT_ROM_CLK_CTRL0__ON_DELAY_MASK                                                                     0x0000000FL
+#define CGTT_ROM_CLK_CTRL0__OFF_HYSTERESIS_MASK                                                               0x00000FF0L
+#define CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE1_MASK                                                               0x40000000L
+#define CGTT_ROM_CLK_CTRL0__SOFT_OVERRIDE0_MASK                                                               0x80000000L
+//ROM_INDEX
+#define ROM_INDEX__ROM_INDEX__SHIFT                                                                           0x0
+#define ROM_INDEX__ROM_INDEX_MASK                                                                             0x01FFFFFFL
+//ROM_DATA
+#define ROM_DATA__ROM_DATA__SHIFT                                                                             0x0
+#define ROM_DATA__ROM_DATA_MASK                                                                               0xFFFFFFFFL
+
+#endif
diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h
index f775aac6c1bd..a41875ac5dfb 100644
--- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h
+++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h
@@ -103,6 +103,7 @@ enum pp_clock_type {
 
 enum amd_pp_sensors {
 	AMDGPU_PP_SENSOR_GFX_SCLK = 0,
+	AMDGPU_PP_SENSOR_CPU_CLK,
 	AMDGPU_PP_SENSOR_VDDNB,
 	AMDGPU_PP_SENSOR_VDDGFX,
 	AMDGPU_PP_SENSOR_UVD_VCLK,
@@ -155,9 +156,11 @@ enum {
 enum PP_OD_DPM_TABLE_COMMAND {
 	PP_OD_EDIT_SCLK_VDDC_TABLE,
 	PP_OD_EDIT_MCLK_VDDC_TABLE,
+	PP_OD_EDIT_CCLK_VDDC_TABLE,
 	PP_OD_EDIT_VDDC_CURVE,
 	PP_OD_RESTORE_DEFAULT_TABLE,
-	PP_OD_COMMIT_DPM_TABLE
+	PP_OD_COMMIT_DPM_TABLE,
+	PP_OD_EDIT_VDDGFX_OFFSET
 };
 
 struct pp_states_info {
diff --git a/drivers/gpu/drm/amd/include/renoir_ip_offset.h b/drivers/gpu/drm/amd/include/renoir_ip_offset.h
index 07633e22e99a..7dff85c81e5a 100644
--- a/drivers/gpu/drm/amd/include/renoir_ip_offset.h
+++ b/drivers/gpu/drm/amd/include/renoir_ip_offset.h
@@ -33,7 +33,7 @@ struct IP_BASE_INSTANCE
 struct IP_BASE
 {
     struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
-};
+} __maybe_unused;
 
 
 static const struct IP_BASE ACP_BASE ={ { { { 0x02403800, 0x00480000, 0, 0, 0 } },
diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
index 7b6ef05a1d35..5fa65f191a37 100644
--- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
@@ -36,6 +36,7 @@
 #include <linux/hwmon-sysfs.h>
 #include <linux/nospec.h>
 #include <linux/pm_runtime.h>
+#include <asm/processor.h>
 #include "hwmgr.h"
 
 static const struct cg_flag_name clocks[] = {
@@ -730,11 +731,18 @@ static ssize_t amdgpu_set_pp_table(struct device *dev,
  *
  * - minimum and maximum engine clock labeled OD_SCLK
  *
- * - maximum memory clock labeled OD_MCLK
+ * - minimum(not available for Vega20 and Navi1x) and maximum memory
+ *   clock labeled OD_MCLK
  *
  * - three <frequency, voltage> points labeled OD_VDDC_CURVE.
  *   They can be used to calibrate the sclk voltage curve.
  *
+ * - voltage offset(in mV) applied on target voltage calculation.
+ *   This is available for Sienna Cichlid, Navy Flounder and Dimgrey
+ *   Cavefish. For these ASICs, the target voltage calculation can be
+ *   illustrated by "voltage = voltage calculated from v/f curve +
+ *   overdrive vddgfx offset"
+ *
  * - a list of valid ranges for sclk, mclk, and voltage curve points
  *   labeled OD_RANGE
  *
@@ -755,6 +763,11 @@ static ssize_t amdgpu_set_pp_table(struct device *dev,
  *   600mV. "vc 2 1000 1000" will update point3 with clock set
  *   as 1000Mhz and voltage 1000mV.
  *
+ *   To update the voltage offset applied for gfxclk/voltage calculation,
+ *   enter the new value by writing a string that contains "vo offset".
+ *   This is supported by Sienna Cichlid, Navy Flounder and Dimgrey Cavefish.
+ *   And the offset can be a positive or negative value.
+ *
  * - When you have edited all of the states as needed, write "c" (commit)
  *   to the file to commit your changes
  *
@@ -787,6 +800,8 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
 
 	if (*buf == 's')
 		type = PP_OD_EDIT_SCLK_VDDC_TABLE;
+	else if (*buf == 'p')
+		type = PP_OD_EDIT_CCLK_VDDC_TABLE;
 	else if (*buf == 'm')
 		type = PP_OD_EDIT_MCLK_VDDC_TABLE;
 	else if(*buf == 'r')
@@ -795,6 +810,8 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
 		type = PP_OD_COMMIT_DPM_TABLE;
 	else if (!strncmp(buf, "vc", 2))
 		type = PP_OD_EDIT_VDDC_CURVE;
+	else if (!strncmp(buf, "vo", 2))
+		type = PP_OD_EDIT_VDDGFX_OFFSET;
 	else
 		return -EINVAL;
 
@@ -802,12 +819,14 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
 
 	tmp_str = buf_cpy;
 
-	if (type == PP_OD_EDIT_VDDC_CURVE)
+	if ((type == PP_OD_EDIT_VDDC_CURVE) ||
+	     (type == PP_OD_EDIT_VDDGFX_OFFSET))
 		tmp_str++;
 	while (isspace(*++tmp_str));
 
-	while (tmp_str[0]) {
-		sub_str = strsep(&tmp_str, delimiter);
+	while ((sub_str = strsep(&tmp_str, delimiter)) != NULL) {
+		if (strlen(sub_str) == 0)
+			continue;
 		ret = kstrtol(sub_str, 0, &parameter[parameter_size]);
 		if (ret)
 			return -EINVAL;
@@ -898,7 +917,9 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev,
 		size = smu_print_clk_levels(&adev->smu, SMU_OD_SCLK, buf);
 		size += smu_print_clk_levels(&adev->smu, SMU_OD_MCLK, buf+size);
 		size += smu_print_clk_levels(&adev->smu, SMU_OD_VDDC_CURVE, buf+size);
+		size += smu_print_clk_levels(&adev->smu, SMU_OD_VDDGFX_OFFSET, buf+size);
 		size += smu_print_clk_levels(&adev->smu, SMU_OD_RANGE, buf+size);
+		size += smu_print_clk_levels(&adev->smu, SMU_OD_CCLK, buf+size);
 	} else if (adev->powerplay.pp_funcs->print_clock_levels) {
 		size = amdgpu_dpm_print_clock_levels(adev, OD_SCLK, buf);
 		size += amdgpu_dpm_print_clock_levels(adev, OD_MCLK, buf+size);
@@ -1074,7 +1095,7 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev,
 static ssize_t amdgpu_read_mask(const char *buf, size_t count, uint32_t *mask)
 {
 	int ret;
-	long level;
+	unsigned long level;
 	char *sub_str = NULL;
 	char *tmp;
 	char buf_cpy[AMDGPU_MASK_BUF_MAX + 1];
@@ -1087,11 +1108,10 @@ static ssize_t amdgpu_read_mask(const char *buf, size_t count, uint32_t *mask)
 	memcpy(buf_cpy, buf, bytes);
 	buf_cpy[bytes] = '\0';
 	tmp = buf_cpy;
-	while (tmp[0]) {
-		sub_str = strsep(&tmp, delimiter);
+	while ((sub_str = strsep(&tmp, delimiter)) != NULL) {
 		if (strlen(sub_str)) {
-			ret = kstrtol(sub_str, 0, &level);
-			if (ret)
+			ret = kstrtoul(sub_str, 0, &level);
+			if (ret || level > 31)
 				return -EINVAL;
 			*mask |= 1 << level;
 		} else
@@ -1346,6 +1366,138 @@ static ssize_t amdgpu_set_pp_dpm_fclk(struct device *dev,
 	return count;
 }
 
+static ssize_t amdgpu_get_pp_dpm_vclk(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	ssize_t size;
+	int ret;
+
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+
+	ret = pm_runtime_get_sync(ddev->dev);
+	if (ret < 0) {
+		pm_runtime_put_autosuspend(ddev->dev);
+		return ret;
+	}
+
+	if (is_support_sw_smu(adev))
+		size = smu_print_clk_levels(&adev->smu, SMU_VCLK, buf);
+	else
+		size = snprintf(buf, PAGE_SIZE, "\n");
+
+	pm_runtime_mark_last_busy(ddev->dev);
+	pm_runtime_put_autosuspend(ddev->dev);
+
+	return size;
+}
+
+static ssize_t amdgpu_set_pp_dpm_vclk(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf,
+		size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	int ret;
+	uint32_t mask = 0;
+
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+
+	ret = amdgpu_read_mask(buf, count, &mask);
+	if (ret)
+		return ret;
+
+	ret = pm_runtime_get_sync(ddev->dev);
+	if (ret < 0) {
+		pm_runtime_put_autosuspend(ddev->dev);
+		return ret;
+	}
+
+	if (is_support_sw_smu(adev))
+		ret = smu_force_clk_levels(&adev->smu, SMU_VCLK, mask);
+	else
+		ret = 0;
+
+	pm_runtime_mark_last_busy(ddev->dev);
+	pm_runtime_put_autosuspend(ddev->dev);
+
+	if (ret)
+		return -EINVAL;
+
+	return count;
+}
+
+static ssize_t amdgpu_get_pp_dpm_dclk(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	ssize_t size;
+	int ret;
+
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+
+	ret = pm_runtime_get_sync(ddev->dev);
+	if (ret < 0) {
+		pm_runtime_put_autosuspend(ddev->dev);
+		return ret;
+	}
+
+	if (is_support_sw_smu(adev))
+		size = smu_print_clk_levels(&adev->smu, SMU_DCLK, buf);
+	else
+		size = snprintf(buf, PAGE_SIZE, "\n");
+
+	pm_runtime_mark_last_busy(ddev->dev);
+	pm_runtime_put_autosuspend(ddev->dev);
+
+	return size;
+}
+
+static ssize_t amdgpu_set_pp_dpm_dclk(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf,
+		size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	int ret;
+	uint32_t mask = 0;
+
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+
+	ret = amdgpu_read_mask(buf, count, &mask);
+	if (ret)
+		return ret;
+
+	ret = pm_runtime_get_sync(ddev->dev);
+	if (ret < 0) {
+		pm_runtime_put_autosuspend(ddev->dev);
+		return ret;
+	}
+
+	if (is_support_sw_smu(adev))
+		ret = smu_force_clk_levels(&adev->smu, SMU_DCLK, mask);
+	else
+		ret = 0;
+
+	pm_runtime_mark_last_busy(ddev->dev);
+	pm_runtime_put_autosuspend(ddev->dev);
+
+	if (ret)
+		return -EINVAL;
+
+	return count;
+}
+
 static ssize_t amdgpu_get_pp_dpm_dcefclk(struct device *dev,
 		struct device_attribute *attr,
 		char *buf)
@@ -1717,8 +1869,9 @@ static ssize_t amdgpu_set_pp_power_profile_mode(struct device *dev,
 			i++;
 		memcpy(buf_cpy, buf, count-i);
 		tmp_str = buf_cpy;
-		while (tmp_str[0]) {
-			sub_str = strsep(&tmp_str, delimiter);
+		while ((sub_str = strsep(&tmp_str, delimiter)) != NULL) {
+			if (strlen(sub_str) == 0)
+				continue;
 			ret = kstrtol(sub_str, 0, &parameter[parameter_size]);
 			if (ret)
 				return -EINVAL;
@@ -2025,6 +2178,8 @@ static struct amdgpu_device_attr amdgpu_device_attrs[] = {
 	AMDGPU_DEVICE_ATTR_RW(pp_dpm_mclk,				ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF),
 	AMDGPU_DEVICE_ATTR_RW(pp_dpm_socclk,				ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF),
 	AMDGPU_DEVICE_ATTR_RW(pp_dpm_fclk,				ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF),
+	AMDGPU_DEVICE_ATTR_RW(pp_dpm_vclk,				ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF),
+	AMDGPU_DEVICE_ATTR_RW(pp_dpm_dclk,				ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF),
 	AMDGPU_DEVICE_ATTR_RW(pp_dpm_dcefclk,				ATTR_FLAG_BASIC),
 	AMDGPU_DEVICE_ATTR_RW(pp_dpm_pcie,				ATTR_FLAG_BASIC),
 	AMDGPU_DEVICE_ATTR_RW(pp_sclk_od,				ATTR_FLAG_BASIC),
@@ -2067,7 +2222,8 @@ static int default_attr_update(struct amdgpu_device *adev, struct amdgpu_device_
 	} else if (DEVICE_ATTR_IS(pp_od_clk_voltage)) {
 		*states = ATTR_STATE_UNSUPPORTED;
 		if ((is_support_sw_smu(adev) && adev->smu.od_enabled) ||
-		    (!is_support_sw_smu(adev) && hwmgr->od_enabled))
+		    (is_support_sw_smu(adev) && adev->smu.is_apu) ||
+			(!is_support_sw_smu(adev) && hwmgr->od_enabled))
 			*states = ATTR_STATE_SUPPORTED;
 	} else if (DEVICE_ATTR_IS(mem_busy_percent)) {
 		if (adev->flags & AMD_IS_APU || asic_type == CHIP_VEGA10)
@@ -2087,6 +2243,12 @@ static int default_attr_update(struct amdgpu_device *adev, struct amdgpu_device_
 	} else if (DEVICE_ATTR_IS(gpu_metrics)) {
 		if (asic_type < CHIP_VEGA12)
 			*states = ATTR_STATE_UNSUPPORTED;
+	} else if (DEVICE_ATTR_IS(pp_dpm_vclk)) {
+		if (!(asic_type == CHIP_VANGOGH))
+			*states = ATTR_STATE_UNSUPPORTED;
+	} else if (DEVICE_ATTR_IS(pp_dpm_dclk)) {
+		if (!(asic_type == CHIP_VANGOGH))
+			*states = ATTR_STATE_UNSUPPORTED;
 	}
 
 	if (asic_type == CHIP_ARCTURUS) {
@@ -2897,7 +3059,8 @@ static ssize_t amdgpu_hwmon_show_power_cap_max(struct device *dev,
 					 char *buf)
 {
 	struct amdgpu_device *adev = dev_get_drvdata(dev);
-	uint32_t limit = 0;
+	int limit_type = to_sensor_dev_attr(attr)->index;
+	uint32_t limit = limit_type << 24;
 	ssize_t size;
 	int r;
 
@@ -2911,7 +3074,7 @@ static ssize_t amdgpu_hwmon_show_power_cap_max(struct device *dev,
 	}
 
 	if (is_support_sw_smu(adev)) {
-		smu_get_power_limit(&adev->smu, &limit, true);
+		smu_get_power_limit(&adev->smu, &limit, SMU_PPT_LIMIT_MAX);
 		size = snprintf(buf, PAGE_SIZE, "%u\n", limit * 1000000);
 	} else if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->get_power_limit) {
 		adev->powerplay.pp_funcs->get_power_limit(adev->powerplay.pp_handle, &limit, true);
@@ -2931,7 +3094,8 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
 					 char *buf)
 {
 	struct amdgpu_device *adev = dev_get_drvdata(dev);
-	uint32_t limit = 0;
+	int limit_type = to_sensor_dev_attr(attr)->index;
+	uint32_t limit = limit_type << 24;
 	ssize_t size;
 	int r;
 
@@ -2945,7 +3109,7 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
 	}
 
 	if (is_support_sw_smu(adev)) {
-		smu_get_power_limit(&adev->smu, &limit, false);
+		smu_get_power_limit(&adev->smu, &limit, SMU_PPT_LIMIT_CURRENT);
 		size = snprintf(buf, PAGE_SIZE, "%u\n", limit * 1000000);
 	} else if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->get_power_limit) {
 		adev->powerplay.pp_funcs->get_power_limit(adev->powerplay.pp_handle, &limit, false);
@@ -2960,6 +3124,15 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
 	return size;
 }
 
+static ssize_t amdgpu_hwmon_show_power_label(struct device *dev,
+					 struct device_attribute *attr,
+					 char *buf)
+{
+	int limit_type = to_sensor_dev_attr(attr)->index;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n",
+		limit_type == SMU_FAST_PPT_LIMIT ? "fastPPT" : "slowPPT");
+}
 
 static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
 		struct device_attribute *attr,
@@ -2967,6 +3140,7 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
 		size_t count)
 {
 	struct amdgpu_device *adev = dev_get_drvdata(dev);
+	int limit_type = to_sensor_dev_attr(attr)->index;
 	int err;
 	u32 value;
 
@@ -2981,7 +3155,7 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
 		return err;
 
 	value = value / 1000000; /* convert to Watt */
-
+	value |= limit_type << 24;
 
 	err = pm_runtime_get_sync(adev_to_drm(adev)->dev);
 	if (err < 0) {
@@ -3193,6 +3367,12 @@ static SENSOR_DEVICE_ATTR(power1_average, S_IRUGO, amdgpu_hwmon_show_power_avg,
 static SENSOR_DEVICE_ATTR(power1_cap_max, S_IRUGO, amdgpu_hwmon_show_power_cap_max, NULL, 0);
 static SENSOR_DEVICE_ATTR(power1_cap_min, S_IRUGO, amdgpu_hwmon_show_power_cap_min, NULL, 0);
 static SENSOR_DEVICE_ATTR(power1_cap, S_IRUGO | S_IWUSR, amdgpu_hwmon_show_power_cap, amdgpu_hwmon_set_power_cap, 0);
+static SENSOR_DEVICE_ATTR(power1_label, S_IRUGO, amdgpu_hwmon_show_power_label, NULL, 0);
+static SENSOR_DEVICE_ATTR(power2_average, S_IRUGO, amdgpu_hwmon_show_power_avg, NULL, 1);
+static SENSOR_DEVICE_ATTR(power2_cap_max, S_IRUGO, amdgpu_hwmon_show_power_cap_max, NULL, 1);
+static SENSOR_DEVICE_ATTR(power2_cap_min, S_IRUGO, amdgpu_hwmon_show_power_cap_min, NULL, 1);
+static SENSOR_DEVICE_ATTR(power2_cap, S_IRUGO | S_IWUSR, amdgpu_hwmon_show_power_cap, amdgpu_hwmon_set_power_cap, 1);
+static SENSOR_DEVICE_ATTR(power2_label, S_IRUGO, amdgpu_hwmon_show_power_label, NULL, 1);
 static SENSOR_DEVICE_ATTR(freq1_input, S_IRUGO, amdgpu_hwmon_show_sclk, NULL, 0);
 static SENSOR_DEVICE_ATTR(freq1_label, S_IRUGO, amdgpu_hwmon_show_sclk_label, NULL, 0);
 static SENSOR_DEVICE_ATTR(freq2_input, S_IRUGO, amdgpu_hwmon_show_mclk, NULL, 0);
@@ -3231,6 +3411,12 @@ static struct attribute *hwmon_attributes[] = {
 	&sensor_dev_attr_power1_cap_max.dev_attr.attr,
 	&sensor_dev_attr_power1_cap_min.dev_attr.attr,
 	&sensor_dev_attr_power1_cap.dev_attr.attr,
+	&sensor_dev_attr_power1_label.dev_attr.attr,
+	&sensor_dev_attr_power2_average.dev_attr.attr,
+	&sensor_dev_attr_power2_cap_max.dev_attr.attr,
+	&sensor_dev_attr_power2_cap_min.dev_attr.attr,
+	&sensor_dev_attr_power2_cap.dev_attr.attr,
+	&sensor_dev_attr_power2_label.dev_attr.attr,
 	&sensor_dev_attr_freq1_input.dev_attr.attr,
 	&sensor_dev_attr_freq1_label.dev_attr.attr,
 	&sensor_dev_attr_freq2_input.dev_attr.attr,
@@ -3323,8 +3509,9 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj,
 			effective_mode &= ~S_IWUSR;
 	}
 
-	if (((adev->flags & AMD_IS_APU) ||
-	     adev->family == AMDGPU_FAMILY_SI) &&	/* not implemented yet */
+	if (((adev->family == AMDGPU_FAMILY_SI) ||
+		 ((adev->flags & AMD_IS_APU) &&
+	      (adev->asic_type != CHIP_VANGOGH))) &&	/* not implemented yet */
 	    (attr == &sensor_dev_attr_power1_cap_max.dev_attr.attr ||
 	     attr == &sensor_dev_attr_power1_cap_min.dev_attr.attr||
 	     attr == &sensor_dev_attr_power1_cap.dev_attr.attr))
@@ -3387,6 +3574,16 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj,
 	     attr == &sensor_dev_attr_temp3_label.dev_attr.attr))
 		return 0;
 
+	/* only Vangogh has fast PPT limit and power labels */
+	if (!(adev->asic_type == CHIP_VANGOGH) &&
+	    (attr == &sensor_dev_attr_power2_average.dev_attr.attr ||
+		 attr == &sensor_dev_attr_power2_cap_max.dev_attr.attr ||
+	     attr == &sensor_dev_attr_power2_cap_min.dev_attr.attr ||
+		 attr == &sensor_dev_attr_power2_cap.dev_attr.attr ||
+		 attr == &sensor_dev_attr_power2_label.dev_attr.attr ||
+		 attr == &sensor_dev_attr_power1_label.dev_attr.attr))
+		return 0;
+
 	return effective_mode;
 }
 
@@ -3465,6 +3662,27 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev)
  */
 #if defined(CONFIG_DEBUG_FS)
 
+static void amdgpu_debugfs_prints_cpu_info(struct seq_file *m,
+					   struct amdgpu_device *adev) {
+	uint16_t *p_val;
+	uint32_t size;
+	int i;
+
+	if (is_support_cclk_dpm(adev)) {
+		p_val = kcalloc(adev->smu.cpu_core_num, sizeof(uint16_t),
+				GFP_KERNEL);
+
+		if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_CPU_CLK,
+					    (void *)p_val, &size)) {
+			for (i = 0; i < adev->smu.cpu_core_num; i++)
+				seq_printf(m, "\t%u MHz (CPU%d)\n",
+					   *(p_val + i), i);
+		}
+
+		kfree(p_val);
+	}
+}
+
 static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *adev)
 {
 	uint32_t value;
@@ -3475,6 +3693,9 @@ static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *a
 	/* GPU Clocks */
 	size = sizeof(value);
 	seq_printf(m, "GFX Clocks and Power:\n");
+
+	amdgpu_debugfs_prints_cpu_info(m, adev);
+
 	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_MCLK, (void *)&value, &size))
 		seq_printf(m, "\t%u MHz (MCLK)\n", value/100);
 	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_SCLK, (void *)&value, &size))
diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
index 4bdbcce7092d..10b0624ade65 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
@@ -33,6 +33,8 @@
 #define SMU_TEMPERATURE_UNITS_PER_CENTIGRADES	1000
 #define SMU_FW_NAME_LEN			0x24
 
+#define SMU_DPM_USER_PROFILE_RESTORE (1 << 0)
+
 struct smu_hw_power_state {
 	unsigned int magic;
 };
@@ -159,6 +161,19 @@ enum smu_power_src_type
 	SMU_POWER_SOURCE_COUNT,
 };
 
+enum smu_ppt_limit_type
+{
+	SMU_DEFAULT_PPT_LIMIT = 0,
+	SMU_FAST_PPT_LIMIT,
+};
+
+enum smu_ppt_limit_level
+{
+	SMU_PPT_LIMIT_MIN = -1,
+	SMU_PPT_LIMIT_CURRENT,
+	SMU_PPT_LIMIT_MAX,
+};
+
 enum smu_memory_pool_size
 {
     SMU_MEMORY_POOL_SIZE_ZERO   = 0,
@@ -168,6 +183,17 @@ enum smu_memory_pool_size
     SMU_MEMORY_POOL_SIZE_2_GB   = 0x80000000,
 };
 
+struct smu_user_dpm_profile {
+	uint32_t fan_mode;
+	uint32_t power_limit;
+	uint32_t fan_speed_percent;
+	uint32_t flags;
+
+	/* user clock state information */
+	uint32_t clk_mask[SMU_CLK_COUNT];
+	uint32_t clk_dependency;
+};
+
 #define SMU_TABLE_INIT(tables, table_id, s, a, d)	\
 	do {						\
 		tables[table_id].size = s;		\
@@ -459,130 +485,672 @@ struct smu_context
 	struct work_struct interrupt_work;
 
 	unsigned fan_max_rpm;
-	unsigned manual_fan_speed_rpm;
+	unsigned manual_fan_speed_percent;
 
 	uint32_t gfx_default_hard_min_freq;
 	uint32_t gfx_default_soft_max_freq;
 	uint32_t gfx_actual_hard_min_freq;
 	uint32_t gfx_actual_soft_max_freq;
+
+	/* APU only */
+	uint32_t cpu_default_soft_min_freq;
+	uint32_t cpu_default_soft_max_freq;
+	uint32_t cpu_actual_soft_min_freq;
+	uint32_t cpu_actual_soft_max_freq;
+	uint32_t cpu_core_id_select;
+	uint16_t cpu_core_num;
+
+	struct smu_user_dpm_profile user_dpm_profile;
 };
 
 struct i2c_adapter;
 
+/**
+ * struct pptable_funcs - Callbacks used to interact with the SMU.
+ */
 struct pptable_funcs {
+	/**
+	 * @run_btc: Calibrate voltage/frequency curve to fit the system's
+	 *           power delivery and voltage margins. Required for adaptive
+	 *           voltage frequency scaling (AVFS).
+	 */
 	int (*run_btc)(struct smu_context *smu);
+
+	/**
+	 * @get_allowed_feature_mask: Get allowed feature mask.
+	 * &feature_mask: Array to store feature mask.
+	 * &num: Elements in &feature_mask.
+	 */
 	int (*get_allowed_feature_mask)(struct smu_context *smu, uint32_t *feature_mask, uint32_t num);
+
+	/**
+	 * @get_current_power_state: Get the current power state.
+	 *
+	 * Return: Current power state on success, negative errno on failure.
+	 */
 	enum amd_pm_state_type (*get_current_power_state)(struct smu_context *smu);
+
+	/**
+	 * @set_default_dpm_table: Retrieve the default overdrive settings from
+	 *                         the SMU.
+	 */
 	int (*set_default_dpm_table)(struct smu_context *smu);
+
 	int (*set_power_state)(struct smu_context *smu);
+
+	/**
+	 * @populate_umd_state_clk: Populate the UMD power state table with
+	 *                          defaults.
+	 */
 	int (*populate_umd_state_clk)(struct smu_context *smu);
+
+	/**
+	 * @print_clk_levels: Print DPM clock levels for a clock domain
+	 *                    to buffer. Star current level.
+	 *
+	 * Used for sysfs interfaces.
+	 */
 	int (*print_clk_levels)(struct smu_context *smu, enum smu_clk_type clk_type, char *buf);
+
+	/**
+	 * @force_clk_levels: Set a range of allowed DPM levels for a clock
+	 *                    domain.
+	 * &clk_type: Clock domain.
+	 * &mask: Range of allowed DPM levels.
+	 */
 	int (*force_clk_levels)(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t mask);
+
+	/**
+	 * @od_edit_dpm_table: Edit the custom overdrive DPM table.
+	 * &type: Type of edit.
+	 * &input: Edit parameters.
+	 * &size: Size of &input.
+	 */
 	int (*od_edit_dpm_table)(struct smu_context *smu,
 				 enum PP_OD_DPM_TABLE_COMMAND type,
 				 long *input, uint32_t size);
+
+	/**
+	 * @get_clock_by_type_with_latency: Get the speed and latency of a clock
+	 *                                  domain.
+	 */
 	int (*get_clock_by_type_with_latency)(struct smu_context *smu,
 					      enum smu_clk_type clk_type,
 					      struct
 					      pp_clock_levels_with_latency
 					      *clocks);
+	/**
+	 * @get_clock_by_type_with_voltage: Get the speed and voltage of a clock
+	 *                                  domain.
+	 */
+	int (*get_clock_by_type_with_voltage)(struct smu_context *smu,
+					      enum amd_pp_clock_type type,
+					      struct
+					      pp_clock_levels_with_voltage
+					      *clocks);
+
+	/**
+	 * @get_power_profile_mode: Print all power profile modes to
+	 *                          buffer. Star current mode.
+	 */
 	int (*get_power_profile_mode)(struct smu_context *smu, char *buf);
+
+	/**
+	 * @set_power_profile_mode: Set a power profile mode. Also used to
+	 *                          create/set custom power profile modes.
+	 * &input: Power profile mode parameters.
+	 * &size: Size of &input.
+	 */
 	int (*set_power_profile_mode)(struct smu_context *smu, long *input, uint32_t size);
+
+	/**
+	 * @dpm_set_vcn_enable: Enable/disable VCN engine dynamic power
+	 *                      management.
+	 */
 	int (*dpm_set_vcn_enable)(struct smu_context *smu, bool enable);
+
+	/**
+	 * @dpm_set_jpeg_enable: Enable/disable JPEG engine dynamic power
+	 *                       management.
+	 */
 	int (*dpm_set_jpeg_enable)(struct smu_context *smu, bool enable);
+
+	/**
+	 * @read_sensor: Read data from a sensor.
+	 * &sensor: Sensor to read data from.
+	 * &data: Sensor reading.
+	 * &size: Size of &data.
+	 */
 	int (*read_sensor)(struct smu_context *smu, enum amd_pp_sensors sensor,
 			   void *data, uint32_t *size);
+
+	/**
+	 * @pre_display_config_changed: Prepare GPU for a display configuration
+	 *                              change.
+	 *
+	 * Disable display tracking and pin memory clock speed to maximum. Used
+	 * in display component synchronization.
+	 */
 	int (*pre_display_config_changed)(struct smu_context *smu);
+
+	/**
+	 * @display_config_changed: Notify the SMU of the current display
+	 *                          configuration.
+	 *
+	 * Allows SMU to properly track blanking periods for memory clock
+	 * adjustment. Used in display component synchronization.
+	 */
 	int (*display_config_changed)(struct smu_context *smu);
+
 	int (*apply_clocks_adjust_rules)(struct smu_context *smu);
+
+	/**
+	 * @notify_smc_display_config: Applies display requirements to the
+	 *                             current power state.
+	 *
+	 * Optimize deep sleep DCEFclk and mclk for the current display
+	 * configuration. Used in display component synchronization.
+	 */
 	int (*notify_smc_display_config)(struct smu_context *smu);
+
+	/**
+	 * @is_dpm_running: Check if DPM is running.
+	 *
+	 * Return: True if DPM is running, false otherwise.
+	 */
 	bool (*is_dpm_running)(struct smu_context *smu);
-	int (*get_fan_speed_rpm)(struct smu_context *smu, uint32_t *speed);
+
+	/**
+	 * @get_fan_speed_percent: Get the current fan speed in percent.
+	 */
+	int (*get_fan_speed_percent)(struct smu_context *smu, uint32_t *speed);
+
+	/**
+	 * @set_watermarks_table: Configure and upload the watermarks tables to
+	 *                        the SMU.
+	 */
 	int (*set_watermarks_table)(struct smu_context *smu,
 				    struct pp_smu_wm_range_sets *clock_ranges);
+
+	/**
+	 * @get_thermal_temperature_range: Get safe thermal limits in Celcius.
+	 */
 	int (*get_thermal_temperature_range)(struct smu_context *smu, struct smu_temperature_range *range);
+
+	/**
+	 * @get_uclk_dpm_states: Get memory clock DPM levels in kHz.
+	 * &clocks_in_khz: Array of DPM levels.
+	 * &num_states: Elements in &clocks_in_khz.
+	 */
 	int (*get_uclk_dpm_states)(struct smu_context *smu, uint32_t *clocks_in_khz, uint32_t *num_states);
+
+	/**
+	 * @set_default_od_settings: Set the overdrive tables to defaults.
+	 */
 	int (*set_default_od_settings)(struct smu_context *smu);
+
+	/**
+	 * @set_performance_level: Set a performance level.
+	 */
 	int (*set_performance_level)(struct smu_context *smu, enum amd_dpm_forced_level level);
+
+	/**
+	 * @display_disable_memory_clock_switch: Enable/disable dynamic memory
+	 *                                       clock switching.
+	 *
+	 * Disabling this feature forces memory clock speed to maximum.
+	 * Enabling sets the minimum memory clock capable of driving the
+	 * current display configuration.
+	 */
 	int (*display_disable_memory_clock_switch)(struct smu_context *smu, bool disable_memory_clock_switch);
+
+	/**
+	 * @dump_pptable: Print the power play table to the system log.
+	 */
 	void (*dump_pptable)(struct smu_context *smu);
+
+	/**
+	 * @get_power_limit: Get the device's power limits.
+	 */
 	int (*get_power_limit)(struct smu_context *smu);
+
+	/**
+	 * @get_ppt_limit: Get the device's ppt limits.
+	 */
+	int (*get_ppt_limit)(struct smu_context *smu, uint32_t *ppt_limit,
+			enum smu_ppt_limit_type limit_type, enum smu_ppt_limit_level limit_level);
+
+	/**
+	 * @set_df_cstate: Set data fabric cstate.
+	 */
 	int (*set_df_cstate)(struct smu_context *smu, enum pp_df_cstate state);
+
+	/**
+	 * @allow_xgmi_power_down: Enable/disable external global memory
+	 *                         interconnect power down.
+	 */
 	int (*allow_xgmi_power_down)(struct smu_context *smu, bool en);
+
+	/**
+	 * @update_pcie_parameters: Update and upload the system's PCIe
+	 *                          capabilites to the SMU.
+	 * &pcie_gen_cap: Maximum allowed PCIe generation.
+	 * &pcie_width_cap: Maximum allowed PCIe width.
+	 */
 	int (*update_pcie_parameters)(struct smu_context *smu, uint32_t pcie_gen_cap, uint32_t pcie_width_cap);
+
+	/**
+	 * @i2c_init: Initialize i2c.
+	 *
+	 * The i2c bus is used internally by the SMU voltage regulators and
+	 * other devices. The i2c's EEPROM also stores bad page tables on boards
+	 * with ECC.
+	 */
 	int (*i2c_init)(struct smu_context *smu, struct i2c_adapter *control);
+
+	/**
+	 * @i2c_fini: Tear down i2c.
+	 */
 	void (*i2c_fini)(struct smu_context *smu, struct i2c_adapter *control);
+
+	/**
+	 * @get_unique_id: Get the GPU's unique id. Used for asset tracking.
+	 */
 	void (*get_unique_id)(struct smu_context *smu);
+
+	/**
+	 * @get_dpm_clock_table: Get a copy of the DPM clock table.
+	 *
+	 * Used by display component in bandwidth and watermark calculations.
+	 */
 	int (*get_dpm_clock_table)(struct smu_context *smu, struct dpm_clocks *clock_table);
+
+	/**
+	 * @init_microcode: Request the SMU's firmware from the kernel.
+	 */
 	int (*init_microcode)(struct smu_context *smu);
+
+	/**
+	 * @load_microcode: Load firmware onto the SMU.
+	 */
 	int (*load_microcode)(struct smu_context *smu);
+
+	/**
+	 * @fini_microcode: Release the SMU's firmware.
+	 */
 	void (*fini_microcode)(struct smu_context *smu);
+
+	/**
+	 * @init_smc_tables: Initialize the SMU tables.
+	 */
 	int (*init_smc_tables)(struct smu_context *smu);
+
+	/**
+	 * @fini_smc_tables: Release the SMU tables.
+	 */
 	int (*fini_smc_tables)(struct smu_context *smu);
+
+	/**
+	 * @init_power: Initialize the power gate table context.
+	 */
 	int (*init_power)(struct smu_context *smu);
+
+	/**
+	 * @fini_power: Release the power gate table context.
+	 */
 	int (*fini_power)(struct smu_context *smu);
+
+	/**
+	 * @check_fw_status: Check the SMU's firmware status.
+	 *
+	 * Return: Zero if check passes, negative errno on failure.
+	 */
 	int (*check_fw_status)(struct smu_context *smu);
+
+	/**
+	 * @setup_pptable: Initialize the power play table and populate it with
+	 *                 default values.
+	 */
 	int (*setup_pptable)(struct smu_context *smu);
+
+	/**
+	 * @get_vbios_bootup_values: Get default boot values from the VBIOS.
+	 */
 	int (*get_vbios_bootup_values)(struct smu_context *smu);
+
+	/**
+	 * @check_fw_version: Print driver and SMU interface versions to the
+	 *                    system log.
+	 *
+	 * Interface mismatch is not a critical failure.
+	 */
 	int (*check_fw_version)(struct smu_context *smu);
+
+	/**
+	 * @powergate_sdma: Power up/down system direct memory access.
+	 */
 	int (*powergate_sdma)(struct smu_context *smu, bool gate);
+
+	/**
+	 * @set_gfx_cgpg: Enable/disable graphics engine course grain power
+	 *                gating.
+	 */
 	int (*set_gfx_cgpg)(struct smu_context *smu, bool enable);
+
+	/**
+	 * @write_pptable: Write the power play table to the SMU.
+	 */
 	int (*write_pptable)(struct smu_context *smu);
+
+	/**
+	 * @set_driver_table_location: Send the location of the driver table to
+	 *                             the SMU.
+	 */
 	int (*set_driver_table_location)(struct smu_context *smu);
+
+	/**
+	 * @set_tool_table_location: Send the location of the tool table to the
+	 *                           SMU.
+	 */
 	int (*set_tool_table_location)(struct smu_context *smu);
+
+	/**
+	 * @notify_memory_pool_location: Send the location of the memory pool to
+	 *                               the SMU.
+	 */
 	int (*notify_memory_pool_location)(struct smu_context *smu);
+
+	/**
+	 * @system_features_control: Enable/disable all SMU features.
+	 */
 	int (*system_features_control)(struct smu_context *smu, bool en);
+
+	/**
+	 * @send_smc_msg_with_param: Send a message with a parameter to the SMU.
+	 * &msg: Type of message.
+	 * &param: Message parameter.
+	 * &read_arg: SMU response (optional).
+	 */
 	int (*send_smc_msg_with_param)(struct smu_context *smu,
 				       enum smu_message_type msg, uint32_t param, uint32_t *read_arg);
+
+	/**
+	 * @send_smc_msg: Send a message to the SMU.
+	 * &msg: Type of message.
+	 * &read_arg: SMU response (optional).
+	 */
 	int (*send_smc_msg)(struct smu_context *smu,
 			    enum smu_message_type msg,
 			    uint32_t *read_arg);
+
+	/**
+	 * @init_display_count: Notify the SMU of the number of display
+	 *                      components in current display configuration.
+	 */
 	int (*init_display_count)(struct smu_context *smu, uint32_t count);
+
+	/**
+	 * @set_allowed_mask: Notify the SMU of the features currently allowed
+	 *                    by the driver.
+	 */
 	int (*set_allowed_mask)(struct smu_context *smu);
+
+	/**
+	 * @get_enabled_mask: Get a mask of features that are currently enabled
+	 *                    on the SMU.
+	 * &feature_mask: Array representing enabled feature mask.
+	 * &num: Elements in &feature_mask.
+	 */
 	int (*get_enabled_mask)(struct smu_context *smu, uint32_t *feature_mask, uint32_t num);
+
+	/**
+	 * @feature_is_enabled: Test if a feature is enabled.
+	 *
+	 * Return: One if enabled, zero if disabled.
+	 */
 	int (*feature_is_enabled)(struct smu_context *smu, enum smu_feature_mask mask);
+
+	/**
+	 * @disable_all_features_with_exception: Disable all features with
+	 *                                       exception to those in &mask.
+	 */
 	int (*disable_all_features_with_exception)(struct smu_context *smu, enum smu_feature_mask mask);
+
+	/**
+	 * @notify_display_change: Enable fast memory clock switching.
+	 *
+	 * Allows for fine grained memory clock switching but has more stringent
+	 * timing requirements.
+	 */
 	int (*notify_display_change)(struct smu_context *smu);
+
+	/**
+	 * @set_power_limit: Set power limit in watts.
+	 */
 	int (*set_power_limit)(struct smu_context *smu, uint32_t n);
+
+	/**
+	 * @init_max_sustainable_clocks: Populate max sustainable clock speed
+	 *                               table with values from the SMU.
+	 */
 	int (*init_max_sustainable_clocks)(struct smu_context *smu);
+
+	/**
+	 * @enable_thermal_alert: Enable thermal alert interrupts.
+	 */
 	int (*enable_thermal_alert)(struct smu_context *smu);
+
+	/**
+	 * @disable_thermal_alert: Disable thermal alert interrupts.
+	 */
 	int (*disable_thermal_alert)(struct smu_context *smu);
+
+	/**
+	 * @set_min_dcef_deep_sleep: Set a minimum display fabric deep sleep
+	 *                           clock speed in MHz.
+	 */
 	int (*set_min_dcef_deep_sleep)(struct smu_context *smu, uint32_t clk);
+
+	/**
+	 * @display_clock_voltage_request: Set a hard minimum frequency
+	 * for a clock domain.
+	 */
 	int (*display_clock_voltage_request)(struct smu_context *smu, struct
 					     pp_display_clock_request
 					     *clock_req);
+
+	/**
+	 * @get_fan_control_mode: Get the current fan control mode.
+	 */
 	uint32_t (*get_fan_control_mode)(struct smu_context *smu);
+
+	/**
+	 * @set_fan_control_mode: Set the fan control mode.
+	 */
 	int (*set_fan_control_mode)(struct smu_context *smu, uint32_t mode);
-	int (*set_fan_speed_rpm)(struct smu_context *smu, uint32_t speed);
+
+	/**
+	 * @set_fan_speed_percent: Set a static fan speed in percent.
+	 */
+	int (*set_fan_speed_percent)(struct smu_context *smu, uint32_t speed);
+
+	/**
+	 * @set_xgmi_pstate: Set inter-chip global memory interconnect pstate.
+	 * &pstate: Pstate to set. D0 if Nonzero, D3 otherwise.
+	 */
 	int (*set_xgmi_pstate)(struct smu_context *smu, uint32_t pstate);
+
+	/**
+	 * @gfx_off_control: Enable/disable graphics engine poweroff.
+	 */
 	int (*gfx_off_control)(struct smu_context *smu, bool enable);
+
+
+	/**
+	 * @get_gfx_off_status: Get graphics engine poweroff status.
+	 *
+	 * Return:
+	 * 0 - GFXOFF(default).
+	 * 1 - Transition out of GFX State.
+	 * 2 - Not in GFXOFF.
+	 * 3 - Transition into GFXOFF.
+	 */
 	uint32_t (*get_gfx_off_status)(struct smu_context *smu);
+
+	/**
+	 * @register_irq_handler: Register interupt request handlers.
+	 */
 	int (*register_irq_handler)(struct smu_context *smu);
+
+	/**
+	 * @set_azalia_d3_pme: Wake the audio decode engine from d3 sleep.
+	 */
 	int (*set_azalia_d3_pme)(struct smu_context *smu);
+
+	/**
+	 * @get_max_sustainable_clocks_by_dc: Get a copy of the max sustainable
+	 *                                    clock speeds table.
+	 *
+	 * Provides a way for the display component (DC) to get the max
+	 * sustainable clocks from the SMU.
+	 */
 	int (*get_max_sustainable_clocks_by_dc)(struct smu_context *smu, struct pp_smu_nv_clock_table *max_clocks);
+
+	/**
+	 * @baco_is_support: Check if GPU supports BACO (Bus Active, Chip Off).
+	 */
 	bool (*baco_is_support)(struct smu_context *smu);
+
+	/**
+	 * @baco_get_state: Get the current BACO state.
+	 *
+	 * Return: Current BACO state.
+	 */
 	enum smu_baco_state (*baco_get_state)(struct smu_context *smu);
+
+	/**
+	 * @baco_set_state: Enter/exit BACO.
+	 */
 	int (*baco_set_state)(struct smu_context *smu, enum smu_baco_state state);
+
+	/**
+	 * @baco_enter: Enter BACO.
+	 */
 	int (*baco_enter)(struct smu_context *smu);
+
+	/**
+	 * @baco_exit: Exit Baco.
+	 */
 	int (*baco_exit)(struct smu_context *smu);
+
+	/**
+	 * @mode1_reset_is_support: Check if GPU supports mode1 reset.
+	 */
 	bool (*mode1_reset_is_support)(struct smu_context *smu);
+
+	/**
+	 * @mode1_reset: Perform mode1 reset.
+	 *
+	 * Complete GPU reset.
+	 */
 	int (*mode1_reset)(struct smu_context *smu);
+
+	/**
+	 * @mode2_reset: Perform mode2 reset.
+	 *
+	 * Mode2 reset generally does not reset as many IPs as mode1 reset. The
+	 * IPs reset varies by asic.
+	 */
 	int (*mode2_reset)(struct smu_context *smu);
+
+	/**
+	 * @get_dpm_ultimate_freq: Get the hard frequency range of a clock
+	 *                         domain in MHz.
+	 */
 	int (*get_dpm_ultimate_freq)(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t *min, uint32_t *max);
+
+	/**
+	 * @set_soft_freq_limited_range: Set the soft frequency range of a clock
+	 *                               domain in MHz.
+	 */
 	int (*set_soft_freq_limited_range)(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t min, uint32_t max);
+
+	/**
+	 * @set_power_source: Notify the SMU of the current power source.
+	 */
 	int (*set_power_source)(struct smu_context *smu, enum smu_power_src_type power_src);
+
+	/**
+	 * @log_thermal_throttling_event: Print a thermal throttling warning to
+	 *                                the system's log.
+	 */
 	void (*log_thermal_throttling_event)(struct smu_context *smu);
+
+	/**
+	 * @get_pp_feature_mask: Print a human readable table of enabled
+	 *                       features to buffer.
+	 */
 	size_t (*get_pp_feature_mask)(struct smu_context *smu, char *buf);
+
+	/**
+	 * @set_pp_feature_mask: Request the SMU enable/disable features to
+	 *                       match those enabled in &new_mask.
+	 */
 	int (*set_pp_feature_mask)(struct smu_context *smu, uint64_t new_mask);
+
+	/**
+	 * @get_gpu_metrics: Get a copy of the GPU metrics table from the SMU.
+	 *
+	 * Return: Size of &table
+	 */
 	ssize_t (*get_gpu_metrics)(struct smu_context *smu, void **table);
+
+	/**
+	 * @enable_mgpu_fan_boost: Enable multi-GPU fan boost.
+	 */
 	int (*enable_mgpu_fan_boost)(struct smu_context *smu);
+
+	/**
+	 * @gfx_ulv_control: Enable/disable ultra low voltage.
+	 */
 	int (*gfx_ulv_control)(struct smu_context *smu, bool enablement);
+
+	/**
+	 * @deep_sleep_control: Enable/disable deep sleep.
+	 */
 	int (*deep_sleep_control)(struct smu_context *smu, bool enablement);
+
+	/**
+	 * @get_fan_parameters: Get fan parameters.
+	 *
+	 * Get maximum fan speed from the power play table.
+	 */
 	int (*get_fan_parameters)(struct smu_context *smu);
+
+	/**
+	 * @post_init: Helper function for asic specific workarounds.
+	 */
 	int (*post_init)(struct smu_context *smu);
+
+	/**
+	 * @interrupt_work: Work task scheduled from SMU interrupt handler.
+	 */
 	void (*interrupt_work)(struct smu_context *smu);
+
+	/**
+	 * @gpo_control: Enable/disable graphics power optimization if supported.
+	 */
 	int (*gpo_control)(struct smu_context *smu, bool enablement);
+
+	/**
+	 * @gfx_state_change_set: Send the current graphics state to the SMU.
+	 */
 	int (*gfx_state_change_set)(struct smu_context *smu, uint32_t state);
+
+	/**
+	 * @set_fine_grain_gfx_freq_parameters: Set fine grain graphics clock
+	 *                                      parameters to defaults.
+	 */
 	int (*set_fine_grain_gfx_freq_parameters)(struct smu_context *smu);
 };
 
@@ -596,6 +1164,7 @@ typedef enum {
 	METRICS_CURR_DCLK1,
 	METRICS_CURR_FCLK,
 	METRICS_CURR_DCEFCLK,
+	METRICS_AVERAGE_CPUCLK,
 	METRICS_AVERAGE_GFXCLK,
 	METRICS_AVERAGE_SOCCLK,
 	METRICS_AVERAGE_FCLK,
@@ -636,6 +1205,12 @@ enum smu_cmn2asic_mapping_type {
 #define FEA_MAP(fea) \
 	[SMU_FEATURE_##fea##_BIT] = {1, FEATURE_##fea##_BIT}
 
+#define FEA_MAP_REVERSE(fea) \
+	[SMU_FEATURE_DPM_##fea##_BIT] = {1, FEATURE_##fea##_DPM_BIT}
+
+#define FEA_MAP_HALF_REVERSE(fea) \
+	[SMU_FEATURE_DPM_##fea##CLK_BIT] = {1, FEATURE_##fea##_DPM_BIT}
+
 #define TAB_MAP(tab) \
 	[SMU_TABLE_##tab] = {1, TABLE_##tab}
 
@@ -662,7 +1237,7 @@ int smu_set_fan_speed_rpm(struct smu_context *smu, uint32_t speed);
 
 int smu_get_power_limit(struct smu_context *smu,
 			uint32_t *limit,
-			bool max_setting);
+			enum smu_ppt_limit_level limit_level);
 
 int smu_set_power_limit(struct smu_context *smu, uint32_t limit);
 int smu_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf);
@@ -718,6 +1293,7 @@ extern const struct amdgpu_ip_block_version smu_v11_0_ip_block;
 extern const struct amdgpu_ip_block_version smu_v12_0_ip_block;
 
 bool is_support_sw_smu(struct amdgpu_device *adev);
+bool is_support_cclk_dpm(struct amdgpu_device *adev);
 int smu_reset(struct smu_context *smu);
 int smu_sys_get_pp_table(struct smu_context *smu, void **table);
 int smu_sys_set_pp_table(struct smu_context *smu,  void *buf, size_t size);
diff --git a/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_vangogh.h b/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_vangogh.h
index 1c19eae93ff1..6e23a3f803a7 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_vangogh.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_vangogh.h
@@ -141,7 +141,6 @@ typedef struct {
   uint32_t MaxGfxClk;
 
   uint8_t NumDfPstatesEnabled;
-  uint8_t NumDpmLevelsEnabled;
   uint8_t NumDcfclkLevelsEnabled;
   uint8_t NumDispClkLevelsEnabled;  //applies to both dispclk and dppclk
   uint8_t NumSocClkLevelsEnabled;
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_types.h b/drivers/gpu/drm/amd/pm/inc/smu_types.h
index 720d15612fe1..aa4822202587 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu_types.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu_types.h
@@ -133,8 +133,6 @@
 	__SMU_DUMMY_MAP(PowerUpSdma),                 \
 	__SMU_DUMMY_MAP(SetHardMinIspclkByFreq),      \
 	__SMU_DUMMY_MAP(SetHardMinVcn),               \
-	__SMU_DUMMY_MAP(Spare1),                      \
-	__SMU_DUMMY_MAP(Spare2),           	      \
 	__SMU_DUMMY_MAP(SetAllowFclkSwitch),          \
 	__SMU_DUMMY_MAP(SetMinVideoGfxclkFreq),       \
 	__SMU_DUMMY_MAP(ActiveProcessNotify),         \
@@ -211,6 +209,11 @@
 	__SMU_DUMMY_MAP(SetGpoFeaturePMask),             \
 	__SMU_DUMMY_MAP(DisallowGpo),                    \
 	__SMU_DUMMY_MAP(Enable2ndUSB20Port),             \
+	__SMU_DUMMY_MAP(RequestActiveWgp),               \
+       __SMU_DUMMY_MAP(SetFastPPTLimit),                \
+       __SMU_DUMMY_MAP(SetSlowPPTLimit),                \
+       __SMU_DUMMY_MAP(GetFastPPTLimit),                \
+       __SMU_DUMMY_MAP(GetSlowPPTLimit),                \
 
 #undef __SMU_DUMMY_MAP
 #define __SMU_DUMMY_MAP(type)	SMU_MSG_##type
@@ -236,10 +239,12 @@ enum smu_clk_type {
 	SMU_SCLK,
 	SMU_MCLK,
 	SMU_PCIE,
+	SMU_OD_CCLK,
 	SMU_OD_SCLK,
 	SMU_OD_MCLK,
 	SMU_OD_VDDC_CURVE,
 	SMU_OD_RANGE,
+	SMU_OD_VDDGFX_OFFSET,
 	SMU_CLK_COUNT,
 };
 
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h b/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h
index 13de692a4213..d4cddd2390a2 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h
@@ -31,7 +31,7 @@
 #define SMU11_DRIVER_IF_VERSION_NV12 0x36
 #define SMU11_DRIVER_IF_VERSION_NV14 0x36
 #define SMU11_DRIVER_IF_VERSION_Sienna_Cichlid 0x3D
-#define SMU11_DRIVER_IF_VERSION_Navy_Flounder 0xC
+#define SMU11_DRIVER_IF_VERSION_Navy_Flounder 0xE
 #define SMU11_DRIVER_IF_VERSION_VANGOGH 0x02
 #define SMU11_DRIVER_IF_VERSION_Dimgrey_Cavefish 0xF
 
@@ -129,6 +129,15 @@ struct smu_11_0_power_context {
 	enum smu_11_0_power_state power_state;
 };
 
+struct smu_11_5_power_context {
+	uint32_t	power_source;
+	uint8_t		in_power_limit_boost_mode;
+	enum smu_11_0_power_state power_state;
+
+	uint32_t	current_fast_ppt_limit;
+	uint32_t	max_fast_ppt_limit;
+};
+
 enum smu_v11_0_baco_seq {
 	BACO_SEQ_BACO = 0,
 	BACO_SEQ_MSR,
@@ -203,11 +212,8 @@ int
 smu_v11_0_set_fan_control_mode(struct smu_context *smu,
 			       uint32_t mode);
 
-int smu_v11_0_set_fan_speed_rpm(struct smu_context *smu,
-				       uint32_t speed);
-
-int smu_v11_0_get_fan_speed_rpm(struct smu_context *smu,
-				uint32_t *speed);
+int smu_v11_0_set_fan_speed_percent(struct smu_context *smu,
+				    uint32_t speed);
 
 int smu_v11_0_set_xgmi_pstate(struct smu_context *smu,
 				     uint32_t pstate);
@@ -275,10 +281,6 @@ int smu_v11_0_get_current_pcie_link_speed_level(struct smu_context *smu);
 
 int smu_v11_0_get_current_pcie_link_speed(struct smu_context *smu);
 
-void smu_v11_0_init_gpu_metrics_v1_0(struct gpu_metrics_v1_0 *gpu_metrics);
-
-void smu_v11_0_init_gpu_metrics_v2_0(struct gpu_metrics_v2_0 *gpu_metrics);
-
 int smu_v11_0_gfx_ulv_control(struct smu_context *smu,
 			      bool enablement);
 
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v11_5_ppsmc.h b/drivers/gpu/drm/amd/pm/inc/smu_v11_5_ppsmc.h
index 55d7892e4e0e..fe130a497d6c 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu_v11_5_ppsmc.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v11_5_ppsmc.h
@@ -104,7 +104,11 @@
 #define PPSMC_MSG_DramLogSetDramBufferSize             0x46
 #define PPSMC_MSG_RequestActiveWgp                     0x47
 #define PPSMC_MSG_QueryActiveWgp                       0x48
-#define PPSMC_Message_Count                            0x49
+#define PPSMC_MSG_SetFastPPTLimit                      0x49
+#define PPSMC_MSG_SetSlowPPTLimit                      0x4A
+#define PPSMC_MSG_GetFastPPTLimit                      0x4B
+#define PPSMC_MSG_GetSlowPPTLimit                      0x4C
+#define PPSMC_Message_Count                            0x4D
 
 //Argument for PPSMC_MSG_GfxDeviceDriverReset
 enum {
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v12_0.h b/drivers/gpu/drm/amd/pm/inc/smu_v12_0.h
index fa2e8cb07967..02de3b6199e5 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu_v12_0.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v12_0.h
@@ -60,7 +60,5 @@ int smu_v12_0_set_soft_freq_limited_range(struct smu_context *smu, enum smu_clk_
 
 int smu_v12_0_set_driver_table_location(struct smu_context *smu);
 
-void smu_v12_0_init_gpu_metrics_v2_0(struct gpu_metrics_v2_0 *gpu_metrics);
-
 #endif
 #endif
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/hwmgr.c
index 6a7de8b898fa..f2cef0930aa9 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/hwmgr.c
@@ -33,6 +33,7 @@
 #include "ppsmc.h"
 #include "amd_acpi.h"
 #include "pp_psm.h"
+#include "vega10_hwmgr.h"
 
 extern const struct pp_smumgr_func ci_smu_funcs;
 extern const struct pp_smumgr_func smu8_smu_funcs;
@@ -46,7 +47,6 @@ extern const struct pp_smumgr_func vega12_smu_funcs;
 extern const struct pp_smumgr_func smu10_smu_funcs;
 extern const struct pp_smumgr_func vega20_smu_funcs;
 
-extern int vega10_hwmgr_init(struct pp_hwmgr *hwmgr);
 extern int smu10_init_function_pointers(struct pp_hwmgr *hwmgr);
 
 static int polaris_set_asic_special_caps(struct pp_hwmgr *hwmgr);
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c
index 83a6504e093c..b1038d30c8dc 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c
@@ -279,7 +279,7 @@ static const ATOM_VOLTAGE_OBJECT_V3 *atomctrl_lookup_voltage_type_v3(
  *
  * @hwmgr:           input parameter: pointer to HwMgr
  * @clock_value:     input parameter: memory clock
- * @dividers:        output parameter: memory PLL dividers
+ * @mpll_param:      output parameter: memory clock parameters
  * @strobe_mode:     input parameter: 1 for strobe mode,  0 for performance mode
  */
 int atomctrl_get_memory_pll_dividers_si(
@@ -332,7 +332,7 @@ int atomctrl_get_memory_pll_dividers_si(
  *
  * @hwmgr:                 input parameter: pointer to HwMgr
  * @clock_value:           input parameter: memory clock
- * @dividers:              output parameter: memory PLL dividers
+ * @mpll_param:            output parameter: memory clock parameters
  */
 int atomctrl_get_memory_pll_dividers_vi(struct pp_hwmgr *hwmgr,
 		uint32_t clock_value, pp_atomctrl_memory_clock_param *mpll_param)
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/process_pptables_v1_0.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/process_pptables_v1_0.c
index 741e03ad5311..f2a55c1413f5 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/process_pptables_v1_0.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/process_pptables_v1_0.c
@@ -1362,6 +1362,7 @@ static int ppt_get_vce_state_table_entry_v1_0(struct pp_hwmgr *hwmgr, uint32_t i
  * @hwmgr: Pointer to the hardware manager.
  * @entry_index: The index of the entry to be extracted from the table.
  * @power_state: The address of the PowerState instance being created.
+ * @call_back_func: The function to call into to fill power state
  * Return: -1 if the entry cannot be retrieved.
  */
 int get_powerplay_table_entry_v1_0(struct pp_hwmgr *hwmgr,
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c
index e57e64bbacdc..ed05a30d1139 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c
@@ -251,7 +251,7 @@ static int smu10_set_hard_min_gfxclk_by_freq(struct pp_hwmgr *hwmgr, uint32_t cl
 		smu10_data->gfx_actual_soft_min_freq = clock;
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 					PPSMC_MSG_SetHardMinGfxClk,
-					smu10_data->gfx_actual_soft_min_freq,
+					clock,
 					NULL);
 	}
 	return 0;
@@ -558,7 +558,8 @@ static int smu10_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 
 	/* enable the pp_od_clk_voltage sysfs file */
 	hwmgr->od_enabled = 1;
-
+	/* disabled fine grain tuning function by default */
+	data->fine_grain_enabled = 0;
 	return result;
 }
 
@@ -597,6 +598,7 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 	uint32_t min_mclk = hwmgr->display_config->min_mem_set_clock/100;
 	uint32_t index_fclk = data->clock_vol_info.vdd_dep_on_fclk->count - 1;
 	uint32_t index_socclk = data->clock_vol_info.vdd_dep_on_socclk->count - 1;
+	uint32_t fine_grain_min_freq = 0, fine_grain_max_freq = 0;
 
 	if (hwmgr->smu_version < 0x1E3700) {
 		pr_info("smu firmware version too old, can not set dpm level\n");
@@ -613,6 +615,14 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 	switch (level) {
 	case AMD_DPM_FORCED_LEVEL_HIGH:
 	case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK:
+		data->fine_grain_enabled = 0;
+
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &fine_grain_min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &fine_grain_max_freq);
+
+		data->gfx_actual_soft_min_freq = fine_grain_min_freq;
+		data->gfx_actual_soft_max_freq = fine_grain_max_freq;
+
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 						PPSMC_MSG_SetHardMinGfxClk,
 						data->gfx_max_freq_limit/100,
@@ -648,6 +658,14 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 						NULL);
 		break;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK:
+		data->fine_grain_enabled = 0;
+
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &fine_grain_min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &fine_grain_max_freq);
+
+		data->gfx_actual_soft_min_freq = fine_grain_min_freq;
+		data->gfx_actual_soft_max_freq = fine_grain_max_freq;
+
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 						PPSMC_MSG_SetHardMinGfxClk,
 						min_sclk,
@@ -658,6 +676,14 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 						NULL);
 		break;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK:
+		data->fine_grain_enabled = 0;
+
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &fine_grain_min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &fine_grain_max_freq);
+
+		data->gfx_actual_soft_min_freq = fine_grain_min_freq;
+		data->gfx_actual_soft_max_freq = fine_grain_max_freq;
+
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 						PPSMC_MSG_SetHardMinFclkByFreq,
 						min_mclk,
@@ -668,6 +694,14 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 						NULL);
 		break;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD:
+		data->fine_grain_enabled = 0;
+
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &fine_grain_min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &fine_grain_max_freq);
+
+		data->gfx_actual_soft_min_freq = fine_grain_min_freq;
+		data->gfx_actual_soft_max_freq = fine_grain_max_freq;
+
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 						PPSMC_MSG_SetHardMinGfxClk,
 						SMU10_UMD_PSTATE_GFXCLK,
@@ -703,6 +737,14 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 						NULL);
 		break;
 	case AMD_DPM_FORCED_LEVEL_AUTO:
+		data->fine_grain_enabled = 0;
+
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &fine_grain_min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &fine_grain_max_freq);
+
+		data->gfx_actual_soft_min_freq = fine_grain_min_freq;
+		data->gfx_actual_soft_max_freq = fine_grain_max_freq;
+
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 						PPSMC_MSG_SetHardMinGfxClk,
 						min_sclk,
@@ -741,6 +783,14 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 						NULL);
 		break;
 	case AMD_DPM_FORCED_LEVEL_LOW:
+		data->fine_grain_enabled = 0;
+
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &fine_grain_min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &fine_grain_max_freq);
+
+		data->gfx_actual_soft_min_freq = fine_grain_min_freq;
+		data->gfx_actual_soft_max_freq = fine_grain_max_freq;
+
 		smum_send_msg_to_smc_with_parameter(hwmgr,
 						PPSMC_MSG_SetHardMinGfxClk,
 						data->gfx_min_freq_limit/100,
@@ -759,6 +809,7 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr,
 						NULL);
 		break;
 	case AMD_DPM_FORCED_LEVEL_MANUAL:
+		data->fine_grain_enabled = 1;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_EXIT:
 	default:
 		break;
@@ -948,6 +999,8 @@ static int smu10_print_clock_levels(struct pp_hwmgr *hwmgr,
 	struct smu10_voltage_dependency_table *mclk_table =
 			data->clock_vol_info.vdd_dep_on_fclk;
 	uint32_t i, now, size = 0;
+	uint32_t min_freq, max_freq = 0;
+	uint32_t ret = 0;
 
 	switch (type) {
 	case PP_SCLK:
@@ -983,18 +1036,28 @@ static int smu10_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 	case OD_SCLK:
 		if (hwmgr->od_enabled) {
-			size = sprintf(buf, "%s:\n", "OD_SCLK");
+			ret = smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &min_freq);
+			if (ret)
+				return ret;
+			ret = smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &max_freq);
+			if (ret)
+				return ret;
 
+			size = sprintf(buf, "%s:\n", "OD_SCLK");
 			size += sprintf(buf + size, "0: %10uMhz\n",
-			(data->gfx_actual_soft_min_freq > 0) ? data->gfx_actual_soft_min_freq : data->gfx_min_freq_limit/100);
-			size += sprintf(buf + size, "1: %10uMhz\n", data->gfx_max_freq_limit/100);
+			(data->gfx_actual_soft_min_freq > 0) ? data->gfx_actual_soft_min_freq : min_freq);
+			size += sprintf(buf + size, "1: %10uMhz\n",
+			(data->gfx_actual_soft_max_freq > 0) ? data->gfx_actual_soft_max_freq : max_freq);
 		}
 		break;
 	case OD_RANGE:
 		if (hwmgr->od_enabled) {
-			uint32_t min_freq, max_freq = 0;
-			smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &min_freq);
-			smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &max_freq);
+			ret = smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &min_freq);
+			if (ret)
+				return ret;
+			ret = smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &max_freq);
+			if (ret)
+				return ret;
 
 			size = sprintf(buf, "%s:\n", "OD_RANGE");
 			size += sprintf(buf + size, "SCLK: %7uMHz %10uMHz\n",
@@ -1414,23 +1477,96 @@ static int smu10_set_fine_grain_clk_vol(struct pp_hwmgr *hwmgr,
 					enum PP_OD_DPM_TABLE_COMMAND type,
 					long *input, uint32_t size)
 {
+	uint32_t min_freq, max_freq = 0;
+	struct smu10_hwmgr *smu10_data = (struct smu10_hwmgr *)(hwmgr->backend);
+	int ret = 0;
+
 	if (!hwmgr->od_enabled) {
 		pr_err("Fine grain not support\n");
 		return -EINVAL;
 	}
 
-	if (size != 2) {
-		pr_err("Input parameter number not correct\n");
+	if (!smu10_data->fine_grain_enabled) {
+		pr_err("pp_od_clk_voltage is not accessible if power_dpm_force_perfomance_level is not in manual mode!\n");
 		return -EINVAL;
 	}
 
 	if (type == PP_OD_EDIT_SCLK_VDDC_TABLE) {
-		if (input[0] == 0)
-			smu10_set_hard_min_gfxclk_by_freq(hwmgr, input[1]);
-		else if (input[0] == 1)
-			smu10_set_soft_max_gfxclk_by_freq(hwmgr, input[1]);
-		else
+		if (size != 2) {
+			pr_err("Input parameter number not correct\n");
 			return -EINVAL;
+		}
+
+		if (input[0] == 0) {
+			smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &min_freq);
+			if (input[1] < min_freq) {
+				pr_err("Fine grain setting minimum sclk (%ld) MHz is less than the minimum allowed (%d) MHz\n",
+					input[1], min_freq);
+				return -EINVAL;
+			}
+			smu10_data->gfx_actual_soft_min_freq = input[1];
+		} else if (input[0] == 1) {
+			smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &max_freq);
+			if (input[1] > max_freq) {
+				pr_err("Fine grain setting maximum sclk (%ld) MHz is greater than the maximum allowed (%d) MHz\n",
+					input[1], max_freq);
+				return -EINVAL;
+			}
+			smu10_data->gfx_actual_soft_max_freq = input[1];
+		} else {
+			return -EINVAL;
+		}
+	} else if (type == PP_OD_RESTORE_DEFAULT_TABLE) {
+		if (size != 0) {
+			pr_err("Input parameter number not correct\n");
+			return -EINVAL;
+		}
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMinGfxclkFrequency, &min_freq);
+		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetMaxGfxclkFrequency, &max_freq);
+
+		smu10_data->gfx_actual_soft_min_freq = min_freq;
+		smu10_data->gfx_actual_soft_max_freq = max_freq;
+
+		ret = smum_send_msg_to_smc_with_parameter(hwmgr,
+					PPSMC_MSG_SetHardMinGfxClk,
+					min_freq,
+					NULL);
+		if (ret)
+			return ret;
+
+		ret = smum_send_msg_to_smc_with_parameter(hwmgr,
+					PPSMC_MSG_SetSoftMaxGfxClk,
+					max_freq,
+					NULL);
+		if (ret)
+			return ret;
+	} else if (type == PP_OD_COMMIT_DPM_TABLE) {
+		if (size != 0) {
+			pr_err("Input parameter number not correct\n");
+			return -EINVAL;
+		}
+
+		if (smu10_data->gfx_actual_soft_min_freq > smu10_data->gfx_actual_soft_max_freq) {
+			pr_err("The setting minimun sclk (%d) MHz is greater than the setting maximum sclk (%d) MHz\n",
+					smu10_data->gfx_actual_soft_min_freq, smu10_data->gfx_actual_soft_max_freq);
+			return -EINVAL;
+		}
+
+		ret = smum_send_msg_to_smc_with_parameter(hwmgr,
+					PPSMC_MSG_SetHardMinGfxClk,
+					smu10_data->gfx_actual_soft_min_freq,
+					NULL);
+		if (ret)
+			return ret;
+
+		ret = smum_send_msg_to_smc_with_parameter(hwmgr,
+					PPSMC_MSG_SetSoftMaxGfxClk,
+					smu10_data->gfx_actual_soft_max_freq,
+					NULL);
+		if (ret)
+			return ret;
+	} else {
+		return -EINVAL;
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.h b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.h
index 6c9b5f060902..808e0ecbe1f0 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.h
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.h
@@ -283,6 +283,7 @@ struct smu10_hwmgr {
 	uint32_t                        vclk_soft_min;
 	uint32_t                        dclk_soft_min;
 	uint32_t                        gfx_actual_soft_min_freq;
+	uint32_t                        gfx_actual_soft_max_freq;
 	uint32_t                        gfx_min_freq_limit;
 	uint32_t                        gfx_max_freq_limit; /* in 10Khz*/
 
@@ -299,6 +300,8 @@ struct smu10_hwmgr {
 	bool need_min_deep_sleep_dcefclk;
 	uint32_t                             deep_sleep_dcefclk;
 	uint32_t                             num_active_display;
+
+	bool							fine_grain_enabled;
 };
 
 struct pp_hwmgr;
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c
index 82676c086ce4..a2681fe875ed 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c
@@ -235,7 +235,7 @@ static int smu7_get_current_pcie_lane_number(struct pp_hwmgr *hwmgr)
 /**
  * smu7_enable_smc_voltage_controller - Enable voltage control
  *
- * @hwmgr  the address of the powerplay hardware manager.
+ * @hwmgr:  the address of the powerplay hardware manager.
  * Return:   always PP_Result_OK
  */
 static int smu7_enable_smc_voltage_controller(struct pp_hwmgr *hwmgr)
@@ -4501,7 +4501,7 @@ static int smu7_display_configuration_changed_task(struct pp_hwmgr *hwmgr)
  * smu7_set_max_fan_rpm_output - Set maximum target operating fan output RPM
  *
  * @hwmgr:  the address of the powerplay hardware manager.
- * @usMaxFanRpm:  max operating fan RPM value.
+ * @us_max_fan_rpm:  max operating fan RPM value.
  * Return:   The response that came from the SMC.
  */
 static int smu7_set_max_fan_rpm_output(struct pp_hwmgr *hwmgr, uint16_t us_max_fan_rpm)
@@ -5216,10 +5216,10 @@ static int smu7_set_watermarks_for_clocks_ranges(struct pp_hwmgr *hwmgr,
 		for (j = 0; j < dep_sclk_table->count; j++) {
 			valid_entry = false;
 			for (k = 0; k < watermarks->num_wm_sets; k++) {
-				if (dep_sclk_table->entries[i].clk / 10 >= watermarks->wm_clk_ranges[k].wm_min_eng_clk_in_khz &&
-				    dep_sclk_table->entries[i].clk / 10 < watermarks->wm_clk_ranges[k].wm_max_eng_clk_in_khz &&
-				    dep_mclk_table->entries[i].clk / 10 >= watermarks->wm_clk_ranges[k].wm_min_mem_clk_in_khz &&
-				    dep_mclk_table->entries[i].clk / 10 < watermarks->wm_clk_ranges[k].wm_max_mem_clk_in_khz) {
+				if (dep_sclk_table->entries[i].clk >= watermarks->wm_clk_ranges[k].wm_min_eng_clk_in_khz / 10 &&
+				    dep_sclk_table->entries[i].clk < watermarks->wm_clk_ranges[k].wm_max_eng_clk_in_khz / 10 &&
+				    dep_mclk_table->entries[i].clk >= watermarks->wm_clk_ranges[k].wm_min_mem_clk_in_khz / 10 &&
+				    dep_mclk_table->entries[i].clk < watermarks->wm_clk_ranges[k].wm_max_mem_clk_in_khz / 10) {
 					valid_entry = true;
 					table->DisplayWatermark[i][j] = watermarks->wm_clk_ranges[k].wm_set_id;
 					break;
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
index 1b47f94e0331..22b636e2b89b 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
@@ -542,11 +542,11 @@ static int vega10_get_socclk_for_voltage_evv(struct pp_hwmgr *hwmgr,
 
 #define ATOM_VIRTUAL_VOLTAGE_ID0             0xff01
 /**
-* Get Leakage VDDC based on leakage ID.
-*
-* @param    hwmgr  the address of the powerplay hardware manager.
-* @return   always 0.
-*/
+ * Get Leakage VDDC based on leakage ID.
+ *
+ * @hwmgr:  the address of the powerplay hardware manager.
+ * return:  always 0.
+ */
 static int vega10_get_evv_voltages(struct pp_hwmgr *hwmgr)
 {
 	struct vega10_hwmgr *data = hwmgr->backend;
@@ -600,9 +600,9 @@ static int vega10_get_evv_voltages(struct pp_hwmgr *hwmgr)
 /**
  * Change virtual leakage voltage to actual value.
  *
- * @param     hwmgr  the address of the powerplay hardware manager.
- * @param     pointer to changing voltage
- * @param     pointer to leakage table
+ * @hwmgr:         the address of the powerplay hardware manager.
+ * @voltage:       pointer to changing voltage
+ * @leakage_table: pointer to leakage table
  */
 static void vega10_patch_with_vdd_leakage(struct pp_hwmgr *hwmgr,
 		uint16_t *voltage, struct vega10_leakage_voltage *leakage_table)
@@ -624,13 +624,13 @@ static void vega10_patch_with_vdd_leakage(struct pp_hwmgr *hwmgr,
 }
 
 /**
-* Patch voltage lookup table by EVV leakages.
-*
-* @param     hwmgr  the address of the powerplay hardware manager.
-* @param     pointer to voltage lookup table
-* @param     pointer to leakage table
-* @return     always 0
-*/
+ * Patch voltage lookup table by EVV leakages.
+ *
+ * @hwmgr:         the address of the powerplay hardware manager.
+ * @lookup_table:  pointer to voltage lookup table
+ * @leakage_table: pointer to leakage table
+ * return:         always 0
+ */
 static int vega10_patch_lookup_table_with_leakage(struct pp_hwmgr *hwmgr,
 		phm_ppt_v1_voltage_lookup_table *lookup_table,
 		struct vega10_leakage_voltage *leakage_table)
@@ -1001,13 +1001,12 @@ static int vega10_setup_asic_task(struct pp_hwmgr *hwmgr)
 }
 
 /**
-* Remove repeated voltage values and create table with unique values.
-*
-* @param    hwmgr  the address of the powerplay hardware manager.
-* @param    vol_table  the pointer to changing voltage table
-* @return    0 in success
-*/
-
+ * Remove repeated voltage values and create table with unique values.
+ *
+ * @hwmgr:      the address of the powerplay hardware manager.
+ * @vol_table:  the pointer to changing voltage table
+ * return:      0 in success
+ */
 static int vega10_trim_voltage_table(struct pp_hwmgr *hwmgr,
 		struct pp_atomfwctrl_voltage_table *vol_table)
 {
@@ -1151,11 +1150,11 @@ static void vega10_trim_voltage_table_to_fit_state_table(
 }
 
 /**
-* Create Voltage Tables.
-*
-* @param    hwmgr  the address of the powerplay hardware manager.
-* @return   always 0
-*/
+ * Create Voltage Tables.
+ *
+ * @hwmgr:  the address of the powerplay hardware manager.
+ * return:  always 0
+ */
 static int vega10_construct_voltage_tables(struct pp_hwmgr *hwmgr)
 {
 	struct vega10_hwmgr *data = hwmgr->backend;
@@ -1212,11 +1211,11 @@ static int vega10_construct_voltage_tables(struct pp_hwmgr *hwmgr)
 }
 
 /*
- * @fn vega10_init_dpm_state
- * @brief Function to initialize all Soft Min/Max and Hard Min/Max to 0xff.
+ * vega10_init_dpm_state
+ * Function to initialize all Soft Min/Max and Hard Min/Max to 0xff.
  *
- * @param    dpm_state - the address of the DPM Table to initiailize.
- * @return   None.
+ * @dpm_state: - the address of the DPM Table to initiailize.
+ * return:   None.
  */
 static void vega10_init_dpm_state(struct vega10_dpm_state *dpm_state)
 {
@@ -1460,11 +1459,11 @@ static int vega10_setup_default_dpm_tables(struct pp_hwmgr *hwmgr)
 }
 
 /*
- * @fn vega10_populate_ulv_state
- * @brief Function to provide parameters for Utral Low Voltage state to SMC.
+ * vega10_populate_ulv_state
+ * Function to provide parameters for Utral Low Voltage state to SMC.
  *
- * @param    hwmgr - the address of the hardware manager.
- * @return   Always 0.
+ * @hwmgr: - the address of the hardware manager.
+ * return:   Always 0.
  */
 static int vega10_populate_ulv_state(struct pp_hwmgr *hwmgr)
 {
@@ -1506,6 +1505,48 @@ static int vega10_populate_single_lclk_level(struct pp_hwmgr *hwmgr,
 	return 0;
 }
 
+static int vega10_override_pcie_parameters(struct pp_hwmgr *hwmgr)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev);
+	struct vega10_hwmgr *data =
+			(struct vega10_hwmgr *)(hwmgr->backend);
+	uint32_t pcie_gen = 0, pcie_width = 0;
+	PPTable_t *pp_table = &(data->smc_state_table.pp_table);
+	int i;
+
+	if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4)
+		pcie_gen = 3;
+	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
+		pcie_gen = 2;
+	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2)
+		pcie_gen = 1;
+	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1)
+		pcie_gen = 0;
+
+	if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X16)
+		pcie_width = 6;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X12)
+		pcie_width = 5;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X8)
+		pcie_width = 4;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X4)
+		pcie_width = 3;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X2)
+		pcie_width = 2;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X1)
+		pcie_width = 1;
+
+	for (i = 0; i < NUM_LINK_LEVELS; i++) {
+		if (pp_table->PcieGenSpeed[i] > pcie_gen)
+			pp_table->PcieGenSpeed[i] = pcie_gen;
+
+		if (pp_table->PcieLaneCount[i] > pcie_width)
+			pp_table->PcieLaneCount[i] = pcie_width;
+	}
+
+	return 0;
+}
+
 static int vega10_populate_smc_link_levels(struct pp_hwmgr *hwmgr)
 {
 	int result = -1;
@@ -1545,13 +1586,13 @@ static int vega10_populate_smc_link_levels(struct pp_hwmgr *hwmgr)
 }
 
 /**
-* Populates single SMC GFXSCLK structure using the provided engine clock
-*
-* @param    hwmgr      the address of the hardware manager
-* @param    gfx_clock  the GFX clock to use to populate the structure.
-* @param    current_gfxclk_level  location in PPTable for the SMC GFXCLK structure.
-*/
-
+ * Populates single SMC GFXSCLK structure using the provided engine clock
+ *
+ * @hwmgr:      the address of the hardware manager
+ * @gfx_clock:  the GFX clock to use to populate the structure.
+ * @current_gfxclk_level:  location in PPTable for the SMC GFXCLK structure.
+ * @acg_freq:   ACG frequenty to return (MHz)
+ */
 static int vega10_populate_single_gfx_level(struct pp_hwmgr *hwmgr,
 		uint32_t gfx_clock, PllSetting_t *current_gfxclk_level,
 		uint32_t *acg_freq)
@@ -1610,12 +1651,13 @@ static int vega10_populate_single_gfx_level(struct pp_hwmgr *hwmgr,
 }
 
 /**
- * @brief Populates single SMC SOCCLK structure using the provided clock.
+ * Populates single SMC SOCCLK structure using the provided clock.
  *
- * @param    hwmgr - the address of the hardware manager.
- * @param    soc_clock - the SOC clock to use to populate the structure.
- * @param    current_socclk_level - location in PPTable for the SMC SOCCLK structure.
- * @return   0 on success..
+ * @hwmgr:     the address of the hardware manager.
+ * @soc_clock: the SOC clock to use to populate the structure.
+ * @current_soc_did:   DFS divider to pass back to caller
+ * @current_vol_index: index of current VDD to pass back to caller
+ * return:      0 on success
  */
 static int vega10_populate_single_soc_level(struct pp_hwmgr *hwmgr,
 		uint32_t soc_clock, uint8_t *current_soc_did,
@@ -1659,10 +1701,10 @@ static int vega10_populate_single_soc_level(struct pp_hwmgr *hwmgr,
 }
 
 /**
-* Populates all SMC SCLK levels' structure based on the trimmed allowed dpm engine clock states
-*
-* @param    hwmgr      the address of the hardware manager
-*/
+ * Populates all SMC SCLK levels' structure based on the trimmed allowed dpm engine clock states
+ *
+ * @hwmgr:      the address of the hardware manager
+ */
 static int vega10_populate_all_graphic_levels(struct pp_hwmgr *hwmgr)
 {
 	struct vega10_hwmgr *data = hwmgr->backend;
@@ -1746,12 +1788,12 @@ static void vega10_populate_vddc_soc_levels(struct pp_hwmgr *hwmgr)
 	}
 }
 
-/**
- * @brief Populates single SMC GFXCLK structure using the provided clock.
+/*
+ * Populates single SMC GFXCLK structure using the provided clock.
  *
- * @param    hwmgr - the address of the hardware manager.
- * @param    mem_clock - the memory clock to use to populate the structure.
- * @return   0 on success..
+ * @hwmgr:     the address of the hardware manager.
+ * @mem_clock: the memory clock to use to populate the structure.
+ * return:     0 on success..
  */
 static int vega10_populate_single_memory_level(struct pp_hwmgr *hwmgr,
 		uint32_t mem_clock, uint8_t *current_mem_vid,
@@ -1808,10 +1850,10 @@ static int vega10_populate_single_memory_level(struct pp_hwmgr *hwmgr,
 }
 
 /**
- * @brief Populates all SMC MCLK levels' structure based on the trimmed allowed dpm memory clock states.
+ * Populates all SMC MCLK levels' structure based on the trimmed allowed dpm memory clock states.
  *
- * @param    pHwMgr - the address of the hardware manager.
- * @return   PP_Result_OK on success.
+ * @hwmgr:  the address of the hardware manager.
+ * return:   PP_Result_OK on success.
  */
 static int vega10_populate_all_memory_levels(struct pp_hwmgr *hwmgr)
 {
@@ -2486,12 +2528,11 @@ static void vega10_check_dpm_table_updated(struct pp_hwmgr *hwmgr)
 }
 
 /**
-* Initializes the SMC table and uploads it
-*
-* @param    hwmgr  the address of the powerplay hardware manager.
-* @param    pInput  the pointer to input data (PowerState)
-* @return   always 0
-*/
+ * Initializes the SMC table and uploads it
+ *
+ * @hwmgr:  the address of the powerplay hardware manager.
+ * return:  always 0
+ */
 static int vega10_init_smc_table(struct pp_hwmgr *hwmgr)
 {
 	int result;
@@ -2557,6 +2598,11 @@ static int vega10_init_smc_table(struct pp_hwmgr *hwmgr)
 			"Failed to initialize Link Level!",
 			return result);
 
+	result = vega10_override_pcie_parameters(hwmgr);
+	PP_ASSERT_WITH_CODE(!result,
+			"Failed to override pcie parameters!",
+			return result);
+
 	result = vega10_populate_all_graphic_levels(hwmgr);
 	PP_ASSERT_WITH_CODE(!result,
 			"Failed to initialize Graphics Level!",
@@ -2864,11 +2910,11 @@ static int vega10_stop_dpm(struct pp_hwmgr *hwmgr, uint32_t bitmap)
 }
 
 /**
- * @brief Tell SMC to enabled the supported DPMs.
+ * Tell SMC to enabled the supported DPMs.
  *
- * @param    hwmgr - the address of the powerplay hardware manager.
- * @Param    bitmap - bitmap for the features to enabled.
- * @return   0 on at least one DPM is successfully enabled.
+ * @hwmgr:   the address of the powerplay hardware manager.
+ * @bitmap:  bitmap for the features to enabled.
+ * return:  0 on at least one DPM is successfully enabled.
  */
 static int vega10_start_dpm(struct pp_hwmgr *hwmgr, uint32_t bitmap)
 {
@@ -2923,6 +2969,7 @@ static int vega10_start_dpm(struct pp_hwmgr *hwmgr, uint32_t bitmap)
 	return 0;
 }
 
+
 static int vega10_enable_disable_PCC_limit_feature(struct pp_hwmgr *hwmgr, bool enable)
 {
 	struct vega10_hwmgr *data = hwmgr->backend;
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.h b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.h
index f752b4ad0c8a..07c06f8c90b0 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.h
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.h
@@ -442,5 +442,6 @@ int vega10_update_uvd_dpm(struct pp_hwmgr *hwmgr, bool bgate);
 int vega10_update_samu_dpm(struct pp_hwmgr *hwmgr, bool bgate);
 int vega10_update_acp_dpm(struct pp_hwmgr *hwmgr, bool bgate);
 int vega10_enable_disable_vce_dpm(struct pp_hwmgr *hwmgr, bool enable);
+int vega10_hwmgr_init(struct pp_hwmgr *hwmgr);
 
 #endif /* _VEGA10_HWMGR_H_ */
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c
index dc206fa88c5e..43e01d880f7c 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c
@@ -481,6 +481,67 @@ static void vega12_init_dpm_state(struct vega12_dpm_state *dpm_state)
 	dpm_state->hard_max_level = 0xffff;
 }
 
+static int vega12_override_pcie_parameters(struct pp_hwmgr *hwmgr)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev);
+	struct vega12_hwmgr *data =
+			(struct vega12_hwmgr *)(hwmgr->backend);
+	uint32_t pcie_gen = 0, pcie_width = 0, smu_pcie_arg, pcie_gen_arg, pcie_width_arg;
+	PPTable_t *pp_table = &(data->smc_state_table.pp_table);
+	int i;
+	int ret;
+
+	if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4)
+		pcie_gen = 3;
+	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
+		pcie_gen = 2;
+	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2)
+		pcie_gen = 1;
+	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1)
+		pcie_gen = 0;
+
+	if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X16)
+		pcie_width = 6;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X12)
+		pcie_width = 5;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X8)
+		pcie_width = 4;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X4)
+		pcie_width = 3;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X2)
+		pcie_width = 2;
+	else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X1)
+		pcie_width = 1;
+
+	/* Bit 31:16: LCLK DPM level. 0 is DPM0, and 1 is DPM1
+	 * Bit 15:8:  PCIE GEN, 0 to 3 corresponds to GEN1 to GEN4
+	 * Bit 7:0:   PCIE lane width, 1 to 7 corresponds is x1 to x32
+	 */
+	for (i = 0; i < NUM_LINK_LEVELS; i++) {
+		pcie_gen_arg = (pp_table->PcieGenSpeed[i] > pcie_gen) ? pcie_gen :
+			pp_table->PcieGenSpeed[i];
+		pcie_width_arg = (pp_table->PcieLaneCount[i] > pcie_width) ? pcie_width :
+			pp_table->PcieLaneCount[i];
+
+		if (pcie_gen_arg != pp_table->PcieGenSpeed[i] || pcie_width_arg !=
+		    pp_table->PcieLaneCount[i]) {
+			smu_pcie_arg = (i << 16) | (pcie_gen_arg << 8) | pcie_width_arg;
+			ret = smum_send_msg_to_smc_with_parameter(hwmgr,
+				PPSMC_MSG_OverridePcieParameters, smu_pcie_arg,
+				NULL);
+			PP_ASSERT_WITH_CODE(!ret,
+				"[OverridePcieParameters] Attempt to override pcie params failed!",
+				return ret);
+		}
+
+		/* update the pptable */
+		pp_table->PcieGenSpeed[i] = pcie_gen_arg;
+		pp_table->PcieLaneCount[i] = pcie_width_arg;
+	}
+
+	return 0;
+}
+
 static int vega12_get_number_of_dpm_level(struct pp_hwmgr *hwmgr,
 		PPCLK_e clk_id, uint32_t *num_of_levels)
 {
@@ -718,12 +779,11 @@ static int vega12_save_default_power_profile(struct pp_hwmgr *hwmgr)
 #endif
 
 /**
-* Initializes the SMC table and uploads it
-*
-* @param    hwmgr  the address of the powerplay hardware manager.
-* @param    pInput  the pointer to input data (PowerState)
-* @return   always 0
-*/
+ * Initializes the SMC table and uploads it
+ *
+ * @hwmgr:  the address of the powerplay hardware manager.
+ * return:  always 0
+ */
 static int vega12_init_smc_table(struct pp_hwmgr *hwmgr)
 {
 	int result;
@@ -969,6 +1029,11 @@ static int vega12_enable_dpm_tasks(struct pp_hwmgr *hwmgr)
 			"Failed to enable all smu features!",
 			return result);
 
+	result = vega12_override_pcie_parameters(hwmgr);
+	PP_ASSERT_WITH_CODE(!result,
+			"[EnableDPMTasks] Failed to override pcie parameters!",
+			return result);
+
 	tmp_result = vega12_power_control_set_level(hwmgr);
 	PP_ASSERT_WITH_CODE(!tmp_result,
 			"Failed to power control set level!",
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c
index da84012b7fd5..f19964c69a00 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c
@@ -771,12 +771,11 @@ static int vega20_setup_default_dpm_tables(struct pp_hwmgr *hwmgr)
 }
 
 /**
-* Initializes the SMC table and uploads it
-*
-* @param    hwmgr  the address of the powerplay hardware manager.
-* @param    pInput  the pointer to input data (PowerState)
-* @return   always 0
-*/
+ * Initializes the SMC table and uploads it
+ *
+ * @hwmgr:  the address of the powerplay hardware manager.
+ * return:  always 0
+ */
 static int vega20_init_smc_table(struct pp_hwmgr *hwmgr)
 {
 	int result;
@@ -832,7 +831,9 @@ static int vega20_override_pcie_parameters(struct pp_hwmgr *hwmgr)
 	struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev);
 	struct vega20_hwmgr *data =
 			(struct vega20_hwmgr *)(hwmgr->backend);
-	uint32_t pcie_gen = 0, pcie_width = 0, smu_pcie_arg;
+	uint32_t pcie_gen = 0, pcie_width = 0, smu_pcie_arg, pcie_gen_arg, pcie_width_arg;
+	PPTable_t *pp_table = &(data->smc_state_table.pp_table);
+	int i;
 	int ret;
 
 	if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4)
@@ -861,17 +862,27 @@ static int vega20_override_pcie_parameters(struct pp_hwmgr *hwmgr)
 	 * Bit 15:8:  PCIE GEN, 0 to 3 corresponds to GEN1 to GEN4
 	 * Bit 7:0:   PCIE lane width, 1 to 7 corresponds is x1 to x32
 	 */
-	smu_pcie_arg = (1 << 16) | (pcie_gen << 8) | pcie_width;
-	ret = smum_send_msg_to_smc_with_parameter(hwmgr,
-			PPSMC_MSG_OverridePcieParameters, smu_pcie_arg,
-			NULL);
-	PP_ASSERT_WITH_CODE(!ret,
-		"[OverridePcieParameters] Attempt to override pcie params failed!",
-		return ret);
+	for (i = 0; i < NUM_LINK_LEVELS; i++) {
+		pcie_gen_arg = (pp_table->PcieGenSpeed[i] > pcie_gen) ? pcie_gen :
+			pp_table->PcieGenSpeed[i];
+		pcie_width_arg = (pp_table->PcieLaneCount[i] > pcie_width) ? pcie_width :
+			pp_table->PcieLaneCount[i];
+
+		if (pcie_gen_arg != pp_table->PcieGenSpeed[i] || pcie_width_arg !=
+		    pp_table->PcieLaneCount[i]) {
+			smu_pcie_arg = (i << 16) | (pcie_gen_arg << 8) | pcie_width_arg;
+			ret = smum_send_msg_to_smc_with_parameter(hwmgr,
+				PPSMC_MSG_OverridePcieParameters, smu_pcie_arg,
+				NULL);
+			PP_ASSERT_WITH_CODE(!ret,
+				"[OverridePcieParameters] Attempt to override pcie params failed!",
+				return ret);
+		}
 
-	data->pcie_parameters_override = true;
-	data->pcie_gen_level1 = pcie_gen;
-	data->pcie_width_level1 = pcie_width;
+		/* update the pptable */
+		pp_table->PcieGenSpeed[i] = pcie_gen_arg;
+		pp_table->PcieLaneCount[i] = pcie_width_arg;
+	}
 
 	return 0;
 }
@@ -3320,9 +3331,7 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 			data->od8_settings.od8_settings_array;
 	OverDriveTable_t *od_table =
 			&(data->smc_state_table.overdrive_table);
-	struct phm_ppt_v3_information *pptable_information =
-		(struct phm_ppt_v3_information *)hwmgr->pptable;
-	PPTable_t *pptable = (PPTable_t *)pptable_information->smc_pptable;
+	PPTable_t *pptable = &(data->smc_state_table.pp_table);
 	struct pp_clock_levels_with_latency clocks;
 	struct vega20_single_dpm_table *fclk_dpm_table =
 			&(data->dpm_table.fclk_table);
@@ -3421,13 +3430,9 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 		current_lane_width =
 			vega20_get_current_pcie_link_width_level(hwmgr);
 		for (i = 0; i < NUM_LINK_LEVELS; i++) {
-			if (i == 1 && data->pcie_parameters_override) {
-				gen_speed = data->pcie_gen_level1;
-				lane_width = data->pcie_width_level1;
-			} else {
-				gen_speed = pptable->PcieGenSpeed[i];
-				lane_width = pptable->PcieLaneCount[i];
-			}
+			gen_speed = pptable->PcieGenSpeed[i];
+			lane_width = pptable->PcieLaneCount[i];
+
 			size += sprintf(buf + size, "%d: %s %s %dMhz %s\n", i,
 					(gen_speed == 0) ? "2.5GT/s," :
 					(gen_speed == 1) ? "5.0GT/s," :
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index 8b867a6d52b5..d143ef1b460b 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -266,6 +266,119 @@ int smu_dpm_set_power_gate(struct smu_context *smu, uint32_t block_type,
 	return ret;
 }
 
+/**
+ * smu_set_user_clk_dependencies - set user profile clock dependencies
+ *
+ * @smu:	smu_context pointer
+ * @clk:	enum smu_clk_type type
+ *
+ * Enable/Disable the clock dependency for the @clk type.
+ */
+static void smu_set_user_clk_dependencies(struct smu_context *smu, enum smu_clk_type clk)
+{
+	if (smu->adev->in_suspend)
+		return;
+
+	/*
+	 * mclk, fclk and socclk are interdependent
+	 * on each other
+	 */
+	if (clk == SMU_MCLK) {
+		/* reset clock dependency */
+		smu->user_dpm_profile.clk_dependency = 0;
+		/* set mclk dependent clocks(fclk and socclk) */
+		smu->user_dpm_profile.clk_dependency = BIT(SMU_FCLK) | BIT(SMU_SOCCLK);
+	} else if (clk == SMU_FCLK) {
+		/* give priority to mclk, if mclk dependent clocks are set */
+		if (smu->user_dpm_profile.clk_dependency == (BIT(SMU_FCLK) | BIT(SMU_SOCCLK)))
+			return;
+
+		/* reset clock dependency */
+		smu->user_dpm_profile.clk_dependency = 0;
+		/* set fclk dependent clocks(mclk and socclk) */
+		smu->user_dpm_profile.clk_dependency = BIT(SMU_MCLK) | BIT(SMU_SOCCLK);
+	} else if (clk == SMU_SOCCLK) {
+		/* give priority to mclk, if mclk dependent clocks are set */
+		if (smu->user_dpm_profile.clk_dependency == (BIT(SMU_FCLK) | BIT(SMU_SOCCLK)))
+			return;
+
+		/* reset clock dependency */
+		smu->user_dpm_profile.clk_dependency = 0;
+		/* set socclk dependent clocks(mclk and fclk) */
+		smu->user_dpm_profile.clk_dependency = BIT(SMU_MCLK) | BIT(SMU_FCLK);
+	} else
+		/* add clk dependencies here, if any */
+		return;
+}
+
+/**
+ * smu_restore_dpm_user_profile - reinstate user dpm profile
+ *
+ * @smu:	smu_context pointer
+ *
+ * Restore the saved user power configurations include power limit,
+ * clock frequencies, fan control mode and fan speed.
+ */
+static void smu_restore_dpm_user_profile(struct smu_context *smu)
+{
+	struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
+	int ret = 0;
+
+	if (!smu->adev->in_suspend)
+		return;
+
+	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
+		return;
+
+	/* Enable restore flag */
+	smu->user_dpm_profile.flags = SMU_DPM_USER_PROFILE_RESTORE;
+
+	/* set the user dpm power limit */
+	if (smu->user_dpm_profile.power_limit) {
+		ret = smu_set_power_limit(smu, smu->user_dpm_profile.power_limit);
+		if (ret)
+			dev_err(smu->adev->dev, "Failed to set power limit value\n");
+	}
+
+	/* set the user dpm clock configurations */
+	if (smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) {
+		enum smu_clk_type clk_type;
+
+		for (clk_type = 0; clk_type < SMU_CLK_COUNT; clk_type++) {
+			/*
+			 * Iterate over smu clk type and force the saved user clk
+			 * configs, skip if clock dependency is enabled
+			 */
+			if (!(smu->user_dpm_profile.clk_dependency & BIT(clk_type)) &&
+					smu->user_dpm_profile.clk_mask[clk_type]) {
+				ret = smu_force_clk_levels(smu, clk_type,
+						smu->user_dpm_profile.clk_mask[clk_type]);
+				if (ret)
+					dev_err(smu->adev->dev, "Failed to set clock type = %d\n",
+							clk_type);
+			}
+		}
+	}
+
+	/* set the user dpm fan configurations */
+	if (smu->user_dpm_profile.fan_mode == AMD_FAN_CTRL_MANUAL) {
+		ret = smu_set_fan_control_mode(smu, smu->user_dpm_profile.fan_mode);
+		if (ret) {
+			dev_err(smu->adev->dev, "Failed to set manual fan control mode\n");
+			return;
+		}
+
+		if (!ret && smu->user_dpm_profile.fan_speed_percent) {
+			ret = smu_set_fan_speed_percent(smu, smu->user_dpm_profile.fan_speed_percent);
+			if (ret)
+				dev_err(smu->adev->dev, "Failed to set manual fan speed\n");
+		}
+	}
+
+	/* Disable restore flag */
+	smu->user_dpm_profile.flags &= ~SMU_DPM_USER_PROFILE_RESTORE;
+}
+
 int smu_get_power_num_states(struct smu_context *smu,
 			     struct pp_states_info *state_info)
 {
@@ -288,6 +401,20 @@ bool is_support_sw_smu(struct amdgpu_device *adev)
 	return false;
 }
 
+bool is_support_cclk_dpm(struct amdgpu_device *adev)
+{
+	struct smu_context *smu = &adev->smu;
+
+	if (!is_support_sw_smu(adev))
+		return false;
+
+	if (!smu_feature_is_enabled(smu, SMU_FEATURE_CCLK_DPM_BIT))
+		return false;
+
+	return true;
+}
+
+
 int smu_sys_get_pp_table(struct smu_context *smu, void **table)
 {
 	struct smu_table_context *smu_table = &smu->smu_table;
@@ -405,8 +532,6 @@ static int smu_set_funcs(struct amdgpu_device *adev)
 		break;
 	case CHIP_VANGOGH:
 		vangogh_set_ppt_funcs(smu);
-		/* enable the OD by default to allow the fine grain tuning function */
-		smu->od_enabled = true;
 		break;
 	default:
 		return -EINVAL;
@@ -478,9 +603,6 @@ static int smu_late_init(void *handle)
 
 	smu_set_fine_grain_gfx_freq_parameters(smu);
 
-	if (adev->asic_type == CHIP_VANGOGH)
-		return 0;
-
 	if (!smu->pm_enabled)
 		return 0;
 
@@ -517,6 +639,8 @@ static int smu_late_init(void *handle)
 			AMD_PP_TASK_COMPLETE_INIT,
 			false);
 
+	smu_restore_dpm_user_profile(smu);
+
 	return 0;
 }
 
@@ -1610,6 +1734,12 @@ int smu_force_performance_level(struct smu_context *smu, enum amd_dpm_forced_lev
 
 	mutex_unlock(&smu->mutex);
 
+	/* reset user dpm clock state */
+	if (!ret && smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) {
+		memset(smu->user_dpm_profile.clk_mask, 0, sizeof(smu->user_dpm_profile.clk_mask));
+		smu->user_dpm_profile.clk_dependency = 0;
+	}
+
 	return ret;
 }
 
@@ -1644,8 +1774,13 @@ int smu_force_clk_levels(struct smu_context *smu,
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs && smu->ppt_funcs->force_clk_levels)
+	if (smu->ppt_funcs && smu->ppt_funcs->force_clk_levels) {
 		ret = smu->ppt_funcs->force_clk_levels(smu, clk_type, mask);
+		if (!ret && smu->user_dpm_profile.flags != SMU_DPM_USER_PROFILE_RESTORE) {
+			smu->user_dpm_profile.clk_mask[clk_type] = mask;
+			smu_set_user_clk_dependencies(smu, clk_type);
+		}
+	}
 
 	mutex_unlock(&smu->mutex);
 
@@ -1887,6 +2022,7 @@ int smu_set_gfx_cgpg(struct smu_context *smu, bool enabled)
 
 int smu_set_fan_speed_rpm(struct smu_context *smu, uint32_t speed)
 {
+	u32 percent;
 	int ret = 0;
 
 	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
@@ -1894,8 +2030,12 @@ int smu_set_fan_speed_rpm(struct smu_context *smu, uint32_t speed)
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs->set_fan_speed_rpm)
-		ret = smu->ppt_funcs->set_fan_speed_rpm(smu, speed);
+	if (smu->ppt_funcs->set_fan_speed_percent) {
+		percent = speed * 100 / smu->fan_max_rpm;
+		ret = smu->ppt_funcs->set_fan_speed_percent(smu, percent);
+		if (!ret && smu->user_dpm_profile.flags != SMU_DPM_USER_PROFILE_RESTORE)
+			smu->user_dpm_profile.fan_speed_percent = percent;
+	}
 
 	mutex_unlock(&smu->mutex);
 
@@ -1904,22 +2044,40 @@ int smu_set_fan_speed_rpm(struct smu_context *smu, uint32_t speed)
 
 int smu_get_power_limit(struct smu_context *smu,
 			uint32_t *limit,
-			bool max_setting)
+			enum smu_ppt_limit_level limit_level)
 {
+	uint32_t limit_type = *limit >> 24;
+	int ret = 0;
+
 	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&smu->mutex);
 
-	*limit = (max_setting ? smu->max_power_limit : smu->current_power_limit);
+	if (limit_type != SMU_DEFAULT_PPT_LIMIT) {
+		if (smu->ppt_funcs->get_ppt_limit)
+			ret = smu->ppt_funcs->get_ppt_limit(smu, limit, limit_type, limit_level);
+	} else {
+		switch (limit_level) {
+		case SMU_PPT_LIMIT_CURRENT:
+			*limit = smu->current_power_limit;
+			break;
+		case SMU_PPT_LIMIT_MAX:
+			*limit = smu->max_power_limit;
+			break;
+		default:
+			break;
+		}
+	}
 
 	mutex_unlock(&smu->mutex);
 
-	return 0;
+	return ret;
 }
 
 int smu_set_power_limit(struct smu_context *smu, uint32_t limit)
 {
+	uint32_t limit_type = limit >> 24;
 	int ret = 0;
 
 	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
@@ -1927,6 +2085,12 @@ int smu_set_power_limit(struct smu_context *smu, uint32_t limit)
 
 	mutex_lock(&smu->mutex);
 
+	if (limit_type != SMU_DEFAULT_PPT_LIMIT)
+		if (smu->ppt_funcs->set_power_limit) {
+			ret = smu->ppt_funcs->set_power_limit(smu, limit);
+			goto out;
+		}
+
 	if (limit > smu->max_power_limit) {
 		dev_err(smu->adev->dev,
 			"New power limit (%d) is over the max allowed %d\n",
@@ -1937,8 +2101,11 @@ int smu_set_power_limit(struct smu_context *smu, uint32_t limit)
 	if (!limit)
 		limit = smu->current_power_limit;
 
-	if (smu->ppt_funcs->set_power_limit)
+	if (smu->ppt_funcs->set_power_limit) {
 		ret = smu->ppt_funcs->set_power_limit(smu, limit);
+		if (!ret && smu->user_dpm_profile.flags != SMU_DPM_USER_PROFILE_RESTORE)
+			smu->user_dpm_profile.power_limit = limit;
+	}
 
 out:
 	mutex_unlock(&smu->mutex);
@@ -2115,11 +2282,19 @@ int smu_set_fan_control_mode(struct smu_context *smu, int value)
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs->set_fan_control_mode)
+	if (smu->ppt_funcs->set_fan_control_mode) {
 		ret = smu->ppt_funcs->set_fan_control_mode(smu, value);
+		if (!ret && smu->user_dpm_profile.flags != SMU_DPM_USER_PROFILE_RESTORE)
+			smu->user_dpm_profile.fan_mode = value;
+	}
 
 	mutex_unlock(&smu->mutex);
 
+	/* reset user dpm fan speed */
+	if (!ret && value != AMD_FAN_CTRL_MANUAL &&
+			smu->user_dpm_profile.flags != SMU_DPM_USER_PROFILE_RESTORE)
+		smu->user_dpm_profile.fan_speed_percent = 0;
+
 	return ret;
 }
 
@@ -2127,17 +2302,15 @@ int smu_get_fan_speed_percent(struct smu_context *smu, uint32_t *speed)
 {
 	int ret = 0;
 	uint32_t percent;
-	uint32_t current_rpm;
 
 	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs->get_fan_speed_rpm) {
-		ret = smu->ppt_funcs->get_fan_speed_rpm(smu, &current_rpm);
+	if (smu->ppt_funcs->get_fan_speed_percent) {
+		ret = smu->ppt_funcs->get_fan_speed_percent(smu, &percent);
 		if (!ret) {
-			percent = current_rpm * 100 / smu->fan_max_rpm;
 			*speed = percent > 100 ? 100 : percent;
 		}
 	}
@@ -2151,18 +2324,18 @@ int smu_get_fan_speed_percent(struct smu_context *smu, uint32_t *speed)
 int smu_set_fan_speed_percent(struct smu_context *smu, uint32_t speed)
 {
 	int ret = 0;
-	uint32_t rpm;
 
 	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs->set_fan_speed_rpm) {
+	if (smu->ppt_funcs->set_fan_speed_percent) {
 		if (speed > 100)
 			speed = 100;
-		rpm = speed * smu->fan_max_rpm / 100;
-		ret = smu->ppt_funcs->set_fan_speed_rpm(smu, rpm);
+		ret = smu->ppt_funcs->set_fan_speed_percent(smu, speed);
+		if (!ret && smu->user_dpm_profile.flags != SMU_DPM_USER_PROFILE_RESTORE)
+			smu->user_dpm_profile.fan_speed_percent = speed;
 	}
 
 	mutex_unlock(&smu->mutex);
@@ -2173,14 +2346,17 @@ int smu_set_fan_speed_percent(struct smu_context *smu, uint32_t speed)
 int smu_get_fan_speed_rpm(struct smu_context *smu, uint32_t *speed)
 {
 	int ret = 0;
+	u32 percent;
 
 	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs->get_fan_speed_rpm)
-		ret = smu->ppt_funcs->get_fan_speed_rpm(smu, speed);
+	if (smu->ppt_funcs->get_fan_speed_percent) {
+		ret = smu->ppt_funcs->get_fan_speed_percent(smu, &percent);
+		*speed = percent * smu->fan_max_rpm / 100;
+	}
 
 	mutex_unlock(&smu->mutex);
 
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
index cd7b411457ff..9f0d03ae3109 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
@@ -1080,15 +1080,27 @@ static int arcturus_read_sensor(struct smu_context *smu,
 	return ret;
 }
 
-static int arcturus_get_fan_speed_rpm(struct smu_context *smu,
-				      uint32_t *speed)
+static int arcturus_get_fan_speed_percent(struct smu_context *smu,
+					  uint32_t *speed)
 {
+	int ret;
+	u32 rpm;
+
 	if (!speed)
 		return -EINVAL;
 
-	return arcturus_get_smu_metrics_data(smu,
-					     METRICS_CURR_FANSPEED,
-					     speed);
+	switch (smu_v11_0_get_fan_control_mode(smu)) {
+	case AMD_FAN_CTRL_AUTO:
+		ret = arcturus_get_smu_metrics_data(smu,
+						    METRICS_CURR_FANSPEED,
+						    &rpm);
+		if (!ret && smu->fan_max_rpm)
+			*speed = rpm * 100 / smu->fan_max_rpm;
+		return ret;
+	default:
+		*speed = smu->user_dpm_profile.fan_speed_percent;
+		return 0;
+	}
 }
 
 static int arcturus_get_fan_parameters(struct smu_context *smu)
@@ -1310,7 +1322,7 @@ static int arcturus_set_power_profile_mode(struct smu_context *smu,
 						       CMN2ASIC_MAPPING_WORKLOAD,
 						       profile_mode);
 	if (workload_type < 0) {
-		dev_err(smu->adev->dev, "Unsupported power profile mode %d on arcturus\n", profile_mode);
+		dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on arcturus\n", profile_mode);
 		return -EINVAL;
 	}
 
@@ -2227,7 +2239,7 @@ static ssize_t arcturus_get_gpu_metrics(struct smu_context *smu,
 	if (ret)
 		return ret;
 
-	smu_v11_0_init_gpu_metrics_v1_0(gpu_metrics);
+	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 0);
 
 	gpu_metrics->temperature_edge = metrics.TemperatureEdge;
 	gpu_metrics->temperature_hotspot = metrics.TemperatureHotspot;
@@ -2264,6 +2276,8 @@ static ssize_t arcturus_get_gpu_metrics(struct smu_context *smu,
 	gpu_metrics->pcie_link_speed =
 			arcturus_get_current_pcie_link_speed(smu);
 
+	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
+
 	*table = (void *)gpu_metrics;
 
 	return sizeof(struct gpu_metrics_v1_0);
@@ -2281,7 +2295,7 @@ static const struct pptable_funcs arcturus_ppt_funcs = {
 	.print_clk_levels = arcturus_print_clk_levels,
 	.force_clk_levels = arcturus_force_clk_levels,
 	.read_sensor = arcturus_read_sensor,
-	.get_fan_speed_rpm = arcturus_get_fan_speed_rpm,
+	.get_fan_speed_percent = arcturus_get_fan_speed_percent,
 	.get_power_profile_mode = arcturus_get_power_profile_mode,
 	.set_power_profile_mode = arcturus_set_power_profile_mode,
 	.set_performance_level = arcturus_set_performance_level,
@@ -2326,7 +2340,7 @@ static const struct pptable_funcs arcturus_ppt_funcs = {
 	.display_clock_voltage_request = smu_v11_0_display_clock_voltage_request,
 	.get_fan_control_mode = smu_v11_0_get_fan_control_mode,
 	.set_fan_control_mode = smu_v11_0_set_fan_control_mode,
-	.set_fan_speed_rpm = smu_v11_0_set_fan_speed_rpm,
+	.set_fan_speed_percent = smu_v11_0_set_fan_speed_percent,
 	.set_xgmi_pstate = smu_v11_0_set_xgmi_pstate,
 	.gfx_off_control = smu_v11_0_gfx_off_control,
 	.register_irq_handler = smu_v11_0_register_irq_handler,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
index 51e83123f72a..6e641f1513d8 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
@@ -1317,15 +1317,27 @@ static bool navi10_is_dpm_running(struct smu_context *smu)
 	return !!(feature_enabled & SMC_DPM_FEATURE);
 }
 
-static int navi10_get_fan_speed_rpm(struct smu_context *smu,
-				    uint32_t *speed)
+static int navi10_get_fan_speed_percent(struct smu_context *smu,
+					uint32_t *speed)
 {
+	int ret;
+	u32 rpm;
+
 	if (!speed)
 		return -EINVAL;
 
-	return navi10_get_smu_metrics_data(smu,
-					   METRICS_CURR_FANSPEED,
-					   speed);
+	switch (smu_v11_0_get_fan_control_mode(smu)) {
+	case AMD_FAN_CTRL_AUTO:
+		ret = navi10_get_smu_metrics_data(smu,
+						  METRICS_CURR_FANSPEED,
+						  &rpm);
+		if (!ret && smu->fan_max_rpm)
+			*speed = rpm * 100 / smu->fan_max_rpm;
+		return ret;
+	default:
+		*speed = smu->user_dpm_profile.fan_speed_percent;
+		return 0;
+	}
 }
 
 static int navi10_get_fan_parameters(struct smu_context *smu)
@@ -1673,7 +1685,7 @@ static int navi10_read_sensor(struct smu_context *smu,
 		*size = 4;
 		break;
 	case AMDGPU_PP_SENSOR_GFX_SCLK:
-		ret = navi10_get_current_clk_freq_by_table(smu, SMU_GFXCLK, (uint32_t *)data);
+		ret = navi10_get_smu_metrics_data(smu, METRICS_AVERAGE_GFXCLK, (uint32_t *)data);
 		*(uint32_t *)data *= 100;
 		*size = 4;
 		break;
@@ -2302,7 +2314,7 @@ static ssize_t navi10_get_gpu_metrics(struct smu_context *smu,
 
 	mutex_unlock(&smu->metrics_lock);
 
-	smu_v11_0_init_gpu_metrics_v1_0(gpu_metrics);
+	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 0);
 
 	gpu_metrics->temperature_edge = metrics.TemperatureEdge;
 	gpu_metrics->temperature_hotspot = metrics.TemperatureHotspot;
@@ -2342,6 +2354,8 @@ static ssize_t navi10_get_gpu_metrics(struct smu_context *smu,
 	gpu_metrics->pcie_link_speed =
 			smu_v11_0_get_current_pcie_link_speed(smu);
 
+	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
+
 	*table = (void *)gpu_metrics;
 
 	return sizeof(struct gpu_metrics_v1_0);
@@ -2413,7 +2427,7 @@ static const struct pptable_funcs navi10_ppt_funcs = {
 	.display_config_changed = navi10_display_config_changed,
 	.notify_smc_display_config = navi10_notify_smc_display_config,
 	.is_dpm_running = navi10_is_dpm_running,
-	.get_fan_speed_rpm = navi10_get_fan_speed_rpm,
+	.get_fan_speed_percent = navi10_get_fan_speed_percent,
 	.get_power_profile_mode = navi10_get_power_profile_mode,
 	.set_power_profile_mode = navi10_set_power_profile_mode,
 	.set_watermarks_table = navi10_set_watermarks_table,
@@ -2456,7 +2470,7 @@ static const struct pptable_funcs navi10_ppt_funcs = {
 	.display_clock_voltage_request = smu_v11_0_display_clock_voltage_request,
 	.get_fan_control_mode = smu_v11_0_get_fan_control_mode,
 	.set_fan_control_mode = smu_v11_0_set_fan_control_mode,
-	.set_fan_speed_rpm = smu_v11_0_set_fan_speed_rpm,
+	.set_fan_speed_percent = smu_v11_0_set_fan_speed_percent,
 	.set_xgmi_pstate = smu_v11_0_set_xgmi_pstate,
 	.gfx_off_control = smu_v11_0_gfx_off_control,
 	.register_irq_handler = smu_v11_0_register_irq_handler,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
index 9608745d732f..af73e1430af5 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
@@ -261,6 +261,11 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_GFX_GPO_BIT);
 	}
 
+	if ((adev->pm.pp_feature & PP_GFX_DCS_MASK) &&
+	    (adev->asic_type > CHIP_SIENNA_CICHLID) &&
+	    !(adev->flags & AMD_IS_APU))
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_GFX_DCS_BIT);
+
 	if (adev->pm.pp_feature & PP_MCLK_DPM_MASK)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_UCLK_BIT)
 					| FEATURE_MASK(FEATURE_MEM_VDDCI_SCALING_BIT)
@@ -294,6 +299,12 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 	    smu->adev->pg_flags & AMD_PG_SUPPORT_JPEG)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_MM_DPM_PG_BIT);
 
+	if (smu->dc_controlled_by_gpio)
+       *(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_ACDC_BIT);
+
+	if (amdgpu_aspm == 1)
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DS_LCLK_BIT);
+
 	return 0;
 }
 
@@ -314,6 +325,12 @@ static int sienna_cichlid_check_powerplay_table(struct smu_context *smu)
 	table_context->thermal_controller_type =
 		powerplay_table->thermal_controller_type;
 
+	/*
+	 * Instead of having its own buffer space and get overdrive_table copied,
+	 * smu->od_settings just points to the actual overdrive_table
+	 */
+	smu->od_settings = &powerplay_table->overdrive_table;
+
 	return 0;
 }
 
@@ -907,6 +924,22 @@ static bool sienna_cichlid_is_support_fine_grained_dpm(struct smu_context *smu,
 	return dpm_desc->SnapToDiscrete == 0 ? true : false;
 }
 
+static bool sienna_cichlid_is_od_feature_supported(struct smu_11_0_7_overdrive_table *od_table,
+						   enum SMU_11_0_7_ODFEATURE_CAP cap)
+{
+	return od_table->cap[cap];
+}
+
+static void sienna_cichlid_get_od_setting_range(struct smu_11_0_7_overdrive_table *od_table,
+						enum SMU_11_0_7_ODSETTING_ID setting,
+						uint32_t *min, uint32_t *max)
+{
+	if (min)
+		*min = od_table->min[setting];
+	if (max)
+		*max = od_table->max[setting];
+}
+
 static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 			enum smu_clk_type clk_type, char *buf)
 {
@@ -915,11 +948,16 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 	struct smu_dpm_context *smu_dpm = &smu->smu_dpm;
 	struct smu_11_0_dpm_context *dpm_context = smu_dpm->dpm_context;
 	PPTable_t *pptable = (PPTable_t *)table_context->driver_pptable;
+	struct smu_11_0_7_overdrive_table *od_settings = smu->od_settings;
+	OverDriveTable_t *od_table =
+		(OverDriveTable_t *)table_context->overdrive_table;
 	int i, size = 0, ret = 0;
 	uint32_t cur_value = 0, value = 0, count = 0;
 	uint32_t freq_values[3] = {0};
 	uint32_t mark_index = 0;
 	uint32_t gen_speed, lane_width;
+	uint32_t min_value, max_value;
+	uint32_t smu_version;
 
 	switch (clk_type) {
 	case SMU_GFXCLK:
@@ -995,6 +1033,70 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 					(lane_width == dpm_context->dpm_tables.pcie_table.pcie_lane[i]) ?
 					"*" : "");
 		break;
+	case SMU_OD_SCLK:
+		if (!smu->od_enabled || !od_table || !od_settings)
+			break;
+
+		if (!sienna_cichlid_is_od_feature_supported(od_settings, SMU_11_0_7_ODCAP_GFXCLK_LIMITS))
+			break;
+
+		size += sprintf(buf + size, "OD_SCLK:\n");
+		size += sprintf(buf + size, "0: %uMhz\n1: %uMhz\n", od_table->GfxclkFmin, od_table->GfxclkFmax);
+		break;
+
+	case SMU_OD_MCLK:
+		if (!smu->od_enabled || !od_table || !od_settings)
+			break;
+
+		if (!sienna_cichlid_is_od_feature_supported(od_settings, SMU_11_0_7_ODCAP_UCLK_LIMITS))
+			break;
+
+		size += sprintf(buf + size, "OD_MCLK:\n");
+		size += sprintf(buf + size, "0: %uMhz\n1: %uMHz\n", od_table->UclkFmin, od_table->UclkFmax);
+		break;
+
+	case SMU_OD_VDDGFX_OFFSET:
+		if (!smu->od_enabled || !od_table || !od_settings)
+			break;
+
+		/*
+		 * OD GFX Voltage Offset functionality is supported only by 58.41.0
+		 * and onwards SMU firmwares.
+		 */
+		smu_cmn_get_smc_version(smu, NULL, &smu_version);
+		if ((adev->asic_type == CHIP_SIENNA_CICHLID) &&
+		     (smu_version < 0x003a2900))
+			break;
+
+		size += sprintf(buf + size, "OD_VDDGFX_OFFSET:\n");
+		size += sprintf(buf + size, "%dmV\n", od_table->VddGfxOffset);
+		break;
+
+	case SMU_OD_RANGE:
+		if (!smu->od_enabled || !od_table || !od_settings)
+			break;
+
+		size = sprintf(buf, "%s:\n", "OD_RANGE");
+
+		if (sienna_cichlid_is_od_feature_supported(od_settings, SMU_11_0_7_ODCAP_GFXCLK_LIMITS)) {
+			sienna_cichlid_get_od_setting_range(od_settings, SMU_11_0_7_ODSETTING_GFXCLKFMIN,
+							    &min_value, NULL);
+			sienna_cichlid_get_od_setting_range(od_settings, SMU_11_0_7_ODSETTING_GFXCLKFMAX,
+							    NULL, &max_value);
+			size += sprintf(buf + size, "SCLK: %7uMhz %10uMhz\n",
+					min_value, max_value);
+		}
+
+		if (sienna_cichlid_is_od_feature_supported(od_settings, SMU_11_0_7_ODCAP_UCLK_LIMITS)) {
+			sienna_cichlid_get_od_setting_range(od_settings, SMU_11_0_7_ODSETTING_UCLKFMIN,
+							    &min_value, NULL);
+			sienna_cichlid_get_od_setting_range(od_settings, SMU_11_0_7_ODSETTING_UCLKFMAX,
+							    NULL, &max_value);
+			size += sprintf(buf + size, "MCLK: %7uMhz %10uMhz\n",
+					min_value, max_value);
+		}
+		break;
+
 	default:
 		break;
 	}
@@ -1146,15 +1248,27 @@ static bool sienna_cichlid_is_dpm_running(struct smu_context *smu)
 	return !!(feature_enabled & SMC_DPM_FEATURE);
 }
 
-static int sienna_cichlid_get_fan_speed_rpm(struct smu_context *smu,
-				    uint32_t *speed)
+static int sienna_cichlid_get_fan_speed_percent(struct smu_context *smu,
+						uint32_t *speed)
 {
+	int ret;
+	u32 rpm;
+
 	if (!speed)
 		return -EINVAL;
 
-	return sienna_cichlid_get_smu_metrics_data(smu,
-						METRICS_CURR_FANSPEED,
-						speed);
+	switch (smu_v11_0_get_fan_control_mode(smu)) {
+	case AMD_FAN_CTRL_AUTO:
+		ret = sienna_cichlid_get_smu_metrics_data(smu,
+							  METRICS_CURR_FANSPEED,
+							  &rpm);
+		if (!ret && smu->fan_max_rpm)
+			*speed = rpm * 100 / smu->fan_max_rpm;
+		return ret;
+	default:
+		*speed = smu->user_dpm_profile.fan_speed_percent;
+		return 0;
+	}
 }
 
 static int sienna_cichlid_get_fan_parameters(struct smu_context *smu)
@@ -1694,6 +1808,243 @@ static int sienna_cichlid_get_dpm_ultimate_freq(struct smu_context *smu,
 	return ret;
 }
 
+static void sienna_cichlid_dump_od_table(struct smu_context *smu,
+					 OverDriveTable_t *od_table)
+{
+	struct amdgpu_device *adev = smu->adev;
+	uint32_t smu_version;
+
+	dev_dbg(smu->adev->dev, "OD: Gfxclk: (%d, %d)\n", od_table->GfxclkFmin,
+							  od_table->GfxclkFmax);
+	dev_dbg(smu->adev->dev, "OD: Uclk: (%d, %d)\n", od_table->UclkFmin,
+							od_table->UclkFmax);
+
+	smu_cmn_get_smc_version(smu, NULL, &smu_version);
+	if (!((adev->asic_type == CHIP_SIENNA_CICHLID) &&
+	       (smu_version < 0x003a2900)))
+		dev_dbg(smu->adev->dev, "OD: VddGfxOffset: %d\n", od_table->VddGfxOffset);
+}
+
+static int sienna_cichlid_set_default_od_settings(struct smu_context *smu)
+{
+	OverDriveTable_t *od_table =
+		(OverDriveTable_t *)smu->smu_table.overdrive_table;
+	OverDriveTable_t *boot_od_table =
+		(OverDriveTable_t *)smu->smu_table.boot_overdrive_table;
+	int ret = 0;
+
+	ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE,
+				   0, (void *)od_table, false);
+	if (ret) {
+		dev_err(smu->adev->dev, "Failed to get overdrive table!\n");
+		return ret;
+	}
+
+	memcpy(boot_od_table, od_table, sizeof(OverDriveTable_t));
+
+	sienna_cichlid_dump_od_table(smu, od_table);
+
+	return 0;
+}
+
+static int sienna_cichlid_od_setting_check_range(struct smu_context *smu,
+						 struct smu_11_0_7_overdrive_table *od_table,
+						 enum SMU_11_0_7_ODSETTING_ID setting,
+						 uint32_t value)
+{
+	if (value < od_table->min[setting]) {
+		dev_warn(smu->adev->dev, "OD setting (%d, %d) is less than the minimum allowed (%d)\n",
+					  setting, value, od_table->min[setting]);
+		return -EINVAL;
+	}
+	if (value > od_table->max[setting]) {
+		dev_warn(smu->adev->dev, "OD setting (%d, %d) is greater than the maximum allowed (%d)\n",
+					  setting, value, od_table->max[setting]);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int sienna_cichlid_od_edit_dpm_table(struct smu_context *smu,
+					    enum PP_OD_DPM_TABLE_COMMAND type,
+					    long input[], uint32_t size)
+{
+	struct smu_table_context *table_context = &smu->smu_table;
+	OverDriveTable_t *od_table =
+		(OverDriveTable_t *)table_context->overdrive_table;
+	struct smu_11_0_7_overdrive_table *od_settings =
+		(struct smu_11_0_7_overdrive_table *)smu->od_settings;
+	struct amdgpu_device *adev = smu->adev;
+	enum SMU_11_0_7_ODSETTING_ID freq_setting;
+	uint16_t *freq_ptr;
+	int i, ret = 0;
+	uint32_t smu_version;
+
+	if (!smu->od_enabled) {
+		dev_warn(smu->adev->dev, "OverDrive is not enabled!\n");
+		return -EINVAL;
+	}
+
+	if (!smu->od_settings) {
+		dev_err(smu->adev->dev, "OD board limits are not set!\n");
+		return -ENOENT;
+	}
+
+	if (!(table_context->overdrive_table && table_context->boot_overdrive_table)) {
+		dev_err(smu->adev->dev, "Overdrive table was not initialized!\n");
+		return -EINVAL;
+	}
+
+	switch (type) {
+	case PP_OD_EDIT_SCLK_VDDC_TABLE:
+		if (!sienna_cichlid_is_od_feature_supported(od_settings,
+							    SMU_11_0_7_ODCAP_GFXCLK_LIMITS)) {
+			dev_warn(smu->adev->dev, "GFXCLK_LIMITS not supported!\n");
+			return -ENOTSUPP;
+		}
+
+		for (i = 0; i < size; i += 2) {
+			if (i + 2 > size) {
+				dev_info(smu->adev->dev, "invalid number of input parameters %d\n", size);
+				return -EINVAL;
+			}
+
+			switch (input[i]) {
+			case 0:
+				if (input[i + 1] > od_table->GfxclkFmax) {
+					dev_info(smu->adev->dev, "GfxclkFmin (%ld) must be <= GfxclkFmax (%u)!\n",
+						input[i + 1], od_table->GfxclkFmax);
+					return -EINVAL;
+				}
+
+				freq_setting = SMU_11_0_7_ODSETTING_GFXCLKFMIN;
+				freq_ptr = &od_table->GfxclkFmin;
+				break;
+
+			case 1:
+				if (input[i + 1] < od_table->GfxclkFmin) {
+					dev_info(smu->adev->dev, "GfxclkFmax (%ld) must be >= GfxclkFmin (%u)!\n",
+						input[i + 1], od_table->GfxclkFmin);
+					return -EINVAL;
+				}
+
+				freq_setting = SMU_11_0_7_ODSETTING_GFXCLKFMAX;
+				freq_ptr = &od_table->GfxclkFmax;
+				break;
+
+			default:
+				dev_info(smu->adev->dev, "Invalid SCLK_VDDC_TABLE index: %ld\n", input[i]);
+				dev_info(smu->adev->dev, "Supported indices: [0:min,1:max]\n");
+				return -EINVAL;
+			}
+
+			ret = sienna_cichlid_od_setting_check_range(smu, od_settings,
+								    freq_setting, input[i + 1]);
+			if (ret)
+				return ret;
+
+			*freq_ptr = (uint16_t)input[i + 1];
+		}
+		break;
+
+	case PP_OD_EDIT_MCLK_VDDC_TABLE:
+		if (!sienna_cichlid_is_od_feature_supported(od_settings, SMU_11_0_7_ODCAP_UCLK_LIMITS)) {
+			dev_warn(smu->adev->dev, "UCLK_LIMITS not supported!\n");
+			return -ENOTSUPP;
+		}
+
+		for (i = 0; i < size; i += 2) {
+			if (i + 2 > size) {
+				dev_info(smu->adev->dev, "invalid number of input parameters %d\n", size);
+				return -EINVAL;
+			}
+
+			switch (input[i]) {
+			case 0:
+				if (input[i + 1] > od_table->UclkFmax) {
+					dev_info(smu->adev->dev, "UclkFmin (%ld) must be <= UclkFmax (%u)!\n",
+						input[i + 1], od_table->UclkFmax);
+					return -EINVAL;
+				}
+
+				freq_setting = SMU_11_0_7_ODSETTING_UCLKFMIN;
+				freq_ptr = &od_table->UclkFmin;
+				break;
+
+			case 1:
+				if (input[i + 1] < od_table->UclkFmin) {
+					dev_info(smu->adev->dev, "UclkFmax (%ld) must be >= UclkFmin (%u)!\n",
+						input[i + 1], od_table->UclkFmin);
+					return -EINVAL;
+				}
+
+				freq_setting = SMU_11_0_7_ODSETTING_UCLKFMAX;
+				freq_ptr = &od_table->UclkFmax;
+				break;
+
+			default:
+				dev_info(smu->adev->dev, "Invalid MCLK_VDDC_TABLE index: %ld\n", input[i]);
+				dev_info(smu->adev->dev, "Supported indices: [0:min,1:max]\n");
+				return -EINVAL;
+			}
+
+			ret = sienna_cichlid_od_setting_check_range(smu, od_settings,
+								    freq_setting, input[i + 1]);
+			if (ret)
+				return ret;
+
+			*freq_ptr = (uint16_t)input[i + 1];
+		}
+		break;
+
+	case PP_OD_RESTORE_DEFAULT_TABLE:
+		memcpy(table_context->overdrive_table,
+				table_context->boot_overdrive_table,
+				sizeof(OverDriveTable_t));
+		fallthrough;
+
+	case PP_OD_COMMIT_DPM_TABLE:
+		sienna_cichlid_dump_od_table(smu, od_table);
+
+		ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE,
+					   0, (void *)od_table, true);
+		if (ret) {
+			dev_err(smu->adev->dev, "Failed to import overdrive table!\n");
+			return ret;
+		}
+		break;
+
+	case PP_OD_EDIT_VDDGFX_OFFSET:
+		if (size != 1) {
+			dev_info(smu->adev->dev, "invalid number of parameters: %d\n", size);
+			return -EINVAL;
+		}
+
+		/*
+		 * OD GFX Voltage Offset functionality is supported only by 58.41.0
+		 * and onwards SMU firmwares.
+		 */
+		smu_cmn_get_smc_version(smu, NULL, &smu_version);
+		if ((adev->asic_type == CHIP_SIENNA_CICHLID) &&
+		     (smu_version < 0x003a2900)) {
+			dev_err(smu->adev->dev, "OD GFX Voltage offset functionality is supported "
+						"only by 58.41.0 and onwards SMU firmwares!\n");
+			return -EOPNOTSUPP;
+		}
+
+		od_table->VddGfxOffset = (int16_t)input[0];
+
+		sienna_cichlid_dump_od_table(smu, od_table);
+		break;
+
+	default:
+		return -ENOSYS;
+	}
+
+	return ret;
+}
+
 static int sienna_cichlid_run_btc(struct smu_context *smu)
 {
 	return smu_cmn_send_smc_msg(smu, SMU_MSG_RunDcBtc, NULL);
@@ -2372,7 +2723,7 @@ static void sienna_cichlid_fill_i2c_req(SwI2cRequest_t  *req, bool write,
 {
 	int i;
 
-	req->I2CcontrollerPort = 0;
+	req->I2CcontrollerPort = 1;
 	req->I2CSpeed = 2;
 	req->SlaveAddress = address;
 	req->NumCmds = numbytes;
@@ -2610,7 +2961,7 @@ static ssize_t sienna_cichlid_get_gpu_metrics(struct smu_context *smu,
 	if (ret)
 		return ret;
 
-	smu_v11_0_init_gpu_metrics_v1_0(gpu_metrics);
+	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 0);
 
 	gpu_metrics->temperature_edge = metrics->TemperatureEdge;
 	gpu_metrics->temperature_hotspot = metrics->TemperatureHotspot;
@@ -2653,6 +3004,8 @@ static ssize_t sienna_cichlid_get_gpu_metrics(struct smu_context *smu,
 	gpu_metrics->pcie_link_speed =
 			smu_v11_0_get_current_pcie_link_speed(smu);
 
+	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
+
 	*table = (void *)gpu_metrics;
 
 	return sizeof(struct gpu_metrics_v1_0);
@@ -2759,7 +3112,7 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.display_config_changed = sienna_cichlid_display_config_changed,
 	.notify_smc_display_config = sienna_cichlid_notify_smc_display_config,
 	.is_dpm_running = sienna_cichlid_is_dpm_running,
-	.get_fan_speed_rpm = sienna_cichlid_get_fan_speed_rpm,
+	.get_fan_speed_percent = sienna_cichlid_get_fan_speed_percent,
 	.get_power_profile_mode = sienna_cichlid_get_power_profile_mode,
 	.set_power_profile_mode = sienna_cichlid_set_power_profile_mode,
 	.set_watermarks_table = sienna_cichlid_set_watermarks_table,
@@ -2802,7 +3155,7 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.display_clock_voltage_request = smu_v11_0_display_clock_voltage_request,
 	.get_fan_control_mode = smu_v11_0_get_fan_control_mode,
 	.set_fan_control_mode = smu_v11_0_set_fan_control_mode,
-	.set_fan_speed_rpm = smu_v11_0_set_fan_speed_rpm,
+	.set_fan_speed_percent = smu_v11_0_set_fan_speed_percent,
 	.set_xgmi_pstate = smu_v11_0_set_xgmi_pstate,
 	.gfx_off_control = smu_v11_0_gfx_off_control,
 	.register_irq_handler = smu_v11_0_register_irq_handler,
@@ -2817,6 +3170,8 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.mode1_reset = smu_v11_0_mode1_reset,
 	.get_dpm_ultimate_freq = sienna_cichlid_get_dpm_ultimate_freq,
 	.set_soft_freq_limited_range = smu_v11_0_set_soft_freq_limited_range,
+	.set_default_od_settings = sienna_cichlid_set_default_od_settings,
+	.od_edit_dpm_table = sienna_cichlid_od_edit_dpm_table,
 	.run_btc = sienna_cichlid_run_btc,
 	.set_power_source = smu_v11_0_set_power_source,
 	.get_pp_feature_mask = smu_cmn_get_pp_feature_mask,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
index b279dbbbce6b..a6211858ead4 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
@@ -78,6 +78,9 @@ MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_smc.bin");
 #define PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK 0xC000
 #define PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE__SHIFT 0xE
 
+#define mmTHM_BACO_CNTL_ARCT			0xA7
+#define mmTHM_BACO_CNTL_ARCT_BASE_IDX		0
+
 static int link_width[] = {0, 1, 2, 4, 8, 12, 16};
 static int link_speed[] = {25, 50, 80, 160};
 
@@ -474,12 +477,14 @@ int smu_v11_0_fini_smc_tables(struct smu_context *smu)
 int smu_v11_0_init_power(struct smu_context *smu)
 {
 	struct smu_power_context *smu_power = &smu->smu_power;
+	size_t size = smu->adev->asic_type == CHIP_VANGOGH ?
+			sizeof(struct smu_11_5_power_context) :
+			sizeof(struct smu_11_0_power_context);
 
-	smu_power->power_context = kzalloc(sizeof(struct smu_11_0_power_context),
-					   GFP_KERNEL);
+	smu_power->power_context = kzalloc(size, GFP_KERNEL);
 	if (!smu_power->power_context)
 		return -ENOMEM;
-	smu_power->power_context_size = sizeof(struct smu_11_0_power_context);
+	smu_power->power_context_size = size;
 
 	return 0;
 }
@@ -1119,6 +1124,7 @@ int smu_v11_0_gfx_off_control(struct smu_context *smu, bool enable)
 	case CHIP_SIENNA_CICHLID:
 	case CHIP_NAVY_FLOUNDER:
 	case CHIP_DIMGREY_CAVEFISH:
+	case CHIP_VANGOGH:
 		if (!(adev->pm.pp_feature & PP_GFXOFF_MASK))
 			return 0;
 		if (enable)
@@ -1136,10 +1142,10 @@ int smu_v11_0_gfx_off_control(struct smu_context *smu, bool enable)
 uint32_t
 smu_v11_0_get_fan_control_mode(struct smu_context *smu)
 {
-	if (!smu_cmn_feature_is_enabled(smu, SMU_FEATURE_FAN_CONTROL_BIT))
-		return AMD_FAN_CTRL_MANUAL;
-	else
+	if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_FAN_CONTROL_BIT))
 		return AMD_FAN_CTRL_AUTO;
+	else
+		return smu->user_dpm_profile.fan_mode;
 }
 
 static int
@@ -1174,6 +1180,35 @@ smu_v11_0_set_fan_static_mode(struct smu_context *smu, uint32_t mode)
 }
 
 int
+smu_v11_0_set_fan_speed_percent(struct smu_context *smu, uint32_t speed)
+{
+	struct amdgpu_device *adev = smu->adev;
+	uint32_t duty100, duty;
+	uint64_t tmp64;
+
+	if (speed > 100)
+		speed = 100;
+
+	if (smu_v11_0_auto_fan_control(smu, 0))
+		return -EINVAL;
+
+	duty100 = REG_GET_FIELD(RREG32_SOC15(THM, 0, mmCG_FDO_CTRL1),
+				CG_FDO_CTRL1, FMAX_DUTY100);
+	if (!duty100)
+		return -EINVAL;
+
+	tmp64 = (uint64_t)speed * duty100;
+	do_div(tmp64, 100);
+	duty = (uint32_t)tmp64;
+
+	WREG32_SOC15(THM, 0, mmCG_FDO_CTRL0,
+		     REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_FDO_CTRL0),
+				   CG_FDO_CTRL0, FDO_STATIC_DUTY, duty));
+
+	return smu_v11_0_set_fan_static_mode(smu, FDO_PWM_MODE_STATIC);
+}
+
+int
 smu_v11_0_set_fan_control_mode(struct smu_context *smu,
 			       uint32_t mode)
 {
@@ -1181,7 +1216,7 @@ smu_v11_0_set_fan_control_mode(struct smu_context *smu,
 
 	switch (mode) {
 	case AMD_FAN_CTRL_NONE:
-		ret = smu_v11_0_set_fan_speed_rpm(smu, smu->fan_max_rpm);
+		ret = smu_v11_0_set_fan_speed_percent(smu, 100);
 		break;
 	case AMD_FAN_CTRL_MANUAL:
 		ret = smu_v11_0_auto_fan_control(smu, 0);
@@ -1201,58 +1236,6 @@ smu_v11_0_set_fan_control_mode(struct smu_context *smu,
 	return ret;
 }
 
-int smu_v11_0_set_fan_speed_rpm(struct smu_context *smu,
-				       uint32_t speed)
-{
-	struct amdgpu_device *adev = smu->adev;
-	int ret;
-	uint32_t tach_period, crystal_clock_freq;
-
-	if (!speed)
-		return -EINVAL;
-
-	ret = smu_v11_0_auto_fan_control(smu, 0);
-	if (ret)
-		return ret;
-
-	/*
-	 * crystal_clock_freq div by 4 is required since the fan control
-	 * module refers to 25MHz
-	 */
-
-	crystal_clock_freq = amdgpu_asic_get_xclk(adev) / 4;
-	tach_period = 60 * crystal_clock_freq * 10000 / (8 * speed);
-	WREG32_SOC15(THM, 0, mmCG_TACH_CTRL,
-		     REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_CTRL),
-				   CG_TACH_CTRL, TARGET_PERIOD,
-				   tach_period));
-
-	ret = smu_v11_0_set_fan_static_mode(smu, FDO_PWM_MODE_STATIC_RPM);
-
-	return ret;
-}
-
-int smu_v11_0_get_fan_speed_rpm(struct smu_context *smu,
-				uint32_t *speed)
-{
-	struct amdgpu_device *adev = smu->adev;
-	uint32_t tach_period, crystal_clock_freq;
-	uint64_t tmp64;
-
-	tach_period = REG_GET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_CTRL),
-				    CG_TACH_CTRL, TARGET_PERIOD);
-	if (!tach_period)
-		return -EINVAL;
-
-	crystal_clock_freq = amdgpu_asic_get_xclk(adev);
-
-	tmp64 = (uint64_t)crystal_clock_freq * 60 * 10000;
-	do_div(tmp64, (tach_period * 8));
-	*speed = (uint32_t)tmp64;
-
-	return 0;
-}
-
 int smu_v11_0_set_xgmi_pstate(struct smu_context *smu,
 				     uint32_t pstate)
 {
@@ -1552,9 +1535,15 @@ int smu_v11_0_baco_set_state(struct smu_context *smu, enum smu_baco_state state)
 			break;
 		default:
 			if (!ras || !ras->supported) {
-				data = RREG32_SOC15(THM, 0, mmTHM_BACO_CNTL);
-				data |= 0x80000000;
-				WREG32_SOC15(THM, 0, mmTHM_BACO_CNTL, data);
+				if (adev->asic_type == CHIP_ARCTURUS) {
+					data = RREG32_SOC15(THM, 0, mmTHM_BACO_CNTL_ARCT);
+					data |= 0x80000000;
+					WREG32_SOC15(THM, 0, mmTHM_BACO_CNTL_ARCT, data);
+				} else {
+					data = RREG32_SOC15(THM, 0, mmTHM_BACO_CNTL);
+					data |= 0x80000000;
+					WREG32_SOC15(THM, 0, mmTHM_BACO_CNTL, data);
+				}
 
 				ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_EnterBaco, 0, NULL);
 			} else {
@@ -2043,30 +2032,6 @@ int smu_v11_0_get_current_pcie_link_speed(struct smu_context *smu)
 	return link_speed[speed_level];
 }
 
-void smu_v11_0_init_gpu_metrics_v1_0(struct gpu_metrics_v1_0 *gpu_metrics)
-{
-	memset(gpu_metrics, 0xFF, sizeof(struct gpu_metrics_v1_0));
-
-	gpu_metrics->common_header.structure_size =
-				sizeof(struct gpu_metrics_v1_0);
-	gpu_metrics->common_header.format_revision = 1;
-	gpu_metrics->common_header.content_revision = 0;
-
-	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
-}
-
-void smu_v11_0_init_gpu_metrics_v2_0(struct gpu_metrics_v2_0 *gpu_metrics)
-{
-	memset(gpu_metrics, 0xFF, sizeof(struct gpu_metrics_v2_0));
-
-	gpu_metrics->common_header.structure_size =
-				sizeof(struct gpu_metrics_v2_0);
-	gpu_metrics->common_header.format_revision = 2;
-	gpu_metrics->common_header.content_revision = 0;
-
-	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
-}
-
 int smu_v11_0_gfx_ulv_control(struct smu_context *smu,
 			      bool enablement)
 {
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c
index 8cb4fcee9a2c..7ddbaecb11c2 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c
@@ -31,6 +31,10 @@
 #include "smu_v11_5_ppsmc.h"
 #include "smu_v11_5_pmfw.h"
 #include "smu_cmn.h"
+#include "soc15_common.h"
+#include "asic_reg/gc/gc_10_3_0_offset.h"
+#include "asic_reg/gc/gc_10_3_0_sh_mask.h"
+#include <asm/processor.h>
 
 /*
  * DO NOT use these for err/warn/info/debug messages.
@@ -59,7 +63,8 @@ static struct cmn2asic_msg_mapping vangogh_message_map[SMU_MSG_MAX_COUNT] = {
 	MSG_MAP(GetSmuVersion,                  PPSMC_MSG_GetSmuVersion,		0),
 	MSG_MAP(GetDriverIfVersion,             PPSMC_MSG_GetDriverIfVersion,	0),
 	MSG_MAP(EnableGfxOff,                   PPSMC_MSG_EnableGfxOff,			0),
-	MSG_MAP(DisallowGfxOff,                 PPSMC_MSG_DisableGfxOff,		0),
+	MSG_MAP(AllowGfxOff,                    PPSMC_MSG_AllowGfxOff,          0),
+	MSG_MAP(DisallowGfxOff,                 PPSMC_MSG_DisallowGfxOff,		0),
 	MSG_MAP(PowerDownIspByTile,             PPSMC_MSG_PowerDownIspByTile,	0),
 	MSG_MAP(PowerUpIspByTile,               PPSMC_MSG_PowerUpIspByTile,		0),
 	MSG_MAP(PowerDownVcn,                   PPSMC_MSG_PowerDownVcn,			0),
@@ -76,7 +81,6 @@ static struct cmn2asic_msg_mapping vangogh_message_map[SMU_MSG_MAX_COUNT] = {
 	MSG_MAP(TransferTableDram2Smu,          PPSMC_MSG_TransferTableDram2Smu,	0),
 	MSG_MAP(GfxDeviceDriverReset,           PPSMC_MSG_GfxDeviceDriverReset,		0),
 	MSG_MAP(GetEnabledSmuFeatures,          PPSMC_MSG_GetEnabledSmuFeatures,	0),
-	MSG_MAP(Spare1,                         PPSMC_MSG_spare1,					0),
 	MSG_MAP(SetHardMinSocclkByFreq,         PPSMC_MSG_SetHardMinSocclkByFreq,	0),
 	MSG_MAP(SetSoftMinFclk,                 PPSMC_MSG_SetSoftMinFclk,		0),
 	MSG_MAP(SetSoftMinVcn,                  PPSMC_MSG_SetSoftMinVcn,		0),
@@ -88,7 +92,6 @@ static struct cmn2asic_msg_mapping vangogh_message_map[SMU_MSG_MAX_COUNT] = {
 	MSG_MAP(SetSoftMaxSocclkByFreq,         PPSMC_MSG_SetSoftMaxSocclkByFreq,	0),
 	MSG_MAP(SetSoftMaxFclkByFreq,           PPSMC_MSG_SetSoftMaxFclkByFreq,		0),
 	MSG_MAP(SetSoftMaxVcn,                  PPSMC_MSG_SetSoftMaxVcn,			0),
-	MSG_MAP(Spare2,                         PPSMC_MSG_spare2,					0),
 	MSG_MAP(SetPowerLimitPercentage,        PPSMC_MSG_SetPowerLimitPercentage,	0),
 	MSG_MAP(PowerDownJpeg,                  PPSMC_MSG_PowerDownJpeg,			0),
 	MSG_MAP(PowerUpJpeg,                    PPSMC_MSG_PowerUpJpeg,				0),
@@ -118,6 +121,11 @@ static struct cmn2asic_msg_mapping vangogh_message_map[SMU_MSG_MAX_COUNT] = {
 	MSG_MAP(StopDramLogging,                    PPSMC_MSG_StopDramLogging,						0),
 	MSG_MAP(SetSoftMinCclk,                     PPSMC_MSG_SetSoftMinCclk,						0),
 	MSG_MAP(SetSoftMaxCclk,                     PPSMC_MSG_SetSoftMaxCclk,						0),
+	MSG_MAP(RequestActiveWgp,                   PPSMC_MSG_RequestActiveWgp,                     0),
+	MSG_MAP(SetFastPPTLimit,                    PPSMC_MSG_SetFastPPTLimit,						0),
+	MSG_MAP(SetSlowPPTLimit,                    PPSMC_MSG_SetSlowPPTLimit,						0),
+	MSG_MAP(GetFastPPTLimit,                    PPSMC_MSG_GetFastPPTLimit,						0),
+	MSG_MAP(GetSlowPPTLimit,                    PPSMC_MSG_GetSlowPPTLimit,						0),
 };
 
 static struct cmn2asic_mapping vangogh_feature_mask_map[SMU_FEATURE_COUNT] = {
@@ -162,6 +170,9 @@ static struct cmn2asic_mapping vangogh_feature_mask_map[SMU_FEATURE_COUNT] = {
 	FEA_MAP(A55_DPM),
 	FEA_MAP(CVIP_DSP_DPM),
 	FEA_MAP(MSMU_LOW_POWER),
+	FEA_MAP_REVERSE(SOCCLK),
+	FEA_MAP_REVERSE(FCLK),
+	FEA_MAP_HALF_REVERSE(GFX),
 };
 
 static struct cmn2asic_mapping vangogh_table_map[SMU_TABLE_COUNT] = {
@@ -171,6 +182,14 @@ static struct cmn2asic_mapping vangogh_table_map[SMU_TABLE_COUNT] = {
 	TAB_MAP_VALID(DPMCLOCKS),
 };
 
+static struct cmn2asic_mapping vangogh_workload_map[PP_SMC_POWER_PROFILE_COUNT] = {
+	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_FULLSCREEN3D,		WORKLOAD_PPLIB_FULL_SCREEN_3D_BIT),
+	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_VIDEO,		WORKLOAD_PPLIB_VIDEO_BIT),
+	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_VR,			WORKLOAD_PPLIB_VR_BIT),
+	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_COMPUTE,		WORKLOAD_PPLIB_COMPUTE_BIT),
+	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_CUSTOM,		WORKLOAD_PPLIB_CUSTOM_BIT),
+};
+
 static int vangogh_tables_init(struct smu_context *smu)
 {
 	struct smu_table_context *smu_table = &smu->smu_table;
@@ -242,6 +261,12 @@ static int vangogh_get_smu_metrics_data(struct smu_context *smu,
 	case METRICS_AVERAGE_SOCCLK:
 		*value = metrics->SocclkFrequency;
 		break;
+	case METRICS_AVERAGE_VCLK:
+		*value = metrics->VclkFrequency;
+		break;
+	case METRICS_AVERAGE_DCLK:
+		*value = metrics->DclkFrequency;
+		break;
 	case METRICS_AVERAGE_UCLK:
 		*value = metrics->MemclkFrequency;
 		break;
@@ -252,7 +277,8 @@ static int vangogh_get_smu_metrics_data(struct smu_context *smu,
 		*value = metrics->UvdActivity;
 		break;
 	case METRICS_AVERAGE_SOCKETPOWER:
-		*value = metrics->CurrentSocketPower;
+		*value = (metrics->CurrentSocketPower << 8) /
+		1000 ;
 		break;
 	case METRICS_TEMPERATURE_EDGE:
 		*value = metrics->GfxTemperature / 100 *
@@ -271,6 +297,10 @@ static int vangogh_get_smu_metrics_data(struct smu_context *smu,
 	case METRICS_VOLTAGE_VDDSOC:
 		*value = metrics->Voltage[1];
 		break;
+	case METRICS_AVERAGE_CPUCLK:
+		memcpy(value, &metrics->CoreFrequency[0],
+		       smu->cpu_core_num * sizeof(uint16_t));
+		break;
 	default:
 		*value = UINT_MAX;
 		break;
@@ -307,6 +337,13 @@ static int vangogh_init_smc_tables(struct smu_context *smu)
 	if (ret)
 		return ret;
 
+#ifdef CONFIG_X86
+	/* AMD x86 APU only */
+	smu->cpu_core_num = boot_cpu_data.x86_max_cores;
+#else
+	smu->cpu_core_num = 4;
+#endif
+
 	return smu_v11_0_init_smc_tables(smu);
 }
 
@@ -316,17 +353,13 @@ static int vangogh_dpm_set_vcn_enable(struct smu_context *smu, bool enable)
 
 	if (enable) {
 		/* vcn dpm on is a prerequisite for vcn power gate messages */
-		if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_VCN_PG_BIT)) {
-			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerUpVcn, 0, NULL);
-			if (ret)
-				return ret;
-		}
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerUpVcn, 0, NULL);
+		if (ret)
+			return ret;
 	} else {
-		if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_VCN_PG_BIT)) {
-			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerDownVcn, 0, NULL);
-			if (ret)
-				return ret;
-		}
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerDownVcn, 0, NULL);
+		if (ret)
+			return ret;
 	}
 
 	return ret;
@@ -337,54 +370,18 @@ static int vangogh_dpm_set_jpeg_enable(struct smu_context *smu, bool enable)
 	int ret = 0;
 
 	if (enable) {
-		if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_JPEG_PG_BIT)) {
-			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerUpJpeg, 0, NULL);
-			if (ret)
-				return ret;
-		}
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerUpJpeg, 0, NULL);
+		if (ret)
+			return ret;
 	} else {
-		if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_JPEG_PG_BIT)) {
-			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerDownJpeg, 0, NULL);
-			if (ret)
-				return ret;
-		}
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_PowerDownJpeg, 0, NULL);
+		if (ret)
+			return ret;
 	}
 
 	return ret;
 }
 
-static int vangogh_get_allowed_feature_mask(struct smu_context *smu,
-					    uint32_t *feature_mask,
-					    uint32_t num)
-{
-	struct amdgpu_device *adev = smu->adev;
-
-	if (num > 2)
-		return -EINVAL;
-
-	memset(feature_mask, 0, sizeof(uint32_t) * num);
-
-	*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_GFX_DPM_BIT)
-				| FEATURE_MASK(FEATURE_MP0CLK_DPM_BIT)
-				| FEATURE_MASK(FEATURE_DS_SOCCLK_BIT)
-				| FEATURE_MASK(FEATURE_PPT_BIT)
-				| FEATURE_MASK(FEATURE_TDC_BIT)
-				| FEATURE_MASK(FEATURE_FAN_CONTROLLER_BIT)
-				| FEATURE_MASK(FEATURE_DS_LCLK_BIT)
-				| FEATURE_MASK(FEATURE_DS_DCFCLK_BIT);
-
-	if (adev->pm.pp_feature & PP_SOCCLK_DPM_MASK)
-		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_SOCCLK_DPM_BIT);
-
-	if (adev->pm.pp_feature & PP_DCEFCLK_DPM_MASK)
-		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DCFCLK_DPM_BIT);
-
-	if (smu->adev->pg_flags & AMD_PG_SUPPORT_ATHUB)
-		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_ATHUB_PG_BIT);
-
-	return 0;
-}
-
 static bool vangogh_is_dpm_running(struct smu_context *smu)
 {
 	int ret = 0;
@@ -402,14 +399,68 @@ static bool vangogh_is_dpm_running(struct smu_context *smu)
 	return !!(feature_enabled & SMC_DPM_FEATURE);
 }
 
+static int vangogh_get_dpm_clk_limited(struct smu_context *smu, enum smu_clk_type clk_type,
+						uint32_t dpm_level, uint32_t *freq)
+{
+	DpmClocks_t *clk_table = smu->smu_table.clocks_table;
+
+	if (!clk_table || clk_type >= SMU_CLK_COUNT)
+		return -EINVAL;
+
+	switch (clk_type) {
+	case SMU_SOCCLK:
+		if (dpm_level >= clk_table->NumSocClkLevelsEnabled)
+			return -EINVAL;
+		*freq = clk_table->SocClocks[dpm_level];
+		break;
+	case SMU_VCLK:
+		if (dpm_level >= clk_table->VcnClkLevelsEnabled)
+			return -EINVAL;
+		*freq = clk_table->VcnClocks[dpm_level].vclk;
+		break;
+	case SMU_DCLK:
+		if (dpm_level >= clk_table->VcnClkLevelsEnabled)
+			return -EINVAL;
+		*freq = clk_table->VcnClocks[dpm_level].dclk;
+		break;
+	case SMU_UCLK:
+	case SMU_MCLK:
+		if (dpm_level >= clk_table->NumDfPstatesEnabled)
+			return -EINVAL;
+		*freq = clk_table->DfPstateTable[dpm_level].memclk;
+
+		break;
+	case SMU_FCLK:
+		if (dpm_level >= clk_table->NumDfPstatesEnabled)
+			return -EINVAL;
+		*freq = clk_table->DfPstateTable[dpm_level].fclk;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int vangogh_print_fine_grain_clk(struct smu_context *smu,
 			enum smu_clk_type clk_type, char *buf)
 {
-	int size = 0;
+	DpmClocks_t *clk_table = smu->smu_table.clocks_table;
+	SmuMetrics_t metrics;
+	struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
+	int i, size = 0, ret = 0;
+	uint32_t cur_value = 0, value = 0, count = 0;
+	bool cur_value_match_level = false;
+
+	memset(&metrics, 0, sizeof(metrics));
+
+	ret = smu_cmn_get_metrics_table(smu, &metrics, false);
+	if (ret)
+		return ret;
 
 	switch (clk_type) {
 	case SMU_OD_SCLK:
-		if (smu->od_enabled) {
+		if (smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) {
 			size = sprintf(buf, "%s:\n", "OD_SCLK");
 			size += sprintf(buf + size, "0: %10uMhz\n",
 			(smu->gfx_actual_hard_min_freq > 0) ? smu->gfx_actual_hard_min_freq : smu->gfx_default_hard_min_freq);
@@ -417,12 +468,71 @@ static int vangogh_print_fine_grain_clk(struct smu_context *smu,
 			(smu->gfx_actual_soft_max_freq > 0) ? smu->gfx_actual_soft_max_freq : smu->gfx_default_soft_max_freq);
 		}
 		break;
+	case SMU_OD_CCLK:
+		if (smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) {
+			size = sprintf(buf, "CCLK_RANGE in Core%d:\n",  smu->cpu_core_id_select);
+			size += sprintf(buf + size, "0: %10uMhz\n",
+			(smu->cpu_actual_soft_min_freq > 0) ? smu->cpu_actual_soft_min_freq : smu->cpu_default_soft_min_freq);
+			size += sprintf(buf + size, "1: %10uMhz\n",
+			(smu->cpu_actual_soft_max_freq > 0) ? smu->cpu_actual_soft_max_freq : smu->cpu_default_soft_max_freq);
+		}
+		break;
 	case SMU_OD_RANGE:
-		if (smu->od_enabled) {
+		if (smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) {
 			size = sprintf(buf, "%s:\n", "OD_RANGE");
 			size += sprintf(buf + size, "SCLK: %7uMhz %10uMhz\n",
 				smu->gfx_default_hard_min_freq, smu->gfx_default_soft_max_freq);
+			size += sprintf(buf + size, "CCLK: %7uMhz %10uMhz\n",
+				smu->cpu_default_soft_min_freq, smu->cpu_default_soft_max_freq);
+		}
+		break;
+	case SMU_SOCCLK:
+		/* the level 3 ~ 6 of socclk use the same frequency for vangogh */
+		count = clk_table->NumSocClkLevelsEnabled;
+		cur_value = metrics.SocclkFrequency;
+		break;
+	case SMU_VCLK:
+		count = clk_table->VcnClkLevelsEnabled;
+		cur_value = metrics.VclkFrequency;
+		break;
+	case SMU_DCLK:
+		count = clk_table->VcnClkLevelsEnabled;
+		cur_value = metrics.DclkFrequency;
+		break;
+	case SMU_MCLK:
+		count = clk_table->NumDfPstatesEnabled;
+		cur_value = metrics.MemclkFrequency;
+		break;
+	case SMU_FCLK:
+		count = clk_table->NumDfPstatesEnabled;
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_GetFclkFrequency, 0, &cur_value);
+		if (ret)
+			return ret;
+		break;
+	default:
+		break;
+	}
+
+	switch (clk_type) {
+	case SMU_SOCCLK:
+	case SMU_VCLK:
+	case SMU_DCLK:
+	case SMU_MCLK:
+	case SMU_FCLK:
+		for (i = 0; i < count; i++) {
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, i, &value);
+			if (ret)
+				return ret;
+			if (!value)
+				continue;
+			size += sprintf(buf + size, "%d: %uMhz %s\n", i, value,
+					cur_value == value ? "*" : "");
+			if (cur_value == value)
+				cur_value_match_level = true;
 		}
+
+		if (!cur_value_match_level)
+			size += sprintf(buf + size, "   %uMhz *\n", cur_value);
 		break;
 	default:
 		break;
@@ -431,6 +541,726 @@ static int vangogh_print_fine_grain_clk(struct smu_context *smu,
 	return size;
 }
 
+static int vangogh_get_profiling_clk_mask(struct smu_context *smu,
+					 enum amd_dpm_forced_level level,
+					 uint32_t *vclk_mask,
+					 uint32_t *dclk_mask,
+					 uint32_t *mclk_mask,
+					 uint32_t *fclk_mask,
+					 uint32_t *soc_mask)
+{
+	DpmClocks_t *clk_table = smu->smu_table.clocks_table;
+
+	if (level == AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK) {
+		if (mclk_mask)
+			*mclk_mask = clk_table->NumDfPstatesEnabled - 1;
+
+		if (fclk_mask)
+			*fclk_mask = clk_table->NumDfPstatesEnabled - 1;
+
+		if (soc_mask)
+			*soc_mask = 0;
+	} else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) {
+		if (mclk_mask)
+			*mclk_mask = 0;
+
+		if (fclk_mask)
+			*fclk_mask = 0;
+
+		if (soc_mask)
+			*soc_mask = 1;
+
+		if (vclk_mask)
+			*vclk_mask = 1;
+
+		if (dclk_mask)
+			*dclk_mask = 1;
+	} else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD) {
+		if (mclk_mask)
+			*mclk_mask = 0;
+
+		if (fclk_mask)
+			*fclk_mask = 0;
+
+		if (soc_mask)
+			*soc_mask = 1;
+
+		if (vclk_mask)
+			*vclk_mask = 1;
+
+		if (dclk_mask)
+			*dclk_mask = 1;
+	}
+
+	return 0;
+}
+
+static bool vangogh_clk_dpm_is_enabled(struct smu_context *smu,
+				enum smu_clk_type clk_type)
+{
+	enum smu_feature_mask feature_id = 0;
+
+	switch (clk_type) {
+	case SMU_MCLK:
+	case SMU_UCLK:
+	case SMU_FCLK:
+		feature_id = SMU_FEATURE_DPM_FCLK_BIT;
+		break;
+	case SMU_GFXCLK:
+	case SMU_SCLK:
+		feature_id = SMU_FEATURE_DPM_GFXCLK_BIT;
+		break;
+	case SMU_SOCCLK:
+		feature_id = SMU_FEATURE_DPM_SOCCLK_BIT;
+		break;
+	case SMU_VCLK:
+	case SMU_DCLK:
+		feature_id = SMU_FEATURE_VCN_DPM_BIT;
+		break;
+	default:
+		return true;
+	}
+
+	if (!smu_cmn_feature_is_enabled(smu, feature_id))
+		return false;
+
+	return true;
+}
+
+static int vangogh_get_dpm_ultimate_freq(struct smu_context *smu,
+					enum smu_clk_type clk_type,
+					uint32_t *min,
+					uint32_t *max)
+{
+	int ret = 0;
+	uint32_t soc_mask;
+	uint32_t vclk_mask;
+	uint32_t dclk_mask;
+	uint32_t mclk_mask;
+	uint32_t fclk_mask;
+	uint32_t clock_limit;
+
+	if (!vangogh_clk_dpm_is_enabled(smu, clk_type)) {
+		switch (clk_type) {
+		case SMU_MCLK:
+		case SMU_UCLK:
+			clock_limit = smu->smu_table.boot_values.uclk;
+			break;
+		case SMU_FCLK:
+			clock_limit = smu->smu_table.boot_values.fclk;
+			break;
+		case SMU_GFXCLK:
+		case SMU_SCLK:
+			clock_limit = smu->smu_table.boot_values.gfxclk;
+			break;
+		case SMU_SOCCLK:
+			clock_limit = smu->smu_table.boot_values.socclk;
+			break;
+		case SMU_VCLK:
+			clock_limit = smu->smu_table.boot_values.vclk;
+			break;
+		case SMU_DCLK:
+			clock_limit = smu->smu_table.boot_values.dclk;
+			break;
+		default:
+			clock_limit = 0;
+			break;
+		}
+
+		/* clock in Mhz unit */
+		if (min)
+			*min = clock_limit / 100;
+		if (max)
+			*max = clock_limit / 100;
+
+		return 0;
+	}
+	if (max) {
+		ret = vangogh_get_profiling_clk_mask(smu,
+							AMD_DPM_FORCED_LEVEL_PROFILE_PEAK,
+							&vclk_mask,
+							&dclk_mask,
+							&mclk_mask,
+							&fclk_mask,
+							&soc_mask);
+		if (ret)
+			goto failed;
+
+		switch (clk_type) {
+		case SMU_UCLK:
+		case SMU_MCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, mclk_mask, max);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_SOCCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, soc_mask, max);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_FCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, fclk_mask, max);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_VCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, vclk_mask, max);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_DCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, dclk_mask, max);
+			if (ret)
+				goto failed;
+			break;
+		default:
+			ret = -EINVAL;
+			goto failed;
+		}
+	}
+	if (min) {
+		switch (clk_type) {
+		case SMU_UCLK:
+		case SMU_MCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, mclk_mask, min);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_SOCCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, soc_mask, min);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_FCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, fclk_mask, min);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_VCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, vclk_mask, min);
+			if (ret)
+				goto failed;
+			break;
+		case SMU_DCLK:
+			ret = vangogh_get_dpm_clk_limited(smu, clk_type, dclk_mask, min);
+			if (ret)
+				goto failed;
+			break;
+		default:
+			ret = -EINVAL;
+			goto failed;
+		}
+	}
+failed:
+	return ret;
+}
+
+static int vangogh_get_power_profile_mode(struct smu_context *smu,
+					   char *buf)
+{
+	static const char *profile_name[] = {
+					"BOOTUP_DEFAULT",
+					"3D_FULL_SCREEN",
+					"POWER_SAVING",
+					"VIDEO",
+					"VR",
+					"COMPUTE",
+					"CUSTOM"};
+	uint32_t i, size = 0;
+	int16_t workload_type = 0;
+
+	if (!buf)
+		return -EINVAL;
+
+	for (i = 0; i <= PP_SMC_POWER_PROFILE_CUSTOM; i++) {
+		/*
+		 * Conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT
+		 * Not all profile modes are supported on vangogh.
+		 */
+		workload_type = smu_cmn_to_asic_specific_index(smu,
+							       CMN2ASIC_MAPPING_WORKLOAD,
+							       i);
+
+		if (workload_type < 0)
+			continue;
+
+		size += sprintf(buf + size, "%2d %14s%s\n",
+			i, profile_name[i], (i == smu->power_profile_mode) ? "*" : " ");
+	}
+
+	return size;
+}
+
+static int vangogh_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
+{
+	int workload_type, ret;
+	uint32_t profile_mode = input[size];
+
+	if (profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) {
+		dev_err(smu->adev->dev, "Invalid power profile mode %d\n", profile_mode);
+		return -EINVAL;
+	}
+
+	if (profile_mode == PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT ||
+			profile_mode == PP_SMC_POWER_PROFILE_POWERSAVING)
+		return 0;
+
+	/* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
+	workload_type = smu_cmn_to_asic_specific_index(smu,
+						       CMN2ASIC_MAPPING_WORKLOAD,
+						       profile_mode);
+	if (workload_type < 0) {
+		dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on VANGOGH\n",
+					profile_mode);
+		return -EINVAL;
+	}
+
+	ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_ActiveProcessNotify,
+				    1 << workload_type,
+				    NULL);
+	if (ret) {
+		dev_err_once(smu->adev->dev, "Fail to set workload type %d\n",
+					workload_type);
+		return ret;
+	}
+
+	smu->power_profile_mode = profile_mode;
+
+	return 0;
+}
+
+static int vangogh_set_soft_freq_limited_range(struct smu_context *smu,
+					  enum smu_clk_type clk_type,
+					  uint32_t min,
+					  uint32_t max)
+{
+	int ret = 0;
+
+	if (!vangogh_clk_dpm_is_enabled(smu, clk_type))
+		return 0;
+
+	switch (clk_type) {
+	case SMU_GFXCLK:
+	case SMU_SCLK:
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetHardMinGfxClk,
+							min, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetSoftMaxGfxClk,
+							max, NULL);
+		if (ret)
+			return ret;
+		break;
+	case SMU_FCLK:
+	case SMU_MCLK:
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetHardMinFclkByFreq,
+							min, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetSoftMaxFclkByFreq,
+							max, NULL);
+		if (ret)
+			return ret;
+		break;
+	case SMU_SOCCLK:
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetHardMinSocclkByFreq,
+							min, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetSoftMaxSocclkByFreq,
+							max, NULL);
+		if (ret)
+			return ret;
+		break;
+	case SMU_VCLK:
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetHardMinVcn,
+							min << 16, NULL);
+		if (ret)
+			return ret;
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetSoftMaxVcn,
+							max << 16, NULL);
+		if (ret)
+			return ret;
+		break;
+	case SMU_DCLK:
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetHardMinVcn,
+							min, NULL);
+		if (ret)
+			return ret;
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetSoftMaxVcn,
+							max, NULL);
+		if (ret)
+			return ret;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+static int vangogh_force_clk_levels(struct smu_context *smu,
+				   enum smu_clk_type clk_type, uint32_t mask)
+{
+	uint32_t soft_min_level = 0, soft_max_level = 0;
+	uint32_t min_freq = 0, max_freq = 0;
+	int ret = 0 ;
+
+	soft_min_level = mask ? (ffs(mask) - 1) : 0;
+	soft_max_level = mask ? (fls(mask) - 1) : 0;
+
+	switch (clk_type) {
+	case SMU_SOCCLK:
+		ret = vangogh_get_dpm_clk_limited(smu, clk_type,
+						soft_min_level, &min_freq);
+		if (ret)
+			return ret;
+		ret = vangogh_get_dpm_clk_limited(smu, clk_type,
+						soft_max_level, &max_freq);
+		if (ret)
+			return ret;
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetSoftMaxSocclkByFreq,
+								max_freq, NULL);
+		if (ret)
+			return ret;
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetHardMinSocclkByFreq,
+								min_freq, NULL);
+		if (ret)
+			return ret;
+		break;
+	case SMU_MCLK:
+	case SMU_FCLK:
+		ret = vangogh_get_dpm_clk_limited(smu,
+							clk_type, soft_min_level, &min_freq);
+		if (ret)
+			return ret;
+		ret = vangogh_get_dpm_clk_limited(smu,
+							clk_type, soft_max_level, &max_freq);
+		if (ret)
+			return ret;
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetSoftMaxFclkByFreq,
+								max_freq, NULL);
+		if (ret)
+			return ret;
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetHardMinFclkByFreq,
+								min_freq, NULL);
+		if (ret)
+			return ret;
+		break;
+	case SMU_VCLK:
+		ret = vangogh_get_dpm_clk_limited(smu,
+							clk_type, soft_min_level, &min_freq);
+		if (ret)
+			return ret;
+
+		ret = vangogh_get_dpm_clk_limited(smu,
+							clk_type, soft_max_level, &max_freq);
+		if (ret)
+			return ret;
+
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetHardMinVcn,
+								min_freq << 16, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetSoftMaxVcn,
+								max_freq << 16, NULL);
+		if (ret)
+			return ret;
+
+		break;
+	case SMU_DCLK:
+		ret = vangogh_get_dpm_clk_limited(smu,
+							clk_type, soft_min_level, &min_freq);
+		if (ret)
+			return ret;
+
+		ret = vangogh_get_dpm_clk_limited(smu,
+							clk_type, soft_max_level, &max_freq);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetHardMinVcn,
+							min_freq, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+							SMU_MSG_SetSoftMaxVcn,
+							max_freq, NULL);
+		if (ret)
+			return ret;
+
+		break;
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+static int vangogh_force_dpm_limit_value(struct smu_context *smu, bool highest)
+{
+	int ret = 0, i = 0;
+	uint32_t min_freq, max_freq, force_freq;
+	enum smu_clk_type clk_type;
+
+	enum smu_clk_type clks[] = {
+		SMU_SOCCLK,
+		SMU_VCLK,
+		SMU_DCLK,
+		SMU_MCLK,
+		SMU_FCLK,
+	};
+
+	for (i = 0; i < ARRAY_SIZE(clks); i++) {
+		clk_type = clks[i];
+		ret = vangogh_get_dpm_ultimate_freq(smu, clk_type, &min_freq, &max_freq);
+		if (ret)
+			return ret;
+
+		force_freq = highest ? max_freq : min_freq;
+		ret = vangogh_set_soft_freq_limited_range(smu, clk_type, force_freq, force_freq);
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+static int vangogh_unforce_dpm_levels(struct smu_context *smu)
+{
+	int ret = 0, i = 0;
+	uint32_t min_freq, max_freq;
+	enum smu_clk_type clk_type;
+
+	struct clk_feature_map {
+		enum smu_clk_type clk_type;
+		uint32_t	feature;
+	} clk_feature_map[] = {
+		{SMU_MCLK,   SMU_FEATURE_DPM_FCLK_BIT},
+		{SMU_FCLK, SMU_FEATURE_DPM_FCLK_BIT},
+		{SMU_SOCCLK, SMU_FEATURE_DPM_SOCCLK_BIT},
+		{SMU_VCLK, SMU_FEATURE_VCN_DPM_BIT},
+		{SMU_DCLK, SMU_FEATURE_VCN_DPM_BIT},
+	};
+
+	for (i = 0; i < ARRAY_SIZE(clk_feature_map); i++) {
+
+		if (!smu_cmn_feature_is_enabled(smu, clk_feature_map[i].feature))
+		    continue;
+
+		clk_type = clk_feature_map[i].clk_type;
+
+		ret = vangogh_get_dpm_ultimate_freq(smu, clk_type, &min_freq, &max_freq);
+
+		if (ret)
+			return ret;
+
+		ret = vangogh_set_soft_freq_limited_range(smu, clk_type, min_freq, max_freq);
+
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+static int vangogh_set_peak_clock_by_device(struct smu_context *smu)
+{
+	int ret = 0;
+	uint32_t socclk_freq = 0, fclk_freq = 0;
+	uint32_t vclk_freq = 0, dclk_freq = 0;
+
+	ret = vangogh_get_dpm_ultimate_freq(smu, SMU_FCLK, NULL, &fclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_set_soft_freq_limited_range(smu, SMU_FCLK, fclk_freq, fclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_get_dpm_ultimate_freq(smu, SMU_SOCCLK, NULL, &socclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_set_soft_freq_limited_range(smu, SMU_SOCCLK, socclk_freq, socclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &vclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_set_soft_freq_limited_range(smu, SMU_VCLK, vclk_freq, vclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &dclk_freq);
+	if (ret)
+		return ret;
+
+	ret = vangogh_set_soft_freq_limited_range(smu, SMU_DCLK, dclk_freq, dclk_freq);
+	if (ret)
+		return ret;
+
+	return ret;
+}
+
+static int vangogh_set_performance_level(struct smu_context *smu,
+					enum amd_dpm_forced_level level)
+{
+	int ret = 0;
+	uint32_t soc_mask, mclk_mask, fclk_mask;
+	uint32_t vclk_mask = 0, dclk_mask = 0;
+
+	switch (level) {
+	case AMD_DPM_FORCED_LEVEL_HIGH:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = vangogh_force_dpm_limit_value(smu, true);
+		break;
+	case AMD_DPM_FORCED_LEVEL_LOW:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = vangogh_force_dpm_limit_value(smu, false);
+		break;
+	case AMD_DPM_FORCED_LEVEL_AUTO:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = vangogh_unforce_dpm_levels(smu);
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+					SMU_MSG_SetHardMinGfxClk,
+					VANGOGH_UMD_PSTATE_STANDARD_GFXCLK, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+					SMU_MSG_SetSoftMaxGfxClk,
+					VANGOGH_UMD_PSTATE_STANDARD_GFXCLK, NULL);
+		if (ret)
+			return ret;
+
+		ret = vangogh_get_profiling_clk_mask(smu, level,
+							&vclk_mask,
+							&dclk_mask,
+							&mclk_mask,
+							&fclk_mask,
+							&soc_mask);
+		if (ret)
+			return ret;
+
+		vangogh_force_clk_levels(smu, SMU_MCLK, 1 << mclk_mask);
+		vangogh_force_clk_levels(smu, SMU_FCLK, 1 << fclk_mask);
+		vangogh_force_clk_levels(smu, SMU_SOCCLK, 1 << soc_mask);
+		vangogh_force_clk_levels(smu, SMU_VCLK, 1 << vclk_mask);
+		vangogh_force_clk_levels(smu, SMU_DCLK, 1 << dclk_mask);
+
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetHardMinVcn,
+								VANGOGH_UMD_PSTATE_PEAK_DCLK, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetSoftMaxVcn,
+								VANGOGH_UMD_PSTATE_PEAK_DCLK, NULL);
+		if (ret)
+			return ret;
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = vangogh_get_profiling_clk_mask(smu, level,
+							NULL,
+							NULL,
+							&mclk_mask,
+							&fclk_mask,
+							NULL);
+		if (ret)
+			return ret;
+
+		vangogh_force_clk_levels(smu, SMU_MCLK, 1 << mclk_mask);
+		vangogh_force_clk_levels(smu, SMU_FCLK, 1 << fclk_mask);
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+		smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetHardMinGfxClk,
+				VANGOGH_UMD_PSTATE_PEAK_GFXCLK, NULL);
+		if (ret)
+			return ret;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetSoftMaxGfxClk,
+				VANGOGH_UMD_PSTATE_PEAK_GFXCLK, NULL);
+		if (ret)
+			return ret;
+
+		ret = vangogh_set_peak_clock_by_device(smu);
+		break;
+	case AMD_DPM_FORCED_LEVEL_MANUAL:
+	case AMD_DPM_FORCED_LEVEL_PROFILE_EXIT:
+	default:
+		break;
+	}
+	return ret;
+}
+
 static int vangogh_read_sensor(struct smu_context *smu,
 				 enum amd_pp_sensors sensor,
 				 void *data, uint32_t *size)
@@ -492,6 +1322,12 @@ static int vangogh_read_sensor(struct smu_context *smu,
 						   (uint32_t *)data);
 		*size = 4;
 		break;
+	case AMDGPU_PP_SENSOR_CPU_CLK:
+		ret = vangogh_get_smu_metrics_data(smu,
+						   METRICS_AVERAGE_CPUCLK,
+						   (uint32_t *)data);
+		*size = smu->cpu_core_num * sizeof(uint16_t);
+		break;
 	default:
 		ret = -EOPNOTSUPP;
 		break;
@@ -513,7 +1349,7 @@ static int vangogh_set_watermarks_table(struct smu_context *smu,
 
 	if (clock_ranges) {
 		if (clock_ranges->num_reader_wm_sets > NUM_WM_RANGES ||
-		    clock_ranges->num_writer_wm_sets > NUM_WM_RANGES)
+			clock_ranges->num_writer_wm_sets > NUM_WM_RANGES)
 			return -EINVAL;
 
 		for (i = 0; i < clock_ranges->num_reader_wm_sets; i++) {
@@ -574,7 +1410,7 @@ static ssize_t vangogh_get_gpu_metrics(struct smu_context *smu,
 	if (ret)
 		return ret;
 
-	smu_v11_0_init_gpu_metrics_v2_0(gpu_metrics);
+	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 2, 0);
 
 	gpu_metrics->temperature_gfx = metrics.GfxTemperature;
 	gpu_metrics->temperature_soc = metrics.SocTemperature;
@@ -590,14 +1426,17 @@ static ssize_t vangogh_get_gpu_metrics(struct smu_context *smu,
 	gpu_metrics->average_socket_power = metrics.CurrentSocketPower;
 	gpu_metrics->average_cpu_power = metrics.Power[0];
 	gpu_metrics->average_soc_power = metrics.Power[1];
+	gpu_metrics->average_gfx_power = metrics.Power[2];
 	memcpy(&gpu_metrics->average_core_power[0],
 		&metrics.CorePower[0],
 		sizeof(uint16_t) * 8);
 
 	gpu_metrics->average_gfxclk_frequency = metrics.GfxclkFrequency;
 	gpu_metrics->average_socclk_frequency = metrics.SocclkFrequency;
+	gpu_metrics->average_uclk_frequency = metrics.MemclkFrequency;
 	gpu_metrics->average_fclk_frequency = metrics.MemclkFrequency;
 	gpu_metrics->average_vclk_frequency = metrics.VclkFrequency;
+	gpu_metrics->average_dclk_frequency = metrics.DclkFrequency;
 
 	memcpy(&gpu_metrics->current_coreclk[0],
 		&metrics.CoreFrequency[0],
@@ -607,22 +1446,55 @@ static ssize_t vangogh_get_gpu_metrics(struct smu_context *smu,
 
 	gpu_metrics->throttle_status = metrics.ThrottlerStatus;
 
+	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
+
 	*table = (void *)gpu_metrics;
 
 	return sizeof(struct gpu_metrics_v2_0);
 }
 
 static int vangogh_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TABLE_COMMAND type,
-							long input[], uint32_t size)
+					long input[], uint32_t size)
 {
 	int ret = 0;
+	int i;
+	struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
 
-	if (!smu->od_enabled) {
-		dev_warn(smu->adev->dev, "Fine grain is not enabled!\n");
+	if (!(smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL)) {
+		dev_warn(smu->adev->dev,
+			"pp_od_clk_voltage is not accessible if power_dpm_force_perfomance_level is not in manual mode!\n");
 		return -EINVAL;
 	}
 
 	switch (type) {
+	case PP_OD_EDIT_CCLK_VDDC_TABLE:
+		if (size != 3) {
+			dev_err(smu->adev->dev, "Input parameter number not correct (should be 4 for processor)\n");
+			return -EINVAL;
+		}
+		if (input[0] >= smu->cpu_core_num) {
+			dev_err(smu->adev->dev, "core index is overflow, should be less than %d\n",
+				smu->cpu_core_num);
+		}
+		smu->cpu_core_id_select = input[0];
+		if (input[1] == 0) {
+			if (input[2] < smu->cpu_default_soft_min_freq) {
+				dev_warn(smu->adev->dev, "Fine grain setting minimum cclk (%ld) MHz is less than the minimum allowed (%d) MHz\n",
+					input[2], smu->cpu_default_soft_min_freq);
+				return -EINVAL;
+			}
+			smu->cpu_actual_soft_min_freq = input[2];
+		} else if (input[1] == 1) {
+			if (input[2] > smu->cpu_default_soft_max_freq) {
+				dev_warn(smu->adev->dev, "Fine grain setting maximum cclk (%ld) MHz is greater than the maximum allowed (%d) MHz\n",
+					input[2], smu->cpu_default_soft_max_freq);
+				return -EINVAL;
+			}
+			smu->cpu_actual_soft_max_freq = input[2];
+		} else {
+			return -EINVAL;
+		}
+		break;
 	case PP_OD_EDIT_SCLK_VDDC_TABLE:
 		if (size != 2) {
 			dev_err(smu->adev->dev, "Input parameter number not correct\n");
@@ -631,14 +1503,16 @@ static int vangogh_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TAB
 
 		if (input[0] == 0) {
 			if (input[1] < smu->gfx_default_hard_min_freq) {
-				dev_warn(smu->adev->dev, "Fine grain setting minimum sclk (%ld) MHz is less than the minimum allowed (%d) MHz\n",
+				dev_warn(smu->adev->dev,
+					"Fine grain setting minimum sclk (%ld) MHz is less than the minimum allowed (%d) MHz\n",
 					input[1], smu->gfx_default_hard_min_freq);
 				return -EINVAL;
 			}
 			smu->gfx_actual_hard_min_freq = input[1];
 		} else if (input[0] == 1) {
 			if (input[1] > smu->gfx_default_soft_max_freq) {
-				dev_warn(smu->adev->dev, "Fine grain setting maximum sclk (%ld) MHz is greater than the maximum allowed (%d) MHz\n",
+				dev_warn(smu->adev->dev,
+					"Fine grain setting maximum sclk (%ld) MHz is greater than the maximum allowed (%d) MHz\n",
 					input[1], smu->gfx_default_soft_max_freq);
 				return -EINVAL;
 			}
@@ -654,6 +1528,8 @@ static int vangogh_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TAB
 		} else {
 			smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
 			smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+			smu->cpu_actual_soft_min_freq = smu->cpu_default_soft_min_freq;
+			smu->cpu_actual_soft_max_freq = smu->cpu_default_soft_max_freq;
 
 			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetHardMinGfxClk,
 									smu->gfx_actual_hard_min_freq, NULL);
@@ -668,6 +1544,29 @@ static int vangogh_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TAB
 				dev_err(smu->adev->dev, "Restore the default soft max sclk failed!");
 				return ret;
 			}
+
+			if (smu->adev->pm.fw_version < 0x43f1b00) {
+				dev_warn(smu->adev->dev, "CPUSoftMax/CPUSoftMin are not supported, please update SBIOS!\n");
+				break;
+			}
+
+			for (i = 0; i < smu->cpu_core_num; i++) {
+				ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetSoftMinCclk,
+								      (i << 20) | smu->cpu_actual_soft_min_freq,
+								      NULL);
+				if (ret) {
+					dev_err(smu->adev->dev, "Set hard min cclk failed!");
+					return ret;
+				}
+
+				ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetSoftMaxCclk,
+								      (i << 20) | smu->cpu_actual_soft_max_freq,
+								      NULL);
+				if (ret) {
+					dev_err(smu->adev->dev, "Set soft max cclk failed!");
+					return ret;
+				}
+			}
 		}
 		break;
 	case PP_OD_COMMIT_DPM_TABLE:
@@ -676,8 +1575,10 @@ static int vangogh_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TAB
 			return -EINVAL;
 		} else {
 			if (smu->gfx_actual_hard_min_freq > smu->gfx_actual_soft_max_freq) {
-				dev_err(smu->adev->dev, "The setting minimun sclk (%d) MHz is greater than the setting maximum sclk (%d) MHz\n",
-				smu->gfx_actual_hard_min_freq, smu->gfx_actual_soft_max_freq);
+				dev_err(smu->adev->dev,
+					"The setting minimun sclk (%d) MHz is greater than the setting maximum sclk (%d) MHz\n",
+					smu->gfx_actual_hard_min_freq,
+					smu->gfx_actual_soft_max_freq);
 				return -EINVAL;
 			}
 
@@ -694,6 +1595,29 @@ static int vangogh_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TAB
 				dev_err(smu->adev->dev, "Set soft max sclk failed!");
 				return ret;
 			}
+
+			if (smu->adev->pm.fw_version < 0x43f1b00) {
+				dev_warn(smu->adev->dev, "CPUSoftMax/CPUSoftMin are not supported, please update SBIOS!\n");
+				break;
+			}
+
+			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetSoftMinCclk,
+							      ((smu->cpu_core_id_select << 20)
+							       | smu->cpu_actual_soft_min_freq),
+							      NULL);
+			if (ret) {
+				dev_err(smu->adev->dev, "Set hard min cclk failed!");
+				return ret;
+			}
+
+			ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetSoftMaxCclk,
+							      ((smu->cpu_core_id_select << 20)
+							       | smu->cpu_actual_soft_max_freq),
+							      NULL);
+			if (ret) {
+				dev_err(smu->adev->dev, "Set soft max cclk failed!");
+				return ret;
+			}
 		}
 		break;
 	default:
@@ -719,18 +1643,245 @@ static int vangogh_set_fine_grain_gfx_freq_parameters(struct smu_context *smu)
 	smu->gfx_actual_hard_min_freq = 0;
 	smu->gfx_actual_soft_max_freq = 0;
 
+	smu->cpu_default_soft_min_freq = 1400;
+	smu->cpu_default_soft_max_freq = 3500;
+	smu->cpu_actual_soft_min_freq = 0;
+	smu->cpu_actual_soft_max_freq = 0;
+
+	return 0;
+}
+
+static int vangogh_get_dpm_clock_table(struct smu_context *smu, struct dpm_clocks *clock_table)
+{
+	DpmClocks_t *table = smu->smu_table.clocks_table;
+	int i;
+
+	if (!clock_table || !table)
+		return -EINVAL;
+
+	for (i = 0; i < NUM_SOCCLK_DPM_LEVELS; i++) {
+		clock_table->SocClocks[i].Freq = table->SocClocks[i];
+		clock_table->SocClocks[i].Vol = table->SocVoltage[i];
+	}
+
+	for (i = 0; i < NUM_FCLK_DPM_LEVELS; i++) {
+		clock_table->FClocks[i].Freq = table->DfPstateTable[i].fclk;
+		clock_table->FClocks[i].Vol = table->DfPstateTable[i].voltage;
+	}
+
+	for (i = 0; i < NUM_FCLK_DPM_LEVELS; i++) {
+		clock_table->MemClocks[i].Freq = table->DfPstateTable[i].memclk;
+		clock_table->MemClocks[i].Vol = table->DfPstateTable[i].voltage;
+	}
+
 	return 0;
 }
 
+
 static int vangogh_system_features_control(struct smu_context *smu, bool en)
 {
 	struct amdgpu_device *adev = smu->adev;
+	struct smu_feature *feature = &smu->smu_feature;
+	uint32_t feature_mask[2];
+	int ret = 0;
+
+	if (adev->pm.fw_version >= 0x43f1700 && !en)
+		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_RlcPowerNotify,
+						      RLC_STATUS_OFF, NULL);
+
+	bitmap_zero(feature->enabled, feature->feature_num);
+	bitmap_zero(feature->supported, feature->feature_num);
+
+	if (!en)
+		return ret;
+
+	ret = smu_cmn_get_enabled_32_bits_mask(smu, feature_mask, 2);
+	if (ret)
+		return ret;
+
+	bitmap_copy(feature->enabled, (unsigned long *)&feature_mask,
+		    feature->feature_num);
+	bitmap_copy(feature->supported, (unsigned long *)&feature_mask,
+		    feature->feature_num);
+
+	return 0;
+}
 
-	if (adev->pm.fw_version >= 0x43f1700)
-		return smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_RlcPowerNotify,
-						en ? RLC_STATUS_NORMAL : RLC_STATUS_OFF, NULL);
-	else
+static int vangogh_post_smu_init(struct smu_context *smu)
+{
+	struct amdgpu_device *adev = smu->adev;
+	uint32_t tmp;
+	int ret = 0;
+	uint8_t aon_bits = 0;
+	/* Two CUs in one WGP */
+	uint32_t req_active_wgps = adev->gfx.cu_info.number/2;
+	uint32_t total_cu = adev->gfx.config.max_cu_per_sh *
+		adev->gfx.config.max_sh_per_se * adev->gfx.config.max_shader_engines;
+
+	/* allow message will be sent after enable message on Vangogh*/
+	if (smu_cmn_feature_is_supported(smu, SMU_FEATURE_DPM_GFXCLK_BIT) &&
+			(adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) {
+		ret = smu_cmn_send_smc_msg(smu, SMU_MSG_EnableGfxOff, NULL);
+		if (ret) {
+			dev_err(adev->dev, "Failed to Enable GfxOff!\n");
+			return ret;
+		}
+	} else {
+		adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
+		dev_info(adev->dev, "If GFX DPM or power gate disabled, disable GFXOFF\n");
+	}
+
+	/* if all CUs are active, no need to power off any WGPs */
+	if (total_cu == adev->gfx.cu_info.number)
 		return 0;
+
+	/*
+	 * Calculate the total bits number of always on WGPs for all SA/SEs in
+	 * RLC_PG_ALWAYS_ON_WGP_MASK.
+	 */
+	tmp = RREG32_KIQ(SOC15_REG_OFFSET(GC, 0, mmRLC_PG_ALWAYS_ON_WGP_MASK));
+	tmp &= RLC_PG_ALWAYS_ON_WGP_MASK__AON_WGP_MASK_MASK;
+
+	aon_bits = hweight32(tmp) * adev->gfx.config.max_sh_per_se * adev->gfx.config.max_shader_engines;
+
+	/* Do not request any WGPs less than set in the AON_WGP_MASK */
+	if (aon_bits > req_active_wgps) {
+		dev_info(adev->dev, "Number of always on WGPs greater than active WGPs: WGP power save not requested.\n");
+		return 0;
+	} else {
+		return smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_RequestActiveWgp, req_active_wgps, NULL);
+	}
+}
+
+static int vangogh_mode_reset(struct smu_context *smu, int type)
+{
+	int ret = 0, index = 0;
+
+	index = smu_cmn_to_asic_specific_index(smu, CMN2ASIC_MAPPING_MSG,
+					       SMU_MSG_GfxDeviceDriverReset);
+	if (index < 0)
+		return index == -EACCES ? 0 : index;
+
+	mutex_lock(&smu->message_lock);
+
+	ret = smu_cmn_send_msg_without_waiting(smu, (uint16_t)index, type);
+
+	mutex_unlock(&smu->message_lock);
+
+	mdelay(10);
+
+	return ret;
+}
+
+static int vangogh_mode2_reset(struct smu_context *smu)
+{
+	return vangogh_mode_reset(smu, SMU_RESET_MODE_2);
+}
+
+static int vangogh_get_power_limit(struct smu_context *smu)
+{
+	struct smu_11_5_power_context *power_context =
+								smu->smu_power.power_context;
+	uint32_t ppt_limit;
+	int ret = 0;
+
+	if (smu->adev->pm.fw_version < 0x43f1e00)
+		return ret;
+
+	ret = smu_cmn_send_smc_msg(smu, SMU_MSG_GetSlowPPTLimit, &ppt_limit);
+	if (ret) {
+		dev_err(smu->adev->dev, "Get slow PPT limit failed!\n");
+		return ret;
+	}
+	/* convert from milliwatt to watt */
+	smu->current_power_limit = ppt_limit / 1000;
+	smu->max_power_limit = 29;
+
+	ret = smu_cmn_send_smc_msg(smu, SMU_MSG_GetFastPPTLimit, &ppt_limit);
+	if (ret) {
+		dev_err(smu->adev->dev, "Get fast PPT limit failed!\n");
+		return ret;
+	}
+	/* convert from milliwatt to watt */
+	power_context->current_fast_ppt_limit = ppt_limit / 1000;
+	power_context->max_fast_ppt_limit = 30;
+
+	return ret;
+}
+
+static int vangogh_get_ppt_limit(struct smu_context *smu,
+								uint32_t *ppt_limit,
+								enum smu_ppt_limit_type type,
+								enum smu_ppt_limit_level level)
+{
+	struct smu_11_5_power_context *power_context =
+							smu->smu_power.power_context;
+
+	if (!power_context)
+		return -EOPNOTSUPP;
+
+	if (type == SMU_FAST_PPT_LIMIT) {
+		switch (level) {
+		case SMU_PPT_LIMIT_MAX:
+			*ppt_limit = power_context->max_fast_ppt_limit;
+			break;
+		case SMU_PPT_LIMIT_CURRENT:
+			*ppt_limit = power_context->current_fast_ppt_limit;
+			break;
+		default:
+			break;
+		}
+	}
+
+	return 0;
+}
+
+static int vangogh_set_power_limit(struct smu_context *smu, uint32_t ppt_limit)
+{
+	struct smu_11_5_power_context *power_context =
+							smu->smu_power.power_context;
+	uint32_t limit_type = ppt_limit >> 24;
+	int ret = 0;
+
+	if (!smu_cmn_feature_is_enabled(smu, SMU_FEATURE_PPT_BIT)) {
+		dev_err(smu->adev->dev, "Setting new power limit is not supported!\n");
+		return -EOPNOTSUPP;
+	}
+
+	switch (limit_type) {
+	case SMU_DEFAULT_PPT_LIMIT:
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+				SMU_MSG_SetSlowPPTLimit,
+				ppt_limit * 1000, /* convert from watt to milliwatt */
+				NULL);
+		if (ret)
+			return ret;
+
+		smu->current_power_limit = ppt_limit;
+		break;
+	case SMU_FAST_PPT_LIMIT:
+		ppt_limit &= ~(SMU_FAST_PPT_LIMIT << 24);
+		if (ppt_limit > power_context->max_fast_ppt_limit) {
+			dev_err(smu->adev->dev,
+				"New power limit (%d) is over the max allowed %d\n",
+				ppt_limit, power_context->max_fast_ppt_limit);
+			return ret;
+		}
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+				SMU_MSG_SetFastPPTLimit,
+				ppt_limit * 1000, /* convert from watt to milliwatt */
+				NULL);
+		if (ret)
+			return ret;
+
+		power_context->current_fast_ppt_limit = ppt_limit;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
 }
 
 static const struct pptable_funcs vangogh_ppt_funcs = {
@@ -742,7 +1893,6 @@ static const struct pptable_funcs vangogh_ppt_funcs = {
 	.init_power = smu_v11_0_init_power,
 	.fini_power = smu_v11_0_fini_power,
 	.register_irq_handler = smu_v11_0_register_irq_handler,
-	.get_allowed_feature_mask = vangogh_get_allowed_feature_mask,
 	.notify_memory_pool_location = smu_v11_0_notify_memory_pool_location,
 	.send_smc_msg_with_param = smu_cmn_send_smc_msg_with_param,
 	.send_smc_msg = smu_cmn_send_smc_msg,
@@ -761,6 +1911,18 @@ static const struct pptable_funcs vangogh_ppt_funcs = {
 	.set_default_dpm_table = vangogh_set_default_dpm_tables,
 	.set_fine_grain_gfx_freq_parameters = vangogh_set_fine_grain_gfx_freq_parameters,
 	.system_features_control = vangogh_system_features_control,
+	.feature_is_enabled = smu_cmn_feature_is_enabled,
+	.set_power_profile_mode = vangogh_set_power_profile_mode,
+	.get_power_profile_mode = vangogh_get_power_profile_mode,
+	.get_dpm_clock_table = vangogh_get_dpm_clock_table,
+	.force_clk_levels = vangogh_force_clk_levels,
+	.set_performance_level = vangogh_set_performance_level,
+	.post_init = vangogh_post_smu_init,
+	.mode2_reset = vangogh_mode2_reset,
+	.gfx_off_control = smu_v11_0_gfx_off_control,
+	.get_ppt_limit = vangogh_get_ppt_limit,
+	.get_power_limit = vangogh_get_power_limit,
+	.set_power_limit = vangogh_set_power_limit,
 };
 
 void vangogh_set_ppt_funcs(struct smu_context *smu)
@@ -769,5 +1931,6 @@ void vangogh_set_ppt_funcs(struct smu_context *smu)
 	smu->message_map = vangogh_message_map;
 	smu->feature_map = vangogh_feature_mask_map;
 	smu->table_map = vangogh_table_map;
+	smu->workload_map = vangogh_workload_map;
 	smu->is_apu = true;
 }
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.h b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.h
index eab455493076..c56d4583dc72 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.h
@@ -28,9 +28,29 @@
 extern void vangogh_set_ppt_funcs(struct smu_context *smu);
 
 /* UMD PState Vangogh Msg Parameters in MHz */
-#define VANGOGH_UMD_PSTATE_GFXCLK       700
-#define VANGOGH_UMD_PSTATE_SOCCLK       678
-#define VANGOGH_UMD_PSTATE_FCLK         800
+#define VANGOGH_UMD_PSTATE_STANDARD_GFXCLK       1100
+#define VANGOGH_UMD_PSTATE_STANDARD_SOCCLK       600
+#define VANGOGH_UMD_PSTATE_STANDARD_FCLK         800
+#define VANGOGH_UMD_PSTATE_STANDARD_VCLK         705
+#define VANGOGH_UMD_PSTATE_STANDARD_DCLK         600
+
+#define VANGOGH_UMD_PSTATE_PEAK_GFXCLK       1300
+#define VANGOGH_UMD_PSTATE_PEAK_SOCCLK       600
+#define VANGOGH_UMD_PSTATE_PEAK_FCLK         800
+#define VANGOGH_UMD_PSTATE_PEAK_VCLK         705
+#define VANGOGH_UMD_PSTATE_PEAK_DCLK         600
+
+#define VANGOGH_UMD_PSTATE_MIN_SCLK_GFXCLK       400
+#define VANGOGH_UMD_PSTATE_MIN_SCLK_SOCCLK       1000
+#define VANGOGH_UMD_PSTATE_MIN_SCLK_FCLK         800
+#define VANGOGH_UMD_PSTATE_MIN_SCLK_VCLK         1000
+#define VANGOGH_UMD_PSTATE_MIN_SCLK_DCLK         800
+
+#define VANGOGH_UMD_PSTATE_MIN_MCLK_GFXCLK       1100
+#define VANGOGH_UMD_PSTATE_MIN_MCLK_SOCCLK       1000
+#define VANGOGH_UMD_PSTATE_MIN_MCLK_FCLK         400
+#define VANGOGH_UMD_PSTATE_MIN_MCLK_VCLK         1000
+#define VANGOGH_UMD_PSTATE_MIN_MCLK_DCLK         800
 
 /* RLC Power Status */
 #define RLC_STATUS_OFF          0
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
index dc75db8af371..5493388fcb10 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
@@ -56,8 +56,6 @@ static struct cmn2asic_msg_mapping renoir_message_map[SMU_MSG_MAX_COUNT] = {
 	MSG_MAP(PowerUpSdma,                    PPSMC_MSG_PowerUpSdma,                  1),
 	MSG_MAP(SetHardMinIspclkByFreq,         PPSMC_MSG_SetHardMinIspclkByFreq,       1),
 	MSG_MAP(SetHardMinVcn,                  PPSMC_MSG_SetHardMinVcn,                1),
-	MSG_MAP(Spare1,                         PPSMC_MSG_spare1,                       1),
-	MSG_MAP(Spare2,                         PPSMC_MSG_spare2,                       1),
 	MSG_MAP(SetAllowFclkSwitch,             PPSMC_MSG_SetAllowFclkSwitch,           1),
 	MSG_MAP(SetMinVideoGfxclkFreq,          PPSMC_MSG_SetMinVideoGfxclkFreq,        1),
 	MSG_MAP(ActiveProcessNotify,            PPSMC_MSG_ActiveProcessNotify,          1),
@@ -188,6 +186,7 @@ static int renoir_get_dpm_clk_limited(struct smu_context *smu, enum smu_clk_type
 			return -EINVAL;
 		*freq = clk_table->SocClocks[dpm_level].Freq;
 		break;
+	case SMU_UCLK:
 	case SMU_MCLK:
 		if (dpm_level >= NUM_FCLK_DPM_LEVELS)
 			return -EINVAL;
@@ -343,12 +342,142 @@ failed:
 	return ret;
 }
 
+static int renoir_od_edit_dpm_table(struct smu_context *smu,
+							enum PP_OD_DPM_TABLE_COMMAND type,
+							long input[], uint32_t size)
+{
+	int ret = 0;
+	struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
+
+	if (!(smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL)) {
+		dev_warn(smu->adev->dev,
+			"pp_od_clk_voltage is not accessible if power_dpm_force_perfomance_level is not in manual mode!\n");
+		return -EINVAL;
+	}
+
+	switch (type) {
+	case PP_OD_EDIT_SCLK_VDDC_TABLE:
+		if (size != 2) {
+			dev_err(smu->adev->dev, "Input parameter number not correct\n");
+			return -EINVAL;
+		}
+
+		if (input[0] == 0) {
+			if (input[1] < smu->gfx_default_hard_min_freq) {
+				dev_warn(smu->adev->dev,
+					"Fine grain setting minimum sclk (%ld) MHz is less than the minimum allowed (%d) MHz\n",
+					input[1], smu->gfx_default_hard_min_freq);
+				return -EINVAL;
+			}
+			smu->gfx_actual_hard_min_freq = input[1];
+		} else if (input[0] == 1) {
+			if (input[1] > smu->gfx_default_soft_max_freq) {
+				dev_warn(smu->adev->dev,
+					"Fine grain setting maximum sclk (%ld) MHz is greater than the maximum allowed (%d) MHz\n",
+					input[1], smu->gfx_default_soft_max_freq);
+				return -EINVAL;
+			}
+			smu->gfx_actual_soft_max_freq = input[1];
+		} else {
+			return -EINVAL;
+		}
+		break;
+	case PP_OD_RESTORE_DEFAULT_TABLE:
+		if (size != 0) {
+			dev_err(smu->adev->dev, "Input parameter number not correct\n");
+			return -EINVAL;
+		}
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetHardMinGfxClk,
+								smu->gfx_actual_hard_min_freq,
+								NULL);
+		if (ret) {
+			dev_err(smu->adev->dev, "Restore the default hard min sclk failed!");
+			return ret;
+		}
+
+		ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetSoftMaxGfxClk,
+								smu->gfx_actual_soft_max_freq,
+								NULL);
+		if (ret) {
+			dev_err(smu->adev->dev, "Restore the default soft max sclk failed!");
+			return ret;
+		}
+		break;
+	case PP_OD_COMMIT_DPM_TABLE:
+		if (size != 0) {
+			dev_err(smu->adev->dev, "Input parameter number not correct\n");
+			return -EINVAL;
+		} else {
+			if (smu->gfx_actual_hard_min_freq > smu->gfx_actual_soft_max_freq) {
+				dev_err(smu->adev->dev,
+					"The setting minimun sclk (%d) MHz is greater than the setting maximum sclk (%d) MHz\n",
+					smu->gfx_actual_hard_min_freq,
+					smu->gfx_actual_soft_max_freq);
+				return -EINVAL;
+			}
+
+			ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetHardMinGfxClk,
+								smu->gfx_actual_hard_min_freq,
+								NULL);
+			if (ret) {
+				dev_err(smu->adev->dev, "Set hard min sclk failed!");
+				return ret;
+			}
+
+			ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_SetSoftMaxGfxClk,
+								smu->gfx_actual_soft_max_freq,
+								NULL);
+			if (ret) {
+				dev_err(smu->adev->dev, "Set soft max sclk failed!");
+				return ret;
+			}
+		}
+		break;
+	default:
+		return -ENOSYS;
+	}
+
+	return ret;
+}
+
+static int renoir_set_fine_grain_gfx_freq_parameters(struct smu_context *smu)
+{
+	uint32_t min = 0, max = 0;
+	uint32_t ret = 0;
+
+	ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_GetMinGfxclkFrequency,
+								0, &min);
+	if (ret)
+		return ret;
+	ret = smu_cmn_send_smc_msg_with_param(smu,
+								SMU_MSG_GetMaxGfxclkFrequency,
+								0, &max);
+	if (ret)
+		return ret;
+
+	smu->gfx_default_hard_min_freq = min;
+	smu->gfx_default_soft_max_freq = max;
+	smu->gfx_actual_hard_min_freq = 0;
+	smu->gfx_actual_soft_max_freq = 0;
+
+	return 0;
+}
+
 static int renoir_print_clk_levels(struct smu_context *smu,
 			enum smu_clk_type clk_type, char *buf)
 {
 	int i, size = 0, ret = 0;
 	uint32_t cur_value = 0, value = 0, count = 0, min = 0, max = 0;
 	SmuMetrics_t metrics;
+	struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
 	bool cur_value_match_level = false;
 
 	memset(&metrics, 0, sizeof(metrics));
@@ -358,6 +487,30 @@ static int renoir_print_clk_levels(struct smu_context *smu,
 		return ret;
 
 	switch (clk_type) {
+	case SMU_OD_RANGE:
+		if (smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) {
+			ret = smu_cmn_send_smc_msg_with_param(smu,
+						SMU_MSG_GetMinGfxclkFrequency,
+						0, &min);
+			if (ret)
+				return ret;
+			ret = smu_cmn_send_smc_msg_with_param(smu,
+						SMU_MSG_GetMaxGfxclkFrequency,
+						0, &max);
+			if (ret)
+				return ret;
+			size += sprintf(buf + size, "OD_RANGE\nSCLK: %10uMhz %10uMhz\n", min, max);
+		}
+		break;
+	case SMU_OD_SCLK:
+		if (smu_dpm_ctx->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) {
+			min = (smu->gfx_actual_hard_min_freq > 0) ? smu->gfx_actual_hard_min_freq : smu->gfx_default_hard_min_freq;
+			max = (smu->gfx_actual_soft_max_freq > 0) ? smu->gfx_actual_soft_max_freq : smu->gfx_default_soft_max_freq;
+			size += sprintf(buf + size, "OD_SCLK\n");
+			size += sprintf(buf + size, "0:%10uMhz\n", min);
+			size += sprintf(buf + size, "1:%10uMhz\n", max);
+		}
+		break;
 	case SMU_GFXCLK:
 	case SMU_SCLK:
 		/* retirve table returned paramters unit is MHz */
@@ -398,23 +551,35 @@ static int renoir_print_clk_levels(struct smu_context *smu,
 		cur_value = metrics.ClockFrequency[CLOCK_FCLK];
 		break;
 	default:
-		return -EINVAL;
+		break;
 	}
 
-	for (i = 0; i < count; i++) {
-		ret = renoir_get_dpm_clk_limited(smu, clk_type, i, &value);
-		if (ret)
-			return ret;
-		if (!value)
-			continue;
-		size += sprintf(buf + size, "%d: %uMhz %s\n", i, value,
-				cur_value == value ? "*" : "");
-		if (cur_value == value)
-			cur_value_match_level = true;
-	}
+	switch (clk_type) {
+	case SMU_GFXCLK:
+	case SMU_SCLK:
+	case SMU_SOCCLK:
+	case SMU_MCLK:
+	case SMU_DCEFCLK:
+	case SMU_FCLK:
+		for (i = 0; i < count; i++) {
+			ret = renoir_get_dpm_clk_limited(smu, clk_type, i, &value);
+			if (ret)
+				return ret;
+			if (!value)
+				continue;
+			size += sprintf(buf + size, "%d: %uMhz %s\n", i, value,
+					cur_value == value ? "*" : "");
+			if (cur_value == value)
+				cur_value_match_level = true;
+		}
 
-	if (!cur_value_match_level)
-		size += sprintf(buf + size, "   %uMhz *\n", cur_value);
+		if (!cur_value_match_level)
+			size += sprintf(buf + size, "   %uMhz *\n", cur_value);
+
+		break;
+	default:
+		break;
+	}
 
 	return size;
 }
@@ -666,6 +831,10 @@ static int renoir_set_power_profile_mode(struct smu_context *smu, long *input, u
 		return -EINVAL;
 	}
 
+	if (profile_mode == PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT ||
+			profile_mode == PP_SMC_POWER_PROFILE_POWERSAVING)
+		return 0;
+
 	/* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
 	workload_type = smu_cmn_to_asic_specific_index(smu,
 						       CMN2ASIC_MAPPING_WORKLOAD,
@@ -675,7 +844,7 @@ static int renoir_set_power_profile_mode(struct smu_context *smu, long *input, u
 		 * TODO: If some case need switch to powersave/default power mode
 		 * then can consider enter WORKLOAD_COMPUTE/WORKLOAD_CUSTOM for power saving.
 		 */
-		dev_err_once(smu->adev->dev, "Unsupported power profile mode %d on RENOIR\n", profile_mode);
+		dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on RENOIR\n", profile_mode);
 		return -EINVAL;
 	}
 
@@ -724,15 +893,27 @@ static int renoir_set_performance_level(struct smu_context *smu,
 
 	switch (level) {
 	case AMD_DPM_FORCED_LEVEL_HIGH:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
 		ret = renoir_force_dpm_limit_value(smu, true);
 		break;
 	case AMD_DPM_FORCED_LEVEL_LOW:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
 		ret = renoir_force_dpm_limit_value(smu, false);
 		break;
 	case AMD_DPM_FORCED_LEVEL_AUTO:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
 		ret = renoir_unforce_dpm_levels(smu);
 		break;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
 		ret = smu_cmn_send_smc_msg_with_param(smu,
 						      SMU_MSG_SetHardMinGfxClk,
 						      RENOIR_UMD_PSTATE_GFXCLK,
@@ -785,6 +966,9 @@ static int renoir_set_performance_level(struct smu_context *smu,
 		break;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK:
 	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
 		ret = renoir_get_profiling_clk_mask(smu, level,
 						    &sclk_mask,
 						    &mclk_mask,
@@ -796,6 +980,9 @@ static int renoir_set_performance_level(struct smu_context *smu,
 		renoir_force_clk_levels(smu, SMU_SOCCLK, 1 << soc_mask);
 		break;
 	case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK:
+		smu->gfx_actual_hard_min_freq = smu->gfx_default_hard_min_freq;
+		smu->gfx_actual_soft_max_freq = smu->gfx_default_soft_max_freq;
+
 		ret = renoir_set_peak_clock_by_device(smu);
 		break;
 	case AMD_DPM_FORCED_LEVEL_MANUAL:
@@ -943,7 +1130,7 @@ static int renoir_get_smu_metrics_data(struct smu_context *smu,
 		*value = metrics->AverageUvdActivity / 100;
 		break;
 	case METRICS_AVERAGE_SOCKETPOWER:
-		*value = metrics->CurrentSocketPower << 8;
+		*value = (metrics->CurrentSocketPower << 8) / 1000;
 		break;
 	case METRICS_TEMPERATURE_EDGE:
 		*value = (metrics->GfxTemperature / 100) *
@@ -1071,7 +1258,7 @@ static ssize_t renoir_get_gpu_metrics(struct smu_context *smu,
 	if (ret)
 		return ret;
 
-	smu_v12_0_init_gpu_metrics_v2_0(gpu_metrics);
+	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 2, 0);
 
 	gpu_metrics->temperature_gfx = metrics.GfxTemperature;
 	gpu_metrics->temperature_soc = metrics.SocTemperature;
@@ -1112,6 +1299,8 @@ static ssize_t renoir_get_gpu_metrics(struct smu_context *smu,
 
 	gpu_metrics->fan_pwm = metrics.FanPwm;
 
+	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
+
 	*table = (void *)gpu_metrics;
 
 	return sizeof(struct gpu_metrics_v2_0);
@@ -1120,7 +1309,7 @@ static ssize_t renoir_get_gpu_metrics(struct smu_context *smu,
 static int renoir_gfx_state_change_set(struct smu_context *smu, uint32_t state)
 {
 
-	return smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_GpuChangeState, state, NULL);
+	return 0;
 }
 
 static const struct pptable_funcs renoir_ppt_funcs = {
@@ -1159,6 +1348,8 @@ static const struct pptable_funcs renoir_ppt_funcs = {
 	.set_pp_feature_mask = smu_cmn_set_pp_feature_mask,
 	.get_gpu_metrics = renoir_get_gpu_metrics,
 	.gfx_state_change_set = renoir_gfx_state_change_set,
+	.set_fine_grain_gfx_freq_parameters = renoir_set_fine_grain_gfx_freq_parameters,
+	.od_edit_dpm_table = renoir_od_edit_dpm_table,
 };
 
 void renoir_set_ppt_funcs(struct smu_context *smu)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu12/smu_v12_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu12/smu_v12_0.c
index 522d55004655..6cc4855c8a37 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu12/smu_v12_0.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu12/smu_v12_0.c
@@ -225,6 +225,7 @@ int smu_v12_0_set_soft_freq_limited_range(struct smu_context *smu, enum smu_clk_
 	break;
 	case SMU_FCLK:
 	case SMU_MCLK:
+	case SMU_UCLK:
 		ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetHardMinFclkByFreq, min, NULL);
 		if (ret)
 			return ret;
@@ -277,15 +278,3 @@ int smu_v12_0_set_driver_table_location(struct smu_context *smu)
 
 	return ret;
 }
-
-void smu_v12_0_init_gpu_metrics_v2_0(struct gpu_metrics_v2_0 *gpu_metrics)
-{
-	memset(gpu_metrics, 0xFF, sizeof(struct gpu_metrics_v2_0));
-
-	gpu_metrics->common_header.structure_size =
-				sizeof(struct gpu_metrics_v2_0);
-	gpu_metrics->common_header.format_revision = 2;
-	gpu_metrics->common_header.content_revision = 0;
-
-	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
-}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
index f8260769061c..bcedd4d92e35 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
@@ -68,14 +68,6 @@ static const char *smu_get_message_name(struct smu_context *smu,
 	return __smu_message_names[type];
 }
 
-static void smu_cmn_send_msg_without_waiting(struct smu_context *smu,
-					     uint16_t msg)
-{
-	struct amdgpu_device *adev = smu->adev;
-
-	WREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_66, msg);
-}
-
 static void smu_cmn_read_arg(struct smu_context *smu,
 			     uint32_t *arg)
 {
@@ -92,7 +84,7 @@ static int smu_cmn_wait_for_response(struct smu_context *smu)
 	for (i = 0; i < timeout; i++) {
 		cur_value = RREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_90);
 		if ((cur_value & MP1_C2PMSG_90__CONTENT_MASK) != 0)
-			return cur_value == 0x1 ? 0 : -EIO;
+			return cur_value;
 
 		udelay(1);
 	}
@@ -101,7 +93,29 @@ static int smu_cmn_wait_for_response(struct smu_context *smu)
 	if (i == timeout)
 		return -ETIME;
 
-	return RREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_90) == 0x1 ? 0 : -EIO;
+	return RREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_90);
+}
+
+int smu_cmn_send_msg_without_waiting(struct smu_context *smu,
+				     uint16_t msg, uint32_t param)
+{
+	struct amdgpu_device *adev = smu->adev;
+	int ret;
+
+	ret = smu_cmn_wait_for_response(smu);
+	if (ret != 0x1) {
+		dev_err(adev->dev, "Msg issuing pre-check failed and "
+		       "SMU may be not in the right state!\n");
+		if (ret != -ETIME)
+			ret = -EIO;
+		return ret;
+	}
+
+	WREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_90, 0);
+	WREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_82, param);
+	WREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_66, msg);
+
+	return 0;
 }
 
 int smu_cmn_send_smc_msg_with_param(struct smu_context *smu,
@@ -122,29 +136,28 @@ int smu_cmn_send_smc_msg_with_param(struct smu_context *smu,
 		return index == -EACCES ? 0 : index;
 
 	mutex_lock(&smu->message_lock);
-	ret = smu_cmn_wait_for_response(smu);
-	if (ret) {
-		dev_err(adev->dev, "Msg issuing pre-check failed and "
-		       "SMU may be not in the right state!\n");
+	ret = smu_cmn_send_msg_without_waiting(smu, (uint16_t)index, param);
+	if (ret)
 		goto out;
-	}
-
-	WREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_90, 0);
-
-	WREG32_SOC15_NO_KIQ(MP1, 0, mmMP1_SMN_C2PMSG_82, param);
-
-	smu_cmn_send_msg_without_waiting(smu, (uint16_t)index);
 
 	ret = smu_cmn_wait_for_response(smu);
-	if (ret) {
-		dev_err(adev->dev, "failed send message: %10s (%d) \tparam: 0x%08x response %#x\n",
-		       smu_get_message_name(smu, msg), index, param, ret);
+	if (ret != 0x1) {
+		if (ret == -ETIME) {
+			dev_err(adev->dev, "message: %15s (%d) \tparam: 0x%08x is timeout (no response)\n",
+				smu_get_message_name(smu, msg), index, param);
+		} else {
+			dev_err(adev->dev, "failed send message: %15s (%d) \tparam: 0x%08x response %#x\n",
+				smu_get_message_name(smu, msg), index, param,
+				ret);
+			ret = -EIO;
+		}
 		goto out;
 	}
 
 	if (read_arg)
 		smu_cmn_read_arg(smu, read_arg);
 
+	ret = 0; /* 0 as driver return value */
 out:
 	mutex_unlock(&smu->message_lock);
 	return ret;
@@ -269,11 +282,13 @@ int smu_cmn_feature_is_enabled(struct smu_context *smu,
 			       enum smu_feature_mask mask)
 {
 	struct smu_feature *feature = &smu->smu_feature;
+	struct amdgpu_device *adev = smu->adev;
 	int feature_id;
 	int ret = 0;
 
-	if (smu->is_apu)
+	if (smu->is_apu && adev->family < AMDGPU_FAMILY_VGH)
 		return 1;
+
 	feature_id = smu_cmn_to_asic_specific_index(smu,
 						    CMN2ASIC_MAPPING_FEATURE,
 						    mask);
@@ -731,3 +746,31 @@ int smu_cmn_get_metrics_table(struct smu_context *smu,
 
 	return ret;
 }
+
+void smu_cmn_init_soft_gpu_metrics(void *table, uint8_t frev, uint8_t crev)
+{
+	struct metrics_table_header *header = (struct metrics_table_header *)table;
+	uint16_t structure_size;
+
+#define METRICS_VERSION(a, b)	((a << 16) | b )
+
+	switch (METRICS_VERSION(frev, crev)) {
+	case METRICS_VERSION(1, 0):
+		structure_size = sizeof(struct gpu_metrics_v1_0);
+		break;
+	case METRICS_VERSION(2, 0):
+		structure_size = sizeof(struct gpu_metrics_v2_0);
+		break;
+	default:
+		return;
+	}
+
+#undef METRICS_VERSION
+
+	memset(header, 0xFF, structure_size);
+
+	header->format_revision = frev;
+	header->content_revision = crev;
+	header->structure_size = structure_size;
+
+}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
index 01e825d83d8d..c69250185575 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
@@ -26,6 +26,8 @@
 #include "amdgpu_smu.h"
 
 #if defined(SWSMU_CODE_LAYER_L2) || defined(SWSMU_CODE_LAYER_L3) || defined(SWSMU_CODE_LAYER_L4)
+int smu_cmn_send_msg_without_waiting(struct smu_context *smu,
+				     uint16_t msg, uint32_t param);
 int smu_cmn_send_smc_msg_with_param(struct smu_context *smu,
 				    enum smu_message_type msg,
 				    uint32_t param,
@@ -95,5 +97,7 @@ int smu_cmn_get_metrics_table(struct smu_context *smu,
 			      void *metrics_table,
 			      bool bypass_cache);
 
+void smu_cmn_init_soft_gpu_metrics(void *table, uint8_t frev, uint8_t crev);
+
 #endif
 #endif
diff --git a/drivers/gpu/drm/bridge/lontium-lt9611uxc.c b/drivers/gpu/drm/bridge/lontium-lt9611uxc.c
index 0c98d27f84ac..fee27952ec6d 100644
--- a/drivers/gpu/drm/bridge/lontium-lt9611uxc.c
+++ b/drivers/gpu/drm/bridge/lontium-lt9611uxc.c
@@ -14,6 +14,7 @@
 #include <linux/regmap.h>
 #include <linux/regulator/consumer.h>
 #include <linux/wait.h>
+#include <linux/workqueue.h>
 
 #include <sound/hdmi-codec.h>
 
@@ -36,6 +37,7 @@ struct lt9611uxc {
 	struct mutex ocm_lock;
 
 	struct wait_queue_head wq;
+	struct work_struct work;
 
 	struct device_node *dsi0_node;
 	struct device_node *dsi1_node;
@@ -52,6 +54,8 @@ struct lt9611uxc {
 
 	bool hpd_supported;
 	bool edid_read;
+	/* can be accessed from different threads, so protect this with ocm_lock */
+	bool hdmi_connected;
 	uint8_t fw_version;
 };
 
@@ -143,21 +147,41 @@ static irqreturn_t lt9611uxc_irq_thread_handler(int irq, void *dev_id)
 	if (irq_status)
 		regmap_write(lt9611uxc->regmap, 0xb022, 0);
 
-	lt9611uxc_unlock(lt9611uxc);
-
-	if (irq_status & BIT(0))
+	if (irq_status & BIT(0)) {
 		lt9611uxc->edid_read = !!(hpd_status & BIT(0));
+		wake_up_all(&lt9611uxc->wq);
+	}
 
 	if (irq_status & BIT(1)) {
-		if (lt9611uxc->connector.dev)
-			drm_kms_helper_hotplug_event(lt9611uxc->connector.dev);
-		else
-			drm_bridge_hpd_notify(&lt9611uxc->bridge, !!(hpd_status & BIT(1)));
+		lt9611uxc->hdmi_connected = hpd_status & BIT(1);
+		schedule_work(&lt9611uxc->work);
 	}
 
+	lt9611uxc_unlock(lt9611uxc);
+
 	return IRQ_HANDLED;
 }
 
+static void lt9611uxc_hpd_work(struct work_struct *work)
+{
+	struct lt9611uxc *lt9611uxc = container_of(work, struct lt9611uxc, work);
+	bool connected;
+
+	if (lt9611uxc->connector.dev)
+		drm_kms_helper_hotplug_event(lt9611uxc->connector.dev);
+	else {
+
+		mutex_lock(&lt9611uxc->ocm_lock);
+		connected = lt9611uxc->hdmi_connected;
+		mutex_unlock(&lt9611uxc->ocm_lock);
+
+		drm_bridge_hpd_notify(&lt9611uxc->bridge,
+				      connected ?
+				      connector_status_connected :
+				      connector_status_disconnected);
+	}
+}
+
 static void lt9611uxc_reset(struct lt9611uxc *lt9611uxc)
 {
 	gpiod_set_value_cansleep(lt9611uxc->reset_gpio, 1);
@@ -445,18 +469,21 @@ static enum drm_connector_status lt9611uxc_bridge_detect(struct drm_bridge *brid
 	struct lt9611uxc *lt9611uxc = bridge_to_lt9611uxc(bridge);
 	unsigned int reg_val = 0;
 	int ret;
-	int connected = 1;
+	bool connected = true;
+
+	lt9611uxc_lock(lt9611uxc);
 
 	if (lt9611uxc->hpd_supported) {
-		lt9611uxc_lock(lt9611uxc);
 		ret = regmap_read(lt9611uxc->regmap, 0xb023, &reg_val);
-		lt9611uxc_unlock(lt9611uxc);
 
 		if (ret)
 			dev_err(lt9611uxc->dev, "failed to read hpd status: %d\n", ret);
 		else
 			connected  = reg_val & BIT(1);
 	}
+	lt9611uxc->hdmi_connected = connected;
+
+	lt9611uxc_unlock(lt9611uxc);
 
 	return connected ?  connector_status_connected :
 				connector_status_disconnected;
@@ -465,7 +492,7 @@ static enum drm_connector_status lt9611uxc_bridge_detect(struct drm_bridge *brid
 static int lt9611uxc_wait_for_edid(struct lt9611uxc *lt9611uxc)
 {
 	return wait_event_interruptible_timeout(lt9611uxc->wq, lt9611uxc->edid_read,
-			msecs_to_jiffies(100));
+			msecs_to_jiffies(500));
 }
 
 static int lt9611uxc_get_edid_block(void *data, u8 *buf, unsigned int block, size_t len)
@@ -503,7 +530,10 @@ static struct edid *lt9611uxc_bridge_get_edid(struct drm_bridge *bridge,
 	ret = lt9611uxc_wait_for_edid(lt9611uxc);
 	if (ret < 0) {
 		dev_err(lt9611uxc->dev, "wait for EDID failed: %d\n", ret);
-		return ERR_PTR(ret);
+		return NULL;
+	} else if (ret == 0) {
+		dev_err(lt9611uxc->dev, "wait for EDID timeout\n");
+		return NULL;
 	}
 
 	return drm_do_get_edid(connector, lt9611uxc_get_edid_block, lt9611uxc);
@@ -926,6 +956,8 @@ retry:
 	lt9611uxc->fw_version = ret;
 
 	init_waitqueue_head(&lt9611uxc->wq);
+	INIT_WORK(&lt9611uxc->work, lt9611uxc_hpd_work);
+
 	ret = devm_request_threaded_irq(dev, client->irq, NULL,
 					lt9611uxc_irq_thread_handler,
 					IRQF_ONESHOT, "lt9611uxc", lt9611uxc);
@@ -962,6 +994,7 @@ static int lt9611uxc_remove(struct i2c_client *client)
 	struct lt9611uxc *lt9611uxc = i2c_get_clientdata(client);
 
 	disable_irq(client->irq);
+	flush_scheduled_work();
 	lt9611uxc_audio_exit(lt9611uxc);
 	drm_bridge_remove(&lt9611uxc->bridge);
 
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index 1a0ddcf104ac..47ced8bf7e36 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -2987,7 +2987,7 @@ int drm_atomic_helper_set_config(struct drm_mode_set *set,
 
 	ret = handle_conflicting_encoders(state, true);
 	if (ret)
-		return ret;
+		goto fail;
 
 	ret = drm_atomic_commit(state);
 
diff --git a/drivers/gpu/drm/drm_crtc_helper_internal.h b/drivers/gpu/drm/drm_crtc_helper_internal.h
index 25ce42e79995..61e09f8a8d0f 100644
--- a/drivers/gpu/drm/drm_crtc_helper_internal.h
+++ b/drivers/gpu/drm/drm_crtc_helper_internal.h
@@ -32,16 +32,6 @@
 #include <drm/drm_encoder.h>
 #include <drm/drm_modes.h>
 
-/* drm_fb_helper.c */
-#ifdef CONFIG_DRM_FBDEV_EMULATION
-int drm_fb_helper_modinit(void);
-#else
-static inline int drm_fb_helper_modinit(void)
-{
-	return 0;
-}
-#endif
-
 /* drm_dp_aux_dev.c */
 #ifdef CONFIG_DRM_DP_AUX_CHARDEV
 int drm_dp_aux_dev_init(void);
diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c
index 3c5b04abb66f..eedbb48815b7 100644
--- a/drivers/gpu/drm/drm_dp_helper.c
+++ b/drivers/gpu/drm/drm_dp_helper.c
@@ -1236,7 +1236,7 @@ bool drm_dp_read_sink_count_cap(struct drm_connector *connector,
 	return connector->connector_type != DRM_MODE_CONNECTOR_eDP &&
 		dpcd[DP_DPCD_REV] >= DP_DPCD_REV_11 &&
 		dpcd[DP_DOWNSTREAMPORT_PRESENT] & DP_DWN_STRM_PORT_PRESENT &&
-		!drm_dp_has_quirk(desc, 0, DP_DPCD_QUIRK_NO_SINK_COUNT);
+		!drm_dp_has_quirk(desc, DP_DPCD_QUIRK_NO_SINK_COUNT);
 }
 EXPORT_SYMBOL(drm_dp_read_sink_count_cap);
 
@@ -1957,87 +1957,6 @@ drm_dp_get_quirks(const struct drm_dp_dpcd_ident *ident, bool is_branch)
 #undef DEVICE_ID_ANY
 #undef DEVICE_ID
 
-struct edid_quirk {
-	u8 mfg_id[2];
-	u8 prod_id[2];
-	u32 quirks;
-};
-
-#define MFG(first, second) { (first), (second) }
-#define PROD_ID(first, second) { (first), (second) }
-
-/*
- * Some devices have unreliable OUIDs where they don't set the device ID
- * correctly, and as a result we need to use the EDID for finding additional
- * DP quirks in such cases.
- */
-static const struct edid_quirk edid_quirk_list[] = {
-	/* Optional 4K AMOLED panel in the ThinkPad X1 Extreme 2nd Generation
-	 * only supports DPCD backlight controls
-	 */
-	{ MFG(0x4c, 0x83), PROD_ID(0x41, 0x41), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-	/*
-	 * Some Dell CML 2020 systems have panels support both AUX and PWM
-	 * backlight control, and some only support AUX backlight control. All
-	 * said panels start up in AUX mode by default, and we don't have any
-	 * support for disabling HDR mode on these panels which would be
-	 * required to switch to PWM backlight control mode (plus, I'm not
-	 * even sure we want PWM backlight controls over DPCD backlight
-	 * controls anyway...). Until we have a better way of detecting these,
-	 * force DPCD backlight mode on all of them.
-	 */
-	{ MFG(0x06, 0xaf), PROD_ID(0x9b, 0x32), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-	{ MFG(0x06, 0xaf), PROD_ID(0xeb, 0x41), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-	{ MFG(0x4d, 0x10), PROD_ID(0xc7, 0x14), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-	{ MFG(0x4d, 0x10), PROD_ID(0xe6, 0x14), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-	{ MFG(0x4c, 0x83), PROD_ID(0x47, 0x41), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-	{ MFG(0x09, 0xe5), PROD_ID(0xde, 0x08), BIT(DP_QUIRK_FORCE_DPCD_BACKLIGHT) },
-};
-
-#undef MFG
-#undef PROD_ID
-
-/**
- * drm_dp_get_edid_quirks() - Check the EDID of a DP device to find additional
- * DP-specific quirks
- * @edid: The EDID to check
- *
- * While OUIDs are meant to be used to recognize a DisplayPort device, a lot
- * of manufacturers don't seem to like following standards and neglect to fill
- * the dev-ID in, making it impossible to only use OUIDs for determining
- * quirks in some cases. This function can be used to check the EDID and look
- * up any additional DP quirks. The bits returned by this function correspond
- * to the quirk bits in &drm_dp_quirk.
- *
- * Returns: a bitmask of quirks, if any. The driver can check this using
- * drm_dp_has_quirk().
- */
-u32 drm_dp_get_edid_quirks(const struct edid *edid)
-{
-	const struct edid_quirk *quirk;
-	u32 quirks = 0;
-	int i;
-
-	if (!edid)
-		return 0;
-
-	for (i = 0; i < ARRAY_SIZE(edid_quirk_list); i++) {
-		quirk = &edid_quirk_list[i];
-		if (memcmp(quirk->mfg_id, edid->mfg_id,
-			   sizeof(edid->mfg_id)) == 0 &&
-		    memcmp(quirk->prod_id, edid->prod_code,
-			   sizeof(edid->prod_code)) == 0)
-			quirks |= quirk->quirks;
-	}
-
-	DRM_DEBUG_KMS("DP sink: EDID mfg %*phD prod-ID %*phD quirks: 0x%04x\n",
-		      (int)sizeof(edid->mfg_id), edid->mfg_id,
-		      (int)sizeof(edid->prod_code), edid->prod_code, quirks);
-
-	return quirks;
-}
-EXPORT_SYMBOL(drm_dp_get_edid_quirks);
-
 /**
  * drm_dp_read_desc - read sink/branch descriptor from DPCD
  * @aux: DisplayPort AUX channel
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index f1068c5bcef8..42a0c6888c33 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -3635,14 +3635,26 @@ static int drm_dp_send_up_ack_reply(struct drm_dp_mst_topology_mgr *mgr,
 	return 0;
 }
 
-static int drm_dp_get_vc_payload_bw(u8 dp_link_bw, u8  dp_link_count)
+/**
+ * drm_dp_get_vc_payload_bw - get the VC payload BW for an MST link
+ * @link_rate: link rate in 10kbits/s units
+ * @link_lane_count: lane count
+ *
+ * Calculate the total bandwidth of a MultiStream Transport link. The returned
+ * value is in units of PBNs/(timeslots/1 MTP). This value can be used to
+ * convert the number of PBNs required for a given stream to the number of
+ * timeslots this stream requires in each MTP.
+ */
+int drm_dp_get_vc_payload_bw(int link_rate, int link_lane_count)
 {
-	if (dp_link_bw == 0 || dp_link_count == 0)
-		DRM_DEBUG_KMS("invalid link bandwidth in DPCD: %x (link count: %d)\n",
-			      dp_link_bw, dp_link_count);
+	if (link_rate == 0 || link_lane_count == 0)
+		DRM_DEBUG_KMS("invalid link rate/lane count: (%d / %d)\n",
+			      link_rate, link_lane_count);
 
-	return dp_link_bw * dp_link_count / 2;
+	/* See DP v2.0 2.6.4.2, VCPayload_Bandwidth_for_OneTimeSlotPer_MTP_Allocation */
+	return link_rate * link_lane_count / 54000;
 }
+EXPORT_SYMBOL(drm_dp_get_vc_payload_bw);
 
 /**
  * drm_dp_read_mst_cap() - check whether or not a sink supports MST
@@ -3698,7 +3710,7 @@ int drm_dp_mst_topology_mgr_set_mst(struct drm_dp_mst_topology_mgr *mgr, bool ms
 			goto out_unlock;
 		}
 
-		mgr->pbn_div = drm_dp_get_vc_payload_bw(mgr->dpcd[1],
+		mgr->pbn_div = drm_dp_get_vc_payload_bw(drm_dp_bw_code_to_link_rate(mgr->dpcd[1]),
 							mgr->dpcd[2] & DP_MAX_LANE_COUNT_MASK);
 		if (mgr->pbn_div == 0) {
 			ret = -EINVAL;
@@ -4218,6 +4230,7 @@ drm_dp_mst_detect_port(struct drm_connector *connector,
 
 	switch (port->pdt) {
 	case DP_PEER_DEVICE_NONE:
+		break;
 	case DP_PEER_DEVICE_MST_BRANCHING:
 		if (!port->mcs)
 			ret = connector_status_connected;
@@ -5834,8 +5847,7 @@ struct drm_dp_aux *drm_dp_mst_dsc_aux_for_port(struct drm_dp_mst_port *port)
 	if (drm_dp_read_desc(port->mgr->aux, &desc, true))
 		return NULL;
 
-	if (drm_dp_has_quirk(&desc, 0,
-			     DP_DPCD_QUIRK_DSC_WITHOUT_VIRTUAL_DPCD) &&
+	if (drm_dp_has_quirk(&desc, DP_DPCD_QUIRK_DSC_WITHOUT_VIRTUAL_DPCD) &&
 	    port->mgr->dpcd[DP_DPCD_REV] >= DP_DPCD_REV_14 &&
 	    port->parent == port->mgr->mst_primary) {
 		u8 downstreamport;
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index a44c3a438059..f6baa2046124 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -2048,7 +2048,7 @@ static void drm_fbdev_cleanup(struct drm_fb_helper *fb_helper)
 
 	if (shadow)
 		vfree(shadow);
-	else
+	else if (fb_helper->buffer)
 		drm_client_buffer_vunmap(fb_helper->buffer);
 
 	drm_client_framebuffer_delete(fb_helper->buffer);
@@ -2514,24 +2514,3 @@ void drm_fbdev_generic_setup(struct drm_device *dev,
 	drm_client_register(&fb_helper->client);
 }
 EXPORT_SYMBOL(drm_fbdev_generic_setup);
-
-/* The Kconfig DRM_KMS_HELPER selects FRAMEBUFFER_CONSOLE (if !EXPERT)
- * but the module doesn't depend on any fb console symbols.  At least
- * attempt to load fbcon to avoid leaving the system without a usable console.
- */
-int __init drm_fb_helper_modinit(void)
-{
-#if defined(CONFIG_FRAMEBUFFER_CONSOLE_MODULE) && !defined(CONFIG_EXPERT)
-	const char name[] = "fbcon";
-	struct module *fbcon;
-
-	mutex_lock(&module_mutex);
-	fbcon = find_module(name);
-	mutex_unlock(&module_mutex);
-
-	if (!fbcon)
-		request_module_nowait(name);
-#endif
-	return 0;
-}
-EXPORT_SYMBOL(drm_fb_helper_modinit);
diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_gem_shmem_helper.c
index c8a6547a1757..6d625cee7a6a 100644
--- a/drivers/gpu/drm/drm_gem_shmem_helper.c
+++ b/drivers/gpu/drm/drm_gem_shmem_helper.c
@@ -526,14 +526,28 @@ static vm_fault_t drm_gem_shmem_fault(struct vm_fault *vmf)
 	struct drm_gem_object *obj = vma->vm_private_data;
 	struct drm_gem_shmem_object *shmem = to_drm_gem_shmem_obj(obj);
 	loff_t num_pages = obj->size >> PAGE_SHIFT;
+	vm_fault_t ret;
 	struct page *page;
+	pgoff_t page_offset;
 
-	if (vmf->pgoff >= num_pages || WARN_ON_ONCE(!shmem->pages))
-		return VM_FAULT_SIGBUS;
+	/* We don't use vmf->pgoff since that has the fake offset */
+	page_offset = (vmf->address - vma->vm_start) >> PAGE_SHIFT;
 
-	page = shmem->pages[vmf->pgoff];
+	mutex_lock(&shmem->pages_lock);
+
+	if (page_offset >= num_pages ||
+	    WARN_ON_ONCE(!shmem->pages) ||
+	    shmem->madv < 0) {
+		ret = VM_FAULT_SIGBUS;
+	} else {
+		page = shmem->pages[page_offset];
+
+		ret = vmf_insert_page(vma, vmf->address, page);
+	}
 
-	return vmf_insert_page(vma, vmf->address, page);
+	mutex_unlock(&shmem->pages_lock);
+
+	return ret;
 }
 
 static void drm_gem_shmem_vm_open(struct vm_area_struct *vma)
@@ -582,9 +596,6 @@ int drm_gem_shmem_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
 	struct drm_gem_shmem_object *shmem;
 	int ret;
 
-	/* Remove the fake offset */
-	vma->vm_pgoff -= drm_vma_node_start(&obj->vma_node);
-
 	if (obj->import_attach) {
 		/* Drop the reference drm_gem_mmap_obj() acquired.*/
 		drm_gem_object_put(obj);
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index d7519b814f1b..2b7c3a07956d 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -385,9 +385,16 @@ static int drm_gem_vram_kmap_locked(struct drm_gem_vram_object *gbo,
 	if (gbo->vmap_use_count > 0)
 		goto out;
 
-	ret = ttm_bo_vmap(&gbo->bo, &gbo->map);
-	if (ret)
-		return ret;
+	/*
+	 * VRAM helpers unmap the BO only on demand. So the previous
+	 * page mapping might still be around. Only vmap if the there's
+	 * no mapping present.
+	 */
+	if (dma_buf_map_is_null(&gbo->map)) {
+		ret = ttm_bo_vmap(&gbo->bo, &gbo->map);
+		if (ret)
+			return ret;
+	}
 
 out:
 	++gbo->vmap_use_count;
@@ -573,6 +580,7 @@ static void drm_gem_vram_bo_driver_move_notify(struct drm_gem_vram_object *gbo)
 		return;
 
 	ttm_bo_vunmap(bo, &gbo->map);
+	dma_buf_map_clear(&gbo->map); /* explicitly clear mapping for next vmap call */
 }
 
 static int drm_gem_vram_bo_driver_move(struct drm_gem_vram_object *gbo,
diff --git a/drivers/gpu/drm/drm_kms_helper_common.c b/drivers/gpu/drm/drm_kms_helper_common.c
index 221a8528c993..f933da1656eb 100644
--- a/drivers/gpu/drm/drm_kms_helper_common.c
+++ b/drivers/gpu/drm/drm_kms_helper_common.c
@@ -64,19 +64,18 @@ MODULE_PARM_DESC(edid_firmware,
 
 static int __init drm_kms_helper_init(void)
 {
-	int ret;
-
-	/* Call init functions from specific kms helpers here */
-	ret = drm_fb_helper_modinit();
-	if (ret < 0)
-		goto out;
-
-	ret = drm_dp_aux_dev_init();
-	if (ret < 0)
-		goto out;
-
-out:
-	return ret;
+	/*
+	 * The Kconfig DRM_KMS_HELPER selects FRAMEBUFFER_CONSOLE (if !EXPERT)
+	 * but the module doesn't depend on any fb console symbols.  At least
+	 * attempt to load fbcon to avoid leaving the system without a usable
+	 * console.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_FBDEV_EMULATION) &&
+	    IS_MODULE(CONFIG_FRAMEBUFFER_CONSOLE) &&
+	    !IS_ENABLED(CONFIG_EXPERT))
+		request_module_nowait("fbcon");
+
+	return drm_dp_aux_dev_init();
 }
 
 static void __exit drm_kms_helper_exit(void)
diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c
index 77c5698153f8..0dd43882fe7c 100644
--- a/drivers/gpu/drm/drm_plane.c
+++ b/drivers/gpu/drm/drm_plane.c
@@ -1305,7 +1305,14 @@ retry:
 	if (ret)
 		goto out;
 
-	if (old_fb->format != fb->format) {
+	/*
+	 * Only check the FOURCC format code, excluding modifiers. This is
+	 * enough for all legacy drivers. Atomic drivers have their own
+	 * checks in their ->atomic_check implementation, which will
+	 * return -EINVAL if any hw or driver constraint is violated due
+	 * to modifier changes.
+	 */
+	if (old_fb->format->format != fb->format->format) {
 		DRM_DEBUG_KMS("Page flip is not allowed to change frame buffer format.\n");
 		ret = -EINVAL;
 		goto out;
diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c
index 9ecfec1e5567..6231a8214c25 100644
--- a/drivers/gpu/drm/drm_syncobj.c
+++ b/drivers/gpu/drm/drm_syncobj.c
@@ -397,19 +397,18 @@ int drm_syncobj_find_fence(struct drm_file *file_private,
 	}
 
 	*fence = drm_syncobj_fence_get(syncobj);
-	drm_syncobj_put(syncobj);
 
 	if (*fence) {
 		ret = dma_fence_chain_find_seqno(fence, point);
 		if (!ret)
-			return 0;
+			goto out;
 		dma_fence_put(*fence);
 	} else {
 		ret = -EINVAL;
 	}
 
 	if (!(flags & DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT))
-		return ret;
+		goto out;
 
 	memset(&wait, 0, sizeof(wait));
 	wait.task = current;
@@ -441,6 +440,9 @@ int drm_syncobj_find_fence(struct drm_file *file_private,
 	if (wait.node.next)
 		drm_syncobj_remove_wait(syncobj, &wait);
 
+out:
+	drm_syncobj_put(syncobj);
+
 	return ret;
 }
 EXPORT_SYMBOL(drm_syncobj_find_fence);
diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig
index 951d5f708e92..6a251e3aa779 100644
--- a/drivers/gpu/drm/exynos/Kconfig
+++ b/drivers/gpu/drm/exynos/Kconfig
@@ -89,7 +89,6 @@ comment "Sub-drivers"
 config DRM_EXYNOS_G2D
 	bool "G2D"
 	depends on VIDEO_SAMSUNG_S5P_G2D=n || COMPILE_TEST
-	select FRAME_VECTOR
 	help
 	  Choose this option if you want to use Exynos G2D for DRM.
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index 967a5cdc120e..1e0c5a7f206e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -205,7 +205,8 @@ struct g2d_cmdlist_userptr {
 	dma_addr_t		dma_addr;
 	unsigned long		userptr;
 	unsigned long		size;
-	struct frame_vector	*vec;
+	struct page		**pages;
+	unsigned int		npages;
 	struct sg_table		*sgt;
 	atomic_t		refcount;
 	bool			in_pool;
@@ -378,7 +379,6 @@ static void g2d_userptr_put_dma_addr(struct g2d_data *g2d,
 					bool force)
 {
 	struct g2d_cmdlist_userptr *g2d_userptr = obj;
-	struct page **pages;
 
 	if (!obj)
 		return;
@@ -398,15 +398,9 @@ out:
 	dma_unmap_sgtable(to_dma_dev(g2d->drm_dev), g2d_userptr->sgt,
 			  DMA_BIDIRECTIONAL, 0);
 
-	pages = frame_vector_pages(g2d_userptr->vec);
-	if (!IS_ERR(pages)) {
-		int i;
-
-		for (i = 0; i < frame_vector_count(g2d_userptr->vec); i++)
-			set_page_dirty_lock(pages[i]);
-	}
-	put_vaddr_frames(g2d_userptr->vec);
-	frame_vector_destroy(g2d_userptr->vec);
+	unpin_user_pages_dirty_lock(g2d_userptr->pages, g2d_userptr->npages,
+				    true);
+	kvfree(g2d_userptr->pages);
 
 	if (!g2d_userptr->out_of_list)
 		list_del_init(&g2d_userptr->list);
@@ -474,35 +468,35 @@ static dma_addr_t *g2d_userptr_get_dma_addr(struct g2d_data *g2d,
 	offset = userptr & ~PAGE_MASK;
 	end = PAGE_ALIGN(userptr + size);
 	npages = (end - start) >> PAGE_SHIFT;
-	g2d_userptr->vec = frame_vector_create(npages);
-	if (!g2d_userptr->vec) {
+	g2d_userptr->pages = kvmalloc_array(npages, sizeof(*g2d_userptr->pages),
+					    GFP_KERNEL);
+	if (!g2d_userptr->pages) {
 		ret = -ENOMEM;
 		goto err_free;
 	}
 
-	ret = get_vaddr_frames(start, npages, FOLL_FORCE | FOLL_WRITE,
-		g2d_userptr->vec);
+	ret = pin_user_pages_fast(start, npages,
+				  FOLL_FORCE | FOLL_WRITE | FOLL_LONGTERM,
+				  g2d_userptr->pages);
 	if (ret != npages) {
 		DRM_DEV_ERROR(g2d->dev,
 			      "failed to get user pages from userptr.\n");
 		if (ret < 0)
-			goto err_destroy_framevec;
-		ret = -EFAULT;
-		goto err_put_framevec;
-	}
-	if (frame_vector_to_pages(g2d_userptr->vec) < 0) {
+			goto err_destroy_pages;
+		npages = ret;
 		ret = -EFAULT;
-		goto err_put_framevec;
+		goto err_unpin_pages;
 	}
+	g2d_userptr->npages = npages;
 
 	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
 	if (!sgt) {
 		ret = -ENOMEM;
-		goto err_put_framevec;
+		goto err_unpin_pages;
 	}
 
 	ret = sg_alloc_table_from_pages(sgt,
-					frame_vector_pages(g2d_userptr->vec),
+					g2d_userptr->pages,
 					npages, offset, size, GFP_KERNEL);
 	if (ret < 0) {
 		DRM_DEV_ERROR(g2d->dev, "failed to get sgt from pages.\n");
@@ -538,11 +532,11 @@ err_sg_free_table:
 err_free_sgt:
 	kfree(sgt);
 
-err_put_framevec:
-	put_vaddr_frames(g2d_userptr->vec);
+err_unpin_pages:
+	unpin_user_pages(g2d_userptr->pages, npages);
 
-err_destroy_framevec:
-	frame_vector_destroy(g2d_userptr->vec);
+err_destroy_pages:
+	kvfree(g2d_userptr->pages);
 
 err_free:
 	kfree(g2d_userptr);
diff --git a/drivers/gpu/drm/gma500/oaktrail_device.c b/drivers/gpu/drm/gma500/oaktrail_device.c
index 9650588c7223..454156fcbec7 100644
--- a/drivers/gpu/drm/gma500/oaktrail_device.c
+++ b/drivers/gpu/drm/gma500/oaktrail_device.c
@@ -10,9 +10,6 @@
 #include <linux/dmi.h>
 #include <linux/module.h>
 
-#include <asm/intel-mid.h>
-#include <asm/intel_scu_ipc.h>
-
 #include <drm/drm.h>
 
 #include "intel_bios.h"
diff --git a/drivers/gpu/drm/gma500/psb_drv.h b/drivers/gpu/drm/gma500/psb_drv.h
index 697f64e0f2f4..49afa577d442 100644
--- a/drivers/gpu/drm/gma500/psb_drv.h
+++ b/drivers/gpu/drm/gma500/psb_drv.h
@@ -386,6 +386,8 @@ struct psb_ops;
 
 #define PSB_NUM_PIPE		3
 
+struct intel_scu_ipc_dev;
+
 struct drm_psb_private {
 	struct drm_device *dev;
 	struct pci_dev *aux_pdev; /* Currently only used by mrst */
@@ -525,6 +527,7 @@ struct drm_psb_private {
 	 * Used for modifying backlight from
 	 * xrandr -- consider removing and using HAL instead
 	 */
+	struct intel_scu_ipc_dev *scu;
 	struct backlight_device *backlight_device;
 	struct drm_property *backlight_property;
 	bool backlight_enabled;
diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug
index 25cd9788a4d5..72a38f28393f 100644
--- a/drivers/gpu/drm/i915/Kconfig.debug
+++ b/drivers/gpu/drm/i915/Kconfig.debug
@@ -19,6 +19,8 @@ config DRM_I915_WERROR
 config DRM_I915_DEBUG
 	bool "Enable additional driver debugging"
 	depends on DRM_I915
+	depends on EXPERT # only for developers
+	depends on !COMPILE_TEST # never built by robots
 	select DEBUG_FS
 	select PREEMPT_COUNT
 	select I2C_CHARDEV
@@ -31,10 +33,13 @@ config DRM_I915_DEBUG
 	select DRM_DEBUG_SELFTEST
 	select DMABUF_SELFTESTS
 	select SW_SYNC # signaling validation framework (igt/syncobj*)
+	select DRM_I915_WERROR
+	select DRM_I915_DEBUG_GEM
+	select DRM_I915_DEBUG_GEM_ONCE
+	select DRM_I915_DEBUG_MMIO
+	select DRM_I915_DEBUG_RUNTIME_PM
 	select DRM_I915_SW_FENCE_DEBUG_OBJECTS
 	select DRM_I915_SELFTEST
-	select DRM_I915_DEBUG_RUNTIME_PM
-	select DRM_I915_DEBUG_MMIO
 	default n
 	help
 	  Choose this option to turn on extra driver debugging that may affect
@@ -69,6 +74,21 @@ config DRM_I915_DEBUG_GEM
 
 	  If in doubt, say "N".
 
+config DRM_I915_DEBUG_GEM_ONCE
+	bool "Make a GEM debug failure fatal"
+	default n
+	depends on DRM_I915_DEBUG_GEM
+	help
+	  During development, we often only want the very first failure
+	  as that would otherwise be lost in the deluge of subsequent
+	  failures. However, more casual testers may not want to trigger
+	  a hard BUG_ON and hope that the system remains sufficiently usable
+	  to capture a bug report in situ.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
 config DRM_I915_ERRLOG_GEM
 	bool "Insert extra logging (very verbose) for common GEM errors"
 	default n
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index c1fa5a9310f9..2385a7505f5d 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -21,7 +21,6 @@ subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable)
 subdir-ccflags-y += $(call cc-disable-warning, sign-compare)
 subdir-ccflags-y += $(call cc-disable-warning, sometimes-uninitialized)
 subdir-ccflags-y += $(call cc-disable-warning, initializer-overrides)
-subdir-ccflags-y += $(call cc-disable-warning, uninitialized)
 subdir-ccflags-y += $(call cc-disable-warning, frame-address)
 subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werror
 
@@ -38,6 +37,7 @@ i915-y += i915_drv.o \
 	  i915_config.o \
 	  i915_irq.o \
 	  i915_getparam.o \
+	  i915_mitigations.o \
 	  i915_params.o \
 	  i915_pci.o \
 	  i915_scatterlist.o \
@@ -58,6 +58,7 @@ i915-y += i915_drv.o \
 
 # core library code
 i915-y += \
+	dma_resv_utils.o \
 	i915_memcpy.o \
 	i915_mm.o \
 	i915_sw_fence.o \
@@ -82,6 +83,7 @@ gt-y += \
 	gt/gen6_engine_cs.o \
 	gt/gen6_ppgtt.o \
 	gt/gen7_renderclear.o \
+	gt/gen8_engine_cs.o \
 	gt/gen8_ppgtt.o \
 	gt/intel_breadcrumbs.o \
 	gt/intel_context.o \
@@ -91,6 +93,7 @@ gt-y += \
 	gt/intel_engine_heartbeat.o \
 	gt/intel_engine_pm.o \
 	gt/intel_engine_user.o \
+	gt/intel_execlists_submission.o \
 	gt/intel_ggtt.o \
 	gt/intel_ggtt_fencing.o \
 	gt/intel_gt.o \
@@ -106,6 +109,7 @@ gt-y += \
 	gt/intel_mocs.o \
 	gt/intel_ppgtt.o \
 	gt/intel_rc6.o \
+	gt/intel_region_lmem.o \
 	gt/intel_renderstate.o \
 	gt/intel_reset.o \
 	gt/intel_ring.o \
@@ -131,6 +135,7 @@ gem-y += \
 	gem/i915_gem_clflush.o \
 	gem/i915_gem_client_blt.o \
 	gem/i915_gem_context.o \
+	gem/i915_gem_create.o \
 	gem/i915_gem_dmabuf.o \
 	gem/i915_gem_domain.o \
 	gem/i915_gem_execbuffer.o \
@@ -166,7 +171,6 @@ i915-y += \
 	  i915_scheduler.o \
 	  i915_trace_points.o \
 	  i915_vma.o \
-	  intel_region_lmem.o \
 	  intel_wopcm.o
 
 # general-purpose microcontroller (GuC) support
@@ -196,14 +200,17 @@ i915-y += \
 	display/intel_color.o \
 	display/intel_combo_phy.o \
 	display/intel_connector.o \
+	display/intel_crtc.o \
 	display/intel_csr.o \
 	display/intel_cursor.o \
 	display/intel_display.o \
 	display/intel_display_power.o \
 	display/intel_dpio_phy.o \
+	display/intel_dpll.o \
 	display/intel_dpll_mgr.o \
 	display/intel_dsb.o \
 	display/intel_fbc.o \
+	display/intel_fdi.o \
 	display/intel_fifo_underrun.o \
 	display/intel_frontbuffer.o \
 	display/intel_global_state.o \
@@ -235,6 +242,7 @@ i915-y += \
 	display/intel_crt.o \
 	display/intel_ddi.o \
 	display/intel_dp.o \
+	display/intel_dp_aux.o \
 	display/intel_dp_aux_backlight.o \
 	display/intel_dp_hdcp.o \
 	display/intel_dp_link_training.o \
@@ -248,9 +256,11 @@ i915-y += \
 	display/intel_lspcon.o \
 	display/intel_lvds.o \
 	display/intel_panel.o \
+	display/intel_pps.o \
 	display/intel_sdvo.o \
 	display/intel_tv.o \
 	display/intel_vdsc.o \
+	display/intel_vrr.o \
 	display/vlv_dsi.o \
 	display/vlv_dsi_pll.o
 
@@ -285,17 +295,9 @@ obj-$(CONFIG_DRM_I915_GVT_KVMGT) += gvt/kvmgt.o
 
 # exclude some broken headers from the test coverage
 no-header-test := \
-	display/intel_vbt_defs.h \
-	gvt/execlist.h \
-	gvt/fb_decoder.h \
-	gvt/gtt.h \
-	gvt/gvt.h \
-	gvt/interrupt.h \
-	gvt/mmio_context.h \
-	gvt/mpt.h \
-	gvt/scheduler.h
+	display/intel_vbt_defs.h
 
-extra-$(CONFIG_DRM_I915_WERROR) += \
+always-$(CONFIG_DRM_I915_WERROR) += \
 	$(patsubst %.h,%.hdrtest, $(filter-out $(no-header-test), \
 		$(shell cd $(srctree)/$(src) && find * -name '*.h')))
 
diff --git a/drivers/gpu/drm/i915/display/i9xx_plane.c b/drivers/gpu/drm/i915/display/i9xx_plane.c
index b78985c855a5..e3e69e6cef65 100644
--- a/drivers/gpu/drm/i915/display/i9xx_plane.c
+++ b/drivers/gpu/drm/i915/display/i9xx_plane.c
@@ -256,6 +256,33 @@ int i9xx_check_plane_surface(struct intel_plane_state *plane_state)
 		offset = 0;
 
 	/*
+	 * When using an X-tiled surface the plane starts to
+	 * misbehave if the x offset + width exceeds the stride.
+	 * hsw/bdw: underrun galore
+	 * ilk/snb/ivb: wrap to the next tile row mid scanout
+	 * i965/g4x: so far appear immune to this
+	 * vlv/chv: TODO check
+	 *
+	 * Linear surfaces seem to work just fine, even on hsw/bdw
+	 * despite them not using the linear offset anymore.
+	 */
+	if (INTEL_GEN(dev_priv) >= 4 && fb->modifier == I915_FORMAT_MOD_X_TILED) {
+		u32 alignment = intel_surf_alignment(fb, 0);
+		int cpp = fb->format->cpp[0];
+
+		while ((src_x + src_w) * cpp > plane_state->color_plane[0].stride) {
+			if (offset == 0) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "Unable to find suitable display surface offset due to X-tiling\n");
+				return -EINVAL;
+			}
+
+			offset = intel_plane_adjust_aligned_offset(&src_x, &src_y, plane_state, 0,
+								   offset, offset - alignment);
+		}
+	}
+
+	/*
 	 * Put the final coordinates back so that the src
 	 * coordinate checks will see the right values.
 	 */
@@ -276,6 +303,13 @@ int i9xx_check_plane_surface(struct intel_plane_state *plane_state)
 		}
 	}
 
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		drm_WARN_ON(&dev_priv->drm, src_x > 8191 || src_y > 4095);
+	} else if (INTEL_GEN(dev_priv) >= 4 &&
+		   fb->modifier == I915_FORMAT_MOD_X_TILED) {
+		drm_WARN_ON(&dev_priv->drm, src_x > 4095 || src_y > 4095);
+	}
+
 	plane_state->color_plane[0].offset = offset;
 	plane_state->color_plane[0].x = src_x;
 	plane_state->color_plane[0].y = src_y;
@@ -488,6 +522,129 @@ static void i9xx_disable_plane(struct intel_plane *plane,
 	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
 }
 
+static void
+g4x_primary_async_flip(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state,
+		       const struct intel_plane_state *plane_state,
+		       bool async_flip)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	u32 dspcntr = plane_state->ctl | i9xx_plane_ctl_crtc(crtc_state);
+	u32 dspaddr_offset = plane_state->color_plane[0].offset;
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	unsigned long irqflags;
+
+	if (async_flip)
+		dspcntr |= DISPPLANE_ASYNC_FLIP;
+
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+	intel_de_write_fw(dev_priv, DSPCNTR(i9xx_plane), dspcntr);
+	intel_de_write_fw(dev_priv, DSPSURF(i9xx_plane),
+			  intel_plane_ggtt_offset(plane_state) + dspaddr_offset);
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+}
+
+static void
+vlv_primary_async_flip(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state,
+		       const struct intel_plane_state *plane_state,
+		       bool async_flip)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	u32 dspaddr_offset = plane_state->color_plane[0].offset;
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+	intel_de_write_fw(dev_priv, DSPADDR_VLV(i9xx_plane),
+			  intel_plane_ggtt_offset(plane_state) + dspaddr_offset);
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+}
+
+static void
+bdw_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_enable_pipe_irq(i915, pipe, GEN8_PIPE_PRIMARY_FLIP_DONE);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+bdw_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_disable_pipe_irq(i915, pipe, GEN8_PIPE_PRIMARY_FLIP_DONE);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ivb_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_enable_display_irq(i915, DE_PLANE_FLIP_DONE_IVB(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ivb_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_disable_display_irq(i915, DE_PLANE_FLIP_DONE_IVB(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ilk_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_enable_display_irq(i915, DE_PLANE_FLIP_DONE(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ilk_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_disable_display_irq(i915, DE_PLANE_FLIP_DONE(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+vlv_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	i915_enable_pipestat(i915, pipe, PLANE_FLIP_DONE_INT_STATUS_VLV);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+vlv_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	i915_disable_pipestat(i915, pipe, PLANE_FLIP_DONE_INT_STATUS_VLV);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
 static bool i9xx_plane_get_hw_state(struct intel_plane *plane,
 				    enum pipe *pipe)
 {
@@ -523,21 +680,56 @@ static bool i9xx_plane_get_hw_state(struct intel_plane *plane,
 	return ret;
 }
 
+static unsigned int
+hsw_primary_max_stride(struct intel_plane *plane,
+		       u32 pixel_format, u64 modifier,
+		       unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 8k pixels to guarantee OFFSET.x doesn't get too big. */
+	return min(8192 * cpp, 32 * 1024);
+}
+
+static unsigned int
+ilk_primary_max_stride(struct intel_plane *plane,
+		       u32 pixel_format, u64 modifier,
+		       unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 4k pixels to guarantee TILEOFF.x doesn't get too big. */
+	if (modifier == I915_FORMAT_MOD_X_TILED)
+		return min(4096 * cpp, 32 * 1024);
+	else
+		return 32 * 1024;
+}
+
 unsigned int
+i965_plane_max_stride(struct intel_plane *plane,
+		      u32 pixel_format, u64 modifier,
+		      unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 4k pixels to guarantee TILEOFF.x doesn't get too big. */
+	if (modifier == I915_FORMAT_MOD_X_TILED)
+		return min(4096 * cpp, 16 * 1024);
+	else
+		return 32 * 1024;
+}
+
+static unsigned int
 i9xx_plane_max_stride(struct intel_plane *plane,
 		      u32 pixel_format, u64 modifier,
 		      unsigned int rotation)
 {
 	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
 
-	if (!HAS_GMCH(dev_priv)) {
-		return 32*1024;
-	} else if (INTEL_GEN(dev_priv) >= 4) {
-		if (modifier == I915_FORMAT_MOD_X_TILED)
-			return 16*1024;
-		else
-			return 32*1024;
-	} else if (INTEL_GEN(dev_priv) >= 3) {
+	if (INTEL_GEN(dev_priv) >= 3) {
 		if (modifier == I915_FORMAT_MOD_X_TILED)
 			return 8*1024;
 		else
@@ -649,12 +841,42 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe)
 	else
 		plane->min_cdclk = i9xx_plane_min_cdclk;
 
-	plane->max_stride = i9xx_plane_max_stride;
+	if (HAS_GMCH(dev_priv)) {
+		if (INTEL_GEN(dev_priv) >= 4)
+			plane->max_stride = i965_plane_max_stride;
+		else
+			plane->max_stride = i9xx_plane_max_stride;
+	} else {
+		if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+			plane->max_stride = hsw_primary_max_stride;
+		else
+			plane->max_stride = ilk_primary_max_stride;
+	}
+
 	plane->update_plane = i9xx_update_plane;
 	plane->disable_plane = i9xx_disable_plane;
 	plane->get_hw_state = i9xx_plane_get_hw_state;
 	plane->check_plane = i9xx_plane_check;
 
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		plane->async_flip = vlv_primary_async_flip;
+		plane->enable_flip_done = vlv_primary_enable_flip_done;
+		plane->disable_flip_done = vlv_primary_disable_flip_done;
+	} else if (IS_BROADWELL(dev_priv)) {
+		plane->need_async_flip_disable_wa = true;
+		plane->async_flip = g4x_primary_async_flip;
+		plane->enable_flip_done = bdw_primary_enable_flip_done;
+		plane->disable_flip_done = bdw_primary_disable_flip_done;
+	} else if (INTEL_GEN(dev_priv) >= 7) {
+		plane->async_flip = g4x_primary_async_flip;
+		plane->enable_flip_done = ivb_primary_enable_flip_done;
+		plane->disable_flip_done = ivb_primary_disable_flip_done;
+	} else if (INTEL_GEN(dev_priv) >= 5) {
+		plane->async_flip = g4x_primary_async_flip;
+		plane->enable_flip_done = ilk_primary_enable_flip_done;
+		plane->disable_flip_done = ilk_primary_disable_flip_done;
+	}
+
 	if (INTEL_GEN(dev_priv) >= 5 || IS_G4X(dev_priv))
 		ret = drm_universal_plane_init(&dev_priv->drm, &plane->base,
 					       0, plane_funcs,
diff --git a/drivers/gpu/drm/i915/display/i9xx_plane.h b/drivers/gpu/drm/i915/display/i9xx_plane.h
index bc2834a62735..ca963c2a8457 100644
--- a/drivers/gpu/drm/i915/display/i9xx_plane.h
+++ b/drivers/gpu/drm/i915/display/i9xx_plane.h
@@ -13,7 +13,7 @@ struct drm_i915_private;
 struct intel_plane;
 struct intel_plane_state;
 
-unsigned int i9xx_plane_max_stride(struct intel_plane *plane,
+unsigned int i965_plane_max_stride(struct intel_plane *plane,
 				   u32 pixel_format, u64 modifier,
 				   unsigned int rotation);
 int i9xx_check_plane_surface(struct intel_plane_state *plane_state);
diff --git a/drivers/gpu/drm/i915/display/intel_atomic_plane.c b/drivers/gpu/drm/i915/display/intel_atomic_plane.c
index b5e1ee99535c..4683f98f7e54 100644
--- a/drivers/gpu/drm/i915/display/intel_atomic_plane.c
+++ b/drivers/gpu/drm/i915/display/intel_atomic_plane.c
@@ -452,7 +452,7 @@ void intel_update_plane(struct intel_plane *plane,
 	trace_intel_update_plane(&plane->base, crtc);
 
 	if (crtc_state->uapi.async_flip && plane->async_flip)
-		plane->async_flip(plane, crtc_state, plane_state);
+		plane->async_flip(plane, crtc_state, plane_state, true);
 	else
 		plane->update_plane(plane, crtc_state, plane_state);
 }
diff --git a/drivers/gpu/drm/i915/display/intel_bios.c b/drivers/gpu/drm/i915/display/intel_bios.c
index 06c3310446a2..987cf509337f 100644
--- a/drivers/gpu/drm/i915/display/intel_bios.c
+++ b/drivers/gpu/drm/i915/display/intel_bios.c
@@ -1623,6 +1623,13 @@ static const u8 icp_ddc_pin_map[] = {
 	[TGL_DDC_BUS_PORT_6] = GMBUS_PIN_14_TC6_TGP,
 };
 
+static const u8 rkl_pch_tgp_ddc_pin_map[] = {
+	[ICL_DDC_BUS_DDI_A] = GMBUS_PIN_1_BXT,
+	[ICL_DDC_BUS_DDI_B] = GMBUS_PIN_2_BXT,
+	[RKL_DDC_BUS_DDI_D] = GMBUS_PIN_9_TC1_ICP,
+	[RKL_DDC_BUS_DDI_E] = GMBUS_PIN_10_TC2_ICP,
+};
+
 static u8 map_ddc_pin(struct drm_i915_private *dev_priv, u8 vbt_pin)
 {
 	const u8 *ddc_pin_map;
@@ -1630,6 +1637,9 @@ static u8 map_ddc_pin(struct drm_i915_private *dev_priv, u8 vbt_pin)
 
 	if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1) {
 		return vbt_pin;
+	} else if (IS_ROCKETLAKE(dev_priv) && INTEL_PCH_TYPE(dev_priv) == PCH_TGP) {
+		ddc_pin_map = rkl_pch_tgp_ddc_pin_map;
+		n_entries = ARRAY_SIZE(rkl_pch_tgp_ddc_pin_map);
 	} else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) {
 		ddc_pin_map = icp_ddc_pin_map;
 		n_entries = ARRAY_SIZE(icp_ddc_pin_map);
diff --git a/drivers/gpu/drm/i915/display/intel_bw.c b/drivers/gpu/drm/i915/display/intel_bw.c
index bd060404d249..4b5a30ac84bc 100644
--- a/drivers/gpu/drm/i915/display/intel_bw.c
+++ b/drivers/gpu/drm/i915/display/intel_bw.c
@@ -20,76 +20,9 @@ struct intel_qgv_point {
 struct intel_qgv_info {
 	struct intel_qgv_point points[I915_NUM_QGV_POINTS];
 	u8 num_points;
-	u8 num_channels;
 	u8 t_bl;
-	enum intel_dram_type dram_type;
 };
 
-static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv,
-					  struct intel_qgv_info *qi)
-{
-	u32 val = 0;
-	int ret;
-
-	ret = sandybridge_pcode_read(dev_priv,
-				     ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
-				     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO,
-				     &val, NULL);
-	if (ret)
-		return ret;
-
-	if (IS_GEN(dev_priv, 12)) {
-		switch (val & 0xf) {
-		case 0:
-			qi->dram_type = INTEL_DRAM_DDR4;
-			break;
-		case 3:
-			qi->dram_type = INTEL_DRAM_LPDDR4;
-			break;
-		case 4:
-			qi->dram_type = INTEL_DRAM_DDR3;
-			break;
-		case 5:
-			qi->dram_type = INTEL_DRAM_LPDDR3;
-			break;
-		default:
-			MISSING_CASE(val & 0xf);
-			break;
-		}
-	} else if (IS_GEN(dev_priv, 11)) {
-		switch (val & 0xf) {
-		case 0:
-			qi->dram_type = INTEL_DRAM_DDR4;
-			break;
-		case 1:
-			qi->dram_type = INTEL_DRAM_DDR3;
-			break;
-		case 2:
-			qi->dram_type = INTEL_DRAM_LPDDR3;
-			break;
-		case 3:
-			qi->dram_type = INTEL_DRAM_LPDDR4;
-			break;
-		default:
-			MISSING_CASE(val & 0xf);
-			break;
-		}
-	} else {
-		MISSING_CASE(INTEL_GEN(dev_priv));
-		qi->dram_type = INTEL_DRAM_LPDDR3; /* Conservative default */
-	}
-
-	qi->num_channels = (val & 0xf0) >> 4;
-	qi->num_points = (val & 0xf00) >> 8;
-
-	if (IS_GEN(dev_priv, 12))
-		qi->t_bl = qi->dram_type == INTEL_DRAM_DDR4 ? 4 : 16;
-	else if (IS_GEN(dev_priv, 11))
-		qi->t_bl = qi->dram_type == INTEL_DRAM_DDR4 ? 4 : 8;
-
-	return 0;
-}
-
 static int icl_pcode_read_qgv_point_info(struct drm_i915_private *dev_priv,
 					 struct intel_qgv_point *sp,
 					 int point)
@@ -139,11 +72,15 @@ int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
 static int icl_get_qgv_points(struct drm_i915_private *dev_priv,
 			      struct intel_qgv_info *qi)
 {
+	const struct dram_info *dram_info = &dev_priv->dram_info;
 	int i, ret;
 
-	ret = icl_pcode_read_mem_global_info(dev_priv, qi);
-	if (ret)
-		return ret;
+	qi->num_points = dram_info->num_qgv_points;
+
+	if (IS_GEN(dev_priv, 12))
+		qi->t_bl = dev_priv->dram_info.type == INTEL_DRAM_DDR4 ? 4 : 16;
+	else if (IS_GEN(dev_priv, 11))
+		qi->t_bl = dev_priv->dram_info.type == INTEL_DRAM_DDR4 ? 4 : 8;
 
 	if (drm_WARN_ON(&dev_priv->drm,
 			qi->num_points > ARRAY_SIZE(qi->points)))
@@ -209,7 +146,7 @@ static int icl_get_bw_info(struct drm_i915_private *dev_priv, const struct intel
 {
 	struct intel_qgv_info qi = {};
 	bool is_y_tile = true; /* assume y tile may be used */
-	int num_channels;
+	int num_channels = dev_priv->dram_info.num_channels;
 	int deinterleave;
 	int ipqdepth, ipqdepthpch;
 	int dclk_max;
@@ -222,7 +159,6 @@ static int icl_get_bw_info(struct drm_i915_private *dev_priv, const struct intel
 			    "Failed to get memory subsystem information, ignoring bandwidth limits");
 		return ret;
 	}
-	num_channels = qi.num_channels;
 
 	deinterleave = DIV_ROUND_UP(num_channels, is_y_tile ? 4 : 2);
 	dclk_max = icl_sagv_max_dclk(&qi);
diff --git a/drivers/gpu/drm/i915/display/intel_color.c b/drivers/gpu/drm/i915/display/intel_color.c
index 172d398081ee..ff7dcb7088bf 100644
--- a/drivers/gpu/drm/i915/display/intel_color.c
+++ b/drivers/gpu/drm/i915/display/intel_color.c
@@ -1485,6 +1485,7 @@ static u32 ivb_csc_mode(const struct intel_crtc_state *crtc_state)
 
 static int ivb_color_check(struct intel_crtc_state *crtc_state)
 {
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	bool limited_color_range = ilk_csc_limited_range(crtc_state);
 	int ret;
 
@@ -1492,6 +1493,13 @@ static int ivb_color_check(struct intel_crtc_state *crtc_state)
 	if (ret)
 		return ret;
 
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB &&
+	    crtc_state->hw.ctm) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "YCBCR and CTM together are not possible\n");
+		return -EINVAL;
+	}
+
 	crtc_state->gamma_enable =
 		(crtc_state->hw.gamma_lut ||
 		 crtc_state->hw.degamma_lut) &&
@@ -1525,12 +1533,20 @@ static u32 glk_gamma_mode(const struct intel_crtc_state *crtc_state)
 
 static int glk_color_check(struct intel_crtc_state *crtc_state)
 {
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int ret;
 
 	ret = check_luts(crtc_state);
 	if (ret)
 		return ret;
 
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB &&
+	    crtc_state->hw.ctm) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "YCBCR and CTM together are not possible\n");
+		return -EINVAL;
+	}
+
 	crtc_state->gamma_enable =
 		crtc_state->hw.gamma_lut &&
 		!crtc_state->c8_planes;
diff --git a/drivers/gpu/drm/i915/display/intel_crtc.c b/drivers/gpu/drm/i915/display/intel_crtc.c
new file mode 100644
index 000000000000..8e77ca7ddf11
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crtc.c
@@ -0,0 +1,324 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include <linux/kernel.h>
+#include <linux/slab.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_plane.h>
+#include <drm/drm_plane_helper.h>
+
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_color.h"
+#include "intel_crtc.h"
+#include "intel_cursor.h"
+#include "intel_display_debugfs.h"
+#include "intel_display_types.h"
+#include "intel_pipe_crc.h"
+#include "intel_sprite.h"
+#include "i9xx_plane.h"
+
+static void assert_vblank_disabled(struct drm_crtc *crtc)
+{
+	if (I915_STATE_WARN_ON(drm_crtc_vblank_get(crtc) == 0))
+		drm_crtc_vblank_put(crtc);
+}
+
+u32 intel_crtc_get_vblank_counter(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_vblank_crtc *vblank = &dev->vblank[drm_crtc_index(&crtc->base)];
+
+	if (!vblank->max_vblank_count)
+		return (u32)drm_crtc_accurate_vblank_count(&crtc->base);
+
+	return crtc->base.funcs->get_vblank_counter(&crtc->base);
+}
+
+u32 intel_crtc_max_vblank_count(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	u32 mode_flags = crtc->mode_flags;
+
+	/*
+	 * From Gen 11, In case of dsi cmd mode, frame counter wouldnt
+	 * have updated at the beginning of TE, if we want to use
+	 * the hw counter, then we would find it updated in only
+	 * the next TE, hence switching to sw counter.
+	 */
+	if (mode_flags & (I915_MODE_FLAG_DSI_USE_TE0 | I915_MODE_FLAG_DSI_USE_TE1))
+		return 0;
+
+	/*
+	 * On i965gm the hardware frame counter reads
+	 * zero when the TV encoder is enabled :(
+	 */
+	if (IS_I965GM(dev_priv) &&
+	    (crtc_state->output_types & BIT(INTEL_OUTPUT_TVOUT)))
+		return 0;
+
+	if (INTEL_GEN(dev_priv) >= 5 || IS_G4X(dev_priv))
+		return 0xffffffff; /* full 32 bit counter */
+	else if (INTEL_GEN(dev_priv) >= 3)
+		return 0xffffff; /* only 24 bits of frame count */
+	else
+		return 0; /* Gen2 doesn't have a hardware frame counter */
+}
+
+void intel_crtc_vblank_on(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	assert_vblank_disabled(&crtc->base);
+	drm_crtc_set_max_vblank_count(&crtc->base,
+				      intel_crtc_max_vblank_count(crtc_state));
+	drm_crtc_vblank_on(&crtc->base);
+}
+
+void intel_crtc_vblank_off(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	drm_crtc_vblank_off(&crtc->base);
+	assert_vblank_disabled(&crtc->base);
+}
+
+struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state;
+
+	crtc_state = kmalloc(sizeof(*crtc_state), GFP_KERNEL);
+
+	if (crtc_state)
+		intel_crtc_state_reset(crtc_state, crtc);
+
+	return crtc_state;
+}
+
+void intel_crtc_state_reset(struct intel_crtc_state *crtc_state,
+			    struct intel_crtc *crtc)
+{
+	memset(crtc_state, 0, sizeof(*crtc_state));
+
+	__drm_atomic_helper_crtc_state_reset(&crtc_state->uapi, &crtc->base);
+
+	crtc_state->cpu_transcoder = INVALID_TRANSCODER;
+	crtc_state->master_transcoder = INVALID_TRANSCODER;
+	crtc_state->hsw_workaround_pipe = INVALID_PIPE;
+	crtc_state->scaler_state.scaler_id = -1;
+	crtc_state->mst_master_transcoder = INVALID_TRANSCODER;
+}
+
+static struct intel_crtc *intel_crtc_alloc(void)
+{
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+
+	crtc = kzalloc(sizeof(*crtc), GFP_KERNEL);
+	if (!crtc)
+		return ERR_PTR(-ENOMEM);
+
+	crtc_state = intel_crtc_state_alloc(crtc);
+	if (!crtc_state) {
+		kfree(crtc);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	crtc->base.state = &crtc_state->uapi;
+	crtc->config = crtc_state;
+
+	return crtc;
+}
+
+static void intel_crtc_free(struct intel_crtc *crtc)
+{
+	intel_crtc_destroy_state(&crtc->base, crtc->base.state);
+	kfree(crtc);
+}
+
+static void intel_crtc_destroy(struct drm_crtc *crtc)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	drm_crtc_cleanup(crtc);
+	kfree(intel_crtc);
+}
+
+static int intel_crtc_late_register(struct drm_crtc *crtc)
+{
+	intel_crtc_debugfs_add(crtc);
+	return 0;
+}
+
+#define INTEL_CRTC_FUNCS \
+	.set_config = drm_atomic_helper_set_config, \
+	.destroy = intel_crtc_destroy, \
+	.page_flip = drm_atomic_helper_page_flip, \
+	.atomic_duplicate_state = intel_crtc_duplicate_state, \
+	.atomic_destroy_state = intel_crtc_destroy_state, \
+	.set_crc_source = intel_crtc_set_crc_source, \
+	.verify_crc_source = intel_crtc_verify_crc_source, \
+	.get_crc_sources = intel_crtc_get_crc_sources, \
+	.late_register = intel_crtc_late_register
+
+static const struct drm_crtc_funcs bdw_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = g4x_get_vblank_counter,
+	.enable_vblank = bdw_enable_vblank,
+	.disable_vblank = bdw_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs ilk_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = g4x_get_vblank_counter,
+	.enable_vblank = ilk_enable_vblank,
+	.disable_vblank = ilk_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs g4x_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = g4x_get_vblank_counter,
+	.enable_vblank = i965_enable_vblank,
+	.disable_vblank = i965_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i965_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = i915_get_vblank_counter,
+	.enable_vblank = i965_enable_vblank,
+	.disable_vblank = i965_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i915gm_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = i915_get_vblank_counter,
+	.enable_vblank = i915gm_enable_vblank,
+	.disable_vblank = i915gm_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i915_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = i915_get_vblank_counter,
+	.enable_vblank = i8xx_enable_vblank,
+	.disable_vblank = i8xx_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i8xx_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	/* no hw vblank counter */
+	.enable_vblank = i8xx_enable_vblank,
+	.disable_vblank = i8xx_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	struct intel_plane *primary, *cursor;
+	const struct drm_crtc_funcs *funcs;
+	struct intel_crtc *crtc;
+	int sprite, ret;
+
+	crtc = intel_crtc_alloc();
+	if (IS_ERR(crtc))
+		return PTR_ERR(crtc);
+
+	crtc->pipe = pipe;
+	crtc->num_scalers = RUNTIME_INFO(dev_priv)->num_scalers[pipe];
+
+	primary = intel_primary_plane_create(dev_priv, pipe);
+	if (IS_ERR(primary)) {
+		ret = PTR_ERR(primary);
+		goto fail;
+	}
+	crtc->plane_ids_mask |= BIT(primary->id);
+
+	for_each_sprite(dev_priv, pipe, sprite) {
+		struct intel_plane *plane;
+
+		plane = intel_sprite_plane_create(dev_priv, pipe, sprite);
+		if (IS_ERR(plane)) {
+			ret = PTR_ERR(plane);
+			goto fail;
+		}
+		crtc->plane_ids_mask |= BIT(plane->id);
+	}
+
+	cursor = intel_cursor_plane_create(dev_priv, pipe);
+	if (IS_ERR(cursor)) {
+		ret = PTR_ERR(cursor);
+		goto fail;
+	}
+	crtc->plane_ids_mask |= BIT(cursor->id);
+
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv) ||
+		    IS_VALLEYVIEW(dev_priv) || IS_G4X(dev_priv))
+			funcs = &g4x_crtc_funcs;
+		else if (IS_GEN(dev_priv, 4))
+			funcs = &i965_crtc_funcs;
+		else if (IS_I945GM(dev_priv) || IS_I915GM(dev_priv))
+			funcs = &i915gm_crtc_funcs;
+		else if (IS_GEN(dev_priv, 3))
+			funcs = &i915_crtc_funcs;
+		else
+			funcs = &i8xx_crtc_funcs;
+	} else {
+		if (INTEL_GEN(dev_priv) >= 8)
+			funcs = &bdw_crtc_funcs;
+		else
+			funcs = &ilk_crtc_funcs;
+	}
+
+	ret = drm_crtc_init_with_planes(&dev_priv->drm, &crtc->base,
+					&primary->base, &cursor->base,
+					funcs, "pipe %c", pipe_name(pipe));
+	if (ret)
+		goto fail;
+
+	BUG_ON(pipe >= ARRAY_SIZE(dev_priv->pipe_to_crtc_mapping) ||
+	       dev_priv->pipe_to_crtc_mapping[pipe] != NULL);
+	dev_priv->pipe_to_crtc_mapping[pipe] = crtc;
+
+	if (INTEL_GEN(dev_priv) < 9) {
+		enum i9xx_plane_id i9xx_plane = primary->i9xx_plane;
+
+		BUG_ON(i9xx_plane >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) ||
+		       dev_priv->plane_to_crtc_mapping[i9xx_plane] != NULL);
+		dev_priv->plane_to_crtc_mapping[i9xx_plane] = crtc;
+	}
+
+	if (INTEL_GEN(dev_priv) >= 10)
+		drm_crtc_create_scaling_filter_property(&crtc->base,
+						BIT(DRM_SCALING_FILTER_DEFAULT) |
+						BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR));
+
+	intel_color_init(crtc);
+
+	intel_crtc_crc_init(crtc);
+
+	drm_WARN_ON(&dev_priv->drm, drm_crtc_index(&crtc->base) != crtc->pipe);
+
+	return 0;
+
+fail:
+	intel_crtc_free(crtc);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_crtc.h b/drivers/gpu/drm/i915/display/intel_crtc.h
new file mode 100644
index 000000000000..08112d557411
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crtc.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_CRTC_H_
+#define _INTEL_CRTC_H_
+
+#include <linux/types.h>
+
+enum pipe;
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_crtc_state;
+
+u32 intel_crtc_max_vblank_count(const struct intel_crtc_state *crtc_state);
+int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe);
+struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc);
+void intel_crtc_state_reset(struct intel_crtc_state *crtc_state,
+			    struct intel_crtc *crtc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_cursor.c b/drivers/gpu/drm/i915/display/intel_cursor.c
index 276d2bb0e2cf..21fe4d2753e9 100644
--- a/drivers/gpu/drm/i915/display/intel_cursor.c
+++ b/drivers/gpu/drm/i915/display/intel_cursor.c
@@ -492,8 +492,8 @@ static void i9xx_update_cursor(struct intel_plane *plane,
 	unsigned long irqflags;
 
 	if (plane_state && plane_state->uapi.visible) {
-		unsigned width = drm_rect_width(&plane_state->uapi.dst);
-		unsigned height = drm_rect_height(&plane_state->uapi.dst);
+		int width = drm_rect_width(&plane_state->uapi.dst);
+		int height = drm_rect_height(&plane_state->uapi.dst);
 
 		cntl = plane_state->ctl |
 			i9xx_cursor_ctl_crtc(crtc_state);
@@ -522,7 +522,7 @@ static void i9xx_update_cursor(struct intel_plane *plane,
 	 * cursor that doesn't appear to move, or even change
 	 * shape. Thus we always write CURBASE.
 	 *
-	 * The other registers are armed by by the CURBASE write
+	 * The other registers are armed by the CURBASE write
 	 * except when the plane is getting enabled at which time
 	 * the CURCNTR write arms the update.
 	 */
diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
index 17eaa56c5a99..1bb40ec5fe5d 100644
--- a/drivers/gpu/drm/i915/display/intel_ddi.c
+++ b/drivers/gpu/drm/i915/display/intel_ddi.c
@@ -46,10 +46,12 @@
 #include "intel_hotplug.h"
 #include "intel_lspcon.h"
 #include "intel_panel.h"
+#include "intel_pps.h"
 #include "intel_psr.h"
 #include "intel_sprite.h"
 #include "intel_tc.h"
 #include "intel_vdsc.h"
+#include "intel_vrr.h"
 
 struct ddi_buf_trans {
 	u32 trans1;	/* balance leg enable, de-emph level */
@@ -611,6 +613,34 @@ static const struct cnl_ddi_buf_trans jsl_combo_phy_ddi_translations_edp_hbr2[]
 	{ 0xA, 0x35, 0x3F, 0x00, 0x00 },        /* 350   350      0.0   */
 };
 
+static const struct cnl_ddi_buf_trans dg1_combo_phy_ddi_translations_dp_rbr_hbr[] = {
+						/* NT mV Trans mV db    */
+	{ 0xA, 0x32, 0x3F, 0x00, 0x00 },	/* 350   350      0.0   */
+	{ 0xA, 0x48, 0x35, 0x00, 0x0A },	/* 350   500      3.1   */
+	{ 0xC, 0x63, 0x2F, 0x00, 0x10 },	/* 350   700      6.0   */
+	{ 0x6, 0x7F, 0x2C, 0x00, 0x13 },	/* 350   900      8.2   */
+	{ 0xA, 0x43, 0x3F, 0x00, 0x00 },	/* 500   500      0.0   */
+	{ 0xC, 0x60, 0x36, 0x00, 0x09 },	/* 500   700      2.9   */
+	{ 0x6, 0x7F, 0x30, 0x00, 0x0F },	/* 500   900      5.1   */
+	{ 0xC, 0x60, 0x3F, 0x00, 0x00 },	/* 650   700      0.6   */
+	{ 0x6, 0x7F, 0x37, 0x00, 0x08 },	/* 600   900      3.5   */
+	{ 0x6, 0x7F, 0x3F, 0x00, 0x00 },	/* 900   900      0.0   */
+};
+
+static const struct cnl_ddi_buf_trans dg1_combo_phy_ddi_translations_dp_hbr2_hbr3[] = {
+						/* NT mV Trans mV db    */
+	{ 0xA, 0x32, 0x3F, 0x00, 0x00 },	/* 350   350      0.0   */
+	{ 0xA, 0x48, 0x35, 0x00, 0x0A },	/* 350   500      3.1   */
+	{ 0xC, 0x63, 0x2F, 0x00, 0x10 },	/* 350   700      6.0   */
+	{ 0x6, 0x7F, 0x2C, 0x00, 0x13 },	/* 350   900      8.2   */
+	{ 0xA, 0x43, 0x3F, 0x00, 0x00 },	/* 500   500      0.0   */
+	{ 0xC, 0x60, 0x36, 0x00, 0x09 },	/* 500   700      2.9   */
+	{ 0x6, 0x7F, 0x30, 0x00, 0x0F },	/* 500   900      5.1   */
+	{ 0xC, 0x58, 0x3F, 0x00, 0x00 },	/* 650   700      0.6   */
+	{ 0x6, 0x7F, 0x35, 0x00, 0x0A },	/* 600   900      3.5   */
+	{ 0x6, 0x7F, 0x3F, 0x00, 0x00 },	/* 900   900      0.0   */
+};
+
 struct icl_mg_phy_ddi_buf_trans {
 	u32 cri_txdeemph_override_11_6;
 	u32 cri_txdeemph_override_5_0;
@@ -766,6 +796,34 @@ static const struct cnl_ddi_buf_trans tgl_combo_phy_ddi_translations_edp_hbr2_ho
 	{ 0x6, 0x7F, 0x3F, 0x00, 0x00 },	/* 2	1	*/
 };
 
+static const struct cnl_ddi_buf_trans rkl_combo_phy_ddi_translations_dp_hbr[] = {
+						/* NT mV Trans mV db    */
+	{ 0xA, 0x2F, 0x3F, 0x00, 0x00 },	/* 350   350      0.0   */
+	{ 0xA, 0x4F, 0x37, 0x00, 0x08 },	/* 350   500      3.1   */
+	{ 0xC, 0x63, 0x2F, 0x00, 0x10 },	/* 350   700      6.0   */
+	{ 0x6, 0x7D, 0x2A, 0x00, 0x15 },	/* 350   900      8.2   */
+	{ 0xA, 0x4C, 0x3F, 0x00, 0x00 },	/* 500   500      0.0   */
+	{ 0xC, 0x73, 0x34, 0x00, 0x0B },	/* 500   700      2.9   */
+	{ 0x6, 0x7F, 0x2F, 0x00, 0x10 },	/* 500   900      5.1   */
+	{ 0xC, 0x6E, 0x3E, 0x00, 0x01 },	/* 650   700      0.6   */
+	{ 0x6, 0x7F, 0x35, 0x00, 0x0A },	/* 600   900      3.5   */
+	{ 0x6, 0x7F, 0x3F, 0x00, 0x00 },	/* 900   900      0.0   */
+};
+
+static const struct cnl_ddi_buf_trans rkl_combo_phy_ddi_translations_dp_hbr2_hbr3[] = {
+						/* NT mV Trans mV db    */
+	{ 0xA, 0x35, 0x3F, 0x00, 0x00 },	/* 350   350      0.0   */
+	{ 0xA, 0x50, 0x38, 0x00, 0x07 },	/* 350   500      3.1   */
+	{ 0xC, 0x61, 0x33, 0x00, 0x0C },	/* 350   700      6.0   */
+	{ 0x6, 0x7F, 0x2E, 0x00, 0x11 },	/* 350   900      8.2   */
+	{ 0xA, 0x47, 0x3F, 0x00, 0x00 },	/* 500   500      0.0   */
+	{ 0xC, 0x5F, 0x38, 0x00, 0x07 },	/* 500   700      2.9   */
+	{ 0x6, 0x7F, 0x2F, 0x00, 0x10 },	/* 500   900      5.1   */
+	{ 0xC, 0x5F, 0x3F, 0x00, 0x00 },	/* 650   700      0.6   */
+	{ 0x6, 0x7E, 0x36, 0x00, 0x09 },	/* 600   900      3.5   */
+	{ 0x6, 0x7F, 0x3F, 0x00, 0x00 },	/* 900   900      0.0   */
+};
+
 static bool is_hobl_buf_trans(const struct cnl_ddi_buf_trans *table)
 {
 	return table == tgl_combo_phy_ddi_translations_edp_hbr2_hobl;
@@ -1093,6 +1151,12 @@ icl_get_combo_buf_trans_edp(struct intel_encoder *encoder,
 	} else if (dev_priv->vbt.edp.low_vswing) {
 		*n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_edp_hbr2);
 		return icl_combo_phy_ddi_translations_edp_hbr2;
+	} else if (IS_DG1(dev_priv) && crtc_state->port_clock > 270000) {
+		*n_entries = ARRAY_SIZE(dg1_combo_phy_ddi_translations_dp_hbr2_hbr3);
+		return dg1_combo_phy_ddi_translations_dp_hbr2_hbr3;
+	} else if (IS_DG1(dev_priv)) {
+		*n_entries = ARRAY_SIZE(dg1_combo_phy_ddi_translations_dp_rbr_hbr);
+		return dg1_combo_phy_ddi_translations_dp_rbr_hbr;
 	}
 
 	return icl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
@@ -1259,7 +1323,10 @@ tgl_get_combo_buf_trans_dp(struct intel_encoder *encoder,
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 
 	if (crtc_state->port_clock > 270000) {
-		if (IS_TGL_U(dev_priv) || IS_TGL_Y(dev_priv)) {
+		if (IS_ROCKETLAKE(dev_priv)) {
+			*n_entries = ARRAY_SIZE(rkl_combo_phy_ddi_translations_dp_hbr2_hbr3);
+			return rkl_combo_phy_ddi_translations_dp_hbr2_hbr3;
+		} else if (IS_TGL_U(dev_priv) || IS_TGL_Y(dev_priv)) {
 			*n_entries = ARRAY_SIZE(tgl_uy_combo_phy_ddi_translations_dp_hbr2);
 			return tgl_uy_combo_phy_ddi_translations_dp_hbr2;
 		} else {
@@ -1267,8 +1334,13 @@ tgl_get_combo_buf_trans_dp(struct intel_encoder *encoder,
 			return tgl_combo_phy_ddi_translations_dp_hbr2;
 		}
 	} else {
-		*n_entries = ARRAY_SIZE(tgl_combo_phy_ddi_translations_dp_hbr);
-		return tgl_combo_phy_ddi_translations_dp_hbr;
+		if (IS_ROCKETLAKE(dev_priv)) {
+			*n_entries = ARRAY_SIZE(rkl_combo_phy_ddi_translations_dp_hbr);
+			return rkl_combo_phy_ddi_translations_dp_hbr;
+		} else {
+			*n_entries = ARRAY_SIZE(tgl_combo_phy_ddi_translations_dp_hbr);
+			return tgl_combo_phy_ddi_translations_dp_hbr;
+		}
 	}
 }
 
@@ -2029,9 +2101,9 @@ void intel_ddi_disable_transcoder_func(const struct intel_crtc_state *crtc_state
 	}
 }
 
-int intel_ddi_toggle_hdcp_signalling(struct intel_encoder *intel_encoder,
-				     enum transcoder cpu_transcoder,
-				     bool enable)
+int intel_ddi_toggle_hdcp_bits(struct intel_encoder *intel_encoder,
+			       enum transcoder cpu_transcoder,
+			       bool enable, u32 hdcp_mask)
 {
 	struct drm_device *dev = intel_encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
@@ -2046,9 +2118,9 @@ int intel_ddi_toggle_hdcp_signalling(struct intel_encoder *intel_encoder,
 
 	tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
 	if (enable)
-		tmp |= TRANS_DDI_HDCP_SIGNALLING;
+		tmp |= hdcp_mask;
 	else
-		tmp &= ~TRANS_DDI_HDCP_SIGNALLING;
+		tmp &= ~hdcp_mask;
 	intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder), tmp);
 	intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref);
 	return ret;
@@ -2631,15 +2703,11 @@ static void icl_ddi_combo_vswing_program(struct intel_encoder *encoder,
 		ddi_translations = ehl_get_combo_buf_trans(encoder, crtc_state, &n_entries);
 	else
 		ddi_translations = icl_get_combo_buf_trans(encoder, crtc_state, &n_entries);
-	if (!ddi_translations)
-		return;
 
-	if (level >= n_entries) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "DDI translation not found for level %d. Using %d instead.",
-			    level, n_entries - 1);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !ddi_translations))
+		return;
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, level >= n_entries))
 		level = n_entries - 1;
-	}
 
 	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP)) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
@@ -2759,14 +2827,15 @@ static void icl_mg_phy_ddi_vswing_sequence(struct intel_encoder *encoder,
 	int n_entries, ln;
 	u32 val;
 
+	if (enc_to_dig_port(encoder)->tc_mode == TC_PORT_TBT_ALT)
+		return;
+
 	ddi_translations = icl_get_mg_buf_trans(encoder, crtc_state, &n_entries);
-	/* The table does not have values for level 3 and level 9. */
-	if (level >= n_entries || level == 3 || level == 9) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "DDI translation not found for level %d. Using %d instead.",
-			    level, n_entries - 2);
-		level = n_entries - 2;
-	}
+
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !ddi_translations))
+		return;
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, level >= n_entries))
+		level = n_entries - 1;
 
 	/* Set MG_TX_LINK_PARAMS cri_use_fs32 to 0. */
 	for (ln = 0; ln < 2; ln++) {
@@ -2896,9 +2965,14 @@ tgl_dkl_phy_ddi_vswing_sequence(struct intel_encoder *encoder,
 	u32 val, dpcnt_mask, dpcnt_val;
 	int n_entries, ln;
 
+	if (enc_to_dig_port(encoder)->tc_mode == TC_PORT_TBT_ALT)
+		return;
+
 	ddi_translations = tgl_get_dkl_buf_trans(encoder, crtc_state, &n_entries);
 
-	if (level >= n_entries)
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !ddi_translations))
+		return;
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, level >= n_entries))
 		level = n_entries - 1;
 
 	dpcnt_mask = (DKL_TX_PRESHOOT_COEFF_MASK |
@@ -3485,6 +3559,22 @@ i915_reg_t dp_tp_status_reg(struct intel_encoder *encoder,
 		return DP_TP_STATUS(encoder->port);
 }
 
+static void intel_dp_sink_set_msa_timing_par_ignore_state(struct intel_dp *intel_dp,
+							  const struct intel_crtc_state *crtc_state,
+							  bool enable)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (!crtc_state->vrr.enable)
+		return;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_DOWNSPREAD_CTRL,
+			       enable ? DP_MSA_TIMING_PAR_IGNORE_EN : 0) <= 0)
+		drm_dbg_kms(&i915->drm,
+			    "Failed to set MSA_TIMING_PAR_IGNORE %s in the sink\n",
+			    enable ? "enable" : "disable");
+}
+
 static void intel_dp_sink_set_fec_ready(struct intel_dp *intel_dp,
 					const struct intel_crtc_state *crtc_state)
 {
@@ -3531,6 +3621,23 @@ static void intel_ddi_disable_fec_state(struct intel_encoder *encoder,
 	intel_de_posting_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
 }
 
+static void intel_ddi_power_up_lanes(struct intel_encoder *encoder,
+				     const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (intel_phy_is_combo(i915, phy)) {
+		bool lane_reversal =
+			dig_port->saved_port_bits & DDI_BUF_PORT_REVERSAL;
+
+		intel_combo_phy_power_up_lanes(i915, phy, false,
+					       crtc_state->lane_count,
+					       lane_reversal);
+	}
+}
+
 static void tgl_ddi_pre_enable_dp(struct intel_atomic_state *state,
 				  struct intel_encoder *encoder,
 				  const struct intel_crtc_state *crtc_state,
@@ -3555,7 +3662,7 @@ static void tgl_ddi_pre_enable_dp(struct intel_atomic_state *state,
 	 */
 
 	/* 2. Enable Panel Power if PPS is required */
-	intel_edp_panel_on(intel_dp);
+	intel_pps_on(intel_dp);
 
 	/*
 	 * 3. For non-TBT Type-C ports, set FIA lane count
@@ -3622,14 +3729,7 @@ static void tgl_ddi_pre_enable_dp(struct intel_atomic_state *state,
 	 * 7.f Combo PHY: Configure PORT_CL_DW10 Static Power Down to power up
 	 * the used lanes of the DDI.
 	 */
-	if (intel_phy_is_combo(dev_priv, phy)) {
-		bool lane_reversal =
-			dig_port->saved_port_bits & DDI_BUF_PORT_REVERSAL;
-
-		intel_combo_phy_power_up_lanes(dev_priv, phy, false,
-					       crtc_state->lane_count,
-					       lane_reversal);
-	}
+	intel_ddi_power_up_lanes(encoder, crtc_state);
 
 	/*
 	 * 7.g Configure and enable DDI_BUF_CTL
@@ -3698,7 +3798,7 @@ static void hsw_ddi_pre_enable_dp(struct intel_atomic_state *state,
 				 crtc_state->port_clock,
 				 crtc_state->lane_count);
 
-	intel_edp_panel_on(intel_dp);
+	intel_pps_on(intel_dp);
 
 	intel_ddi_clk_select(encoder, crtc_state);
 
@@ -3720,14 +3820,7 @@ static void hsw_ddi_pre_enable_dp(struct intel_atomic_state *state,
 	else
 		intel_prepare_dp_ddi_buffers(encoder, crtc_state);
 
-	if (intel_phy_is_combo(dev_priv, phy)) {
-		bool lane_reversal =
-			dig_port->saved_port_bits & DDI_BUF_PORT_REVERSAL;
-
-		intel_combo_phy_power_up_lanes(dev_priv, phy, false,
-					       crtc_state->lane_count,
-					       lane_reversal);
-	}
+	intel_ddi_power_up_lanes(encoder, crtc_state);
 
 	intel_ddi_init_dp_buf_reg(encoder, crtc_state);
 	if (!is_mst)
@@ -3940,8 +4033,8 @@ static void intel_ddi_post_disable_dp(struct intel_atomic_state *state,
 	if (INTEL_GEN(dev_priv) >= 12)
 		intel_ddi_disable_pipe_clock(old_crtc_state);
 
-	intel_edp_panel_vdd_on(intel_dp);
-	intel_edp_panel_off(intel_dp);
+	intel_pps_vdd_on(intel_dp);
+	intel_pps_off(intel_dp);
 
 	if (!intel_phy_is_tc(dev_priv, phy) ||
 	    dig_port->tc_mode != TC_PORT_TBT_ALT)
@@ -3992,6 +4085,8 @@ static void intel_ddi_post_disable(struct intel_atomic_state *state,
 
 		intel_disable_pipe(old_crtc_state);
 
+		intel_vrr_disable(old_crtc_state);
+
 		intel_ddi_disable_transcoder_func(old_crtc_state);
 
 		intel_dsc_disable(old_crtc_state);
@@ -4222,6 +4317,8 @@ static void intel_enable_ddi_hdmi(struct intel_atomic_state *state,
 		intel_de_write(dev_priv, reg, val);
 	}
 
+	intel_ddi_power_up_lanes(encoder, crtc_state);
+
 	/* In HDMI/DVI mode, the port width, and swing/emphasis values
 	 * are ignored so nothing special needs to be done besides
 	 * enabling the port.
@@ -4243,6 +4340,8 @@ static void intel_enable_ddi(struct intel_atomic_state *state,
 	if (!crtc_state->bigjoiner_slave)
 		intel_ddi_enable_transcoder_func(encoder, crtc_state);
 
+	intel_vrr_enable(encoder, crtc_state);
+
 	intel_enable_pipe(crtc_state);
 
 	intel_crtc_vblank_on(crtc_state);
@@ -4256,7 +4355,7 @@ static void intel_enable_ddi(struct intel_atomic_state *state,
 	if (conn_state->content_protection ==
 	    DRM_MODE_CONTENT_PROTECTION_DESIRED)
 		intel_hdcp_enable(to_intel_connector(conn_state->connector),
-				  crtc_state->cpu_transcoder,
+				  crtc_state,
 				  (u8)conn_state->hdcp_content_type);
 }
 
@@ -4279,6 +4378,9 @@ static void intel_disable_ddi_dp(struct intel_atomic_state *state,
 	/* Disable the decompression in DP Sink */
 	intel_dp_sink_set_decompression_state(intel_dp, old_crtc_state,
 					      false);
+	/* Disable Ignore_MSA bit in DP Sink */
+	intel_dp_sink_set_msa_timing_par_ignore_state(intel_dp, old_crtc_state,
+						      false);
 }
 
 static void intel_disable_ddi_hdmi(struct intel_atomic_state *state,
@@ -4969,6 +5071,8 @@ static void intel_ddi_encoder_destroy(struct drm_encoder *encoder)
 	intel_dp_encoder_flush_work(encoder);
 
 	drm_encoder_cleanup(encoder);
+	if (dig_port)
+		kfree(dig_port->hdcp_port_data.streams);
 	kfree(dig_port);
 }
 
@@ -5122,12 +5226,20 @@ intel_ddi_hotplug(struct intel_encoder *encoder,
 {
 	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
 	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_dp *intel_dp = &dig_port->dp;
 	enum phy phy = intel_port_to_phy(i915, encoder->port);
 	bool is_tc = intel_phy_is_tc(i915, phy);
 	struct drm_modeset_acquire_ctx ctx;
 	enum intel_hotplug_state state;
 	int ret;
 
+	if (intel_dp->compliance.test_active &&
+	    intel_dp->compliance.test_type == DP_TEST_LINK_PHY_TEST_PATTERN) {
+		intel_dp_phy_test(encoder);
+		/* just do the PHY test and nothing else */
+		return INTEL_HOTPLUG_UNCHANGED;
+	}
+
 	state = intel_encoder_hotplug(encoder, connector);
 
 	drm_modeset_acquire_init(&ctx, 0);
@@ -5285,8 +5397,7 @@ intel_ddi_max_lanes(struct intel_digital_port *dig_port)
 static bool hti_uses_phy(struct drm_i915_private *i915, enum phy phy)
 {
 	return i915->hti_state & HDPORT_ENABLED &&
-		(i915->hti_state & HDPORT_PHY_USED_DP(phy) ||
-		 i915->hti_state & HDPORT_PHY_USED_HDMI(phy));
+	       i915->hti_state & HDPORT_DDI_USED(phy);
 }
 
 static enum hpd_pin dg1_hpd_pin(struct drm_i915_private *dev_priv,
diff --git a/drivers/gpu/drm/i915/display/intel_ddi.h b/drivers/gpu/drm/i915/display/intel_ddi.h
index dcc711cfe4fe..a4dd815c0000 100644
--- a/drivers/gpu/drm/i915/display/intel_ddi.h
+++ b/drivers/gpu/drm/i915/display/intel_ddi.h
@@ -50,9 +50,9 @@ u32 bxt_signal_levels(struct intel_dp *intel_dp,
 		      const struct intel_crtc_state *crtc_state);
 u32 ddi_signal_levels(struct intel_dp *intel_dp,
 		      const struct intel_crtc_state *crtc_state);
-int intel_ddi_toggle_hdcp_signalling(struct intel_encoder *intel_encoder,
-				     enum transcoder cpu_transcoder,
-				     bool enable);
+int intel_ddi_toggle_hdcp_bits(struct intel_encoder *intel_encoder,
+			       enum transcoder cpu_transcoder,
+			       bool enable, u32 hdcp_mask);
 void icl_sanitize_encoder_pll_mapping(struct intel_encoder *encoder);
 
 #endif /* __INTEL_DDI_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
index b5c9a059a3f6..25fed00b5121 100644
--- a/drivers/gpu/drm/i915/display/intel_display.c
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -48,6 +48,7 @@
 #include "display/intel_display_debugfs.h"
 #include "display/intel_dp.h"
 #include "display/intel_dp_mst.h"
+#include "display/intel_dpll.h"
 #include "display/intel_dpll_mgr.h"
 #include "display/intel_dsi.h"
 #include "display/intel_dvo.h"
@@ -57,6 +58,9 @@
 #include "display/intel_sdvo.h"
 #include "display/intel_tv.h"
 #include "display/intel_vdsc.h"
+#include "display/intel_vrr.h"
+
+#include "gem/i915_gem_object.h"
 
 #include "gt/intel_rps.h"
 
@@ -68,11 +72,12 @@
 #include "intel_bw.h"
 #include "intel_cdclk.h"
 #include "intel_color.h"
+#include "intel_crtc.h"
 #include "intel_csr.h"
-#include "intel_cursor.h"
 #include "intel_display_types.h"
 #include "intel_dp_link_training.h"
 #include "intel_fbc.h"
+#include "intel_fdi.h"
 #include "intel_fbdev.h"
 #include "intel_fifo_underrun.h"
 #include "intel_frontbuffer.h"
@@ -81,6 +86,7 @@
 #include "intel_overlay.h"
 #include "intel_pipe_crc.h"
 #include "intel_pm.h"
+#include "intel_pps.h"
 #include "intel_psr.h"
 #include "intel_quirks.h"
 #include "intel_sideband.h"
@@ -114,18 +120,6 @@ static void skl_pfit_enable(const struct intel_crtc_state *crtc_state);
 static void ilk_pfit_enable(const struct intel_crtc_state *crtc_state);
 static void intel_modeset_setup_hw_state(struct drm_device *dev,
 					 struct drm_modeset_acquire_ctx *ctx);
-static struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc);
-
-struct intel_limit {
-	struct {
-		int min, max;
-	} dot, vco, n, m, m1, m2, p, p1;
-
-	struct {
-		int dot_limit;
-		int p2_slow, p2_fast;
-	} p2;
-};
 
 /* returns HPLL frequency in kHz */
 int vlv_get_hpll_vco(struct drm_i915_private *dev_priv)
@@ -184,281 +178,6 @@ static void intel_update_czclk(struct drm_i915_private *dev_priv)
 		dev_priv->czclk_freq);
 }
 
-/* units of 100MHz */
-static u32 intel_fdi_link_freq(struct drm_i915_private *dev_priv,
-			       const struct intel_crtc_state *pipe_config)
-{
-	if (HAS_DDI(dev_priv))
-		return pipe_config->port_clock; /* SPLL */
-	else
-		return dev_priv->fdi_pll_freq;
-}
-
-static const struct intel_limit intel_limits_i8xx_dac = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 908000, .max = 1512000 },
-	.n = { .min = 2, .max = 16 },
-	.m = { .min = 96, .max = 140 },
-	.m1 = { .min = 18, .max = 26 },
-	.m2 = { .min = 6, .max = 16 },
-	.p = { .min = 4, .max = 128 },
-	.p1 = { .min = 2, .max = 33 },
-	.p2 = { .dot_limit = 165000,
-		.p2_slow = 4, .p2_fast = 2 },
-};
-
-static const struct intel_limit intel_limits_i8xx_dvo = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 908000, .max = 1512000 },
-	.n = { .min = 2, .max = 16 },
-	.m = { .min = 96, .max = 140 },
-	.m1 = { .min = 18, .max = 26 },
-	.m2 = { .min = 6, .max = 16 },
-	.p = { .min = 4, .max = 128 },
-	.p1 = { .min = 2, .max = 33 },
-	.p2 = { .dot_limit = 165000,
-		.p2_slow = 4, .p2_fast = 4 },
-};
-
-static const struct intel_limit intel_limits_i8xx_lvds = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 908000, .max = 1512000 },
-	.n = { .min = 2, .max = 16 },
-	.m = { .min = 96, .max = 140 },
-	.m1 = { .min = 18, .max = 26 },
-	.m2 = { .min = 6, .max = 16 },
-	.p = { .min = 4, .max = 128 },
-	.p1 = { .min = 1, .max = 6 },
-	.p2 = { .dot_limit = 165000,
-		.p2_slow = 14, .p2_fast = 7 },
-};
-
-static const struct intel_limit intel_limits_i9xx_sdvo = {
-	.dot = { .min = 20000, .max = 400000 },
-	.vco = { .min = 1400000, .max = 2800000 },
-	.n = { .min = 1, .max = 6 },
-	.m = { .min = 70, .max = 120 },
-	.m1 = { .min = 8, .max = 18 },
-	.m2 = { .min = 3, .max = 7 },
-	.p = { .min = 5, .max = 80 },
-	.p1 = { .min = 1, .max = 8 },
-	.p2 = { .dot_limit = 200000,
-		.p2_slow = 10, .p2_fast = 5 },
-};
-
-static const struct intel_limit intel_limits_i9xx_lvds = {
-	.dot = { .min = 20000, .max = 400000 },
-	.vco = { .min = 1400000, .max = 2800000 },
-	.n = { .min = 1, .max = 6 },
-	.m = { .min = 70, .max = 120 },
-	.m1 = { .min = 8, .max = 18 },
-	.m2 = { .min = 3, .max = 7 },
-	.p = { .min = 7, .max = 98 },
-	.p1 = { .min = 1, .max = 8 },
-	.p2 = { .dot_limit = 112000,
-		.p2_slow = 14, .p2_fast = 7 },
-};
-
-
-static const struct intel_limit intel_limits_g4x_sdvo = {
-	.dot = { .min = 25000, .max = 270000 },
-	.vco = { .min = 1750000, .max = 3500000},
-	.n = { .min = 1, .max = 4 },
-	.m = { .min = 104, .max = 138 },
-	.m1 = { .min = 17, .max = 23 },
-	.m2 = { .min = 5, .max = 11 },
-	.p = { .min = 10, .max = 30 },
-	.p1 = { .min = 1, .max = 3},
-	.p2 = { .dot_limit = 270000,
-		.p2_slow = 10,
-		.p2_fast = 10
-	},
-};
-
-static const struct intel_limit intel_limits_g4x_hdmi = {
-	.dot = { .min = 22000, .max = 400000 },
-	.vco = { .min = 1750000, .max = 3500000},
-	.n = { .min = 1, .max = 4 },
-	.m = { .min = 104, .max = 138 },
-	.m1 = { .min = 16, .max = 23 },
-	.m2 = { .min = 5, .max = 11 },
-	.p = { .min = 5, .max = 80 },
-	.p1 = { .min = 1, .max = 8},
-	.p2 = { .dot_limit = 165000,
-		.p2_slow = 10, .p2_fast = 5 },
-};
-
-static const struct intel_limit intel_limits_g4x_single_channel_lvds = {
-	.dot = { .min = 20000, .max = 115000 },
-	.vco = { .min = 1750000, .max = 3500000 },
-	.n = { .min = 1, .max = 3 },
-	.m = { .min = 104, .max = 138 },
-	.m1 = { .min = 17, .max = 23 },
-	.m2 = { .min = 5, .max = 11 },
-	.p = { .min = 28, .max = 112 },
-	.p1 = { .min = 2, .max = 8 },
-	.p2 = { .dot_limit = 0,
-		.p2_slow = 14, .p2_fast = 14
-	},
-};
-
-static const struct intel_limit intel_limits_g4x_dual_channel_lvds = {
-	.dot = { .min = 80000, .max = 224000 },
-	.vco = { .min = 1750000, .max = 3500000 },
-	.n = { .min = 1, .max = 3 },
-	.m = { .min = 104, .max = 138 },
-	.m1 = { .min = 17, .max = 23 },
-	.m2 = { .min = 5, .max = 11 },
-	.p = { .min = 14, .max = 42 },
-	.p1 = { .min = 2, .max = 6 },
-	.p2 = { .dot_limit = 0,
-		.p2_slow = 7, .p2_fast = 7
-	},
-};
-
-static const struct intel_limit pnv_limits_sdvo = {
-	.dot = { .min = 20000, .max = 400000},
-	.vco = { .min = 1700000, .max = 3500000 },
-	/* Pineview's Ncounter is a ring counter */
-	.n = { .min = 3, .max = 6 },
-	.m = { .min = 2, .max = 256 },
-	/* Pineview only has one combined m divider, which we treat as m2. */
-	.m1 = { .min = 0, .max = 0 },
-	.m2 = { .min = 0, .max = 254 },
-	.p = { .min = 5, .max = 80 },
-	.p1 = { .min = 1, .max = 8 },
-	.p2 = { .dot_limit = 200000,
-		.p2_slow = 10, .p2_fast = 5 },
-};
-
-static const struct intel_limit pnv_limits_lvds = {
-	.dot = { .min = 20000, .max = 400000 },
-	.vco = { .min = 1700000, .max = 3500000 },
-	.n = { .min = 3, .max = 6 },
-	.m = { .min = 2, .max = 256 },
-	.m1 = { .min = 0, .max = 0 },
-	.m2 = { .min = 0, .max = 254 },
-	.p = { .min = 7, .max = 112 },
-	.p1 = { .min = 1, .max = 8 },
-	.p2 = { .dot_limit = 112000,
-		.p2_slow = 14, .p2_fast = 14 },
-};
-
-/* Ironlake / Sandybridge
- *
- * We calculate clock using (register_value + 2) for N/M1/M2, so here
- * the range value for them is (actual_value - 2).
- */
-static const struct intel_limit ilk_limits_dac = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000 },
-	.n = { .min = 1, .max = 5 },
-	.m = { .min = 79, .max = 127 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 5, .max = 80 },
-	.p1 = { .min = 1, .max = 8 },
-	.p2 = { .dot_limit = 225000,
-		.p2_slow = 10, .p2_fast = 5 },
-};
-
-static const struct intel_limit ilk_limits_single_lvds = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000 },
-	.n = { .min = 1, .max = 3 },
-	.m = { .min = 79, .max = 118 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 28, .max = 112 },
-	.p1 = { .min = 2, .max = 8 },
-	.p2 = { .dot_limit = 225000,
-		.p2_slow = 14, .p2_fast = 14 },
-};
-
-static const struct intel_limit ilk_limits_dual_lvds = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000 },
-	.n = { .min = 1, .max = 3 },
-	.m = { .min = 79, .max = 127 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 14, .max = 56 },
-	.p1 = { .min = 2, .max = 8 },
-	.p2 = { .dot_limit = 225000,
-		.p2_slow = 7, .p2_fast = 7 },
-};
-
-/* LVDS 100mhz refclk limits. */
-static const struct intel_limit ilk_limits_single_lvds_100m = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000 },
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 79, .max = 126 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 28, .max = 112 },
-	.p1 = { .min = 2, .max = 8 },
-	.p2 = { .dot_limit = 225000,
-		.p2_slow = 14, .p2_fast = 14 },
-};
-
-static const struct intel_limit ilk_limits_dual_lvds_100m = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000 },
-	.n = { .min = 1, .max = 3 },
-	.m = { .min = 79, .max = 126 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 14, .max = 42 },
-	.p1 = { .min = 2, .max = 6 },
-	.p2 = { .dot_limit = 225000,
-		.p2_slow = 7, .p2_fast = 7 },
-};
-
-static const struct intel_limit intel_limits_vlv = {
-	 /*
-	  * These are the data rate limits (measured in fast clocks)
-	  * since those are the strictest limits we have. The fast
-	  * clock and actual rate limits are more relaxed, so checking
-	  * them would make no difference.
-	  */
-	.dot = { .min = 25000 * 5, .max = 270000 * 5 },
-	.vco = { .min = 4000000, .max = 6000000 },
-	.n = { .min = 1, .max = 7 },
-	.m1 = { .min = 2, .max = 3 },
-	.m2 = { .min = 11, .max = 156 },
-	.p1 = { .min = 2, .max = 3 },
-	.p2 = { .p2_slow = 2, .p2_fast = 20 }, /* slow=min, fast=max */
-};
-
-static const struct intel_limit intel_limits_chv = {
-	/*
-	 * These are the data rate limits (measured in fast clocks)
-	 * since those are the strictest limits we have.  The fast
-	 * clock and actual rate limits are more relaxed, so checking
-	 * them would make no difference.
-	 */
-	.dot = { .min = 25000 * 5, .max = 540000 * 5},
-	.vco = { .min = 4800000, .max = 6480000 },
-	.n = { .min = 1, .max = 1 },
-	.m1 = { .min = 2, .max = 2 },
-	.m2 = { .min = 24 << 22, .max = 175 << 22 },
-	.p1 = { .min = 2, .max = 4 },
-	.p2 = {	.p2_slow = 1, .p2_fast = 14 },
-};
-
-static const struct intel_limit intel_limits_bxt = {
-	/* FIXME: find real dot limits */
-	.dot = { .min = 0, .max = INT_MAX },
-	.vco = { .min = 4800000, .max = 6700000 },
-	.n = { .min = 1, .max = 1 },
-	.m1 = { .min = 2, .max = 2 },
-	/* FIXME: find real m2 limits */
-	.m2 = { .min = 2 << 22, .max = 255 << 22 },
-	.p1 = { .min = 2, .max = 4 },
-	.p2 = { .p2_slow = 1, .p2_fast = 20 },
-};
-
 /* WA Display #0827: Gen9:all */
 static void
 skl_wa_827(struct drm_i915_private *dev_priv, enum pipe pipe, bool enable)
@@ -503,483 +222,6 @@ is_trans_port_sync_mode(const struct intel_crtc_state *crtc_state)
 		is_trans_port_sync_slave(crtc_state);
 }
 
-/*
- * Platform specific helpers to calculate the port PLL loopback- (clock.m),
- * and post-divider (clock.p) values, pre- (clock.vco) and post-divided fast
- * (clock.dot) clock rates. This fast dot clock is fed to the port's IO logic.
- * The helpers' return value is the rate of the clock that is fed to the
- * display engine's pipe which can be the above fast dot clock rate or a
- * divided-down version of it.
- */
-/* m1 is reserved as 0 in Pineview, n is a ring counter */
-static int pnv_calc_dpll_params(int refclk, struct dpll *clock)
-{
-	clock->m = clock->m2 + 2;
-	clock->p = clock->p1 * clock->p2;
-	if (WARN_ON(clock->n == 0 || clock->p == 0))
-		return 0;
-	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n);
-	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
-
-	return clock->dot;
-}
-
-static u32 i9xx_dpll_compute_m(struct dpll *dpll)
-{
-	return 5 * (dpll->m1 + 2) + (dpll->m2 + 2);
-}
-
-static int i9xx_calc_dpll_params(int refclk, struct dpll *clock)
-{
-	clock->m = i9xx_dpll_compute_m(clock);
-	clock->p = clock->p1 * clock->p2;
-	if (WARN_ON(clock->n + 2 == 0 || clock->p == 0))
-		return 0;
-	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n + 2);
-	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
-
-	return clock->dot;
-}
-
-static int vlv_calc_dpll_params(int refclk, struct dpll *clock)
-{
-	clock->m = clock->m1 * clock->m2;
-	clock->p = clock->p1 * clock->p2;
-	if (WARN_ON(clock->n == 0 || clock->p == 0))
-		return 0;
-	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n);
-	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
-
-	return clock->dot / 5;
-}
-
-int chv_calc_dpll_params(int refclk, struct dpll *clock)
-{
-	clock->m = clock->m1 * clock->m2;
-	clock->p = clock->p1 * clock->p2;
-	if (WARN_ON(clock->n == 0 || clock->p == 0))
-		return 0;
-	clock->vco = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(refclk, clock->m),
-					   clock->n << 22);
-	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
-
-	return clock->dot / 5;
-}
-
-/*
- * Returns whether the given set of divisors are valid for a given refclk with
- * the given connectors.
- */
-static bool intel_pll_is_valid(struct drm_i915_private *dev_priv,
-			       const struct intel_limit *limit,
-			       const struct dpll *clock)
-{
-	if (clock->n < limit->n.min || limit->n.max < clock->n)
-		return false;
-	if (clock->p1 < limit->p1.min || limit->p1.max < clock->p1)
-		return false;
-	if (clock->m2 < limit->m2.min || limit->m2.max < clock->m2)
-		return false;
-	if (clock->m1 < limit->m1.min || limit->m1.max < clock->m1)
-		return false;
-
-	if (!IS_PINEVIEW(dev_priv) && !IS_VALLEYVIEW(dev_priv) &&
-	    !IS_CHERRYVIEW(dev_priv) && !IS_GEN9_LP(dev_priv))
-		if (clock->m1 <= clock->m2)
-			return false;
-
-	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv) &&
-	    !IS_GEN9_LP(dev_priv)) {
-		if (clock->p < limit->p.min || limit->p.max < clock->p)
-			return false;
-		if (clock->m < limit->m.min || limit->m.max < clock->m)
-			return false;
-	}
-
-	if (clock->vco < limit->vco.min || limit->vco.max < clock->vco)
-		return false;
-	/* XXX: We may need to be checking "Dot clock" depending on the multiplier,
-	 * connector, etc., rather than just a single range.
-	 */
-	if (clock->dot < limit->dot.min || limit->dot.max < clock->dot)
-		return false;
-
-	return true;
-}
-
-static int
-i9xx_select_p2_div(const struct intel_limit *limit,
-		   const struct intel_crtc_state *crtc_state,
-		   int target)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		/*
-		 * For LVDS just rely on its current settings for dual-channel.
-		 * We haven't figured out how to reliably set up different
-		 * single/dual channel state, if we even can.
-		 */
-		if (intel_is_dual_link_lvds(dev_priv))
-			return limit->p2.p2_fast;
-		else
-			return limit->p2.p2_slow;
-	} else {
-		if (target < limit->p2.dot_limit)
-			return limit->p2.p2_slow;
-		else
-			return limit->p2.p2_fast;
-	}
-}
-
-/*
- * Returns a set of divisors for the desired target clock with the given
- * refclk, or FALSE.  The returned values represent the clock equation:
- * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
- *
- * Target and reference clocks are specified in kHz.
- *
- * If match_clock is provided, then best_clock P divider must match the P
- * divider from @match_clock used for LVDS downclocking.
- */
-static bool
-i9xx_find_best_dpll(const struct intel_limit *limit,
-		    struct intel_crtc_state *crtc_state,
-		    int target, int refclk, struct dpll *match_clock,
-		    struct dpll *best_clock)
-{
-	struct drm_device *dev = crtc_state->uapi.crtc->dev;
-	struct dpll clock;
-	int err = target;
-
-	memset(best_clock, 0, sizeof(*best_clock));
-
-	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
-
-	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
-	     clock.m1++) {
-		for (clock.m2 = limit->m2.min;
-		     clock.m2 <= limit->m2.max; clock.m2++) {
-			if (clock.m2 >= clock.m1)
-				break;
-			for (clock.n = limit->n.min;
-			     clock.n <= limit->n.max; clock.n++) {
-				for (clock.p1 = limit->p1.min;
-					clock.p1 <= limit->p1.max; clock.p1++) {
-					int this_err;
-
-					i9xx_calc_dpll_params(refclk, &clock);
-					if (!intel_pll_is_valid(to_i915(dev),
-								limit,
-								&clock))
-						continue;
-					if (match_clock &&
-					    clock.p != match_clock->p)
-						continue;
-
-					this_err = abs(clock.dot - target);
-					if (this_err < err) {
-						*best_clock = clock;
-						err = this_err;
-					}
-				}
-			}
-		}
-	}
-
-	return (err != target);
-}
-
-/*
- * Returns a set of divisors for the desired target clock with the given
- * refclk, or FALSE.  The returned values represent the clock equation:
- * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
- *
- * Target and reference clocks are specified in kHz.
- *
- * If match_clock is provided, then best_clock P divider must match the P
- * divider from @match_clock used for LVDS downclocking.
- */
-static bool
-pnv_find_best_dpll(const struct intel_limit *limit,
-		   struct intel_crtc_state *crtc_state,
-		   int target, int refclk, struct dpll *match_clock,
-		   struct dpll *best_clock)
-{
-	struct drm_device *dev = crtc_state->uapi.crtc->dev;
-	struct dpll clock;
-	int err = target;
-
-	memset(best_clock, 0, sizeof(*best_clock));
-
-	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
-
-	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
-	     clock.m1++) {
-		for (clock.m2 = limit->m2.min;
-		     clock.m2 <= limit->m2.max; clock.m2++) {
-			for (clock.n = limit->n.min;
-			     clock.n <= limit->n.max; clock.n++) {
-				for (clock.p1 = limit->p1.min;
-					clock.p1 <= limit->p1.max; clock.p1++) {
-					int this_err;
-
-					pnv_calc_dpll_params(refclk, &clock);
-					if (!intel_pll_is_valid(to_i915(dev),
-								limit,
-								&clock))
-						continue;
-					if (match_clock &&
-					    clock.p != match_clock->p)
-						continue;
-
-					this_err = abs(clock.dot - target);
-					if (this_err < err) {
-						*best_clock = clock;
-						err = this_err;
-					}
-				}
-			}
-		}
-	}
-
-	return (err != target);
-}
-
-/*
- * Returns a set of divisors for the desired target clock with the given
- * refclk, or FALSE.  The returned values represent the clock equation:
- * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
- *
- * Target and reference clocks are specified in kHz.
- *
- * If match_clock is provided, then best_clock P divider must match the P
- * divider from @match_clock used for LVDS downclocking.
- */
-static bool
-g4x_find_best_dpll(const struct intel_limit *limit,
-		   struct intel_crtc_state *crtc_state,
-		   int target, int refclk, struct dpll *match_clock,
-		   struct dpll *best_clock)
-{
-	struct drm_device *dev = crtc_state->uapi.crtc->dev;
-	struct dpll clock;
-	int max_n;
-	bool found = false;
-	/* approximately equals target * 0.00585 */
-	int err_most = (target >> 8) + (target >> 9);
-
-	memset(best_clock, 0, sizeof(*best_clock));
-
-	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
-
-	max_n = limit->n.max;
-	/* based on hardware requirement, prefer smaller n to precision */
-	for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) {
-		/* based on hardware requirement, prefere larger m1,m2 */
-		for (clock.m1 = limit->m1.max;
-		     clock.m1 >= limit->m1.min; clock.m1--) {
-			for (clock.m2 = limit->m2.max;
-			     clock.m2 >= limit->m2.min; clock.m2--) {
-				for (clock.p1 = limit->p1.max;
-				     clock.p1 >= limit->p1.min; clock.p1--) {
-					int this_err;
-
-					i9xx_calc_dpll_params(refclk, &clock);
-					if (!intel_pll_is_valid(to_i915(dev),
-								limit,
-								&clock))
-						continue;
-
-					this_err = abs(clock.dot - target);
-					if (this_err < err_most) {
-						*best_clock = clock;
-						err_most = this_err;
-						max_n = clock.n;
-						found = true;
-					}
-				}
-			}
-		}
-	}
-	return found;
-}
-
-/*
- * Check if the calculated PLL configuration is more optimal compared to the
- * best configuration and error found so far. Return the calculated error.
- */
-static bool vlv_PLL_is_optimal(struct drm_device *dev, int target_freq,
-			       const struct dpll *calculated_clock,
-			       const struct dpll *best_clock,
-			       unsigned int best_error_ppm,
-			       unsigned int *error_ppm)
-{
-	/*
-	 * For CHV ignore the error and consider only the P value.
-	 * Prefer a bigger P value based on HW requirements.
-	 */
-	if (IS_CHERRYVIEW(to_i915(dev))) {
-		*error_ppm = 0;
-
-		return calculated_clock->p > best_clock->p;
-	}
-
-	if (drm_WARN_ON_ONCE(dev, !target_freq))
-		return false;
-
-	*error_ppm = div_u64(1000000ULL *
-				abs(target_freq - calculated_clock->dot),
-			     target_freq);
-	/*
-	 * Prefer a better P value over a better (smaller) error if the error
-	 * is small. Ensure this preference for future configurations too by
-	 * setting the error to 0.
-	 */
-	if (*error_ppm < 100 && calculated_clock->p > best_clock->p) {
-		*error_ppm = 0;
-
-		return true;
-	}
-
-	return *error_ppm + 10 < best_error_ppm;
-}
-
-/*
- * Returns a set of divisors for the desired target clock with the given
- * refclk, or FALSE.  The returned values represent the clock equation:
- * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
- */
-static bool
-vlv_find_best_dpll(const struct intel_limit *limit,
-		   struct intel_crtc_state *crtc_state,
-		   int target, int refclk, struct dpll *match_clock,
-		   struct dpll *best_clock)
-{
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
-	struct drm_device *dev = crtc->base.dev;
-	struct dpll clock;
-	unsigned int bestppm = 1000000;
-	/* min update 19.2 MHz */
-	int max_n = min(limit->n.max, refclk / 19200);
-	bool found = false;
-
-	target *= 5; /* fast clock */
-
-	memset(best_clock, 0, sizeof(*best_clock));
-
-	/* based on hardware requirement, prefer smaller n to precision */
-	for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) {
-		for (clock.p1 = limit->p1.max; clock.p1 >= limit->p1.min; clock.p1--) {
-			for (clock.p2 = limit->p2.p2_fast; clock.p2 >= limit->p2.p2_slow;
-			     clock.p2 -= clock.p2 > 10 ? 2 : 1) {
-				clock.p = clock.p1 * clock.p2;
-				/* based on hardware requirement, prefer bigger m1,m2 values */
-				for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max; clock.m1++) {
-					unsigned int ppm;
-
-					clock.m2 = DIV_ROUND_CLOSEST(target * clock.p * clock.n,
-								     refclk * clock.m1);
-
-					vlv_calc_dpll_params(refclk, &clock);
-
-					if (!intel_pll_is_valid(to_i915(dev),
-								limit,
-								&clock))
-						continue;
-
-					if (!vlv_PLL_is_optimal(dev, target,
-								&clock,
-								best_clock,
-								bestppm, &ppm))
-						continue;
-
-					*best_clock = clock;
-					bestppm = ppm;
-					found = true;
-				}
-			}
-		}
-	}
-
-	return found;
-}
-
-/*
- * Returns a set of divisors for the desired target clock with the given
- * refclk, or FALSE.  The returned values represent the clock equation:
- * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
- */
-static bool
-chv_find_best_dpll(const struct intel_limit *limit,
-		   struct intel_crtc_state *crtc_state,
-		   int target, int refclk, struct dpll *match_clock,
-		   struct dpll *best_clock)
-{
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
-	struct drm_device *dev = crtc->base.dev;
-	unsigned int best_error_ppm;
-	struct dpll clock;
-	u64 m2;
-	int found = false;
-
-	memset(best_clock, 0, sizeof(*best_clock));
-	best_error_ppm = 1000000;
-
-	/*
-	 * Based on hardware doc, the n always set to 1, and m1 always
-	 * set to 2.  If requires to support 200Mhz refclk, we need to
-	 * revisit this because n may not 1 anymore.
-	 */
-	clock.n = 1;
-	clock.m1 = 2;
-	target *= 5;	/* fast clock */
-
-	for (clock.p1 = limit->p1.max; clock.p1 >= limit->p1.min; clock.p1--) {
-		for (clock.p2 = limit->p2.p2_fast;
-				clock.p2 >= limit->p2.p2_slow;
-				clock.p2 -= clock.p2 > 10 ? 2 : 1) {
-			unsigned int error_ppm;
-
-			clock.p = clock.p1 * clock.p2;
-
-			m2 = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(target, clock.p * clock.n) << 22,
-						   refclk * clock.m1);
-
-			if (m2 > INT_MAX/clock.m1)
-				continue;
-
-			clock.m2 = m2;
-
-			chv_calc_dpll_params(refclk, &clock);
-
-			if (!intel_pll_is_valid(to_i915(dev), limit, &clock))
-				continue;
-
-			if (!vlv_PLL_is_optimal(dev, target, &clock, best_clock,
-						best_error_ppm, &error_ppm))
-				continue;
-
-			*best_clock = clock;
-			best_error_ppm = error_ppm;
-			found = true;
-		}
-	}
-
-	return found;
-}
-
-bool bxt_find_best_dpll(struct intel_crtc_state *crtc_state,
-			struct dpll *best_clock)
-{
-	int refclk = 100000;
-	const struct intel_limit *limit = &intel_limits_bxt;
-
-	return chv_find_best_dpll(limit, crtc_state,
-				  crtc_state->port_clock, refclk,
-				  NULL, best_clock);
-}
-
 static bool pipe_scanline_is_moving(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
@@ -1253,12 +495,6 @@ static void assert_planes_disabled(struct intel_crtc *crtc)
 		assert_plane_disabled(plane);
 }
 
-static void assert_vblank_disabled(struct drm_crtc *crtc)
-{
-	if (I915_STATE_WARN_ON(drm_crtc_vblank_get(crtc) == 0))
-		drm_crtc_vblank_put(crtc);
-}
-
 void assert_pch_transcoder_disabled(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
@@ -1610,7 +846,7 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state)
 		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
 		/* Configure frame start delay to match the CPU */
 		val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
-		val |= TRANS_CHICKEN2_FRAME_START_DELAY(0);
+		val |= TRANS_CHICKEN2_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 		intel_de_write(dev_priv, reg, val);
 	}
 
@@ -1621,7 +857,7 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state)
 	if (HAS_PCH_IBX(dev_priv)) {
 		/* Configure frame start delay to match the CPU */
 		val &= ~TRANS_FRAME_START_DELAY_MASK;
-		val |= TRANS_FRAME_START_DELAY(0);
+		val |= TRANS_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 
 		/*
 		 * Make the BPC in transcoder be consistent with
@@ -1666,7 +902,7 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
 	/* Configure frame start delay to match the CPU */
 	val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
-	val |= TRANS_CHICKEN2_FRAME_START_DELAY(0);
+	val |= TRANS_CHICKEN2_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 	intel_de_write(dev_priv, TRANS_CHICKEN2(PIPE_A), val);
 
 	val = TRANS_ENABLE;
@@ -1743,55 +979,6 @@ enum pipe intel_crtc_pch_transcoder(struct intel_crtc *crtc)
 		return crtc->pipe;
 }
 
-static u32 intel_crtc_max_vblank_count(const struct intel_crtc_state *crtc_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
-	u32 mode_flags = crtc->mode_flags;
-
-	/*
-	 * From Gen 11, In case of dsi cmd mode, frame counter wouldnt
-	 * have updated at the beginning of TE, if we want to use
-	 * the hw counter, then we would find it updated in only
-	 * the next TE, hence switching to sw counter.
-	 */
-	if (mode_flags & (I915_MODE_FLAG_DSI_USE_TE0 | I915_MODE_FLAG_DSI_USE_TE1))
-		return 0;
-
-	/*
-	 * On i965gm the hardware frame counter reads
-	 * zero when the TV encoder is enabled :(
-	 */
-	if (IS_I965GM(dev_priv) &&
-	    (crtc_state->output_types & BIT(INTEL_OUTPUT_TVOUT)))
-		return 0;
-
-	if (INTEL_GEN(dev_priv) >= 5 || IS_G4X(dev_priv))
-		return 0xffffffff; /* full 32 bit counter */
-	else if (INTEL_GEN(dev_priv) >= 3)
-		return 0xffffff; /* only 24 bits of frame count */
-	else
-		return 0; /* Gen2 doesn't have a hardware frame counter */
-}
-
-void intel_crtc_vblank_on(const struct intel_crtc_state *crtc_state)
-{
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
-
-	assert_vblank_disabled(&crtc->base);
-	drm_crtc_set_max_vblank_count(&crtc->base,
-				      intel_crtc_max_vblank_count(crtc_state));
-	drm_crtc_vblank_on(&crtc->base);
-}
-
-void intel_crtc_vblank_off(const struct intel_crtc_state *crtc_state)
-{
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
-
-	drm_crtc_vblank_off(&crtc->base);
-	assert_vblank_disabled(&crtc->base);
-}
-
 void intel_enable_pipe(const struct intel_crtc_state *new_crtc_state)
 {
 	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
@@ -1906,8 +1093,8 @@ static bool is_ccs_plane(const struct drm_framebuffer *fb, int plane)
 static bool is_gen12_ccs_modifier(u64 modifier)
 {
 	return modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS ||
+	       modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC ||
 	       modifier == I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS;
-
 }
 
 static bool is_gen12_ccs_plane(const struct drm_framebuffer *fb, int plane)
@@ -1915,6 +1102,12 @@ static bool is_gen12_ccs_plane(const struct drm_framebuffer *fb, int plane)
 	return is_gen12_ccs_modifier(fb->modifier) && is_ccs_plane(fb, plane);
 }
 
+static bool is_gen12_ccs_cc_plane(const struct drm_framebuffer *fb, int plane)
+{
+	return fb->modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC &&
+	       plane == 2;
+}
+
 static bool is_aux_plane(const struct drm_framebuffer *fb, int plane)
 {
 	if (is_ccs_modifier(fb->modifier))
@@ -1936,6 +1129,9 @@ static int ccs_to_main_plane(const struct drm_framebuffer *fb, int ccs_plane)
 	drm_WARN_ON(fb->dev, !is_ccs_modifier(fb->modifier) ||
 		    ccs_plane < fb->format->num_planes / 2);
 
+	if (is_gen12_ccs_cc_plane(fb, ccs_plane))
+		return 0;
+
 	return ccs_plane - fb->format->num_planes / 2;
 }
 
@@ -1954,7 +1150,7 @@ int intel_main_to_aux_plane(const struct drm_framebuffer *fb, int main_plane)
 
 bool
 intel_format_info_is_yuv_semiplanar(const struct drm_format_info *info,
-				    uint64_t modifier)
+				    u64 modifier)
 {
 	return info->is_yuv &&
 	       info->num_planes == (is_ccs_modifier(modifier) ? 4 : 2);
@@ -1986,6 +1182,7 @@ intel_tile_width_bytes(const struct drm_framebuffer *fb, int color_plane)
 			return 128;
 		fallthrough;
 	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
 	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
 		if (is_ccs_plane(fb, color_plane))
 			return 64;
@@ -2120,8 +1317,13 @@ static unsigned int intel_linear_alignment(const struct drm_i915_private *dev_pr
 		return 0;
 }
 
-static unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
-					 int color_plane)
+static bool has_async_flips(struct drm_i915_private *i915)
+{
+	return INTEL_GEN(i915) >= 5;
+}
+
+unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
+				  int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(fb->dev);
 
@@ -2134,7 +1336,7 @@ static unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
 	case DRM_FORMAT_MOD_LINEAR:
 		return intel_linear_alignment(dev_priv);
 	case I915_FORMAT_MOD_X_TILED:
-		if (INTEL_GEN(dev_priv) >= 9)
+		if (has_async_flips(dev_priv))
 			return 256 * 1024;
 		return 0;
 	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
@@ -2142,6 +1344,7 @@ static unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
 			return intel_tile_row_size(fb, color_plane);
 		fallthrough;
 	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
 		return 16 * 1024;
 	case I915_FORMAT_MOD_Y_TILED_CCS:
 	case I915_FORMAT_MOD_Yf_TILED_CCS:
@@ -2247,7 +1450,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb,
 		 */
 		ret = i915_vma_pin_fence(vma);
 		if (ret != 0 && INTEL_GEN(dev_priv) < 4) {
-			i915_gem_object_unpin_from_display_plane(vma);
+			i915_vma_unpin(vma);
 			vma = ERR_PTR(ret);
 			goto err;
 		}
@@ -2265,12 +1468,9 @@ err:
 
 void intel_unpin_fb_vma(struct i915_vma *vma, unsigned long flags)
 {
-	i915_gem_object_lock(vma->obj, NULL);
 	if (flags & PLANE_HAS_FENCE)
 		i915_vma_unpin_fence(vma);
-	i915_gem_object_unpin_from_display_plane(vma);
-	i915_gem_object_unlock(vma->obj);
-
+	i915_vma_unpin(vma);
 	i915_vma_put(vma);
 }
 
@@ -2390,10 +1590,10 @@ static u32 intel_adjust_aligned_offset(int *x, int *y,
  * Adjust the tile offset by moving the difference into
  * the x/y offsets.
  */
-static u32 intel_plane_adjust_aligned_offset(int *x, int *y,
-					     const struct intel_plane_state *state,
-					     int color_plane,
-					     u32 old_offset, u32 new_offset)
+u32 intel_plane_adjust_aligned_offset(int *x, int *y,
+				      const struct intel_plane_state *state,
+				      int color_plane,
+				      u32 old_offset, u32 new_offset)
 {
 	return intel_adjust_aligned_offset(x, y, state->hw.fb, color_plane,
 					   state->hw.rotation,
@@ -2546,6 +1746,7 @@ static unsigned int intel_fb_modifier_to_tiling(u64 fb_modifier)
 	case I915_FORMAT_MOD_Y_TILED:
 	case I915_FORMAT_MOD_Y_TILED_CCS:
 	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
 	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
 		return I915_TILING_Y;
 	default:
@@ -2624,6 +1825,25 @@ static const struct drm_format_info gen12_ccs_formats[] = {
 	  .hsub = 2, .vsub = 2, .is_yuv = true },
 };
 
+/*
+ * Same as gen12_ccs_formats[] above, but with additional surface used
+ * to pass Clear Color information in plane 2 with 64 bits of data.
+ */
+static const struct drm_format_info gen12_ccs_cc_formats[] = {
+	{ .format = DRM_FORMAT_XRGB8888, .depth = 24, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_XBGR8888, .depth = 24, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_ARGB8888, .depth = 32, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+	{ .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+};
+
 static const struct drm_format_info *
 lookup_format_info(const struct drm_format_info formats[],
 		   int num_formats, u32 format)
@@ -2652,6 +1872,10 @@ intel_get_format_info(const struct drm_mode_fb_cmd2 *cmd)
 		return lookup_format_info(gen12_ccs_formats,
 					  ARRAY_SIZE(gen12_ccs_formats),
 					  cmd->pixel_format);
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
+		return lookup_format_info(gen12_ccs_cc_formats,
+					  ARRAY_SIZE(gen12_ccs_cc_formats),
+					  cmd->pixel_format);
 	default:
 		return NULL;
 	}
@@ -2660,6 +1884,7 @@ intel_get_format_info(const struct drm_mode_fb_cmd2 *cmd)
 bool is_ccs_modifier(u64 modifier)
 {
 	return modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS ||
+	       modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC ||
 	       modifier == I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS ||
 	       modifier == I915_FORMAT_MOD_Y_TILED_CCS ||
 	       modifier == I915_FORMAT_MOD_Yf_TILED_CCS;
@@ -2878,7 +2103,7 @@ intel_fb_check_ccs_xy(struct drm_framebuffer *fb, int ccs_plane, int x, int y)
 	int ccs_x, ccs_y;
 	int main_x, main_y;
 
-	if (!is_ccs_plane(fb, ccs_plane))
+	if (!is_ccs_plane(fb, ccs_plane) || is_gen12_ccs_cc_plane(fb, ccs_plane))
 		return 0;
 
 	intel_tile_dims(fb, ccs_plane, &tile_width, &tile_height);
@@ -3005,6 +2230,18 @@ intel_fill_fb_info(struct drm_i915_private *dev_priv,
 		int x, y;
 		int ret;
 
+		/*
+		 * Plane 2 of Render Compression with Clear Color fb modifier
+		 * is consumed by the driver and not passed to DE. Skip the
+		 * arithmetic related to alignment and offset calculation.
+		 */
+		if (is_gen12_ccs_cc_plane(fb, i)) {
+			if (IS_ALIGNED(fb->offsets[i], PAGE_SIZE))
+				continue;
+			else
+				return -EINVAL;
+		}
+
 		cpp = fb->format->cpp[i];
 		intel_fb_plane_dims(&width, &height, fb, i);
 
@@ -3752,33 +2989,19 @@ static int intel_plane_max_height(struct intel_plane *plane,
 		return INT_MAX;
 }
 
-static int skl_check_main_surface(struct intel_plane_state *plane_state)
+int skl_calc_main_surface_offset(const struct intel_plane_state *plane_state,
+				 int *x, int *y, u32 *offset)
 {
 	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
 	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
 	const struct drm_framebuffer *fb = plane_state->hw.fb;
-	unsigned int rotation = plane_state->hw.rotation;
-	int x = plane_state->uapi.src.x1 >> 16;
-	int y = plane_state->uapi.src.y1 >> 16;
-	int w = drm_rect_width(&plane_state->uapi.src) >> 16;
-	int h = drm_rect_height(&plane_state->uapi.src) >> 16;
-	int min_width = intel_plane_min_width(plane, fb, 0, rotation);
-	int max_width = intel_plane_max_width(plane, fb, 0, rotation);
-	int max_height = intel_plane_max_height(plane, fb, 0, rotation);
-	int aux_plane = intel_main_to_aux_plane(fb, 0);
-	u32 aux_offset = plane_state->color_plane[aux_plane].offset;
-	u32 alignment, offset;
-
-	if (w > max_width || w < min_width || h > max_height) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "requested Y/RGB source size %dx%d outside limits (min: %dx1 max: %dx%d)\n",
-			    w, h, min_width, max_width, max_height);
-		return -EINVAL;
-	}
+	const int aux_plane = intel_main_to_aux_plane(fb, 0);
+	const u32 aux_offset = plane_state->color_plane[aux_plane].offset;
+	const u32 alignment = intel_surf_alignment(fb, 0);
+	const int w = drm_rect_width(&plane_state->uapi.src) >> 16;
 
-	intel_add_fb_offsets(&x, &y, plane_state, 0);
-	offset = intel_plane_compute_aligned_offset(&x, &y, plane_state, 0);
-	alignment = intel_surf_alignment(fb, 0);
+	intel_add_fb_offsets(x, y, plane_state, 0);
+	*offset = intel_plane_compute_aligned_offset(x, y, plane_state, 0);
 	if (drm_WARN_ON(&dev_priv->drm, alignment && !is_power_of_2(alignment)))
 		return -EINVAL;
 
@@ -3787,9 +3010,10 @@ static int skl_check_main_surface(struct intel_plane_state *plane_state)
 	 * main surface offset, and it must be non-negative. Make
 	 * sure that is what we will get.
 	 */
-	if (aux_plane && offset > aux_offset)
-		offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0,
-							   offset, aux_offset & ~(alignment - 1));
+	if (aux_plane && *offset > aux_offset)
+		*offset = intel_plane_adjust_aligned_offset(x, y, plane_state, 0,
+							    *offset,
+							    aux_offset & ~(alignment - 1));
 
 	/*
 	 * When using an X-tiled surface, the plane blows up
@@ -3800,18 +3024,51 @@ static int skl_check_main_surface(struct intel_plane_state *plane_state)
 	if (fb->modifier == I915_FORMAT_MOD_X_TILED) {
 		int cpp = fb->format->cpp[0];
 
-		while ((x + w) * cpp > plane_state->color_plane[0].stride) {
-			if (offset == 0) {
+		while ((*x + w) * cpp > plane_state->color_plane[0].stride) {
+			if (*offset == 0) {
 				drm_dbg_kms(&dev_priv->drm,
 					    "Unable to find suitable display surface offset due to X-tiling\n");
 				return -EINVAL;
 			}
 
-			offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0,
-								   offset, offset - alignment);
+			*offset = intel_plane_adjust_aligned_offset(x, y, plane_state, 0,
+								    *offset,
+								    *offset - alignment);
 		}
 	}
 
+	return 0;
+}
+
+static int skl_check_main_surface(struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	const unsigned int rotation = plane_state->hw.rotation;
+	int x = plane_state->uapi.src.x1 >> 16;
+	int y = plane_state->uapi.src.y1 >> 16;
+	const int w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	const int h = drm_rect_height(&plane_state->uapi.src) >> 16;
+	const int min_width = intel_plane_min_width(plane, fb, 0, rotation);
+	const int max_width = intel_plane_max_width(plane, fb, 0, rotation);
+	const int max_height = intel_plane_max_height(plane, fb, 0, rotation);
+	const int aux_plane = intel_main_to_aux_plane(fb, 0);
+	const u32 alignment = intel_surf_alignment(fb, 0);
+	u32 offset;
+	int ret;
+
+	if (w > max_width || w < min_width || h > max_height) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "requested Y/RGB source size %dx%d outside limits (min: %dx1 max: %dx%d)\n",
+			    w, h, min_width, max_width, max_height);
+		return -EINVAL;
+	}
+
+	ret = skl_calc_main_surface_offset(plane_state, &x, &y, &offset);
+	if (ret)
+		return ret;
+
 	/*
 	 * CCS AUX surface doesn't have its own x/y offsets, we must make sure
 	 * they match with the main surface x/y offsets.
@@ -3834,6 +3091,8 @@ static int skl_check_main_surface(struct intel_plane_state *plane_state)
 		}
 	}
 
+	drm_WARN_ON(&dev_priv->drm, x > 8191 || y > 8191);
+
 	plane_state->color_plane[0].offset = offset;
 	plane_state->color_plane[0].x = x;
 	plane_state->color_plane[0].y = y;
@@ -3906,6 +3165,8 @@ static int skl_check_nv12_aux_surface(struct intel_plane_state *plane_state)
 		}
 	}
 
+	drm_WARN_ON(&i915->drm, x > 8191 || y > 8191);
+
 	plane_state->color_plane[uv_plane].offset = offset;
 	plane_state->color_plane[uv_plane].x = x;
 	plane_state->color_plane[uv_plane].y = y;
@@ -3926,7 +3187,8 @@ static int skl_check_ccs_aux_surface(struct intel_plane_state *plane_state)
 		int hsub, vsub;
 		int x, y;
 
-		if (!is_ccs_plane(fb, ccs_plane))
+		if (!is_ccs_plane(fb, ccs_plane) ||
+		    is_gen12_ccs_cc_plane(fb, ccs_plane))
 			continue;
 
 		intel_fb_plane_get_subsampling(&main_hsub, &main_vsub, fb,
@@ -4166,6 +3428,7 @@ static u32 skl_plane_ctl_tiling(u64 fb_modifier)
 	case I915_FORMAT_MOD_Y_TILED:
 		return PLANE_CTL_TILED_Y;
 	case I915_FORMAT_MOD_Y_TILED_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
 		return PLANE_CTL_TILED_Y | PLANE_CTL_RENDER_DECOMPRESSION_ENABLE;
 	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
 		return PLANE_CTL_TILED_Y |
@@ -4226,9 +3489,6 @@ u32 skl_plane_ctl_crtc(const struct intel_crtc_state *crtc_state)
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	u32 plane_ctl = 0;
 
-	if (crtc_state->uapi.async_flip)
-		plane_ctl |= PLANE_CTL_ASYNC_FLIP;
-
 	if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv))
 		return plane_ctl;
 
@@ -4326,6 +3586,8 @@ u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state,
 			plane_color_ctl |= PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE;
 	} else if (fb->format->is_yuv) {
 		plane_color_ctl |= PLANE_COLOR_INPUT_CSC_ENABLE;
+		if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
+			plane_color_ctl |= PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE;
 	}
 
 	return plane_color_ctl;
@@ -4515,532 +3777,6 @@ static void icl_set_pipe_chicken(struct intel_crtc *crtc)
 	intel_de_write(dev_priv, PIPE_CHICKEN(pipe), tmp);
 }
 
-static void intel_fdi_normal_train(struct intel_crtc *crtc)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum pipe pipe = crtc->pipe;
-	i915_reg_t reg;
-	u32 temp;
-
-	/* enable normal train */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	if (IS_IVYBRIDGE(dev_priv)) {
-		temp &= ~FDI_LINK_TRAIN_NONE_IVB;
-		temp |= FDI_LINK_TRAIN_NONE_IVB | FDI_TX_ENHANCE_FRAME_ENABLE;
-	} else {
-		temp &= ~FDI_LINK_TRAIN_NONE;
-		temp |= FDI_LINK_TRAIN_NONE | FDI_TX_ENHANCE_FRAME_ENABLE;
-	}
-	intel_de_write(dev_priv, reg, temp);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	if (HAS_PCH_CPT(dev_priv)) {
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp |= FDI_LINK_TRAIN_NORMAL_CPT;
-	} else {
-		temp &= ~FDI_LINK_TRAIN_NONE;
-		temp |= FDI_LINK_TRAIN_NONE;
-	}
-	intel_de_write(dev_priv, reg, temp | FDI_RX_ENHANCE_FRAME_ENABLE);
-
-	/* wait one idle pattern time */
-	intel_de_posting_read(dev_priv, reg);
-	udelay(1000);
-
-	/* IVB wants error correction enabled */
-	if (IS_IVYBRIDGE(dev_priv))
-		intel_de_write(dev_priv, reg,
-		               intel_de_read(dev_priv, reg) | FDI_FS_ERRC_ENABLE | FDI_FE_ERRC_ENABLE);
-}
-
-/* The FDI link training functions for ILK/Ibexpeak. */
-static void ilk_fdi_link_train(struct intel_crtc *crtc,
-			       const struct intel_crtc_state *crtc_state)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum pipe pipe = crtc->pipe;
-	i915_reg_t reg;
-	u32 temp, tries;
-
-	/* FDI needs bits from pipe first */
-	assert_pipe_enabled(dev_priv, crtc_state->cpu_transcoder);
-
-	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
-	   for train result */
-	reg = FDI_RX_IMR(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_RX_SYMBOL_LOCK;
-	temp &= ~FDI_RX_BIT_LOCK;
-	intel_de_write(dev_priv, reg, temp);
-	intel_de_read(dev_priv, reg);
-	udelay(150);
-
-	/* enable CPU FDI TX and PCH FDI RX */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_DP_PORT_WIDTH_MASK;
-	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_1;
-	intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_1;
-	intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(150);
-
-	/* Ironlake workaround, enable clock pointer after FDI enable*/
-	intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
-		       FDI_RX_PHASE_SYNC_POINTER_OVR);
-	intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
-		       FDI_RX_PHASE_SYNC_POINTER_OVR | FDI_RX_PHASE_SYNC_POINTER_EN);
-
-	reg = FDI_RX_IIR(pipe);
-	for (tries = 0; tries < 5; tries++) {
-		temp = intel_de_read(dev_priv, reg);
-		drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
-
-		if ((temp & FDI_RX_BIT_LOCK)) {
-			drm_dbg_kms(&dev_priv->drm, "FDI train 1 done.\n");
-			intel_de_write(dev_priv, reg, temp | FDI_RX_BIT_LOCK);
-			break;
-		}
-	}
-	if (tries == 5)
-		drm_err(&dev_priv->drm, "FDI train 1 fail!\n");
-
-	/* Train 2 */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_2;
-	intel_de_write(dev_priv, reg, temp);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_2;
-	intel_de_write(dev_priv, reg, temp);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(150);
-
-	reg = FDI_RX_IIR(pipe);
-	for (tries = 0; tries < 5; tries++) {
-		temp = intel_de_read(dev_priv, reg);
-		drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
-
-		if (temp & FDI_RX_SYMBOL_LOCK) {
-			intel_de_write(dev_priv, reg,
-				       temp | FDI_RX_SYMBOL_LOCK);
-			drm_dbg_kms(&dev_priv->drm, "FDI train 2 done.\n");
-			break;
-		}
-	}
-	if (tries == 5)
-		drm_err(&dev_priv->drm, "FDI train 2 fail!\n");
-
-	drm_dbg_kms(&dev_priv->drm, "FDI train done\n");
-
-}
-
-static const int snb_b_fdi_train_param[] = {
-	FDI_LINK_TRAIN_400MV_0DB_SNB_B,
-	FDI_LINK_TRAIN_400MV_6DB_SNB_B,
-	FDI_LINK_TRAIN_600MV_3_5DB_SNB_B,
-	FDI_LINK_TRAIN_800MV_0DB_SNB_B,
-};
-
-/* The FDI link training functions for SNB/Cougarpoint. */
-static void gen6_fdi_link_train(struct intel_crtc *crtc,
-				const struct intel_crtc_state *crtc_state)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum pipe pipe = crtc->pipe;
-	i915_reg_t reg;
-	u32 temp, i, retry;
-
-	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
-	   for train result */
-	reg = FDI_RX_IMR(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_RX_SYMBOL_LOCK;
-	temp &= ~FDI_RX_BIT_LOCK;
-	intel_de_write(dev_priv, reg, temp);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(150);
-
-	/* enable CPU FDI TX and PCH FDI RX */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_DP_PORT_WIDTH_MASK;
-	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_1;
-	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
-	/* SNB-B */
-	temp |= FDI_LINK_TRAIN_400MV_0DB_SNB_B;
-	intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
-
-	intel_de_write(dev_priv, FDI_RX_MISC(pipe),
-		       FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	if (HAS_PCH_CPT(dev_priv)) {
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
-	} else {
-		temp &= ~FDI_LINK_TRAIN_NONE;
-		temp |= FDI_LINK_TRAIN_PATTERN_1;
-	}
-	intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(150);
-
-	for (i = 0; i < 4; i++) {
-		reg = FDI_TX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
-		temp |= snb_b_fdi_train_param[i];
-		intel_de_write(dev_priv, reg, temp);
-
-		intel_de_posting_read(dev_priv, reg);
-		udelay(500);
-
-		for (retry = 0; retry < 5; retry++) {
-			reg = FDI_RX_IIR(pipe);
-			temp = intel_de_read(dev_priv, reg);
-			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
-			if (temp & FDI_RX_BIT_LOCK) {
-				intel_de_write(dev_priv, reg,
-					       temp | FDI_RX_BIT_LOCK);
-				drm_dbg_kms(&dev_priv->drm,
-					    "FDI train 1 done.\n");
-				break;
-			}
-			udelay(50);
-		}
-		if (retry < 5)
-			break;
-	}
-	if (i == 4)
-		drm_err(&dev_priv->drm, "FDI train 1 fail!\n");
-
-	/* Train 2 */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_2;
-	if (IS_GEN(dev_priv, 6)) {
-		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
-		/* SNB-B */
-		temp |= FDI_LINK_TRAIN_400MV_0DB_SNB_B;
-	}
-	intel_de_write(dev_priv, reg, temp);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	if (HAS_PCH_CPT(dev_priv)) {
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp |= FDI_LINK_TRAIN_PATTERN_2_CPT;
-	} else {
-		temp &= ~FDI_LINK_TRAIN_NONE;
-		temp |= FDI_LINK_TRAIN_PATTERN_2;
-	}
-	intel_de_write(dev_priv, reg, temp);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(150);
-
-	for (i = 0; i < 4; i++) {
-		reg = FDI_TX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
-		temp |= snb_b_fdi_train_param[i];
-		intel_de_write(dev_priv, reg, temp);
-
-		intel_de_posting_read(dev_priv, reg);
-		udelay(500);
-
-		for (retry = 0; retry < 5; retry++) {
-			reg = FDI_RX_IIR(pipe);
-			temp = intel_de_read(dev_priv, reg);
-			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
-			if (temp & FDI_RX_SYMBOL_LOCK) {
-				intel_de_write(dev_priv, reg,
-					       temp | FDI_RX_SYMBOL_LOCK);
-				drm_dbg_kms(&dev_priv->drm,
-					    "FDI train 2 done.\n");
-				break;
-			}
-			udelay(50);
-		}
-		if (retry < 5)
-			break;
-	}
-	if (i == 4)
-		drm_err(&dev_priv->drm, "FDI train 2 fail!\n");
-
-	drm_dbg_kms(&dev_priv->drm, "FDI train done.\n");
-}
-
-/* Manual link training for Ivy Bridge A0 parts */
-static void ivb_manual_fdi_link_train(struct intel_crtc *crtc,
-				      const struct intel_crtc_state *crtc_state)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum pipe pipe = crtc->pipe;
-	i915_reg_t reg;
-	u32 temp, i, j;
-
-	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
-	   for train result */
-	reg = FDI_RX_IMR(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_RX_SYMBOL_LOCK;
-	temp &= ~FDI_RX_BIT_LOCK;
-	intel_de_write(dev_priv, reg, temp);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(150);
-
-	drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR before link train 0x%x\n",
-		    intel_de_read(dev_priv, FDI_RX_IIR(pipe)));
-
-	/* Try each vswing and preemphasis setting twice before moving on */
-	for (j = 0; j < ARRAY_SIZE(snb_b_fdi_train_param) * 2; j++) {
-		/* disable first in case we need to retry */
-		reg = FDI_TX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~(FDI_LINK_TRAIN_AUTO | FDI_LINK_TRAIN_NONE_IVB);
-		temp &= ~FDI_TX_ENABLE;
-		intel_de_write(dev_priv, reg, temp);
-
-		reg = FDI_RX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~FDI_LINK_TRAIN_AUTO;
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp &= ~FDI_RX_ENABLE;
-		intel_de_write(dev_priv, reg, temp);
-
-		/* enable CPU FDI TX and PCH FDI RX */
-		reg = FDI_TX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~FDI_DP_PORT_WIDTH_MASK;
-		temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
-		temp |= FDI_LINK_TRAIN_PATTERN_1_IVB;
-		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
-		temp |= snb_b_fdi_train_param[j/2];
-		temp |= FDI_COMPOSITE_SYNC;
-		intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
-
-		intel_de_write(dev_priv, FDI_RX_MISC(pipe),
-			       FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
-
-		reg = FDI_RX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
-		temp |= FDI_COMPOSITE_SYNC;
-		intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
-
-		intel_de_posting_read(dev_priv, reg);
-		udelay(1); /* should be 0.5us */
-
-		for (i = 0; i < 4; i++) {
-			reg = FDI_RX_IIR(pipe);
-			temp = intel_de_read(dev_priv, reg);
-			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
-
-			if (temp & FDI_RX_BIT_LOCK ||
-			    (intel_de_read(dev_priv, reg) & FDI_RX_BIT_LOCK)) {
-				intel_de_write(dev_priv, reg,
-					       temp | FDI_RX_BIT_LOCK);
-				drm_dbg_kms(&dev_priv->drm,
-					    "FDI train 1 done, level %i.\n",
-					    i);
-				break;
-			}
-			udelay(1); /* should be 0.5us */
-		}
-		if (i == 4) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "FDI train 1 fail on vswing %d\n", j / 2);
-			continue;
-		}
-
-		/* Train 2 */
-		reg = FDI_TX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~FDI_LINK_TRAIN_NONE_IVB;
-		temp |= FDI_LINK_TRAIN_PATTERN_2_IVB;
-		intel_de_write(dev_priv, reg, temp);
-
-		reg = FDI_RX_CTL(pipe);
-		temp = intel_de_read(dev_priv, reg);
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp |= FDI_LINK_TRAIN_PATTERN_2_CPT;
-		intel_de_write(dev_priv, reg, temp);
-
-		intel_de_posting_read(dev_priv, reg);
-		udelay(2); /* should be 1.5us */
-
-		for (i = 0; i < 4; i++) {
-			reg = FDI_RX_IIR(pipe);
-			temp = intel_de_read(dev_priv, reg);
-			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
-
-			if (temp & FDI_RX_SYMBOL_LOCK ||
-			    (intel_de_read(dev_priv, reg) & FDI_RX_SYMBOL_LOCK)) {
-				intel_de_write(dev_priv, reg,
-					       temp | FDI_RX_SYMBOL_LOCK);
-				drm_dbg_kms(&dev_priv->drm,
-					    "FDI train 2 done, level %i.\n",
-					    i);
-				goto train_done;
-			}
-			udelay(2); /* should be 1.5us */
-		}
-		if (i == 4)
-			drm_dbg_kms(&dev_priv->drm,
-				    "FDI train 2 fail on vswing %d\n", j / 2);
-	}
-
-train_done:
-	drm_dbg_kms(&dev_priv->drm, "FDI train done.\n");
-}
-
-static void ilk_fdi_pll_enable(const struct intel_crtc_state *crtc_state)
-{
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->uapi.crtc);
-	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
-	enum pipe pipe = intel_crtc->pipe;
-	i915_reg_t reg;
-	u32 temp;
-
-	/* enable PCH FDI RX PLL, wait warmup plus DMI latency */
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~(FDI_DP_PORT_WIDTH_MASK | (0x7 << 16));
-	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
-	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
-	intel_de_write(dev_priv, reg, temp | FDI_RX_PLL_ENABLE);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(200);
-
-	/* Switch from Rawclk to PCDclk */
-	temp = intel_de_read(dev_priv, reg);
-	intel_de_write(dev_priv, reg, temp | FDI_PCDCLK);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(200);
-
-	/* Enable CPU FDI TX PLL, always on for Ironlake */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	if ((temp & FDI_TX_PLL_ENABLE) == 0) {
-		intel_de_write(dev_priv, reg, temp | FDI_TX_PLL_ENABLE);
-
-		intel_de_posting_read(dev_priv, reg);
-		udelay(100);
-	}
-}
-
-static void ilk_fdi_pll_disable(struct intel_crtc *intel_crtc)
-{
-	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum pipe pipe = intel_crtc->pipe;
-	i915_reg_t reg;
-	u32 temp;
-
-	/* Switch from PCDclk to Rawclk */
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	intel_de_write(dev_priv, reg, temp & ~FDI_PCDCLK);
-
-	/* Disable CPU FDI TX PLL */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	intel_de_write(dev_priv, reg, temp & ~FDI_TX_PLL_ENABLE);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(100);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	intel_de_write(dev_priv, reg, temp & ~FDI_RX_PLL_ENABLE);
-
-	/* Wait for the clocks to turn off. */
-	intel_de_posting_read(dev_priv, reg);
-	udelay(100);
-}
-
-static void ilk_fdi_disable(struct intel_crtc *crtc)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	enum pipe pipe = crtc->pipe;
-	i915_reg_t reg;
-	u32 temp;
-
-	/* disable CPU FDI tx and PCH FDI rx */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	intel_de_write(dev_priv, reg, temp & ~FDI_TX_ENABLE);
-	intel_de_posting_read(dev_priv, reg);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~(0x7 << 16);
-	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
-	intel_de_write(dev_priv, reg, temp & ~FDI_RX_ENABLE);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(100);
-
-	/* Ironlake workaround, disable clock pointer after downing FDI */
-	if (HAS_PCH_IBX(dev_priv))
-		intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
-			       FDI_RX_PHASE_SYNC_POINTER_OVR);
-
-	/* still set train pattern 1 */
-	reg = FDI_TX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_PATTERN_1;
-	intel_de_write(dev_priv, reg, temp);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = intel_de_read(dev_priv, reg);
-	if (HAS_PCH_CPT(dev_priv)) {
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
-	} else {
-		temp &= ~FDI_LINK_TRAIN_NONE;
-		temp |= FDI_LINK_TRAIN_PATTERN_1;
-	}
-	/* BPC in FDI rx is consistent with that in PIPECONF */
-	temp &= ~(0x07 << 16);
-	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
-	intel_de_write(dev_priv, reg, temp);
-
-	intel_de_posting_read(dev_priv, reg);
-	udelay(100);
-}
-
 bool intel_has_pending_fb_unpin(struct drm_i915_private *dev_priv)
 {
 	struct drm_crtc *crtc;
@@ -5271,7 +4007,7 @@ static void ivb_update_fdi_bc_bifurcation(const struct intel_crtc_state *crtc_st
  * Finds the encoder associated with the given CRTC. This can only be
  * used when we know that the CRTC isn't feeding multiple encoders!
  */
-static struct intel_encoder *
+struct intel_encoder *
 intel_get_crtc_new_encoder(const struct intel_atomic_state *state,
 			   const struct intel_crtc_state *crtc_state)
 {
@@ -5790,7 +4526,7 @@ static void cnl_program_nearest_filter_coefs(struct drm_i915_private *dev_priv,
 	intel_de_write_fw(dev_priv, CNL_PS_COEF_INDEX_SET(pipe, id, set), 0);
 }
 
-inline u32 skl_scaler_get_filter_select(enum drm_scaling_filter filter, int set)
+u32 skl_scaler_get_filter_select(enum drm_scaling_filter filter, int set)
 {
 	if (filter == DRM_SCALING_FILTER_NEAREST_NEIGHBOR) {
 		return (PS_FILTER_PROGRAMMED |
@@ -6109,41 +4845,72 @@ static void intel_post_plane_update(struct intel_atomic_state *state,
 		icl_wa_scalerclkgating(dev_priv, pipe, false);
 }
 
-static void skl_disable_async_flip_wa(struct intel_atomic_state *state,
-				      struct intel_crtc *crtc,
-				      const struct intel_crtc_state *new_crtc_state)
+static void intel_crtc_enable_flip_done(struct intel_atomic_state *state,
+					struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u8 update_planes = crtc_state->update_planes;
+	const struct intel_plane_state *plane_state;
 	struct intel_plane *plane;
-	struct intel_plane_state *new_plane_state;
 	int i;
 
-	for_each_new_intel_plane_in_state(state, plane, new_plane_state, i) {
-		u32 update_mask = new_crtc_state->update_planes;
-		u32 plane_ctl, surf_addr;
-		enum plane_id plane_id;
-		unsigned long irqflags;
-		enum pipe pipe;
-
-		if (crtc->pipe != plane->pipe ||
-		    !(update_mask & BIT(plane->id)))
-			continue;
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->enable_flip_done &&
+		    plane->pipe == crtc->pipe &&
+		    update_planes & BIT(plane->id))
+			plane->enable_flip_done(plane);
+	}
+}
 
-		plane_id = plane->id;
-		pipe = plane->pipe;
+static void intel_crtc_disable_flip_done(struct intel_atomic_state *state,
+					 struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u8 update_planes = crtc_state->update_planes;
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
 
-		spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
-		plane_ctl = intel_de_read_fw(dev_priv, PLANE_CTL(pipe, plane_id));
-		surf_addr = intel_de_read_fw(dev_priv, PLANE_SURF(pipe, plane_id));
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->disable_flip_done &&
+		    plane->pipe == crtc->pipe &&
+		    update_planes & BIT(plane->id))
+			plane->disable_flip_done(plane);
+	}
+}
 
-		plane_ctl &= ~PLANE_CTL_ASYNC_FLIP;
+static void intel_crtc_async_flip_disable_wa(struct intel_atomic_state *state,
+					     struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u8 update_planes = new_crtc_state->update_planes;
+	const struct intel_plane_state *old_plane_state;
+	struct intel_plane *plane;
+	bool need_vbl_wait = false;
+	int i;
 
-		intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), plane_ctl);
-		intel_de_write_fw(dev_priv, PLANE_SURF(pipe, plane_id), surf_addr);
-		spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+	for_each_old_intel_plane_in_state(state, plane, old_plane_state, i) {
+		if (plane->need_async_flip_disable_wa &&
+		    plane->pipe == crtc->pipe &&
+		    update_planes & BIT(plane->id)) {
+			/*
+			 * Apart from the async flip bit we want to
+			 * preserve the old state for the plane.
+			 */
+			plane->async_flip(plane, old_crtc_state,
+					  old_plane_state, false);
+			need_vbl_wait = true;
+		}
 	}
 
-	intel_wait_for_vblank(dev_priv, crtc->pipe);
+	if (need_vbl_wait)
+		intel_wait_for_vblank(i915, crtc->pipe);
 }
 
 static void intel_pre_plane_update(struct intel_atomic_state *state,
@@ -6236,10 +5003,8 @@ static void intel_pre_plane_update(struct intel_atomic_state *state,
 	 * WA for platforms where async address update enable bit
 	 * is double buffered and only latched at start of vblank.
 	 */
-	if (old_crtc_state->uapi.async_flip &&
-	    !new_crtc_state->uapi.async_flip &&
-	    IS_GEN_RANGE(dev_priv, 9, 10))
-		skl_disable_async_flip_wa(state, crtc, new_crtc_state);
+	if (old_crtc_state->uapi.async_flip && !new_crtc_state->uapi.async_flip)
+		intel_crtc_async_flip_disable_wa(state, crtc);
 }
 
 static void intel_crtc_disable_planes(struct intel_atomic_state *state,
@@ -6659,7 +5424,7 @@ static void hsw_set_frame_start_delay(const struct intel_crtc_state *crtc_state)
 
 	val = intel_de_read(dev_priv, reg);
 	val &= ~HSW_FRAME_START_DELAY_MASK;
-	val |= HSW_FRAME_START_DELAY(0);
+	val |= HSW_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 	intel_de_write(dev_priv, reg, val);
 }
 
@@ -7447,143 +6212,6 @@ static void intel_connector_verify_state(struct intel_crtc_state *crtc_state,
 	}
 }
 
-static int pipe_required_fdi_lanes(struct intel_crtc_state *crtc_state)
-{
-	if (crtc_state->hw.enable && crtc_state->has_pch_encoder)
-		return crtc_state->fdi_lanes;
-
-	return 0;
-}
-
-static int ilk_check_fdi_lanes(struct drm_device *dev, enum pipe pipe,
-			       struct intel_crtc_state *pipe_config)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct drm_atomic_state *state = pipe_config->uapi.state;
-	struct intel_crtc *other_crtc;
-	struct intel_crtc_state *other_crtc_state;
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "checking fdi config on pipe %c, lanes %i\n",
-		    pipe_name(pipe), pipe_config->fdi_lanes);
-	if (pipe_config->fdi_lanes > 4) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "invalid fdi lane config on pipe %c: %i lanes\n",
-			    pipe_name(pipe), pipe_config->fdi_lanes);
-		return -EINVAL;
-	}
-
-	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
-		if (pipe_config->fdi_lanes > 2) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "only 2 lanes on haswell, required: %i lanes\n",
-				    pipe_config->fdi_lanes);
-			return -EINVAL;
-		} else {
-			return 0;
-		}
-	}
-
-	if (INTEL_NUM_PIPES(dev_priv) == 2)
-		return 0;
-
-	/* Ivybridge 3 pipe is really complicated */
-	switch (pipe) {
-	case PIPE_A:
-		return 0;
-	case PIPE_B:
-		if (pipe_config->fdi_lanes <= 2)
-			return 0;
-
-		other_crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_C);
-		other_crtc_state =
-			intel_atomic_get_crtc_state(state, other_crtc);
-		if (IS_ERR(other_crtc_state))
-			return PTR_ERR(other_crtc_state);
-
-		if (pipe_required_fdi_lanes(other_crtc_state) > 0) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "invalid shared fdi lane config on pipe %c: %i lanes\n",
-				    pipe_name(pipe), pipe_config->fdi_lanes);
-			return -EINVAL;
-		}
-		return 0;
-	case PIPE_C:
-		if (pipe_config->fdi_lanes > 2) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "only 2 lanes on pipe %c: required %i lanes\n",
-				    pipe_name(pipe), pipe_config->fdi_lanes);
-			return -EINVAL;
-		}
-
-		other_crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_B);
-		other_crtc_state =
-			intel_atomic_get_crtc_state(state, other_crtc);
-		if (IS_ERR(other_crtc_state))
-			return PTR_ERR(other_crtc_state);
-
-		if (pipe_required_fdi_lanes(other_crtc_state) > 2) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "fdi link B uses too many lanes to enable link C\n");
-			return -EINVAL;
-		}
-		return 0;
-	default:
-		BUG();
-	}
-}
-
-#define RETRY 1
-static int ilk_fdi_compute_config(struct intel_crtc *intel_crtc,
-				  struct intel_crtc_state *pipe_config)
-{
-	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *i915 = to_i915(dev);
-	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
-	int lane, link_bw, fdi_dotclock, ret;
-	bool needs_recompute = false;
-
-retry:
-	/* FDI is a binary signal running at ~2.7GHz, encoding
-	 * each output octet as 10 bits. The actual frequency
-	 * is stored as a divider into a 100MHz clock, and the
-	 * mode pixel clock is stored in units of 1KHz.
-	 * Hence the bw of each lane in terms of the mode signal
-	 * is:
-	 */
-	link_bw = intel_fdi_link_freq(i915, pipe_config);
-
-	fdi_dotclock = adjusted_mode->crtc_clock;
-
-	lane = ilk_get_lanes_required(fdi_dotclock, link_bw,
-				      pipe_config->pipe_bpp);
-
-	pipe_config->fdi_lanes = lane;
-
-	intel_link_compute_m_n(pipe_config->pipe_bpp, lane, fdi_dotclock,
-			       link_bw, &pipe_config->fdi_m_n, false, false);
-
-	ret = ilk_check_fdi_lanes(dev, intel_crtc->pipe, pipe_config);
-	if (ret == -EDEADLK)
-		return ret;
-
-	if (ret == -EINVAL && pipe_config->pipe_bpp > 6*3) {
-		pipe_config->pipe_bpp -= 2*3;
-		drm_dbg_kms(&i915->drm,
-			    "fdi link bw constraint, reducing pipe bpp to %i\n",
-			    pipe_config->pipe_bpp);
-		needs_recompute = true;
-		pipe_config->bw_constrained = true;
-
-		goto retry;
-	}
-
-	if (needs_recompute)
-		return RETRY;
-
-	return ret;
-}
-
 bool hsw_crtc_state_ips_capable(const struct intel_crtc_state *crtc_state)
 {
 	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
@@ -7815,19 +6443,6 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 		return -EINVAL;
 	}
 
-	if ((pipe_config->output_format == INTEL_OUTPUT_FORMAT_YCBCR420 ||
-	     pipe_config->output_format == INTEL_OUTPUT_FORMAT_YCBCR444) &&
-	     pipe_config->hw.ctm) {
-		/*
-		 * There is only one pipe CSC unit per pipe, and we need that
-		 * for output conversion from RGB->YCBCR. So if CTM is already
-		 * applied we can't support YCBCR420 output.
-		 */
-		drm_dbg_kms(&dev_priv->drm,
-			    "YCBCR420 and CTM together are not possible\n");
-		return -EINVAL;
-	}
-
 	/*
 	 * Pipe horizontal size must be even in:
 	 * - DVO ganged mode
@@ -7939,51 +6554,6 @@ static void intel_panel_sanitize_ssc(struct drm_i915_private *dev_priv)
 	}
 }
 
-static bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
-{
-	if (dev_priv->params.panel_use_ssc >= 0)
-		return dev_priv->params.panel_use_ssc != 0;
-	return dev_priv->vbt.lvds_use_ssc
-		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
-}
-
-static u32 pnv_dpll_compute_fp(struct dpll *dpll)
-{
-	return (1 << dpll->n) << 16 | dpll->m2;
-}
-
-static u32 i9xx_dpll_compute_fp(struct dpll *dpll)
-{
-	return dpll->n << 16 | dpll->m1 << 8 | dpll->m2;
-}
-
-static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
-				     struct intel_crtc_state *crtc_state,
-				     struct dpll *reduced_clock)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	u32 fp, fp2 = 0;
-
-	if (IS_PINEVIEW(dev_priv)) {
-		fp = pnv_dpll_compute_fp(&crtc_state->dpll);
-		if (reduced_clock)
-			fp2 = pnv_dpll_compute_fp(reduced_clock);
-	} else {
-		fp = i9xx_dpll_compute_fp(&crtc_state->dpll);
-		if (reduced_clock)
-			fp2 = i9xx_dpll_compute_fp(reduced_clock);
-	}
-
-	crtc_state->dpll_hw_state.fp0 = fp;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
-	    reduced_clock) {
-		crtc_state->dpll_hw_state.fp1 = fp2;
-	} else {
-		crtc_state->dpll_hw_state.fp1 = fp;
-	}
-}
-
 static void vlv_pllb_recal_opamp(struct drm_i915_private *dev_priv, enum pipe
 		pipe)
 {
@@ -8108,39 +6678,6 @@ void intel_dp_set_m_n(const struct intel_crtc_state *crtc_state, enum link_m_n_s
 		intel_cpu_transcoder_set_m_n(crtc_state, dp_m_n, dp_m2_n2);
 }
 
-static void vlv_compute_dpll(struct intel_crtc *crtc,
-			     struct intel_crtc_state *pipe_config)
-{
-	pipe_config->dpll_hw_state.dpll = DPLL_INTEGRATED_REF_CLK_VLV |
-		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
-	if (crtc->pipe != PIPE_A)
-		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
-
-	/* DPLL not used with DSI, but still need the rest set up */
-	if (!intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DSI))
-		pipe_config->dpll_hw_state.dpll |= DPLL_VCO_ENABLE |
-			DPLL_EXT_BUFFER_ENABLE_VLV;
-
-	pipe_config->dpll_hw_state.dpll_md =
-		(pipe_config->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
-}
-
-static void chv_compute_dpll(struct intel_crtc *crtc,
-			     struct intel_crtc_state *pipe_config)
-{
-	pipe_config->dpll_hw_state.dpll = DPLL_SSC_REF_CLK_CHV |
-		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
-	if (crtc->pipe != PIPE_A)
-		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
-
-	/* DPLL not used with DSI, but still need the rest set up */
-	if (!intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DSI))
-		pipe_config->dpll_hw_state.dpll |= DPLL_VCO_ENABLE;
-
-	pipe_config->dpll_hw_state.dpll_md =
-		(pipe_config->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
-}
-
 static void vlv_prepare_pll(struct intel_crtc *crtc,
 			    const struct intel_crtc_state *pipe_config)
 {
@@ -8400,128 +6937,7 @@ void vlv_force_pll_off(struct drm_i915_private *dev_priv, enum pipe pipe)
 		vlv_disable_pll(dev_priv, pipe);
 }
 
-static void i9xx_compute_dpll(struct intel_crtc *crtc,
-			      struct intel_crtc_state *crtc_state,
-			      struct dpll *reduced_clock)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	u32 dpll;
-	struct dpll *clock = &crtc_state->dpll;
-
-	i9xx_update_pll_dividers(crtc, crtc_state, reduced_clock);
-
-	dpll = DPLL_VGA_MODE_DIS;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
-		dpll |= DPLLB_MODE_LVDS;
-	else
-		dpll |= DPLLB_MODE_DAC_SERIAL;
-
-	if (IS_I945G(dev_priv) || IS_I945GM(dev_priv) ||
-	    IS_G33(dev_priv) || IS_PINEVIEW(dev_priv)) {
-		dpll |= (crtc_state->pixel_multiplier - 1)
-			<< SDVO_MULTIPLIER_SHIFT_HIRES;
-	}
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
-	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
-		dpll |= DPLL_SDVO_HIGH_SPEED;
-
-	if (intel_crtc_has_dp_encoder(crtc_state))
-		dpll |= DPLL_SDVO_HIGH_SPEED;
-
-	/* compute bitmask from p1 value */
-	if (IS_PINEVIEW(dev_priv))
-		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT_PINEVIEW;
-	else {
-		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
-		if (IS_G4X(dev_priv) && reduced_clock)
-			dpll |= (1 << (reduced_clock->p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
-	}
-	switch (clock->p2) {
-	case 5:
-		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
-		break;
-	case 7:
-		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7;
-		break;
-	case 10:
-		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10;
-		break;
-	case 14:
-		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14;
-		break;
-	}
-	if (INTEL_GEN(dev_priv) >= 4)
-		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
-
-	if (crtc_state->sdvo_tv_clock)
-		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
-		 intel_panel_use_ssc(dev_priv))
-		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
-	else
-		dpll |= PLL_REF_INPUT_DREFCLK;
-
-	dpll |= DPLL_VCO_ENABLE;
-	crtc_state->dpll_hw_state.dpll = dpll;
-
-	if (INTEL_GEN(dev_priv) >= 4) {
-		u32 dpll_md = (crtc_state->pixel_multiplier - 1)
-			<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-		crtc_state->dpll_hw_state.dpll_md = dpll_md;
-	}
-}
-
-static void i8xx_compute_dpll(struct intel_crtc *crtc,
-			      struct intel_crtc_state *crtc_state,
-			      struct dpll *reduced_clock)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	u32 dpll;
-	struct dpll *clock = &crtc_state->dpll;
-
-	i9xx_update_pll_dividers(crtc, crtc_state, reduced_clock);
 
-	dpll = DPLL_VGA_MODE_DIS;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
-	} else {
-		if (clock->p1 == 2)
-			dpll |= PLL_P1_DIVIDE_BY_TWO;
-		else
-			dpll |= (clock->p1 - 2) << DPLL_FPA01_P1_POST_DIV_SHIFT;
-		if (clock->p2 == 4)
-			dpll |= PLL_P2_DIVIDE_BY_4;
-	}
-
-	/*
-	 * Bspec:
-	 * "[Almador Errata}: For the correct operation of the muxed DVO pins
-	 *  (GDEVSELB/I2Cdata, GIRDBY/I2CClk) and (GFRAMEB/DVI_Data,
-	 *  GTRDYB/DVI_Clk): Bit 31 (DPLL VCO Enable) and Bit 30 (2X Clock
-	 *  Enable) must be set to “1� in both the DPLL A Control Register
-	 *  (06014h-06017h) and DPLL B Control Register (06018h-0601Bh)."
-	 *
-	 * For simplicity We simply keep both bits always enabled in
-	 * both DPLLS. The spec says we should disable the DVO 2X clock
-	 * when not needed, but this seems to work fine in practice.
-	 */
-	if (IS_I830(dev_priv) ||
-	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO))
-		dpll |= DPLL_DVO_2X_MODE;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
-	    intel_panel_use_ssc(dev_priv))
-		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
-	else
-		dpll |= PLL_REF_INPUT_DREFCLK;
-
-	dpll |= DPLL_VCO_ENABLE;
-	crtc_state->dpll_hw_state.dpll = dpll;
-}
 
 static void intel_set_transcoder_timings(const struct intel_crtc_state *crtc_state)
 {
@@ -8721,214 +7137,12 @@ static void i9xx_set_pipeconf(const struct intel_crtc_state *crtc_state)
 
 	pipeconf |= PIPECONF_GAMMA_MODE(crtc_state->gamma_mode);
 
-	pipeconf |= PIPECONF_FRAME_START_DELAY(0);
+	pipeconf |= PIPECONF_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 
 	intel_de_write(dev_priv, PIPECONF(crtc->pipe), pipeconf);
 	intel_de_posting_read(dev_priv, PIPECONF(crtc->pipe));
 }
 
-static int i8xx_crtc_compute_clock(struct intel_crtc *crtc,
-				   struct intel_crtc_state *crtc_state)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	const struct intel_limit *limit;
-	int refclk = 48000;
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		if (intel_panel_use_ssc(dev_priv)) {
-			refclk = dev_priv->vbt.lvds_ssc_freq;
-			drm_dbg_kms(&dev_priv->drm,
-				    "using SSC reference clock of %d kHz\n",
-				    refclk);
-		}
-
-		limit = &intel_limits_i8xx_lvds;
-	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO)) {
-		limit = &intel_limits_i8xx_dvo;
-	} else {
-		limit = &intel_limits_i8xx_dac;
-	}
-
-	if (!crtc_state->clock_set &&
-	    !i9xx_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				 refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&dev_priv->drm,
-			"Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	i8xx_compute_dpll(crtc, crtc_state, NULL);
-
-	return 0;
-}
-
-static int g4x_crtc_compute_clock(struct intel_crtc *crtc,
-				  struct intel_crtc_state *crtc_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	const struct intel_limit *limit;
-	int refclk = 96000;
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		if (intel_panel_use_ssc(dev_priv)) {
-			refclk = dev_priv->vbt.lvds_ssc_freq;
-			drm_dbg_kms(&dev_priv->drm,
-				    "using SSC reference clock of %d kHz\n",
-				    refclk);
-		}
-
-		if (intel_is_dual_link_lvds(dev_priv))
-			limit = &intel_limits_g4x_dual_channel_lvds;
-		else
-			limit = &intel_limits_g4x_single_channel_lvds;
-	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI) ||
-		   intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG)) {
-		limit = &intel_limits_g4x_hdmi;
-	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO)) {
-		limit = &intel_limits_g4x_sdvo;
-	} else {
-		/* The option is for other outputs */
-		limit = &intel_limits_i9xx_sdvo;
-	}
-
-	if (!crtc_state->clock_set &&
-	    !g4x_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&dev_priv->drm,
-			"Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	i9xx_compute_dpll(crtc, crtc_state, NULL);
-
-	return 0;
-}
-
-static int pnv_crtc_compute_clock(struct intel_crtc *crtc,
-				  struct intel_crtc_state *crtc_state)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	const struct intel_limit *limit;
-	int refclk = 96000;
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		if (intel_panel_use_ssc(dev_priv)) {
-			refclk = dev_priv->vbt.lvds_ssc_freq;
-			drm_dbg_kms(&dev_priv->drm,
-				    "using SSC reference clock of %d kHz\n",
-				    refclk);
-		}
-
-		limit = &pnv_limits_lvds;
-	} else {
-		limit = &pnv_limits_sdvo;
-	}
-
-	if (!crtc_state->clock_set &&
-	    !pnv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&dev_priv->drm,
-			"Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	i9xx_compute_dpll(crtc, crtc_state, NULL);
-
-	return 0;
-}
-
-static int i9xx_crtc_compute_clock(struct intel_crtc *crtc,
-				   struct intel_crtc_state *crtc_state)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	const struct intel_limit *limit;
-	int refclk = 96000;
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		if (intel_panel_use_ssc(dev_priv)) {
-			refclk = dev_priv->vbt.lvds_ssc_freq;
-			drm_dbg_kms(&dev_priv->drm,
-				    "using SSC reference clock of %d kHz\n",
-				    refclk);
-		}
-
-		limit = &intel_limits_i9xx_lvds;
-	} else {
-		limit = &intel_limits_i9xx_sdvo;
-	}
-
-	if (!crtc_state->clock_set &&
-	    !i9xx_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				 refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&dev_priv->drm,
-			"Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	i9xx_compute_dpll(crtc, crtc_state, NULL);
-
-	return 0;
-}
-
-static int chv_crtc_compute_clock(struct intel_crtc *crtc,
-				  struct intel_crtc_state *crtc_state)
-{
-	int refclk = 100000;
-	const struct intel_limit *limit = &intel_limits_chv;
-	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	if (!crtc_state->clock_set &&
-	    !chv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&i915->drm, "Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	chv_compute_dpll(crtc, crtc_state);
-
-	return 0;
-}
-
-static int vlv_crtc_compute_clock(struct intel_crtc *crtc,
-				  struct intel_crtc_state *crtc_state)
-{
-	int refclk = 100000;
-	const struct intel_limit *limit = &intel_limits_vlv;
-	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	if (!crtc_state->clock_set &&
-	    !vlv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&i915->drm,  "Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	vlv_compute_dpll(crtc, crtc_state);
-
-	return 0;
-}
-
 static bool i9xx_has_pfit(struct drm_i915_private *dev_priv)
 {
 	if (IS_I830(dev_priv))
@@ -9830,7 +8044,7 @@ static void ilk_set_pipeconf(const struct intel_crtc_state *crtc_state)
 
 	val |= PIPECONF_GAMMA_MODE(crtc_state->gamma_mode);
 
-	val |= PIPECONF_FRAME_START_DELAY(0);
+	val |= PIPECONF_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 
 	intel_de_write(dev_priv, PIPECONF(pipe), val);
 	intel_de_posting_read(dev_priv, PIPECONF(pipe));
@@ -9938,172 +8152,6 @@ int ilk_get_lanes_required(int target_clock, int link_bw, int bpp)
 	return DIV_ROUND_UP(bps, link_bw * 8);
 }
 
-static bool ilk_needs_fb_cb_tune(struct dpll *dpll, int factor)
-{
-	return i9xx_dpll_compute_m(dpll) < factor * dpll->n;
-}
-
-static void ilk_compute_dpll(struct intel_crtc *crtc,
-			     struct intel_crtc_state *crtc_state,
-			     struct dpll *reduced_clock)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	u32 dpll, fp, fp2;
-	int factor;
-
-	/* Enable autotuning of the PLL clock (if permissible) */
-	factor = 21;
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		if ((intel_panel_use_ssc(dev_priv) &&
-		     dev_priv->vbt.lvds_ssc_freq == 100000) ||
-		    (HAS_PCH_IBX(dev_priv) &&
-		     intel_is_dual_link_lvds(dev_priv)))
-			factor = 25;
-	} else if (crtc_state->sdvo_tv_clock) {
-		factor = 20;
-	}
-
-	fp = i9xx_dpll_compute_fp(&crtc_state->dpll);
-
-	if (ilk_needs_fb_cb_tune(&crtc_state->dpll, factor))
-		fp |= FP_CB_TUNE;
-
-	if (reduced_clock) {
-		fp2 = i9xx_dpll_compute_fp(reduced_clock);
-
-		if (reduced_clock->m < factor * reduced_clock->n)
-			fp2 |= FP_CB_TUNE;
-	} else {
-		fp2 = fp;
-	}
-
-	dpll = 0;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
-		dpll |= DPLLB_MODE_LVDS;
-	else
-		dpll |= DPLLB_MODE_DAC_SERIAL;
-
-	dpll |= (crtc_state->pixel_multiplier - 1)
-		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
-	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
-		dpll |= DPLL_SDVO_HIGH_SPEED;
-
-	if (intel_crtc_has_dp_encoder(crtc_state))
-		dpll |= DPLL_SDVO_HIGH_SPEED;
-
-	/*
-	 * The high speed IO clock is only really required for
-	 * SDVO/HDMI/DP, but we also enable it for CRT to make it
-	 * possible to share the DPLL between CRT and HDMI. Enabling
-	 * the clock needlessly does no real harm, except use up a
-	 * bit of power potentially.
-	 *
-	 * We'll limit this to IVB with 3 pipes, since it has only two
-	 * DPLLs and so DPLL sharing is the only way to get three pipes
-	 * driving PCH ports at the same time. On SNB we could do this,
-	 * and potentially avoid enabling the second DPLL, but it's not
-	 * clear if it''s a win or loss power wise. No point in doing
-	 * this on ILK at all since it has a fixed DPLL<->pipe mapping.
-	 */
-	if (INTEL_NUM_PIPES(dev_priv) == 3 &&
-	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
-		dpll |= DPLL_SDVO_HIGH_SPEED;
-
-	/* compute bitmask from p1 value */
-	dpll |= (1 << (crtc_state->dpll.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
-	/* also FPA1 */
-	dpll |= (1 << (crtc_state->dpll.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
-
-	switch (crtc_state->dpll.p2) {
-	case 5:
-		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
-		break;
-	case 7:
-		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7;
-		break;
-	case 10:
-		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10;
-		break;
-	case 14:
-		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14;
-		break;
-	}
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
-	    intel_panel_use_ssc(dev_priv))
-		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
-	else
-		dpll |= PLL_REF_INPUT_DREFCLK;
-
-	dpll |= DPLL_VCO_ENABLE;
-
-	crtc_state->dpll_hw_state.dpll = dpll;
-	crtc_state->dpll_hw_state.fp0 = fp;
-	crtc_state->dpll_hw_state.fp1 = fp2;
-}
-
-static int ilk_crtc_compute_clock(struct intel_crtc *crtc,
-				  struct intel_crtc_state *crtc_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	struct intel_atomic_state *state =
-		to_intel_atomic_state(crtc_state->uapi.state);
-	const struct intel_limit *limit;
-	int refclk = 120000;
-
-	memset(&crtc_state->dpll_hw_state, 0,
-	       sizeof(crtc_state->dpll_hw_state));
-
-	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
-	if (!crtc_state->has_pch_encoder)
-		return 0;
-
-	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
-		if (intel_panel_use_ssc(dev_priv)) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "using SSC reference clock of %d kHz\n",
-				    dev_priv->vbt.lvds_ssc_freq);
-			refclk = dev_priv->vbt.lvds_ssc_freq;
-		}
-
-		if (intel_is_dual_link_lvds(dev_priv)) {
-			if (refclk == 100000)
-				limit = &ilk_limits_dual_lvds_100m;
-			else
-				limit = &ilk_limits_dual_lvds;
-		} else {
-			if (refclk == 100000)
-				limit = &ilk_limits_single_lvds_100m;
-			else
-				limit = &ilk_limits_single_lvds;
-		}
-	} else {
-		limit = &ilk_limits_dac;
-	}
-
-	if (!crtc_state->clock_set &&
-	    !g4x_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
-				refclk, NULL, &crtc_state->dpll)) {
-		drm_err(&dev_priv->drm,
-			"Couldn't find PLL settings for mode!\n");
-		return -EINVAL;
-	}
-
-	ilk_compute_dpll(crtc, crtc_state, NULL);
-
-	if (!intel_reserve_shared_dplls(state, crtc, NULL)) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "failed to find PLL for pipe %c\n",
-			    pipe_name(crtc->pipe));
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
 static void intel_pch_transcoder_get_m_n(struct intel_crtc *crtc,
 					 struct intel_link_m_n *m_n)
 {
@@ -10516,29 +8564,6 @@ out:
 	return ret;
 }
 
-static int hsw_crtc_compute_clock(struct intel_crtc *crtc,
-				  struct intel_crtc_state *crtc_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	struct intel_atomic_state *state =
-		to_intel_atomic_state(crtc_state->uapi.state);
-
-	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI) ||
-	    INTEL_GEN(dev_priv) >= 11) {
-		struct intel_encoder *encoder =
-			intel_get_crtc_new_encoder(state, crtc_state);
-
-		if (!intel_reserve_shared_dplls(state, crtc, encoder)) {
-			drm_dbg_kms(&dev_priv->drm,
-				    "failed to find PLL for pipe %c\n",
-				    pipe_name(crtc->pipe));
-			return -EINVAL;
-		}
-	}
-
-	return 0;
-}
-
 static void dg1_get_ddi_pll(struct drm_i915_private *dev_priv, enum port port,
 			    struct intel_crtc_state *pipe_config)
 {
@@ -10932,8 +8957,6 @@ static bool hsw_get_pipe_config(struct intel_crtc *crtc,
 	bool active;
 	u32 tmp;
 
-	pipe_config->master_transcoder = INVALID_TRANSCODER;
-
 	if (!intel_display_power_get_in_set_if_enabled(dev_priv, &power_domain_set,
 						       POWER_DOMAIN_PIPE(crtc->pipe)))
 		return false;
@@ -10966,6 +8989,9 @@ static bool hsw_get_pipe_config(struct intel_crtc *crtc,
 		intel_get_transcoder_timings(crtc, pipe_config);
 	}
 
+	if (HAS_VRR(dev_priv) && !transcoder_is_dsi(pipe_config->cpu_transcoder))
+		intel_vrr_get_config(crtc, pipe_config);
+
 	intel_get_pipe_src_size(crtc, pipe_config);
 
 	if (IS_HASWELL(dev_priv)) {
@@ -11444,33 +9470,6 @@ static void ilk_pch_clock_get(struct intel_crtc *crtc,
 					 &pipe_config->fdi_m_n);
 }
 
-static void intel_crtc_state_reset(struct intel_crtc_state *crtc_state,
-				   struct intel_crtc *crtc)
-{
-	memset(crtc_state, 0, sizeof(*crtc_state));
-
-	__drm_atomic_helper_crtc_state_reset(&crtc_state->uapi, &crtc->base);
-
-	crtc_state->cpu_transcoder = INVALID_TRANSCODER;
-	crtc_state->master_transcoder = INVALID_TRANSCODER;
-	crtc_state->hsw_workaround_pipe = INVALID_PIPE;
-	crtc_state->output_format = INTEL_OUTPUT_FORMAT_INVALID;
-	crtc_state->scaler_state.scaler_id = -1;
-	crtc_state->mst_master_transcoder = INVALID_TRANSCODER;
-}
-
-static struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc)
-{
-	struct intel_crtc_state *crtc_state;
-
-	crtc_state = kmalloc(sizeof(*crtc_state), GFP_KERNEL);
-
-	if (crtc_state)
-		intel_crtc_state_reset(crtc_state, crtc);
-
-	return crtc_state;
-}
-
 /* Returns the currently programmed mode of the given encoder. */
 struct drm_display_mode *
 intel_encoder_current_mode(struct intel_encoder *encoder)
@@ -11511,14 +9510,6 @@ intel_encoder_current_mode(struct intel_encoder *encoder)
 	return mode;
 }
 
-static void intel_crtc_destroy(struct drm_crtc *crtc)
-{
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	drm_crtc_cleanup(crtc);
-	kfree(intel_crtc);
-}
-
 /**
  * intel_wm_need_update - Check whether watermarks need updating
  * @cur: current plane state
@@ -12220,7 +10211,6 @@ static void snprintf_output_types(char *buf, size_t len,
 }
 
 static const char * const output_format_str[] = {
-	[INTEL_OUTPUT_FORMAT_INVALID] = "Invalid",
 	[INTEL_OUTPUT_FORMAT_RGB] = "RGB",
 	[INTEL_OUTPUT_FORMAT_YCBCR420] = "YCBCR4:2:0",
 	[INTEL_OUTPUT_FORMAT_YCBCR444] = "YCBCR4:4:4",
@@ -12229,7 +10219,7 @@ static const char * const output_format_str[] = {
 static const char *output_formats(enum intel_output_format format)
 {
 	if (format >= ARRAY_SIZE(output_format_str))
-		format = INTEL_OUTPUT_FORMAT_INVALID;
+		return "invalid";
 	return output_format_str[format];
 }
 
@@ -12345,6 +10335,13 @@ static void intel_dump_pipe_config(const struct intel_crtc_state *pipe_config,
 	    intel_hdmi_infoframe_enable(DP_SDP_VSC))
 		intel_dump_dp_vsc_sdp(dev_priv, &pipe_config->infoframes.vsc);
 
+	drm_dbg_kms(&dev_priv->drm, "vrr: %s, vmin: %d, vmax: %d, pipeline full: %d, flipline: %d, vmin vblank: %d, vmax vblank: %d\n",
+		    yesno(pipe_config->vrr.enable),
+		    pipe_config->vrr.vmin, pipe_config->vrr.vmax,
+		    pipe_config->vrr.pipeline_full, pipe_config->vrr.flipline,
+		    intel_vrr_vmin_vblank_start(pipe_config),
+		    intel_vrr_vmax_vblank_start(pipe_config));
+
 	drm_dbg_kms(&dev_priv->drm, "requested mode:\n");
 	drm_mode_debug_printmodeline(&pipe_config->hw.mode);
 	drm_dbg_kms(&dev_priv->drm, "adjusted mode:\n");
@@ -12732,7 +10729,7 @@ encoder_retry:
 		return ret;
 	}
 
-	if (ret == RETRY) {
+	if (ret == I915_DISPLAY_CONFIG_RETRY) {
 		if (drm_WARN(&i915->drm, !retry,
 			     "loop in pipe configuration computation\n"))
 			return -EINVAL;
@@ -13338,6 +11335,12 @@ intel_pipe_config_compare(const struct intel_crtc_state *current_config,
 
 	PIPE_CONF_CHECK_I(mst_master_transcoder);
 
+	PIPE_CONF_CHECK_BOOL(vrr.enable);
+	PIPE_CONF_CHECK_I(vrr.vmin);
+	PIPE_CONF_CHECK_I(vrr.vmax);
+	PIPE_CONF_CHECK_I(vrr.flipline);
+	PIPE_CONF_CHECK_I(vrr.pipeline_full);
+
 #undef PIPE_CONF_CHECK_X
 #undef PIPE_CONF_CHECK_I
 #undef PIPE_CONF_CHECK_BOOL
@@ -13796,10 +11799,17 @@ intel_crtc_update_active_timings(const struct intel_crtc_state *crtc_state)
 {
 	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
 	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	const struct drm_display_mode *adjusted_mode =
-		&crtc_state->hw.adjusted_mode;
+	struct drm_display_mode adjusted_mode =
+		crtc_state->hw.adjusted_mode;
 
-	drm_calc_timestamping_constants(&crtc->base, adjusted_mode);
+	if (crtc_state->vrr.enable) {
+		adjusted_mode.crtc_vtotal = crtc_state->vrr.vmax;
+		adjusted_mode.crtc_vblank_end = crtc_state->vrr.vmax;
+		adjusted_mode.crtc_vblank_start = intel_vrr_vmin_vblank_start(crtc_state);
+		crtc->vmax_vblank_start = intel_vrr_vmax_vblank_start(crtc_state);
+	}
+
+	drm_calc_timestamping_constants(&crtc->base, &adjusted_mode);
 
 	crtc->mode_flags = crtc_state->mode_flags;
 
@@ -13833,8 +11843,8 @@ intel_crtc_update_active_timings(const struct intel_crtc_state *crtc_state)
 	if (IS_GEN(dev_priv, 2)) {
 		int vtotal;
 
-		vtotal = adjusted_mode->crtc_vtotal;
-		if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+		vtotal = adjusted_mode.crtc_vtotal;
+		if (adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 			vtotal /= 2;
 
 		crtc->scanline_offset = vtotal - 1;
@@ -14307,7 +12317,7 @@ static void kill_bigjoiner_slave(struct intel_atomic_state *state,
  * Async flip can only change the plane surface address, so anything else
  * changing is rejected from the intel_atomic_check_async() function.
  * Once this check is cleared, flip done interrupt is enabled using
- * the skl_enable_flip_done() function.
+ * the intel_crtc_enable_flip_done() function.
  *
  * As soon as the surface address register is written, flip done interrupt is
  * generated and the requested events are sent to the usersapce in the interrupt
@@ -14351,7 +12361,7 @@ static int intel_atomic_check_async(struct intel_atomic_state *state)
 		 * this(vlv/chv and icl+) should be added when async flip is
 		 * enabled in the atomic IOCTL path.
 		 */
-		if (plane->id != PLANE_PRIMARY)
+		if (!plane->async_flip)
 			return -EINVAL;
 
 		/*
@@ -14495,6 +12505,8 @@ static int intel_atomic_check(struct drm_device *dev,
 			new_crtc_state->uapi.mode_changed = true;
 	}
 
+	intel_vrr_check_modeset(state);
+
 	ret = drm_atomic_helper_check_modeset(dev, &state->base);
 	if (ret)
 		goto fail;
@@ -14621,20 +12633,6 @@ static int intel_atomic_check(struct drm_device *dev,
 	if (ret)
 		goto fail;
 
-	/*
-	 * distrust_bios_wm will force a full dbuf recomputation
-	 * but the hardware state will only get updated accordingly
-	 * if state->modeset==true. Hence distrust_bios_wm==true &&
-	 * state->modeset==false is an invalid combination which
-	 * would cause the hardware and software dbuf state to get
-	 * out of sync. We must prevent that.
-	 *
-	 * FIXME clean up this mess and introduce better
-	 * state tracking for dbuf.
-	 */
-	if (dev_priv->wm.distrust_bios_wm)
-		any_ms = true;
-
 	intel_fbc_choose_crtc(dev_priv, state);
 	ret = calc_watermark_data(state);
 	if (ret)
@@ -14720,17 +12718,6 @@ static int intel_atomic_prepare_commit(struct intel_atomic_state *state)
 	return 0;
 }
 
-u32 intel_crtc_get_vblank_counter(struct intel_crtc *crtc)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_vblank_crtc *vblank = &dev->vblank[drm_crtc_index(&crtc->base)];
-
-	if (!vblank->max_vblank_count)
-		return (u32)drm_crtc_accurate_vblank_count(&crtc->base);
-
-	return crtc->base.funcs->get_vblank_counter(&crtc->base);
-}
-
 void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc,
 				  struct intel_crtc_state *crtc_state)
 {
@@ -15198,6 +13185,43 @@ static void intel_atomic_cleanup_work(struct work_struct *work)
 	intel_atomic_helper_free_state(i915);
 }
 
+static void intel_atomic_prepare_plane_clear_colors(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_plane *plane;
+	struct intel_plane_state *plane_state;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		struct drm_framebuffer *fb = plane_state->hw.fb;
+		int ret;
+
+		if (!fb ||
+		    fb->modifier != I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC)
+			continue;
+
+		/*
+		 * The layout of the fast clear color value expected by HW
+		 * (the DRM ABI requiring this value to be located in fb at offset 0 of plane#2):
+		 * - 4 x 4 bytes per-channel value
+		 *   (in surface type specific float/int format provided by the fb user)
+		 * - 8 bytes native color value used by the display
+		 *   (converted/written by GPU during a fast clear operation using the
+		 *    above per-channel values)
+		 *
+		 * The commit's FB prepare hook already ensured that FB obj is pinned and the
+		 * caller made sure that the object is synced wrt. the related color clear value
+		 * GPU write on it.
+		 */
+		ret = i915_gem_object_read_from_page(intel_fb_obj(fb),
+						     fb->offsets[2] + 16,
+						     &plane_state->ccval,
+						     sizeof(plane_state->ccval));
+		/* The above could only fail if the FB obj has an unexpected backing store type. */
+		drm_WARN_ON(&i915->drm, ret);
+	}
+}
+
 static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 {
 	struct drm_device *dev = state->base.dev;
@@ -15215,6 +13239,8 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 	if (state->modeset)
 		wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_MODESET);
 
+	intel_atomic_prepare_plane_clear_colors(state);
+
 	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
 					    new_crtc_state, i) {
 		if (intel_crtc_needs_modeset(new_crtc_state) ||
@@ -15263,7 +13289,7 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 
 	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
 		if (new_crtc_state->uapi.async_flip)
-			skl_enable_flip_done(crtc);
+			intel_crtc_enable_flip_done(state, crtc);
 	}
 
 	/* Now enable the clocks, plane, pipe, and connectors that we set up. */
@@ -15288,7 +13314,7 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 
 	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
 		if (new_crtc_state->uapi.async_flip)
-			skl_disable_flip_done(crtc);
+			intel_crtc_disable_flip_done(state, crtc);
 
 		if (new_crtc_state->hw.active &&
 		    !intel_crtc_needs_modeset(new_crtc_state) &&
@@ -15491,7 +13517,6 @@ static int intel_atomic_commit(struct drm_device *dev,
 		intel_runtime_pm_put(&dev_priv->runtime_pm, state->wakeref);
 		return ret;
 	}
-	dev_priv->wm.distrust_bios_wm = false;
 	intel_shared_dpll_swap_state(state);
 	intel_atomic_track_fbs(state);
 
@@ -15609,15 +13634,6 @@ void intel_plane_unpin_fb(struct intel_plane_state *old_plane_state)
 		intel_unpin_fb_vma(vma, old_plane_state->flags);
 }
 
-static void fb_obj_bump_render_priority(struct drm_i915_gem_object *obj)
-{
-	struct i915_sched_attr attr = {
-		.priority = I915_USER_PRIORITY(I915_PRIORITY_DISPLAY),
-	};
-
-	i915_gem_object_wait_priority(obj, 0, &attr);
-}
-
 /**
  * intel_prepare_plane_fb - Prepare fb for usage on plane
  * @_plane: drm plane to prepare for
@@ -15634,6 +13650,9 @@ int
 intel_prepare_plane_fb(struct drm_plane *_plane,
 		       struct drm_plane_state *_new_plane_state)
 {
+	struct i915_sched_attr attr = {
+		.priority = I915_USER_PRIORITY(I915_PRIORITY_DISPLAY),
+	};
 	struct intel_plane *plane = to_intel_plane(_plane);
 	struct intel_plane_state *new_plane_state =
 		to_intel_plane_state(_new_plane_state);
@@ -15673,6 +13692,8 @@ intel_prepare_plane_fb(struct drm_plane *_plane,
 	}
 
 	if (new_plane_state->uapi.fence) { /* explicit fencing */
+		i915_gem_fence_wait_priority(new_plane_state->uapi.fence,
+					     &attr);
 		ret = i915_sw_fence_await_dma_fence(&state->commit_ready,
 						    new_plane_state->uapi.fence,
 						    i915_fence_timeout(dev_priv),
@@ -15694,7 +13715,7 @@ intel_prepare_plane_fb(struct drm_plane *_plane,
 	if (ret)
 		return ret;
 
-	fb_obj_bump_render_priority(obj);
+	i915_gem_object_wait_priority(obj, 0, &attr);
 	i915_gem_object_flush_frontbuffer(obj, ORIGIN_DIRTYFB);
 
 	if (!new_plane_state->uapi.fence) { /* implicit fencing */
@@ -15783,113 +13804,6 @@ void intel_plane_destroy(struct drm_plane *plane)
 	kfree(to_intel_plane(plane));
 }
 
-static int intel_crtc_late_register(struct drm_crtc *crtc)
-{
-	intel_crtc_debugfs_add(crtc);
-	return 0;
-}
-
-#define INTEL_CRTC_FUNCS \
-	.set_config = drm_atomic_helper_set_config, \
-	.destroy = intel_crtc_destroy, \
-	.page_flip = drm_atomic_helper_page_flip, \
-	.atomic_duplicate_state = intel_crtc_duplicate_state, \
-	.atomic_destroy_state = intel_crtc_destroy_state, \
-	.set_crc_source = intel_crtc_set_crc_source, \
-	.verify_crc_source = intel_crtc_verify_crc_source, \
-	.get_crc_sources = intel_crtc_get_crc_sources, \
-	.late_register = intel_crtc_late_register
-
-static const struct drm_crtc_funcs bdw_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	.get_vblank_counter = g4x_get_vblank_counter,
-	.enable_vblank = bdw_enable_vblank,
-	.disable_vblank = bdw_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static const struct drm_crtc_funcs ilk_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	.get_vblank_counter = g4x_get_vblank_counter,
-	.enable_vblank = ilk_enable_vblank,
-	.disable_vblank = ilk_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static const struct drm_crtc_funcs g4x_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	.get_vblank_counter = g4x_get_vblank_counter,
-	.enable_vblank = i965_enable_vblank,
-	.disable_vblank = i965_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static const struct drm_crtc_funcs i965_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	.get_vblank_counter = i915_get_vblank_counter,
-	.enable_vblank = i965_enable_vblank,
-	.disable_vblank = i965_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static const struct drm_crtc_funcs i915gm_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	.get_vblank_counter = i915_get_vblank_counter,
-	.enable_vblank = i915gm_enable_vblank,
-	.disable_vblank = i915gm_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static const struct drm_crtc_funcs i915_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	.get_vblank_counter = i915_get_vblank_counter,
-	.enable_vblank = i8xx_enable_vblank,
-	.disable_vblank = i8xx_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static const struct drm_crtc_funcs i8xx_crtc_funcs = {
-	INTEL_CRTC_FUNCS,
-
-	/* no hw vblank counter */
-	.enable_vblank = i8xx_enable_vblank,
-	.disable_vblank = i8xx_disable_vblank,
-	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
-};
-
-static struct intel_crtc *intel_crtc_alloc(void)
-{
-	struct intel_crtc_state *crtc_state;
-	struct intel_crtc *crtc;
-
-	crtc = kzalloc(sizeof(*crtc), GFP_KERNEL);
-	if (!crtc)
-		return ERR_PTR(-ENOMEM);
-
-	crtc_state = intel_crtc_state_alloc(crtc);
-	if (!crtc_state) {
-		kfree(crtc);
-		return ERR_PTR(-ENOMEM);
-	}
-
-	crtc->base.state = &crtc_state->uapi;
-	crtc->config = crtc_state;
-
-	return crtc;
-}
-
-static void intel_crtc_free(struct intel_crtc *crtc)
-{
-	intel_crtc_destroy_state(&crtc->base, crtc->base.state);
-	kfree(crtc);
-}
-
 static void intel_plane_possible_crtcs_init(struct drm_i915_private *dev_priv)
 {
 	struct intel_plane *plane;
@@ -15902,100 +13816,6 @@ static void intel_plane_possible_crtcs_init(struct drm_i915_private *dev_priv)
 	}
 }
 
-static int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe)
-{
-	struct intel_plane *primary, *cursor;
-	const struct drm_crtc_funcs *funcs;
-	struct intel_crtc *crtc;
-	int sprite, ret;
-
-	crtc = intel_crtc_alloc();
-	if (IS_ERR(crtc))
-		return PTR_ERR(crtc);
-
-	crtc->pipe = pipe;
-	crtc->num_scalers = RUNTIME_INFO(dev_priv)->num_scalers[pipe];
-
-	primary = intel_primary_plane_create(dev_priv, pipe);
-	if (IS_ERR(primary)) {
-		ret = PTR_ERR(primary);
-		goto fail;
-	}
-	crtc->plane_ids_mask |= BIT(primary->id);
-
-	for_each_sprite(dev_priv, pipe, sprite) {
-		struct intel_plane *plane;
-
-		plane = intel_sprite_plane_create(dev_priv, pipe, sprite);
-		if (IS_ERR(plane)) {
-			ret = PTR_ERR(plane);
-			goto fail;
-		}
-		crtc->plane_ids_mask |= BIT(plane->id);
-	}
-
-	cursor = intel_cursor_plane_create(dev_priv, pipe);
-	if (IS_ERR(cursor)) {
-		ret = PTR_ERR(cursor);
-		goto fail;
-	}
-	crtc->plane_ids_mask |= BIT(cursor->id);
-
-	if (HAS_GMCH(dev_priv)) {
-		if (IS_CHERRYVIEW(dev_priv) ||
-		    IS_VALLEYVIEW(dev_priv) || IS_G4X(dev_priv))
-			funcs = &g4x_crtc_funcs;
-		else if (IS_GEN(dev_priv, 4))
-			funcs = &i965_crtc_funcs;
-		else if (IS_I945GM(dev_priv) || IS_I915GM(dev_priv))
-			funcs = &i915gm_crtc_funcs;
-		else if (IS_GEN(dev_priv, 3))
-			funcs = &i915_crtc_funcs;
-		else
-			funcs = &i8xx_crtc_funcs;
-	} else {
-		if (INTEL_GEN(dev_priv) >= 8)
-			funcs = &bdw_crtc_funcs;
-		else
-			funcs = &ilk_crtc_funcs;
-	}
-
-	ret = drm_crtc_init_with_planes(&dev_priv->drm, &crtc->base,
-					&primary->base, &cursor->base,
-					funcs, "pipe %c", pipe_name(pipe));
-	if (ret)
-		goto fail;
-
-	BUG_ON(pipe >= ARRAY_SIZE(dev_priv->pipe_to_crtc_mapping) ||
-	       dev_priv->pipe_to_crtc_mapping[pipe] != NULL);
-	dev_priv->pipe_to_crtc_mapping[pipe] = crtc;
-
-	if (INTEL_GEN(dev_priv) < 9) {
-		enum i9xx_plane_id i9xx_plane = primary->i9xx_plane;
-
-		BUG_ON(i9xx_plane >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) ||
-		       dev_priv->plane_to_crtc_mapping[i9xx_plane] != NULL);
-		dev_priv->plane_to_crtc_mapping[i9xx_plane] = crtc;
-	}
-
-	if (INTEL_GEN(dev_priv) >= 10)
-		drm_crtc_create_scaling_filter_property(&crtc->base,
-						BIT(DRM_SCALING_FILTER_DEFAULT) |
-						BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR));
-
-	intel_color_init(crtc);
-
-	intel_crtc_crc_init(crtc);
-
-	drm_WARN_ON(&dev_priv->drm, drm_crtc_index(&crtc->base) != crtc->pipe);
-
-	return 0;
-
-fail:
-	intel_crtc_free(crtc);
-
-	return ret;
-}
 
 int intel_get_pipe_from_crtc_id_ioctl(struct drm_device *dev, void *data,
 				      struct drm_file *file)
@@ -16078,48 +13898,12 @@ static bool intel_ddi_crt_present(struct drm_i915_private *dev_priv)
 	return true;
 }
 
-void intel_pps_unlock_regs_wa(struct drm_i915_private *dev_priv)
-{
-	int pps_num;
-	int pps_idx;
-
-	if (HAS_DDI(dev_priv))
-		return;
-	/*
-	 * This w/a is needed at least on CPT/PPT, but to be sure apply it
-	 * everywhere where registers can be write protected.
-	 */
-	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		pps_num = 2;
-	else
-		pps_num = 1;
-
-	for (pps_idx = 0; pps_idx < pps_num; pps_idx++) {
-		u32 val = intel_de_read(dev_priv, PP_CONTROL(pps_idx));
-
-		val = (val & ~PANEL_UNLOCK_MASK) | PANEL_UNLOCK_REGS;
-		intel_de_write(dev_priv, PP_CONTROL(pps_idx), val);
-	}
-}
-
-static void intel_pps_init(struct drm_i915_private *dev_priv)
-{
-	if (HAS_PCH_SPLIT(dev_priv) || IS_GEN9_LP(dev_priv))
-		dev_priv->pps_mmio_base = PCH_PPS_BASE;
-	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		dev_priv->pps_mmio_base = VLV_PPS_BASE;
-	else
-		dev_priv->pps_mmio_base = PPS_BASE;
-
-	intel_pps_unlock_regs_wa(dev_priv);
-}
-
 static void intel_setup_outputs(struct drm_i915_private *dev_priv)
 {
 	struct intel_encoder *encoder;
 	bool dpd_is_edp = false;
 
-	intel_pps_init(dev_priv);
+	intel_pps_unlock_regs_wa(dev_priv);
 
 	if (!HAS_DISPLAY(dev_priv))
 		return;
@@ -16516,7 +14300,7 @@ static int intel_framebuffer_init(struct intel_framebuffer *intel_fb,
 			goto err;
 		}
 
-		if (is_gen12_ccs_plane(fb, i)) {
+		if (is_gen12_ccs_plane(fb, i) && !is_gen12_ccs_cc_plane(fb, i)) {
 			int ccs_aux_stride = gen12_ccs_aux_stride(fb, i);
 
 			if (fb->pitches[i] != ccs_aux_stride) {
@@ -16714,86 +14498,40 @@ void intel_init_display_hooks(struct drm_i915_private *dev_priv)
 {
 	intel_init_cdclk_hooks(dev_priv);
 
+	intel_dpll_init_clock_hook(dev_priv);
+
 	if (INTEL_GEN(dev_priv) >= 9) {
 		dev_priv->display.get_pipe_config = hsw_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			skl_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = hsw_crtc_compute_clock;
 		dev_priv->display.crtc_enable = hsw_crtc_enable;
 		dev_priv->display.crtc_disable = hsw_crtc_disable;
 	} else if (HAS_DDI(dev_priv)) {
 		dev_priv->display.get_pipe_config = hsw_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock =
-			hsw_crtc_compute_clock;
 		dev_priv->display.crtc_enable = hsw_crtc_enable;
 		dev_priv->display.crtc_disable = hsw_crtc_disable;
 	} else if (HAS_PCH_SPLIT(dev_priv)) {
 		dev_priv->display.get_pipe_config = ilk_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock =
-			ilk_crtc_compute_clock;
 		dev_priv->display.crtc_enable = ilk_crtc_enable;
 		dev_priv->display.crtc_disable = ilk_crtc_disable;
-	} else if (IS_CHERRYVIEW(dev_priv)) {
-		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = chv_crtc_compute_clock;
-		dev_priv->display.crtc_enable = valleyview_crtc_enable;
-		dev_priv->display.crtc_disable = i9xx_crtc_disable;
-	} else if (IS_VALLEYVIEW(dev_priv)) {
+	} else if (IS_CHERRYVIEW(dev_priv) ||
+		   IS_VALLEYVIEW(dev_priv)) {
 		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = vlv_crtc_compute_clock;
 		dev_priv->display.crtc_enable = valleyview_crtc_enable;
 		dev_priv->display.crtc_disable = i9xx_crtc_disable;
-	} else if (IS_G4X(dev_priv)) {
-		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = g4x_crtc_compute_clock;
-		dev_priv->display.crtc_enable = i9xx_crtc_enable;
-		dev_priv->display.crtc_disable = i9xx_crtc_disable;
-	} else if (IS_PINEVIEW(dev_priv)) {
-		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = pnv_crtc_compute_clock;
-		dev_priv->display.crtc_enable = i9xx_crtc_enable;
-		dev_priv->display.crtc_disable = i9xx_crtc_disable;
-	} else if (!IS_GEN(dev_priv, 2)) {
-		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock;
-		dev_priv->display.crtc_enable = i9xx_crtc_enable;
-		dev_priv->display.crtc_disable = i9xx_crtc_disable;
 	} else {
 		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
-		dev_priv->display.get_initial_plane_config =
-			i9xx_get_initial_plane_config;
-		dev_priv->display.crtc_compute_clock = i8xx_crtc_compute_clock;
 		dev_priv->display.crtc_enable = i9xx_crtc_enable;
 		dev_priv->display.crtc_disable = i9xx_crtc_disable;
 	}
 
-	if (IS_GEN(dev_priv, 5)) {
-		dev_priv->display.fdi_link_train = ilk_fdi_link_train;
-	} else if (IS_GEN(dev_priv, 6)) {
-		dev_priv->display.fdi_link_train = gen6_fdi_link_train;
-	} else if (IS_IVYBRIDGE(dev_priv)) {
-		/* FIXME: detect B0+ stepping and use auto training */
-		dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
-	}
+	intel_fdi_init_hook(dev_priv);
 
-	if (INTEL_GEN(dev_priv) >= 9)
+	if (INTEL_GEN(dev_priv) >= 9) {
 		dev_priv->display.commit_modeset_enables = skl_commit_modeset_enables;
-	else
+		dev_priv->display.get_initial_plane_config = skl_get_initial_plane_config;
+	} else {
 		dev_priv->display.commit_modeset_enables = intel_commit_modeset_enables;
+		dev_priv->display.get_initial_plane_config = i9xx_get_initial_plane_config;
+	}
 
 }
 
@@ -16801,14 +14539,10 @@ void intel_modeset_init_hw(struct drm_i915_private *i915)
 {
 	struct intel_cdclk_state *cdclk_state =
 		to_intel_cdclk_state(i915->cdclk.obj.state);
-	struct intel_dbuf_state *dbuf_state =
-		to_intel_dbuf_state(i915->dbuf.obj.state);
 
 	intel_update_cdclk(i915);
 	intel_dump_cdclk_config(&i915->cdclk.hw, "Current CDCLK");
 	cdclk_state->logical = cdclk_state->actual = i915->cdclk.hw;
-
-	dbuf_state->enabled_slices = i915->dbuf.enabled_slices;
 }
 
 static int sanitize_watermarks_add_affected(struct drm_atomic_state *state)
@@ -17041,8 +14775,7 @@ static void intel_mode_config_init(struct drm_i915_private *i915)
 
 	mode_config->funcs = &intel_mode_funcs;
 
-	if (INTEL_GEN(i915) >= 9)
-		mode_config->async_page_flip = true;
+	mode_config->async_page_flip = has_async_flips(i915);
 
 	/*
 	 * Maximum framebuffer dimensions, chosen to match
@@ -17127,6 +14860,8 @@ int intel_modeset_init_noirq(struct drm_i915_private *i915)
 	i915->flip_wq = alloc_workqueue("i915_flip", WQ_HIGHPRI |
 					WQ_UNBOUND, WQ_UNBOUND_MAX_ACTIVE);
 
+	i915->framestart_delay = 1; /* 1-4 */
+
 	intel_mode_config_init(i915);
 
 	ret = intel_cdclk_init(i915);
@@ -17173,6 +14908,8 @@ int intel_modeset_init_nogem(struct drm_i915_private *i915)
 
 	intel_panel_sanitize_ssc(i915);
 
+	intel_pps_setup(i915);
+
 	intel_gmbus_setup(i915);
 
 	drm_dbg_kms(&i915->drm, "%d display pipe%s available.\n",
@@ -17461,7 +15198,7 @@ static void intel_sanitize_frame_start_delay(const struct intel_crtc_state *crtc
 
 		val = intel_de_read(dev_priv, reg);
 		val &= ~HSW_FRAME_START_DELAY_MASK;
-		val |= HSW_FRAME_START_DELAY(0);
+		val |= HSW_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 		intel_de_write(dev_priv, reg, val);
 	} else {
 		i915_reg_t reg = PIPECONF(cpu_transcoder);
@@ -17469,7 +15206,7 @@ static void intel_sanitize_frame_start_delay(const struct intel_crtc_state *crtc
 
 		val = intel_de_read(dev_priv, reg);
 		val &= ~PIPECONF_FRAME_START_DELAY_MASK;
-		val |= PIPECONF_FRAME_START_DELAY(0);
+		val |= PIPECONF_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 		intel_de_write(dev_priv, reg, val);
 	}
 
@@ -17482,7 +15219,7 @@ static void intel_sanitize_frame_start_delay(const struct intel_crtc_state *crtc
 
 		val = intel_de_read(dev_priv, reg);
 		val &= ~TRANS_FRAME_START_DELAY_MASK;
-		val |= TRANS_FRAME_START_DELAY(0);
+		val |= TRANS_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 		intel_de_write(dev_priv, reg, val);
 	} else {
 		enum pipe pch_transcoder = intel_crtc_pch_transcoder(crtc);
@@ -17491,7 +15228,7 @@ static void intel_sanitize_frame_start_delay(const struct intel_crtc_state *crtc
 
 		val = intel_de_read(dev_priv, reg);
 		val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
-		val |= TRANS_CHICKEN2_FRAME_START_DELAY(0);
+		val |= TRANS_CHICKEN2_FRAME_START_DELAY(dev_priv->framestart_delay - 1);
 		intel_de_write(dev_priv, reg, val);
 	}
 }
diff --git a/drivers/gpu/drm/i915/display/intel_display.h b/drivers/gpu/drm/i915/display/intel_display.h
index f7e49a39bd92..76f8a805b0a3 100644
--- a/drivers/gpu/drm/i915/display/intel_display.h
+++ b/drivers/gpu/drm/i915/display/intel_display.h
@@ -546,7 +546,6 @@ unsigned int intel_rotation_info_size(const struct intel_rotation_info *rot_info
 unsigned int intel_remapped_info_size(const struct intel_remapped_info *rem_info);
 bool intel_has_pending_fb_unpin(struct drm_i915_private *dev_priv);
 int intel_display_suspend(struct drm_device *dev);
-void intel_pps_unlock_regs_wa(struct drm_i915_private *dev_priv);
 void intel_encoder_destroy(struct drm_encoder *encoder);
 struct drm_display_mode *
 intel_encoder_current_mode(struct intel_encoder *encoder);
@@ -630,6 +629,8 @@ u32 skl_plane_ctl_crtc(const struct intel_crtc_state *crtc_state);
 u32 skl_plane_stride(const struct intel_plane_state *plane_state,
 		     int plane);
 int skl_check_plane_surface(struct intel_plane_state *plane_state);
+int skl_calc_main_surface_offset(const struct intel_plane_state *plane_state,
+				 int *x, int *y, u32 *offset);
 int skl_format_to_fourcc(int format, bool rgb_order, bool alpha);
 int bdw_get_pipemisc_bpp(struct intel_crtc *crtc);
 unsigned int intel_plane_fence_y_offset(const struct intel_plane_state *plane_state);
@@ -641,7 +642,7 @@ void intel_display_print_error_state(struct drm_i915_error_state_buf *e,
 
 bool
 intel_format_info_is_yuv_semiplanar(const struct drm_format_info *info,
-				    uint64_t modifier);
+				    u64 modifier);
 
 int intel_plane_compute_gtt(struct intel_plane_state *plane_state);
 u32 intel_plane_compute_aligned_offset(int *x, int *y,
@@ -649,6 +650,15 @@ u32 intel_plane_compute_aligned_offset(int *x, int *y,
 				       int color_plane);
 int intel_plane_pin_fb(struct intel_plane_state *plane_state);
 void intel_plane_unpin_fb(struct intel_plane_state *old_plane_state);
+struct intel_encoder *
+intel_get_crtc_new_encoder(const struct intel_atomic_state *state,
+			   const struct intel_crtc_state *crtc_state);
+unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
+				  int color_plane);
+u32 intel_plane_adjust_aligned_offset(int *x, int *y,
+				      const struct intel_plane_state *state,
+				      int color_plane,
+				      u32 old_offset, u32 new_offset);
 
 /* modesetting */
 void intel_modeset_init_hw(struct drm_i915_private *i915);
diff --git a/drivers/gpu/drm/i915/display/intel_display_debugfs.c b/drivers/gpu/drm/i915/display/intel_display_debugfs.c
index b02610fa8c5d..0e364dc3e7c3 100644
--- a/drivers/gpu/drm/i915/display/intel_display_debugfs.c
+++ b/drivers/gpu/drm/i915/display/intel_display_debugfs.c
@@ -1135,7 +1135,6 @@ static ssize_t i915_ipc_status_write(struct file *file, const char __user *ubuf,
 		if (!dev_priv->ipc_enabled && enable)
 			drm_info(&dev_priv->drm,
 				 "Enabling IPC: WM will be proper only after next commit\n");
-		dev_priv->wm.distrust_bios_wm = true;
 		dev_priv->ipc_enabled = enable;
 		intel_enable_ipc(dev_priv);
 	}
@@ -2151,13 +2150,13 @@ static int i915_panel_show(struct seq_file *m, void *data)
 		return -ENODEV;
 
 	seq_printf(m, "Panel power up delay: %d\n",
-		   intel_dp->panel_power_up_delay);
+		   intel_dp->pps.panel_power_up_delay);
 	seq_printf(m, "Panel power down delay: %d\n",
-		   intel_dp->panel_power_down_delay);
+		   intel_dp->pps.panel_power_down_delay);
 	seq_printf(m, "Backlight on delay: %d\n",
-		   intel_dp->backlight_on_delay);
+		   intel_dp->pps.backlight_on_delay);
 	seq_printf(m, "Backlight off delay: %d\n",
-		   intel_dp->backlight_off_delay);
+		   intel_dp->pps.backlight_off_delay);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c
index d52374f01316..c11c37c65d86 100644
--- a/drivers/gpu/drm/i915/display/intel_display_power.c
+++ b/drivers/gpu/drm/i915/display/intel_display_power.c
@@ -4,7 +4,6 @@
  */
 
 #include "display/intel_crt.h"
-#include "display/intel_dp.h"
 
 #include "i915_drv.h"
 #include "i915_irq.h"
@@ -16,6 +15,7 @@
 #include "intel_dpio_phy.h"
 #include "intel_hotplug.h"
 #include "intel_pm.h"
+#include "intel_pps.h"
 #include "intel_sideband.h"
 #include "intel_tc.h"
 #include "intel_vga.h"
@@ -936,7 +936,7 @@ static void bxt_enable_dc9(struct drm_i915_private *dev_priv)
 	 * because PPS registers are always on.
 	 */
 	if (!HAS_PCH_SPLIT(dev_priv))
-		intel_power_sequencer_reset(dev_priv);
+		intel_pps_reset_all(dev_priv);
 	gen9_set_dc_state(dev_priv, DC_STATE_EN_DC9);
 }
 
@@ -1446,7 +1446,7 @@ static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv)
 	/* make sure we're done processing display irqs */
 	intel_synchronize_irq(dev_priv);
 
-	intel_power_sequencer_reset(dev_priv);
+	intel_pps_reset_all(dev_priv);
 
 	/* Prevent us from re-enabling polling on accident in late suspend */
 	if (!dev_priv->drm.dev->power.is_suspended)
diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
index 1067bd073c95..184ecbbcec99 100644
--- a/drivers/gpu/drm/i915/display/intel_display_types.h
+++ b/drivers/gpu/drm/i915/display/intel_display_types.h
@@ -228,7 +228,7 @@ struct intel_encoder {
 struct intel_panel_bl_funcs {
 	/* Connector and platform specific backlight functions */
 	int (*setup)(struct intel_connector *connector, enum pipe pipe);
-	u32 (*get)(struct intel_connector *connector);
+	u32 (*get)(struct intel_connector *connector, enum pipe pipe);
 	void (*set)(const struct drm_connector_state *conn_state, u32 level);
 	void (*disable)(const struct drm_connector_state *conn_state, u32 level);
 	void (*enable)(const struct intel_crtc_state *crtc_state,
@@ -252,17 +252,28 @@ struct intel_panel {
 		bool alternate_pwm_increment;	/* lpt+ */
 
 		/* PWM chip */
+		u32 pwm_level_min;
+		u32 pwm_level_max;
+		bool pwm_enabled;
 		bool util_pin_active_low;	/* bxt+ */
 		u8 controller;		/* bxt+ only */
 		struct pwm_device *pwm;
 		struct pwm_state pwm_state;
 
 		/* DPCD backlight */
-		u8 pwmgen_bit_count;
+		union {
+			struct {
+				u8 pwmgen_bit_count;
+			} vesa;
+			struct {
+				bool sdr_uses_aux;
+			} intel;
+		} edp;
 
 		struct backlight_device *device;
 
 		const struct intel_panel_bl_funcs *funcs;
+		const struct intel_panel_bl_funcs *pwm_funcs;
 		void (*power)(struct intel_connector *, bool enable);
 	} backlight;
 };
@@ -343,6 +354,10 @@ struct intel_hdcp_shim {
 				 enum transcoder cpu_transcoder,
 				 bool enable);
 
+	/* Enable/Disable stream encryption on DP MST Transport Link */
+	int (*stream_encryption)(struct intel_connector *connector,
+				 bool enable);
+
 	/* Ensures the link is still protected */
 	bool (*check_link)(struct intel_digital_port *dig_port,
 			   struct intel_connector *connector);
@@ -374,8 +389,13 @@ struct intel_hdcp_shim {
 	int (*config_stream_type)(struct intel_digital_port *dig_port,
 				  bool is_repeater, u8 type);
 
+	/* Enable/Disable HDCP 2.2 stream encryption on DP MST Transport Link */
+	int (*stream_2_2_encryption)(struct intel_connector *connector,
+				     bool enable);
+
 	/* HDCP2.2 Link Integrity Check */
-	int (*check_2_2_link)(struct intel_digital_port *dig_port);
+	int (*check_2_2_link)(struct intel_digital_port *dig_port,
+			      struct intel_connector *connector);
 };
 
 struct intel_hdcp {
@@ -402,7 +422,6 @@ struct intel_hdcp {
 	 * content can flow only through a link protected by HDCP2.2.
 	 */
 	u8 content_type;
-	struct hdcp_port_data port_data;
 
 	bool is_paired;
 	bool is_repeater;
@@ -436,6 +455,8 @@ struct intel_hdcp {
 	 * Hence caching the transcoder here.
 	 */
 	enum transcoder cpu_transcoder;
+	/* Only used for DP MST stream encryption */
+	enum transcoder stream_transcoder;
 };
 
 struct intel_connector {
@@ -535,7 +556,7 @@ struct intel_plane_state {
 		struct drm_framebuffer *fb;
 
 		u16 alpha;
-		uint16_t pixel_blend_mode;
+		u16 pixel_blend_mode;
 		unsigned int rotation;
 		enum drm_color_encoding color_encoding;
 		enum drm_color_range color_range;
@@ -610,6 +631,9 @@ struct intel_plane_state {
 	struct drm_intel_sprite_colorkey ckey;
 
 	struct drm_rect psr2_sel_fetch_area;
+
+	/* Clear Color Value */
+	u64 ccval;
 };
 
 struct intel_initial_plane_config {
@@ -669,6 +693,8 @@ struct intel_crtc_scaler_state {
 #define I915_MODE_FLAG_DSI_USE_TE1 (1<<4)
 /* Flag to indicate mipi dsi periodic command mode where we do not get TE */
 #define I915_MODE_FLAG_DSI_PERIODIC_CMD_MODE (1<<5)
+/* Do tricks to make vblank timestamps sane with VRR? */
+#define I915_MODE_FLAG_VRR (1<<6)
 
 struct intel_wm_level {
 	bool enable;
@@ -804,7 +830,6 @@ struct intel_crtc_wm_state {
 };
 
 enum intel_output_format {
-	INTEL_OUTPUT_FORMAT_INVALID,
 	INTEL_OUTPUT_FORMAT_RGB,
 	INTEL_OUTPUT_FORMAT_YCBCR420,
 	INTEL_OUTPUT_FORMAT_YCBCR444,
@@ -1127,6 +1152,13 @@ struct intel_crtc_state {
 	struct intel_dsb *dsb;
 
 	u32 psr2_man_track_ctl;
+
+	/* Variable Refresh Rate state */
+	struct {
+		bool enable;
+		u8 pipeline_full;
+		u16 flipline, vmin, vmax;
+	} vrr;
 };
 
 enum intel_pipe_crc_source {
@@ -1169,6 +1201,8 @@ struct intel_crtc {
 	/* I915_MODE_FLAG_* */
 	u8 mode_flags;
 
+	u16 vmax_vblank_start;
+
 	struct intel_display_power_domain_set enabled_power_domains;
 	struct intel_overlay *overlay;
 
@@ -1221,6 +1255,7 @@ struct intel_plane {
 	enum pipe pipe;
 	bool has_fbc;
 	bool has_ccs;
+	bool need_async_flip_disable_wa;
 	u32 frontbuffer_bit;
 
 	struct {
@@ -1257,7 +1292,10 @@ struct intel_plane {
 			 const struct intel_plane_state *plane_state);
 	void (*async_flip)(struct intel_plane *plane,
 			   const struct intel_crtc_state *crtc_state,
-			   const struct intel_plane_state *plane_state);
+			   const struct intel_plane_state *plane_state,
+			   bool async_flip);
+	void (*enable_flip_done)(struct intel_plane *plane);
+	void (*disable_flip_done)(struct intel_plane *plane);
 };
 
 struct intel_watermark_params {
@@ -1344,6 +1382,38 @@ struct intel_dp_pcon_frl {
 	int trained_rate_gbps;
 };
 
+struct intel_pps {
+	int panel_power_up_delay;
+	int panel_power_down_delay;
+	int panel_power_cycle_delay;
+	int backlight_on_delay;
+	int backlight_off_delay;
+	struct delayed_work panel_vdd_work;
+	bool want_panel_vdd;
+	unsigned long last_power_on;
+	unsigned long last_backlight_off;
+	ktime_t panel_power_off_time;
+	intel_wakeref_t vdd_wakeref;
+
+	/*
+	 * Pipe whose power sequencer is currently locked into
+	 * this port. Only relevant on VLV/CHV.
+	 */
+	enum pipe pps_pipe;
+	/*
+	 * Pipe currently driving the port. Used for preventing
+	 * the use of the PPS for any pipe currentrly driving
+	 * external DP as that will mess things up on VLV.
+	 */
+	enum pipe active_pipe;
+	/*
+	 * Set if the sequencer may be reset due to a power transition,
+	 * requiring a reinitialization. Only relevant on BXT.
+	 */
+	bool pps_reset;
+	struct edp_power_seq pps_delays;
+};
+
 struct intel_dp {
 	i915_reg_t output_reg;
 	u32 DP;
@@ -1354,6 +1424,7 @@ struct intel_dp {
 	bool has_hdmi_sink;
 	bool has_audio;
 	bool reset_link_params;
+	bool use_max_params;
 	u8 dpcd[DP_RECEIVER_CAP_SIZE];
 	u8 psr_dpcd[EDP_PSR_RECEIVER_CAP_SIZE];
 	u8 downstream_ports[DP_MAX_DOWNSTREAM_PORTS];
@@ -1379,39 +1450,11 @@ struct intel_dp {
 	int max_link_rate;
 	/* sink or branch descriptor */
 	struct drm_dp_desc desc;
-	u32 edid_quirks;
 	struct drm_dp_aux aux;
 	u32 aux_busy_last_status;
 	u8 train_set[4];
-	int panel_power_up_delay;
-	int panel_power_down_delay;
-	int panel_power_cycle_delay;
-	int backlight_on_delay;
-	int backlight_off_delay;
-	struct delayed_work panel_vdd_work;
-	bool want_panel_vdd;
-	unsigned long last_power_on;
-	unsigned long last_backlight_off;
-	ktime_t panel_power_off_time;
-	intel_wakeref_t vdd_wakeref;
 
-	/*
-	 * Pipe whose power sequencer is currently locked into
-	 * this port. Only relevant on VLV/CHV.
-	 */
-	enum pipe pps_pipe;
-	/*
-	 * Pipe currently driving the port. Used for preventing
-	 * the use of the PPS for any pipe currentrly driving
-	 * external DP as that will mess things up on VLV.
-	 */
-	enum pipe active_pipe;
-	/*
-	 * Set if the sequencer may be reset due to a power transition,
-	 * requiring a reinitialization. Only relevant on BXT.
-	 */
-	bool pps_reset;
-	struct edp_power_seq pps_delays;
+	struct intel_pps pps;
 
 	bool can_mst; /* this port supports mst */
 	bool is_mst;
@@ -1510,10 +1553,14 @@ struct intel_digital_port {
 	enum phy_fia tc_phy_fia;
 	u8 tc_phy_fia_idx;
 
-	/* protects num_hdcp_streams reference count */
+	/* protects num_hdcp_streams reference count, hdcp_port_data and hdcp_auth_status */
 	struct mutex hdcp_mutex;
 	/* the number of pipes using HDCP signalling out of this port */
 	unsigned int num_hdcp_streams;
+	/* port HDCP auth status */
+	bool hdcp_auth_status;
+	/* HDCP port data need to pass to security f/w */
+	struct hdcp_port_data hdcp_port_data;
 
 	void (*write_infoframe)(struct intel_encoder *encoder,
 				const struct intel_crtc_state *crtc_state,
@@ -1824,4 +1871,26 @@ to_intel_frontbuffer(struct drm_framebuffer *fb)
 	return fb ? to_intel_framebuffer(fb)->frontbuffer : NULL;
 }
 
+static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
+{
+	if (dev_priv->params.panel_use_ssc >= 0)
+		return dev_priv->params.panel_use_ssc != 0;
+	return dev_priv->vbt.lvds_use_ssc
+		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
+}
+
+static inline u32 i9xx_dpll_compute_fp(struct dpll *dpll)
+{
+	return dpll->n << 16 | dpll->m1 << 8 | dpll->m2;
+}
+
+static inline u32 intel_fdi_link_freq(struct drm_i915_private *dev_priv,
+				      const struct intel_crtc_state *pipe_config)
+{
+	if (HAS_DDI(dev_priv))
+		return pipe_config->port_clock; /* SPLL */
+	else
+		return dev_priv->fdi_pll_freq;
+}
+
 #endif /*  __INTEL_DISPLAY_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
index 8a00e609085f..8c12d5375607 100644
--- a/drivers/gpu/drm/i915/display/intel_dp.c
+++ b/drivers/gpu/drm/i915/display/intel_dp.c
@@ -41,13 +41,13 @@
 
 #include "i915_debugfs.h"
 #include "i915_drv.h"
-#include "i915_trace.h"
 #include "intel_atomic.h"
 #include "intel_audio.h"
 #include "intel_connector.h"
 #include "intel_ddi.h"
 #include "intel_display_types.h"
 #include "intel_dp.h"
+#include "intel_dp_aux.h"
 #include "intel_dp_link_training.h"
 #include "intel_dp_mst.h"
 #include "intel_dpio_phy.h"
@@ -58,10 +58,12 @@
 #include "intel_lspcon.h"
 #include "intel_lvds.h"
 #include "intel_panel.h"
+#include "intel_pps.h"
 #include "intel_psr.h"
 #include "intel_sideband.h"
 #include "intel_tc.h"
 #include "intel_vdsc.h"
+#include "intel_vrr.h"
 
 #define DP_DPRX_ESI_LEN 14
 
@@ -121,6 +123,11 @@ static const struct dp_link_dpll chv_dpll[] = {
 		{ .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } },
 };
 
+const struct dpll *vlv_get_dpll(struct drm_i915_private *i915)
+{
+	return IS_CHERRYVIEW(i915) ? &chv_dpll[0].dpll : &vlv_dpll[0].dpll;
+}
+
 /* Constants for DP DSC configurations */
 static const u8 valid_dsc_bpp[] = {6, 8, 10, 12, 15};
 
@@ -145,12 +152,6 @@ bool intel_dp_is_edp(struct intel_dp *intel_dp)
 
 static void intel_dp_link_down(struct intel_encoder *encoder,
 			       const struct intel_crtc_state *old_crtc_state);
-static bool edp_panel_vdd_on(struct intel_dp *intel_dp);
-static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
-static void vlv_init_panel_power_sequencer(struct intel_encoder *encoder,
-					   const struct intel_crtc_state *crtc_state);
-static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv,
-				      enum pipe pipe);
 static void intel_dp_unset_edid(struct intel_dp *intel_dp);
 
 /* update sink rates from dpcd */
@@ -162,8 +163,7 @@ static void intel_dp_set_sink_rates(struct intel_dp *intel_dp)
 	int i, max_rate;
 	int max_lttpr_rate;
 
-	if (drm_dp_has_quirk(&intel_dp->desc, 0,
-			     DP_DPCD_QUIRK_CAN_DO_MAX_LINK_RATE_3_24_GBPS)) {
+	if (drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_CAN_DO_MAX_LINK_RATE_3_24_GBPS)) {
 		/* Needed, e.g., for Apple MBP 2017, 15 inch eDP Retina panel */
 		static const int quirk_rates[] = { 162000, 270000, 324000 };
 
@@ -480,6 +480,13 @@ int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,
 		return -1;
 	}
 
+	if (intel_dp_is_edp(intel_dp) && !intel_dp->use_max_params) {
+		drm_dbg_kms(&i915->drm,
+			    "Retrying Link training for eDP with max parameters\n");
+		intel_dp->use_max_params = true;
+		return 0;
+	}
+
 	index = intel_dp_rate_index(intel_dp->common_rates,
 				    intel_dp->num_common_rates,
 				    link_rate);
@@ -856,1129 +863,6 @@ intel_dp_mode_valid(struct drm_connector *connector,
 	return intel_mode_valid_max_plane_size(dev_priv, mode, bigjoiner);
 }
 
-u32 intel_dp_pack_aux(const u8 *src, int src_bytes)
-{
-	int i;
-	u32 v = 0;
-
-	if (src_bytes > 4)
-		src_bytes = 4;
-	for (i = 0; i < src_bytes; i++)
-		v |= ((u32)src[i]) << ((3 - i) * 8);
-	return v;
-}
-
-static void intel_dp_unpack_aux(u32 src, u8 *dst, int dst_bytes)
-{
-	int i;
-	if (dst_bytes > 4)
-		dst_bytes = 4;
-	for (i = 0; i < dst_bytes; i++)
-		dst[i] = src >> ((3-i) * 8);
-}
-
-static void
-intel_dp_init_panel_power_sequencer(struct intel_dp *intel_dp);
-static void
-intel_dp_init_panel_power_sequencer_registers(struct intel_dp *intel_dp,
-					      bool force_disable_vdd);
-static void
-intel_dp_pps_init(struct intel_dp *intel_dp);
-
-static intel_wakeref_t
-pps_lock(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	intel_wakeref_t wakeref;
-
-	/*
-	 * See intel_power_sequencer_reset() why we need
-	 * a power domain reference here.
-	 */
-	wakeref = intel_display_power_get(dev_priv,
-					  intel_aux_power_domain(dp_to_dig_port(intel_dp)));
-
-	mutex_lock(&dev_priv->pps_mutex);
-
-	return wakeref;
-}
-
-static intel_wakeref_t
-pps_unlock(struct intel_dp *intel_dp, intel_wakeref_t wakeref)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	mutex_unlock(&dev_priv->pps_mutex);
-	intel_display_power_put(dev_priv,
-				intel_aux_power_domain(dp_to_dig_port(intel_dp)),
-				wakeref);
-	return 0;
-}
-
-#define with_pps_lock(dp, wf) \
-	for ((wf) = pps_lock(dp); (wf); (wf) = pps_unlock((dp), (wf)))
-
-static void
-vlv_power_sequencer_kick(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum pipe pipe = intel_dp->pps_pipe;
-	bool pll_enabled, release_cl_override = false;
-	enum dpio_phy phy = DPIO_PHY(pipe);
-	enum dpio_channel ch = vlv_pipe_to_channel(pipe);
-	u32 DP;
-
-	if (drm_WARN(&dev_priv->drm,
-		     intel_de_read(dev_priv, intel_dp->output_reg) & DP_PORT_EN,
-		     "skipping pipe %c power sequencer kick due to [ENCODER:%d:%s] being active\n",
-		     pipe_name(pipe), dig_port->base.base.base.id,
-		     dig_port->base.base.name))
-		return;
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "kicking pipe %c power sequencer for [ENCODER:%d:%s]\n",
-		    pipe_name(pipe), dig_port->base.base.base.id,
-		    dig_port->base.base.name);
-
-	/* Preserve the BIOS-computed detected bit. This is
-	 * supposed to be read-only.
-	 */
-	DP = intel_de_read(dev_priv, intel_dp->output_reg) & DP_DETECTED;
-	DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
-	DP |= DP_PORT_WIDTH(1);
-	DP |= DP_LINK_TRAIN_PAT_1;
-
-	if (IS_CHERRYVIEW(dev_priv))
-		DP |= DP_PIPE_SEL_CHV(pipe);
-	else
-		DP |= DP_PIPE_SEL(pipe);
-
-	pll_enabled = intel_de_read(dev_priv, DPLL(pipe)) & DPLL_VCO_ENABLE;
-
-	/*
-	 * The DPLL for the pipe must be enabled for this to work.
-	 * So enable temporarily it if it's not already enabled.
-	 */
-	if (!pll_enabled) {
-		release_cl_override = IS_CHERRYVIEW(dev_priv) &&
-			!chv_phy_powergate_ch(dev_priv, phy, ch, true);
-
-		if (vlv_force_pll_on(dev_priv, pipe, IS_CHERRYVIEW(dev_priv) ?
-				     &chv_dpll[0].dpll : &vlv_dpll[0].dpll)) {
-			drm_err(&dev_priv->drm,
-				"Failed to force on pll for pipe %c!\n",
-				pipe_name(pipe));
-			return;
-		}
-	}
-
-	/*
-	 * Similar magic as in intel_dp_enable_port().
-	 * We _must_ do this port enable + disable trick
-	 * to make this power sequencer lock onto the port.
-	 * Otherwise even VDD force bit won't work.
-	 */
-	intel_de_write(dev_priv, intel_dp->output_reg, DP);
-	intel_de_posting_read(dev_priv, intel_dp->output_reg);
-
-	intel_de_write(dev_priv, intel_dp->output_reg, DP | DP_PORT_EN);
-	intel_de_posting_read(dev_priv, intel_dp->output_reg);
-
-	intel_de_write(dev_priv, intel_dp->output_reg, DP & ~DP_PORT_EN);
-	intel_de_posting_read(dev_priv, intel_dp->output_reg);
-
-	if (!pll_enabled) {
-		vlv_force_pll_off(dev_priv, pipe);
-
-		if (release_cl_override)
-			chv_phy_powergate_ch(dev_priv, phy, ch, false);
-	}
-}
-
-static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv)
-{
-	struct intel_encoder *encoder;
-	unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B);
-
-	/*
-	 * We don't have power sequencer currently.
-	 * Pick one that's not used by other ports.
-	 */
-	for_each_intel_dp(&dev_priv->drm, encoder) {
-		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-
-		if (encoder->type == INTEL_OUTPUT_EDP) {
-			drm_WARN_ON(&dev_priv->drm,
-				    intel_dp->active_pipe != INVALID_PIPE &&
-				    intel_dp->active_pipe !=
-				    intel_dp->pps_pipe);
-
-			if (intel_dp->pps_pipe != INVALID_PIPE)
-				pipes &= ~(1 << intel_dp->pps_pipe);
-		} else {
-			drm_WARN_ON(&dev_priv->drm,
-				    intel_dp->pps_pipe != INVALID_PIPE);
-
-			if (intel_dp->active_pipe != INVALID_PIPE)
-				pipes &= ~(1 << intel_dp->active_pipe);
-		}
-	}
-
-	if (pipes == 0)
-		return INVALID_PIPE;
-
-	return ffs(pipes) - 1;
-}
-
-static enum pipe
-vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum pipe pipe;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	/* We should never land here with regular DP ports */
-	drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp));
-
-	drm_WARN_ON(&dev_priv->drm, intel_dp->active_pipe != INVALID_PIPE &&
-		    intel_dp->active_pipe != intel_dp->pps_pipe);
-
-	if (intel_dp->pps_pipe != INVALID_PIPE)
-		return intel_dp->pps_pipe;
-
-	pipe = vlv_find_free_pps(dev_priv);
-
-	/*
-	 * Didn't find one. This should not happen since there
-	 * are two power sequencers and up to two eDP ports.
-	 */
-	if (drm_WARN_ON(&dev_priv->drm, pipe == INVALID_PIPE))
-		pipe = PIPE_A;
-
-	vlv_steal_power_sequencer(dev_priv, pipe);
-	intel_dp->pps_pipe = pipe;
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "picked pipe %c power sequencer for [ENCODER:%d:%s]\n",
-		    pipe_name(intel_dp->pps_pipe),
-		    dig_port->base.base.base.id,
-		    dig_port->base.base.name);
-
-	/* init power sequencer on this pipe and port */
-	intel_dp_init_panel_power_sequencer(intel_dp);
-	intel_dp_init_panel_power_sequencer_registers(intel_dp, true);
-
-	/*
-	 * Even vdd force doesn't work until we've made
-	 * the power sequencer lock in on the port.
-	 */
-	vlv_power_sequencer_kick(intel_dp);
-
-	return intel_dp->pps_pipe;
-}
-
-static int
-bxt_power_sequencer_idx(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	int backlight_controller = dev_priv->vbt.backlight.controller;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	/* We should never land here with regular DP ports */
-	drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp));
-
-	if (!intel_dp->pps_reset)
-		return backlight_controller;
-
-	intel_dp->pps_reset = false;
-
-	/*
-	 * Only the HW needs to be reprogrammed, the SW state is fixed and
-	 * has been setup during connector init.
-	 */
-	intel_dp_init_panel_power_sequencer_registers(intel_dp, false);
-
-	return backlight_controller;
-}
-
-typedef bool (*vlv_pipe_check)(struct drm_i915_private *dev_priv,
-			       enum pipe pipe);
-
-static bool vlv_pipe_has_pp_on(struct drm_i915_private *dev_priv,
-			       enum pipe pipe)
-{
-	return intel_de_read(dev_priv, PP_STATUS(pipe)) & PP_ON;
-}
-
-static bool vlv_pipe_has_vdd_on(struct drm_i915_private *dev_priv,
-				enum pipe pipe)
-{
-	return intel_de_read(dev_priv, PP_CONTROL(pipe)) & EDP_FORCE_VDD;
-}
-
-static bool vlv_pipe_any(struct drm_i915_private *dev_priv,
-			 enum pipe pipe)
-{
-	return true;
-}
-
-static enum pipe
-vlv_initial_pps_pipe(struct drm_i915_private *dev_priv,
-		     enum port port,
-		     vlv_pipe_check pipe_check)
-{
-	enum pipe pipe;
-
-	for (pipe = PIPE_A; pipe <= PIPE_B; pipe++) {
-		u32 port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(pipe)) &
-			PANEL_PORT_SELECT_MASK;
-
-		if (port_sel != PANEL_PORT_SELECT_VLV(port))
-			continue;
-
-		if (!pipe_check(dev_priv, pipe))
-			continue;
-
-		return pipe;
-	}
-
-	return INVALID_PIPE;
-}
-
-static void
-vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum port port = dig_port->base.port;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	/* try to find a pipe with this port selected */
-	/* first pick one where the panel is on */
-	intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
-						  vlv_pipe_has_pp_on);
-	/* didn't find one? pick one where vdd is on */
-	if (intel_dp->pps_pipe == INVALID_PIPE)
-		intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
-							  vlv_pipe_has_vdd_on);
-	/* didn't find one? pick one with just the correct port */
-	if (intel_dp->pps_pipe == INVALID_PIPE)
-		intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
-							  vlv_pipe_any);
-
-	/* didn't find one? just let vlv_power_sequencer_pipe() pick one when needed */
-	if (intel_dp->pps_pipe == INVALID_PIPE) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "no initial power sequencer for [ENCODER:%d:%s]\n",
-			    dig_port->base.base.base.id,
-			    dig_port->base.base.name);
-		return;
-	}
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "initial power sequencer for [ENCODER:%d:%s]: pipe %c\n",
-		    dig_port->base.base.base.id,
-		    dig_port->base.base.name,
-		    pipe_name(intel_dp->pps_pipe));
-
-	intel_dp_init_panel_power_sequencer(intel_dp);
-	intel_dp_init_panel_power_sequencer_registers(intel_dp, false);
-}
-
-void intel_power_sequencer_reset(struct drm_i915_private *dev_priv)
-{
-	struct intel_encoder *encoder;
-
-	if (drm_WARN_ON(&dev_priv->drm,
-			!(IS_VALLEYVIEW(dev_priv) ||
-			  IS_CHERRYVIEW(dev_priv) ||
-			  IS_GEN9_LP(dev_priv))))
-		return;
-
-	/*
-	 * We can't grab pps_mutex here due to deadlock with power_domain
-	 * mutex when power_domain functions are called while holding pps_mutex.
-	 * That also means that in order to use pps_pipe the code needs to
-	 * hold both a power domain reference and pps_mutex, and the power domain
-	 * reference get/put must be done while _not_ holding pps_mutex.
-	 * pps_{lock,unlock}() do these steps in the correct order, so one
-	 * should use them always.
-	 */
-
-	for_each_intel_dp(&dev_priv->drm, encoder) {
-		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-
-		drm_WARN_ON(&dev_priv->drm,
-			    intel_dp->active_pipe != INVALID_PIPE);
-
-		if (encoder->type != INTEL_OUTPUT_EDP)
-			continue;
-
-		if (IS_GEN9_LP(dev_priv))
-			intel_dp->pps_reset = true;
-		else
-			intel_dp->pps_pipe = INVALID_PIPE;
-	}
-}
-
-struct pps_registers {
-	i915_reg_t pp_ctrl;
-	i915_reg_t pp_stat;
-	i915_reg_t pp_on;
-	i915_reg_t pp_off;
-	i915_reg_t pp_div;
-};
-
-static void intel_pps_get_registers(struct intel_dp *intel_dp,
-				    struct pps_registers *regs)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	int pps_idx = 0;
-
-	memset(regs, 0, sizeof(*regs));
-
-	if (IS_GEN9_LP(dev_priv))
-		pps_idx = bxt_power_sequencer_idx(intel_dp);
-	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		pps_idx = vlv_power_sequencer_pipe(intel_dp);
-
-	regs->pp_ctrl = PP_CONTROL(pps_idx);
-	regs->pp_stat = PP_STATUS(pps_idx);
-	regs->pp_on = PP_ON_DELAYS(pps_idx);
-	regs->pp_off = PP_OFF_DELAYS(pps_idx);
-
-	/* Cycle delay moved from PP_DIVISOR to PP_CONTROL */
-	if (IS_GEN9_LP(dev_priv) || INTEL_PCH_TYPE(dev_priv) >= PCH_CNP)
-		regs->pp_div = INVALID_MMIO_REG;
-	else
-		regs->pp_div = PP_DIVISOR(pps_idx);
-}
-
-static i915_reg_t
-_pp_ctrl_reg(struct intel_dp *intel_dp)
-{
-	struct pps_registers regs;
-
-	intel_pps_get_registers(intel_dp, &regs);
-
-	return regs.pp_ctrl;
-}
-
-static i915_reg_t
-_pp_stat_reg(struct intel_dp *intel_dp)
-{
-	struct pps_registers regs;
-
-	intel_pps_get_registers(intel_dp, &regs);
-
-	return regs.pp_stat;
-}
-
-static bool edp_have_panel_power(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
-	    intel_dp->pps_pipe == INVALID_PIPE)
-		return false;
-
-	return (intel_de_read(dev_priv, _pp_stat_reg(intel_dp)) & PP_ON) != 0;
-}
-
-static bool edp_have_panel_vdd(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
-	    intel_dp->pps_pipe == INVALID_PIPE)
-		return false;
-
-	return intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD;
-}
-
-static void
-intel_dp_check_edp(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	if (!edp_have_panel_power(intel_dp) && !edp_have_panel_vdd(intel_dp)) {
-		drm_WARN(&dev_priv->drm, 1,
-			 "eDP powered off while attempting aux channel communication.\n");
-		drm_dbg_kms(&dev_priv->drm, "Status 0x%08x Control 0x%08x\n",
-			    intel_de_read(dev_priv, _pp_stat_reg(intel_dp)),
-			    intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)));
-	}
-}
-
-static u32
-intel_dp_aux_wait_done(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
-	i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp);
-	const unsigned int timeout_ms = 10;
-	u32 status;
-	bool done;
-
-#define C (((status = intel_uncore_read_notrace(&i915->uncore, ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
-	done = wait_event_timeout(i915->gmbus_wait_queue, C,
-				  msecs_to_jiffies_timeout(timeout_ms));
-
-	/* just trace the final value */
-	trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true);
-
-	if (!done)
-		drm_err(&i915->drm,
-			"%s: did not complete or timeout within %ums (status 0x%08x)\n",
-			intel_dp->aux.name, timeout_ms, status);
-#undef C
-
-	return status;
-}
-
-static u32 g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	if (index)
-		return 0;
-
-	/*
-	 * The clock divider is based off the hrawclk, and would like to run at
-	 * 2MHz.  So, take the hrawclk value and divide by 2000 and use that
-	 */
-	return DIV_ROUND_CLOSEST(RUNTIME_INFO(dev_priv)->rawclk_freq, 2000);
-}
-
-static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	u32 freq;
-
-	if (index)
-		return 0;
-
-	/*
-	 * The clock divider is based off the cdclk or PCH rawclk, and would
-	 * like to run at 2MHz.  So, take the cdclk or PCH rawclk value and
-	 * divide by 2000 and use that
-	 */
-	if (dig_port->aux_ch == AUX_CH_A)
-		freq = dev_priv->cdclk.hw.cdclk;
-	else
-		freq = RUNTIME_INFO(dev_priv)->rawclk_freq;
-	return DIV_ROUND_CLOSEST(freq, 2000);
-}
-
-static u32 hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-
-	if (dig_port->aux_ch != AUX_CH_A && HAS_PCH_LPT_H(dev_priv)) {
-		/* Workaround for non-ULT HSW */
-		switch (index) {
-		case 0: return 63;
-		case 1: return 72;
-		default: return 0;
-		}
-	}
-
-	return ilk_get_aux_clock_divider(intel_dp, index);
-}
-
-static u32 skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
-{
-	/*
-	 * SKL doesn't need us to program the AUX clock divider (Hardware will
-	 * derive the clock from CDCLK automatically). We still implement the
-	 * get_aux_clock_divider vfunc to plug-in into the existing code.
-	 */
-	return index ? 0 : 1;
-}
-
-static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp,
-				int send_bytes,
-				u32 aux_clock_divider)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_i915_private *dev_priv =
-			to_i915(dig_port->base.base.dev);
-	u32 precharge, timeout;
-
-	if (IS_GEN(dev_priv, 6))
-		precharge = 3;
-	else
-		precharge = 5;
-
-	if (IS_BROADWELL(dev_priv))
-		timeout = DP_AUX_CH_CTL_TIME_OUT_600us;
-	else
-		timeout = DP_AUX_CH_CTL_TIME_OUT_400us;
-
-	return DP_AUX_CH_CTL_SEND_BUSY |
-	       DP_AUX_CH_CTL_DONE |
-	       DP_AUX_CH_CTL_INTERRUPT |
-	       DP_AUX_CH_CTL_TIME_OUT_ERROR |
-	       timeout |
-	       DP_AUX_CH_CTL_RECEIVE_ERROR |
-	       (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
-	       (precharge << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
-	       (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT);
-}
-
-static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp,
-				int send_bytes,
-				u32 unused)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_i915_private *i915 =
-			to_i915(dig_port->base.base.dev);
-	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
-	u32 ret;
-
-	ret = DP_AUX_CH_CTL_SEND_BUSY |
-	      DP_AUX_CH_CTL_DONE |
-	      DP_AUX_CH_CTL_INTERRUPT |
-	      DP_AUX_CH_CTL_TIME_OUT_ERROR |
-	      DP_AUX_CH_CTL_TIME_OUT_MAX |
-	      DP_AUX_CH_CTL_RECEIVE_ERROR |
-	      (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
-	      DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(32) |
-	      DP_AUX_CH_CTL_SYNC_PULSE_SKL(32);
-
-	if (intel_phy_is_tc(i915, phy) &&
-	    dig_port->tc_mode == TC_PORT_TBT_ALT)
-		ret |= DP_AUX_CH_CTL_TBT_IO;
-
-	return ret;
-}
-
-static int
-intel_dp_aux_xfer(struct intel_dp *intel_dp,
-		  const u8 *send, int send_bytes,
-		  u8 *recv, int recv_size,
-		  u32 aux_send_ctl_flags)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_i915_private *i915 =
-			to_i915(dig_port->base.base.dev);
-	struct intel_uncore *uncore = &i915->uncore;
-	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
-	bool is_tc_port = intel_phy_is_tc(i915, phy);
-	i915_reg_t ch_ctl, ch_data[5];
-	u32 aux_clock_divider;
-	enum intel_display_power_domain aux_domain;
-	intel_wakeref_t aux_wakeref;
-	intel_wakeref_t pps_wakeref;
-	int i, ret, recv_bytes;
-	int try, clock = 0;
-	u32 status;
-	bool vdd;
-
-	ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp);
-	for (i = 0; i < ARRAY_SIZE(ch_data); i++)
-		ch_data[i] = intel_dp->aux_ch_data_reg(intel_dp, i);
-
-	if (is_tc_port)
-		intel_tc_port_lock(dig_port);
-
-	aux_domain = intel_aux_power_domain(dig_port);
-
-	aux_wakeref = intel_display_power_get(i915, aux_domain);
-	pps_wakeref = pps_lock(intel_dp);
-
-	/*
-	 * We will be called with VDD already enabled for dpcd/edid/oui reads.
-	 * In such cases we want to leave VDD enabled and it's up to upper layers
-	 * to turn it off. But for eg. i2c-dev access we need to turn it on/off
-	 * ourselves.
-	 */
-	vdd = edp_panel_vdd_on(intel_dp);
-
-	/* dp aux is extremely sensitive to irq latency, hence request the
-	 * lowest possible wakeup latency and so prevent the cpu from going into
-	 * deep sleep states.
-	 */
-	cpu_latency_qos_update_request(&intel_dp->pm_qos, 0);
-
-	intel_dp_check_edp(intel_dp);
-
-	/* Try to wait for any previous AUX channel activity */
-	for (try = 0; try < 3; try++) {
-		status = intel_uncore_read_notrace(uncore, ch_ctl);
-		if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0)
-			break;
-		msleep(1);
-	}
-	/* just trace the final value */
-	trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true);
-
-	if (try == 3) {
-		const u32 status = intel_uncore_read(uncore, ch_ctl);
-
-		if (status != intel_dp->aux_busy_last_status) {
-			drm_WARN(&i915->drm, 1,
-				 "%s: not started (status 0x%08x)\n",
-				 intel_dp->aux.name, status);
-			intel_dp->aux_busy_last_status = status;
-		}
-
-		ret = -EBUSY;
-		goto out;
-	}
-
-	/* Only 5 data registers! */
-	if (drm_WARN_ON(&i915->drm, send_bytes > 20 || recv_size > 20)) {
-		ret = -E2BIG;
-		goto out;
-	}
-
-	while ((aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, clock++))) {
-		u32 send_ctl = intel_dp->get_aux_send_ctl(intel_dp,
-							  send_bytes,
-							  aux_clock_divider);
-
-		send_ctl |= aux_send_ctl_flags;
-
-		/* Must try at least 3 times according to DP spec */
-		for (try = 0; try < 5; try++) {
-			/* Load the send data into the aux channel data registers */
-			for (i = 0; i < send_bytes; i += 4)
-				intel_uncore_write(uncore,
-						   ch_data[i >> 2],
-						   intel_dp_pack_aux(send + i,
-								     send_bytes - i));
-
-			/* Send the command and wait for it to complete */
-			intel_uncore_write(uncore, ch_ctl, send_ctl);
-
-			status = intel_dp_aux_wait_done(intel_dp);
-
-			/* Clear done status and any errors */
-			intel_uncore_write(uncore,
-					   ch_ctl,
-					   status |
-					   DP_AUX_CH_CTL_DONE |
-					   DP_AUX_CH_CTL_TIME_OUT_ERROR |
-					   DP_AUX_CH_CTL_RECEIVE_ERROR);
-
-			/* DP CTS 1.2 Core Rev 1.1, 4.2.1.1 & 4.2.1.2
-			 *   400us delay required for errors and timeouts
-			 *   Timeout errors from the HW already meet this
-			 *   requirement so skip to next iteration
-			 */
-			if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR)
-				continue;
-
-			if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
-				usleep_range(400, 500);
-				continue;
-			}
-			if (status & DP_AUX_CH_CTL_DONE)
-				goto done;
-		}
-	}
-
-	if ((status & DP_AUX_CH_CTL_DONE) == 0) {
-		drm_err(&i915->drm, "%s: not done (status 0x%08x)\n",
-			intel_dp->aux.name, status);
-		ret = -EBUSY;
-		goto out;
-	}
-
-done:
-	/* Check for timeout or receive error.
-	 * Timeouts occur when the sink is not connected
-	 */
-	if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
-		drm_err(&i915->drm, "%s: receive error (status 0x%08x)\n",
-			intel_dp->aux.name, status);
-		ret = -EIO;
-		goto out;
-	}
-
-	/* Timeouts occur when the device isn't connected, so they're
-	 * "normal" -- don't fill the kernel log with these */
-	if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR) {
-		drm_dbg_kms(&i915->drm, "%s: timeout (status 0x%08x)\n",
-			    intel_dp->aux.name, status);
-		ret = -ETIMEDOUT;
-		goto out;
-	}
-
-	/* Unload any bytes sent back from the other side */
-	recv_bytes = ((status & DP_AUX_CH_CTL_MESSAGE_SIZE_MASK) >>
-		      DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT);
-
-	/*
-	 * By BSpec: "Message sizes of 0 or >20 are not allowed."
-	 * We have no idea of what happened so we return -EBUSY so
-	 * drm layer takes care for the necessary retries.
-	 */
-	if (recv_bytes == 0 || recv_bytes > 20) {
-		drm_dbg_kms(&i915->drm,
-			    "%s: Forbidden recv_bytes = %d on aux transaction\n",
-			    intel_dp->aux.name, recv_bytes);
-		ret = -EBUSY;
-		goto out;
-	}
-
-	if (recv_bytes > recv_size)
-		recv_bytes = recv_size;
-
-	for (i = 0; i < recv_bytes; i += 4)
-		intel_dp_unpack_aux(intel_uncore_read(uncore, ch_data[i >> 2]),
-				    recv + i, recv_bytes - i);
-
-	ret = recv_bytes;
-out:
-	cpu_latency_qos_update_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
-
-	if (vdd)
-		edp_panel_vdd_off(intel_dp, false);
-
-	pps_unlock(intel_dp, pps_wakeref);
-	intel_display_power_put_async(i915, aux_domain, aux_wakeref);
-
-	if (is_tc_port)
-		intel_tc_port_unlock(dig_port);
-
-	return ret;
-}
-
-#define BARE_ADDRESS_SIZE	3
-#define HEADER_SIZE		(BARE_ADDRESS_SIZE + 1)
-
-static void
-intel_dp_aux_header(u8 txbuf[HEADER_SIZE],
-		    const struct drm_dp_aux_msg *msg)
-{
-	txbuf[0] = (msg->request << 4) | ((msg->address >> 16) & 0xf);
-	txbuf[1] = (msg->address >> 8) & 0xff;
-	txbuf[2] = msg->address & 0xff;
-	txbuf[3] = msg->size - 1;
-}
-
-static u32 intel_dp_aux_xfer_flags(const struct drm_dp_aux_msg *msg)
-{
-	/*
-	 * If we're trying to send the HDCP Aksv, we need to set a the Aksv
-	 * select bit to inform the hardware to send the Aksv after our header
-	 * since we can't access that data from software.
-	 */
-	if ((msg->request & ~DP_AUX_I2C_MOT) == DP_AUX_NATIVE_WRITE &&
-	    msg->address == DP_AUX_HDCP_AKSV)
-		return DP_AUX_CH_CTL_AUX_AKSV_SELECT;
-
-	return 0;
-}
-
-static ssize_t
-intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg)
-{
-	struct intel_dp *intel_dp = container_of(aux, struct intel_dp, aux);
-	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
-	u8 txbuf[20], rxbuf[20];
-	size_t txsize, rxsize;
-	u32 flags = intel_dp_aux_xfer_flags(msg);
-	int ret;
-
-	intel_dp_aux_header(txbuf, msg);
-
-	switch (msg->request & ~DP_AUX_I2C_MOT) {
-	case DP_AUX_NATIVE_WRITE:
-	case DP_AUX_I2C_WRITE:
-	case DP_AUX_I2C_WRITE_STATUS_UPDATE:
-		txsize = msg->size ? HEADER_SIZE + msg->size : BARE_ADDRESS_SIZE;
-		rxsize = 2; /* 0 or 1 data bytes */
-
-		if (drm_WARN_ON(&i915->drm, txsize > 20))
-			return -E2BIG;
-
-		drm_WARN_ON(&i915->drm, !msg->buffer != !msg->size);
-
-		if (msg->buffer)
-			memcpy(txbuf + HEADER_SIZE, msg->buffer, msg->size);
-
-		ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize,
-					rxbuf, rxsize, flags);
-		if (ret > 0) {
-			msg->reply = rxbuf[0] >> 4;
-
-			if (ret > 1) {
-				/* Number of bytes written in a short write. */
-				ret = clamp_t(int, rxbuf[1], 0, msg->size);
-			} else {
-				/* Return payload size. */
-				ret = msg->size;
-			}
-		}
-		break;
-
-	case DP_AUX_NATIVE_READ:
-	case DP_AUX_I2C_READ:
-		txsize = msg->size ? HEADER_SIZE : BARE_ADDRESS_SIZE;
-		rxsize = msg->size + 1;
-
-		if (drm_WARN_ON(&i915->drm, rxsize > 20))
-			return -E2BIG;
-
-		ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize,
-					rxbuf, rxsize, flags);
-		if (ret > 0) {
-			msg->reply = rxbuf[0] >> 4;
-			/*
-			 * Assume happy day, and copy the data. The caller is
-			 * expected to check msg->reply before touching it.
-			 *
-			 * Return payload size.
-			 */
-			ret--;
-			memcpy(msg->buffer, rxbuf + 1, ret);
-		}
-		break;
-
-	default:
-		ret = -EINVAL;
-		break;
-	}
-
-	return ret;
-}
-
-
-static i915_reg_t g4x_aux_ctl_reg(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_D:
-		return DP_AUX_CH_CTL(aux_ch);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_CTL(AUX_CH_B);
-	}
-}
-
-static i915_reg_t g4x_aux_data_reg(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_D:
-		return DP_AUX_CH_DATA(aux_ch, index);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_DATA(AUX_CH_B, index);
-	}
-}
-
-static i915_reg_t ilk_aux_ctl_reg(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_A:
-		return DP_AUX_CH_CTL(aux_ch);
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_D:
-		return PCH_DP_AUX_CH_CTL(aux_ch);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_CTL(AUX_CH_A);
-	}
-}
-
-static i915_reg_t ilk_aux_data_reg(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_A:
-		return DP_AUX_CH_DATA(aux_ch, index);
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_D:
-		return PCH_DP_AUX_CH_DATA(aux_ch, index);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_DATA(AUX_CH_A, index);
-	}
-}
-
-static i915_reg_t skl_aux_ctl_reg(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_A:
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_D:
-	case AUX_CH_E:
-	case AUX_CH_F:
-		return DP_AUX_CH_CTL(aux_ch);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_CTL(AUX_CH_A);
-	}
-}
-
-static i915_reg_t skl_aux_data_reg(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_A:
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_D:
-	case AUX_CH_E:
-	case AUX_CH_F:
-		return DP_AUX_CH_DATA(aux_ch, index);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_DATA(AUX_CH_A, index);
-	}
-}
-
-static i915_reg_t tgl_aux_ctl_reg(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_A:
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_USBC1:
-	case AUX_CH_USBC2:
-	case AUX_CH_USBC3:
-	case AUX_CH_USBC4:
-	case AUX_CH_USBC5:
-	case AUX_CH_USBC6:
-		return DP_AUX_CH_CTL(aux_ch);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_CTL(AUX_CH_A);
-	}
-}
-
-static i915_reg_t tgl_aux_data_reg(struct intel_dp *intel_dp, int index)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	switch (aux_ch) {
-	case AUX_CH_A:
-	case AUX_CH_B:
-	case AUX_CH_C:
-	case AUX_CH_USBC1:
-	case AUX_CH_USBC2:
-	case AUX_CH_USBC3:
-	case AUX_CH_USBC4:
-	case AUX_CH_USBC5:
-	case AUX_CH_USBC6:
-		return DP_AUX_CH_DATA(aux_ch, index);
-	default:
-		MISSING_CASE(aux_ch);
-		return DP_AUX_CH_DATA(AUX_CH_A, index);
-	}
-}
-
-static void
-intel_dp_aux_fini(struct intel_dp *intel_dp)
-{
-	if (cpu_latency_qos_request_active(&intel_dp->pm_qos))
-		cpu_latency_qos_remove_request(&intel_dp->pm_qos);
-
-	kfree(intel_dp->aux.name);
-}
-
-static void
-intel_dp_aux_init(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *encoder = &dig_port->base;
-	enum aux_ch aux_ch = dig_port->aux_ch;
-
-	if (INTEL_GEN(dev_priv) >= 12) {
-		intel_dp->aux_ch_ctl_reg = tgl_aux_ctl_reg;
-		intel_dp->aux_ch_data_reg = tgl_aux_data_reg;
-	} else if (INTEL_GEN(dev_priv) >= 9) {
-		intel_dp->aux_ch_ctl_reg = skl_aux_ctl_reg;
-		intel_dp->aux_ch_data_reg = skl_aux_data_reg;
-	} else if (HAS_PCH_SPLIT(dev_priv)) {
-		intel_dp->aux_ch_ctl_reg = ilk_aux_ctl_reg;
-		intel_dp->aux_ch_data_reg = ilk_aux_data_reg;
-	} else {
-		intel_dp->aux_ch_ctl_reg = g4x_aux_ctl_reg;
-		intel_dp->aux_ch_data_reg = g4x_aux_data_reg;
-	}
-
-	if (INTEL_GEN(dev_priv) >= 9)
-		intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider;
-	else if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
-		intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider;
-	else if (HAS_PCH_SPLIT(dev_priv))
-		intel_dp->get_aux_clock_divider = ilk_get_aux_clock_divider;
-	else
-		intel_dp->get_aux_clock_divider = g4x_get_aux_clock_divider;
-
-	if (INTEL_GEN(dev_priv) >= 9)
-		intel_dp->get_aux_send_ctl = skl_get_aux_send_ctl;
-	else
-		intel_dp->get_aux_send_ctl = g4x_get_aux_send_ctl;
-
-	drm_dp_aux_init(&intel_dp->aux);
-
-	/* Failure to allocate our preferred name is not critical */
-	if (INTEL_GEN(dev_priv) >= 12 && aux_ch >= AUX_CH_USBC1)
-		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX USBC%c/%s",
-					       aux_ch - AUX_CH_USBC1 + '1',
-					       encoder->base.name);
-	else
-		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX %c/%s",
-					       aux_ch_name(aux_ch),
-					       encoder->base.name);
-
-	intel_dp->aux.transfer = intel_dp_aux_transfer;
-	cpu_latency_qos_add_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
-}
-
 bool intel_dp_source_supports_hbr2(struct intel_dp *intel_dp)
 {
 	int max_rate = intel_dp->source_rates[intel_dp->num_source_rates - 1];
@@ -2290,6 +1174,44 @@ intel_dp_compute_link_config_wide(struct intel_dp *intel_dp,
 	return -EINVAL;
 }
 
+/* Optimize link config in order: max bpp, min lanes, min clock */
+static int
+intel_dp_compute_link_config_fast(struct intel_dp *intel_dp,
+				  struct intel_crtc_state *pipe_config,
+				  const struct link_config_limits *limits)
+{
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	int bpp, clock, lane_count;
+	int mode_rate, link_clock, link_avail;
+
+	for (bpp = limits->max_bpp; bpp >= limits->min_bpp; bpp -= 2 * 3) {
+		int output_bpp = intel_dp_output_bpp(pipe_config->output_format, bpp);
+
+		mode_rate = intel_dp_link_required(adjusted_mode->crtc_clock,
+						   output_bpp);
+
+		for (lane_count = limits->min_lane_count;
+		     lane_count <= limits->max_lane_count;
+		     lane_count <<= 1) {
+			for (clock = limits->min_clock; clock <= limits->max_clock; clock++) {
+				link_clock = intel_dp->common_rates[clock];
+				link_avail = intel_dp_max_data_rate(link_clock,
+								    lane_count);
+
+				if (mode_rate <= link_avail) {
+					pipe_config->lane_count = lane_count;
+					pipe_config->pipe_bpp = bpp;
+					pipe_config->port_clock = link_clock;
+
+					return 0;
+				}
+			}
+		}
+	}
+
+	return -EINVAL;
+}
+
 static int intel_dp_dsc_compute_bpp(struct intel_dp *intel_dp, u8 dsc_max_bpc)
 {
 	int i, num_bpc;
@@ -2513,13 +1435,14 @@ intel_dp_compute_link_config(struct intel_encoder *encoder,
 	limits.min_bpp = intel_dp_min_bpp(pipe_config->output_format);
 	limits.max_bpp = intel_dp_max_bpp(intel_dp, pipe_config);
 
-	if (intel_dp_is_edp(intel_dp)) {
+	if (intel_dp->use_max_params) {
 		/*
 		 * Use the maximum clock and number of lanes the eDP panel
-		 * advertizes being capable of. The panels are generally
+		 * advertizes being capable of in case the initial fast
+		 * optimal params failed us. The panels are generally
 		 * designed to support only a single clock and lane
-		 * configuration, and typically these values correspond to the
-		 * native resolution of the panel.
+		 * configuration, and typically on older panels these
+		 * values correspond to the native resolution of the panel.
 		 */
 		limits.min_lane_count = limits.max_lane_count;
 		limits.min_clock = limits.max_clock;
@@ -2538,11 +1461,22 @@ intel_dp_compute_link_config(struct intel_encoder *encoder,
 	    intel_dp_can_bigjoiner(intel_dp))
 		pipe_config->bigjoiner = true;
 
-	/*
-	 * Optimize for slow and wide. This is the place to add alternative
-	 * optimization policy.
-	 */
-	ret = intel_dp_compute_link_config_wide(intel_dp, pipe_config, &limits);
+	if (intel_dp_is_edp(intel_dp))
+		/*
+		 * Optimize for fast and narrow. eDP 1.3 section 3.3 and eDP 1.4
+		 * section A.1: "It is recommended that the minimum number of
+		 * lanes be used, using the minimum link rate allowed for that
+		 * lane configuration."
+		 *
+		 * Note that we fall back to the max clock and lane count for eDP
+		 * panels that fail with the fast optimal settings (see
+		 * intel_dp->use_max_params), in which case the fast vs. wide
+		 * choice doesn't matter.
+		 */
+		ret = intel_dp_compute_link_config_fast(intel_dp, pipe_config, &limits);
+	else
+		/* Optimize for slow and wide. */
+		ret = intel_dp_compute_link_config_wide(intel_dp, pipe_config, &limits);
 
 	/* enable compression if the mode doesn't fit available BW */
 	drm_dbg_kms(&i915->drm, "Force DSC en = %d\n", intel_dp->force_dsc_en);
@@ -2791,6 +1725,9 @@ intel_dp_drrs_compute_config(struct intel_dp *intel_dp,
 	struct intel_connector *intel_connector = intel_dp->attached_connector;
 	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
 
+	if (pipe_config->vrr.enable)
+		return;
+
 	/*
 	 * DRRS and PSR can't be enable together, so giving preference to PSR
 	 * as it allows more power-savings by complete shutting down display,
@@ -2823,8 +1760,7 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	struct intel_connector *intel_connector = intel_dp->attached_connector;
 	struct intel_digital_connector_state *intel_conn_state =
 		to_intel_digital_connector_state(conn_state);
-	bool constant_n = drm_dp_has_quirk(&intel_dp->desc, 0,
-					   DP_DPCD_QUIRK_CONSTANT_N);
+	bool constant_n = drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_CONSTANT_N);
 	int ret = 0, output_bpp;
 
 	if (HAS_PCH_SPLIT(dev_priv) && !HAS_DDI(dev_priv) && port != PORT_A)
@@ -2894,6 +1830,7 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	if (!HAS_DDI(dev_priv))
 		intel_dp_set_clock(encoder, pipe_config);
 
+	intel_vrr_compute_config(pipe_config, conn_state);
 	intel_psr_compute_config(intel_dp, pipe_config);
 	intel_dp_drrs_compute_config(intel_dp, pipe_config, output_bpp,
 				     constant_n);
@@ -2994,431 +1931,6 @@ static void intel_dp_prepare(struct intel_encoder *encoder,
 	}
 }
 
-#define IDLE_ON_MASK		(PP_ON | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
-#define IDLE_ON_VALUE   	(PP_ON | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_ON_IDLE)
-
-#define IDLE_OFF_MASK		(PP_ON | PP_SEQUENCE_MASK | 0                     | 0)
-#define IDLE_OFF_VALUE		(0     | PP_SEQUENCE_NONE | 0                     | 0)
-
-#define IDLE_CYCLE_MASK		(PP_ON | PP_SEQUENCE_MASK | PP_CYCLE_DELAY_ACTIVE | PP_SEQUENCE_STATE_MASK)
-#define IDLE_CYCLE_VALUE	(0     | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_OFF_IDLE)
-
-static void intel_pps_verify_state(struct intel_dp *intel_dp);
-
-static void wait_panel_status(struct intel_dp *intel_dp,
-				       u32 mask,
-				       u32 value)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	i915_reg_t pp_stat_reg, pp_ctrl_reg;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	intel_pps_verify_state(intel_dp);
-
-	pp_stat_reg = _pp_stat_reg(intel_dp);
-	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "mask %08x value %08x status %08x control %08x\n",
-		    mask, value,
-		    intel_de_read(dev_priv, pp_stat_reg),
-		    intel_de_read(dev_priv, pp_ctrl_reg));
-
-	if (intel_de_wait_for_register(dev_priv, pp_stat_reg,
-				       mask, value, 5000))
-		drm_err(&dev_priv->drm,
-			"Panel status timeout: status %08x control %08x\n",
-			intel_de_read(dev_priv, pp_stat_reg),
-			intel_de_read(dev_priv, pp_ctrl_reg));
-
-	drm_dbg_kms(&dev_priv->drm, "Wait complete\n");
-}
-
-static void wait_panel_on(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
-
-	drm_dbg_kms(&i915->drm, "Wait for panel power on\n");
-	wait_panel_status(intel_dp, IDLE_ON_MASK, IDLE_ON_VALUE);
-}
-
-static void wait_panel_off(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
-
-	drm_dbg_kms(&i915->drm, "Wait for panel power off time\n");
-	wait_panel_status(intel_dp, IDLE_OFF_MASK, IDLE_OFF_VALUE);
-}
-
-static void wait_panel_power_cycle(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
-	ktime_t panel_power_on_time;
-	s64 panel_power_off_duration;
-
-	drm_dbg_kms(&i915->drm, "Wait for panel power cycle\n");
-
-	/* take the difference of currrent time and panel power off time
-	 * and then make panel wait for t11_t12 if needed. */
-	panel_power_on_time = ktime_get_boottime();
-	panel_power_off_duration = ktime_ms_delta(panel_power_on_time, intel_dp->panel_power_off_time);
-
-	/* When we disable the VDD override bit last we have to do the manual
-	 * wait. */
-	if (panel_power_off_duration < (s64)intel_dp->panel_power_cycle_delay)
-		wait_remaining_ms_from_jiffies(jiffies,
-				       intel_dp->panel_power_cycle_delay - panel_power_off_duration);
-
-	wait_panel_status(intel_dp, IDLE_CYCLE_MASK, IDLE_CYCLE_VALUE);
-}
-
-static void wait_backlight_on(struct intel_dp *intel_dp)
-{
-	wait_remaining_ms_from_jiffies(intel_dp->last_power_on,
-				       intel_dp->backlight_on_delay);
-}
-
-static void edp_wait_backlight_off(struct intel_dp *intel_dp)
-{
-	wait_remaining_ms_from_jiffies(intel_dp->last_backlight_off,
-				       intel_dp->backlight_off_delay);
-}
-
-/* Read the current pp_control value, unlocking the register if it
- * is locked
- */
-
-static  u32 ilk_get_pp_control(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	u32 control;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	control = intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp));
-	if (drm_WARN_ON(&dev_priv->drm, !HAS_DDI(dev_priv) &&
-			(control & PANEL_UNLOCK_MASK) != PANEL_UNLOCK_REGS)) {
-		control &= ~PANEL_UNLOCK_MASK;
-		control |= PANEL_UNLOCK_REGS;
-	}
-	return control;
-}
-
-/*
- * Must be paired with edp_panel_vdd_off().
- * Must hold pps_mutex around the whole on/off sequence.
- * Can be nested with intel_edp_panel_vdd_{on,off}() calls.
- */
-static bool edp_panel_vdd_on(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	u32 pp;
-	i915_reg_t pp_stat_reg, pp_ctrl_reg;
-	bool need_to_disable = !intel_dp->want_panel_vdd;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if (!intel_dp_is_edp(intel_dp))
-		return false;
-
-	cancel_delayed_work(&intel_dp->panel_vdd_work);
-	intel_dp->want_panel_vdd = true;
-
-	if (edp_have_panel_vdd(intel_dp))
-		return need_to_disable;
-
-	drm_WARN_ON(&dev_priv->drm, intel_dp->vdd_wakeref);
-	intel_dp->vdd_wakeref = intel_display_power_get(dev_priv,
-							intel_aux_power_domain(dig_port));
-
-	drm_dbg_kms(&dev_priv->drm, "Turning [ENCODER:%d:%s] VDD on\n",
-		    dig_port->base.base.base.id,
-		    dig_port->base.base.name);
-
-	if (!edp_have_panel_power(intel_dp))
-		wait_panel_power_cycle(intel_dp);
-
-	pp = ilk_get_pp_control(intel_dp);
-	pp |= EDP_FORCE_VDD;
-
-	pp_stat_reg = _pp_stat_reg(intel_dp);
-	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-
-	intel_de_write(dev_priv, pp_ctrl_reg, pp);
-	intel_de_posting_read(dev_priv, pp_ctrl_reg);
-	drm_dbg_kms(&dev_priv->drm, "PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
-		    intel_de_read(dev_priv, pp_stat_reg),
-		    intel_de_read(dev_priv, pp_ctrl_reg));
-	/*
-	 * If the panel wasn't on, delay before accessing aux channel
-	 */
-	if (!edp_have_panel_power(intel_dp)) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "[ENCODER:%d:%s] panel power wasn't enabled\n",
-			    dig_port->base.base.base.id,
-			    dig_port->base.base.name);
-		msleep(intel_dp->panel_power_up_delay);
-	}
-
-	return need_to_disable;
-}
-
-/*
- * Must be paired with intel_edp_panel_vdd_off() or
- * intel_edp_panel_off().
- * Nested calls to these functions are not allowed since
- * we drop the lock. Caller must use some higher level
- * locking to prevent nested calls from other threads.
- */
-void intel_edp_panel_vdd_on(struct intel_dp *intel_dp)
-{
-	intel_wakeref_t wakeref;
-	bool vdd;
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	vdd = false;
-	with_pps_lock(intel_dp, wakeref)
-		vdd = edp_panel_vdd_on(intel_dp);
-	I915_STATE_WARN(!vdd, "[ENCODER:%d:%s] VDD already requested on\n",
-			dp_to_dig_port(intel_dp)->base.base.base.id,
-			dp_to_dig_port(intel_dp)->base.base.name);
-}
-
-static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port =
-		dp_to_dig_port(intel_dp);
-	u32 pp;
-	i915_reg_t pp_stat_reg, pp_ctrl_reg;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	drm_WARN_ON(&dev_priv->drm, intel_dp->want_panel_vdd);
-
-	if (!edp_have_panel_vdd(intel_dp))
-		return;
-
-	drm_dbg_kms(&dev_priv->drm, "Turning [ENCODER:%d:%s] VDD off\n",
-		    dig_port->base.base.base.id,
-		    dig_port->base.base.name);
-
-	pp = ilk_get_pp_control(intel_dp);
-	pp &= ~EDP_FORCE_VDD;
-
-	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-	pp_stat_reg = _pp_stat_reg(intel_dp);
-
-	intel_de_write(dev_priv, pp_ctrl_reg, pp);
-	intel_de_posting_read(dev_priv, pp_ctrl_reg);
-
-	/* Make sure sequencer is idle before allowing subsequent activity */
-	drm_dbg_kms(&dev_priv->drm, "PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
-		    intel_de_read(dev_priv, pp_stat_reg),
-		    intel_de_read(dev_priv, pp_ctrl_reg));
-
-	if ((pp & PANEL_POWER_ON) == 0)
-		intel_dp->panel_power_off_time = ktime_get_boottime();
-
-	intel_display_power_put(dev_priv,
-				intel_aux_power_domain(dig_port),
-				fetch_and_zero(&intel_dp->vdd_wakeref));
-}
-
-static void edp_panel_vdd_work(struct work_struct *__work)
-{
-	struct intel_dp *intel_dp =
-		container_of(to_delayed_work(__work),
-			     struct intel_dp, panel_vdd_work);
-	intel_wakeref_t wakeref;
-
-	with_pps_lock(intel_dp, wakeref) {
-		if (!intel_dp->want_panel_vdd)
-			edp_panel_vdd_off_sync(intel_dp);
-	}
-}
-
-static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
-{
-	unsigned long delay;
-
-	/*
-	 * Queue the timer to fire a long time from now (relative to the power
-	 * down delay) to keep the panel power up across a sequence of
-	 * operations.
-	 */
-	delay = msecs_to_jiffies(intel_dp->panel_power_cycle_delay * 5);
-	schedule_delayed_work(&intel_dp->panel_vdd_work, delay);
-}
-
-/*
- * Must be paired with edp_panel_vdd_on().
- * Must hold pps_mutex around the whole on/off sequence.
- * Can be nested with intel_edp_panel_vdd_{on,off}() calls.
- */
-static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	I915_STATE_WARN(!intel_dp->want_panel_vdd, "[ENCODER:%d:%s] VDD not forced on",
-			dp_to_dig_port(intel_dp)->base.base.base.id,
-			dp_to_dig_port(intel_dp)->base.base.name);
-
-	intel_dp->want_panel_vdd = false;
-
-	if (sync)
-		edp_panel_vdd_off_sync(intel_dp);
-	else
-		edp_panel_vdd_schedule_off(intel_dp);
-}
-
-static void edp_panel_on(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	u32 pp;
-	i915_reg_t pp_ctrl_reg;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	drm_dbg_kms(&dev_priv->drm, "Turn [ENCODER:%d:%s] panel power on\n",
-		    dp_to_dig_port(intel_dp)->base.base.base.id,
-		    dp_to_dig_port(intel_dp)->base.base.name);
-
-	if (drm_WARN(&dev_priv->drm, edp_have_panel_power(intel_dp),
-		     "[ENCODER:%d:%s] panel power already on\n",
-		     dp_to_dig_port(intel_dp)->base.base.base.id,
-		     dp_to_dig_port(intel_dp)->base.base.name))
-		return;
-
-	wait_panel_power_cycle(intel_dp);
-
-	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-	pp = ilk_get_pp_control(intel_dp);
-	if (IS_GEN(dev_priv, 5)) {
-		/* ILK workaround: disable reset around power sequence */
-		pp &= ~PANEL_POWER_RESET;
-		intel_de_write(dev_priv, pp_ctrl_reg, pp);
-		intel_de_posting_read(dev_priv, pp_ctrl_reg);
-	}
-
-	pp |= PANEL_POWER_ON;
-	if (!IS_GEN(dev_priv, 5))
-		pp |= PANEL_POWER_RESET;
-
-	intel_de_write(dev_priv, pp_ctrl_reg, pp);
-	intel_de_posting_read(dev_priv, pp_ctrl_reg);
-
-	wait_panel_on(intel_dp);
-	intel_dp->last_power_on = jiffies;
-
-	if (IS_GEN(dev_priv, 5)) {
-		pp |= PANEL_POWER_RESET; /* restore panel reset bit */
-		intel_de_write(dev_priv, pp_ctrl_reg, pp);
-		intel_de_posting_read(dev_priv, pp_ctrl_reg);
-	}
-}
-
-void intel_edp_panel_on(struct intel_dp *intel_dp)
-{
-	intel_wakeref_t wakeref;
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	with_pps_lock(intel_dp, wakeref)
-		edp_panel_on(intel_dp);
-}
-
-
-static void edp_panel_off(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	u32 pp;
-	i915_reg_t pp_ctrl_reg;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	drm_dbg_kms(&dev_priv->drm, "Turn [ENCODER:%d:%s] panel power off\n",
-		    dig_port->base.base.base.id, dig_port->base.base.name);
-
-	drm_WARN(&dev_priv->drm, !intel_dp->want_panel_vdd,
-		 "Need [ENCODER:%d:%s] VDD to turn off panel\n",
-		 dig_port->base.base.base.id, dig_port->base.base.name);
-
-	pp = ilk_get_pp_control(intel_dp);
-	/* We need to switch off panel power _and_ force vdd, for otherwise some
-	 * panels get very unhappy and cease to work. */
-	pp &= ~(PANEL_POWER_ON | PANEL_POWER_RESET | EDP_FORCE_VDD |
-		EDP_BLC_ENABLE);
-
-	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-
-	intel_dp->want_panel_vdd = false;
-
-	intel_de_write(dev_priv, pp_ctrl_reg, pp);
-	intel_de_posting_read(dev_priv, pp_ctrl_reg);
-
-	wait_panel_off(intel_dp);
-	intel_dp->panel_power_off_time = ktime_get_boottime();
-
-	/* We got a reference when we enabled the VDD. */
-	intel_display_power_put(dev_priv,
-				intel_aux_power_domain(dig_port),
-				fetch_and_zero(&intel_dp->vdd_wakeref));
-}
-
-void intel_edp_panel_off(struct intel_dp *intel_dp)
-{
-	intel_wakeref_t wakeref;
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	with_pps_lock(intel_dp, wakeref)
-		edp_panel_off(intel_dp);
-}
-
-/* Enable backlight in the panel power control. */
-static void _intel_edp_backlight_on(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	intel_wakeref_t wakeref;
-
-	/*
-	 * If we enable the backlight right away following a panel power
-	 * on, we may see slight flicker as the panel syncs with the eDP
-	 * link.  So delay a bit to make sure the image is solid before
-	 * allowing it to appear.
-	 */
-	wait_backlight_on(intel_dp);
-
-	with_pps_lock(intel_dp, wakeref) {
-		i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-		u32 pp;
-
-		pp = ilk_get_pp_control(intel_dp);
-		pp |= EDP_BLC_ENABLE;
-
-		intel_de_write(dev_priv, pp_ctrl_reg, pp);
-		intel_de_posting_read(dev_priv, pp_ctrl_reg);
-	}
-}
 
 /* Enable backlight PWM and backlight PP control. */
 void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state,
@@ -3433,31 +1945,7 @@ void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state,
 	drm_dbg_kms(&i915->drm, "\n");
 
 	intel_panel_enable_backlight(crtc_state, conn_state);
-	_intel_edp_backlight_on(intel_dp);
-}
-
-/* Disable backlight in the panel power control. */
-static void _intel_edp_backlight_off(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	intel_wakeref_t wakeref;
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	with_pps_lock(intel_dp, wakeref) {
-		i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-		u32 pp;
-
-		pp = ilk_get_pp_control(intel_dp);
-		pp &= ~EDP_BLC_ENABLE;
-
-		intel_de_write(dev_priv, pp_ctrl_reg, pp);
-		intel_de_posting_read(dev_priv, pp_ctrl_reg);
-	}
-
-	intel_dp->last_backlight_off = jiffies;
-	edp_wait_backlight_off(intel_dp);
+	intel_pps_backlight_on(intel_dp);
 }
 
 /* Disable backlight PP control and backlight PWM. */
@@ -3471,37 +1959,10 @@ void intel_edp_backlight_off(const struct drm_connector_state *old_conn_state)
 
 	drm_dbg_kms(&i915->drm, "\n");
 
-	_intel_edp_backlight_off(intel_dp);
+	intel_pps_backlight_off(intel_dp);
 	intel_panel_disable_backlight(old_conn_state);
 }
 
-/*
- * Hook for controlling the panel power control backlight through the bl_power
- * sysfs attribute. Take care to handle multiple calls.
- */
-static void intel_edp_backlight_power(struct intel_connector *connector,
-				      bool enable)
-{
-	struct drm_i915_private *i915 = to_i915(connector->base.dev);
-	struct intel_dp *intel_dp = intel_attached_dp(connector);
-	intel_wakeref_t wakeref;
-	bool is_enabled;
-
-	is_enabled = false;
-	with_pps_lock(intel_dp, wakeref)
-		is_enabled = ilk_get_pp_control(intel_dp) & EDP_BLC_ENABLE;
-	if (is_enabled == enable)
-		return;
-
-	drm_dbg_kms(&i915->drm, "panel power control backlight %s\n",
-		    enable ? "enable" : "disable");
-
-	if (enable)
-		_intel_edp_backlight_on(intel_dp);
-	else
-		_intel_edp_backlight_off(intel_dp);
-}
-
 static void assert_dp_port(struct intel_dp *intel_dp, bool state)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
@@ -3922,10 +2383,10 @@ static void intel_disable_dp(struct intel_atomic_state *state,
 
 	/* Make sure the panel is off before trying to change the mode. But also
 	 * ensure that we have vdd while we switch off the panel. */
-	intel_edp_panel_vdd_on(intel_dp);
+	intel_pps_vdd_on(intel_dp);
 	intel_edp_backlight_off(old_conn_state);
 	intel_dp_set_power(intel_dp, DP_SET_POWER_D3);
-	intel_edp_panel_off(intel_dp);
+	intel_pps_off(intel_dp);
 	intel_dp->frl.is_trained = false;
 	intel_dp->frl.trained_rate_gbps = 0;
 }
@@ -4372,8 +2833,8 @@ void intel_dp_configure_protocol_converter(struct intel_dp *intel_dp,
 		drm_dbg_kms(&i915->drm, "Failed to set protocol converter HDMI mode to %s\n",
 			    enableddisabled(intel_dp->has_hdmi_sink));
 
-	tmp = intel_dp->dfp.ycbcr_444_to_420 ?
-		DP_CONVERSION_TO_YCBCR420_ENABLE : 0;
+	tmp = crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444 &&
+		intel_dp->dfp.ycbcr_444_to_420 ? DP_CONVERSION_TO_YCBCR420_ENABLE : 0;
 
 	if (drm_dp_dpcd_writeb(&intel_dp->aux,
 			       DP_PROTOCOL_CONVERTER_CONTROL_1, tmp) != 1)
@@ -4435,15 +2896,15 @@ static void intel_enable_dp(struct intel_atomic_state *state,
 	if (drm_WARN_ON(&dev_priv->drm, dp_reg & DP_PORT_EN))
 		return;
 
-	with_pps_lock(intel_dp, wakeref) {
+	with_intel_pps_lock(intel_dp, wakeref) {
 		if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-			vlv_init_panel_power_sequencer(encoder, pipe_config);
+			vlv_pps_init(encoder, pipe_config);
 
 		intel_dp_enable_port(intel_dp, pipe_config);
 
-		edp_panel_vdd_on(intel_dp);
-		edp_panel_on(intel_dp);
-		edp_panel_vdd_off(intel_dp, true);
+		intel_pps_vdd_on_unlocked(intel_dp);
+		intel_pps_on_unlocked(intel_dp);
+		intel_pps_vdd_off_unlocked(intel_dp, true);
 	}
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
@@ -4502,112 +2963,6 @@ static void g4x_pre_enable_dp(struct intel_atomic_state *state,
 		ilk_edp_pll_on(intel_dp, pipe_config);
 }
 
-static void vlv_detach_power_sequencer(struct intel_dp *intel_dp)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
-	enum pipe pipe = intel_dp->pps_pipe;
-	i915_reg_t pp_on_reg = PP_ON_DELAYS(pipe);
-
-	drm_WARN_ON(&dev_priv->drm, intel_dp->active_pipe != INVALID_PIPE);
-
-	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
-		return;
-
-	edp_panel_vdd_off_sync(intel_dp);
-
-	/*
-	 * VLV seems to get confused when multiple power sequencers
-	 * have the same port selected (even if only one has power/vdd
-	 * enabled). The failure manifests as vlv_wait_port_ready() failing
-	 * CHV on the other hand doesn't seem to mind having the same port
-	 * selected in multiple power sequencers, but let's clear the
-	 * port select always when logically disconnecting a power sequencer
-	 * from a port.
-	 */
-	drm_dbg_kms(&dev_priv->drm,
-		    "detaching pipe %c power sequencer from [ENCODER:%d:%s]\n",
-		    pipe_name(pipe), dig_port->base.base.base.id,
-		    dig_port->base.base.name);
-	intel_de_write(dev_priv, pp_on_reg, 0);
-	intel_de_posting_read(dev_priv, pp_on_reg);
-
-	intel_dp->pps_pipe = INVALID_PIPE;
-}
-
-static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv,
-				      enum pipe pipe)
-{
-	struct intel_encoder *encoder;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	for_each_intel_dp(&dev_priv->drm, encoder) {
-		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-
-		drm_WARN(&dev_priv->drm, intel_dp->active_pipe == pipe,
-			 "stealing pipe %c power sequencer from active [ENCODER:%d:%s]\n",
-			 pipe_name(pipe), encoder->base.base.id,
-			 encoder->base.name);
-
-		if (intel_dp->pps_pipe != pipe)
-			continue;
-
-		drm_dbg_kms(&dev_priv->drm,
-			    "stealing pipe %c power sequencer from [ENCODER:%d:%s]\n",
-			    pipe_name(pipe), encoder->base.base.id,
-			    encoder->base.name);
-
-		/* make sure vdd is off before we steal it */
-		vlv_detach_power_sequencer(intel_dp);
-	}
-}
-
-static void vlv_init_panel_power_sequencer(struct intel_encoder *encoder,
-					   const struct intel_crtc_state *crtc_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
-	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	drm_WARN_ON(&dev_priv->drm, intel_dp->active_pipe != INVALID_PIPE);
-
-	if (intel_dp->pps_pipe != INVALID_PIPE &&
-	    intel_dp->pps_pipe != crtc->pipe) {
-		/*
-		 * If another power sequencer was being used on this
-		 * port previously make sure to turn off vdd there while
-		 * we still have control of it.
-		 */
-		vlv_detach_power_sequencer(intel_dp);
-	}
-
-	/*
-	 * We may be stealing the power
-	 * sequencer from another port.
-	 */
-	vlv_steal_power_sequencer(dev_priv, crtc->pipe);
-
-	intel_dp->active_pipe = crtc->pipe;
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	/* now it's all ours */
-	intel_dp->pps_pipe = crtc->pipe;
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "initializing pipe %c power sequencer for [ENCODER:%d:%s]\n",
-		    pipe_name(intel_dp->pps_pipe), encoder->base.base.id,
-		    encoder->base.name);
-
-	/* init power sequencer on this pipe and port */
-	intel_dp_init_panel_power_sequencer(intel_dp);
-	intel_dp_init_panel_power_sequencer_registers(intel_dp, true);
-}
-
 static void vlv_pre_enable_dp(struct intel_atomic_state *state,
 			      struct intel_encoder *encoder,
 			      const struct intel_crtc_state *pipe_config,
@@ -5007,22 +3362,19 @@ ivb_cpu_edp_set_signal_levels(struct intel_dp *intel_dp,
 	intel_de_posting_read(dev_priv, intel_dp->output_reg);
 }
 
-void intel_dp_set_signal_levels(struct intel_dp *intel_dp,
-				const struct intel_crtc_state *crtc_state)
+static char dp_training_pattern_name(u8 train_pat)
 {
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	u8 train_set = intel_dp->train_set[0];
-
-	drm_dbg_kms(&dev_priv->drm, "Using vswing level %d%s\n",
-		    train_set & DP_TRAIN_VOLTAGE_SWING_MASK,
-		    train_set & DP_TRAIN_MAX_SWING_REACHED ? " (max)" : "");
-	drm_dbg_kms(&dev_priv->drm, "Using pre-emphasis level %d%s\n",
-		    (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) >>
-		    DP_TRAIN_PRE_EMPHASIS_SHIFT,
-		    train_set & DP_TRAIN_MAX_PRE_EMPHASIS_REACHED ?
-		    " (max)" : "");
-
-	intel_dp->set_signal_levels(intel_dp, crtc_state);
+	switch (train_pat) {
+	case DP_TRAINING_PATTERN_1:
+	case DP_TRAINING_PATTERN_2:
+	case DP_TRAINING_PATTERN_3:
+		return '0' + train_pat;
+	case DP_TRAINING_PATTERN_4:
+		return '4';
+	default:
+		MISSING_CASE(train_pat);
+		return '?';
+	}
 }
 
 void
@@ -5030,13 +3382,15 @@ intel_dp_program_link_training_pattern(struct intel_dp *intel_dp,
 				       const struct intel_crtc_state *crtc_state,
 				       u8 dp_train_pat)
 {
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u8 train_pat = intel_dp_training_pattern_symbol(dp_train_pat);
 
-	if ((intel_dp_training_pattern_symbol(dp_train_pat)) !=
-	    DP_TRAINING_PATTERN_DISABLE)
+	if (train_pat != DP_TRAINING_PATTERN_DISABLE)
 		drm_dbg_kms(&dev_priv->drm,
-			    "Using DP training pattern TPS%d\n",
-			    intel_dp_training_pattern_symbol(dp_train_pat));
+			    "[ENCODER:%d:%s] Using DP training pattern TPS%c\n",
+			    encoder->base.base.id, encoder->base.name,
+			    dp_training_pattern_name(train_pat));
 
 	intel_dp->set_link_train(intel_dp, crtc_state, dp_train_pat);
 }
@@ -5102,15 +3456,15 @@ intel_dp_link_down(struct intel_encoder *encoder,
 		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 	}
 
-	msleep(intel_dp->panel_power_down_delay);
+	msleep(intel_dp->pps.panel_power_down_delay);
 
 	intel_dp->DP = DP;
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
 		intel_wakeref_t wakeref;
 
-		with_pps_lock(intel_dp, wakeref)
-			intel_dp->active_pipe = INVALID_PIPE;
+		with_intel_pps_lock(intel_dp, wakeref)
+			intel_dp->pps.active_pipe = INVALID_PIPE;
 	}
 }
 
@@ -6079,7 +4433,7 @@ static void intel_dp_process_phy_request(struct intel_dp *intel_dp,
 
 	intel_dp_autotest_phy_ddi_disable(intel_dp, crtc_state);
 
-	intel_dp_set_signal_levels(intel_dp, crtc_state);
+	intel_dp_set_signal_levels(intel_dp, crtc_state, DP_PHY_DPRX);
 
 	intel_dp_phy_pattern_update(intel_dp, crtc_state);
 
@@ -6152,6 +4506,17 @@ update_status:
 			    "Could not write test response to sink\n");
 }
 
+static void
+intel_dp_mst_hpd_irq(struct intel_dp *intel_dp, u8 *esi, bool *handled)
+{
+		drm_dp_mst_hpd_irq(&intel_dp->mst_mgr, esi, handled);
+
+		if (esi[1] & DP_CP_IRQ) {
+			intel_hdcp_handle_cp_irq(intel_dp->attached_connector);
+			*handled = true;
+		}
+}
+
 /**
  * intel_dp_check_mst_status - service any pending MST interrupts, check link status
  * @intel_dp: Intel DP struct
@@ -6196,7 +4561,8 @@ intel_dp_check_mst_status(struct intel_dp *intel_dp)
 
 		drm_dbg_kms(&i915->drm, "got esi %3ph\n", esi);
 
-		drm_dp_mst_hpd_irq(&intel_dp->mst_mgr, esi, &handled);
+		intel_dp_mst_hpd_irq(intel_dp, esi, &handled);
+
 		if (!handled)
 			break;
 
@@ -6520,7 +4886,7 @@ static int intel_dp_do_phy_test(struct intel_encoder *encoder,
 	return 0;
 }
 
-static void intel_dp_phy_test(struct intel_encoder *encoder)
+void intel_dp_phy_test(struct intel_encoder *encoder)
 {
 	struct drm_modeset_acquire_ctx ctx;
 	int ret;
@@ -6955,8 +5321,8 @@ intel_dp_update_420(struct intel_dp *intel_dp)
 							intel_dp->downstream_ports);
 	rgb_to_ycbcr = drm_dp_downstream_rgb_to_ycbcr_conversion(intel_dp->dpcd,
 								 intel_dp->downstream_ports,
-								 DP_DS_HDMI_BT601_RGB_YCBCR_CONV ||
-								 DP_DS_HDMI_BT709_RGB_YCBCR_CONV ||
+								 DP_DS_HDMI_BT601_RGB_YCBCR_CONV |
+								 DP_DS_HDMI_BT709_RGB_YCBCR_CONV |
 								 DP_DS_HDMI_BT2020_RGB_YCBCR_CONV);
 
 	if (INTEL_GEN(i915) >= 11) {
@@ -7007,7 +5373,6 @@ intel_dp_set_edid(struct intel_dp *intel_dp)
 	}
 
 	drm_dp_cec_set_edid(&intel_dp->aux, edid);
-	intel_dp->edid_quirks = drm_dp_get_edid_quirks(edid);
 }
 
 static void
@@ -7021,7 +5386,6 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
 
 	intel_dp->has_hdmi_sink = false;
 	intel_dp->has_audio = false;
-	intel_dp->edid_quirks = 0;
 
 	intel_dp->dfp.max_bpc = 0;
 	intel_dp->dfp.max_dotclock = 0;
@@ -7188,6 +5552,10 @@ static int intel_dp_get_modes(struct drm_connector *connector)
 	edid = intel_connector->detect_edid;
 	if (edid) {
 		int ret = intel_connector_update_modes(connector, edid);
+
+		if (intel_vrr_is_capable(connector))
+			drm_connector_set_vrr_capable_property(connector,
+							       true);
 		if (ret)
 			return ret;
 	}
@@ -7276,17 +5644,8 @@ void intel_dp_encoder_flush_work(struct drm_encoder *encoder)
 	struct intel_dp *intel_dp = &dig_port->dp;
 
 	intel_dp_mst_encoder_cleanup(dig_port);
-	if (intel_dp_is_edp(intel_dp)) {
-		intel_wakeref_t wakeref;
 
-		cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-		/*
-		 * vdd might still be enabled do to the delayed vdd off.
-		 * Make sure vdd is actually turned off here.
-		 */
-		with_pps_lock(intel_dp, wakeref)
-			edp_panel_vdd_off_sync(intel_dp);
-	}
+	intel_pps_vdd_off_sync(intel_dp);
 
 	intel_dp_aux_fini(intel_dp);
 }
@@ -7302,55 +5661,15 @@ static void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(intel_encoder);
-	intel_wakeref_t wakeref;
 
-	if (!intel_dp_is_edp(intel_dp))
-		return;
-
-	/*
-	 * vdd might still be enabled do to the delayed vdd off.
-	 * Make sure vdd is actually turned off here.
-	 */
-	cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-	with_pps_lock(intel_dp, wakeref)
-		edp_panel_vdd_off_sync(intel_dp);
+	intel_pps_vdd_off_sync(intel_dp);
 }
 
 void intel_dp_encoder_shutdown(struct intel_encoder *intel_encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(intel_encoder);
-	intel_wakeref_t wakeref;
-
-	if (!intel_dp_is_edp(intel_dp))
-		return;
 
-	with_pps_lock(intel_dp, wakeref)
-		wait_panel_power_cycle(intel_dp);
-}
-
-static void intel_edp_panel_vdd_sanitize(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	if (!edp_have_panel_vdd(intel_dp))
-		return;
-
-	/*
-	 * The VDD bit needs a power domain reference, so if the bit is
-	 * already enabled when we boot or resume, grab this reference and
-	 * schedule a vdd off, so we don't hold on to the reference
-	 * indefinitely.
-	 */
-	drm_dbg_kms(&dev_priv->drm,
-		    "VDD left on by BIOS, adjusting state tracking\n");
-	drm_WARN_ON(&dev_priv->drm, intel_dp->vdd_wakeref);
-	intel_dp->vdd_wakeref = intel_display_power_get(dev_priv,
-							intel_aux_power_domain(dig_port));
-
-	edp_panel_vdd_schedule_off(intel_dp);
+	intel_pps_wait_power_cycle(intel_dp);
 }
 
 static enum pipe vlv_active_pipe(struct intel_dp *intel_dp)
@@ -7370,30 +5689,20 @@ void intel_dp_encoder_reset(struct drm_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_dp *intel_dp = enc_to_intel_dp(to_intel_encoder(encoder));
-	intel_wakeref_t wakeref;
 
 	if (!HAS_DDI(dev_priv))
 		intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg);
 
 	intel_dp->reset_link_params = true;
 
-	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv) &&
-	    !intel_dp_is_edp(intel_dp))
-		return;
-
-	with_pps_lock(intel_dp, wakeref) {
-		if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-			intel_dp->active_pipe = vlv_active_pipe(intel_dp);
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		intel_wakeref_t wakeref;
 
-		if (intel_dp_is_edp(intel_dp)) {
-			/*
-			 * Reinit the power sequencer, in case BIOS did
-			 * something nasty with it.
-			 */
-			intel_dp_pps_init(intel_dp);
-			intel_edp_panel_vdd_sanitize(intel_dp);
-		}
+		with_intel_pps_lock(intel_dp, wakeref)
+			intel_dp->pps.active_pipe = vlv_active_pipe(intel_dp);
 	}
+
+	intel_pps_encoder_reset(intel_dp);
 }
 
 static int intel_modeset_tile_group(struct intel_atomic_state *state,
@@ -7558,19 +5867,6 @@ static const struct drm_encoder_funcs intel_dp_enc_funcs = {
 	.destroy = intel_dp_encoder_destroy,
 };
 
-static bool intel_edp_have_power(struct intel_dp *intel_dp)
-{
-	intel_wakeref_t wakeref;
-	bool have_power = false;
-
-	with_pps_lock(intel_dp, wakeref) {
-		have_power = edp_have_panel_power(intel_dp) &&
-						  edp_have_panel_vdd(intel_dp);
-	}
-
-	return have_power;
-}
-
 enum irqreturn
 intel_dp_hpd_pulse(struct intel_digital_port *dig_port, bool long_hpd)
 {
@@ -7578,7 +5874,7 @@ intel_dp_hpd_pulse(struct intel_digital_port *dig_port, bool long_hpd)
 	struct intel_dp *intel_dp = &dig_port->dp;
 
 	if (dig_port->base.type == INTEL_OUTPUT_EDP &&
-	    (long_hpd || !intel_edp_have_power(intel_dp))) {
+	    (long_hpd || !intel_pps_have_power(intel_dp))) {
 		/*
 		 * vdd off can generate a long/short pulse on eDP which
 		 * would require vdd on to handle it, and thus we
@@ -7672,277 +5968,9 @@ intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connect
 		connector->state->scaling_mode = DRM_MODE_SCALE_ASPECT;
 
 	}
-}
-
-static void intel_dp_init_panel_power_timestamps(struct intel_dp *intel_dp)
-{
-	intel_dp->panel_power_off_time = ktime_get_boottime();
-	intel_dp->last_power_on = jiffies;
-	intel_dp->last_backlight_off = jiffies;
-}
-
-static void
-intel_pps_readout_hw_state(struct intel_dp *intel_dp, struct edp_power_seq *seq)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	u32 pp_on, pp_off, pp_ctl;
-	struct pps_registers regs;
-
-	intel_pps_get_registers(intel_dp, &regs);
-
-	pp_ctl = ilk_get_pp_control(intel_dp);
-
-	/* Ensure PPS is unlocked */
-	if (!HAS_DDI(dev_priv))
-		intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl);
-
-	pp_on = intel_de_read(dev_priv, regs.pp_on);
-	pp_off = intel_de_read(dev_priv, regs.pp_off);
-
-	/* Pull timing values out of registers */
-	seq->t1_t3 = REG_FIELD_GET(PANEL_POWER_UP_DELAY_MASK, pp_on);
-	seq->t8 = REG_FIELD_GET(PANEL_LIGHT_ON_DELAY_MASK, pp_on);
-	seq->t9 = REG_FIELD_GET(PANEL_LIGHT_OFF_DELAY_MASK, pp_off);
-	seq->t10 = REG_FIELD_GET(PANEL_POWER_DOWN_DELAY_MASK, pp_off);
-
-	if (i915_mmio_reg_valid(regs.pp_div)) {
-		u32 pp_div;
-
-		pp_div = intel_de_read(dev_priv, regs.pp_div);
-
-		seq->t11_t12 = REG_FIELD_GET(PANEL_POWER_CYCLE_DELAY_MASK, pp_div) * 1000;
-	} else {
-		seq->t11_t12 = REG_FIELD_GET(BXT_POWER_CYCLE_DELAY_MASK, pp_ctl) * 1000;
-	}
-}
-
-static void
-intel_pps_dump_state(const char *state_name, const struct edp_power_seq *seq)
-{
-	DRM_DEBUG_KMS("%s t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
-		      state_name,
-		      seq->t1_t3, seq->t8, seq->t9, seq->t10, seq->t11_t12);
-}
-
-static void
-intel_pps_verify_state(struct intel_dp *intel_dp)
-{
-	struct edp_power_seq hw;
-	struct edp_power_seq *sw = &intel_dp->pps_delays;
 
-	intel_pps_readout_hw_state(intel_dp, &hw);
-
-	if (hw.t1_t3 != sw->t1_t3 || hw.t8 != sw->t8 || hw.t9 != sw->t9 ||
-	    hw.t10 != sw->t10 || hw.t11_t12 != sw->t11_t12) {
-		DRM_ERROR("PPS state mismatch\n");
-		intel_pps_dump_state("sw", sw);
-		intel_pps_dump_state("hw", &hw);
-	}
-}
-
-static void
-intel_dp_init_panel_power_sequencer(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	struct edp_power_seq cur, vbt, spec,
-		*final = &intel_dp->pps_delays;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	/* already initialized? */
-	if (final->t11_t12 != 0)
-		return;
-
-	intel_pps_readout_hw_state(intel_dp, &cur);
-
-	intel_pps_dump_state("cur", &cur);
-
-	vbt = dev_priv->vbt.edp.pps;
-	/* On Toshiba Satellite P50-C-18C system the VBT T12 delay
-	 * of 500ms appears to be too short. Ocassionally the panel
-	 * just fails to power back on. Increasing the delay to 800ms
-	 * seems sufficient to avoid this problem.
-	 */
-	if (dev_priv->quirks & QUIRK_INCREASE_T12_DELAY) {
-		vbt.t11_t12 = max_t(u16, vbt.t11_t12, 1300 * 10);
-		drm_dbg_kms(&dev_priv->drm,
-			    "Increasing T12 panel delay as per the quirk to %d\n",
-			    vbt.t11_t12);
-	}
-	/* T11_T12 delay is special and actually in units of 100ms, but zero
-	 * based in the hw (so we need to add 100 ms). But the sw vbt
-	 * table multiplies it with 1000 to make it in units of 100usec,
-	 * too. */
-	vbt.t11_t12 += 100 * 10;
-
-	/* Upper limits from eDP 1.3 spec. Note that we use the clunky units of
-	 * our hw here, which are all in 100usec. */
-	spec.t1_t3 = 210 * 10;
-	spec.t8 = 50 * 10; /* no limit for t8, use t7 instead */
-	spec.t9 = 50 * 10; /* no limit for t9, make it symmetric with t8 */
-	spec.t10 = 500 * 10;
-	/* This one is special and actually in units of 100ms, but zero
-	 * based in the hw (so we need to add 100 ms). But the sw vbt
-	 * table multiplies it with 1000 to make it in units of 100usec,
-	 * too. */
-	spec.t11_t12 = (510 + 100) * 10;
-
-	intel_pps_dump_state("vbt", &vbt);
-
-	/* Use the max of the register settings and vbt. If both are
-	 * unset, fall back to the spec limits. */
-#define assign_final(field)	final->field = (max(cur.field, vbt.field) == 0 ? \
-				       spec.field : \
-				       max(cur.field, vbt.field))
-	assign_final(t1_t3);
-	assign_final(t8);
-	assign_final(t9);
-	assign_final(t10);
-	assign_final(t11_t12);
-#undef assign_final
-
-#define get_delay(field)	(DIV_ROUND_UP(final->field, 10))
-	intel_dp->panel_power_up_delay = get_delay(t1_t3);
-	intel_dp->backlight_on_delay = get_delay(t8);
-	intel_dp->backlight_off_delay = get_delay(t9);
-	intel_dp->panel_power_down_delay = get_delay(t10);
-	intel_dp->panel_power_cycle_delay = get_delay(t11_t12);
-#undef get_delay
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "panel power up delay %d, power down delay %d, power cycle delay %d\n",
-		    intel_dp->panel_power_up_delay,
-		    intel_dp->panel_power_down_delay,
-		    intel_dp->panel_power_cycle_delay);
-
-	drm_dbg_kms(&dev_priv->drm, "backlight on delay %d, off delay %d\n",
-		    intel_dp->backlight_on_delay,
-		    intel_dp->backlight_off_delay);
-
-	/*
-	 * We override the HW backlight delays to 1 because we do manual waits
-	 * on them. For T8, even BSpec recommends doing it. For T9, if we
-	 * don't do this, we'll end up waiting for the backlight off delay
-	 * twice: once when we do the manual sleep, and once when we disable
-	 * the panel and wait for the PP_STATUS bit to become zero.
-	 */
-	final->t8 = 1;
-	final->t9 = 1;
-
-	/*
-	 * HW has only a 100msec granularity for t11_t12 so round it up
-	 * accordingly.
-	 */
-	final->t11_t12 = roundup(final->t11_t12, 100 * 10);
-}
-
-static void
-intel_dp_init_panel_power_sequencer_registers(struct intel_dp *intel_dp,
-					      bool force_disable_vdd)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-	u32 pp_on, pp_off, port_sel = 0;
-	int div = RUNTIME_INFO(dev_priv)->rawclk_freq / 1000;
-	struct pps_registers regs;
-	enum port port = dp_to_dig_port(intel_dp)->base.port;
-	const struct edp_power_seq *seq = &intel_dp->pps_delays;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	intel_pps_get_registers(intel_dp, &regs);
-
-	/*
-	 * On some VLV machines the BIOS can leave the VDD
-	 * enabled even on power sequencers which aren't
-	 * hooked up to any port. This would mess up the
-	 * power domain tracking the first time we pick
-	 * one of these power sequencers for use since
-	 * edp_panel_vdd_on() would notice that the VDD was
-	 * already on and therefore wouldn't grab the power
-	 * domain reference. Disable VDD first to avoid this.
-	 * This also avoids spuriously turning the VDD on as
-	 * soon as the new power sequencer gets initialized.
-	 */
-	if (force_disable_vdd) {
-		u32 pp = ilk_get_pp_control(intel_dp);
-
-		drm_WARN(&dev_priv->drm, pp & PANEL_POWER_ON,
-			 "Panel power already on\n");
-
-		if (pp & EDP_FORCE_VDD)
-			drm_dbg_kms(&dev_priv->drm,
-				    "VDD already on, disabling first\n");
-
-		pp &= ~EDP_FORCE_VDD;
-
-		intel_de_write(dev_priv, regs.pp_ctrl, pp);
-	}
-
-	pp_on = REG_FIELD_PREP(PANEL_POWER_UP_DELAY_MASK, seq->t1_t3) |
-		REG_FIELD_PREP(PANEL_LIGHT_ON_DELAY_MASK, seq->t8);
-	pp_off = REG_FIELD_PREP(PANEL_LIGHT_OFF_DELAY_MASK, seq->t9) |
-		REG_FIELD_PREP(PANEL_POWER_DOWN_DELAY_MASK, seq->t10);
-
-	/* Haswell doesn't have any port selection bits for the panel
-	 * power sequencer any more. */
-	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
-		port_sel = PANEL_PORT_SELECT_VLV(port);
-	} else if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) {
-		switch (port) {
-		case PORT_A:
-			port_sel = PANEL_PORT_SELECT_DPA;
-			break;
-		case PORT_C:
-			port_sel = PANEL_PORT_SELECT_DPC;
-			break;
-		case PORT_D:
-			port_sel = PANEL_PORT_SELECT_DPD;
-			break;
-		default:
-			MISSING_CASE(port);
-			break;
-		}
-	}
-
-	pp_on |= port_sel;
-
-	intel_de_write(dev_priv, regs.pp_on, pp_on);
-	intel_de_write(dev_priv, regs.pp_off, pp_off);
-
-	/*
-	 * Compute the divisor for the pp clock, simply match the Bspec formula.
-	 */
-	if (i915_mmio_reg_valid(regs.pp_div)) {
-		intel_de_write(dev_priv, regs.pp_div,
-			       REG_FIELD_PREP(PP_REFERENCE_DIVIDER_MASK, (100 * div) / 2 - 1) | REG_FIELD_PREP(PANEL_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000)));
-	} else {
-		u32 pp_ctl;
-
-		pp_ctl = intel_de_read(dev_priv, regs.pp_ctrl);
-		pp_ctl &= ~BXT_POWER_CYCLE_DELAY_MASK;
-		pp_ctl |= REG_FIELD_PREP(BXT_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000));
-		intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl);
-	}
-
-	drm_dbg_kms(&dev_priv->drm,
-		    "panel power sequencer register settings: PP_ON %#x, PP_OFF %#x, PP_DIV %#x\n",
-		    intel_de_read(dev_priv, regs.pp_on),
-		    intel_de_read(dev_priv, regs.pp_off),
-		    i915_mmio_reg_valid(regs.pp_div) ?
-		    intel_de_read(dev_priv, regs.pp_div) :
-		    (intel_de_read(dev_priv, regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK));
-}
-
-static void intel_dp_pps_init(struct intel_dp *intel_dp)
-{
-	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
-
-	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
-		vlv_initial_power_sequencer_setup(intel_dp);
-	} else {
-		intel_dp_init_panel_power_sequencer(intel_dp);
-		intel_dp_init_panel_power_sequencer_registers(intel_dp, false);
-	}
+	if (HAS_VRR(dev_priv))
+		drm_connector_attach_vrr_capable_property(connector);
 }
 
 /**
@@ -8381,14 +6409,11 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	struct drm_display_mode *downclock_mode = NULL;
 	bool has_dpcd;
 	enum pipe pipe = INVALID_PIPE;
-	intel_wakeref_t wakeref;
 	struct edid *edid;
 
 	if (!intel_dp_is_edp(intel_dp))
 		return true;
 
-	INIT_DELAYED_WORK(&intel_dp->panel_vdd_work, edp_panel_vdd_work);
-
 	/*
 	 * On IBX/CPT we may get here with LVDS already registered. Since the
 	 * driver uses the only internal power sequencer available for both
@@ -8404,11 +6429,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 		return false;
 	}
 
-	with_pps_lock(intel_dp, wakeref) {
-		intel_dp_init_panel_power_timestamps(intel_dp);
-		intel_dp_pps_init(intel_dp);
-		intel_edp_panel_vdd_sanitize(intel_dp);
-	}
+	intel_pps_init(intel_dp);
 
 	/* Cache DPCD and EDID for edp. */
 	has_dpcd = intel_edp_init_dpcd(intel_dp);
@@ -8425,7 +6446,6 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	if (edid) {
 		if (drm_add_edid_modes(connector, edid)) {
 			drm_connector_update_edid_property(connector, edid);
-			intel_dp->edid_quirks = drm_dp_get_edid_quirks(edid);
 		} else {
 			kfree(edid);
 			edid = ERR_PTR(-EINVAL);
@@ -8453,7 +6473,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 		pipe = vlv_active_pipe(intel_dp);
 
 		if (pipe != PIPE_A && pipe != PIPE_B)
-			pipe = intel_dp->pps_pipe;
+			pipe = intel_dp->pps.pps_pipe;
 
 		if (pipe != PIPE_A && pipe != PIPE_B)
 			pipe = PIPE_A;
@@ -8464,7 +6484,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	}
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, downclock_mode);
-	intel_connector->panel.backlight.power = intel_edp_backlight_power;
+	intel_connector->panel.backlight.power = intel_pps_backlight_power;
 	intel_panel_setup_backlight(connector, pipe);
 
 	if (fixed_mode) {
@@ -8476,13 +6496,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	return true;
 
 out_vdd_off:
-	cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-	/*
-	 * vdd might still be enabled do to the delayed vdd off.
-	 * Make sure vdd is actually turned off here.
-	 */
-	with_pps_lock(intel_dp, wakeref)
-		edp_panel_vdd_off_sync(intel_dp);
+	intel_pps_vdd_off_sync(intel_dp);
 
 	return false;
 }
@@ -8536,8 +6550,8 @@ intel_dp_init_connector(struct intel_digital_port *dig_port,
 	intel_dp_set_source_rates(intel_dp);
 
 	intel_dp->reset_link_params = true;
-	intel_dp->pps_pipe = INVALID_PIPE;
-	intel_dp->active_pipe = INVALID_PIPE;
+	intel_dp->pps.pps_pipe = INVALID_PIPE;
+	intel_dp->pps.active_pipe = INVALID_PIPE;
 
 	/* Preserve the current hw state. */
 	intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg);
@@ -8555,7 +6569,7 @@ intel_dp_init_connector(struct intel_digital_port *dig_port,
 	}
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		intel_dp->active_pipe = vlv_active_pipe(intel_dp);
+		intel_dp->pps.active_pipe = vlv_active_pipe(intel_dp);
 
 	/*
 	 * For eDP we always set the encoder type to INTEL_OUTPUT_EDP, but
diff --git a/drivers/gpu/drm/i915/display/intel_dp.h b/drivers/gpu/drm/i915/display/intel_dp.h
index 4280a09fd8fd..d80839139bfb 100644
--- a/drivers/gpu/drm/i915/display/intel_dp.h
+++ b/drivers/gpu/drm/i915/display/intel_dp.h
@@ -70,16 +70,11 @@ enum irqreturn intel_dp_hpd_pulse(struct intel_digital_port *dig_port,
 void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state,
 			    const struct drm_connector_state *conn_state);
 void intel_edp_backlight_off(const struct drm_connector_state *conn_state);
-void intel_edp_panel_vdd_on(struct intel_dp *intel_dp);
-void intel_edp_panel_on(struct intel_dp *intel_dp);
-void intel_edp_panel_off(struct intel_dp *intel_dp);
 void intel_dp_mst_suspend(struct drm_i915_private *dev_priv);
 void intel_dp_mst_resume(struct drm_i915_private *dev_priv);
 int intel_dp_max_link_rate(struct intel_dp *intel_dp);
 int intel_dp_max_lane_count(struct intel_dp *intel_dp);
 int intel_dp_rate_select(struct intel_dp *intel_dp, int rate);
-void intel_power_sequencer_reset(struct drm_i915_private *dev_priv);
-u32 intel_dp_pack_aux(const u8 *src, int src_bytes);
 
 void intel_edp_drrs_enable(struct intel_dp *intel_dp,
 			   const struct intel_crtc_state *crtc_state);
@@ -96,9 +91,6 @@ void
 intel_dp_program_link_training_pattern(struct intel_dp *intel_dp,
 				       const struct intel_crtc_state *crtc_state,
 				       u8 dp_train_pat);
-void
-intel_dp_set_signal_levels(struct intel_dp *intel_dp,
-			   const struct intel_crtc_state *crtc_state);
 void intel_dp_compute_rate(struct intel_dp *intel_dp, int port_clock,
 			   u8 *link_bw, u8 *rate_select);
 bool intel_dp_source_supports_hbr2(struct intel_dp *intel_dp);
@@ -144,9 +136,11 @@ bool intel_dp_initial_fastset_check(struct intel_encoder *encoder,
 				    struct intel_crtc_state *crtc_state);
 void intel_dp_sync_state(struct intel_encoder *encoder,
 			 const struct intel_crtc_state *crtc_state);
+const struct dpll *vlv_get_dpll(struct drm_i915_private *i915);
 
 void intel_dp_check_frl_training(struct intel_dp *intel_dp);
 void intel_dp_pcon_dsc_configure(struct intel_dp *intel_dp,
 				 const struct intel_crtc_state *crtc_state);
+void intel_dp_phy_test(struct intel_encoder *encoder);
 
 #endif /* __INTEL_DP_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c
new file mode 100644
index 000000000000..eaebf123310a
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c
@@ -0,0 +1,692 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020-2021 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "intel_display_types.h"
+#include "intel_dp_aux.h"
+#include "intel_pps.h"
+#include "intel_tc.h"
+
+u32 intel_dp_pack_aux(const u8 *src, int src_bytes)
+{
+	int i;
+	u32 v = 0;
+
+	if (src_bytes > 4)
+		src_bytes = 4;
+	for (i = 0; i < src_bytes; i++)
+		v |= ((u32)src[i]) << ((3 - i) * 8);
+	return v;
+}
+
+static void intel_dp_unpack_aux(u32 src, u8 *dst, int dst_bytes)
+{
+	int i;
+
+	if (dst_bytes > 4)
+		dst_bytes = 4;
+	for (i = 0; i < dst_bytes; i++)
+		dst[i] = src >> ((3 - i) * 8);
+}
+
+static u32
+intel_dp_aux_wait_done(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp);
+	const unsigned int timeout_ms = 10;
+	u32 status;
+	bool done;
+
+#define C (((status = intel_uncore_read_notrace(&i915->uncore, ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
+	done = wait_event_timeout(i915->gmbus_wait_queue, C,
+				  msecs_to_jiffies_timeout(timeout_ms));
+
+	/* just trace the final value */
+	trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true);
+
+	if (!done)
+		drm_err(&i915->drm,
+			"%s: did not complete or timeout within %ums (status 0x%08x)\n",
+			intel_dp->aux.name, timeout_ms, status);
+#undef C
+
+	return status;
+}
+
+static u32 g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (index)
+		return 0;
+
+	/*
+	 * The clock divider is based off the hrawclk, and would like to run at
+	 * 2MHz.  So, take the hrawclk value and divide by 2000 and use that
+	 */
+	return DIV_ROUND_CLOSEST(RUNTIME_INFO(dev_priv)->rawclk_freq, 2000);
+}
+
+static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u32 freq;
+
+	if (index)
+		return 0;
+
+	/*
+	 * The clock divider is based off the cdclk or PCH rawclk, and would
+	 * like to run at 2MHz.  So, take the cdclk or PCH rawclk value and
+	 * divide by 2000 and use that
+	 */
+	if (dig_port->aux_ch == AUX_CH_A)
+		freq = dev_priv->cdclk.hw.cdclk;
+	else
+		freq = RUNTIME_INFO(dev_priv)->rawclk_freq;
+	return DIV_ROUND_CLOSEST(freq, 2000);
+}
+
+static u32 hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+
+	if (dig_port->aux_ch != AUX_CH_A && HAS_PCH_LPT_H(dev_priv)) {
+		/* Workaround for non-ULT HSW */
+		switch (index) {
+		case 0: return 63;
+		case 1: return 72;
+		default: return 0;
+		}
+	}
+
+	return ilk_get_aux_clock_divider(intel_dp, index);
+}
+
+static u32 skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	/*
+	 * SKL doesn't need us to program the AUX clock divider (Hardware will
+	 * derive the clock from CDCLK automatically). We still implement the
+	 * get_aux_clock_divider vfunc to plug-in into the existing code.
+	 */
+	return index ? 0 : 1;
+}
+
+static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp,
+				int send_bytes,
+				u32 aux_clock_divider)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv =
+			to_i915(dig_port->base.base.dev);
+	u32 precharge, timeout;
+
+	if (IS_GEN(dev_priv, 6))
+		precharge = 3;
+	else
+		precharge = 5;
+
+	if (IS_BROADWELL(dev_priv))
+		timeout = DP_AUX_CH_CTL_TIME_OUT_600us;
+	else
+		timeout = DP_AUX_CH_CTL_TIME_OUT_400us;
+
+	return DP_AUX_CH_CTL_SEND_BUSY |
+	       DP_AUX_CH_CTL_DONE |
+	       DP_AUX_CH_CTL_INTERRUPT |
+	       DP_AUX_CH_CTL_TIME_OUT_ERROR |
+	       timeout |
+	       DP_AUX_CH_CTL_RECEIVE_ERROR |
+	       (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
+	       (precharge << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
+	       (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT);
+}
+
+static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp,
+				int send_bytes,
+				u32 unused)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *i915 =
+			to_i915(dig_port->base.base.dev);
+	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
+	u32 ret;
+
+	ret = DP_AUX_CH_CTL_SEND_BUSY |
+	      DP_AUX_CH_CTL_DONE |
+	      DP_AUX_CH_CTL_INTERRUPT |
+	      DP_AUX_CH_CTL_TIME_OUT_ERROR |
+	      DP_AUX_CH_CTL_TIME_OUT_MAX |
+	      DP_AUX_CH_CTL_RECEIVE_ERROR |
+	      (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
+	      DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(32) |
+	      DP_AUX_CH_CTL_SYNC_PULSE_SKL(32);
+
+	if (intel_phy_is_tc(i915, phy) &&
+	    dig_port->tc_mode == TC_PORT_TBT_ALT)
+		ret |= DP_AUX_CH_CTL_TBT_IO;
+
+	return ret;
+}
+
+static int
+intel_dp_aux_xfer(struct intel_dp *intel_dp,
+		  const u8 *send, int send_bytes,
+		  u8 *recv, int recv_size,
+		  u32 aux_send_ctl_flags)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *i915 =
+			to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
+	bool is_tc_port = intel_phy_is_tc(i915, phy);
+	i915_reg_t ch_ctl, ch_data[5];
+	u32 aux_clock_divider;
+	enum intel_display_power_domain aux_domain;
+	intel_wakeref_t aux_wakeref;
+	intel_wakeref_t pps_wakeref;
+	int i, ret, recv_bytes;
+	int try, clock = 0;
+	u32 status;
+	bool vdd;
+
+	ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp);
+	for (i = 0; i < ARRAY_SIZE(ch_data); i++)
+		ch_data[i] = intel_dp->aux_ch_data_reg(intel_dp, i);
+
+	if (is_tc_port)
+		intel_tc_port_lock(dig_port);
+
+	aux_domain = intel_aux_power_domain(dig_port);
+
+	aux_wakeref = intel_display_power_get(i915, aux_domain);
+	pps_wakeref = intel_pps_lock(intel_dp);
+
+	/*
+	 * We will be called with VDD already enabled for dpcd/edid/oui reads.
+	 * In such cases we want to leave VDD enabled and it's up to upper layers
+	 * to turn it off. But for eg. i2c-dev access we need to turn it on/off
+	 * ourselves.
+	 */
+	vdd = intel_pps_vdd_on_unlocked(intel_dp);
+
+	/*
+	 * dp aux is extremely sensitive to irq latency, hence request the
+	 * lowest possible wakeup latency and so prevent the cpu from going into
+	 * deep sleep states.
+	 */
+	cpu_latency_qos_update_request(&intel_dp->pm_qos, 0);
+
+	intel_pps_check_power_unlocked(intel_dp);
+
+	/* Try to wait for any previous AUX channel activity */
+	for (try = 0; try < 3; try++) {
+		status = intel_uncore_read_notrace(uncore, ch_ctl);
+		if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0)
+			break;
+		msleep(1);
+	}
+	/* just trace the final value */
+	trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true);
+
+	if (try == 3) {
+		const u32 status = intel_uncore_read(uncore, ch_ctl);
+
+		if (status != intel_dp->aux_busy_last_status) {
+			drm_WARN(&i915->drm, 1,
+				 "%s: not started (status 0x%08x)\n",
+				 intel_dp->aux.name, status);
+			intel_dp->aux_busy_last_status = status;
+		}
+
+		ret = -EBUSY;
+		goto out;
+	}
+
+	/* Only 5 data registers! */
+	if (drm_WARN_ON(&i915->drm, send_bytes > 20 || recv_size > 20)) {
+		ret = -E2BIG;
+		goto out;
+	}
+
+	while ((aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, clock++))) {
+		u32 send_ctl = intel_dp->get_aux_send_ctl(intel_dp,
+							  send_bytes,
+							  aux_clock_divider);
+
+		send_ctl |= aux_send_ctl_flags;
+
+		/* Must try at least 3 times according to DP spec */
+		for (try = 0; try < 5; try++) {
+			/* Load the send data into the aux channel data registers */
+			for (i = 0; i < send_bytes; i += 4)
+				intel_uncore_write(uncore,
+						   ch_data[i >> 2],
+						   intel_dp_pack_aux(send + i,
+								     send_bytes - i));
+
+			/* Send the command and wait for it to complete */
+			intel_uncore_write(uncore, ch_ctl, send_ctl);
+
+			status = intel_dp_aux_wait_done(intel_dp);
+
+			/* Clear done status and any errors */
+			intel_uncore_write(uncore,
+					   ch_ctl,
+					   status |
+					   DP_AUX_CH_CTL_DONE |
+					   DP_AUX_CH_CTL_TIME_OUT_ERROR |
+					   DP_AUX_CH_CTL_RECEIVE_ERROR);
+
+			/*
+			 * DP CTS 1.2 Core Rev 1.1, 4.2.1.1 & 4.2.1.2
+			 *   400us delay required for errors and timeouts
+			 *   Timeout errors from the HW already meet this
+			 *   requirement so skip to next iteration
+			 */
+			if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR)
+				continue;
+
+			if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
+				usleep_range(400, 500);
+				continue;
+			}
+			if (status & DP_AUX_CH_CTL_DONE)
+				goto done;
+		}
+	}
+
+	if ((status & DP_AUX_CH_CTL_DONE) == 0) {
+		drm_err(&i915->drm, "%s: not done (status 0x%08x)\n",
+			intel_dp->aux.name, status);
+		ret = -EBUSY;
+		goto out;
+	}
+
+done:
+	/*
+	 * Check for timeout or receive error. Timeouts occur when the sink is
+	 * not connected.
+	 */
+	if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
+		drm_err(&i915->drm, "%s: receive error (status 0x%08x)\n",
+			intel_dp->aux.name, status);
+		ret = -EIO;
+		goto out;
+	}
+
+	/*
+	 * Timeouts occur when the device isn't connected, so they're "normal"
+	 * -- don't fill the kernel log with these
+	 */
+	if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR) {
+		drm_dbg_kms(&i915->drm, "%s: timeout (status 0x%08x)\n",
+			    intel_dp->aux.name, status);
+		ret = -ETIMEDOUT;
+		goto out;
+	}
+
+	/* Unload any bytes sent back from the other side */
+	recv_bytes = ((status & DP_AUX_CH_CTL_MESSAGE_SIZE_MASK) >>
+		      DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT);
+
+	/*
+	 * By BSpec: "Message sizes of 0 or >20 are not allowed."
+	 * We have no idea of what happened so we return -EBUSY so
+	 * drm layer takes care for the necessary retries.
+	 */
+	if (recv_bytes == 0 || recv_bytes > 20) {
+		drm_dbg_kms(&i915->drm,
+			    "%s: Forbidden recv_bytes = %d on aux transaction\n",
+			    intel_dp->aux.name, recv_bytes);
+		ret = -EBUSY;
+		goto out;
+	}
+
+	if (recv_bytes > recv_size)
+		recv_bytes = recv_size;
+
+	for (i = 0; i < recv_bytes; i += 4)
+		intel_dp_unpack_aux(intel_uncore_read(uncore, ch_data[i >> 2]),
+				    recv + i, recv_bytes - i);
+
+	ret = recv_bytes;
+out:
+	cpu_latency_qos_update_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
+
+	if (vdd)
+		intel_pps_vdd_off_unlocked(intel_dp, false);
+
+	intel_pps_unlock(intel_dp, pps_wakeref);
+	intel_display_power_put_async(i915, aux_domain, aux_wakeref);
+
+	if (is_tc_port)
+		intel_tc_port_unlock(dig_port);
+
+	return ret;
+}
+
+#define BARE_ADDRESS_SIZE	3
+#define HEADER_SIZE		(BARE_ADDRESS_SIZE + 1)
+
+static void
+intel_dp_aux_header(u8 txbuf[HEADER_SIZE],
+		    const struct drm_dp_aux_msg *msg)
+{
+	txbuf[0] = (msg->request << 4) | ((msg->address >> 16) & 0xf);
+	txbuf[1] = (msg->address >> 8) & 0xff;
+	txbuf[2] = msg->address & 0xff;
+	txbuf[3] = msg->size - 1;
+}
+
+static u32 intel_dp_aux_xfer_flags(const struct drm_dp_aux_msg *msg)
+{
+	/*
+	 * If we're trying to send the HDCP Aksv, we need to set a the Aksv
+	 * select bit to inform the hardware to send the Aksv after our header
+	 * since we can't access that data from software.
+	 */
+	if ((msg->request & ~DP_AUX_I2C_MOT) == DP_AUX_NATIVE_WRITE &&
+	    msg->address == DP_AUX_HDCP_AKSV)
+		return DP_AUX_CH_CTL_AUX_AKSV_SELECT;
+
+	return 0;
+}
+
+static ssize_t
+intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg)
+{
+	struct intel_dp *intel_dp = container_of(aux, struct intel_dp, aux);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 txbuf[20], rxbuf[20];
+	size_t txsize, rxsize;
+	u32 flags = intel_dp_aux_xfer_flags(msg);
+	int ret;
+
+	intel_dp_aux_header(txbuf, msg);
+
+	switch (msg->request & ~DP_AUX_I2C_MOT) {
+	case DP_AUX_NATIVE_WRITE:
+	case DP_AUX_I2C_WRITE:
+	case DP_AUX_I2C_WRITE_STATUS_UPDATE:
+		txsize = msg->size ? HEADER_SIZE + msg->size : BARE_ADDRESS_SIZE;
+		rxsize = 2; /* 0 or 1 data bytes */
+
+		if (drm_WARN_ON(&i915->drm, txsize > 20))
+			return -E2BIG;
+
+		drm_WARN_ON(&i915->drm, !msg->buffer != !msg->size);
+
+		if (msg->buffer)
+			memcpy(txbuf + HEADER_SIZE, msg->buffer, msg->size);
+
+		ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize,
+					rxbuf, rxsize, flags);
+		if (ret > 0) {
+			msg->reply = rxbuf[0] >> 4;
+
+			if (ret > 1) {
+				/* Number of bytes written in a short write. */
+				ret = clamp_t(int, rxbuf[1], 0, msg->size);
+			} else {
+				/* Return payload size. */
+				ret = msg->size;
+			}
+		}
+		break;
+
+	case DP_AUX_NATIVE_READ:
+	case DP_AUX_I2C_READ:
+		txsize = msg->size ? HEADER_SIZE : BARE_ADDRESS_SIZE;
+		rxsize = msg->size + 1;
+
+		if (drm_WARN_ON(&i915->drm, rxsize > 20))
+			return -E2BIG;
+
+		ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize,
+					rxbuf, rxsize, flags);
+		if (ret > 0) {
+			msg->reply = rxbuf[0] >> 4;
+			/*
+			 * Assume happy day, and copy the data. The caller is
+			 * expected to check msg->reply before touching it.
+			 *
+			 * Return payload size.
+			 */
+			ret--;
+			memcpy(msg->buffer, rxbuf + 1, ret);
+		}
+		break;
+
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static i915_reg_t g4x_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_B);
+	}
+}
+
+static i915_reg_t g4x_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_B, index);
+	}
+}
+
+static i915_reg_t ilk_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+		return DP_AUX_CH_CTL(aux_ch);
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return PCH_DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t ilk_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return PCH_DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+static i915_reg_t skl_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+	case AUX_CH_E:
+	case AUX_CH_F:
+		return DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t skl_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+	case AUX_CH_E:
+	case AUX_CH_F:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+static i915_reg_t tgl_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_USBC1:
+	case AUX_CH_USBC2:
+	case AUX_CH_USBC3:
+	case AUX_CH_USBC4:
+	case AUX_CH_USBC5:
+	case AUX_CH_USBC6:
+		return DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t tgl_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_USBC1:
+	case AUX_CH_USBC2:
+	case AUX_CH_USBC3:
+	case AUX_CH_USBC4:
+	case AUX_CH_USBC5:
+	case AUX_CH_USBC6:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+void intel_dp_aux_fini(struct intel_dp *intel_dp)
+{
+	if (cpu_latency_qos_request_active(&intel_dp->pm_qos))
+		cpu_latency_qos_remove_request(&intel_dp->pm_qos);
+
+	kfree(intel_dp->aux.name);
+}
+
+void intel_dp_aux_init(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &dig_port->base;
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	if (INTEL_GEN(dev_priv) >= 12) {
+		intel_dp->aux_ch_ctl_reg = tgl_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = tgl_aux_data_reg;
+	} else if (INTEL_GEN(dev_priv) >= 9) {
+		intel_dp->aux_ch_ctl_reg = skl_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = skl_aux_data_reg;
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		intel_dp->aux_ch_ctl_reg = ilk_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = ilk_aux_data_reg;
+	} else {
+		intel_dp->aux_ch_ctl_reg = g4x_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = g4x_aux_data_reg;
+	}
+
+	if (INTEL_GEN(dev_priv) >= 9)
+		intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider;
+	else if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider;
+	else if (HAS_PCH_SPLIT(dev_priv))
+		intel_dp->get_aux_clock_divider = ilk_get_aux_clock_divider;
+	else
+		intel_dp->get_aux_clock_divider = g4x_get_aux_clock_divider;
+
+	if (INTEL_GEN(dev_priv) >= 9)
+		intel_dp->get_aux_send_ctl = skl_get_aux_send_ctl;
+	else
+		intel_dp->get_aux_send_ctl = g4x_get_aux_send_ctl;
+
+	drm_dp_aux_init(&intel_dp->aux);
+
+	/* Failure to allocate our preferred name is not critical */
+	if (INTEL_GEN(dev_priv) >= 12 && aux_ch >= AUX_CH_USBC1)
+		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX USBC%c/%s",
+					       aux_ch - AUX_CH_USBC1 + '1',
+					       encoder->base.name);
+	else
+		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX %c/%s",
+					       aux_ch_name(aux_ch),
+					       encoder->base.name);
+
+	intel_dp->aux.transfer = intel_dp_aux_transfer;
+	cpu_latency_qos_add_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.h b/drivers/gpu/drm/i915/display/intel_dp_aux.h
new file mode 100644
index 000000000000..4afbe76217b9
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux.h
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020-2021 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_AUX_H__
+#define __INTEL_DP_AUX_H__
+
+#include <linux/types.h>
+
+struct intel_dp;
+
+u32 intel_dp_pack_aux(const u8 *src, int src_bytes);
+
+void intel_dp_aux_fini(struct intel_dp *intel_dp);
+void intel_dp_aux_init(struct intel_dp *intel_dp);
+
+#endif /* __INTEL_DP_AUX_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c
index 9775f33d1aac..651884390137 100644
--- a/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c
@@ -22,8 +22,26 @@
  *
  */
 
+/*
+ * Laptops with Intel GPUs which have panels that support controlling the
+ * backlight through DP AUX can actually use two different interfaces: Intel's
+ * proprietary DP AUX backlight interface, and the standard VESA backlight
+ * interface. Unfortunately, at the time of writing this a lot of laptops will
+ * advertise support for the standard VESA backlight interface when they
+ * don't properly support it. However, on these systems the Intel backlight
+ * interface generally does work properly. Additionally, these systems will
+ * usually just indicate that they use PWM backlight controls in their VBIOS
+ * for some reason.
+ */
+
 #include "intel_display_types.h"
 #include "intel_dp_aux_backlight.h"
+#include "intel_panel.h"
+
+/* TODO:
+ * Implement HDR, right now we just implement the bare minimum to bring us back into SDR mode so we
+ * can make people's backlights work in the mean time
+ */
 
 /*
  * DP AUX registers for Intel's proprietary HDR backlight interface. We define
@@ -77,6 +95,178 @@
 
 #define INTEL_EDP_BRIGHTNESS_OPTIMIZATION_1                            0x359
 
+/* Intel EDP backlight callbacks */
+static bool
+intel_dp_aux_supports_hdr_backlight(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	struct drm_dp_aux *aux = &intel_dp->aux;
+	struct intel_panel *panel = &connector->panel;
+	int ret;
+	u8 tcon_cap[4];
+
+	ret = drm_dp_dpcd_read(aux, INTEL_EDP_HDR_TCON_CAP0, tcon_cap, sizeof(tcon_cap));
+	if (ret < 0)
+		return false;
+
+	if (!(tcon_cap[1] & INTEL_EDP_HDR_TCON_BRIGHTNESS_NITS_CAP))
+		return false;
+
+	if (tcon_cap[0] >= 1) {
+		drm_dbg_kms(&i915->drm, "Detected Intel HDR backlight interface version %d\n",
+			    tcon_cap[0]);
+	} else {
+		drm_dbg_kms(&i915->drm, "Detected unsupported HDR backlight interface version %d\n",
+			    tcon_cap[0]);
+		return false;
+	}
+
+	panel->backlight.edp.intel.sdr_uses_aux =
+		tcon_cap[2] & INTEL_EDP_SDR_TCON_BRIGHTNESS_AUX_CAP;
+
+	return true;
+}
+
+static u32
+intel_dp_aux_hdr_get_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	u8 tmp;
+	u8 buf[2] = { 0 };
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, INTEL_EDP_HDR_GETSET_CTRL_PARAMS, &tmp) < 0) {
+		drm_err(&i915->drm, "Failed to read current backlight mode from DPCD\n");
+		return 0;
+	}
+
+	if (!(tmp & INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE)) {
+		if (!panel->backlight.edp.intel.sdr_uses_aux) {
+			u32 pwm_level = panel->backlight.pwm_funcs->get(connector, pipe);
+
+			return intel_panel_backlight_level_from_pwm(connector, pwm_level);
+		}
+
+		/* Assume 100% brightness if backlight controls aren't enabled yet */
+		return panel->backlight.max;
+	}
+
+	if (drm_dp_dpcd_read(&intel_dp->aux, INTEL_EDP_BRIGHTNESS_NITS_LSB, buf, sizeof(buf)) < 0) {
+		drm_err(&i915->drm, "Failed to read brightness from DPCD\n");
+		return 0;
+	}
+
+	return (buf[1] << 8 | buf[0]);
+}
+
+static void
+intel_dp_aux_hdr_set_aux_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_device *dev = connector->base.dev;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	u8 buf[4] = { 0 };
+
+	buf[0] = level & 0xFF;
+	buf[1] = (level & 0xFF00) >> 8;
+
+	if (drm_dp_dpcd_write(&intel_dp->aux, INTEL_EDP_BRIGHTNESS_NITS_LSB, buf, 4) < 0)
+		drm_err(dev, "Failed to write brightness level to DPCD\n");
+}
+
+static void
+intel_dp_aux_hdr_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	if (panel->backlight.edp.intel.sdr_uses_aux) {
+		intel_dp_aux_hdr_set_aux_backlight(conn_state, level);
+	} else {
+		const u32 pwm_level = intel_panel_backlight_level_to_pwm(connector, level);
+
+		intel_panel_set_pwm_level(conn_state, pwm_level);
+	}
+}
+
+static void
+intel_dp_aux_hdr_enable_backlight(const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	int ret;
+	u8 old_ctrl, ctrl;
+
+	ret = drm_dp_dpcd_readb(&intel_dp->aux, INTEL_EDP_HDR_GETSET_CTRL_PARAMS, &old_ctrl);
+	if (ret < 0) {
+		drm_err(&i915->drm, "Failed to read current backlight control mode: %d\n", ret);
+		return;
+	}
+
+	ctrl = old_ctrl;
+	if (panel->backlight.edp.intel.sdr_uses_aux) {
+		ctrl |= INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE;
+		intel_dp_aux_hdr_set_aux_backlight(conn_state, level);
+	} else {
+		u32 pwm_level = intel_panel_backlight_level_to_pwm(connector, level);
+
+		panel->backlight.pwm_funcs->enable(crtc_state, conn_state, pwm_level);
+
+		ctrl &= ~INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE;
+	}
+
+	if (ctrl != old_ctrl)
+		if (drm_dp_dpcd_writeb(&intel_dp->aux, INTEL_EDP_HDR_GETSET_CTRL_PARAMS, ctrl) < 0)
+			drm_err(&i915->drm, "Failed to configure DPCD brightness controls\n");
+}
+
+static void
+intel_dp_aux_hdr_disable_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	/* Nothing to do for AUX based backlight controls */
+	if (panel->backlight.edp.intel.sdr_uses_aux)
+		return;
+
+	/* Note we want the actual pwm_level to be 0, regardless of pwm_min */
+	panel->backlight.pwm_funcs->disable(conn_state, intel_panel_invert_pwm_level(connector, 0));
+}
+
+static int
+intel_dp_aux_hdr_setup_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	int ret;
+
+	if (panel->backlight.edp.intel.sdr_uses_aux) {
+		drm_dbg_kms(&i915->drm, "SDR backlight is controlled through DPCD\n");
+	} else {
+		drm_dbg_kms(&i915->drm, "SDR backlight is controlled through PWM\n");
+
+		ret = panel->backlight.pwm_funcs->setup(connector, pipe);
+		if (ret < 0) {
+			drm_err(&i915->drm,
+				"Failed to setup SDR backlight controls through PWM: %d\n", ret);
+			return ret;
+		}
+	}
+
+	panel->backlight.max = 512;
+	panel->backlight.min = 0;
+	panel->backlight.level = intel_dp_aux_hdr_get_backlight(connector, pipe);
+	panel->backlight.enabled = panel->backlight.level != 0;
+
+	return 0;
+}
+
 /* VESA backlight callbacks */
 static void set_vesa_backlight_enable(struct intel_dp *intel_dp, bool enable)
 {
@@ -128,7 +318,7 @@ static bool intel_dp_aux_vesa_backlight_dpcd_mode(struct intel_connector *connec
  * Read the current backlight value from DPCD register(s) based
  * on if 8-bit(MSB) or 16-bit(MSB and LSB) values are supported
  */
-static u32 intel_dp_aux_vesa_get_backlight(struct intel_connector *connector)
+static u32 intel_dp_aux_vesa_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
 	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
@@ -195,7 +385,7 @@ static bool intel_dp_aux_vesa_set_pwm_freq(struct intel_connector *connector)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
-	const u8 pn = connector->panel.backlight.pwmgen_bit_count;
+	const u8 pn = connector->panel.backlight.edp.vesa.pwmgen_bit_count;
 	int freq, fxp, f, fxp_actual, fxp_min, fxp_max;
 
 	freq = dev_priv->vbt.backlight.pwm_freq_hz;
@@ -236,6 +426,7 @@ intel_dp_aux_vesa_enable_backlight(const struct intel_crtc_state *crtc_state,
 	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
 	struct intel_panel *panel = &connector->panel;
 	u8 dpcd_buf, new_dpcd_buf, edp_backlight_mode;
+	u8 pwmgen_bit_count = panel->backlight.edp.vesa.pwmgen_bit_count;
 
 	if (drm_dp_dpcd_readb(&intel_dp->aux,
 			DP_EDP_BACKLIGHT_MODE_SET_REGISTER, &dpcd_buf) != 1) {
@@ -256,7 +447,7 @@ intel_dp_aux_vesa_enable_backlight(const struct intel_crtc_state *crtc_state,
 
 		if (drm_dp_dpcd_writeb(&intel_dp->aux,
 				       DP_EDP_PWMGEN_BIT_COUNT,
-				       panel->backlight.pwmgen_bit_count) < 0)
+				       pwmgen_bit_count) < 0)
 			drm_dbg_kms(&i915->drm,
 				    "Failed to write aux pwmgen bit count\n");
 
@@ -364,7 +555,7 @@ static u32 intel_dp_aux_vesa_calc_max_backlight(struct intel_connector *connecto
 			    "Failed to write aux pwmgen bit count\n");
 		return max_backlight;
 	}
-	panel->backlight.pwmgen_bit_count = pn;
+	panel->backlight.edp.vesa.pwmgen_bit_count = pn;
 
 	max_backlight = (1 << pn) - 1;
 
@@ -381,7 +572,7 @@ static int intel_dp_aux_vesa_setup_backlight(struct intel_connector *connector,
 		return -ENODEV;
 
 	panel->backlight.min = 0;
-	panel->backlight.level = intel_dp_aux_vesa_get_backlight(connector);
+	panel->backlight.level = intel_dp_aux_vesa_get_backlight(connector, pipe);
 	panel->backlight.enabled = intel_dp_aux_vesa_backlight_dpcd_mode(connector) &&
 				   panel->backlight.level != 0;
 
@@ -395,9 +586,14 @@ intel_dp_aux_supports_vesa_backlight(struct intel_connector *connector)
 	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
 
 	/* Check the eDP Display control capabilities registers to determine if
-	 * the panel can support backlight control over the aux channel
+	 * the panel can support backlight control over the aux channel.
+	 *
+	 * TODO: We currently only support AUX only backlight configurations, not backlights which
+	 * require a mix of PWM and AUX controls to work. In the mean time, these machines typically
+	 * work just fine using normal PWM controls anyway.
 	 */
 	if (intel_dp->edp_dpcd[1] & DP_EDP_TCON_BACKLIGHT_ADJUSTMENT_CAP &&
+	    (intel_dp->edp_dpcd[1] & DP_EDP_BACKLIGHT_AUX_ENABLE_CAP) &&
 	    (intel_dp->edp_dpcd[2] & DP_EDP_BACKLIGHT_BRIGHTNESS_AUX_SET_CAP)) {
 		drm_dbg_kms(&i915->drm, "AUX Backlight Control Supported!\n");
 		return true;
@@ -405,6 +601,14 @@ intel_dp_aux_supports_vesa_backlight(struct intel_connector *connector)
 	return false;
 }
 
+static const struct intel_panel_bl_funcs intel_dp_hdr_bl_funcs = {
+	.setup = intel_dp_aux_hdr_setup_backlight,
+	.enable = intel_dp_aux_hdr_enable_backlight,
+	.disable = intel_dp_aux_hdr_disable_backlight,
+	.set = intel_dp_aux_hdr_set_backlight,
+	.get = intel_dp_aux_hdr_get_backlight,
+};
+
 static const struct intel_panel_bl_funcs intel_dp_vesa_bl_funcs = {
 	.setup = intel_dp_aux_vesa_setup_backlight,
 	.enable = intel_dp_aux_vesa_enable_backlight,
@@ -413,36 +617,73 @@ static const struct intel_panel_bl_funcs intel_dp_vesa_bl_funcs = {
 	.get = intel_dp_aux_vesa_get_backlight,
 };
 
-int intel_dp_aux_init_backlight_funcs(struct intel_connector *intel_connector)
+enum intel_dp_aux_backlight_modparam {
+	INTEL_DP_AUX_BACKLIGHT_AUTO = -1,
+	INTEL_DP_AUX_BACKLIGHT_OFF = 0,
+	INTEL_DP_AUX_BACKLIGHT_ON = 1,
+	INTEL_DP_AUX_BACKLIGHT_FORCE_VESA = 2,
+	INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL = 3,
+};
+
+int intel_dp_aux_init_backlight_funcs(struct intel_connector *connector)
 {
-	struct intel_panel *panel = &intel_connector->panel;
-	struct intel_dp *intel_dp = enc_to_intel_dp(intel_connector->encoder);
+	struct drm_device *dev = connector->base.dev;
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
 	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	bool try_intel_interface = false, try_vesa_interface = false;
 
-	if (i915->params.enable_dpcd_backlight == 0 ||
-	    !intel_dp_aux_supports_vesa_backlight(intel_connector))
+	/* Check the VBT and user's module parameters to figure out which
+	 * interfaces to probe
+	 */
+	switch (i915->params.enable_dpcd_backlight) {
+	case INTEL_DP_AUX_BACKLIGHT_OFF:
 		return -ENODEV;
+	case INTEL_DP_AUX_BACKLIGHT_AUTO:
+		switch (i915->vbt.backlight.type) {
+		case INTEL_BACKLIGHT_VESA_EDP_AUX_INTERFACE:
+			try_vesa_interface = true;
+			break;
+		case INTEL_BACKLIGHT_DISPLAY_DDI:
+			try_intel_interface = true;
+			try_vesa_interface = true;
+			break;
+		default:
+			return -ENODEV;
+		}
+		break;
+	case INTEL_DP_AUX_BACKLIGHT_ON:
+		if (i915->vbt.backlight.type != INTEL_BACKLIGHT_VESA_EDP_AUX_INTERFACE)
+			try_intel_interface = true;
+
+		try_vesa_interface = true;
+		break;
+	case INTEL_DP_AUX_BACKLIGHT_FORCE_VESA:
+		try_vesa_interface = true;
+		break;
+	case INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL:
+		try_intel_interface = true;
+		break;
+	}
 
 	/*
-	 * There are a lot of machines that don't advertise the backlight
-	 * control interface to use properly in their VBIOS, :\
+	 * A lot of eDP panels in the wild will report supporting both the
+	 * Intel proprietary backlight control interface, and the VESA
+	 * backlight control interface. Many of these panels are liars though,
+	 * and will only work with the Intel interface. So, always probe for
+	 * that first.
 	 */
-	if (i915->vbt.backlight.type !=
-	    INTEL_BACKLIGHT_VESA_EDP_AUX_INTERFACE &&
-	    i915->params.enable_dpcd_backlight != 1 &&
-	    !drm_dp_has_quirk(&intel_dp->desc, intel_dp->edid_quirks,
-			      DP_QUIRK_FORCE_DPCD_BACKLIGHT)) {
-		drm_info(&i915->drm,
-			 "Panel advertises DPCD backlight support, but "
-			 "VBT disagrees. If your backlight controls "
-			 "don't work try booting with "
-			 "i915.enable_dpcd_backlight=1. If your machine "
-			 "needs this, please file a _new_ bug report on "
-			 "drm/i915, see " FDO_BUG_URL " for details.\n");
-		return -ENODEV;
+	if (try_intel_interface && intel_dp_aux_supports_hdr_backlight(connector)) {
+		drm_dbg_kms(dev, "Using Intel proprietary eDP backlight controls\n");
+		panel->backlight.funcs = &intel_dp_hdr_bl_funcs;
+		return 0;
 	}
 
-	panel->backlight.funcs = &intel_dp_vesa_bl_funcs;
+	if (try_vesa_interface && intel_dp_aux_supports_vesa_backlight(connector)) {
+		drm_dbg_kms(dev, "Using VESA eDP backlight controls\n");
+		panel->backlight.funcs = &intel_dp_vesa_bl_funcs;
+		return 0;
+	}
 
-	return 0;
+	return -ENODEV;
 }
diff --git a/drivers/gpu/drm/i915/display/intel_dp_hdcp.c b/drivers/gpu/drm/i915/display/intel_dp_hdcp.c
index 03424d20e9f7..4dba5bb15af5 100644
--- a/drivers/gpu/drm/i915/display/intel_dp_hdcp.c
+++ b/drivers/gpu/drm/i915/display/intel_dp_hdcp.c
@@ -16,6 +16,30 @@
 #include "intel_dp.h"
 #include "intel_hdcp.h"
 
+static unsigned int transcoder_to_stream_enc_status(enum transcoder cpu_transcoder)
+{
+	u32 stream_enc_mask;
+
+	switch (cpu_transcoder) {
+	case TRANSCODER_A:
+		stream_enc_mask = HDCP_STATUS_STREAM_A_ENC;
+		break;
+	case TRANSCODER_B:
+		stream_enc_mask = HDCP_STATUS_STREAM_B_ENC;
+		break;
+	case TRANSCODER_C:
+		stream_enc_mask = HDCP_STATUS_STREAM_C_ENC;
+		break;
+	case TRANSCODER_D:
+		stream_enc_mask = HDCP_STATUS_STREAM_D_ENC;
+		break;
+	default:
+		stream_enc_mask = 0;
+	}
+
+	return stream_enc_mask;
+}
+
 static void intel_dp_hdcp_wait_for_cp_irq(struct intel_hdcp *hdcp, int timeout)
 {
 	long ret;
@@ -561,7 +585,8 @@ int intel_dp_hdcp2_config_stream_type(struct intel_digital_port *dig_port,
 }
 
 static
-int intel_dp_hdcp2_check_link(struct intel_digital_port *dig_port)
+int intel_dp_hdcp2_check_link(struct intel_digital_port *dig_port,
+			      struct intel_connector *connector)
 {
 	u8 rx_status;
 	int ret;
@@ -622,48 +647,143 @@ static const struct intel_hdcp_shim intel_dp_hdcp_shim = {
 };
 
 static int
-intel_dp_mst_hdcp_toggle_signalling(struct intel_digital_port *dig_port,
-				    enum transcoder cpu_transcoder,
-				    bool enable)
+intel_dp_mst_toggle_hdcp_stream_select(struct intel_connector *connector,
+				       bool enable)
 {
-	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
 	int ret;
 
-	if (!enable)
-		usleep_range(6, 60); /* Bspec says >= 6us */
-
-	ret = intel_ddi_toggle_hdcp_signalling(&dig_port->base,
-					       cpu_transcoder, enable);
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base,
+					 hdcp->stream_transcoder, enable,
+					 TRANS_DDI_HDCP_SELECT);
 	if (ret)
-		drm_dbg_kms(&i915->drm, "%s HDCP signalling failed (%d)\n",
-			      enable ? "Enable" : "Disable", ret);
+		drm_err(&i915->drm, "%s HDCP stream select failed (%d)\n",
+			enable ? "Enable" : "Disable", ret);
 	return ret;
 }
 
-static
-bool intel_dp_mst_hdcp_check_link(struct intel_digital_port *dig_port,
-				  struct intel_connector *connector)
+static int
+intel_dp_mst_hdcp_stream_encryption(struct intel_connector *connector,
+				    bool enable)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder = hdcp->stream_transcoder;
+	u32 stream_enc_status;
+	int ret;
+
+	ret = intel_dp_mst_toggle_hdcp_stream_select(connector, enable);
+	if (ret)
+		return ret;
+
+	stream_enc_status =  transcoder_to_stream_enc_status(cpu_transcoder);
+	if (!stream_enc_status)
+		return -EINVAL;
+
+	/* Wait for encryption confirmation */
+	if (intel_de_wait_for_register(i915,
+				       HDCP_STATUS(i915, cpu_transcoder, port),
+				       stream_enc_status,
+				       enable ? stream_enc_status : 0,
+				       HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+		drm_err(&i915->drm, "Timed out waiting for transcoder: %s stream encryption %s\n",
+			transcoder_name(cpu_transcoder), enable ? "enabled" : "disabled");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static bool intel_dp_mst_get_qses_status(struct intel_digital_port *dig_port,
+					 struct intel_connector *connector)
 {
 	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
-	struct intel_dp *intel_dp = &dig_port->dp;
 	struct drm_dp_query_stream_enc_status_ack_reply reply;
+	struct intel_dp *intel_dp = &dig_port->dp;
 	int ret;
 
-	if (!intel_dp_hdcp_check_link(dig_port, connector))
-		return false;
-
 	ret = drm_dp_send_query_stream_enc_status(&intel_dp->mst_mgr,
 						  connector->port, &reply);
 	if (ret) {
 		drm_dbg_kms(&i915->drm,
-			    "[CONNECTOR:%d:%s] failed QSES ret=%d\n",
-			    connector->base.base.id, connector->base.name, ret);
+			    "[%s:%d] failed QSES ret=%d\n",
+			    connector->base.name, connector->base.base.id, ret);
 		return false;
 	}
 
+	drm_dbg_kms(&i915->drm, "[%s:%d] QSES stream auth: %d stream enc: %d\n",
+		    connector->base.name, connector->base.base.id,
+		    reply.auth_completed, reply.encryption_enabled);
+
 	return reply.auth_completed && reply.encryption_enabled;
 }
 
+static int
+intel_dp_mst_hdcp2_stream_encryption(struct intel_connector *connector,
+				     bool enable)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum transcoder cpu_transcoder = hdcp->stream_transcoder;
+	enum pipe pipe = (enum pipe)cpu_transcoder;
+	enum port port = dig_port->base.port;
+	int ret;
+
+	drm_WARN_ON(&i915->drm, enable &&
+		    !!(intel_de_read(i915, HDCP2_AUTH_STREAM(i915, cpu_transcoder, port))
+		    & AUTH_STREAM_TYPE) != data->streams[0].stream_type);
+
+	ret = intel_dp_mst_toggle_hdcp_stream_select(connector, enable);
+	if (ret)
+		return ret;
+
+	/* Wait for encryption confirmation */
+	if (intel_de_wait_for_register(i915,
+				       HDCP2_STREAM_STATUS(i915, cpu_transcoder, pipe),
+				       STREAM_ENCRYPTION_STATUS,
+				       enable ? STREAM_ENCRYPTION_STATUS : 0,
+				       HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+		drm_err(&i915->drm, "Timed out waiting for transcoder: %s stream encryption %s\n",
+			transcoder_name(cpu_transcoder), enable ? "enabled" : "disabled");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+/*
+ * DP v2.0 I.3.3 ignore the stream signature L' in QSES reply msg reply.
+ * I.3.5 MST source device may use a QSES msg to query downstream status
+ * for a particular stream.
+ */
+static
+int intel_dp_mst_hdcp2_check_link(struct intel_digital_port *dig_port,
+				  struct intel_connector *connector)
+{
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	/*
+	 * We do need to do the Link Check only for the connector involved with
+	 * HDCP port authentication and encryption.
+	 * We can re-use the hdcp->is_repeater flag to know that the connector
+	 * involved with HDCP port authentication and encryption.
+	 */
+	if (hdcp->is_repeater) {
+		ret = intel_dp_hdcp2_check_link(dig_port, connector);
+		if (ret)
+			return ret;
+	}
+
+	return intel_dp_mst_get_qses_status(dig_port, connector) ? 0 : -EINVAL;
+}
+
 static const struct intel_hdcp_shim intel_dp_mst_hdcp_shim = {
 	.write_an_aksv = intel_dp_hdcp_write_an_aksv,
 	.read_bksv = intel_dp_hdcp_read_bksv,
@@ -673,10 +793,16 @@ static const struct intel_hdcp_shim intel_dp_mst_hdcp_shim = {
 	.read_ksv_ready = intel_dp_hdcp_read_ksv_ready,
 	.read_ksv_fifo = intel_dp_hdcp_read_ksv_fifo,
 	.read_v_prime_part = intel_dp_hdcp_read_v_prime_part,
-	.toggle_signalling = intel_dp_mst_hdcp_toggle_signalling,
-	.check_link = intel_dp_mst_hdcp_check_link,
+	.toggle_signalling = intel_dp_hdcp_toggle_signalling,
+	.stream_encryption = intel_dp_mst_hdcp_stream_encryption,
+	.check_link = intel_dp_hdcp_check_link,
 	.hdcp_capable = intel_dp_hdcp_capable,
-
+	.write_2_2_msg = intel_dp_hdcp2_write_msg,
+	.read_2_2_msg = intel_dp_hdcp2_read_msg,
+	.config_stream_type = intel_dp_hdcp2_config_stream_type,
+	.stream_2_2_encryption = intel_dp_mst_hdcp2_stream_encryption,
+	.check_2_2_link = intel_dp_mst_hdcp2_check_link,
+	.hdcp_2_2_capable = intel_dp_hdcp2_capable,
 	.protocol = HDCP_PROTOCOL_DP,
 };
 
@@ -693,10 +819,10 @@ int intel_dp_init_hdcp(struct intel_digital_port *dig_port,
 		return 0;
 
 	if (intel_connector->mst_port)
-		return intel_hdcp_init(intel_connector, port,
+		return intel_hdcp_init(intel_connector, dig_port,
 				       &intel_dp_mst_hdcp_shim);
 	else if (!intel_dp_is_edp(intel_dp))
-		return intel_hdcp_init(intel_connector, port,
+		return intel_hdcp_init(intel_connector, dig_port,
 				       &intel_dp_hdcp_shim);
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
index 91d3979902d0..892d7db7d94f 100644
--- a/drivers/gpu/drm/i915/display/intel_dp_link_training.c
+++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
@@ -34,18 +34,6 @@ intel_dp_dump_link_status(const u8 link_status[DP_LINK_STATUS_SIZE])
 		      link_status[3], link_status[4], link_status[5]);
 }
 
-static int intel_dp_lttpr_count(struct intel_dp *intel_dp)
-{
-	int count = drm_dp_lttpr_count(intel_dp->lttpr_common_caps);
-
-	/*
-	 * Pretend no LTTPRs in case of LTTPR detection error, or
-	 * if too many (>8) LTTPRs are detected. This translates to link
-	 * training in transparent mode.
-	 */
-	return count <= 0 ? 0 : count;
-}
-
 static void intel_dp_reset_lttpr_count(struct intel_dp *intel_dp)
 {
 	intel_dp->lttpr_common_caps[DP_PHY_REPEATER_CNT -
@@ -142,6 +130,17 @@ int intel_dp_lttpr_init(struct intel_dp *intel_dp)
 		return 0;
 
 	ret = intel_dp_read_lttpr_common_caps(intel_dp);
+	if (!ret)
+		return 0;
+
+	lttpr_count = drm_dp_lttpr_count(intel_dp->lttpr_common_caps);
+	/*
+	 * Prevent setting LTTPR transparent mode explicitly if no LTTPRs are
+	 * detected as this breaks link training at least on the Dell WD19TB
+	 * dock.
+	 */
+	if (lttpr_count == 0)
+		return 0;
 
 	/*
 	 * See DP Standard v2.0 3.6.6.1. about the explicit disabling of
@@ -150,17 +149,12 @@ int intel_dp_lttpr_init(struct intel_dp *intel_dp)
 	 */
 	intel_dp_set_lttpr_transparent_mode(intel_dp, true);
 
-	if (!ret)
-		return 0;
-
-	lttpr_count = intel_dp_lttpr_count(intel_dp);
-
 	/*
 	 * In case of unsupported number of LTTPRs or failing to switch to
 	 * non-transparent mode fall-back to transparent link training mode,
 	 * still taking into account any LTTPR common lane- rate/count limits.
 	 */
-	if (lttpr_count == 0)
+	if (lttpr_count < 0)
 		return 0;
 
 	if (!intel_dp_set_lttpr_transparent_mode(intel_dp, false)) {
@@ -222,11 +216,11 @@ intel_dp_phy_is_downstream_of_source(struct intel_dp *intel_dp,
 				     enum drm_dp_phy dp_phy)
 {
 	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
-	int lttpr_count = intel_dp_lttpr_count(intel_dp);
+	int lttpr_count = drm_dp_lttpr_count(intel_dp->lttpr_common_caps);
 
-	drm_WARN_ON_ONCE(&i915->drm, lttpr_count == 0 && dp_phy != DP_PHY_DPRX);
+	drm_WARN_ON_ONCE(&i915->drm, lttpr_count <= 0 && dp_phy != DP_PHY_DPRX);
 
-	return lttpr_count == 0 || dp_phy == DP_PHY_LTTPR(lttpr_count - 1);
+	return lttpr_count <= 0 || dp_phy == DP_PHY_LTTPR(lttpr_count - 1);
 }
 
 static u8 intel_dp_phy_voltage_max(struct intel_dp *intel_dp,
@@ -334,6 +328,27 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 	return drm_dp_dpcd_write(&intel_dp->aux, reg, buf, len) == len;
 }
 
+void intel_dp_set_signal_levels(struct intel_dp *intel_dp,
+				const struct intel_crtc_state *crtc_state,
+				enum drm_dp_phy dp_phy)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u8 train_set = intel_dp->train_set[0];
+	char phy_name[10];
+
+	drm_dbg_kms(&dev_priv->drm, "Using vswing level %d%s, pre-emphasis level %d%s, at %s\n",
+		    train_set & DP_TRAIN_VOLTAGE_SWING_MASK,
+		    train_set & DP_TRAIN_MAX_SWING_REACHED ? " (max)" : "",
+		    (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) >>
+		    DP_TRAIN_PRE_EMPHASIS_SHIFT,
+		    train_set & DP_TRAIN_MAX_PRE_EMPHASIS_REACHED ?
+		    " (max)" : "",
+		    intel_dp_phy_name(dp_phy, phy_name, sizeof(phy_name)));
+
+	if (intel_dp_phy_is_downstream_of_source(intel_dp, dp_phy))
+		intel_dp->set_signal_levels(intel_dp, crtc_state);
+}
+
 static bool
 intel_dp_reset_link_train(struct intel_dp *intel_dp,
 			  const struct intel_crtc_state *crtc_state,
@@ -341,7 +356,7 @@ intel_dp_reset_link_train(struct intel_dp *intel_dp,
 			  u8 dp_train_pat)
 {
 	memset(intel_dp->train_set, 0, sizeof(intel_dp->train_set));
-	intel_dp_set_signal_levels(intel_dp, crtc_state);
+	intel_dp_set_signal_levels(intel_dp, crtc_state, dp_phy);
 	return intel_dp_set_link_train(intel_dp, crtc_state, dp_phy, dp_train_pat);
 }
 
@@ -355,7 +370,7 @@ intel_dp_update_link_train(struct intel_dp *intel_dp,
 			    DP_TRAINING_LANE0_SET_PHY_REPEATER(dp_phy);
 	int ret;
 
-	intel_dp_set_signal_levels(intel_dp, crtc_state);
+	intel_dp_set_signal_levels(intel_dp, crtc_state, dp_phy);
 
 	ret = drm_dp_dpcd_write(&intel_dp->aux, reg,
 				intel_dp->train_set, crtc_state->lane_count);
@@ -413,7 +428,7 @@ intel_dp_prepare_link_train(struct intel_dp *intel_dp,
 		drm_dp_dpcd_write(&intel_dp->aux, DP_LINK_RATE_SET,
 				  &rate_select, 1);
 
-	link_config[0] = 0;
+	link_config[0] = crtc_state->vrr.enable ? DP_MSA_TIMING_PAR_IGNORE_EN : 0;
 	link_config[1] = DP_SET_ANSI_8B10B;
 	drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2);
 
@@ -676,9 +691,9 @@ static bool intel_dp_disable_dpcd_training_pattern(struct intel_dp *intel_dp,
  * @intel_dp: DP struct
  * @crtc_state: state for CRTC attached to the encoder
  *
- * Stop the link training of the @intel_dp port, disabling the test pattern
- * symbol generation on the port and disabling the training pattern in
- * the sink's DPCD.
+ * Stop the link training of the @intel_dp port, disabling the training
+ * pattern in the sink's DPCD, and disabling the test pattern symbol
+ * generation on the port.
  *
  * What symbols are output on the port after this point is
  * platform specific: On DDI/VLV/CHV platforms it will be the idle pattern
@@ -692,10 +707,9 @@ void intel_dp_stop_link_train(struct intel_dp *intel_dp,
 {
 	intel_dp->link_trained = true;
 
-	intel_dp_program_link_training_pattern(intel_dp,
-					       crtc_state,
-					       DP_TRAINING_PATTERN_DISABLE);
 	intel_dp_disable_dpcd_training_pattern(intel_dp, DP_PHY_DPRX);
+	intel_dp_program_link_training_pattern(intel_dp, crtc_state,
+					       DP_TRAINING_PATTERN_DISABLE);
 }
 
 static bool
diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.h b/drivers/gpu/drm/i915/display/intel_dp_link_training.h
index 86905aa24db7..6a1f76bd8c75 100644
--- a/drivers/gpu/drm/i915/display/intel_dp_link_training.h
+++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.h
@@ -17,6 +17,9 @@ void intel_dp_get_adjust_train(struct intel_dp *intel_dp,
 			       const struct intel_crtc_state *crtc_state,
 			       enum drm_dp_phy dp_phy,
 			       const u8 link_status[DP_LINK_STATUS_SIZE]);
+void intel_dp_set_signal_levels(struct intel_dp *intel_dp,
+				const struct intel_crtc_state *crtc_state,
+				enum drm_dp_phy dp_phy);
 void intel_dp_start_link_train(struct intel_dp *intel_dp,
 			       const struct intel_crtc_state *crtc_state);
 void intel_dp_stop_link_train(struct intel_dp *intel_dp,
diff --git a/drivers/gpu/drm/i915/display/intel_dp_mst.c b/drivers/gpu/drm/i915/display/intel_dp_mst.c
index 27f04aed8764..b4621ed0127e 100644
--- a/drivers/gpu/drm/i915/display/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/display/intel_dp_mst.c
@@ -53,8 +53,7 @@ static int intel_dp_mst_compute_link_config(struct intel_encoder *encoder,
 	struct drm_i915_private *i915 = to_i915(connector->base.dev);
 	const struct drm_display_mode *adjusted_mode =
 		&crtc_state->hw.adjusted_mode;
-	bool constant_n = drm_dp_has_quirk(&intel_dp->desc, 0,
-					   DP_DPCD_QUIRK_CONSTANT_N);
+	bool constant_n = drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_CONSTANT_N);
 	int bpp, slots = -EINVAL;
 
 	crtc_state->lane_count = limits->max_lane_count;
@@ -69,7 +68,9 @@ static int intel_dp_mst_compute_link_config(struct intel_encoder *encoder,
 
 		slots = drm_dp_atomic_find_vcpi_slots(state, &intel_dp->mst_mgr,
 						      connector->port,
-						      crtc_state->pbn, 0);
+						      crtc_state->pbn,
+						      drm_dp_get_vc_payload_bw(crtc_state->port_clock,
+									       crtc_state->lane_count));
 		if (slots == -EDEADLK)
 			return slots;
 		if (slots >= 0)
@@ -569,7 +570,7 @@ static void intel_mst_enable_dp(struct intel_atomic_state *state,
 	if (conn_state->content_protection ==
 	    DRM_MODE_CONTENT_PROTECTION_DESIRED)
 		intel_hdcp_enable(to_intel_connector(conn_state->connector),
-				  pipe_config->cpu_transcoder,
+				  pipe_config,
 				  (u8)conn_state->hdcp_content_type);
 }
 
@@ -829,12 +830,11 @@ static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topolo
 	intel_attach_force_audio_property(connector);
 	intel_attach_broadcast_rgb_property(connector);
 
-
-	/* TODO: Figure out how to make HDCP work on GEN12+ */
-	if (INTEL_GEN(dev_priv) < 12) {
+	if (INTEL_GEN(dev_priv) <= 12) {
 		ret = intel_dp_init_hdcp(dig_port, intel_connector);
 		if (ret)
-			DRM_DEBUG_KMS("HDCP init failed, skipping.\n");
+			drm_dbg_kms(&dev_priv->drm, "[%s:%d] HDCP MST init failed, skipping.\n",
+				    connector->name, connector->base.id);
 	}
 
 	/*
diff --git a/drivers/gpu/drm/i915/display/intel_dpll.c b/drivers/gpu/drm/i915/display/intel_dpll.c
new file mode 100644
index 000000000000..7ba7f315aaee
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpll.c
@@ -0,0 +1,1363 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include <linux/kernel.h>
+#include "intel_display_types.h"
+#include "intel_display.h"
+#include "intel_dpll.h"
+#include "intel_lvds.h"
+#include "intel_panel.h"
+
+struct intel_limit {
+	struct {
+		int min, max;
+	} dot, vco, n, m, m1, m2, p, p1;
+
+	struct {
+		int dot_limit;
+		int p2_slow, p2_fast;
+	} p2;
+};
+static const struct intel_limit intel_limits_i8xx_dac = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 908000, .max = 1512000 },
+	.n = { .min = 2, .max = 16 },
+	.m = { .min = 96, .max = 140 },
+	.m1 = { .min = 18, .max = 26 },
+	.m2 = { .min = 6, .max = 16 },
+	.p = { .min = 4, .max = 128 },
+	.p1 = { .min = 2, .max = 33 },
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 4, .p2_fast = 2 },
+};
+
+static const struct intel_limit intel_limits_i8xx_dvo = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 908000, .max = 1512000 },
+	.n = { .min = 2, .max = 16 },
+	.m = { .min = 96, .max = 140 },
+	.m1 = { .min = 18, .max = 26 },
+	.m2 = { .min = 6, .max = 16 },
+	.p = { .min = 4, .max = 128 },
+	.p1 = { .min = 2, .max = 33 },
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 4, .p2_fast = 4 },
+};
+
+static const struct intel_limit intel_limits_i8xx_lvds = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 908000, .max = 1512000 },
+	.n = { .min = 2, .max = 16 },
+	.m = { .min = 96, .max = 140 },
+	.m1 = { .min = 18, .max = 26 },
+	.m2 = { .min = 6, .max = 16 },
+	.p = { .min = 4, .max = 128 },
+	.p1 = { .min = 1, .max = 6 },
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 14, .p2_fast = 7 },
+};
+
+static const struct intel_limit intel_limits_i9xx_sdvo = {
+	.dot = { .min = 20000, .max = 400000 },
+	.vco = { .min = 1400000, .max = 2800000 },
+	.n = { .min = 1, .max = 6 },
+	.m = { .min = 70, .max = 120 },
+	.m1 = { .min = 8, .max = 18 },
+	.m2 = { .min = 3, .max = 7 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 200000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit intel_limits_i9xx_lvds = {
+	.dot = { .min = 20000, .max = 400000 },
+	.vco = { .min = 1400000, .max = 2800000 },
+	.n = { .min = 1, .max = 6 },
+	.m = { .min = 70, .max = 120 },
+	.m1 = { .min = 8, .max = 18 },
+	.m2 = { .min = 3, .max = 7 },
+	.p = { .min = 7, .max = 98 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 112000,
+		.p2_slow = 14, .p2_fast = 7 },
+};
+
+
+static const struct intel_limit intel_limits_g4x_sdvo = {
+	.dot = { .min = 25000, .max = 270000 },
+	.vco = { .min = 1750000, .max = 3500000},
+	.n = { .min = 1, .max = 4 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 17, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 10, .max = 30 },
+	.p1 = { .min = 1, .max = 3},
+	.p2 = { .dot_limit = 270000,
+		.p2_slow = 10,
+		.p2_fast = 10
+	},
+};
+
+static const struct intel_limit intel_limits_g4x_hdmi = {
+	.dot = { .min = 22000, .max = 400000 },
+	.vco = { .min = 1750000, .max = 3500000},
+	.n = { .min = 1, .max = 4 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 16, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8},
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit intel_limits_g4x_single_channel_lvds = {
+	.dot = { .min = 20000, .max = 115000 },
+	.vco = { .min = 1750000, .max = 3500000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 17, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 28, .max = 112 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 0,
+		.p2_slow = 14, .p2_fast = 14
+	},
+};
+
+static const struct intel_limit intel_limits_g4x_dual_channel_lvds = {
+	.dot = { .min = 80000, .max = 224000 },
+	.vco = { .min = 1750000, .max = 3500000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 17, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 14, .max = 42 },
+	.p1 = { .min = 2, .max = 6 },
+	.p2 = { .dot_limit = 0,
+		.p2_slow = 7, .p2_fast = 7
+	},
+};
+
+static const struct intel_limit pnv_limits_sdvo = {
+	.dot = { .min = 20000, .max = 400000},
+	.vco = { .min = 1700000, .max = 3500000 },
+	/* Pineview's Ncounter is a ring counter */
+	.n = { .min = 3, .max = 6 },
+	.m = { .min = 2, .max = 256 },
+	/* Pineview only has one combined m divider, which we treat as m2. */
+	.m1 = { .min = 0, .max = 0 },
+	.m2 = { .min = 0, .max = 254 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 200000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit pnv_limits_lvds = {
+	.dot = { .min = 20000, .max = 400000 },
+	.vco = { .min = 1700000, .max = 3500000 },
+	.n = { .min = 3, .max = 6 },
+	.m = { .min = 2, .max = 256 },
+	.m1 = { .min = 0, .max = 0 },
+	.m2 = { .min = 0, .max = 254 },
+	.p = { .min = 7, .max = 112 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 112000,
+		.p2_slow = 14, .p2_fast = 14 },
+};
+
+/* Ironlake / Sandybridge
+ *
+ * We calculate clock using (register_value + 2) for N/M1/M2, so here
+ * the range value for them is (actual_value - 2).
+ */
+static const struct intel_limit ilk_limits_dac = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 5 },
+	.m = { .min = 79, .max = 127 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit ilk_limits_single_lvds = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 79, .max = 118 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 28, .max = 112 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 14, .p2_fast = 14 },
+};
+
+static const struct intel_limit ilk_limits_dual_lvds = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 79, .max = 127 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 14, .max = 56 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 7, .p2_fast = 7 },
+};
+
+/* LVDS 100mhz refclk limits. */
+static const struct intel_limit ilk_limits_single_lvds_100m = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 2 },
+	.m = { .min = 79, .max = 126 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 28, .max = 112 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 14, .p2_fast = 14 },
+};
+
+static const struct intel_limit ilk_limits_dual_lvds_100m = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 79, .max = 126 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 14, .max = 42 },
+	.p1 = { .min = 2, .max = 6 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 7, .p2_fast = 7 },
+};
+
+static const struct intel_limit intel_limits_vlv = {
+	 /*
+	  * These are the data rate limits (measured in fast clocks)
+	  * since those are the strictest limits we have. The fast
+	  * clock and actual rate limits are more relaxed, so checking
+	  * them would make no difference.
+	  */
+	.dot = { .min = 25000 * 5, .max = 270000 * 5 },
+	.vco = { .min = 4000000, .max = 6000000 },
+	.n = { .min = 1, .max = 7 },
+	.m1 = { .min = 2, .max = 3 },
+	.m2 = { .min = 11, .max = 156 },
+	.p1 = { .min = 2, .max = 3 },
+	.p2 = { .p2_slow = 2, .p2_fast = 20 }, /* slow=min, fast=max */
+};
+
+static const struct intel_limit intel_limits_chv = {
+	/*
+	 * These are the data rate limits (measured in fast clocks)
+	 * since those are the strictest limits we have.  The fast
+	 * clock and actual rate limits are more relaxed, so checking
+	 * them would make no difference.
+	 */
+	.dot = { .min = 25000 * 5, .max = 540000 * 5},
+	.vco = { .min = 4800000, .max = 6480000 },
+	.n = { .min = 1, .max = 1 },
+	.m1 = { .min = 2, .max = 2 },
+	.m2 = { .min = 24 << 22, .max = 175 << 22 },
+	.p1 = { .min = 2, .max = 4 },
+	.p2 = {	.p2_slow = 1, .p2_fast = 14 },
+};
+
+static const struct intel_limit intel_limits_bxt = {
+	/* FIXME: find real dot limits */
+	.dot = { .min = 0, .max = INT_MAX },
+	.vco = { .min = 4800000, .max = 6700000 },
+	.n = { .min = 1, .max = 1 },
+	.m1 = { .min = 2, .max = 2 },
+	/* FIXME: find real m2 limits */
+	.m2 = { .min = 2 << 22, .max = 255 << 22 },
+	.p1 = { .min = 2, .max = 4 },
+	.p2 = { .p2_slow = 1, .p2_fast = 20 },
+};
+
+/*
+ * Platform specific helpers to calculate the port PLL loopback- (clock.m),
+ * and post-divider (clock.p) values, pre- (clock.vco) and post-divided fast
+ * (clock.dot) clock rates. This fast dot clock is fed to the port's IO logic.
+ * The helpers' return value is the rate of the clock that is fed to the
+ * display engine's pipe which can be the above fast dot clock rate or a
+ * divided-down version of it.
+ */
+/* m1 is reserved as 0 in Pineview, n is a ring counter */
+int pnv_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = clock->m2 + 2;
+	clock->p = clock->p1 * clock->p2;
+	if (WARN_ON(clock->n == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot;
+}
+
+static u32 i9xx_dpll_compute_m(struct dpll *dpll)
+{
+	return 5 * (dpll->m1 + 2) + (dpll->m2 + 2);
+}
+
+int i9xx_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = i9xx_dpll_compute_m(clock);
+	clock->p = clock->p1 * clock->p2;
+	if (WARN_ON(clock->n + 2 == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n + 2);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot;
+}
+
+int vlv_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = clock->m1 * clock->m2;
+	clock->p = clock->p1 * clock->p2;
+	if (WARN_ON(clock->n == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot / 5;
+}
+
+int chv_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = clock->m1 * clock->m2;
+	clock->p = clock->p1 * clock->p2;
+	if (WARN_ON(clock->n == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(refclk, clock->m),
+					   clock->n << 22);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot / 5;
+}
+
+/*
+ * Returns whether the given set of divisors are valid for a given refclk with
+ * the given connectors.
+ */
+static bool intel_pll_is_valid(struct drm_i915_private *dev_priv,
+			       const struct intel_limit *limit,
+			       const struct dpll *clock)
+{
+	if (clock->n < limit->n.min || limit->n.max < clock->n)
+		return false;
+	if (clock->p1 < limit->p1.min || limit->p1.max < clock->p1)
+		return false;
+	if (clock->m2 < limit->m2.min || limit->m2.max < clock->m2)
+		return false;
+	if (clock->m1 < limit->m1.min || limit->m1.max < clock->m1)
+		return false;
+
+	if (!IS_PINEVIEW(dev_priv) && !IS_VALLEYVIEW(dev_priv) &&
+	    !IS_CHERRYVIEW(dev_priv) && !IS_GEN9_LP(dev_priv))
+		if (clock->m1 <= clock->m2)
+			return false;
+
+	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv) &&
+	    !IS_GEN9_LP(dev_priv)) {
+		if (clock->p < limit->p.min || limit->p.max < clock->p)
+			return false;
+		if (clock->m < limit->m.min || limit->m.max < clock->m)
+			return false;
+	}
+
+	if (clock->vco < limit->vco.min || limit->vco.max < clock->vco)
+		return false;
+	/* XXX: We may need to be checking "Dot clock" depending on the multiplier,
+	 * connector, etc., rather than just a single range.
+	 */
+	if (clock->dot < limit->dot.min || limit->dot.max < clock->dot)
+		return false;
+
+	return true;
+}
+
+static int
+i9xx_select_p2_div(const struct intel_limit *limit,
+		   const struct intel_crtc_state *crtc_state,
+		   int target)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		/*
+		 * For LVDS just rely on its current settings for dual-channel.
+		 * We haven't figured out how to reliably set up different
+		 * single/dual channel state, if we even can.
+		 */
+		if (intel_is_dual_link_lvds(dev_priv))
+			return limit->p2.p2_fast;
+		else
+			return limit->p2.p2_slow;
+	} else {
+		if (target < limit->p2.dot_limit)
+			return limit->p2.p2_slow;
+		else
+			return limit->p2.p2_fast;
+	}
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.  The returned values represent the clock equation:
+ * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
+ *
+ * Target and reference clocks are specified in kHz.
+ *
+ * If match_clock is provided, then best_clock P divider must match the P
+ * divider from @match_clock used for LVDS downclocking.
+ */
+static bool
+i9xx_find_best_dpll(const struct intel_limit *limit,
+		    struct intel_crtc_state *crtc_state,
+		    int target, int refclk, struct dpll *match_clock,
+		    struct dpll *best_clock)
+{
+	struct drm_device *dev = crtc_state->uapi.crtc->dev;
+	struct dpll clock;
+	int err = target;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
+
+	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
+	     clock.m1++) {
+		for (clock.m2 = limit->m2.min;
+		     clock.m2 <= limit->m2.max; clock.m2++) {
+			if (clock.m2 >= clock.m1)
+				break;
+			for (clock.n = limit->n.min;
+			     clock.n <= limit->n.max; clock.n++) {
+				for (clock.p1 = limit->p1.min;
+					clock.p1 <= limit->p1.max; clock.p1++) {
+					int this_err;
+
+					i9xx_calc_dpll_params(refclk, &clock);
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err) {
+						*best_clock = clock;
+						err = this_err;
+					}
+				}
+			}
+		}
+	}
+
+	return (err != target);
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.  The returned values represent the clock equation:
+ * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
+ *
+ * Target and reference clocks are specified in kHz.
+ *
+ * If match_clock is provided, then best_clock P divider must match the P
+ * divider from @match_clock used for LVDS downclocking.
+ */
+static bool
+pnv_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk, struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct drm_device *dev = crtc_state->uapi.crtc->dev;
+	struct dpll clock;
+	int err = target;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
+
+	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
+	     clock.m1++) {
+		for (clock.m2 = limit->m2.min;
+		     clock.m2 <= limit->m2.max; clock.m2++) {
+			for (clock.n = limit->n.min;
+			     clock.n <= limit->n.max; clock.n++) {
+				for (clock.p1 = limit->p1.min;
+					clock.p1 <= limit->p1.max; clock.p1++) {
+					int this_err;
+
+					pnv_calc_dpll_params(refclk, &clock);
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err) {
+						*best_clock = clock;
+						err = this_err;
+					}
+				}
+			}
+		}
+	}
+
+	return (err != target);
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.  The returned values represent the clock equation:
+ * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
+ *
+ * Target and reference clocks are specified in kHz.
+ *
+ * If match_clock is provided, then best_clock P divider must match the P
+ * divider from @match_clock used for LVDS downclocking.
+ */
+static bool
+g4x_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk, struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct drm_device *dev = crtc_state->uapi.crtc->dev;
+	struct dpll clock;
+	int max_n;
+	bool found = false;
+	/* approximately equals target * 0.00585 */
+	int err_most = (target >> 8) + (target >> 9);
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
+
+	max_n = limit->n.max;
+	/* based on hardware requirement, prefer smaller n to precision */
+	for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) {
+		/* based on hardware requirement, prefere larger m1,m2 */
+		for (clock.m1 = limit->m1.max;
+		     clock.m1 >= limit->m1.min; clock.m1--) {
+			for (clock.m2 = limit->m2.max;
+			     clock.m2 >= limit->m2.min; clock.m2--) {
+				for (clock.p1 = limit->p1.max;
+				     clock.p1 >= limit->p1.min; clock.p1--) {
+					int this_err;
+
+					i9xx_calc_dpll_params(refclk, &clock);
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err_most) {
+						*best_clock = clock;
+						err_most = this_err;
+						max_n = clock.n;
+						found = true;
+					}
+				}
+			}
+		}
+	}
+	return found;
+}
+
+/*
+ * Check if the calculated PLL configuration is more optimal compared to the
+ * best configuration and error found so far. Return the calculated error.
+ */
+static bool vlv_PLL_is_optimal(struct drm_device *dev, int target_freq,
+			       const struct dpll *calculated_clock,
+			       const struct dpll *best_clock,
+			       unsigned int best_error_ppm,
+			       unsigned int *error_ppm)
+{
+	/*
+	 * For CHV ignore the error and consider only the P value.
+	 * Prefer a bigger P value based on HW requirements.
+	 */
+	if (IS_CHERRYVIEW(to_i915(dev))) {
+		*error_ppm = 0;
+
+		return calculated_clock->p > best_clock->p;
+	}
+
+	if (drm_WARN_ON_ONCE(dev, !target_freq))
+		return false;
+
+	*error_ppm = div_u64(1000000ULL *
+				abs(target_freq - calculated_clock->dot),
+			     target_freq);
+	/*
+	 * Prefer a better P value over a better (smaller) error if the error
+	 * is small. Ensure this preference for future configurations too by
+	 * setting the error to 0.
+	 */
+	if (*error_ppm < 100 && calculated_clock->p > best_clock->p) {
+		*error_ppm = 0;
+
+		return true;
+	}
+
+	return *error_ppm + 10 < best_error_ppm;
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.  The returned values represent the clock equation:
+ * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
+ */
+static bool
+vlv_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk, struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_device *dev = crtc->base.dev;
+	struct dpll clock;
+	unsigned int bestppm = 1000000;
+	/* min update 19.2 MHz */
+	int max_n = min(limit->n.max, refclk / 19200);
+	bool found = false;
+
+	target *= 5; /* fast clock */
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	/* based on hardware requirement, prefer smaller n to precision */
+	for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) {
+		for (clock.p1 = limit->p1.max; clock.p1 >= limit->p1.min; clock.p1--) {
+			for (clock.p2 = limit->p2.p2_fast; clock.p2 >= limit->p2.p2_slow;
+			     clock.p2 -= clock.p2 > 10 ? 2 : 1) {
+				clock.p = clock.p1 * clock.p2;
+				/* based on hardware requirement, prefer bigger m1,m2 values */
+				for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max; clock.m1++) {
+					unsigned int ppm;
+
+					clock.m2 = DIV_ROUND_CLOSEST(target * clock.p * clock.n,
+								     refclk * clock.m1);
+
+					vlv_calc_dpll_params(refclk, &clock);
+
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+
+					if (!vlv_PLL_is_optimal(dev, target,
+								&clock,
+								best_clock,
+								bestppm, &ppm))
+						continue;
+
+					*best_clock = clock;
+					bestppm = ppm;
+					found = true;
+				}
+			}
+		}
+	}
+
+	return found;
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.  The returned values represent the clock equation:
+ * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
+ */
+static bool
+chv_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk, struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_device *dev = crtc->base.dev;
+	unsigned int best_error_ppm;
+	struct dpll clock;
+	u64 m2;
+	int found = false;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+	best_error_ppm = 1000000;
+
+	/*
+	 * Based on hardware doc, the n always set to 1, and m1 always
+	 * set to 2.  If requires to support 200Mhz refclk, we need to
+	 * revisit this because n may not 1 anymore.
+	 */
+	clock.n = 1;
+	clock.m1 = 2;
+	target *= 5;	/* fast clock */
+
+	for (clock.p1 = limit->p1.max; clock.p1 >= limit->p1.min; clock.p1--) {
+		for (clock.p2 = limit->p2.p2_fast;
+				clock.p2 >= limit->p2.p2_slow;
+				clock.p2 -= clock.p2 > 10 ? 2 : 1) {
+			unsigned int error_ppm;
+
+			clock.p = clock.p1 * clock.p2;
+
+			m2 = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(target, clock.p * clock.n) << 22,
+						   refclk * clock.m1);
+
+			if (m2 > INT_MAX/clock.m1)
+				continue;
+
+			clock.m2 = m2;
+
+			chv_calc_dpll_params(refclk, &clock);
+
+			if (!intel_pll_is_valid(to_i915(dev), limit, &clock))
+				continue;
+
+			if (!vlv_PLL_is_optimal(dev, target, &clock, best_clock,
+						best_error_ppm, &error_ppm))
+				continue;
+
+			*best_clock = clock;
+			best_error_ppm = error_ppm;
+			found = true;
+		}
+	}
+
+	return found;
+}
+
+bool bxt_find_best_dpll(struct intel_crtc_state *crtc_state,
+			struct dpll *best_clock)
+{
+	int refclk = 100000;
+	const struct intel_limit *limit = &intel_limits_bxt;
+
+	return chv_find_best_dpll(limit, crtc_state,
+				  crtc_state->port_clock, refclk,
+				  NULL, best_clock);
+}
+
+static u32 pnv_dpll_compute_fp(struct dpll *dpll)
+{
+	return (1 << dpll->n) << 16 | dpll->m2;
+}
+
+static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
+				     struct intel_crtc_state *crtc_state,
+				     struct dpll *reduced_clock)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 fp, fp2 = 0;
+
+	if (IS_PINEVIEW(dev_priv)) {
+		fp = pnv_dpll_compute_fp(&crtc_state->dpll);
+		if (reduced_clock)
+			fp2 = pnv_dpll_compute_fp(reduced_clock);
+	} else {
+		fp = i9xx_dpll_compute_fp(&crtc_state->dpll);
+		if (reduced_clock)
+			fp2 = i9xx_dpll_compute_fp(reduced_clock);
+	}
+
+	crtc_state->dpll_hw_state.fp0 = fp;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	    reduced_clock) {
+		crtc_state->dpll_hw_state.fp1 = fp2;
+	} else {
+		crtc_state->dpll_hw_state.fp1 = fp;
+	}
+}
+
+static void i9xx_compute_dpll(struct intel_crtc *crtc,
+			      struct intel_crtc_state *crtc_state,
+			      struct dpll *reduced_clock)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dpll;
+	struct dpll *clock = &crtc_state->dpll;
+
+	i9xx_update_pll_dividers(crtc, crtc_state, reduced_clock);
+
+	dpll = DPLL_VGA_MODE_DIS;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
+		dpll |= DPLLB_MODE_LVDS;
+	else
+		dpll |= DPLLB_MODE_DAC_SERIAL;
+
+	if (IS_I945G(dev_priv) || IS_I945GM(dev_priv) ||
+	    IS_G33(dev_priv) || IS_PINEVIEW(dev_priv)) {
+		dpll |= (crtc_state->pixel_multiplier - 1)
+			<< SDVO_MULTIPLIER_SHIFT_HIRES;
+	}
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	/* compute bitmask from p1 value */
+	if (IS_PINEVIEW(dev_priv))
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT_PINEVIEW;
+	else {
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+		if (IS_G4X(dev_priv) && reduced_clock)
+			dpll |= (1 << (reduced_clock->p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+	}
+	switch (clock->p2) {
+	case 5:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
+		break;
+	case 7:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7;
+		break;
+	case 10:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10;
+		break;
+	case 14:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14;
+		break;
+	}
+	if (INTEL_GEN(dev_priv) >= 4)
+		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
+
+	if (crtc_state->sdvo_tv_clock)
+		dpll |= PLL_REF_INPUT_TVCLKINBC;
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+		 intel_panel_use_ssc(dev_priv))
+		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
+	else
+		dpll |= PLL_REF_INPUT_DREFCLK;
+
+	dpll |= DPLL_VCO_ENABLE;
+	crtc_state->dpll_hw_state.dpll = dpll;
+
+	if (INTEL_GEN(dev_priv) >= 4) {
+		u32 dpll_md = (crtc_state->pixel_multiplier - 1)
+			<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+		crtc_state->dpll_hw_state.dpll_md = dpll_md;
+	}
+}
+
+static void i8xx_compute_dpll(struct intel_crtc *crtc,
+			      struct intel_crtc_state *crtc_state,
+			      struct dpll *reduced_clock)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 dpll;
+	struct dpll *clock = &crtc_state->dpll;
+
+	i9xx_update_pll_dividers(crtc, crtc_state, reduced_clock);
+
+	dpll = DPLL_VGA_MODE_DIS;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	} else {
+		if (clock->p1 == 2)
+			dpll |= PLL_P1_DIVIDE_BY_TWO;
+		else
+			dpll |= (clock->p1 - 2) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+		if (clock->p2 == 4)
+			dpll |= PLL_P2_DIVIDE_BY_4;
+	}
+
+	/*
+	 * Bspec:
+	 * "[Almador Errata}: For the correct operation of the muxed DVO pins
+	 *  (GDEVSELB/I2Cdata, GIRDBY/I2CClk) and (GFRAMEB/DVI_Data,
+	 *  GTRDYB/DVI_Clk): Bit 31 (DPLL VCO Enable) and Bit 30 (2X Clock
+	 *  Enable) must be set to “1� in both the DPLL A Control Register
+	 *  (06014h-06017h) and DPLL B Control Register (06018h-0601Bh)."
+	 *
+	 * For simplicity We simply keep both bits always enabled in
+	 * both DPLLS. The spec says we should disable the DVO 2X clock
+	 * when not needed, but this seems to work fine in practice.
+	 */
+	if (IS_I830(dev_priv) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO))
+		dpll |= DPLL_DVO_2X_MODE;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	    intel_panel_use_ssc(dev_priv))
+		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
+	else
+		dpll |= PLL_REF_INPUT_DREFCLK;
+
+	dpll |= DPLL_VCO_ENABLE;
+	crtc_state->dpll_hw_state.dpll = dpll;
+}
+
+static int hsw_crtc_compute_clock(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(crtc_state->uapi.state);
+
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI) ||
+	    INTEL_GEN(dev_priv) >= 11) {
+		struct intel_encoder *encoder =
+			intel_get_crtc_new_encoder(state, crtc_state);
+
+		if (!intel_reserve_shared_dplls(state, crtc, encoder)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "failed to find PLL for pipe %c\n",
+				    pipe_name(crtc->pipe));
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static bool ilk_needs_fb_cb_tune(struct dpll *dpll, int factor)
+{
+	return i9xx_dpll_compute_m(dpll) < factor * dpll->n;
+}
+
+
+static void ilk_compute_dpll(struct intel_crtc *crtc,
+			     struct intel_crtc_state *crtc_state,
+			     struct dpll *reduced_clock)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dpll, fp, fp2;
+	int factor;
+
+	/* Enable autotuning of the PLL clock (if permissible) */
+	factor = 21;
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if ((intel_panel_use_ssc(dev_priv) &&
+		     dev_priv->vbt.lvds_ssc_freq == 100000) ||
+		    (HAS_PCH_IBX(dev_priv) &&
+		     intel_is_dual_link_lvds(dev_priv)))
+			factor = 25;
+	} else if (crtc_state->sdvo_tv_clock) {
+		factor = 20;
+	}
+
+	fp = i9xx_dpll_compute_fp(&crtc_state->dpll);
+
+	if (ilk_needs_fb_cb_tune(&crtc_state->dpll, factor))
+		fp |= FP_CB_TUNE;
+
+	if (reduced_clock) {
+		fp2 = i9xx_dpll_compute_fp(reduced_clock);
+
+		if (reduced_clock->m < factor * reduced_clock->n)
+			fp2 |= FP_CB_TUNE;
+	} else {
+		fp2 = fp;
+	}
+
+	dpll = 0;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
+		dpll |= DPLLB_MODE_LVDS;
+	else
+		dpll |= DPLLB_MODE_DAC_SERIAL;
+
+	dpll |= (crtc_state->pixel_multiplier - 1)
+		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	/*
+	 * The high speed IO clock is only really required for
+	 * SDVO/HDMI/DP, but we also enable it for CRT to make it
+	 * possible to share the DPLL between CRT and HDMI. Enabling
+	 * the clock needlessly does no real harm, except use up a
+	 * bit of power potentially.
+	 *
+	 * We'll limit this to IVB with 3 pipes, since it has only two
+	 * DPLLs and so DPLL sharing is the only way to get three pipes
+	 * driving PCH ports at the same time. On SNB we could do this,
+	 * and potentially avoid enabling the second DPLL, but it's not
+	 * clear if it''s a win or loss power wise. No point in doing
+	 * this on ILK at all since it has a fixed DPLL<->pipe mapping.
+	 */
+	if (INTEL_NUM_PIPES(dev_priv) == 3 &&
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	/* compute bitmask from p1 value */
+	dpll |= (1 << (crtc_state->dpll.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	/* also FPA1 */
+	dpll |= (1 << (crtc_state->dpll.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+
+	switch (crtc_state->dpll.p2) {
+	case 5:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
+		break;
+	case 7:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7;
+		break;
+	case 10:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10;
+		break;
+	case 14:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14;
+		break;
+	}
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	    intel_panel_use_ssc(dev_priv))
+		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
+	else
+		dpll |= PLL_REF_INPUT_DREFCLK;
+
+	dpll |= DPLL_VCO_ENABLE;
+
+	crtc_state->dpll_hw_state.dpll = dpll;
+	crtc_state->dpll_hw_state.fp0 = fp;
+	crtc_state->dpll_hw_state.fp1 = fp2;
+}
+
+static int ilk_crtc_compute_clock(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(crtc_state->uapi.state);
+	const struct intel_limit *limit;
+	int refclk = 120000;
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
+	if (!crtc_state->has_pch_encoder)
+		return 0;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    dev_priv->vbt.lvds_ssc_freq);
+			refclk = dev_priv->vbt.lvds_ssc_freq;
+		}
+
+		if (intel_is_dual_link_lvds(dev_priv)) {
+			if (refclk == 100000)
+				limit = &ilk_limits_dual_lvds_100m;
+			else
+				limit = &ilk_limits_dual_lvds;
+		} else {
+			if (refclk == 100000)
+				limit = &ilk_limits_single_lvds_100m;
+			else
+				limit = &ilk_limits_single_lvds;
+		}
+	} else {
+		limit = &ilk_limits_dac;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !g4x_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&dev_priv->drm,
+			"Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	ilk_compute_dpll(crtc, crtc_state, NULL);
+
+	if (!intel_reserve_shared_dplls(state, crtc, NULL)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "failed to find PLL for pipe %c\n",
+			    pipe_name(crtc->pipe));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+void vlv_compute_dpll(struct intel_crtc *crtc,
+		      struct intel_crtc_state *pipe_config)
+{
+	pipe_config->dpll_hw_state.dpll = DPLL_INTEGRATED_REF_CLK_VLV |
+		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+	if (crtc->pipe != PIPE_A)
+		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	/* DPLL not used with DSI, but still need the rest set up */
+	if (!intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DSI))
+		pipe_config->dpll_hw_state.dpll |= DPLL_VCO_ENABLE |
+			DPLL_EXT_BUFFER_ENABLE_VLV;
+
+	pipe_config->dpll_hw_state.dpll_md =
+		(pipe_config->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
+}
+
+void chv_compute_dpll(struct intel_crtc *crtc,
+		      struct intel_crtc_state *pipe_config)
+{
+	pipe_config->dpll_hw_state.dpll = DPLL_SSC_REF_CLK_CHV |
+		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+	if (crtc->pipe != PIPE_A)
+		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	/* DPLL not used with DSI, but still need the rest set up */
+	if (!intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DSI))
+		pipe_config->dpll_hw_state.dpll |= DPLL_VCO_ENABLE;
+
+	pipe_config->dpll_hw_state.dpll_md =
+		(pipe_config->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
+}
+
+static int chv_crtc_compute_clock(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	int refclk = 100000;
+	const struct intel_limit *limit = &intel_limits_chv;
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (!crtc_state->clock_set &&
+	    !chv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&i915->drm, "Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	chv_compute_dpll(crtc, crtc_state);
+
+	return 0;
+}
+
+static int vlv_crtc_compute_clock(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	int refclk = 100000;
+	const struct intel_limit *limit = &intel_limits_vlv;
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (!crtc_state->clock_set &&
+	    !vlv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&i915->drm,  "Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	vlv_compute_dpll(crtc, crtc_state);
+
+	return 0;
+}
+
+static int g4x_crtc_compute_clock(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_limit *limit;
+	int refclk = 96000;
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		if (intel_is_dual_link_lvds(dev_priv))
+			limit = &intel_limits_g4x_dual_channel_lvds;
+		else
+			limit = &intel_limits_g4x_single_channel_lvds;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI) ||
+		   intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG)) {
+		limit = &intel_limits_g4x_hdmi;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO)) {
+		limit = &intel_limits_g4x_sdvo;
+	} else {
+		/* The option is for other outputs */
+		limit = &intel_limits_i9xx_sdvo;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !g4x_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&dev_priv->drm,
+			"Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	i9xx_compute_dpll(crtc, crtc_state, NULL);
+
+	return 0;
+}
+
+static int pnv_crtc_compute_clock(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	const struct intel_limit *limit;
+	int refclk = 96000;
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		limit = &pnv_limits_lvds;
+	} else {
+		limit = &pnv_limits_sdvo;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !pnv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&dev_priv->drm,
+			"Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	i9xx_compute_dpll(crtc, crtc_state, NULL);
+
+	return 0;
+}
+
+static int i9xx_crtc_compute_clock(struct intel_crtc *crtc,
+				   struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	const struct intel_limit *limit;
+	int refclk = 96000;
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		limit = &intel_limits_i9xx_lvds;
+	} else {
+		limit = &intel_limits_i9xx_sdvo;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !i9xx_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				 refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&dev_priv->drm,
+			"Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	i9xx_compute_dpll(crtc, crtc_state, NULL);
+
+	return 0;
+}
+
+static int i8xx_crtc_compute_clock(struct intel_crtc *crtc,
+				   struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	const struct intel_limit *limit;
+	int refclk = 48000;
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		limit = &intel_limits_i8xx_lvds;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO)) {
+		limit = &intel_limits_i8xx_dvo;
+	} else {
+		limit = &intel_limits_i8xx_dac;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !i9xx_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				 refclk, NULL, &crtc_state->dpll)) {
+		drm_err(&dev_priv->drm,
+			"Couldn't find PLL settings for mode!\n");
+		return -EINVAL;
+	}
+
+	i8xx_compute_dpll(crtc, crtc_state, NULL);
+
+	return 0;
+}
+
+void
+intel_dpll_init_clock_hook(struct drm_i915_private *dev_priv)
+{
+	if (INTEL_GEN(dev_priv) >= 9 || HAS_DDI(dev_priv))
+		dev_priv->display.crtc_compute_clock = hsw_crtc_compute_clock;
+	else if (HAS_PCH_SPLIT(dev_priv))
+		dev_priv->display.crtc_compute_clock = ilk_crtc_compute_clock;
+	else if (IS_CHERRYVIEW(dev_priv))
+		dev_priv->display.crtc_compute_clock = chv_crtc_compute_clock;
+	else if (IS_VALLEYVIEW(dev_priv))
+		dev_priv->display.crtc_compute_clock = vlv_crtc_compute_clock;
+	else if (IS_G4X(dev_priv))
+		dev_priv->display.crtc_compute_clock = g4x_crtc_compute_clock;
+	else if (IS_PINEVIEW(dev_priv))
+		dev_priv->display.crtc_compute_clock = pnv_crtc_compute_clock;
+	else if (!IS_GEN(dev_priv, 2))
+		dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock;
+	else
+		dev_priv->display.crtc_compute_clock = i8xx_crtc_compute_clock;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dpll.h b/drivers/gpu/drm/i915/display/intel_dpll.h
new file mode 100644
index 000000000000..caf4615092e1
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpll.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_DPLL_H_
+#define _INTEL_DPLL_H_
+
+struct dpll;
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_crtc_state;
+
+void intel_dpll_init_clock_hook(struct drm_i915_private *dev_priv);
+int vlv_calc_dpll_params(int refclk, struct dpll *clock);
+int pnv_calc_dpll_params(int refclk, struct dpll *clock);
+int i9xx_calc_dpll_params(int refclk, struct dpll *clock);
+void vlv_compute_dpll(struct intel_crtc *crtc,
+		      struct intel_crtc_state *pipe_config);
+void chv_compute_dpll(struct intel_crtc *crtc,
+		      struct intel_crtc_state *pipe_config);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c b/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c
index 88628764956d..584c14c4cbd0 100644
--- a/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c
+++ b/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c
@@ -43,7 +43,7 @@
 
 #define PANEL_PWM_MAX_VALUE		0xFF
 
-static u32 dcs_get_backlight(struct intel_connector *connector)
+static u32 dcs_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct intel_encoder *encoder = intel_attached_encoder(connector);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
diff --git a/drivers/gpu/drm/i915/display/intel_fbc.c b/drivers/gpu/drm/i915/display/intel_fbc.c
index 33200b5cfad0..5fd4fa4805ef 100644
--- a/drivers/gpu/drm/i915/display/intel_fbc.c
+++ b/drivers/gpu/drm/i915/display/intel_fbc.c
@@ -676,7 +676,7 @@ static bool intel_fbc_hw_tracking_covers_screen(struct intel_crtc *crtc)
 }
 
 static bool tiling_is_valid(struct drm_i915_private *dev_priv,
-			    uint64_t modifier)
+			    u64 modifier)
 {
 	switch (modifier) {
 	case DRM_FORMAT_MOD_LINEAR:
diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.c b/drivers/gpu/drm/i915/display/intel_fbdev.c
index 842c04e63214..84f853f113b9 100644
--- a/drivers/gpu/drm/i915/display/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/display/intel_fbdev.c
@@ -256,7 +256,7 @@ static int intelfb_create(struct drm_fb_helper *helper,
 	 * If the object is stolen however, it will be full of whatever
 	 * garbage was left in there.
 	 */
-	if (vma->obj->stolen && !prealloc)
+	if (!i915_gem_object_is_shmem(vma->obj) && !prealloc)
 		memset_io(info->screen_base, 0, info->screen_size);
 
 	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
@@ -595,7 +595,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous
 	 * full of whatever garbage was left in there.
 	 */
 	if (state == FBINFO_STATE_RUNNING &&
-	    intel_fb_obj(&ifbdev->fb->base)->stolen)
+	    !i915_gem_object_is_shmem(intel_fb_obj(&ifbdev->fb->base)))
 		memset_io(info->screen_base, 0, info->screen_size);
 
 	drm_fb_helper_set_suspend(&ifbdev->helper, state);
diff --git a/drivers/gpu/drm/i915/display/intel_fdi.c b/drivers/gpu/drm/i915/display/intel_fdi.c
new file mode 100644
index 000000000000..b2eb96ae10a2
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fdi.c
@@ -0,0 +1,683 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include "intel_atomic.h"
+#include "intel_display_types.h"
+#include "intel_fdi.h"
+
+/* units of 100MHz */
+static int pipe_required_fdi_lanes(struct intel_crtc_state *crtc_state)
+{
+	if (crtc_state->hw.enable && crtc_state->has_pch_encoder)
+		return crtc_state->fdi_lanes;
+
+	return 0;
+}
+
+static int ilk_check_fdi_lanes(struct drm_device *dev, enum pipe pipe,
+			       struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_atomic_state *state = pipe_config->uapi.state;
+	struct intel_crtc *other_crtc;
+	struct intel_crtc_state *other_crtc_state;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "checking fdi config on pipe %c, lanes %i\n",
+		    pipe_name(pipe), pipe_config->fdi_lanes);
+	if (pipe_config->fdi_lanes > 4) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "invalid fdi lane config on pipe %c: %i lanes\n",
+			    pipe_name(pipe), pipe_config->fdi_lanes);
+		return -EINVAL;
+	}
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		if (pipe_config->fdi_lanes > 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "only 2 lanes on haswell, required: %i lanes\n",
+				    pipe_config->fdi_lanes);
+			return -EINVAL;
+		} else {
+			return 0;
+		}
+	}
+
+	if (INTEL_NUM_PIPES(dev_priv) == 2)
+		return 0;
+
+	/* Ivybridge 3 pipe is really complicated */
+	switch (pipe) {
+	case PIPE_A:
+		return 0;
+	case PIPE_B:
+		if (pipe_config->fdi_lanes <= 2)
+			return 0;
+
+		other_crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_C);
+		other_crtc_state =
+			intel_atomic_get_crtc_state(state, other_crtc);
+		if (IS_ERR(other_crtc_state))
+			return PTR_ERR(other_crtc_state);
+
+		if (pipe_required_fdi_lanes(other_crtc_state) > 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "invalid shared fdi lane config on pipe %c: %i lanes\n",
+				    pipe_name(pipe), pipe_config->fdi_lanes);
+			return -EINVAL;
+		}
+		return 0;
+	case PIPE_C:
+		if (pipe_config->fdi_lanes > 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "only 2 lanes on pipe %c: required %i lanes\n",
+				    pipe_name(pipe), pipe_config->fdi_lanes);
+			return -EINVAL;
+		}
+
+		other_crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_B);
+		other_crtc_state =
+			intel_atomic_get_crtc_state(state, other_crtc);
+		if (IS_ERR(other_crtc_state))
+			return PTR_ERR(other_crtc_state);
+
+		if (pipe_required_fdi_lanes(other_crtc_state) > 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "fdi link B uses too many lanes to enable link C\n");
+			return -EINVAL;
+		}
+		return 0;
+	default:
+		BUG();
+	}
+}
+
+int ilk_fdi_compute_config(struct intel_crtc *intel_crtc,
+				  struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *i915 = to_i915(dev);
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	int lane, link_bw, fdi_dotclock, ret;
+	bool needs_recompute = false;
+
+retry:
+	/* FDI is a binary signal running at ~2.7GHz, encoding
+	 * each output octet as 10 bits. The actual frequency
+	 * is stored as a divider into a 100MHz clock, and the
+	 * mode pixel clock is stored in units of 1KHz.
+	 * Hence the bw of each lane in terms of the mode signal
+	 * is:
+	 */
+	link_bw = intel_fdi_link_freq(i915, pipe_config);
+
+	fdi_dotclock = adjusted_mode->crtc_clock;
+
+	lane = ilk_get_lanes_required(fdi_dotclock, link_bw,
+				      pipe_config->pipe_bpp);
+
+	pipe_config->fdi_lanes = lane;
+
+	intel_link_compute_m_n(pipe_config->pipe_bpp, lane, fdi_dotclock,
+			       link_bw, &pipe_config->fdi_m_n, false, false);
+
+	ret = ilk_check_fdi_lanes(dev, intel_crtc->pipe, pipe_config);
+	if (ret == -EDEADLK)
+		return ret;
+
+	if (ret == -EINVAL && pipe_config->pipe_bpp > 6*3) {
+		pipe_config->pipe_bpp -= 2*3;
+		drm_dbg_kms(&i915->drm,
+			    "fdi link bw constraint, reducing pipe bpp to %i\n",
+			    pipe_config->pipe_bpp);
+		needs_recompute = true;
+		pipe_config->bw_constrained = true;
+
+		goto retry;
+	}
+
+	if (needs_recompute)
+		return I915_DISPLAY_CONFIG_RETRY;
+
+	return ret;
+}
+
+void intel_fdi_normal_train(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* enable normal train */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (IS_IVYBRIDGE(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_NONE_IVB;
+		temp |= FDI_LINK_TRAIN_NONE_IVB | FDI_TX_ENHANCE_FRAME_ENABLE;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_NONE | FDI_TX_ENHANCE_FRAME_ENABLE;
+	}
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_NORMAL_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_NONE;
+	}
+	intel_de_write(dev_priv, reg, temp | FDI_RX_ENHANCE_FRAME_ENABLE);
+
+	/* wait one idle pattern time */
+	intel_de_posting_read(dev_priv, reg);
+	udelay(1000);
+
+	/* IVB wants error correction enabled */
+	if (IS_IVYBRIDGE(dev_priv))
+		intel_de_write(dev_priv, reg,
+			       intel_de_read(dev_priv, reg) | FDI_FS_ERRC_ENABLE | FDI_FE_ERRC_ENABLE);
+}
+
+/* The FDI link training functions for ILK/Ibexpeak. */
+static void ilk_fdi_link_train(struct intel_crtc *crtc,
+			       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp, tries;
+
+	/* FDI needs bits from pipe first */
+	assert_pipe_enabled(dev_priv, crtc_state->cpu_transcoder);
+
+	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
+	   for train result */
+	reg = FDI_RX_IMR(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_RX_SYMBOL_LOCK;
+	temp &= ~FDI_RX_BIT_LOCK;
+	intel_de_write(dev_priv, reg, temp);
+	intel_de_read(dev_priv, reg);
+	udelay(150);
+
+	/* enable CPU FDI TX and PCH FDI RX */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	/* Ironlake workaround, enable clock pointer after FDI enable*/
+	intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
+		       FDI_RX_PHASE_SYNC_POINTER_OVR);
+	intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
+		       FDI_RX_PHASE_SYNC_POINTER_OVR | FDI_RX_PHASE_SYNC_POINTER_EN);
+
+	reg = FDI_RX_IIR(pipe);
+	for (tries = 0; tries < 5; tries++) {
+		temp = intel_de_read(dev_priv, reg);
+		drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+		if ((temp & FDI_RX_BIT_LOCK)) {
+			drm_dbg_kms(&dev_priv->drm, "FDI train 1 done.\n");
+			intel_de_write(dev_priv, reg, temp | FDI_RX_BIT_LOCK);
+			break;
+		}
+	}
+	if (tries == 5)
+		drm_err(&dev_priv->drm, "FDI train 1 fail!\n");
+
+	/* Train 2 */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_2;
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_2;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	reg = FDI_RX_IIR(pipe);
+	for (tries = 0; tries < 5; tries++) {
+		temp = intel_de_read(dev_priv, reg);
+		drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+		if (temp & FDI_RX_SYMBOL_LOCK) {
+			intel_de_write(dev_priv, reg,
+				       temp | FDI_RX_SYMBOL_LOCK);
+			drm_dbg_kms(&dev_priv->drm, "FDI train 2 done.\n");
+			break;
+		}
+	}
+	if (tries == 5)
+		drm_err(&dev_priv->drm, "FDI train 2 fail!\n");
+
+	drm_dbg_kms(&dev_priv->drm, "FDI train done\n");
+
+}
+
+static const int snb_b_fdi_train_param[] = {
+	FDI_LINK_TRAIN_400MV_0DB_SNB_B,
+	FDI_LINK_TRAIN_400MV_6DB_SNB_B,
+	FDI_LINK_TRAIN_600MV_3_5DB_SNB_B,
+	FDI_LINK_TRAIN_800MV_0DB_SNB_B,
+};
+
+/* The FDI link training functions for SNB/Cougarpoint. */
+static void gen6_fdi_link_train(struct intel_crtc *crtc,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp, i, retry;
+
+	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
+	   for train result */
+	reg = FDI_RX_IMR(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_RX_SYMBOL_LOCK;
+	temp &= ~FDI_RX_BIT_LOCK;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	/* enable CPU FDI TX and PCH FDI RX */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+	/* SNB-B */
+	temp |= FDI_LINK_TRAIN_400MV_0DB_SNB_B;
+	intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
+
+	intel_de_write(dev_priv, FDI_RX_MISC(pipe),
+		       FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_1;
+	}
+	intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	for (i = 0; i < 4; i++) {
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		temp |= snb_b_fdi_train_param[i];
+		intel_de_write(dev_priv, reg, temp);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(500);
+
+		for (retry = 0; retry < 5; retry++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+			if (temp & FDI_RX_BIT_LOCK) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_BIT_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 1 done.\n");
+				break;
+			}
+			udelay(50);
+		}
+		if (retry < 5)
+			break;
+	}
+	if (i == 4)
+		drm_err(&dev_priv->drm, "FDI train 1 fail!\n");
+
+	/* Train 2 */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_2;
+	if (IS_GEN(dev_priv, 6)) {
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		/* SNB-B */
+		temp |= FDI_LINK_TRAIN_400MV_0DB_SNB_B;
+	}
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_2_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_2;
+	}
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	for (i = 0; i < 4; i++) {
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		temp |= snb_b_fdi_train_param[i];
+		intel_de_write(dev_priv, reg, temp);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(500);
+
+		for (retry = 0; retry < 5; retry++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+			if (temp & FDI_RX_SYMBOL_LOCK) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_SYMBOL_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 2 done.\n");
+				break;
+			}
+			udelay(50);
+		}
+		if (retry < 5)
+			break;
+	}
+	if (i == 4)
+		drm_err(&dev_priv->drm, "FDI train 2 fail!\n");
+
+	drm_dbg_kms(&dev_priv->drm, "FDI train done.\n");
+}
+
+/* Manual link training for Ivy Bridge A0 parts */
+static void ivb_manual_fdi_link_train(struct intel_crtc *crtc,
+				      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp, i, j;
+
+	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
+	   for train result */
+	reg = FDI_RX_IMR(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_RX_SYMBOL_LOCK;
+	temp &= ~FDI_RX_BIT_LOCK;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR before link train 0x%x\n",
+		    intel_de_read(dev_priv, FDI_RX_IIR(pipe)));
+
+	/* Try each vswing and preemphasis setting twice before moving on */
+	for (j = 0; j < ARRAY_SIZE(snb_b_fdi_train_param) * 2; j++) {
+		/* disable first in case we need to retry */
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~(FDI_LINK_TRAIN_AUTO | FDI_LINK_TRAIN_NONE_IVB);
+		temp &= ~FDI_TX_ENABLE;
+		intel_de_write(dev_priv, reg, temp);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_AUTO;
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp &= ~FDI_RX_ENABLE;
+		intel_de_write(dev_priv, reg, temp);
+
+		/* enable CPU FDI TX and PCH FDI RX */
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_DP_PORT_WIDTH_MASK;
+		temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+		temp |= FDI_LINK_TRAIN_PATTERN_1_IVB;
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		temp |= snb_b_fdi_train_param[j/2];
+		temp |= FDI_COMPOSITE_SYNC;
+		intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
+
+		intel_de_write(dev_priv, FDI_RX_MISC(pipe),
+			       FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+		temp |= FDI_COMPOSITE_SYNC;
+		intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(1); /* should be 0.5us */
+
+		for (i = 0; i < 4; i++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+			if (temp & FDI_RX_BIT_LOCK ||
+			    (intel_de_read(dev_priv, reg) & FDI_RX_BIT_LOCK)) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_BIT_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 1 done, level %i.\n",
+					    i);
+				break;
+			}
+			udelay(1); /* should be 0.5us */
+		}
+		if (i == 4) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "FDI train 1 fail on vswing %d\n", j / 2);
+			continue;
+		}
+
+		/* Train 2 */
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_NONE_IVB;
+		temp |= FDI_LINK_TRAIN_PATTERN_2_IVB;
+		intel_de_write(dev_priv, reg, temp);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_2_CPT;
+		intel_de_write(dev_priv, reg, temp);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(2); /* should be 1.5us */
+
+		for (i = 0; i < 4; i++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+			if (temp & FDI_RX_SYMBOL_LOCK ||
+			    (intel_de_read(dev_priv, reg) & FDI_RX_SYMBOL_LOCK)) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_SYMBOL_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 2 done, level %i.\n",
+					    i);
+				goto train_done;
+			}
+			udelay(2); /* should be 1.5us */
+		}
+		if (i == 4)
+			drm_dbg_kms(&dev_priv->drm,
+				    "FDI train 2 fail on vswing %d\n", j / 2);
+	}
+
+train_done:
+	drm_dbg_kms(&dev_priv->drm, "FDI train done.\n");
+}
+
+void ilk_fdi_pll_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
+	enum pipe pipe = intel_crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* enable PCH FDI RX PLL, wait warmup plus DMI latency */
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~(FDI_DP_PORT_WIDTH_MASK | (0x7 << 16));
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
+	intel_de_write(dev_priv, reg, temp | FDI_RX_PLL_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(200);
+
+	/* Switch from Rawclk to PCDclk */
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp | FDI_PCDCLK);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(200);
+
+	/* Enable CPU FDI TX PLL, always on for Ironlake */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if ((temp & FDI_TX_PLL_ENABLE) == 0) {
+		intel_de_write(dev_priv, reg, temp | FDI_TX_PLL_ENABLE);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(100);
+	}
+}
+
+void ilk_fdi_pll_disable(struct intel_crtc *intel_crtc)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = intel_crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* Switch from PCDclk to Rawclk */
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_PCDCLK);
+
+	/* Disable CPU FDI TX PLL */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_TX_PLL_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_RX_PLL_ENABLE);
+
+	/* Wait for the clocks to turn off. */
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+}
+
+void ilk_fdi_disable(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* disable CPU FDI tx and PCH FDI rx */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_TX_ENABLE);
+	intel_de_posting_read(dev_priv, reg);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~(0x7 << 16);
+	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
+	intel_de_write(dev_priv, reg, temp & ~FDI_RX_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+
+	/* Ironlake workaround, disable clock pointer after downing FDI */
+	if (HAS_PCH_IBX(dev_priv))
+		intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
+			       FDI_RX_PHASE_SYNC_POINTER_OVR);
+
+	/* still set train pattern 1 */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_1;
+	}
+	/* BPC in FDI rx is consistent with that in PIPECONF */
+	temp &= ~(0x07 << 16);
+	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+}
+
+void
+intel_fdi_init_hook(struct drm_i915_private *dev_priv)
+{
+	if (IS_GEN(dev_priv, 5)) {
+		dev_priv->display.fdi_link_train = ilk_fdi_link_train;
+	} else if (IS_GEN(dev_priv, 6)) {
+		dev_priv->display.fdi_link_train = gen6_fdi_link_train;
+	} else if (IS_IVYBRIDGE(dev_priv)) {
+		/* FIXME: detect B0+ stepping and use auto training */
+		dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fdi.h b/drivers/gpu/drm/i915/display/intel_fdi.h
new file mode 100644
index 000000000000..a9cd21663eb8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fdi.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_FDI_H_
+#define _INTEL_FDI_H_
+
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_crtc_state;
+
+#define I915_DISPLAY_CONFIG_RETRY 1
+int ilk_fdi_compute_config(struct intel_crtc *intel_crtc,
+			   struct intel_crtc_state *pipe_config);
+void intel_fdi_normal_train(struct intel_crtc *crtc);
+void ilk_fdi_disable(struct intel_crtc *crtc);
+void ilk_fdi_pll_disable(struct intel_crtc *intel_crtc);
+void ilk_fdi_pll_enable(const struct intel_crtc_state *crtc_state);
+void intel_fdi_init_hook(struct drm_i915_private *dev_priv);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.c b/drivers/gpu/drm/i915/display/intel_frontbuffer.c
index d898b370d7a4..7b38eee9980f 100644
--- a/drivers/gpu/drm/i915/display/intel_frontbuffer.c
+++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c
@@ -225,8 +225,10 @@ static void frontbuffer_release(struct kref *ref)
 	struct i915_vma *vma;
 
 	spin_lock(&obj->vma.lock);
-	for_each_ggtt_vma(vma, obj)
+	for_each_ggtt_vma(vma, obj) {
+		i915_vma_clear_scanout(vma);
 		vma->display_alignment = I915_GTT_MIN_ALIGNMENT;
+	}
 	spin_unlock(&obj->vma.lock);
 
 	RCU_INIT_POINTER(obj->frontbuffer, NULL);
diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c
index b2a4bbcfdcd2..ae1371c36a32 100644
--- a/drivers/gpu/drm/i915/display/intel_hdcp.c
+++ b/drivers/gpu/drm/i915/display/intel_hdcp.c
@@ -15,6 +15,7 @@
 #include <drm/drm_hdcp.h>
 #include <drm/i915_component.h>
 
+#include "i915_drv.h"
 #include "i915_reg.h"
 #include "intel_display_power.h"
 #include "intel_display_types.h"
@@ -23,9 +24,78 @@
 #include "intel_connector.h"
 
 #define KEY_LOAD_TRIES	5
-#define ENCRYPT_STATUS_CHANGE_TIMEOUT_MS	50
 #define HDCP2_LC_RETRY_CNT			3
 
+static int intel_conn_to_vcpi(struct intel_connector *connector)
+{
+	/* For HDMI this is forced to be 0x0. For DP SST also this is 0x0. */
+	return connector->port	? connector->port->vcpi.vcpi : 0;
+}
+
+/*
+ * intel_hdcp_required_content_stream selects the most highest common possible HDCP
+ * content_type for all streams in DP MST topology because security f/w doesn't
+ * have any provision to mark content_type for each stream separately, it marks
+ * all available streams with the content_type proivided at the time of port
+ * authentication. This may prohibit the userspace to use type1 content on
+ * HDCP 2.2 capable sink because of other sink are not capable of HDCP 2.2 in
+ * DP MST topology. Though it is not compulsory, security fw should change its
+ * policy to mark different content_types for different streams.
+ */
+static int
+intel_hdcp_required_content_stream(struct intel_digital_port *dig_port)
+{
+	struct drm_connector_list_iter conn_iter;
+	struct intel_digital_port *conn_dig_port;
+	struct intel_connector *connector;
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	bool enforce_type0 = false;
+	int k;
+
+	data->k = 0;
+
+	if (dig_port->hdcp_auth_status)
+		return 0;
+
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		if (connector->base.status == connector_status_disconnected)
+			continue;
+
+		if (!intel_encoder_is_mst(intel_attached_encoder(connector)))
+			continue;
+
+		conn_dig_port = intel_attached_dig_port(connector);
+		if (conn_dig_port != dig_port)
+			continue;
+
+		if (!enforce_type0 && !intel_hdcp2_capable(connector))
+			enforce_type0 = true;
+
+		data->streams[data->k].stream_id = intel_conn_to_vcpi(connector);
+		data->k++;
+
+		/* if there is only one active stream */
+		if (dig_port->dp.active_mst_links <= 1)
+			break;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	if (drm_WARN_ON(&i915->drm, data->k > INTEL_NUM_PIPES(i915) || data->k == 0))
+		return -EINVAL;
+
+	/*
+	 * Apply common protection level across all streams in DP MST Topology.
+	 * Use highest supported content type for all streams in DP MST Topology.
+	 */
+	for (k = 0; k < data->k; k++)
+		data->streams[k].stream_type =
+			enforce_type0 ? DRM_MODE_HDCP_CONTENT_TYPE0 : DRM_MODE_HDCP_CONTENT_TYPE1;
+
+	return 0;
+}
+
 static
 bool intel_hdcp_is_ksv_valid(u8 *ksv)
 {
@@ -762,15 +832,22 @@ static int intel_hdcp_auth(struct intel_connector *connector)
 	if (intel_de_wait_for_set(dev_priv,
 				  HDCP_STATUS(dev_priv, cpu_transcoder, port),
 				  HDCP_STATUS_ENC,
-				  ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+				  HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
 		drm_err(&dev_priv->drm, "Timed out waiting for encryption\n");
 		return -ETIMEDOUT;
 	}
 
-	/*
-	 * XXX: If we have MST-connected devices, we need to enable encryption
-	 * on those as well.
-	 */
+	/* DP MST Auth Part 1 Step 2.a and Step 2.b */
+	if (shim->stream_encryption) {
+		ret = shim->stream_encryption(connector, true);
+		if (ret) {
+			drm_err(&dev_priv->drm, "[%s:%d] Failed to enable HDCP 1.4 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&dev_priv->drm, "HDCP 1.4 transcoder: %s stream encrypted\n",
+			    transcoder_name(hdcp->stream_transcoder));
+	}
 
 	if (repeater_present)
 		return intel_hdcp_auth_downstream(connector);
@@ -792,24 +869,29 @@ static int _intel_hdcp_disable(struct intel_connector *connector)
 	drm_dbg_kms(&dev_priv->drm, "[%s:%d] HDCP is being disabled...\n",
 		    connector->base.name, connector->base.base.id);
 
-	/*
-	 * If there are other connectors on this port using HDCP, don't disable
-	 * it. Instead, toggle the HDCP signalling off on that particular
-	 * connector/pipe and exit.
-	 */
-	if (dig_port->num_hdcp_streams > 0) {
-		ret = hdcp->shim->toggle_signalling(dig_port,
-						    cpu_transcoder, false);
-		if (ret)
-			DRM_ERROR("Failed to disable HDCP signalling\n");
-		return ret;
+	if (hdcp->shim->stream_encryption) {
+		ret = hdcp->shim->stream_encryption(connector, false);
+		if (ret) {
+			drm_err(&dev_priv->drm, "[%s:%d] Failed to disable HDCP 1.4 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&dev_priv->drm, "HDCP 1.4 transcoder: %s stream encryption disabled\n",
+			    transcoder_name(hdcp->stream_transcoder));
+		/*
+		 * If there are other connectors on this port using HDCP,
+		 * don't disable it until it disabled HDCP encryption for
+		 * all connectors in MST topology.
+		 */
+		if (dig_port->num_hdcp_streams > 0)
+			return 0;
 	}
 
 	hdcp->hdcp_encrypted = false;
 	intel_de_write(dev_priv, HDCP_CONF(dev_priv, cpu_transcoder, port), 0);
 	if (intel_de_wait_for_clear(dev_priv,
 				    HDCP_STATUS(dev_priv, cpu_transcoder, port),
-				    ~0, ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+				    ~0, HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
 		drm_err(&dev_priv->drm,
 			"Failed to disable HDCP, timeout clearing status\n");
 		return -ETIMEDOUT;
@@ -1014,7 +1096,8 @@ static int
 hdcp2_prepare_ake_init(struct intel_connector *connector,
 		       struct hdcp2_ake_init *ake_data)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1043,7 +1126,8 @@ hdcp2_verify_rx_cert_prepare_km(struct intel_connector *connector,
 				struct hdcp2_ake_no_stored_km *ek_pub_km,
 				size_t *msg_sz)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1070,7 +1154,8 @@ hdcp2_verify_rx_cert_prepare_km(struct intel_connector *connector,
 static int hdcp2_verify_hprime(struct intel_connector *connector,
 			       struct hdcp2_ake_send_hprime *rx_hprime)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1095,7 +1180,8 @@ static int
 hdcp2_store_pairing_info(struct intel_connector *connector,
 			 struct hdcp2_ake_send_pairing_info *pairing_info)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1121,7 +1207,8 @@ static int
 hdcp2_prepare_lc_init(struct intel_connector *connector,
 		      struct hdcp2_lc_init *lc_init)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1147,7 +1234,8 @@ static int
 hdcp2_verify_lprime(struct intel_connector *connector,
 		    struct hdcp2_lc_send_lprime *rx_lprime)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1172,7 +1260,8 @@ hdcp2_verify_lprime(struct intel_connector *connector,
 static int hdcp2_prepare_skey(struct intel_connector *connector,
 			      struct hdcp2_ske_send_eks *ske_data)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1200,7 +1289,8 @@ hdcp2_verify_rep_topology_prepare_ack(struct intel_connector *connector,
 								*rep_topology,
 				      struct hdcp2_rep_send_ack *rep_send_ack)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1228,7 +1318,8 @@ static int
 hdcp2_verify_mprime(struct intel_connector *connector,
 		    struct hdcp2_rep_stream_ready *stream_ready)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1251,7 +1342,8 @@ hdcp2_verify_mprime(struct intel_connector *connector,
 
 static int hdcp2_authenticate_port(struct intel_connector *connector)
 {
-	struct hdcp_port_data *data = &connector->hdcp.port_data;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1275,6 +1367,7 @@ static int hdcp2_authenticate_port(struct intel_connector *connector)
 
 static int hdcp2_close_mei_session(struct intel_connector *connector)
 {
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct i915_hdcp_comp_master *comp;
 	int ret;
@@ -1288,7 +1381,7 @@ static int hdcp2_close_mei_session(struct intel_connector *connector)
 	}
 
 	ret = comp->ops->close_hdcp_session(comp->mei_dev,
-					     &connector->hdcp.port_data);
+					     &dig_port->hdcp_port_data);
 	mutex_unlock(&dev_priv->hdcp_comp_mutex);
 
 	return ret;
@@ -1448,13 +1541,14 @@ static
 int _hdcp2_propagate_stream_management_info(struct intel_connector *connector)
 {
 	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct intel_hdcp *hdcp = &connector->hdcp;
 	union {
 		struct hdcp2_rep_stream_manage stream_manage;
 		struct hdcp2_rep_stream_ready stream_ready;
 	} msgs;
 	const struct intel_hdcp_shim *shim = hdcp->shim;
-	int ret;
+	int ret, streams_size_delta, i;
 
 	if (connector->hdcp.seq_num_m > HDCP_2_2_SEQ_NUM_MAX)
 		return -ERANGE;
@@ -1463,16 +1557,18 @@ int _hdcp2_propagate_stream_management_info(struct intel_connector *connector)
 	msgs.stream_manage.msg_id = HDCP_2_2_REP_STREAM_MANAGE;
 	drm_hdcp_cpu_to_be24(msgs.stream_manage.seq_num_m, hdcp->seq_num_m);
 
-	/* K no of streams is fixed as 1. Stored as big-endian. */
-	msgs.stream_manage.k = cpu_to_be16(1);
+	msgs.stream_manage.k = cpu_to_be16(data->k);
 
-	/* For HDMI this is forced to be 0x0. For DP SST also this is 0x0. */
-	msgs.stream_manage.streams[0].stream_id = 0;
-	msgs.stream_manage.streams[0].stream_type = hdcp->content_type;
+	for (i = 0; i < data->k; i++) {
+		msgs.stream_manage.streams[i].stream_id = data->streams[i].stream_id;
+		msgs.stream_manage.streams[i].stream_type = data->streams[i].stream_type;
+	}
 
+	streams_size_delta = (HDCP_2_2_MAX_CONTENT_STREAMS_CNT - data->k) *
+				sizeof(struct hdcp2_streamid_type);
 	/* Send it to Repeater */
 	ret = shim->write_2_2_msg(dig_port, &msgs.stream_manage,
-				  sizeof(msgs.stream_manage));
+				  sizeof(msgs.stream_manage) - streams_size_delta);
 	if (ret < 0)
 		goto out;
 
@@ -1481,8 +1577,8 @@ int _hdcp2_propagate_stream_management_info(struct intel_connector *connector)
 	if (ret < 0)
 		goto out;
 
-	hdcp->port_data.seq_num_m = hdcp->seq_num_m;
-	hdcp->port_data.streams[0].stream_type = hdcp->content_type;
+	data->seq_num_m = hdcp->seq_num_m;
+
 	ret = hdcp2_verify_mprime(connector, &msgs.stream_ready);
 
 out:
@@ -1606,6 +1702,36 @@ static int hdcp2_authenticate_sink(struct intel_connector *connector)
 	return ret;
 }
 
+static int hdcp2_enable_stream_encryption(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum transcoder cpu_transcoder = hdcp->cpu_transcoder;
+	enum port port = dig_port->base.port;
+	int ret = 0;
+
+	if (!(intel_de_read(dev_priv, HDCP2_STATUS(dev_priv, cpu_transcoder, port)) &
+			    LINK_ENCRYPTION_STATUS)) {
+		drm_err(&dev_priv->drm, "[%s:%d] HDCP 2.2 Link is not encrypted\n",
+			connector->base.name, connector->base.base.id);
+		return -EPERM;
+	}
+
+	if (hdcp->shim->stream_2_2_encryption) {
+		ret = hdcp->shim->stream_2_2_encryption(connector, true);
+		if (ret) {
+			drm_err(&dev_priv->drm, "[%s:%d] Failed to enable HDCP 2.2 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&dev_priv->drm, "HDCP 2.2 transcoder: %s stream encrypted\n",
+			    transcoder_name(hdcp->stream_transcoder));
+	}
+
+	return ret;
+}
+
 static int hdcp2_enable_encryption(struct intel_connector *connector)
 {
 	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
@@ -1641,7 +1767,8 @@ static int hdcp2_enable_encryption(struct intel_connector *connector)
 				    HDCP2_STATUS(dev_priv, cpu_transcoder,
 						 port),
 				    LINK_ENCRYPTION_STATUS,
-				    ENCRYPT_STATUS_CHANGE_TIMEOUT_MS);
+				    HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS);
+	dig_port->hdcp_auth_status = true;
 
 	return ret;
 }
@@ -1665,7 +1792,7 @@ static int hdcp2_disable_encryption(struct intel_connector *connector)
 				      HDCP2_STATUS(dev_priv, cpu_transcoder,
 						   port),
 				      LINK_ENCRYPTION_STATUS,
-				      ENCRYPT_STATUS_CHANGE_TIMEOUT_MS);
+				      HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS);
 	if (ret == -ETIMEDOUT)
 		drm_dbg_kms(&dev_priv->drm, "Disable Encryption Timedout");
 
@@ -1714,11 +1841,11 @@ hdcp2_propagate_stream_management_info(struct intel_connector *connector)
 
 static int hdcp2_authenticate_and_encrypt(struct intel_connector *connector)
 {
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
 	struct drm_i915_private *i915 = to_i915(connector->base.dev);
-	struct intel_hdcp *hdcp = &connector->hdcp;
-	int ret, i, tries = 3;
+	int ret = 0, i, tries = 3;
 
-	for (i = 0; i < tries; i++) {
+	for (i = 0; i < tries && !dig_port->hdcp_auth_status; i++) {
 		ret = hdcp2_authenticate_sink(connector);
 		if (!ret) {
 			ret = hdcp2_propagate_stream_management_info(connector);
@@ -1728,8 +1855,7 @@ static int hdcp2_authenticate_and_encrypt(struct intel_connector *connector)
 					    ret);
 				break;
 			}
-			hdcp->port_data.streams[0].stream_type =
-							hdcp->content_type;
+
 			ret = hdcp2_authenticate_port(connector);
 			if (!ret)
 				break;
@@ -1744,7 +1870,7 @@ static int hdcp2_authenticate_and_encrypt(struct intel_connector *connector)
 			drm_dbg_kms(&i915->drm, "Port deauth failed.\n");
 	}
 
-	if (!ret) {
+	if (!ret && !dig_port->hdcp_auth_status) {
 		/*
 		 * Ensuring the required 200mSec min time interval between
 		 * Session Key Exchange and encryption.
@@ -1759,12 +1885,16 @@ static int hdcp2_authenticate_and_encrypt(struct intel_connector *connector)
 		}
 	}
 
+	ret = hdcp2_enable_stream_encryption(connector);
+
 	return ret;
 }
 
 static int _intel_hdcp2_enable(struct intel_connector *connector)
 {
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
 	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct intel_hdcp *hdcp = &connector->hdcp;
 	int ret;
 
@@ -1772,6 +1902,16 @@ static int _intel_hdcp2_enable(struct intel_connector *connector)
 		    connector->base.name, connector->base.base.id,
 		    hdcp->content_type);
 
+	/* Stream which requires encryption */
+	if (!intel_encoder_is_mst(intel_attached_encoder(connector))) {
+		data->k = 1;
+		data->streams[0].stream_type = hdcp->content_type;
+	} else {
+		ret = intel_hdcp_required_content_stream(dig_port);
+		if (ret)
+			return ret;
+	}
+
 	ret = hdcp2_authenticate_and_encrypt(connector);
 	if (ret) {
 		drm_dbg_kms(&i915->drm, "HDCP2 Type%d  Enabling Failed. (%d)\n",
@@ -1789,18 +1929,37 @@ static int _intel_hdcp2_enable(struct intel_connector *connector)
 
 static int _intel_hdcp2_disable(struct intel_connector *connector)
 {
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
 	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
 	int ret;
 
 	drm_dbg_kms(&i915->drm, "[%s:%d] HDCP2.2 is being Disabled\n",
 		    connector->base.name, connector->base.base.id);
 
+	if (hdcp->shim->stream_2_2_encryption) {
+		ret = hdcp->shim->stream_2_2_encryption(connector, false);
+		if (ret) {
+			drm_err(&i915->drm, "[%s:%d] Failed to disable HDCP 2.2 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&i915->drm, "HDCP 2.2 transcoder: %s stream encryption disabled\n",
+			    transcoder_name(hdcp->stream_transcoder));
+
+		if (dig_port->num_hdcp_streams > 0)
+			return 0;
+	}
+
 	ret = hdcp2_disable_encryption(connector);
 
 	if (hdcp2_deauthenticate_port(connector) < 0)
 		drm_dbg_kms(&i915->drm, "Port deauth failed.\n");
 
 	connector->hdcp.hdcp2_encrypted = false;
+	dig_port->hdcp_auth_status = false;
+	data->k = 0;
 
 	return ret;
 }
@@ -1816,6 +1975,7 @@ static int intel_hdcp2_check_link(struct intel_connector *connector)
 	int ret = 0;
 
 	mutex_lock(&hdcp->mutex);
+	mutex_lock(&dig_port->hdcp_mutex);
 	cpu_transcoder = hdcp->cpu_transcoder;
 
 	/* hdcp2_check_link is expected only when HDCP2.2 is Enabled */
@@ -1837,7 +1997,7 @@ static int intel_hdcp2_check_link(struct intel_connector *connector)
 		goto out;
 	}
 
-	ret = hdcp->shim->check_2_2_link(dig_port);
+	ret = hdcp->shim->check_2_2_link(dig_port, connector);
 	if (ret == HDCP_LINK_PROTECTED) {
 		if (hdcp->value != DRM_MODE_CONTENT_PROTECTION_UNDESIRED) {
 			intel_hdcp_update_value(connector,
@@ -1893,6 +2053,7 @@ static int intel_hdcp2_check_link(struct intel_connector *connector)
 	}
 
 out:
+	mutex_unlock(&dig_port->hdcp_mutex);
 	mutex_unlock(&hdcp->mutex);
 	return ret;
 }
@@ -1968,12 +2129,13 @@ static enum mei_fw_tc intel_get_mei_fw_tc(enum transcoder cpu_transcoder)
 }
 
 static int initialize_hdcp_port_data(struct intel_connector *connector,
-				     enum port port,
+				     struct intel_digital_port *dig_port,
 				     const struct intel_hdcp_shim *shim)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
 	struct intel_hdcp *hdcp = &connector->hdcp;
-	struct hdcp_port_data *data = &hdcp->port_data;
+	enum port port = dig_port->base.port;
 
 	if (INTEL_GEN(dev_priv) < 12)
 		data->fw_ddi = intel_get_mei_fw_ddi_index(port);
@@ -1994,16 +2156,15 @@ static int initialize_hdcp_port_data(struct intel_connector *connector,
 	data->port_type = (u8)HDCP_PORT_TYPE_INTEGRATED;
 	data->protocol = (u8)shim->protocol;
 
-	data->k = 1;
 	if (!data->streams)
-		data->streams = kcalloc(data->k,
+		data->streams = kcalloc(INTEL_NUM_PIPES(dev_priv),
 					sizeof(struct hdcp2_streamid_type),
 					GFP_KERNEL);
 	if (!data->streams) {
 		drm_err(&dev_priv->drm, "Out of Memory\n");
 		return -ENOMEM;
 	}
-
+	/* For SST */
 	data->streams[0].stream_id = 0;
 	data->streams[0].stream_type = hdcp->content_type;
 
@@ -2046,14 +2207,15 @@ void intel_hdcp_component_init(struct drm_i915_private *dev_priv)
 	}
 }
 
-static void intel_hdcp2_init(struct intel_connector *connector, enum port port,
+static void intel_hdcp2_init(struct intel_connector *connector,
+			     struct intel_digital_port *dig_port,
 			     const struct intel_hdcp_shim *shim)
 {
 	struct drm_i915_private *i915 = to_i915(connector->base.dev);
 	struct intel_hdcp *hdcp = &connector->hdcp;
 	int ret;
 
-	ret = initialize_hdcp_port_data(connector, port, shim);
+	ret = initialize_hdcp_port_data(connector, dig_port, shim);
 	if (ret) {
 		drm_dbg_kms(&i915->drm, "Mei hdcp data init failed\n");
 		return;
@@ -2063,7 +2225,7 @@ static void intel_hdcp2_init(struct intel_connector *connector, enum port port,
 }
 
 int intel_hdcp_init(struct intel_connector *connector,
-		    enum port port,
+		    struct intel_digital_port *dig_port,
 		    const struct intel_hdcp_shim *shim)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
@@ -2073,15 +2235,15 @@ int intel_hdcp_init(struct intel_connector *connector,
 	if (!shim)
 		return -EINVAL;
 
-	if (is_hdcp2_supported(dev_priv) && !connector->mst_port)
-		intel_hdcp2_init(connector, port, shim);
+	if (is_hdcp2_supported(dev_priv))
+		intel_hdcp2_init(connector, dig_port, shim);
 
 	ret =
 	drm_connector_attach_content_protection_property(&connector->base,
 							 hdcp->hdcp2_supported);
 	if (ret) {
 		hdcp->hdcp2_supported = false;
-		kfree(hdcp->port_data.streams);
+		kfree(dig_port->hdcp_port_data.streams);
 		return ret;
 	}
 
@@ -2095,7 +2257,7 @@ int intel_hdcp_init(struct intel_connector *connector,
 }
 
 int intel_hdcp_enable(struct intel_connector *connector,
-		      enum transcoder cpu_transcoder, u8 content_type)
+		      const struct intel_crtc_state *pipe_config, u8 content_type)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
@@ -2106,15 +2268,28 @@ int intel_hdcp_enable(struct intel_connector *connector,
 	if (!hdcp->shim)
 		return -ENOENT;
 
+	if (!connector->encoder) {
+		drm_err(&dev_priv->drm, "[%s:%d] encoder is not initialized\n",
+			connector->base.name, connector->base.base.id);
+		return -ENODEV;
+	}
+
 	mutex_lock(&hdcp->mutex);
 	mutex_lock(&dig_port->hdcp_mutex);
 	drm_WARN_ON(&dev_priv->drm,
 		    hdcp->value == DRM_MODE_CONTENT_PROTECTION_ENABLED);
 	hdcp->content_type = content_type;
-	hdcp->cpu_transcoder = cpu_transcoder;
+
+	if (intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DP_MST)) {
+		hdcp->cpu_transcoder = pipe_config->mst_master_transcoder;
+		hdcp->stream_transcoder = pipe_config->cpu_transcoder;
+	} else {
+		hdcp->cpu_transcoder = pipe_config->cpu_transcoder;
+		hdcp->stream_transcoder = INVALID_TRANSCODER;
+	}
 
 	if (INTEL_GEN(dev_priv) >= 12)
-		hdcp->port_data.fw_tc = intel_get_mei_fw_tc(cpu_transcoder);
+		dig_port->hdcp_port_data.fw_tc = intel_get_mei_fw_tc(hdcp->cpu_transcoder);
 
 	/*
 	 * Considering that HDCP2.2 is more secure than HDCP1.4, If the setup
@@ -2210,6 +2385,7 @@ void intel_hdcp_update_pipe(struct intel_atomic_state *state,
 	if (content_protection_type_changed) {
 		mutex_lock(&hdcp->mutex);
 		hdcp->value = DRM_MODE_CONTENT_PROTECTION_DESIRED;
+		drm_connector_get(&connector->base);
 		schedule_work(&hdcp->prop_work);
 		mutex_unlock(&hdcp->mutex);
 	}
@@ -2221,11 +2397,19 @@ void intel_hdcp_update_pipe(struct intel_atomic_state *state,
 		desired_and_not_enabled =
 			hdcp->value != DRM_MODE_CONTENT_PROTECTION_ENABLED;
 		mutex_unlock(&hdcp->mutex);
+		/*
+		 * If HDCP already ENABLED and CP property is DESIRED, schedule
+		 * prop_work to update correct CP property to user space.
+		 */
+		if (!desired_and_not_enabled && !content_protection_type_changed) {
+			drm_connector_get(&connector->base);
+			schedule_work(&hdcp->prop_work);
+		}
 	}
 
 	if (desired_and_not_enabled || content_protection_type_changed)
 		intel_hdcp_enable(connector,
-				  crtc_state->cpu_transcoder,
+				  crtc_state,
 				  (u8)conn_state->hdcp_content_type);
 }
 
@@ -2275,7 +2459,6 @@ void intel_hdcp_cleanup(struct intel_connector *connector)
 	drm_WARN_ON(connector->base.dev, work_pending(&hdcp->prop_work));
 
 	mutex_lock(&hdcp->mutex);
-	kfree(hdcp->port_data.streams);
 	hdcp->shim = NULL;
 	mutex_unlock(&hdcp->mutex);
 }
diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.h b/drivers/gpu/drm/i915/display/intel_hdcp.h
index 1bbf5b67ed0a..8f53b0c7fe5c 100644
--- a/drivers/gpu/drm/i915/display/intel_hdcp.h
+++ b/drivers/gpu/drm/i915/display/intel_hdcp.h
@@ -8,6 +8,8 @@
 
 #include <linux/types.h>
 
+#define HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS	50
+
 struct drm_connector;
 struct drm_connector_state;
 struct drm_i915_private;
@@ -16,16 +18,18 @@ struct intel_connector;
 struct intel_crtc_state;
 struct intel_encoder;
 struct intel_hdcp_shim;
+struct intel_digital_port;
 enum port;
 enum transcoder;
 
 void intel_hdcp_atomic_check(struct drm_connector *connector,
 			     struct drm_connector_state *old_state,
 			     struct drm_connector_state *new_state);
-int intel_hdcp_init(struct intel_connector *connector, enum port port,
+int intel_hdcp_init(struct intel_connector *connector,
+		    struct intel_digital_port *dig_port,
 		    const struct intel_hdcp_shim *hdcp_shim);
 int intel_hdcp_enable(struct intel_connector *connector,
-		      enum transcoder cpu_transcoder, u8 content_type);
+		      const struct intel_crtc_state *pipe_config, u8 content_type);
 int intel_hdcp_disable(struct intel_connector *connector);
 void intel_hdcp_update_pipe(struct intel_atomic_state *state,
 			    struct intel_encoder *encoder,
diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c
index c5959590562b..95919d325b0b 100644
--- a/drivers/gpu/drm/i915/display/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/display/intel_hdmi.c
@@ -1494,15 +1494,16 @@ static int kbl_repositioning_enc_en_signal(struct intel_connector *connector,
 		usleep_range(25, 50);
 	}
 
-	ret = intel_ddi_toggle_hdcp_signalling(&dig_port->base, cpu_transcoder,
-					       false);
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base, cpu_transcoder,
+					 false, TRANS_DDI_HDCP_SIGNALLING);
 	if (ret) {
 		drm_err(&dev_priv->drm,
 			"Disable HDCP signalling failed (%d)\n", ret);
 		return ret;
 	}
-	ret = intel_ddi_toggle_hdcp_signalling(&dig_port->base, cpu_transcoder,
-					       true);
+
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base, cpu_transcoder,
+					 true, TRANS_DDI_HDCP_SIGNALLING);
 	if (ret) {
 		drm_err(&dev_priv->drm,
 			"Enable HDCP signalling failed (%d)\n", ret);
@@ -1525,8 +1526,9 @@ int intel_hdmi_hdcp_toggle_signalling(struct intel_digital_port *dig_port,
 	if (!enable)
 		usleep_range(6, 60); /* Bspec says >= 6us */
 
-	ret = intel_ddi_toggle_hdcp_signalling(&dig_port->base, cpu_transcoder,
-					       enable);
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base,
+					 cpu_transcoder, enable,
+					 TRANS_DDI_HDCP_SIGNALLING);
 	if (ret) {
 		drm_err(&dev_priv->drm, "%s HDCP signalling failed (%d)\n",
 			enable ? "Enable" : "Disable", ret);
@@ -1731,7 +1733,8 @@ int intel_hdmi_hdcp2_read_msg(struct intel_digital_port *dig_port,
 }
 
 static
-int intel_hdmi_hdcp2_check_link(struct intel_digital_port *dig_port)
+int intel_hdmi_hdcp2_check_link(struct intel_digital_port *dig_port,
+				struct intel_connector *connector)
 {
 	u8 rx_status[HDCP_2_2_HDMI_RXSTATUS_LEN];
 	int ret;
@@ -2215,7 +2218,11 @@ hdmi_port_clock_valid(struct intel_hdmi *hdmi,
 					  has_hdmi_sink))
 		return MODE_CLOCK_HIGH;
 
-	/* BXT DPLL can't generate 223-240 MHz */
+	/* GLK DPLL can't generate 446-480 MHz */
+	if (IS_GEMINILAKE(dev_priv) && clock > 446666 && clock < 480000)
+		return MODE_CLOCK_RANGE;
+
+	/* BXT/GLK DPLL can't generate 223-240 MHz */
 	if (IS_GEN9_LP(dev_priv) && clock > 223333 && clock < 240000)
 		return MODE_CLOCK_RANGE;
 
@@ -3290,7 +3297,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *dig_port,
 	intel_hdmi->attached_connector = intel_connector;
 
 	if (is_hdcp_supported(dev_priv, port)) {
-		int ret = intel_hdcp_init(intel_connector, port,
+		int ret = intel_hdcp_init(intel_connector, dig_port,
 					  &intel_hdmi_hdcp_shim);
 		if (ret)
 			drm_dbg_kms(&dev_priv->drm,
diff --git a/drivers/gpu/drm/i915/display/intel_overlay.c b/drivers/gpu/drm/i915/display/intel_overlay.c
index 52b4f6193b4c..f455040fa989 100644
--- a/drivers/gpu/drm/i915/display/intel_overlay.c
+++ b/drivers/gpu/drm/i915/display/intel_overlay.c
@@ -29,6 +29,7 @@
 #include <drm/drm_fourcc.h>
 
 #include "gem/i915_gem_pm.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_ring.h"
 
 #include "i915_drv.h"
@@ -182,6 +183,7 @@ struct intel_overlay {
 	struct intel_crtc *crtc;
 	struct i915_vma *vma;
 	struct i915_vma *old_vma;
+	struct intel_frontbuffer *frontbuffer;
 	bool active;
 	bool pfit_active;
 	u32 pfit_vscale_ratio; /* shifted-point number, (1<<12) == 1.0 */
@@ -282,21 +284,19 @@ static void intel_overlay_flip_prepare(struct intel_overlay *overlay,
 				       struct i915_vma *vma)
 {
 	enum pipe pipe = overlay->crtc->pipe;
-	struct intel_frontbuffer *from = NULL, *to = NULL;
+	struct intel_frontbuffer *frontbuffer = NULL;
 
 	drm_WARN_ON(&overlay->i915->drm, overlay->old_vma);
 
-	if (overlay->vma)
-		from = intel_frontbuffer_get(overlay->vma->obj);
 	if (vma)
-		to = intel_frontbuffer_get(vma->obj);
+		frontbuffer = intel_frontbuffer_get(vma->obj);
 
-	intel_frontbuffer_track(from, to, INTEL_FRONTBUFFER_OVERLAY(pipe));
+	intel_frontbuffer_track(overlay->frontbuffer, frontbuffer,
+				INTEL_FRONTBUFFER_OVERLAY(pipe));
 
-	if (to)
-		intel_frontbuffer_put(to);
-	if (from)
-		intel_frontbuffer_put(from);
+	if (overlay->frontbuffer)
+		intel_frontbuffer_put(overlay->frontbuffer);
+	overlay->frontbuffer = frontbuffer;
 
 	intel_frontbuffer_flip_prepare(overlay->i915,
 				       INTEL_FRONTBUFFER_OVERLAY(pipe));
@@ -359,7 +359,7 @@ static void intel_overlay_release_old_vma(struct intel_overlay *overlay)
 	intel_frontbuffer_flip_complete(overlay->i915,
 					INTEL_FRONTBUFFER_OVERLAY(overlay->crtc->pipe));
 
-	i915_gem_object_unpin_from_display_plane(vma);
+	i915_vma_unpin(vma);
 	i915_vma_put(vma);
 }
 
@@ -860,7 +860,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	return 0;
 
 out_unpin:
-	i915_gem_object_unpin_from_display_plane(vma);
+	i915_vma_unpin(vma);
 out_pin_section:
 	atomic_dec(&dev_priv->gpu_error.pending_fb_pin);
 
diff --git a/drivers/gpu/drm/i915/display/intel_panel.c b/drivers/gpu/drm/i915/display/intel_panel.c
index 67f81ae995c4..5fdf52643150 100644
--- a/drivers/gpu/drm/i915/display/intel_panel.c
+++ b/drivers/gpu/drm/i915/display/intel_panel.c
@@ -511,40 +511,79 @@ static u32 scale_hw_to_user(struct intel_connector *connector,
 		     0, user_max);
 }
 
-static u32 intel_panel_compute_brightness(struct intel_connector *connector,
-					  u32 val)
+u32 intel_panel_invert_pwm_level(struct intel_connector *connector, u32 val)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
 
-	drm_WARN_ON(&dev_priv->drm, panel->backlight.max == 0);
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.pwm_level_max == 0);
 
 	if (dev_priv->params.invert_brightness < 0)
 		return val;
 
 	if (dev_priv->params.invert_brightness > 0 ||
 	    dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS) {
-		return panel->backlight.max - val + panel->backlight.min;
+		return panel->backlight.pwm_level_max - val + panel->backlight.pwm_level_min;
 	}
 
 	return val;
 }
 
-static u32 lpt_get_backlight(struct intel_connector *connector)
+void intel_panel_set_pwm_level(const struct drm_connector_state *conn_state, u32 val)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_dbg_kms(&i915->drm, "set backlight PWM = %d\n", val);
+	panel->backlight.pwm_funcs->set(conn_state, val);
+}
+
+u32 intel_panel_backlight_level_to_pwm(struct intel_connector *connector, u32 val)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 panel->backlight.max == 0 || panel->backlight.pwm_level_max == 0);
+
+	val = scale(val, panel->backlight.min, panel->backlight.max,
+		    panel->backlight.pwm_level_min, panel->backlight.pwm_level_max);
+
+	return intel_panel_invert_pwm_level(connector, val);
+}
+
+u32 intel_panel_backlight_level_from_pwm(struct intel_connector *connector, u32 val)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 panel->backlight.max == 0 || panel->backlight.pwm_level_max == 0);
+
+	if (dev_priv->params.invert_brightness > 0 ||
+	    (dev_priv->params.invert_brightness == 0 && dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS))
+		val = panel->backlight.pwm_level_max - (val - panel->backlight.pwm_level_min);
+
+	return scale(val, panel->backlight.pwm_level_min, panel->backlight.pwm_level_max,
+		     panel->backlight.min, panel->backlight.max);
+}
+
+static u32 lpt_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 
 	return intel_de_read(dev_priv, BLC_PWM_PCH_CTL2) & BACKLIGHT_DUTY_CYCLE_MASK;
 }
 
-static u32 pch_get_backlight(struct intel_connector *connector)
+static u32 pch_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 
 	return intel_de_read(dev_priv, BLC_PWM_CPU_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
 }
 
-static u32 i9xx_get_backlight(struct intel_connector *connector)
+static u32 i9xx_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
@@ -564,23 +603,17 @@ static u32 i9xx_get_backlight(struct intel_connector *connector)
 	return val;
 }
 
-static u32 _vlv_get_backlight(struct drm_i915_private *dev_priv, enum pipe pipe)
+static u32 vlv_get_backlight(struct intel_connector *connector, enum pipe pipe)
 {
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
 	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
 		return 0;
 
 	return intel_de_read(dev_priv, VLV_BLC_PWM_CTL(pipe)) & BACKLIGHT_DUTY_CYCLE_MASK;
 }
 
-static u32 vlv_get_backlight(struct intel_connector *connector)
-{
-	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
-	enum pipe pipe = intel_connector_get_pipe(connector);
-
-	return _vlv_get_backlight(dev_priv, pipe);
-}
-
-static u32 bxt_get_backlight(struct intel_connector *connector)
+static u32 bxt_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
@@ -589,7 +622,7 @@ static u32 bxt_get_backlight(struct intel_connector *connector)
 			     BXT_BLC_PWM_DUTY(panel->backlight.controller));
 }
 
-static u32 ext_pwm_get_backlight(struct intel_connector *connector)
+static u32 ext_pwm_get_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct intel_panel *panel = &connector->panel;
 	struct pwm_state state;
@@ -624,12 +657,12 @@ static void i9xx_set_backlight(const struct drm_connector_state *conn_state, u32
 	struct intel_panel *panel = &connector->panel;
 	u32 tmp, mask;
 
-	drm_WARN_ON(&dev_priv->drm, panel->backlight.max == 0);
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.pwm_level_max == 0);
 
 	if (panel->backlight.combination_mode) {
 		u8 lbpc;
 
-		lbpc = level * 0xfe / panel->backlight.max + 1;
+		lbpc = level * 0xfe / panel->backlight.pwm_level_max + 1;
 		level /= lbpc;
 		pci_write_config_byte(dev_priv->drm.pdev, LBPC, lbpc);
 	}
@@ -681,9 +714,8 @@ intel_panel_actually_set_backlight(const struct drm_connector_state *conn_state,
 	struct drm_i915_private *i915 = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
 
-	drm_dbg_kms(&i915->drm, "set backlight PWM = %d\n", level);
+	drm_dbg_kms(&i915->drm, "set backlight level = %d\n", level);
 
-	level = intel_panel_compute_brightness(connector, level);
 	panel->backlight.funcs->set(conn_state, level);
 }
 
@@ -732,7 +764,7 @@ static void lpt_disable_backlight(const struct drm_connector_state *old_conn_sta
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	u32 tmp;
 
-	intel_panel_actually_set_backlight(old_conn_state, level);
+	intel_panel_set_pwm_level(old_conn_state, level);
 
 	/*
 	 * Although we don't support or enable CPU PWM with LPT/SPT based
@@ -760,7 +792,7 @@ static void pch_disable_backlight(const struct drm_connector_state *old_conn_sta
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	u32 tmp;
 
-	intel_panel_actually_set_backlight(old_conn_state, val);
+	intel_panel_set_pwm_level(old_conn_state, val);
 
 	tmp = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
 	intel_de_write(dev_priv, BLC_PWM_CPU_CTL2, tmp & ~BLM_PWM_ENABLE);
@@ -771,7 +803,7 @@ static void pch_disable_backlight(const struct drm_connector_state *old_conn_sta
 
 static void i9xx_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
 {
-	intel_panel_actually_set_backlight(old_conn_state, val);
+	intel_panel_set_pwm_level(old_conn_state, val);
 }
 
 static void i965_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
@@ -779,7 +811,7 @@ static void i965_disable_backlight(const struct drm_connector_state *old_conn_st
 	struct drm_i915_private *dev_priv = to_i915(old_conn_state->connector->dev);
 	u32 tmp;
 
-	intel_panel_actually_set_backlight(old_conn_state, val);
+	intel_panel_set_pwm_level(old_conn_state, val);
 
 	tmp = intel_de_read(dev_priv, BLC_PWM_CTL2);
 	intel_de_write(dev_priv, BLC_PWM_CTL2, tmp & ~BLM_PWM_ENABLE);
@@ -792,7 +824,7 @@ static void vlv_disable_backlight(const struct drm_connector_state *old_conn_sta
 	enum pipe pipe = to_intel_crtc(old_conn_state->crtc)->pipe;
 	u32 tmp;
 
-	intel_panel_actually_set_backlight(old_conn_state, val);
+	intel_panel_set_pwm_level(old_conn_state, val);
 
 	tmp = intel_de_read(dev_priv, VLV_BLC_PWM_CTL2(pipe));
 	intel_de_write(dev_priv, VLV_BLC_PWM_CTL2(pipe),
@@ -806,7 +838,7 @@ static void bxt_disable_backlight(const struct drm_connector_state *old_conn_sta
 	struct intel_panel *panel = &connector->panel;
 	u32 tmp;
 
-	intel_panel_actually_set_backlight(old_conn_state, val);
+	intel_panel_set_pwm_level(old_conn_state, val);
 
 	tmp = intel_de_read(dev_priv,
 			    BXT_BLC_PWM_CTL(panel->backlight.controller));
@@ -827,7 +859,7 @@ static void cnp_disable_backlight(const struct drm_connector_state *old_conn_sta
 	struct intel_panel *panel = &connector->panel;
 	u32 tmp;
 
-	intel_panel_actually_set_backlight(old_conn_state, val);
+	intel_panel_set_pwm_level(old_conn_state, val);
 
 	tmp = intel_de_read(dev_priv,
 			    BXT_BLC_PWM_CTL(panel->backlight.controller));
@@ -906,7 +938,7 @@ static void lpt_enable_backlight(const struct intel_crtc_state *crtc_state,
 		intel_de_write(dev_priv, SOUTH_CHICKEN1, schicken);
 	}
 
-	pch_ctl2 = panel->backlight.max << 16;
+	pch_ctl2 = panel->backlight.pwm_level_max << 16;
 	intel_de_write(dev_priv, BLC_PWM_PCH_CTL2, pch_ctl2);
 
 	pch_ctl1 = 0;
@@ -923,7 +955,7 @@ static void lpt_enable_backlight(const struct intel_crtc_state *crtc_state,
 		       pch_ctl1 | BLM_PCH_PWM_ENABLE);
 
 	/* This won't stick until the above enable. */
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 }
 
 static void pch_enable_backlight(const struct intel_crtc_state *crtc_state,
@@ -958,9 +990,9 @@ static void pch_enable_backlight(const struct intel_crtc_state *crtc_state,
 	intel_de_write(dev_priv, BLC_PWM_CPU_CTL2, cpu_ctl2 | BLM_PWM_ENABLE);
 
 	/* This won't stick until the above enable. */
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 
-	pch_ctl2 = panel->backlight.max << 16;
+	pch_ctl2 = panel->backlight.pwm_level_max << 16;
 	intel_de_write(dev_priv, BLC_PWM_PCH_CTL2, pch_ctl2);
 
 	pch_ctl1 = 0;
@@ -987,7 +1019,7 @@ static void i9xx_enable_backlight(const struct intel_crtc_state *crtc_state,
 		intel_de_write(dev_priv, BLC_PWM_CTL, 0);
 	}
 
-	freq = panel->backlight.max;
+	freq = panel->backlight.pwm_level_max;
 	if (panel->backlight.combination_mode)
 		freq /= 0xff;
 
@@ -1001,7 +1033,7 @@ static void i9xx_enable_backlight(const struct intel_crtc_state *crtc_state,
 	intel_de_posting_read(dev_priv, BLC_PWM_CTL);
 
 	/* XXX: combine this into above write? */
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 
 	/*
 	 * Needed to enable backlight on some 855gm models. BLC_HIST_CTL is
@@ -1028,7 +1060,7 @@ static void i965_enable_backlight(const struct intel_crtc_state *crtc_state,
 		intel_de_write(dev_priv, BLC_PWM_CTL2, ctl2);
 	}
 
-	freq = panel->backlight.max;
+	freq = panel->backlight.pwm_level_max;
 	if (panel->backlight.combination_mode)
 		freq /= 0xff;
 
@@ -1044,7 +1076,7 @@ static void i965_enable_backlight(const struct intel_crtc_state *crtc_state,
 	intel_de_posting_read(dev_priv, BLC_PWM_CTL2);
 	intel_de_write(dev_priv, BLC_PWM_CTL2, ctl2 | BLM_PWM_ENABLE);
 
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 }
 
 static void vlv_enable_backlight(const struct intel_crtc_state *crtc_state,
@@ -1063,11 +1095,11 @@ static void vlv_enable_backlight(const struct intel_crtc_state *crtc_state,
 		intel_de_write(dev_priv, VLV_BLC_PWM_CTL2(pipe), ctl2);
 	}
 
-	ctl = panel->backlight.max << 16;
+	ctl = panel->backlight.pwm_level_max << 16;
 	intel_de_write(dev_priv, VLV_BLC_PWM_CTL(pipe), ctl);
 
 	/* XXX: combine this into above write? */
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 
 	ctl2 = 0;
 	if (panel->backlight.active_low_pwm)
@@ -1116,9 +1148,9 @@ static void bxt_enable_backlight(const struct intel_crtc_state *crtc_state,
 
 	intel_de_write(dev_priv,
 		       BXT_BLC_PWM_FREQ(panel->backlight.controller),
-		       panel->backlight.max);
+		       panel->backlight.pwm_level_max);
 
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 
 	pwm_ctl = 0;
 	if (panel->backlight.active_low_pwm)
@@ -1152,9 +1184,9 @@ static void cnp_enable_backlight(const struct intel_crtc_state *crtc_state,
 
 	intel_de_write(dev_priv,
 		       BXT_BLC_PWM_FREQ(panel->backlight.controller),
-		       panel->backlight.max);
+		       panel->backlight.pwm_level_max);
 
-	intel_panel_actually_set_backlight(conn_state, level);
+	intel_panel_set_pwm_level(conn_state, level);
 
 	pwm_ctl = 0;
 	if (panel->backlight.active_low_pwm)
@@ -1174,7 +1206,6 @@ static void ext_pwm_enable_backlight(const struct intel_crtc_state *crtc_state,
 	struct intel_connector *connector = to_intel_connector(conn_state->connector);
 	struct intel_panel *panel = &connector->panel;
 
-	level = intel_panel_compute_brightness(connector, level);
 	pwm_set_relative_duty_cycle(&panel->backlight.pwm_state, level, 100);
 	panel->backlight.pwm_state.enabled = true;
 	pwm_apply_state(panel->backlight.pwm, &panel->backlight.pwm_state);
@@ -1232,10 +1263,8 @@ static u32 intel_panel_get_backlight(struct intel_connector *connector)
 
 	mutex_lock(&dev_priv->backlight_lock);
 
-	if (panel->backlight.enabled) {
-		val = panel->backlight.funcs->get(connector);
-		val = intel_panel_compute_brightness(connector, val);
-	}
+	if (panel->backlight.enabled)
+		val = panel->backlight.funcs->get(connector, intel_connector_get_pipe(connector));
 
 	mutex_unlock(&dev_priv->backlight_lock);
 
@@ -1566,13 +1595,13 @@ static u32 get_backlight_max_vbt(struct intel_connector *connector)
 	u16 pwm_freq_hz = get_vbt_pwm_freq(dev_priv);
 	u32 pwm;
 
-	if (!panel->backlight.funcs->hz_to_pwm) {
+	if (!panel->backlight.pwm_funcs->hz_to_pwm) {
 		drm_dbg_kms(&dev_priv->drm,
 			    "backlight frequency conversion not supported\n");
 		return 0;
 	}
 
-	pwm = panel->backlight.funcs->hz_to_pwm(connector, pwm_freq_hz);
+	pwm = panel->backlight.pwm_funcs->hz_to_pwm(connector, pwm_freq_hz);
 	if (!pwm) {
 		drm_dbg_kms(&dev_priv->drm,
 			    "backlight frequency conversion failed\n");
@@ -1591,7 +1620,7 @@ static u32 get_backlight_min_vbt(struct intel_connector *connector)
 	struct intel_panel *panel = &connector->panel;
 	int min;
 
-	drm_WARN_ON(&dev_priv->drm, panel->backlight.max == 0);
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.pwm_level_max == 0);
 
 	/*
 	 * XXX: If the vbt value is 255, it makes min equal to max, which leads
@@ -1608,7 +1637,7 @@ static u32 get_backlight_min_vbt(struct intel_connector *connector)
 	}
 
 	/* vbt value is a coefficient in range [0..255] */
-	return scale(min, 0, 255, 0, panel->backlight.max);
+	return scale(min, 0, 255, 0, panel->backlight.pwm_level_max);
 }
 
 static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unused)
@@ -1628,37 +1657,32 @@ static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unus
 	panel->backlight.active_low_pwm = pch_ctl1 & BLM_PCH_POLARITY;
 
 	pch_ctl2 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL2);
-	panel->backlight.max = pch_ctl2 >> 16;
+	panel->backlight.pwm_level_max = pch_ctl2 >> 16;
 
 	cpu_ctl2 = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
 
-	if (!panel->backlight.max)
-		panel->backlight.max = get_backlight_max_vbt(connector);
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
-	panel->backlight.min = get_backlight_min_vbt(connector);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
-	panel->backlight.enabled = pch_ctl1 & BLM_PCH_PWM_ENABLE;
+	panel->backlight.pwm_enabled = pch_ctl1 & BLM_PCH_PWM_ENABLE;
 
-	cpu_mode = panel->backlight.enabled && HAS_PCH_LPT(dev_priv) &&
+	cpu_mode = panel->backlight.pwm_enabled && HAS_PCH_LPT(dev_priv) &&
 		   !(pch_ctl1 & BLM_PCH_OVERRIDE_ENABLE) &&
 		   (cpu_ctl2 & BLM_PWM_ENABLE);
-	if (cpu_mode)
-		val = pch_get_backlight(connector);
-	else
-		val = lpt_get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
 
 	if (cpu_mode) {
+		val = pch_get_backlight(connector, unused);
+
 		drm_dbg_kms(&dev_priv->drm,
 			    "CPU backlight register was enabled, switching to PCH override\n");
 
 		/* Write converted CPU PWM value to PCH override register */
-		lpt_set_backlight(connector->base.state, panel->backlight.level);
+		lpt_set_backlight(connector->base.state, val);
 		intel_de_write(dev_priv, BLC_PWM_PCH_CTL1,
 			       pch_ctl1 | BLM_PCH_OVERRIDE_ENABLE);
 
@@ -1673,29 +1697,24 @@ static int pch_setup_backlight(struct intel_connector *connector, enum pipe unus
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
-	u32 cpu_ctl2, pch_ctl1, pch_ctl2, val;
+	u32 cpu_ctl2, pch_ctl1, pch_ctl2;
 
 	pch_ctl1 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
 	panel->backlight.active_low_pwm = pch_ctl1 & BLM_PCH_POLARITY;
 
 	pch_ctl2 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL2);
-	panel->backlight.max = pch_ctl2 >> 16;
+	panel->backlight.pwm_level_max = pch_ctl2 >> 16;
 
-	if (!panel->backlight.max)
-		panel->backlight.max = get_backlight_max_vbt(connector);
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
-	panel->backlight.min = get_backlight_min_vbt(connector);
-
-	val = pch_get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
 	cpu_ctl2 = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
-	panel->backlight.enabled = (cpu_ctl2 & BLM_PWM_ENABLE) &&
+	panel->backlight.pwm_enabled = (cpu_ctl2 & BLM_PWM_ENABLE) &&
 		(pch_ctl1 & BLM_PCH_PWM_ENABLE);
 
 	return 0;
@@ -1715,27 +1734,26 @@ static int i9xx_setup_backlight(struct intel_connector *connector, enum pipe unu
 	if (IS_PINEVIEW(dev_priv))
 		panel->backlight.active_low_pwm = ctl & BLM_POLARITY_PNV;
 
-	panel->backlight.max = ctl >> 17;
+	panel->backlight.pwm_level_max = ctl >> 17;
 
-	if (!panel->backlight.max) {
-		panel->backlight.max = get_backlight_max_vbt(connector);
-		panel->backlight.max >>= 1;
+	if (!panel->backlight.pwm_level_max) {
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+		panel->backlight.pwm_level_max >>= 1;
 	}
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
 	if (panel->backlight.combination_mode)
-		panel->backlight.max *= 0xff;
+		panel->backlight.pwm_level_max *= 0xff;
 
-	panel->backlight.min = get_backlight_min_vbt(connector);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
-	val = i9xx_get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
+	val = i9xx_get_backlight(connector, unused);
+	val = intel_panel_invert_pwm_level(connector, val);
+	val = clamp(val, panel->backlight.pwm_level_min, panel->backlight.pwm_level_max);
 
-	panel->backlight.enabled = val != 0;
+	panel->backlight.pwm_enabled = val != 0;
 
 	return 0;
 }
@@ -1744,32 +1762,27 @@ static int i965_setup_backlight(struct intel_connector *connector, enum pipe unu
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
-	u32 ctl, ctl2, val;
+	u32 ctl, ctl2;
 
 	ctl2 = intel_de_read(dev_priv, BLC_PWM_CTL2);
 	panel->backlight.combination_mode = ctl2 & BLM_COMBINATION_MODE;
 	panel->backlight.active_low_pwm = ctl2 & BLM_POLARITY_I965;
 
 	ctl = intel_de_read(dev_priv, BLC_PWM_CTL);
-	panel->backlight.max = ctl >> 16;
+	panel->backlight.pwm_level_max = ctl >> 16;
 
-	if (!panel->backlight.max)
-		panel->backlight.max = get_backlight_max_vbt(connector);
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
 	if (panel->backlight.combination_mode)
-		panel->backlight.max *= 0xff;
-
-	panel->backlight.min = get_backlight_min_vbt(connector);
+		panel->backlight.pwm_level_max *= 0xff;
 
-	val = i9xx_get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
-	panel->backlight.enabled = ctl2 & BLM_PWM_ENABLE;
+	panel->backlight.pwm_enabled = ctl2 & BLM_PWM_ENABLE;
 
 	return 0;
 }
@@ -1778,7 +1791,7 @@ static int vlv_setup_backlight(struct intel_connector *connector, enum pipe pipe
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
-	u32 ctl, ctl2, val;
+	u32 ctl, ctl2;
 
 	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
 		return -ENODEV;
@@ -1787,22 +1800,17 @@ static int vlv_setup_backlight(struct intel_connector *connector, enum pipe pipe
 	panel->backlight.active_low_pwm = ctl2 & BLM_POLARITY_I965;
 
 	ctl = intel_de_read(dev_priv, VLV_BLC_PWM_CTL(pipe));
-	panel->backlight.max = ctl >> 16;
+	panel->backlight.pwm_level_max = ctl >> 16;
 
-	if (!panel->backlight.max)
-		panel->backlight.max = get_backlight_max_vbt(connector);
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
-	panel->backlight.min = get_backlight_min_vbt(connector);
-
-	val = _vlv_get_backlight(dev_priv, pipe);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
-	panel->backlight.enabled = ctl2 & BLM_PWM_ENABLE;
+	panel->backlight.pwm_enabled = ctl2 & BLM_PWM_ENABLE;
 
 	return 0;
 }
@@ -1827,24 +1835,18 @@ bxt_setup_backlight(struct intel_connector *connector, enum pipe unused)
 	}
 
 	panel->backlight.active_low_pwm = pwm_ctl & BXT_BLC_PWM_POLARITY;
-	panel->backlight.max =
-		intel_de_read(dev_priv,
-			      BXT_BLC_PWM_FREQ(panel->backlight.controller));
+	panel->backlight.pwm_level_max =
+		intel_de_read(dev_priv, BXT_BLC_PWM_FREQ(panel->backlight.controller));
 
-	if (!panel->backlight.max)
-		panel->backlight.max = get_backlight_max_vbt(connector);
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
-	panel->backlight.min = get_backlight_min_vbt(connector);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
-	val = bxt_get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
-
-	panel->backlight.enabled = pwm_ctl & BXT_BLC_PWM_ENABLE;
+	panel->backlight.pwm_enabled = pwm_ctl & BXT_BLC_PWM_ENABLE;
 
 	return 0;
 }
@@ -1854,7 +1856,7 @@ cnp_setup_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 	struct intel_panel *panel = &connector->panel;
-	u32 pwm_ctl, val;
+	u32 pwm_ctl;
 
 	/*
 	 * CNP has the BXT implementation of backlight, but with only one
@@ -1867,24 +1869,18 @@ cnp_setup_backlight(struct intel_connector *connector, enum pipe unused)
 				BXT_BLC_PWM_CTL(panel->backlight.controller));
 
 	panel->backlight.active_low_pwm = pwm_ctl & BXT_BLC_PWM_POLARITY;
-	panel->backlight.max =
-		intel_de_read(dev_priv,
-			      BXT_BLC_PWM_FREQ(panel->backlight.controller));
+	panel->backlight.pwm_level_max =
+		intel_de_read(dev_priv, BXT_BLC_PWM_FREQ(panel->backlight.controller));
 
-	if (!panel->backlight.max)
-		panel->backlight.max = get_backlight_max_vbt(connector);
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
 
-	if (!panel->backlight.max)
+	if (!panel->backlight.pwm_level_max)
 		return -ENODEV;
 
-	panel->backlight.min = get_backlight_min_vbt(connector);
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
-	val = bxt_get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
-	panel->backlight.level = clamp(val, panel->backlight.min,
-				       panel->backlight.max);
-
-	panel->backlight.enabled = pwm_ctl & BXT_BLC_PWM_ENABLE;
+	panel->backlight.pwm_enabled = pwm_ctl & BXT_BLC_PWM_ENABLE;
 
 	return 0;
 }
@@ -1914,8 +1910,8 @@ static int ext_pwm_setup_backlight(struct intel_connector *connector,
 		return -ENODEV;
 	}
 
-	panel->backlight.max = 100; /* 100% */
-	panel->backlight.min = get_backlight_min_vbt(connector);
+	panel->backlight.pwm_level_max = 100; /* 100% */
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
 
 	if (pwm_is_enabled(panel->backlight.pwm)) {
 		/* PWM is already enabled, use existing settings */
@@ -1923,10 +1919,8 @@ static int ext_pwm_setup_backlight(struct intel_connector *connector,
 
 		level = pwm_get_relative_duty_cycle(&panel->backlight.pwm_state,
 						    100);
-		level = intel_panel_compute_brightness(connector, level);
-		panel->backlight.level = clamp(level, panel->backlight.min,
-					       panel->backlight.max);
-		panel->backlight.enabled = true;
+		level = intel_panel_invert_pwm_level(connector, level);
+		panel->backlight.pwm_enabled = true;
 
 		drm_dbg_kms(&dev_priv->drm, "PWM already enabled at freq %ld, VBT freq %d, level %d\n",
 			    NSEC_PER_SEC / (unsigned long)panel->backlight.pwm_state.period,
@@ -1942,6 +1936,58 @@ static int ext_pwm_setup_backlight(struct intel_connector *connector,
 	return 0;
 }
 
+static void intel_pwm_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	panel->backlight.pwm_funcs->set(conn_state,
+				       intel_panel_invert_pwm_level(connector, level));
+}
+
+static u32 intel_pwm_get_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	return intel_panel_invert_pwm_level(connector,
+					    panel->backlight.pwm_funcs->get(connector, pipe));
+}
+
+static void intel_pwm_enable_backlight(const struct intel_crtc_state *crtc_state,
+				       const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	panel->backlight.pwm_funcs->enable(crtc_state, conn_state,
+					   intel_panel_invert_pwm_level(connector, level));
+}
+
+static void intel_pwm_disable_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	panel->backlight.pwm_funcs->disable(conn_state,
+					    intel_panel_invert_pwm_level(connector, level));
+}
+
+static int intel_pwm_setup_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct intel_panel *panel = &connector->panel;
+	int ret = panel->backlight.pwm_funcs->setup(connector, pipe);
+
+	if (ret < 0)
+		return ret;
+
+	panel->backlight.min = panel->backlight.pwm_level_min;
+	panel->backlight.max = panel->backlight.pwm_level_max;
+	panel->backlight.level = intel_pwm_get_backlight(connector, pipe);
+	panel->backlight.enabled = panel->backlight.pwm_enabled;
+
+	return 0;
+}
+
 void intel_panel_update_backlight(struct intel_atomic_state *state,
 				  struct intel_encoder *encoder,
 				  const struct intel_crtc_state *crtc_state,
@@ -2015,7 +2061,7 @@ static void intel_panel_destroy_backlight(struct intel_panel *panel)
 	panel->backlight.present = false;
 }
 
-static const struct intel_panel_bl_funcs bxt_funcs = {
+static const struct intel_panel_bl_funcs bxt_pwm_funcs = {
 	.setup = bxt_setup_backlight,
 	.enable = bxt_enable_backlight,
 	.disable = bxt_disable_backlight,
@@ -2024,7 +2070,7 @@ static const struct intel_panel_bl_funcs bxt_funcs = {
 	.hz_to_pwm = bxt_hz_to_pwm,
 };
 
-static const struct intel_panel_bl_funcs cnp_funcs = {
+static const struct intel_panel_bl_funcs cnp_pwm_funcs = {
 	.setup = cnp_setup_backlight,
 	.enable = cnp_enable_backlight,
 	.disable = cnp_disable_backlight,
@@ -2033,7 +2079,7 @@ static const struct intel_panel_bl_funcs cnp_funcs = {
 	.hz_to_pwm = cnp_hz_to_pwm,
 };
 
-static const struct intel_panel_bl_funcs lpt_funcs = {
+static const struct intel_panel_bl_funcs lpt_pwm_funcs = {
 	.setup = lpt_setup_backlight,
 	.enable = lpt_enable_backlight,
 	.disable = lpt_disable_backlight,
@@ -2042,7 +2088,7 @@ static const struct intel_panel_bl_funcs lpt_funcs = {
 	.hz_to_pwm = lpt_hz_to_pwm,
 };
 
-static const struct intel_panel_bl_funcs spt_funcs = {
+static const struct intel_panel_bl_funcs spt_pwm_funcs = {
 	.setup = lpt_setup_backlight,
 	.enable = lpt_enable_backlight,
 	.disable = lpt_disable_backlight,
@@ -2051,7 +2097,7 @@ static const struct intel_panel_bl_funcs spt_funcs = {
 	.hz_to_pwm = spt_hz_to_pwm,
 };
 
-static const struct intel_panel_bl_funcs pch_funcs = {
+static const struct intel_panel_bl_funcs pch_pwm_funcs = {
 	.setup = pch_setup_backlight,
 	.enable = pch_enable_backlight,
 	.disable = pch_disable_backlight,
@@ -2068,7 +2114,7 @@ static const struct intel_panel_bl_funcs ext_pwm_funcs = {
 	.get = ext_pwm_get_backlight,
 };
 
-static const struct intel_panel_bl_funcs vlv_funcs = {
+static const struct intel_panel_bl_funcs vlv_pwm_funcs = {
 	.setup = vlv_setup_backlight,
 	.enable = vlv_enable_backlight,
 	.disable = vlv_disable_backlight,
@@ -2077,7 +2123,7 @@ static const struct intel_panel_bl_funcs vlv_funcs = {
 	.hz_to_pwm = vlv_hz_to_pwm,
 };
 
-static const struct intel_panel_bl_funcs i965_funcs = {
+static const struct intel_panel_bl_funcs i965_pwm_funcs = {
 	.setup = i965_setup_backlight,
 	.enable = i965_enable_backlight,
 	.disable = i965_disable_backlight,
@@ -2086,7 +2132,7 @@ static const struct intel_panel_bl_funcs i965_funcs = {
 	.hz_to_pwm = i965_hz_to_pwm,
 };
 
-static const struct intel_panel_bl_funcs i9xx_funcs = {
+static const struct intel_panel_bl_funcs i9xx_pwm_funcs = {
 	.setup = i9xx_setup_backlight,
 	.enable = i9xx_enable_backlight,
 	.disable = i9xx_disable_backlight,
@@ -2095,6 +2141,14 @@ static const struct intel_panel_bl_funcs i9xx_funcs = {
 	.hz_to_pwm = i9xx_hz_to_pwm,
 };
 
+static const struct intel_panel_bl_funcs pwm_bl_funcs = {
+	.setup = intel_pwm_setup_backlight,
+	.enable = intel_pwm_enable_backlight,
+	.disable = intel_pwm_disable_backlight,
+	.set = intel_pwm_set_backlight,
+	.get = intel_pwm_get_backlight,
+};
+
 /* Set up chip specific backlight functions */
 static void
 intel_panel_init_backlight_funcs(struct intel_panel *panel)
@@ -2103,36 +2157,39 @@ intel_panel_init_backlight_funcs(struct intel_panel *panel)
 		container_of(panel, struct intel_connector, panel);
 	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
 
-	if (connector->base.connector_type == DRM_MODE_CONNECTOR_eDP &&
-	    intel_dp_aux_init_backlight_funcs(connector) == 0)
-		return;
-
 	if (connector->base.connector_type == DRM_MODE_CONNECTOR_DSI &&
 	    intel_dsi_dcs_init_backlight_funcs(connector) == 0)
 		return;
 
 	if (IS_GEN9_LP(dev_priv)) {
-		panel->backlight.funcs = &bxt_funcs;
+		panel->backlight.pwm_funcs = &bxt_pwm_funcs;
 	} else if (INTEL_PCH_TYPE(dev_priv) >= PCH_CNP) {
-		panel->backlight.funcs = &cnp_funcs;
+		panel->backlight.pwm_funcs = &cnp_pwm_funcs;
 	} else if (INTEL_PCH_TYPE(dev_priv) >= PCH_LPT) {
 		if (HAS_PCH_LPT(dev_priv))
-			panel->backlight.funcs = &lpt_funcs;
+			panel->backlight.pwm_funcs = &lpt_pwm_funcs;
 		else
-			panel->backlight.funcs = &spt_funcs;
+			panel->backlight.pwm_funcs = &spt_pwm_funcs;
 	} else if (HAS_PCH_SPLIT(dev_priv)) {
-		panel->backlight.funcs = &pch_funcs;
+		panel->backlight.pwm_funcs = &pch_pwm_funcs;
 	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
 		if (connector->base.connector_type == DRM_MODE_CONNECTOR_DSI) {
-			panel->backlight.funcs = &ext_pwm_funcs;
+			panel->backlight.pwm_funcs = &ext_pwm_funcs;
 		} else {
-			panel->backlight.funcs = &vlv_funcs;
+			panel->backlight.pwm_funcs = &vlv_pwm_funcs;
 		}
 	} else if (IS_GEN(dev_priv, 4)) {
-		panel->backlight.funcs = &i965_funcs;
+		panel->backlight.pwm_funcs = &i965_pwm_funcs;
 	} else {
-		panel->backlight.funcs = &i9xx_funcs;
+		panel->backlight.pwm_funcs = &i9xx_pwm_funcs;
 	}
+
+	if (connector->base.connector_type == DRM_MODE_CONNECTOR_eDP &&
+	    intel_dp_aux_init_backlight_funcs(connector) == 0)
+		return;
+
+	/* We're using a standard PWM backlight interface */
+	panel->backlight.funcs = &pwm_bl_funcs;
 }
 
 enum drm_connector_status
diff --git a/drivers/gpu/drm/i915/display/intel_panel.h b/drivers/gpu/drm/i915/display/intel_panel.h
index 5b813fe90557..1d340f77bffc 100644
--- a/drivers/gpu/drm/i915/display/intel_panel.h
+++ b/drivers/gpu/drm/i915/display/intel_panel.h
@@ -49,6 +49,10 @@ struct drm_display_mode *
 intel_panel_edid_fixed_mode(struct intel_connector *connector);
 struct drm_display_mode *
 intel_panel_vbt_fixed_mode(struct intel_connector *connector);
+void intel_panel_set_pwm_level(const struct drm_connector_state *conn_state, u32 level);
+u32 intel_panel_invert_pwm_level(struct intel_connector *connector, u32 level);
+u32 intel_panel_backlight_level_to_pwm(struct intel_connector *connector, u32 level);
+u32 intel_panel_backlight_level_from_pwm(struct intel_connector *connector, u32 val);
 
 #if IS_ENABLED(CONFIG_BACKLIGHT_CLASS_DEVICE)
 int intel_backlight_device_register(struct intel_connector *connector);
diff --git a/drivers/gpu/drm/i915/display/intel_pps.c b/drivers/gpu/drm/i915/display/intel_pps.c
new file mode 100644
index 000000000000..c4867a8020a5
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pps.c
@@ -0,0 +1,1406 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_pps.h"
+
+static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv,
+				      enum pipe pipe);
+
+static void pps_init_delays(struct intel_dp *intel_dp);
+static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd);
+
+intel_wakeref_t intel_pps_lock(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	/*
+	 * See intel_pps_reset_all() why we need a power domain reference here.
+	 */
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_DISPLAY_CORE);
+	mutex_lock(&dev_priv->pps_mutex);
+
+	return wakeref;
+}
+
+intel_wakeref_t intel_pps_unlock(struct intel_dp *intel_dp,
+				 intel_wakeref_t wakeref)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	mutex_unlock(&dev_priv->pps_mutex);
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return 0;
+}
+
+static void
+vlv_power_sequencer_kick(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum pipe pipe = intel_dp->pps.pps_pipe;
+	bool pll_enabled, release_cl_override = false;
+	enum dpio_phy phy = DPIO_PHY(pipe);
+	enum dpio_channel ch = vlv_pipe_to_channel(pipe);
+	u32 DP;
+
+	if (drm_WARN(&dev_priv->drm,
+		     intel_de_read(dev_priv, intel_dp->output_reg) & DP_PORT_EN,
+		     "skipping pipe %c power sequencer kick due to [ENCODER:%d:%s] being active\n",
+		     pipe_name(pipe), dig_port->base.base.base.id,
+		     dig_port->base.base.name))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "kicking pipe %c power sequencer for [ENCODER:%d:%s]\n",
+		    pipe_name(pipe), dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	/* Preserve the BIOS-computed detected bit. This is
+	 * supposed to be read-only.
+	 */
+	DP = intel_de_read(dev_priv, intel_dp->output_reg) & DP_DETECTED;
+	DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	DP |= DP_PORT_WIDTH(1);
+	DP |= DP_LINK_TRAIN_PAT_1;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		DP |= DP_PIPE_SEL_CHV(pipe);
+	else
+		DP |= DP_PIPE_SEL(pipe);
+
+	pll_enabled = intel_de_read(dev_priv, DPLL(pipe)) & DPLL_VCO_ENABLE;
+
+	/*
+	 * The DPLL for the pipe must be enabled for this to work.
+	 * So enable temporarily it if it's not already enabled.
+	 */
+	if (!pll_enabled) {
+		release_cl_override = IS_CHERRYVIEW(dev_priv) &&
+			!chv_phy_powergate_ch(dev_priv, phy, ch, true);
+
+		if (vlv_force_pll_on(dev_priv, pipe, vlv_get_dpll(dev_priv))) {
+			drm_err(&dev_priv->drm,
+				"Failed to force on pll for pipe %c!\n",
+				pipe_name(pipe));
+			return;
+		}
+	}
+
+	/*
+	 * Similar magic as in intel_dp_enable_port().
+	 * We _must_ do this port enable + disable trick
+	 * to make this power sequencer lock onto the port.
+	 * Otherwise even VDD force bit won't work.
+	 */
+	intel_de_write(dev_priv, intel_dp->output_reg, DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	intel_de_write(dev_priv, intel_dp->output_reg, DP | DP_PORT_EN);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	intel_de_write(dev_priv, intel_dp->output_reg, DP & ~DP_PORT_EN);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	if (!pll_enabled) {
+		vlv_force_pll_off(dev_priv, pipe);
+
+		if (release_cl_override)
+			chv_phy_powergate_ch(dev_priv, phy, ch, false);
+	}
+}
+
+static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+	unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B);
+
+	/*
+	 * We don't have power sequencer currently.
+	 * Pick one that's not used by other ports.
+	 */
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		if (encoder->type == INTEL_OUTPUT_EDP) {
+			drm_WARN_ON(&dev_priv->drm,
+				    intel_dp->pps.active_pipe != INVALID_PIPE &&
+				    intel_dp->pps.active_pipe !=
+				    intel_dp->pps.pps_pipe);
+
+			if (intel_dp->pps.pps_pipe != INVALID_PIPE)
+				pipes &= ~(1 << intel_dp->pps.pps_pipe);
+		} else {
+			drm_WARN_ON(&dev_priv->drm,
+				    intel_dp->pps.pps_pipe != INVALID_PIPE);
+
+			if (intel_dp->pps.active_pipe != INVALID_PIPE)
+				pipes &= ~(1 << intel_dp->pps.active_pipe);
+		}
+	}
+
+	if (pipes == 0)
+		return INVALID_PIPE;
+
+	return ffs(pipes) - 1;
+}
+
+static enum pipe
+vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum pipe pipe;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* We should never land here with regular DP ports */
+	drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp));
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE &&
+		    intel_dp->pps.active_pipe != intel_dp->pps.pps_pipe);
+
+	if (intel_dp->pps.pps_pipe != INVALID_PIPE)
+		return intel_dp->pps.pps_pipe;
+
+	pipe = vlv_find_free_pps(dev_priv);
+
+	/*
+	 * Didn't find one. This should not happen since there
+	 * are two power sequencers and up to two eDP ports.
+	 */
+	if (drm_WARN_ON(&dev_priv->drm, pipe == INVALID_PIPE))
+		pipe = PIPE_A;
+
+	vlv_steal_power_sequencer(dev_priv, pipe);
+	intel_dp->pps.pps_pipe = pipe;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "picked pipe %c power sequencer for [ENCODER:%d:%s]\n",
+		    pipe_name(intel_dp->pps.pps_pipe),
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	/* init power sequencer on this pipe and port */
+	pps_init_delays(intel_dp);
+	pps_init_registers(intel_dp, true);
+
+	/*
+	 * Even vdd force doesn't work until we've made
+	 * the power sequencer lock in on the port.
+	 */
+	vlv_power_sequencer_kick(intel_dp);
+
+	return intel_dp->pps.pps_pipe;
+}
+
+static int
+bxt_power_sequencer_idx(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	int backlight_controller = dev_priv->vbt.backlight.controller;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* We should never land here with regular DP ports */
+	drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp));
+
+	if (!intel_dp->pps.pps_reset)
+		return backlight_controller;
+
+	intel_dp->pps.pps_reset = false;
+
+	/*
+	 * Only the HW needs to be reprogrammed, the SW state is fixed and
+	 * has been setup during connector init.
+	 */
+	pps_init_registers(intel_dp, false);
+
+	return backlight_controller;
+}
+
+typedef bool (*vlv_pipe_check)(struct drm_i915_private *dev_priv,
+			       enum pipe pipe);
+
+static bool vlv_pipe_has_pp_on(struct drm_i915_private *dev_priv,
+			       enum pipe pipe)
+{
+	return intel_de_read(dev_priv, PP_STATUS(pipe)) & PP_ON;
+}
+
+static bool vlv_pipe_has_vdd_on(struct drm_i915_private *dev_priv,
+				enum pipe pipe)
+{
+	return intel_de_read(dev_priv, PP_CONTROL(pipe)) & EDP_FORCE_VDD;
+}
+
+static bool vlv_pipe_any(struct drm_i915_private *dev_priv,
+			 enum pipe pipe)
+{
+	return true;
+}
+
+static enum pipe
+vlv_initial_pps_pipe(struct drm_i915_private *dev_priv,
+		     enum port port,
+		     vlv_pipe_check pipe_check)
+{
+	enum pipe pipe;
+
+	for (pipe = PIPE_A; pipe <= PIPE_B; pipe++) {
+		u32 port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(pipe)) &
+			PANEL_PORT_SELECT_MASK;
+
+		if (port_sel != PANEL_PORT_SELECT_VLV(port))
+			continue;
+
+		if (!pipe_check(dev_priv, pipe))
+			continue;
+
+		return pipe;
+	}
+
+	return INVALID_PIPE;
+}
+
+static void
+vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum port port = dig_port->base.port;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* try to find a pipe with this port selected */
+	/* first pick one where the panel is on */
+	intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+						      vlv_pipe_has_pp_on);
+	/* didn't find one? pick one where vdd is on */
+	if (intel_dp->pps.pps_pipe == INVALID_PIPE)
+		intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							      vlv_pipe_has_vdd_on);
+	/* didn't find one? pick one with just the correct port */
+	if (intel_dp->pps.pps_pipe == INVALID_PIPE)
+		intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							      vlv_pipe_any);
+
+	/* didn't find one? just let vlv_power_sequencer_pipe() pick one when needed */
+	if (intel_dp->pps.pps_pipe == INVALID_PIPE) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "no initial power sequencer for [ENCODER:%d:%s]\n",
+			    dig_port->base.base.base.id,
+			    dig_port->base.base.name);
+		return;
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "initial power sequencer for [ENCODER:%d:%s]: pipe %c\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name,
+		    pipe_name(intel_dp->pps.pps_pipe));
+}
+
+void intel_pps_reset_all(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			!(IS_VALLEYVIEW(dev_priv) ||
+			  IS_CHERRYVIEW(dev_priv) ||
+			  IS_GEN9_LP(dev_priv))))
+		return;
+
+	/*
+	 * We can't grab pps_mutex here due to deadlock with power_domain
+	 * mutex when power_domain functions are called while holding pps_mutex.
+	 * That also means that in order to use pps_pipe the code needs to
+	 * hold both a power domain reference and pps_mutex, and the power domain
+	 * reference get/put must be done while _not_ holding pps_mutex.
+	 * pps_{lock,unlock}() do these steps in the correct order, so one
+	 * should use them always.
+	 */
+
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		drm_WARN_ON(&dev_priv->drm,
+			    intel_dp->pps.active_pipe != INVALID_PIPE);
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		if (IS_GEN9_LP(dev_priv))
+			intel_dp->pps.pps_reset = true;
+		else
+			intel_dp->pps.pps_pipe = INVALID_PIPE;
+	}
+}
+
+struct pps_registers {
+	i915_reg_t pp_ctrl;
+	i915_reg_t pp_stat;
+	i915_reg_t pp_on;
+	i915_reg_t pp_off;
+	i915_reg_t pp_div;
+};
+
+static void intel_pps_get_registers(struct intel_dp *intel_dp,
+				    struct pps_registers *regs)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	int pps_idx = 0;
+
+	memset(regs, 0, sizeof(*regs));
+
+	if (IS_GEN9_LP(dev_priv))
+		pps_idx = bxt_power_sequencer_idx(intel_dp);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		pps_idx = vlv_power_sequencer_pipe(intel_dp);
+
+	regs->pp_ctrl = PP_CONTROL(pps_idx);
+	regs->pp_stat = PP_STATUS(pps_idx);
+	regs->pp_on = PP_ON_DELAYS(pps_idx);
+	regs->pp_off = PP_OFF_DELAYS(pps_idx);
+
+	/* Cycle delay moved from PP_DIVISOR to PP_CONTROL */
+	if (IS_GEN9_LP(dev_priv) || INTEL_PCH_TYPE(dev_priv) >= PCH_CNP)
+		regs->pp_div = INVALID_MMIO_REG;
+	else
+		regs->pp_div = PP_DIVISOR(pps_idx);
+}
+
+static i915_reg_t
+_pp_ctrl_reg(struct intel_dp *intel_dp)
+{
+	struct pps_registers regs;
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	return regs.pp_ctrl;
+}
+
+static i915_reg_t
+_pp_stat_reg(struct intel_dp *intel_dp)
+{
+	struct pps_registers regs;
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	return regs.pp_stat;
+}
+
+static bool edp_have_panel_power(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    intel_dp->pps.pps_pipe == INVALID_PIPE)
+		return false;
+
+	return (intel_de_read(dev_priv, _pp_stat_reg(intel_dp)) & PP_ON) != 0;
+}
+
+static bool edp_have_panel_vdd(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    intel_dp->pps.pps_pipe == INVALID_PIPE)
+		return false;
+
+	return intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD;
+}
+
+void intel_pps_check_power_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	if (!edp_have_panel_power(intel_dp) && !edp_have_panel_vdd(intel_dp)) {
+		drm_WARN(&dev_priv->drm, 1,
+			 "eDP powered off while attempting aux channel communication.\n");
+		drm_dbg_kms(&dev_priv->drm, "Status 0x%08x Control 0x%08x\n",
+			    intel_de_read(dev_priv, _pp_stat_reg(intel_dp)),
+			    intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)));
+	}
+}
+
+#define IDLE_ON_MASK		(PP_ON | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
+#define IDLE_ON_VALUE   	(PP_ON | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_ON_IDLE)
+
+#define IDLE_OFF_MASK		(PP_ON | PP_SEQUENCE_MASK | 0                     | 0)
+#define IDLE_OFF_VALUE		(0     | PP_SEQUENCE_NONE | 0                     | 0)
+
+#define IDLE_CYCLE_MASK		(PP_ON | PP_SEQUENCE_MASK | PP_CYCLE_DELAY_ACTIVE | PP_SEQUENCE_STATE_MASK)
+#define IDLE_CYCLE_VALUE	(0     | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_OFF_IDLE)
+
+static void intel_pps_verify_state(struct intel_dp *intel_dp);
+
+static void wait_panel_status(struct intel_dp *intel_dp,
+				       u32 mask,
+				       u32 value)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	i915_reg_t pp_stat_reg, pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	intel_pps_verify_state(intel_dp);
+
+	pp_stat_reg = _pp_stat_reg(intel_dp);
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "mask %08x value %08x status %08x control %08x\n",
+		    mask, value,
+		    intel_de_read(dev_priv, pp_stat_reg),
+		    intel_de_read(dev_priv, pp_ctrl_reg));
+
+	if (intel_de_wait_for_register(dev_priv, pp_stat_reg,
+				       mask, value, 5000))
+		drm_err(&dev_priv->drm,
+			"Panel status timeout: status %08x control %08x\n",
+			intel_de_read(dev_priv, pp_stat_reg),
+			intel_de_read(dev_priv, pp_ctrl_reg));
+
+	drm_dbg_kms(&dev_priv->drm, "Wait complete\n");
+}
+
+static void wait_panel_on(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_dbg_kms(&i915->drm, "Wait for panel power on\n");
+	wait_panel_status(intel_dp, IDLE_ON_MASK, IDLE_ON_VALUE);
+}
+
+static void wait_panel_off(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_dbg_kms(&i915->drm, "Wait for panel power off time\n");
+	wait_panel_status(intel_dp, IDLE_OFF_MASK, IDLE_OFF_VALUE);
+}
+
+static void wait_panel_power_cycle(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	ktime_t panel_power_on_time;
+	s64 panel_power_off_duration;
+
+	drm_dbg_kms(&i915->drm, "Wait for panel power cycle\n");
+
+	/* take the difference of currrent time and panel power off time
+	 * and then make panel wait for t11_t12 if needed. */
+	panel_power_on_time = ktime_get_boottime();
+	panel_power_off_duration = ktime_ms_delta(panel_power_on_time, intel_dp->pps.panel_power_off_time);
+
+	/* When we disable the VDD override bit last we have to do the manual
+	 * wait. */
+	if (panel_power_off_duration < (s64)intel_dp->pps.panel_power_cycle_delay)
+		wait_remaining_ms_from_jiffies(jiffies,
+				       intel_dp->pps.panel_power_cycle_delay - panel_power_off_duration);
+
+	wait_panel_status(intel_dp, IDLE_CYCLE_MASK, IDLE_CYCLE_VALUE);
+}
+
+void intel_pps_wait_power_cycle(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref)
+		wait_panel_power_cycle(intel_dp);
+}
+
+static void wait_backlight_on(struct intel_dp *intel_dp)
+{
+	wait_remaining_ms_from_jiffies(intel_dp->pps.last_power_on,
+				       intel_dp->pps.backlight_on_delay);
+}
+
+static void edp_wait_backlight_off(struct intel_dp *intel_dp)
+{
+	wait_remaining_ms_from_jiffies(intel_dp->pps.last_backlight_off,
+				       intel_dp->pps.backlight_off_delay);
+}
+
+/* Read the current pp_control value, unlocking the register if it
+ * is locked
+ */
+
+static  u32 ilk_get_pp_control(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 control;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	control = intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp));
+	if (drm_WARN_ON(&dev_priv->drm, !HAS_DDI(dev_priv) &&
+			(control & PANEL_UNLOCK_MASK) != PANEL_UNLOCK_REGS)) {
+		control &= ~PANEL_UNLOCK_MASK;
+		control |= PANEL_UNLOCK_REGS;
+	}
+	return control;
+}
+
+/*
+ * Must be paired with intel_pps_vdd_off_unlocked().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_pps_vdd_{on,off}() calls.
+ */
+bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u32 pp;
+	i915_reg_t pp_stat_reg, pp_ctrl_reg;
+	bool need_to_disable = !intel_dp->pps.want_panel_vdd;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return false;
+
+	cancel_delayed_work(&intel_dp->pps.panel_vdd_work);
+	intel_dp->pps.want_panel_vdd = true;
+
+	if (edp_have_panel_vdd(intel_dp))
+		return need_to_disable;
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.vdd_wakeref);
+	intel_dp->pps.vdd_wakeref = intel_display_power_get(dev_priv,
+							    intel_aux_power_domain(dig_port));
+
+	drm_dbg_kms(&dev_priv->drm, "Turning [ENCODER:%d:%s] VDD on\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	if (!edp_have_panel_power(intel_dp))
+		wait_panel_power_cycle(intel_dp);
+
+	pp = ilk_get_pp_control(intel_dp);
+	pp |= EDP_FORCE_VDD;
+
+	pp_stat_reg = _pp_stat_reg(intel_dp);
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	drm_dbg_kms(&dev_priv->drm, "PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+		    intel_de_read(dev_priv, pp_stat_reg),
+		    intel_de_read(dev_priv, pp_ctrl_reg));
+	/*
+	 * If the panel wasn't on, delay before accessing aux channel
+	 */
+	if (!edp_have_panel_power(intel_dp)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[ENCODER:%d:%s] panel power wasn't enabled\n",
+			    dig_port->base.base.base.id,
+			    dig_port->base.base.name);
+		msleep(intel_dp->pps.panel_power_up_delay);
+	}
+
+	return need_to_disable;
+}
+
+/*
+ * Must be paired with intel_pps_off().
+ * Nested calls to these functions are not allowed since
+ * we drop the lock. Caller must use some higher level
+ * locking to prevent nested calls from other threads.
+ */
+void intel_pps_vdd_on(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+	bool vdd;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	vdd = false;
+	with_intel_pps_lock(intel_dp, wakeref)
+		vdd = intel_pps_vdd_on_unlocked(intel_dp);
+	I915_STATE_WARN(!vdd, "[ENCODER:%d:%s] VDD already requested on\n",
+			dp_to_dig_port(intel_dp)->base.base.base.id,
+			dp_to_dig_port(intel_dp)->base.base.name);
+}
+
+static void intel_pps_vdd_off_sync_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port =
+		dp_to_dig_port(intel_dp);
+	u32 pp;
+	i915_reg_t pp_stat_reg, pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.want_panel_vdd);
+
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Turning [ENCODER:%d:%s] VDD off\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	pp = ilk_get_pp_control(intel_dp);
+	pp &= ~EDP_FORCE_VDD;
+
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+	pp_stat_reg = _pp_stat_reg(intel_dp);
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+
+	/* Make sure sequencer is idle before allowing subsequent activity */
+	drm_dbg_kms(&dev_priv->drm, "PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+		    intel_de_read(dev_priv, pp_stat_reg),
+		    intel_de_read(dev_priv, pp_ctrl_reg));
+
+	if ((pp & PANEL_POWER_ON) == 0)
+		intel_dp->pps.panel_power_off_time = ktime_get_boottime();
+
+	intel_display_power_put(dev_priv,
+				intel_aux_power_domain(dig_port),
+				fetch_and_zero(&intel_dp->pps.vdd_wakeref));
+}
+
+void intel_pps_vdd_off_sync(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	cancel_delayed_work_sync(&intel_dp->pps.panel_vdd_work);
+	/*
+	 * vdd might still be enabled due to the delayed vdd off.
+	 * Make sure vdd is actually turned off here.
+	 */
+	with_intel_pps_lock(intel_dp, wakeref)
+		intel_pps_vdd_off_sync_unlocked(intel_dp);
+}
+
+static void edp_panel_vdd_work(struct work_struct *__work)
+{
+	struct intel_pps *pps = container_of(to_delayed_work(__work),
+					     struct intel_pps, panel_vdd_work);
+	struct intel_dp *intel_dp = container_of(pps, struct intel_dp, pps);
+	intel_wakeref_t wakeref;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		if (!intel_dp->pps.want_panel_vdd)
+			intel_pps_vdd_off_sync_unlocked(intel_dp);
+	}
+}
+
+static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
+{
+	unsigned long delay;
+
+	/*
+	 * Queue the timer to fire a long time from now (relative to the power
+	 * down delay) to keep the panel power up across a sequence of
+	 * operations.
+	 */
+	delay = msecs_to_jiffies(intel_dp->pps.panel_power_cycle_delay * 5);
+	schedule_delayed_work(&intel_dp->pps.panel_vdd_work, delay);
+}
+
+/*
+ * Must be paired with edp_panel_vdd_on().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_pps_vdd_{on,off}() calls.
+ */
+void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	I915_STATE_WARN(!intel_dp->pps.want_panel_vdd, "[ENCODER:%d:%s] VDD not forced on",
+			dp_to_dig_port(intel_dp)->base.base.base.id,
+			dp_to_dig_port(intel_dp)->base.base.name);
+
+	intel_dp->pps.want_panel_vdd = false;
+
+	if (sync)
+		intel_pps_vdd_off_sync_unlocked(intel_dp);
+	else
+		edp_panel_vdd_schedule_off(intel_dp);
+}
+
+void intel_pps_on_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 pp;
+	i915_reg_t pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Turn [ENCODER:%d:%s] panel power on\n",
+		    dp_to_dig_port(intel_dp)->base.base.base.id,
+		    dp_to_dig_port(intel_dp)->base.base.name);
+
+	if (drm_WARN(&dev_priv->drm, edp_have_panel_power(intel_dp),
+		     "[ENCODER:%d:%s] panel power already on\n",
+		     dp_to_dig_port(intel_dp)->base.base.base.id,
+		     dp_to_dig_port(intel_dp)->base.base.name))
+		return;
+
+	wait_panel_power_cycle(intel_dp);
+
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+	pp = ilk_get_pp_control(intel_dp);
+	if (IS_GEN(dev_priv, 5)) {
+		/* ILK workaround: disable reset around power sequence */
+		pp &= ~PANEL_POWER_RESET;
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+
+	pp |= PANEL_POWER_ON;
+	if (!IS_GEN(dev_priv, 5))
+		pp |= PANEL_POWER_RESET;
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+
+	wait_panel_on(intel_dp);
+	intel_dp->pps.last_power_on = jiffies;
+
+	if (IS_GEN(dev_priv, 5)) {
+		pp |= PANEL_POWER_RESET; /* restore panel reset bit */
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+}
+
+void intel_pps_on(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref)
+		intel_pps_on_unlocked(intel_dp);
+}
+
+void intel_pps_off_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u32 pp;
+	i915_reg_t pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Turn [ENCODER:%d:%s] panel power off\n",
+		    dig_port->base.base.base.id, dig_port->base.base.name);
+
+	drm_WARN(&dev_priv->drm, !intel_dp->pps.want_panel_vdd,
+		 "Need [ENCODER:%d:%s] VDD to turn off panel\n",
+		 dig_port->base.base.base.id, dig_port->base.base.name);
+
+	pp = ilk_get_pp_control(intel_dp);
+	/* We need to switch off panel power _and_ force vdd, for otherwise some
+	 * panels get very unhappy and cease to work. */
+	pp &= ~(PANEL_POWER_ON | PANEL_POWER_RESET | EDP_FORCE_VDD |
+		EDP_BLC_ENABLE);
+
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+
+	intel_dp->pps.want_panel_vdd = false;
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+
+	wait_panel_off(intel_dp);
+	intel_dp->pps.panel_power_off_time = ktime_get_boottime();
+
+	/* We got a reference when we enabled the VDD. */
+	intel_display_power_put(dev_priv,
+				intel_aux_power_domain(dig_port),
+				fetch_and_zero(&intel_dp->pps.vdd_wakeref));
+}
+
+void intel_pps_off(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref)
+		intel_pps_off_unlocked(intel_dp);
+}
+
+/* Enable backlight in the panel power control. */
+void intel_pps_backlight_on(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	/*
+	 * If we enable the backlight right away following a panel power
+	 * on, we may see slight flicker as the panel syncs with the eDP
+	 * link.  So delay a bit to make sure the image is solid before
+	 * allowing it to appear.
+	 */
+	wait_backlight_on(intel_dp);
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+		u32 pp;
+
+		pp = ilk_get_pp_control(intel_dp);
+		pp |= EDP_BLC_ENABLE;
+
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+}
+
+/* Disable backlight in the panel power control. */
+void intel_pps_backlight_off(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+		u32 pp;
+
+		pp = ilk_get_pp_control(intel_dp);
+		pp &= ~EDP_BLC_ENABLE;
+
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+
+	intel_dp->pps.last_backlight_off = jiffies;
+	edp_wait_backlight_off(intel_dp);
+}
+
+/*
+ * Hook for controlling the panel power control backlight through the bl_power
+ * sysfs attribute. Take care to handle multiple calls.
+ */
+void intel_pps_backlight_power(struct intel_connector *connector, bool enable)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	intel_wakeref_t wakeref;
+	bool is_enabled;
+
+	is_enabled = false;
+	with_intel_pps_lock(intel_dp, wakeref)
+		is_enabled = ilk_get_pp_control(intel_dp) & EDP_BLC_ENABLE;
+	if (is_enabled == enable)
+		return;
+
+	drm_dbg_kms(&i915->drm, "panel power control backlight %s\n",
+		    enable ? "enable" : "disable");
+
+	if (enable)
+		intel_pps_backlight_on(intel_dp);
+	else
+		intel_pps_backlight_off(intel_dp);
+}
+
+static void vlv_detach_power_sequencer(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	enum pipe pipe = intel_dp->pps.pps_pipe;
+	i915_reg_t pp_on_reg = PP_ON_DELAYS(pipe);
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE);
+
+	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
+		return;
+
+	intel_pps_vdd_off_sync_unlocked(intel_dp);
+
+	/*
+	 * VLV seems to get confused when multiple power sequencers
+	 * have the same port selected (even if only one has power/vdd
+	 * enabled). The failure manifests as vlv_wait_port_ready() failing
+	 * CHV on the other hand doesn't seem to mind having the same port
+	 * selected in multiple power sequencers, but let's clear the
+	 * port select always when logically disconnecting a power sequencer
+	 * from a port.
+	 */
+	drm_dbg_kms(&dev_priv->drm,
+		    "detaching pipe %c power sequencer from [ENCODER:%d:%s]\n",
+		    pipe_name(pipe), dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+	intel_de_write(dev_priv, pp_on_reg, 0);
+	intel_de_posting_read(dev_priv, pp_on_reg);
+
+	intel_dp->pps.pps_pipe = INVALID_PIPE;
+}
+
+static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv,
+				      enum pipe pipe)
+{
+	struct intel_encoder *encoder;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		drm_WARN(&dev_priv->drm, intel_dp->pps.active_pipe == pipe,
+			 "stealing pipe %c power sequencer from active [ENCODER:%d:%s]\n",
+			 pipe_name(pipe), encoder->base.base.id,
+			 encoder->base.name);
+
+		if (intel_dp->pps.pps_pipe != pipe)
+			continue;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "stealing pipe %c power sequencer from [ENCODER:%d:%s]\n",
+			    pipe_name(pipe), encoder->base.base.id,
+			    encoder->base.name);
+
+		/* make sure vdd is off before we steal it */
+		vlv_detach_power_sequencer(intel_dp);
+	}
+}
+
+void vlv_pps_init(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE);
+
+	if (intel_dp->pps.pps_pipe != INVALID_PIPE &&
+	    intel_dp->pps.pps_pipe != crtc->pipe) {
+		/*
+		 * If another power sequencer was being used on this
+		 * port previously make sure to turn off vdd there while
+		 * we still have control of it.
+		 */
+		vlv_detach_power_sequencer(intel_dp);
+	}
+
+	/*
+	 * We may be stealing the power
+	 * sequencer from another port.
+	 */
+	vlv_steal_power_sequencer(dev_priv, crtc->pipe);
+
+	intel_dp->pps.active_pipe = crtc->pipe;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	/* now it's all ours */
+	intel_dp->pps.pps_pipe = crtc->pipe;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "initializing pipe %c power sequencer for [ENCODER:%d:%s]\n",
+		    pipe_name(intel_dp->pps.pps_pipe), encoder->base.base.id,
+		    encoder->base.name);
+
+	/* init power sequencer on this pipe and port */
+	pps_init_delays(intel_dp);
+	pps_init_registers(intel_dp, true);
+}
+
+static void intel_pps_vdd_sanitize(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
+
+	/*
+	 * The VDD bit needs a power domain reference, so if the bit is
+	 * already enabled when we boot or resume, grab this reference and
+	 * schedule a vdd off, so we don't hold on to the reference
+	 * indefinitely.
+	 */
+	drm_dbg_kms(&dev_priv->drm,
+		    "VDD left on by BIOS, adjusting state tracking\n");
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.vdd_wakeref);
+	intel_dp->pps.vdd_wakeref = intel_display_power_get(dev_priv,
+							    intel_aux_power_domain(dig_port));
+
+	edp_panel_vdd_schedule_off(intel_dp);
+}
+
+bool intel_pps_have_power(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+	bool have_power = false;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		have_power = edp_have_panel_power(intel_dp) &&
+						  edp_have_panel_vdd(intel_dp);
+	}
+
+	return have_power;
+}
+
+static void pps_init_timestamps(struct intel_dp *intel_dp)
+{
+	intel_dp->pps.panel_power_off_time = ktime_get_boottime();
+	intel_dp->pps.last_power_on = jiffies;
+	intel_dp->pps.last_backlight_off = jiffies;
+}
+
+static void
+intel_pps_readout_hw_state(struct intel_dp *intel_dp, struct edp_power_seq *seq)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 pp_on, pp_off, pp_ctl;
+	struct pps_registers regs;
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	pp_ctl = ilk_get_pp_control(intel_dp);
+
+	/* Ensure PPS is unlocked */
+	if (!HAS_DDI(dev_priv))
+		intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl);
+
+	pp_on = intel_de_read(dev_priv, regs.pp_on);
+	pp_off = intel_de_read(dev_priv, regs.pp_off);
+
+	/* Pull timing values out of registers */
+	seq->t1_t3 = REG_FIELD_GET(PANEL_POWER_UP_DELAY_MASK, pp_on);
+	seq->t8 = REG_FIELD_GET(PANEL_LIGHT_ON_DELAY_MASK, pp_on);
+	seq->t9 = REG_FIELD_GET(PANEL_LIGHT_OFF_DELAY_MASK, pp_off);
+	seq->t10 = REG_FIELD_GET(PANEL_POWER_DOWN_DELAY_MASK, pp_off);
+
+	if (i915_mmio_reg_valid(regs.pp_div)) {
+		u32 pp_div;
+
+		pp_div = intel_de_read(dev_priv, regs.pp_div);
+
+		seq->t11_t12 = REG_FIELD_GET(PANEL_POWER_CYCLE_DELAY_MASK, pp_div) * 1000;
+	} else {
+		seq->t11_t12 = REG_FIELD_GET(BXT_POWER_CYCLE_DELAY_MASK, pp_ctl) * 1000;
+	}
+}
+
+static void
+intel_pps_dump_state(const char *state_name, const struct edp_power_seq *seq)
+{
+	DRM_DEBUG_KMS("%s t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
+		      state_name,
+		      seq->t1_t3, seq->t8, seq->t9, seq->t10, seq->t11_t12);
+}
+
+static void
+intel_pps_verify_state(struct intel_dp *intel_dp)
+{
+	struct edp_power_seq hw;
+	struct edp_power_seq *sw = &intel_dp->pps.pps_delays;
+
+	intel_pps_readout_hw_state(intel_dp, &hw);
+
+	if (hw.t1_t3 != sw->t1_t3 || hw.t8 != sw->t8 || hw.t9 != sw->t9 ||
+	    hw.t10 != sw->t10 || hw.t11_t12 != sw->t11_t12) {
+		DRM_ERROR("PPS state mismatch\n");
+		intel_pps_dump_state("sw", sw);
+		intel_pps_dump_state("hw", &hw);
+	}
+}
+
+static void pps_init_delays(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct edp_power_seq cur, vbt, spec,
+		*final = &intel_dp->pps.pps_delays;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* already initialized? */
+	if (final->t11_t12 != 0)
+		return;
+
+	intel_pps_readout_hw_state(intel_dp, &cur);
+
+	intel_pps_dump_state("cur", &cur);
+
+	vbt = dev_priv->vbt.edp.pps;
+	/* On Toshiba Satellite P50-C-18C system the VBT T12 delay
+	 * of 500ms appears to be too short. Ocassionally the panel
+	 * just fails to power back on. Increasing the delay to 800ms
+	 * seems sufficient to avoid this problem.
+	 */
+	if (dev_priv->quirks & QUIRK_INCREASE_T12_DELAY) {
+		vbt.t11_t12 = max_t(u16, vbt.t11_t12, 1300 * 10);
+		drm_dbg_kms(&dev_priv->drm,
+			    "Increasing T12 panel delay as per the quirk to %d\n",
+			    vbt.t11_t12);
+	}
+	/* T11_T12 delay is special and actually in units of 100ms, but zero
+	 * based in the hw (so we need to add 100 ms). But the sw vbt
+	 * table multiplies it with 1000 to make it in units of 100usec,
+	 * too. */
+	vbt.t11_t12 += 100 * 10;
+
+	/* Upper limits from eDP 1.3 spec. Note that we use the clunky units of
+	 * our hw here, which are all in 100usec. */
+	spec.t1_t3 = 210 * 10;
+	spec.t8 = 50 * 10; /* no limit for t8, use t7 instead */
+	spec.t9 = 50 * 10; /* no limit for t9, make it symmetric with t8 */
+	spec.t10 = 500 * 10;
+	/* This one is special and actually in units of 100ms, but zero
+	 * based in the hw (so we need to add 100 ms). But the sw vbt
+	 * table multiplies it with 1000 to make it in units of 100usec,
+	 * too. */
+	spec.t11_t12 = (510 + 100) * 10;
+
+	intel_pps_dump_state("vbt", &vbt);
+
+	/* Use the max of the register settings and vbt. If both are
+	 * unset, fall back to the spec limits. */
+#define assign_final(field)	final->field = (max(cur.field, vbt.field) == 0 ? \
+				       spec.field : \
+				       max(cur.field, vbt.field))
+	assign_final(t1_t3);
+	assign_final(t8);
+	assign_final(t9);
+	assign_final(t10);
+	assign_final(t11_t12);
+#undef assign_final
+
+#define get_delay(field)	(DIV_ROUND_UP(final->field, 10))
+	intel_dp->pps.panel_power_up_delay = get_delay(t1_t3);
+	intel_dp->pps.backlight_on_delay = get_delay(t8);
+	intel_dp->pps.backlight_off_delay = get_delay(t9);
+	intel_dp->pps.panel_power_down_delay = get_delay(t10);
+	intel_dp->pps.panel_power_cycle_delay = get_delay(t11_t12);
+#undef get_delay
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "panel power up delay %d, power down delay %d, power cycle delay %d\n",
+		    intel_dp->pps.panel_power_up_delay,
+		    intel_dp->pps.panel_power_down_delay,
+		    intel_dp->pps.panel_power_cycle_delay);
+
+	drm_dbg_kms(&dev_priv->drm, "backlight on delay %d, off delay %d\n",
+		    intel_dp->pps.backlight_on_delay,
+		    intel_dp->pps.backlight_off_delay);
+
+	/*
+	 * We override the HW backlight delays to 1 because we do manual waits
+	 * on them. For T8, even BSpec recommends doing it. For T9, if we
+	 * don't do this, we'll end up waiting for the backlight off delay
+	 * twice: once when we do the manual sleep, and once when we disable
+	 * the panel and wait for the PP_STATUS bit to become zero.
+	 */
+	final->t8 = 1;
+	final->t9 = 1;
+
+	/*
+	 * HW has only a 100msec granularity for t11_t12 so round it up
+	 * accordingly.
+	 */
+	final->t11_t12 = roundup(final->t11_t12, 100 * 10);
+}
+
+static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 pp_on, pp_off, port_sel = 0;
+	int div = RUNTIME_INFO(dev_priv)->rawclk_freq / 1000;
+	struct pps_registers regs;
+	enum port port = dp_to_dig_port(intel_dp)->base.port;
+	const struct edp_power_seq *seq = &intel_dp->pps.pps_delays;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	/*
+	 * On some VLV machines the BIOS can leave the VDD
+	 * enabled even on power sequencers which aren't
+	 * hooked up to any port. This would mess up the
+	 * power domain tracking the first time we pick
+	 * one of these power sequencers for use since
+	 * intel_pps_vdd_on_unlocked() would notice that the VDD was
+	 * already on and therefore wouldn't grab the power
+	 * domain reference. Disable VDD first to avoid this.
+	 * This also avoids spuriously turning the VDD on as
+	 * soon as the new power sequencer gets initialized.
+	 */
+	if (force_disable_vdd) {
+		u32 pp = ilk_get_pp_control(intel_dp);
+
+		drm_WARN(&dev_priv->drm, pp & PANEL_POWER_ON,
+			 "Panel power already on\n");
+
+		if (pp & EDP_FORCE_VDD)
+			drm_dbg_kms(&dev_priv->drm,
+				    "VDD already on, disabling first\n");
+
+		pp &= ~EDP_FORCE_VDD;
+
+		intel_de_write(dev_priv, regs.pp_ctrl, pp);
+	}
+
+	pp_on = REG_FIELD_PREP(PANEL_POWER_UP_DELAY_MASK, seq->t1_t3) |
+		REG_FIELD_PREP(PANEL_LIGHT_ON_DELAY_MASK, seq->t8);
+	pp_off = REG_FIELD_PREP(PANEL_LIGHT_OFF_DELAY_MASK, seq->t9) |
+		REG_FIELD_PREP(PANEL_POWER_DOWN_DELAY_MASK, seq->t10);
+
+	/* Haswell doesn't have any port selection bits for the panel
+	 * power sequencer any more. */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		port_sel = PANEL_PORT_SELECT_VLV(port);
+	} else if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) {
+		switch (port) {
+		case PORT_A:
+			port_sel = PANEL_PORT_SELECT_DPA;
+			break;
+		case PORT_C:
+			port_sel = PANEL_PORT_SELECT_DPC;
+			break;
+		case PORT_D:
+			port_sel = PANEL_PORT_SELECT_DPD;
+			break;
+		default:
+			MISSING_CASE(port);
+			break;
+		}
+	}
+
+	pp_on |= port_sel;
+
+	intel_de_write(dev_priv, regs.pp_on, pp_on);
+	intel_de_write(dev_priv, regs.pp_off, pp_off);
+
+	/*
+	 * Compute the divisor for the pp clock, simply match the Bspec formula.
+	 */
+	if (i915_mmio_reg_valid(regs.pp_div)) {
+		intel_de_write(dev_priv, regs.pp_div,
+			       REG_FIELD_PREP(PP_REFERENCE_DIVIDER_MASK, (100 * div) / 2 - 1) | REG_FIELD_PREP(PANEL_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000)));
+	} else {
+		u32 pp_ctl;
+
+		pp_ctl = intel_de_read(dev_priv, regs.pp_ctrl);
+		pp_ctl &= ~BXT_POWER_CYCLE_DELAY_MASK;
+		pp_ctl |= REG_FIELD_PREP(BXT_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000));
+		intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "panel power sequencer register settings: PP_ON %#x, PP_OFF %#x, PP_DIV %#x\n",
+		    intel_de_read(dev_priv, regs.pp_on),
+		    intel_de_read(dev_priv, regs.pp_off),
+		    i915_mmio_reg_valid(regs.pp_div) ?
+		    intel_de_read(dev_priv, regs.pp_div) :
+		    (intel_de_read(dev_priv, regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK));
+}
+
+void intel_pps_encoder_reset(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		/*
+		 * Reinit the power sequencer also on the resume path, in case
+		 * BIOS did something nasty with it.
+		 */
+		if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+			vlv_initial_power_sequencer_setup(intel_dp);
+
+		pps_init_delays(intel_dp);
+		pps_init_registers(intel_dp, false);
+
+		intel_pps_vdd_sanitize(intel_dp);
+	}
+}
+
+void intel_pps_init(struct intel_dp *intel_dp)
+{
+	INIT_DELAYED_WORK(&intel_dp->pps.panel_vdd_work, edp_panel_vdd_work);
+
+	pps_init_timestamps(intel_dp);
+
+	intel_pps_encoder_reset(intel_dp);
+}
+
+void intel_pps_unlock_regs_wa(struct drm_i915_private *dev_priv)
+{
+	int pps_num;
+	int pps_idx;
+
+	if (HAS_DDI(dev_priv))
+		return;
+	/*
+	 * This w/a is needed at least on CPT/PPT, but to be sure apply it
+	 * everywhere where registers can be write protected.
+	 */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		pps_num = 2;
+	else
+		pps_num = 1;
+
+	for (pps_idx = 0; pps_idx < pps_num; pps_idx++) {
+		u32 val = intel_de_read(dev_priv, PP_CONTROL(pps_idx));
+
+		val = (val & ~PANEL_UNLOCK_MASK) | PANEL_UNLOCK_REGS;
+		intel_de_write(dev_priv, PP_CONTROL(pps_idx), val);
+	}
+}
+
+void intel_pps_setup(struct drm_i915_private *i915)
+{
+	if (HAS_PCH_SPLIT(i915) || IS_GEN9_LP(i915))
+		i915->pps_mmio_base = PCH_PPS_BASE;
+	else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		i915->pps_mmio_base = VLV_PPS_BASE;
+	else
+		i915->pps_mmio_base = PPS_BASE;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_pps.h b/drivers/gpu/drm/i915/display/intel_pps.h
new file mode 100644
index 000000000000..fbbcca782e7b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pps.h
@@ -0,0 +1,52 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_PPS_H__
+#define __INTEL_PPS_H__
+
+#include <linux/types.h>
+
+#include "intel_wakeref.h"
+
+struct drm_i915_private;
+struct intel_connector;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_encoder;
+
+intel_wakeref_t intel_pps_lock(struct intel_dp *intel_dp);
+intel_wakeref_t intel_pps_unlock(struct intel_dp *intel_dp, intel_wakeref_t wakeref);
+
+#define with_intel_pps_lock(dp, wf)						\
+	for ((wf) = intel_pps_lock(dp); (wf); (wf) = intel_pps_unlock((dp), (wf)))
+
+void intel_pps_backlight_on(struct intel_dp *intel_dp);
+void intel_pps_backlight_off(struct intel_dp *intel_dp);
+void intel_pps_backlight_power(struct intel_connector *connector, bool enable);
+
+bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp);
+void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync);
+void intel_pps_on_unlocked(struct intel_dp *intel_dp);
+void intel_pps_off_unlocked(struct intel_dp *intel_dp);
+void intel_pps_check_power_unlocked(struct intel_dp *intel_dp);
+
+void intel_pps_vdd_on(struct intel_dp *intel_dp);
+void intel_pps_on(struct intel_dp *intel_dp);
+void intel_pps_off(struct intel_dp *intel_dp);
+void intel_pps_vdd_off_sync(struct intel_dp *intel_dp);
+bool intel_pps_have_power(struct intel_dp *intel_dp);
+void intel_pps_wait_power_cycle(struct intel_dp *intel_dp);
+
+void intel_pps_init(struct intel_dp *intel_dp);
+void intel_pps_encoder_reset(struct intel_dp *intel_dp);
+void intel_pps_reset_all(struct drm_i915_private *i915);
+
+void vlv_pps_init(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state);
+
+void intel_pps_unlock_regs_wa(struct drm_i915_private *i915);
+void intel_pps_setup(struct drm_i915_private *i915);
+
+#endif /* __INTEL_PPS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c
index d9a395c486d3..850cb7f5b332 100644
--- a/drivers/gpu/drm/i915/display/intel_psr.c
+++ b/drivers/gpu/drm/i915/display/intel_psr.c
@@ -28,9 +28,10 @@
 #include "i915_drv.h"
 #include "intel_atomic.h"
 #include "intel_display_types.h"
+#include "intel_dp_aux.h"
+#include "intel_hdmi.h"
 #include "intel_psr.h"
 #include "intel_sprite.h"
-#include "intel_hdmi.h"
 
 /**
  * DOC: Panel Self Refresh (PSR/SRD)
@@ -305,7 +306,7 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp)
 	drm_dbg_kms(&dev_priv->drm, "eDP panel supports PSR version %x\n",
 		    intel_dp->psr_dpcd[0]);
 
-	if (drm_dp_has_quirk(&intel_dp->desc, 0, DP_DPCD_QUIRK_NO_PSR)) {
+	if (drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_NO_PSR)) {
 		drm_dbg_kms(&dev_priv->drm,
 			    "PSR support not currently available for this panel\n");
 		return;
@@ -811,6 +812,13 @@ void intel_psr_compute_config(struct intel_dp *intel_dp,
 		&crtc_state->hw.adjusted_mode;
 	int psr_setup_time;
 
+	/*
+	 * Current PSR panels dont work reliably with VRR enabled
+	 * So if VRR is enabled, do not enable PSR.
+	 */
+	if (crtc_state->vrr.enable)
+		return;
+
 	if (!CAN_PSR(dev_priv))
 		return;
 
@@ -1186,7 +1194,8 @@ void intel_psr2_program_plane_sel_fetch(struct intel_plane *plane,
 	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
 	enum pipe pipe = plane->pipe;
 	const struct drm_rect *clip;
-	u32 val;
+	u32 val, offset;
+	int ret, x, y;
 
 	if (!crtc_state->enable_psr2_sel_fetch)
 		return;
@@ -1203,9 +1212,14 @@ void intel_psr2_program_plane_sel_fetch(struct intel_plane *plane,
 	val |= plane_state->uapi.dst.x1;
 	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_POS(pipe, plane->id), val);
 
-	/* TODO: consider tiling and auxiliary surfaces */
-	val = (clip->y1 + plane_state->color_plane[color_plane].y) << 16;
-	val |= plane_state->color_plane[color_plane].x;
+	/* TODO: consider auxiliary surfaces */
+	x = plane_state->uapi.src.x1 >> 16;
+	y = (plane_state->uapi.src.y1 >> 16) + clip->y1;
+	ret = skl_calc_main_surface_offset(plane_state, &x, &y, &offset);
+	if (ret)
+		drm_warn_once(&dev_priv->drm, "skl_calc_main_surface_offset() returned %i\n",
+			      ret);
+	val = y << 16 | x;
 	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_OFFSET(pipe, plane->id),
 			  val);
 
@@ -1242,9 +1256,11 @@ static void psr2_man_trk_ctl_calc(struct intel_crtc_state *crtc_state,
 	if (clip->y1 == -1)
 		goto exit;
 
+	drm_WARN_ON(crtc_state->uapi.crtc->dev, clip->y1 % 4 || clip->y2 % 4);
+
 	val |= PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE;
 	val |= PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR(clip->y1 / 4 + 1);
-	val |= PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(DIV_ROUND_UP(clip->y2, 4) + 1);
+	val |= PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(clip->y2 / 4 + 1);
 exit:
 	crtc_state->psr2_man_track_ctl = val;
 }
@@ -1269,8 +1285,8 @@ int intel_psr2_sel_fetch_update(struct intel_atomic_state *state,
 				struct intel_crtc *crtc)
 {
 	struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_rect pipe_clip = { .x1 = 0, .y1 = -1, .x2 = INT_MAX, .y2 = -1 };
 	struct intel_plane_state *new_plane_state, *old_plane_state;
-	struct drm_rect pipe_clip = { .y1 = -1 };
 	struct intel_plane *plane;
 	bool full_update = false;
 	int i, ret;
@@ -1282,13 +1298,25 @@ int intel_psr2_sel_fetch_update(struct intel_atomic_state *state,
 	if (ret)
 		return ret;
 
+	/*
+	 * Calculate minimal selective fetch area of each plane and calculate
+	 * the pipe damaged area.
+	 * In the next loop the plane selective fetch area will actually be set
+	 * using whole pipe damaged area.
+	 */
 	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
 					     new_plane_state, i) {
-		struct drm_rect *sel_fetch_area, temp;
+		struct drm_rect src, damaged_area = { .y1 = -1 };
+		struct drm_mode_rect *damaged_clips;
+		u32 num_clips, j;
 
 		if (new_plane_state->uapi.crtc != crtc_state->uapi.crtc)
 			continue;
 
+		if (!new_plane_state->uapi.visible &&
+		    !old_plane_state->uapi.visible)
+			continue;
+
 		/*
 		 * TODO: Not clear how to handle planes with negative position,
 		 * also planes are not updated if they have a negative X
@@ -1300,23 +1328,94 @@ int intel_psr2_sel_fetch_update(struct intel_atomic_state *state,
 			break;
 		}
 
-		if (!new_plane_state->uapi.visible)
-			continue;
+		num_clips = drm_plane_get_damage_clips_count(&new_plane_state->uapi);
 
 		/*
-		 * For now doing a selective fetch in the whole plane area,
-		 * optimizations will come in the future.
+		 * If visibility or plane moved, mark the whole plane area as
+		 * damaged as it needs to be complete redraw in the new and old
+		 * position.
 		 */
-		sel_fetch_area = &new_plane_state->psr2_sel_fetch_area;
-		sel_fetch_area->y1 = new_plane_state->uapi.src.y1 >> 16;
-		sel_fetch_area->y2 = new_plane_state->uapi.src.y2 >> 16;
+		if (new_plane_state->uapi.visible != old_plane_state->uapi.visible ||
+		    !drm_rect_equals(&new_plane_state->uapi.dst,
+				     &old_plane_state->uapi.dst)) {
+			if (old_plane_state->uapi.visible) {
+				damaged_area.y1 = old_plane_state->uapi.dst.y1;
+				damaged_area.y2 = old_plane_state->uapi.dst.y2;
+				clip_area_update(&pipe_clip, &damaged_area);
+			}
+
+			if (new_plane_state->uapi.visible) {
+				damaged_area.y1 = new_plane_state->uapi.dst.y1;
+				damaged_area.y2 = new_plane_state->uapi.dst.y2;
+				clip_area_update(&pipe_clip, &damaged_area);
+			}
+			continue;
+		} else if (new_plane_state->uapi.alpha != old_plane_state->uapi.alpha ||
+			   (!num_clips &&
+			    new_plane_state->uapi.fb != old_plane_state->uapi.fb)) {
+			/*
+			 * If the plane don't have damaged areas but the
+			 * framebuffer changed or alpha changed, mark the whole
+			 * plane area as damaged.
+			 */
+			damaged_area.y1 = new_plane_state->uapi.dst.y1;
+			damaged_area.y2 = new_plane_state->uapi.dst.y2;
+			clip_area_update(&pipe_clip, &damaged_area);
+			continue;
+		}
 
-		temp = *sel_fetch_area;
-		temp.y1 += new_plane_state->uapi.dst.y1;
-		temp.y2 += new_plane_state->uapi.dst.y2;
-		clip_area_update(&pipe_clip, &temp);
+		drm_rect_fp_to_int(&src, &new_plane_state->uapi.src);
+		damaged_clips = drm_plane_get_damage_clips(&new_plane_state->uapi);
+
+		for (j = 0; j < num_clips; j++) {
+			struct drm_rect clip;
+
+			clip.x1 = damaged_clips[j].x1;
+			clip.y1 = damaged_clips[j].y1;
+			clip.x2 = damaged_clips[j].x2;
+			clip.y2 = damaged_clips[j].y2;
+			if (drm_rect_intersect(&clip, &src))
+				clip_area_update(&damaged_area, &clip);
+		}
+
+		if (damaged_area.y1 == -1)
+			continue;
+
+		damaged_area.y1 += new_plane_state->uapi.dst.y1 - src.y1;
+		damaged_area.y2 += new_plane_state->uapi.dst.y1 - src.y1;
+		clip_area_update(&pipe_clip, &damaged_area);
+	}
+
+	if (full_update)
+		goto skip_sel_fetch_set_loop;
+
+	/* It must be aligned to 4 lines */
+	pipe_clip.y1 -= pipe_clip.y1 % 4;
+	if (pipe_clip.y2 % 4)
+		pipe_clip.y2 = ((pipe_clip.y2 / 4) + 1) * 4;
+
+	/*
+	 * Now that we have the pipe damaged area check if it intersect with
+	 * every plane, if it does set the plane selective fetch area.
+	 */
+	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
+					     new_plane_state, i) {
+		struct drm_rect *sel_fetch_area, inter;
+
+		if (new_plane_state->uapi.crtc != crtc_state->uapi.crtc ||
+		    !new_plane_state->uapi.visible)
+			continue;
+
+		inter = pipe_clip;
+		if (!drm_rect_intersect(&inter, &new_plane_state->uapi.dst))
+			continue;
+
+		sel_fetch_area = &new_plane_state->psr2_sel_fetch_area;
+		sel_fetch_area->y1 = inter.y1 - new_plane_state->uapi.dst.y1;
+		sel_fetch_area->y2 = inter.y2 - new_plane_state->uapi.dst.y1;
 	}
 
+skip_sel_fetch_set_loop:
 	psr2_man_trk_ctl_calc(crtc_state, &pipe_clip, full_update);
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/display/intel_sprite.c b/drivers/gpu/drm/i915/display/intel_sprite.c
index 8582b0ec6964..72ed728060e3 100644
--- a/drivers/gpu/drm/i915/display/intel_sprite.c
+++ b/drivers/gpu/drm/i915/display/intel_sprite.c
@@ -50,6 +50,7 @@
 #include "intel_dsi.h"
 #include "intel_sprite.h"
 #include "i9xx_plane.h"
+#include "intel_vrr.h"
 
 int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode,
 			     int usecs)
@@ -62,6 +63,16 @@ int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode,
 			    1000 * adjusted_mode->crtc_htotal);
 }
 
+static int intel_mode_vblank_start(const struct drm_display_mode *mode)
+{
+	int vblank_start = mode->crtc_vblank_start;
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		vblank_start = DIV_ROUND_UP(vblank_start, 2);
+
+	return vblank_start;
+}
+
 /**
  * intel_pipe_update_start() - start update of a set of display registers
  * @new_crtc_state: the new crtc state
@@ -90,9 +101,10 @@ void intel_pipe_update_start(const struct intel_crtc_state *new_crtc_state)
 	if (new_crtc_state->uapi.async_flip)
 		return;
 
-	vblank_start = adjusted_mode->crtc_vblank_start;
-	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
-		vblank_start = DIV_ROUND_UP(vblank_start, 2);
+	if (new_crtc_state->vrr.enable)
+		vblank_start = intel_vrr_vmax_vblank_start(new_crtc_state);
+	else
+		vblank_start = intel_mode_vblank_start(adjusted_mode);
 
 	/* FIXME needs to be calibrated sensibly */
 	min = vblank_start - intel_usecs_to_scanlines(adjusted_mode,
@@ -258,6 +270,9 @@ void intel_pipe_update_end(struct intel_crtc_state *new_crtc_state)
 
 	local_irq_enable();
 
+	/* Send VRR Push to terminate Vblank */
+	intel_vrr_send_push(new_crtc_state);
+
 	if (intel_vgpu_active(dev_priv))
 		return;
 
@@ -634,13 +649,19 @@ skl_program_scaler(struct intel_plane *plane,
 
 /* Preoffset values for YUV to RGB Conversion */
 #define PREOFF_YUV_TO_RGB_HI		0x1800
-#define PREOFF_YUV_TO_RGB_ME		0x1F00
+#define PREOFF_YUV_TO_RGB_ME		0x0000
 #define PREOFF_YUV_TO_RGB_LO		0x1800
 
 #define  ROFF(x)          (((x) & 0xffff) << 16)
 #define  GOFF(x)          (((x) & 0xffff) << 0)
 #define  BOFF(x)          (((x) & 0xffff) << 16)
 
+/*
+ * Programs the input color space conversion stage for ICL HDR planes.
+ * Note that it is assumed that this stage always happens after YUV
+ * range correction. Thus, the input to this stage is assumed to be
+ * in full-range YCbCr.
+ */
 static void
 icl_program_input_csc(struct intel_plane *plane,
 		      const struct intel_crtc_state *crtc_state,
@@ -688,52 +709,7 @@ icl_program_input_csc(struct intel_plane *plane,
 			0x0, 0x7800, 0x7F10,
 		},
 	};
-
-	/* Matrix for Limited Range to Full Range Conversion */
-	static const u16 input_csc_matrix_lr[][9] = {
-		/*
-		 * BT.601 Limted range YCbCr -> full range RGB
-		 * The matrix required is :
-		 * [1.164384, 0.000, 1.596027,
-		 *  1.164384, -0.39175, -0.812813,
-		 *  1.164384, 2.017232, 0.0000]
-		 */
-		[DRM_COLOR_YCBCR_BT601] = {
-			0x7CC8, 0x7950, 0x0,
-			0x8D00, 0x7950, 0x9C88,
-			0x0, 0x7950, 0x6810,
-		},
-		/*
-		 * BT.709 Limited range YCbCr -> full range RGB
-		 * The matrix required is :
-		 * [1.164384, 0.000, 1.792741,
-		 *  1.164384, -0.213249, -0.532909,
-		 *  1.164384, 2.112402, 0.0000]
-		 */
-		[DRM_COLOR_YCBCR_BT709] = {
-			0x7E58, 0x7950, 0x0,
-			0x8888, 0x7950, 0xADA8,
-			0x0, 0x7950,  0x6870,
-		},
-		/*
-		 * BT.2020 Limited range YCbCr -> full range RGB
-		 * The matrix required is :
-		 * [1.164, 0.000, 1.678,
-		 *  1.164, -0.1873, -0.6504,
-		 *  1.164, 2.1417, 0.0000]
-		 */
-		[DRM_COLOR_YCBCR_BT2020] = {
-			0x7D70, 0x7950, 0x0,
-			0x8A68, 0x7950, 0xAC00,
-			0x0, 0x7950, 0x6890,
-		},
-	};
-	const u16 *csc;
-
-	if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
-		csc = input_csc_matrix[plane_state->hw.color_encoding];
-	else
-		csc = input_csc_matrix_lr[plane_state->hw.color_encoding];
+	const u16 *csc = input_csc_matrix[plane_state->hw.color_encoding];
 
 	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 0),
 			  ROFF(csc[0]) | GOFF(csc[1]));
@@ -750,14 +726,8 @@ icl_program_input_csc(struct intel_plane *plane,
 
 	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 0),
 			  PREOFF_YUV_TO_RGB_HI);
-	if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
-		intel_de_write_fw(dev_priv,
-				  PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 1),
-				  0);
-	else
-		intel_de_write_fw(dev_priv,
-				  PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 1),
-				  PREOFF_YUV_TO_RGB_ME);
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 1),
+			  PREOFF_YUV_TO_RGB_ME);
 	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 2),
 			  PREOFF_YUV_TO_RGB_LO);
 	intel_de_write_fw(dev_priv,
@@ -771,7 +741,8 @@ icl_program_input_csc(struct intel_plane *plane,
 static void
 skl_plane_async_flip(struct intel_plane *plane,
 		     const struct intel_crtc_state *crtc_state,
-		     const struct intel_plane_state *plane_state)
+		     const struct intel_plane_state *plane_state,
+		     bool async_flip)
 {
 	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
 	unsigned long irqflags;
@@ -782,6 +753,9 @@ skl_plane_async_flip(struct intel_plane *plane,
 
 	plane_ctl |= skl_plane_ctl_crtc(crtc_state);
 
+	if (async_flip)
+		plane_ctl |= PLANE_CTL_ASYNC_FLIP;
+
 	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
 
 	intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), plane_ctl);
@@ -867,6 +841,10 @@ skl_program_plane(struct intel_plane *plane,
 	if (fb->format->is_yuv && icl_is_hdr_plane(dev_priv, plane_id))
 		icl_program_input_csc(plane, crtc_state, plane_state);
 
+	if (fb->modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC)
+		intel_uncore_write64_fw(&dev_priv->uncore,
+					PLANE_CC_VAL(pipe, plane_id), plane_state->ccval);
+
 	skl_write_plane_wm(plane, crtc_state);
 
 	intel_de_write_fw(dev_priv, PLANE_KEYVAL(pipe, plane_id),
@@ -958,6 +936,28 @@ skl_plane_get_hw_state(struct intel_plane *plane,
 	return ret;
 }
 
+static void
+skl_plane_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_enable_pipe_irq(i915, pipe, GEN9_PIPE_PLANE_FLIP_DONE(plane->id));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+skl_plane_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_disable_pipe_irq(i915, pipe, GEN9_PIPE_PLANE_FLIP_DONE(plane->id));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
 static void i9xx_plane_linear_gamma(u16 gamma[8])
 {
 	/* The points are not evenly spaced. */
@@ -1851,7 +1851,26 @@ g4x_sprite_max_stride(struct intel_plane *plane,
 		      u32 pixel_format, u64 modifier,
 		      unsigned int rotation)
 {
-	return 16384;
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 4k pixels to guarantee TILEOFF.x doesn't get too big. */
+	if (modifier == I915_FORMAT_MOD_X_TILED)
+		return min(4096 * cpp, 16 * 1024);
+	else
+		return 16 * 1024;
+}
+
+static unsigned int
+hsw_sprite_max_stride(struct intel_plane *plane,
+		      u32 pixel_format, u64 modifier,
+		      unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 8k pixels to guarantee OFFSET.x doesn't get too big. */
+	return min(8192 * cpp, 16 * 1024);
 }
 
 static u32 g4x_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state)
@@ -2364,7 +2383,8 @@ static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state,
 	     fb->modifier == I915_FORMAT_MOD_Y_TILED_CCS ||
 	     fb->modifier == I915_FORMAT_MOD_Yf_TILED_CCS ||
 	     fb->modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS ||
-	     fb->modifier == I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS)) {
+	     fb->modifier == I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS ||
+	     fb->modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC)) {
 		drm_dbg_kms(&dev_priv->drm,
 			    "Y/Yf tiling not supported in IF-ID mode\n");
 		return -EINVAL;
@@ -2854,6 +2874,7 @@ static const u64 skl_plane_format_modifiers_ccs[] = {
 static const u64 gen12_plane_format_modifiers_mc_ccs[] = {
 	I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS,
 	I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS,
+	I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC,
 	I915_FORMAT_MOD_Y_TILED,
 	I915_FORMAT_MOD_X_TILED,
 	DRM_FORMAT_MOD_LINEAR,
@@ -2862,6 +2883,7 @@ static const u64 gen12_plane_format_modifiers_mc_ccs[] = {
 
 static const u64 gen12_plane_format_modifiers_rc_ccs[] = {
 	I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS,
+	I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC,
 	I915_FORMAT_MOD_Y_TILED,
 	I915_FORMAT_MOD_X_TILED,
 	DRM_FORMAT_MOD_LINEAR,
@@ -3052,6 +3074,7 @@ static bool gen12_plane_format_mod_supported(struct drm_plane *_plane,
 	case I915_FORMAT_MOD_X_TILED:
 	case I915_FORMAT_MOD_Y_TILED:
 	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
 		break;
 	default:
 		return false;
@@ -3288,7 +3311,13 @@ skl_universal_plane_create(struct drm_i915_private *dev_priv,
 	plane->get_hw_state = skl_plane_get_hw_state;
 	plane->check_plane = skl_plane_check;
 	plane->min_cdclk = skl_plane_min_cdclk;
-	plane->async_flip = skl_plane_async_flip;
+
+	if (plane_id == PLANE_PRIMARY) {
+		plane->need_async_flip_disable_wa = IS_GEN_RANGE(dev_priv, 9, 10);
+		plane->async_flip = skl_plane_async_flip;
+		plane->enable_flip_done = skl_plane_enable_flip_done;
+		plane->disable_flip_done = skl_plane_disable_flip_done;
+	}
 
 	if (INTEL_GEN(dev_priv) >= 11)
 		formats = icl_get_plane_formats(dev_priv, pipe,
@@ -3396,11 +3425,11 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv,
 		return plane;
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
-		plane->max_stride = i9xx_plane_max_stride;
 		plane->update_plane = vlv_update_plane;
 		plane->disable_plane = vlv_disable_plane;
 		plane->get_hw_state = vlv_plane_get_hw_state;
 		plane->check_plane = vlv_sprite_check;
+		plane->max_stride = i965_plane_max_stride;
 		plane->min_cdclk = vlv_plane_min_cdclk;
 
 		if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B) {
@@ -3414,16 +3443,18 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv,
 
 		plane_funcs = &vlv_sprite_funcs;
 	} else if (INTEL_GEN(dev_priv) >= 7) {
-		plane->max_stride = g4x_sprite_max_stride;
 		plane->update_plane = ivb_update_plane;
 		plane->disable_plane = ivb_disable_plane;
 		plane->get_hw_state = ivb_plane_get_hw_state;
 		plane->check_plane = g4x_sprite_check;
 
-		if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv)) {
+			plane->max_stride = hsw_sprite_max_stride;
 			plane->min_cdclk = hsw_plane_min_cdclk;
-		else
+		} else {
+			plane->max_stride = g4x_sprite_max_stride;
 			plane->min_cdclk = ivb_sprite_min_cdclk;
+		}
 
 		formats = snb_plane_formats;
 		num_formats = ARRAY_SIZE(snb_plane_formats);
@@ -3431,11 +3462,11 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv,
 
 		plane_funcs = &snb_sprite_funcs;
 	} else {
-		plane->max_stride = g4x_sprite_max_stride;
 		plane->update_plane = g4x_update_plane;
 		plane->disable_plane = g4x_disable_plane;
 		plane->get_hw_state = g4x_plane_get_hw_state;
 		plane->check_plane = g4x_sprite_check;
+		plane->max_stride = g4x_sprite_max_stride;
 		plane->min_cdclk = g4x_sprite_min_cdclk;
 
 		modifiers = i9xx_plane_format_modifiers;
diff --git a/drivers/gpu/drm/i915/display/intel_tc.c b/drivers/gpu/drm/i915/display/intel_tc.c
index 27dc2dad6809..2cefc13535a0 100644
--- a/drivers/gpu/drm/i915/display/intel_tc.c
+++ b/drivers/gpu/drm/i915/display/intel_tc.c
@@ -23,36 +23,6 @@ static const char *tc_port_mode_name(enum tc_port_mode mode)
 	return names[mode];
 }
 
-static void
-tc_port_load_fia_params(struct drm_i915_private *i915,
-			struct intel_digital_port *dig_port)
-{
-	enum port port = dig_port->base.port;
-	enum tc_port tc_port = intel_port_to_tc(i915, port);
-	u32 modular_fia;
-
-	if (INTEL_INFO(i915)->display.has_modular_fia) {
-		modular_fia = intel_uncore_read(&i915->uncore,
-						PORT_TX_DFLEXDPSP(FIA1));
-		drm_WARN_ON(&i915->drm, modular_fia == 0xffffffff);
-		modular_fia &= MODULAR_FIA_MASK;
-	} else {
-		modular_fia = 0;
-	}
-
-	/*
-	 * Each Modular FIA instance houses 2 TC ports. In SOC that has more
-	 * than two TC ports, there are multiple instances of Modular FIA.
-	 */
-	if (modular_fia) {
-		dig_port->tc_phy_fia = tc_port / 2;
-		dig_port->tc_phy_fia_idx = tc_port % 2;
-	} else {
-		dig_port->tc_phy_fia = FIA1;
-		dig_port->tc_phy_fia_idx = tc_port;
-	}
-}
-
 static enum intel_display_power_domain
 tc_cold_get_power_domain(struct intel_digital_port *dig_port)
 {
@@ -646,6 +616,43 @@ void intel_tc_port_put_link(struct intel_digital_port *dig_port)
 	mutex_unlock(&dig_port->tc_lock);
 }
 
+static bool
+tc_has_modular_fia(struct drm_i915_private *i915, struct intel_digital_port *dig_port)
+{
+	intel_wakeref_t wakeref;
+	u32 val;
+
+	if (!INTEL_INFO(i915)->display.has_modular_fia)
+		return false;
+
+	wakeref = tc_cold_block(dig_port);
+	val = intel_uncore_read(&i915->uncore, PORT_TX_DFLEXDPSP(FIA1));
+	tc_cold_unblock(dig_port, wakeref);
+
+	drm_WARN_ON(&i915->drm, val == 0xffffffff);
+
+	return val & MODULAR_FIA_MASK;
+}
+
+static void
+tc_port_load_fia_params(struct drm_i915_private *i915, struct intel_digital_port *dig_port)
+{
+	enum port port = dig_port->base.port;
+	enum tc_port tc_port = intel_port_to_tc(i915, port);
+
+	/*
+	 * Each Modular FIA instance houses 2 TC ports. In SOC that has more
+	 * than two TC ports, there are multiple instances of Modular FIA.
+	 */
+	if (tc_has_modular_fia(i915, dig_port)) {
+		dig_port->tc_phy_fia = tc_port / 2;
+		dig_port->tc_phy_fia_idx = tc_port % 2;
+	} else {
+		dig_port->tc_phy_fia = FIA1;
+		dig_port->tc_phy_fia_idx = tc_port;
+	}
+}
+
 void intel_tc_port_init(struct intel_digital_port *dig_port, bool is_legacy)
 {
 	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
diff --git a/drivers/gpu/drm/i915/display/intel_vbt_defs.h b/drivers/gpu/drm/i915/display/intel_vbt_defs.h
index 49b4b5fca941..187ec573de59 100644
--- a/drivers/gpu/drm/i915/display/intel_vbt_defs.h
+++ b/drivers/gpu/drm/i915/display/intel_vbt_defs.h
@@ -319,6 +319,8 @@ enum vbt_gmbus_ddi {
 	ICL_DDC_BUS_DDI_A = 0x1,
 	ICL_DDC_BUS_DDI_B,
 	TGL_DDC_BUS_DDI_C,
+	RKL_DDC_BUS_DDI_D = 0x3,
+	RKL_DDC_BUS_DDI_E,
 	ICL_DDC_BUS_PORT_1 = 0x4,
 	ICL_DDC_BUS_PORT_2,
 	ICL_DDC_BUS_PORT_3,
diff --git a/drivers/gpu/drm/i915/display/intel_vrr.c b/drivers/gpu/drm/i915/display/intel_vrr.c
new file mode 100644
index 000000000000..a9c2b2fd9252
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vrr.c
@@ -0,0 +1,209 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_vrr.h"
+
+bool intel_vrr_is_capable(struct drm_connector *connector)
+{
+	struct intel_dp *intel_dp;
+	const struct drm_display_info *info = &connector->display_info;
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+
+	if (connector->connector_type != DRM_MODE_CONNECTOR_eDP &&
+	    connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort)
+		return false;
+
+	intel_dp = intel_attached_dp(to_intel_connector(connector));
+	/*
+	 * DP Sink is capable of VRR video timings if
+	 * Ignore MSA bit is set in DPCD.
+	 * EDID monitor range also should be atleast 10 for reasonable
+	 * Adaptive Sync or Variable Refresh Rate end user experience.
+	 */
+	return HAS_VRR(i915) &&
+		drm_dp_sink_can_do_video_without_timing_msa(intel_dp->dpcd) &&
+		info->monitor_range.max_vfreq - info->monitor_range.min_vfreq > 10;
+}
+
+void
+intel_vrr_check_modeset(struct intel_atomic_state *state)
+{
+	int i;
+	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	struct intel_crtc *crtc;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (new_crtc_state->uapi.vrr_enabled !=
+		    old_crtc_state->uapi.vrr_enabled)
+			new_crtc_state->uapi.mode_changed = true;
+	}
+}
+
+/*
+ * Without VRR registers get latched at:
+ *  vblank_start
+ *
+ * With VRR the earliest registers can get latched is:
+ *  intel_vrr_vmin_vblank_start(), which if we want to maintain
+ *  the correct min vtotal is >=vblank_start+1
+ *
+ * The latest point registers can get latched is the vmax decision boundary:
+ *  intel_vrr_vmax_vblank_start()
+ *
+ * Between those two points the vblank exit starts (and hence registers get
+ * latched) ASAP after a push is sent.
+ *
+ * framestart_delay is programmable 0-3.
+ */
+static int intel_vrr_vblank_exit_length(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	/* The hw imposes the extra scanline before frame start */
+	return crtc_state->vrr.pipeline_full + i915->framestart_delay + 1;
+}
+
+int intel_vrr_vmin_vblank_start(const struct intel_crtc_state *crtc_state)
+{
+	/* Min vblank actually determined by flipline that is always >=vmin+1 */
+	return crtc_state->vrr.vmin + 1 - intel_vrr_vblank_exit_length(crtc_state);
+}
+
+int intel_vrr_vmax_vblank_start(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->vrr.vmax - intel_vrr_vblank_exit_length(crtc_state);
+}
+
+void
+intel_vrr_compute_config(struct intel_crtc_state *crtc_state,
+			 struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector =
+		to_intel_connector(conn_state->connector);
+	struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	const struct drm_display_info *info = &connector->base.display_info;
+	int vmin, vmax;
+
+	if (!intel_vrr_is_capable(&connector->base))
+		return;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return;
+
+	if (!crtc_state->uapi.vrr_enabled)
+		return;
+
+	vmin = DIV_ROUND_UP(adjusted_mode->crtc_clock * 1000,
+			    adjusted_mode->crtc_htotal * info->monitor_range.max_vfreq);
+	vmax = adjusted_mode->crtc_clock * 1000 /
+		(adjusted_mode->crtc_htotal * info->monitor_range.min_vfreq);
+
+	vmin = max_t(int, vmin, adjusted_mode->crtc_vtotal);
+	vmax = max_t(int, vmax, adjusted_mode->crtc_vtotal);
+
+	if (vmin >= vmax)
+		return;
+
+	/*
+	 * flipline determines the min vblank length the hardware will
+	 * generate, and flipline>=vmin+1, hence we reduce vmin by one
+	 * to make sure we can get the actual min vblank length.
+	 */
+	crtc_state->vrr.vmin = vmin - 1;
+	crtc_state->vrr.vmax = vmax;
+	crtc_state->vrr.enable = true;
+
+	crtc_state->vrr.flipline = crtc_state->vrr.vmin + 1;
+
+	/*
+	 * FIXME: s/4/framestart_delay+1/ to get consistent
+	 * earliest/latest points for register latching regardless
+	 * of the framestart_delay used?
+	 *
+	 * FIXME: this really needs the extra scanline to provide consistent
+	 * behaviour for all framestart_delay values. Otherwise with
+	 * framestart_delay==3 we will end up extending the min vblank by
+	 * one extra line.
+	 */
+	crtc_state->vrr.pipeline_full =
+		min(255, crtc_state->vrr.vmin - adjusted_mode->crtc_vdisplay - 4 - 1);
+
+	crtc_state->mode_flags |= I915_MODE_FLAG_VRR;
+}
+
+void intel_vrr_enable(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 trans_vrr_ctl;
+
+	if (!crtc_state->vrr.enable)
+		return;
+
+	trans_vrr_ctl = VRR_CTL_VRR_ENABLE |
+		VRR_CTL_IGN_MAX_SHIFT | VRR_CTL_FLIP_LINE_EN |
+		VRR_CTL_PIPELINE_FULL(crtc_state->vrr.pipeline_full) |
+		VRR_CTL_PIPELINE_FULL_OVERRIDE;
+
+	intel_de_write(dev_priv, TRANS_VRR_VMIN(cpu_transcoder), crtc_state->vrr.vmin - 1);
+	intel_de_write(dev_priv, TRANS_VRR_VMAX(cpu_transcoder), crtc_state->vrr.vmax - 1);
+	intel_de_write(dev_priv, TRANS_VRR_CTL(cpu_transcoder), trans_vrr_ctl);
+	intel_de_write(dev_priv, TRANS_VRR_FLIPLINE(cpu_transcoder), crtc_state->vrr.flipline - 1);
+	intel_de_write(dev_priv, TRANS_PUSH(cpu_transcoder), TRANS_PUSH_EN);
+}
+
+void intel_vrr_send_push(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (!crtc_state->vrr.enable)
+		return;
+
+	intel_de_write(dev_priv, TRANS_PUSH(cpu_transcoder),
+		       TRANS_PUSH_EN | TRANS_PUSH_SEND);
+}
+
+void intel_vrr_disable(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder;
+
+	if (!old_crtc_state->vrr.enable)
+		return;
+
+	intel_de_write(dev_priv, TRANS_VRR_CTL(cpu_transcoder), 0);
+	intel_de_write(dev_priv, TRANS_PUSH(cpu_transcoder), 0);
+}
+
+void intel_vrr_get_config(struct intel_crtc *crtc,
+			  struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 trans_vrr_ctl;
+
+	trans_vrr_ctl = intel_de_read(dev_priv, TRANS_VRR_CTL(cpu_transcoder));
+	crtc_state->vrr.enable = trans_vrr_ctl & VRR_CTL_VRR_ENABLE;
+	if (!crtc_state->vrr.enable)
+		return;
+
+	if (trans_vrr_ctl & VRR_CTL_PIPELINE_FULL_OVERRIDE)
+		crtc_state->vrr.pipeline_full = REG_FIELD_GET(VRR_CTL_PIPELINE_FULL_MASK, trans_vrr_ctl);
+	if (trans_vrr_ctl & VRR_CTL_FLIP_LINE_EN)
+		crtc_state->vrr.flipline = intel_de_read(dev_priv, TRANS_VRR_FLIPLINE(cpu_transcoder)) + 1;
+	crtc_state->vrr.vmax = intel_de_read(dev_priv, TRANS_VRR_VMAX(cpu_transcoder)) + 1;
+	crtc_state->vrr.vmin = intel_de_read(dev_priv, TRANS_VRR_VMIN(cpu_transcoder)) + 1;
+
+	crtc_state->mode_flags |= I915_MODE_FLAG_VRR;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_vrr.h b/drivers/gpu/drm/i915/display/intel_vrr.h
new file mode 100644
index 000000000000..fac01bf4ab50
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vrr.h
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_VRR_H__
+#define __INTEL_VRR_H__
+
+#include <linux/types.h>
+
+struct drm_connector;
+struct drm_connector_state;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_encoder;
+struct intel_crtc;
+
+bool intel_vrr_is_capable(struct drm_connector *connector);
+void intel_vrr_check_modeset(struct intel_atomic_state *state);
+void intel_vrr_compute_config(struct intel_crtc_state *crtc_state,
+			      struct drm_connector_state *conn_state);
+void intel_vrr_enable(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state);
+void intel_vrr_send_push(const struct intel_crtc_state *crtc_state);
+void intel_vrr_disable(const struct intel_crtc_state *old_crtc_state);
+void intel_vrr_get_config(struct intel_crtc *crtc,
+			  struct intel_crtc_state *crtc_state);
+int intel_vrr_vmax_vblank_start(const struct intel_crtc_state *crtc_state);
+int intel_vrr_vmin_vblank_start(const struct intel_crtc_state *crtc_state);
+
+#endif /* __INTEL_VRR_H__ */
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi.c b/drivers/gpu/drm/i915/display/vlv_dsi.c
index d52f9c177908..f94025ec603a 100644
--- a/drivers/gpu/drm/i915/display/vlv_dsi.c
+++ b/drivers/gpu/drm/i915/display/vlv_dsi.c
@@ -812,10 +812,20 @@ static void intel_dsi_pre_enable(struct intel_atomic_state *state,
 		intel_dsi_prepare(encoder, pipe_config);
 
 	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_POWER_ON);
-	intel_dsi_msleep(intel_dsi, intel_dsi->panel_on_delay);
 
-	/* Deassert reset */
-	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DEASSERT_RESET);
+	/*
+	 * Give the panel time to power-on and then deassert its reset.
+	 * Depending on the VBT MIPI sequences version the deassert-seq
+	 * may contain the necessary delay, intel_dsi_msleep() will skip
+	 * the delay in that case. If there is no deassert-seq, then an
+	 * unconditional msleep is used to give the panel time to power-on.
+	 */
+	if (dev_priv->vbt.dsi.sequence[MIPI_SEQ_DEASSERT_RESET]) {
+		intel_dsi_msleep(intel_dsi, intel_dsi->panel_on_delay);
+		intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DEASSERT_RESET);
+	} else {
+		msleep(intel_dsi->panel_on_delay);
+	}
 
 	if (IS_GEMINILAKE(dev_priv)) {
 		glk_cold_boot = glk_dsi_enable_io(encoder);
diff --git a/drivers/gpu/drm/i915/dma_resv_utils.c b/drivers/gpu/drm/i915/dma_resv_utils.c
new file mode 100644
index 000000000000..9e508e7d4629
--- /dev/null
+++ b/drivers/gpu/drm/i915/dma_resv_utils.c
@@ -0,0 +1,17 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/dma-resv.h>
+
+#include "dma_resv_utils.h"
+
+void dma_resv_prune(struct dma_resv *resv)
+{
+	if (dma_resv_trylock(resv)) {
+		if (dma_resv_test_signaled_rcu(resv, true))
+			dma_resv_add_excl_fence(resv, NULL);
+		dma_resv_unlock(resv);
+	}
+}
diff --git a/drivers/gpu/drm/i915/dma_resv_utils.h b/drivers/gpu/drm/i915/dma_resv_utils.h
new file mode 100644
index 000000000000..b9d8fb5f8367
--- /dev/null
+++ b/drivers/gpu/drm/i915/dma_resv_utils.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef DMA_RESV_UTILS_H
+#define DMA_RESV_UTILS_H
+
+struct dma_resv;
+
+void dma_resv_prune(struct dma_resv *resv);
+
+#endif /* DMA_RESV_UTILS_H */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index 4fd38101bb56..4d2f40cf237b 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -72,6 +72,8 @@
 #include "gt/intel_context_param.h"
 #include "gt/intel_engine_heartbeat.h"
 #include "gt/intel_engine_user.h"
+#include "gt/intel_execlists_submission.h" /* virtual_engine */
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_ring.h"
 
 #include "i915_gem_context.h"
@@ -333,13 +335,12 @@ static struct i915_gem_engines *default_engines(struct i915_gem_context *ctx)
 	return e;
 }
 
-static void i915_gem_context_free(struct i915_gem_context *ctx)
+void i915_gem_context_release(struct kref *ref)
 {
-	GEM_BUG_ON(!i915_gem_context_is_closed(ctx));
+	struct i915_gem_context *ctx = container_of(ref, typeof(*ctx), ref);
 
-	spin_lock(&ctx->i915->gem.contexts.lock);
-	list_del(&ctx->link);
-	spin_unlock(&ctx->i915->gem.contexts.lock);
+	trace_i915_context_free(ctx);
+	GEM_BUG_ON(!i915_gem_context_is_closed(ctx));
 
 	mutex_destroy(&ctx->engines_mutex);
 	mutex_destroy(&ctx->lut_mutex);
@@ -353,37 +354,6 @@ static void i915_gem_context_free(struct i915_gem_context *ctx)
 	kfree_rcu(ctx, rcu);
 }
 
-static void contexts_free_all(struct llist_node *list)
-{
-	struct i915_gem_context *ctx, *cn;
-
-	llist_for_each_entry_safe(ctx, cn, list, free_link)
-		i915_gem_context_free(ctx);
-}
-
-static void contexts_flush_free(struct i915_gem_contexts *gc)
-{
-	contexts_free_all(llist_del_all(&gc->free_list));
-}
-
-static void contexts_free_worker(struct work_struct *work)
-{
-	struct i915_gem_contexts *gc =
-		container_of(work, typeof(*gc), free_work);
-
-	contexts_flush_free(gc);
-}
-
-void i915_gem_context_release(struct kref *ref)
-{
-	struct i915_gem_context *ctx = container_of(ref, typeof(*ctx), ref);
-	struct i915_gem_contexts *gc = &ctx->i915->gem.contexts;
-
-	trace_i915_context_free(ctx);
-	if (llist_add(&ctx->free_link, &gc->free_list))
-		schedule_work(&gc->free_work);
-}
-
 static inline struct i915_gem_engines *
 __context_engines_static(const struct i915_gem_context *ctx)
 {
@@ -438,7 +408,7 @@ __active_engine(struct i915_request *rq, struct intel_engine_cs **active)
 	}
 
 	if (i915_request_is_active(rq)) {
-		if (!i915_request_completed(rq))
+		if (!__i915_request_is_complete(rq))
 			*active = locked;
 		ret = true;
 	}
@@ -453,6 +423,9 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
 	struct intel_engine_cs *engine = NULL;
 	struct i915_request *rq;
 
+	if (intel_context_has_inflight(ce))
+		return intel_context_inflight(ce);
+
 	if (!ce->timeline)
 		return NULL;
 
@@ -632,6 +605,10 @@ static void context_close(struct i915_gem_context *ctx)
 	 */
 	lut_close(ctx);
 
+	spin_lock(&ctx->i915->gem.contexts.lock);
+	list_del(&ctx->link);
+	spin_unlock(&ctx->i915->gem.contexts.lock);
+
 	mutex_unlock(&ctx->mutex);
 
 	/*
@@ -740,7 +717,8 @@ err_free:
 }
 
 static inline struct i915_gem_engines *
-__context_engines_await(const struct i915_gem_context *ctx)
+__context_engines_await(const struct i915_gem_context *ctx,
+			bool *user_engines)
 {
 	struct i915_gem_engines *engines;
 
@@ -749,6 +727,10 @@ __context_engines_await(const struct i915_gem_context *ctx)
 		engines = rcu_dereference(ctx->engines);
 		GEM_BUG_ON(!engines);
 
+		if (user_engines)
+			*user_engines = i915_gem_context_user_engines(ctx);
+
+		/* successful await => strong mb */
 		if (unlikely(!i915_sw_fence_await(&engines->fence)))
 			continue;
 
@@ -772,7 +754,7 @@ context_apply_all(struct i915_gem_context *ctx,
 	struct intel_context *ce;
 	int err = 0;
 
-	e = __context_engines_await(ctx);
+	e = __context_engines_await(ctx, NULL);
 	for_each_gem_engine(ce, e, it) {
 		err = fn(ce, data);
 		if (err)
@@ -849,9 +831,6 @@ i915_gem_create_context(struct drm_i915_private *i915, unsigned int flags)
 	    !HAS_EXECLISTS(i915))
 		return ERR_PTR(-EINVAL);
 
-	/* Reap the stale contexts */
-	contexts_flush_free(&i915->gem.contexts);
-
 	ctx = __create_context(i915);
 	if (IS_ERR(ctx))
 		return ctx;
@@ -896,23 +875,11 @@ static void init_contexts(struct i915_gem_contexts *gc)
 {
 	spin_lock_init(&gc->lock);
 	INIT_LIST_HEAD(&gc->list);
-
-	INIT_WORK(&gc->free_work, contexts_free_worker);
-	init_llist_head(&gc->free_list);
 }
 
 void i915_gem_init__contexts(struct drm_i915_private *i915)
 {
 	init_contexts(&i915->gem.contexts);
-	drm_dbg(&i915->drm, "%s context support initialized\n",
-		DRIVER_CAPS(i915)->has_logical_contexts ?
-		"logical" : "fake");
-}
-
-void i915_gem_driver_release__contexts(struct drm_i915_private *i915)
-{
-	flush_work(&i915->gem.contexts.free_work);
-	rcu_barrier(); /* and flush the left over RCU frees */
 }
 
 static int gem_context_register(struct i915_gem_context *ctx,
@@ -988,7 +955,6 @@ err:
 void i915_gem_context_close(struct drm_file *file)
 {
 	struct drm_i915_file_private *file_priv = file->driver_priv;
-	struct drm_i915_private *i915 = file_priv->dev_priv;
 	struct i915_address_space *vm;
 	struct i915_gem_context *ctx;
 	unsigned long idx;
@@ -1000,8 +966,6 @@ void i915_gem_context_close(struct drm_file *file)
 	xa_for_each(&file_priv->vm_xa, idx, vm)
 		i915_vm_put(vm);
 	xa_destroy(&file_priv->vm_xa);
-
-	contexts_flush_free(&i915->gem.contexts);
 }
 
 int i915_gem_vm_create_ioctl(struct drm_device *dev, void *data,
@@ -1116,7 +1080,7 @@ static int context_barrier_task(struct i915_gem_context *ctx,
 		return err;
 	}
 
-	e = __context_engines_await(ctx);
+	e = __context_engines_await(ctx, NULL);
 	if (!e) {
 		i915_active_release(&cb->base);
 		return -ENOENT;
@@ -1879,27 +1843,6 @@ replace:
 	return 0;
 }
 
-static struct i915_gem_engines *
-__copy_engines(struct i915_gem_engines *e)
-{
-	struct i915_gem_engines *copy;
-	unsigned int n;
-
-	copy = alloc_engines(e->num_engines);
-	if (!copy)
-		return ERR_PTR(-ENOMEM);
-
-	for (n = 0; n < e->num_engines; n++) {
-		if (e->engines[n])
-			copy->engines[n] = intel_context_get(e->engines[n]);
-		else
-			copy->engines[n] = NULL;
-	}
-	copy->num_engines = n;
-
-	return copy;
-}
-
 static int
 get_engines(struct i915_gem_context *ctx,
 	    struct drm_i915_gem_context_param *args)
@@ -1907,19 +1850,17 @@ get_engines(struct i915_gem_context *ctx,
 	struct i915_context_param_engines __user *user;
 	struct i915_gem_engines *e;
 	size_t n, count, size;
+	bool user_engines;
 	int err = 0;
 
-	err = mutex_lock_interruptible(&ctx->engines_mutex);
-	if (err)
-		return err;
+	e = __context_engines_await(ctx, &user_engines);
+	if (!e)
+		return -ENOENT;
 
-	e = NULL;
-	if (i915_gem_context_user_engines(ctx))
-		e = __copy_engines(i915_gem_context_engines(ctx));
-	mutex_unlock(&ctx->engines_mutex);
-	if (IS_ERR_OR_NULL(e)) {
+	if (!user_engines) {
+		i915_sw_fence_complete(&e->fence);
 		args->size = 0;
-		return PTR_ERR_OR_ZERO(e);
+		return 0;
 	}
 
 	count = e->num_engines;
@@ -1970,7 +1911,7 @@ get_engines(struct i915_gem_context *ctx,
 	args->size = size;
 
 err_free:
-	free_engines(e);
+	i915_sw_fence_complete(&e->fence);
 	return err;
 }
 
@@ -2136,11 +2077,14 @@ static int copy_ring_size(struct intel_context *dst,
 static int clone_engines(struct i915_gem_context *dst,
 			 struct i915_gem_context *src)
 {
-	struct i915_gem_engines *e = i915_gem_context_lock_engines(src);
-	struct i915_gem_engines *clone;
+	struct i915_gem_engines *clone, *e;
 	bool user_engines;
 	unsigned long n;
 
+	e = __context_engines_await(src, &user_engines);
+	if (!e)
+		return -ENOENT;
+
 	clone = alloc_engines(e->num_engines);
 	if (!clone)
 		goto err_unlock;
@@ -2182,9 +2126,7 @@ static int clone_engines(struct i915_gem_context *dst,
 		}
 	}
 	clone->num_engines = n;
-
-	user_engines = i915_gem_context_user_engines(src);
-	i915_gem_context_unlock_engines(src);
+	i915_sw_fence_complete(&e->fence);
 
 	/* Serialised by constructor */
 	engines_idle_release(dst, rcu_replace_pointer(dst->engines, clone, 1));
@@ -2195,7 +2137,7 @@ static int clone_engines(struct i915_gem_context *dst,
 	return 0;
 
 err_unlock:
-	i915_gem_context_unlock_engines(src);
+	i915_sw_fence_complete(&e->fence);
 	return -ENOMEM;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.h b/drivers/gpu/drm/i915/gem/i915_gem_context.h
index a133f92bbedb..b5c908f3f4f2 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.h
@@ -110,7 +110,6 @@ i915_gem_context_clear_user_engines(struct i915_gem_context *ctx)
 
 /* i915_gem_context.c */
 void i915_gem_init__contexts(struct drm_i915_private *i915);
-void i915_gem_driver_release__contexts(struct drm_i915_private *i915);
 
 int i915_gem_context_open(struct drm_i915_private *i915,
 			  struct drm_file *file);
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h
index ae14ca24a11f..1449f54924e0 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h
@@ -108,7 +108,6 @@ struct i915_gem_context {
 
 	/** link: place with &drm_i915_private.context_list */
 	struct list_head link;
-	struct llist_node free_link;
 
 	/**
 	 * @ref: reference count
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_create.c b/drivers/gpu/drm/i915/gem/i915_gem_create.c
new file mode 100644
index 000000000000..45d60e3d98e3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_create.c
@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "gem/i915_gem_ioctls.h"
+#include "gem/i915_gem_region.h"
+
+#include "i915_drv.h"
+
+static int
+i915_gem_create(struct drm_file *file,
+		struct intel_memory_region *mr,
+		u64 *size_p,
+		u32 *handle_p)
+{
+	struct drm_i915_gem_object *obj;
+	u32 handle;
+	u64 size;
+	int ret;
+
+	GEM_BUG_ON(!is_power_of_2(mr->min_page_size));
+	size = round_up(*size_p, mr->min_page_size);
+	if (size == 0)
+		return -EINVAL;
+
+	/* For most of the ABI (e.g. mmap) we think in system pages */
+	GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE));
+
+	/* Allocate the new object */
+	obj = i915_gem_object_create_region(mr, size, 0);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	GEM_BUG_ON(size != obj->base.size);
+
+	ret = drm_gem_handle_create(file, &obj->base, &handle);
+	/* drop reference from allocate - handle holds it now */
+	i915_gem_object_put(obj);
+	if (ret)
+		return ret;
+
+	*handle_p = handle;
+	*size_p = size;
+	return 0;
+}
+
+int
+i915_gem_dumb_create(struct drm_file *file,
+		     struct drm_device *dev,
+		     struct drm_mode_create_dumb *args)
+{
+	enum intel_memory_type mem_type;
+	int cpp = DIV_ROUND_UP(args->bpp, 8);
+	u32 format;
+
+	switch (cpp) {
+	case 1:
+		format = DRM_FORMAT_C8;
+		break;
+	case 2:
+		format = DRM_FORMAT_RGB565;
+		break;
+	case 4:
+		format = DRM_FORMAT_XRGB8888;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/* have to work out size/pitch and return them */
+	args->pitch = ALIGN(args->width * cpp, 64);
+
+	/* align stride to page size so that we can remap */
+	if (args->pitch > intel_plane_fb_max_stride(to_i915(dev), format,
+						    DRM_FORMAT_MOD_LINEAR))
+		args->pitch = ALIGN(args->pitch, 4096);
+
+	if (args->pitch < args->width)
+		return -EINVAL;
+
+	args->size = mul_u32_u32(args->pitch, args->height);
+
+	mem_type = INTEL_MEMORY_SYSTEM;
+	if (HAS_LMEM(to_i915(dev)))
+		mem_type = INTEL_MEMORY_LOCAL;
+
+	return i915_gem_create(file,
+			       intel_memory_region_by_type(to_i915(dev),
+							   mem_type),
+			       &args->size, &args->handle);
+}
+
+/**
+ * Creates a new mm object and returns a handle to it.
+ * @dev: drm device pointer
+ * @data: ioctl data blob
+ * @file: drm file pointer
+ */
+int
+i915_gem_create_ioctl(struct drm_device *dev, void *data,
+		      struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_create *args = data;
+
+	i915_gem_flush_free_objects(i915);
+
+	return i915_gem_create(file,
+			       intel_memory_region_by_type(i915,
+							   INTEL_MEMORY_SYSTEM),
+			       &args->size, &args->handle);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_domain.c b/drivers/gpu/drm/i915/gem/i915_gem_domain.c
index fcce6909f201..36f54cedaaeb 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_domain.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.c
@@ -5,6 +5,7 @@
  */
 
 #include "display/intel_frontbuffer.h"
+#include "gt/intel_gt.h"
 
 #include "i915_drv.h"
 #include "i915_gem_clflush.h"
@@ -15,13 +16,58 @@
 #include "i915_gem_lmem.h"
 #include "i915_gem_mman.h"
 
+static bool gpu_write_needs_clflush(struct drm_i915_gem_object *obj)
+{
+	return !(obj->cache_level == I915_CACHE_NONE ||
+		 obj->cache_level == I915_CACHE_WT);
+}
+
+static void
+flush_write_domain(struct drm_i915_gem_object *obj, unsigned int flush_domains)
+{
+	struct i915_vma *vma;
+
+	assert_object_held(obj);
+
+	if (!(obj->write_domain & flush_domains))
+		return;
+
+	switch (obj->write_domain) {
+	case I915_GEM_DOMAIN_GTT:
+		spin_lock(&obj->vma.lock);
+		for_each_ggtt_vma(vma, obj) {
+			if (i915_vma_unset_ggtt_write(vma))
+				intel_gt_flush_ggtt_writes(vma->vm->gt);
+		}
+		spin_unlock(&obj->vma.lock);
+
+		i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
+		break;
+
+	case I915_GEM_DOMAIN_WC:
+		wmb();
+		break;
+
+	case I915_GEM_DOMAIN_CPU:
+		i915_gem_clflush_object(obj, I915_CLFLUSH_SYNC);
+		break;
+
+	case I915_GEM_DOMAIN_RENDER:
+		if (gpu_write_needs_clflush(obj))
+			obj->cache_dirty = true;
+		break;
+	}
+
+	obj->write_domain = 0;
+}
+
 static void __i915_gem_object_flush_for_display(struct drm_i915_gem_object *obj)
 {
 	/*
 	 * We manually flush the CPU domain so that we can override and
 	 * force the flush for the display, and perform it asyncrhonously.
 	 */
-	i915_gem_object_flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
 	if (obj->cache_dirty)
 		i915_gem_clflush_object(obj, I915_CLFLUSH_FORCE);
 	obj->write_domain = 0;
@@ -80,7 +126,7 @@ i915_gem_object_set_to_wc_domain(struct drm_i915_gem_object *obj, bool write)
 	if (ret)
 		return ret;
 
-	i915_gem_object_flush_write_domain(obj, ~I915_GEM_DOMAIN_WC);
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_WC);
 
 	/* Serialise direct access to this object with the barriers for
 	 * coherent writes from the GPU, by effectively invalidating the
@@ -141,7 +187,7 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 	if (ret)
 		return ret;
 
-	i915_gem_object_flush_write_domain(obj, ~I915_GEM_DOMAIN_GTT);
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_GTT);
 
 	/* Serialise direct access to this object with the barriers for
 	 * coherent writes from the GPU, by effectively invalidating the
@@ -370,6 +416,7 @@ retry:
 	}
 
 	vma->display_alignment = max_t(u64, vma->display_alignment, alignment);
+	i915_vma_mark_scanout(vma);
 
 	i915_gem_object_flush_if_display_locked(obj);
 
@@ -387,48 +434,6 @@ err:
 	return vma;
 }
 
-static void i915_gem_object_bump_inactive_ggtt(struct drm_i915_gem_object *obj)
-{
-	struct drm_i915_private *i915 = to_i915(obj->base.dev);
-	struct i915_vma *vma;
-
-	if (list_empty(&obj->vma.list))
-		return;
-
-	mutex_lock(&i915->ggtt.vm.mutex);
-	spin_lock(&obj->vma.lock);
-	for_each_ggtt_vma(vma, obj) {
-		if (!drm_mm_node_allocated(&vma->node))
-			continue;
-
-		GEM_BUG_ON(vma->vm != &i915->ggtt.vm);
-		list_move_tail(&vma->vm_link, &vma->vm->bound_list);
-	}
-	spin_unlock(&obj->vma.lock);
-	mutex_unlock(&i915->ggtt.vm.mutex);
-
-	if (i915_gem_object_is_shrinkable(obj)) {
-		unsigned long flags;
-
-		spin_lock_irqsave(&i915->mm.obj_lock, flags);
-
-		if (obj->mm.madv == I915_MADV_WILLNEED &&
-		    !atomic_read(&obj->mm.shrink_pin))
-			list_move_tail(&obj->mm.link, &i915->mm.shrink_list);
-
-		spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
-	}
-}
-
-void
-i915_gem_object_unpin_from_display_plane(struct i915_vma *vma)
-{
-	/* Bump the LRU to try and avoid premature eviction whilst flipping  */
-	i915_gem_object_bump_inactive_ggtt(vma->obj);
-
-	i915_vma_unpin(vma);
-}
-
 /**
  * Moves a single object to the CPU read, and possibly write domain.
  * @obj: object to act on
@@ -451,7 +456,7 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
 	if (ret)
 		return ret;
 
-	i915_gem_object_flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
 
 	/* Flush the CPU cache if it's still invalid. */
 	if ((obj->read_domains & I915_GEM_DOMAIN_CPU) == 0) {
@@ -569,9 +574,6 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 	else
 		err = i915_gem_object_set_to_cpu_domain(obj, write_domain);
 
-	/* And bump the LRU for this access */
-	i915_gem_object_bump_inactive_ggtt(obj);
-
 	i915_gem_object_unlock(obj);
 
 	if (write_domain)
@@ -619,7 +621,7 @@ int i915_gem_object_prepare_read(struct drm_i915_gem_object *obj,
 			goto out;
 	}
 
-	i915_gem_object_flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
 
 	/* If we're not in the cpu read domain, set ourself into the gtt
 	 * read domain and manually flush cachelines (if required). This
@@ -670,7 +672,7 @@ int i915_gem_object_prepare_write(struct drm_i915_gem_object *obj,
 			goto out;
 	}
 
-	i915_gem_object_flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
 
 	/* If we're not in the cpu write domain, set ourself into the
 	 * gtt write domain and manually flush cachelines (as required).
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index bcc80f428172..d70ca36f74f6 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -15,6 +15,7 @@
 
 #include "gem/i915_gem_ioctls.h"
 #include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_buffer_pool.h"
 #include "gt/intel_gt_pm.h"
@@ -534,8 +535,6 @@ eb_add_vma(struct i915_execbuffer *eb,
 	struct drm_i915_gem_exec_object2 *entry = &eb->exec[i];
 	struct eb_vma *ev = &eb->vma[i];
 
-	GEM_BUG_ON(i915_vma_is_closed(vma));
-
 	ev->vma = vma;
 	ev->exec = entry;
 	ev->flags = entry->flags;
@@ -1046,7 +1045,7 @@ static void reloc_gpu_flush(struct i915_execbuffer *eb, struct reloc_cache *cach
 	GEM_BUG_ON(cache->rq_size >= obj->base.size / sizeof(u32));
 	cache->rq_cmd[cache->rq_size] = MI_BATCH_BUFFER_END;
 
-	__i915_gem_object_flush_map(obj, 0, sizeof(u32) * (cache->rq_size + 1));
+	i915_gem_object_flush_map(obj);
 	i915_gem_object_unpin_map(obj);
 
 	intel_gt_chipset_flush(cache->rq->engine->gt);
@@ -1277,7 +1276,10 @@ static int __reloc_gpu_alloc(struct i915_execbuffer *eb,
 	int err;
 
 	if (!pool) {
-		pool = intel_gt_get_buffer_pool(engine->gt, PAGE_SIZE);
+		pool = intel_gt_get_buffer_pool(engine->gt, PAGE_SIZE,
+						cache->has_llc ?
+						I915_MAP_WB :
+						I915_MAP_WC);
 		if (IS_ERR(pool))
 			return PTR_ERR(pool);
 	}
@@ -1287,15 +1289,14 @@ static int __reloc_gpu_alloc(struct i915_execbuffer *eb,
 	if (err)
 		goto err_pool;
 
-	cmd = i915_gem_object_pin_map(pool->obj,
-				      cache->has_llc ?
-				      I915_MAP_FORCE_WB :
-				      I915_MAP_FORCE_WC);
+	cmd = i915_gem_object_pin_map(pool->obj, pool->type);
 	if (IS_ERR(cmd)) {
 		err = PTR_ERR(cmd);
 		goto err_pool;
 	}
 
+	memset32(cmd, 0, pool->obj->base.size / sizeof(u32));
+
 	batch = i915_vma_instance(pool->obj, vma->vm, NULL);
 	if (IS_ERR(batch)) {
 		err = PTR_ERR(batch);
@@ -2457,7 +2458,8 @@ static int eb_parse(struct i915_execbuffer *eb)
 		return -EINVAL;
 
 	if (!pool) {
-		pool = intel_gt_get_buffer_pool(eb->engine->gt, len);
+		pool = intel_gt_get_buffer_pool(eb->engine->gt, len,
+						I915_MAP_WB);
 		if (IS_ERR(pool))
 			return PTR_ERR(pool);
 		eb->batch_pool = pool;
@@ -2533,6 +2535,9 @@ static int eb_submit(struct i915_execbuffer *eb, struct i915_vma *batch)
 {
 	int err;
 
+	if (intel_context_nopreempt(eb->context))
+		__set_bit(I915_FENCE_FLAG_NOPREEMPT, &eb->request->fence.flags);
+
 	err = eb_move_to_gpu(eb);
 	if (err)
 		return err;
@@ -2573,15 +2578,12 @@ static int eb_submit(struct i915_execbuffer *eb, struct i915_vma *batch)
 			return err;
 	}
 
-	if (intel_context_nopreempt(eb->context))
-		__set_bit(I915_FENCE_FLAG_NOPREEMPT, &eb->request->fence.flags);
-
 	return 0;
 }
 
 static int num_vcs_engines(const struct drm_i915_private *i915)
 {
-	return hweight64(VDBOX_MASK(&i915->gt));
+	return hweight_long(VDBOX_MASK(&i915->gt));
 }
 
 /*
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_lmem.c b/drivers/gpu/drm/i915/gem/i915_gem_lmem.c
index 932ee21e6609..194f35342710 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_lmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_lmem.c
@@ -31,18 +31,13 @@ i915_gem_object_create_lmem(struct drm_i915_private *i915,
 					     size, flags);
 }
 
-struct drm_i915_gem_object *
-__i915_gem_lmem_object_create(struct intel_memory_region *mem,
-			      resource_size_t size,
-			      unsigned int flags)
+int __i915_gem_lmem_object_init(struct intel_memory_region *mem,
+				struct drm_i915_gem_object *obj,
+				resource_size_t size,
+				unsigned int flags)
 {
 	static struct lock_class_key lock_class;
 	struct drm_i915_private *i915 = mem->i915;
-	struct drm_i915_gem_object *obj;
-
-	obj = i915_gem_object_alloc();
-	if (!obj)
-		return ERR_PTR(-ENOMEM);
 
 	drm_gem_private_object_init(&i915->drm, &obj->base, size);
 	i915_gem_object_init(obj, &i915_gem_lmem_obj_ops, &lock_class);
@@ -53,5 +48,5 @@ __i915_gem_lmem_object_create(struct intel_memory_region *mem,
 
 	i915_gem_object_init_memory_region(obj, mem, flags);
 
-	return obj;
+	return 0;
 }
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_lmem.h b/drivers/gpu/drm/i915/gem/i915_gem_lmem.h
index fc3f15580fe3..036d53c01de9 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_lmem.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_lmem.h
@@ -21,9 +21,9 @@ i915_gem_object_create_lmem(struct drm_i915_private *i915,
 			    resource_size_t size,
 			    unsigned int flags);
 
-struct drm_i915_gem_object *
-__i915_gem_lmem_object_create(struct intel_memory_region *mem,
-			      resource_size_t size,
-			      unsigned int flags);
+int __i915_gem_lmem_object_init(struct intel_memory_region *mem,
+				struct drm_i915_gem_object *obj,
+				resource_size_t size,
+				unsigned int flags);
 
 #endif /* !__I915_GEM_LMEM_H */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c
index 00d24000b5e8..70f798405f7f 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
@@ -25,13 +25,13 @@
 #include <linux/sched/mm.h>
 
 #include "display/intel_frontbuffer.h"
-#include "gt/intel_gt.h"
 #include "i915_drv.h"
 #include "i915_gem_clflush.h"
 #include "i915_gem_context.h"
 #include "i915_gem_mman.h"
 #include "i915_gem_object.h"
 #include "i915_globals.h"
+#include "i915_memcpy.h"
 #include "i915_trace.h"
 
 static struct i915_global_object {
@@ -313,52 +313,6 @@ static void i915_gem_free_object(struct drm_gem_object *gem_obj)
 		queue_work(i915->wq, &i915->mm.free_work);
 }
 
-static bool gpu_write_needs_clflush(struct drm_i915_gem_object *obj)
-{
-	return !(obj->cache_level == I915_CACHE_NONE ||
-		 obj->cache_level == I915_CACHE_WT);
-}
-
-void
-i915_gem_object_flush_write_domain(struct drm_i915_gem_object *obj,
-				   unsigned int flush_domains)
-{
-	struct i915_vma *vma;
-
-	assert_object_held(obj);
-
-	if (!(obj->write_domain & flush_domains))
-		return;
-
-	switch (obj->write_domain) {
-	case I915_GEM_DOMAIN_GTT:
-		spin_lock(&obj->vma.lock);
-		for_each_ggtt_vma(vma, obj) {
-			if (i915_vma_unset_ggtt_write(vma))
-				intel_gt_flush_ggtt_writes(vma->vm->gt);
-		}
-		spin_unlock(&obj->vma.lock);
-
-		i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
-		break;
-
-	case I915_GEM_DOMAIN_WC:
-		wmb();
-		break;
-
-	case I915_GEM_DOMAIN_CPU:
-		i915_gem_clflush_object(obj, I915_CLFLUSH_SYNC);
-		break;
-
-	case I915_GEM_DOMAIN_RENDER:
-		if (gpu_write_needs_clflush(obj))
-			obj->cache_dirty = true;
-		break;
-	}
-
-	obj->write_domain = 0;
-}
-
 void __i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
 					 enum fb_op_origin origin)
 {
@@ -383,6 +337,70 @@ void __i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
 	}
 }
 
+static void
+i915_gem_object_read_from_page_kmap(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
+{
+	void *src_map;
+	void *src_ptr;
+
+	src_map = kmap_atomic(i915_gem_object_get_page(obj, offset >> PAGE_SHIFT));
+
+	src_ptr = src_map + offset_in_page(offset);
+	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
+		drm_clflush_virt_range(src_ptr, size);
+	memcpy(dst, src_ptr, size);
+
+	kunmap_atomic(src_map);
+}
+
+static void
+i915_gem_object_read_from_page_iomap(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
+{
+	void __iomem *src_map;
+	void __iomem *src_ptr;
+	dma_addr_t dma = i915_gem_object_get_dma_address(obj, offset >> PAGE_SHIFT);
+
+	src_map = io_mapping_map_wc(&obj->mm.region->iomap,
+				    dma - obj->mm.region->region.start,
+				    PAGE_SIZE);
+
+	src_ptr = src_map + offset_in_page(offset);
+	if (!i915_memcpy_from_wc(dst, (void __force *)src_ptr, size))
+		memcpy_fromio(dst, src_ptr, size);
+
+	io_mapping_unmap(src_map);
+}
+
+/**
+ * i915_gem_object_read_from_page - read data from the page of a GEM object
+ * @obj: GEM object to read from
+ * @offset: offset within the object
+ * @dst: buffer to store the read data
+ * @size: size to read
+ *
+ * Reads data from @obj at the specified offset. The requested region to read
+ * from can't cross a page boundary. The caller must ensure that @obj pages
+ * are pinned and that @obj is synced wrt. any related writes.
+ *
+ * Returns 0 on success or -ENODEV if the type of @obj's backing store is
+ * unsupported.
+ */
+int i915_gem_object_read_from_page(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
+{
+	GEM_BUG_ON(offset >= obj->base.size);
+	GEM_BUG_ON(offset_in_page(offset) > PAGE_SIZE - size);
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+
+	if (i915_gem_object_has_struct_page(obj))
+		i915_gem_object_read_from_page_kmap(obj, offset, dst, size);
+	else if (i915_gem_object_has_iomem(obj))
+		i915_gem_object_read_from_page_iomap(obj, offset, dst, size);
+	else
+		return -ENODEV;
+
+	return 0;
+}
+
 void i915_gem_init__objects(struct drm_i915_private *i915)
 {
 	INIT_WORK(&i915->mm.free_work, __i915_gem_free_work);
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.h b/drivers/gpu/drm/i915/gem/i915_gem_object.h
index be14486f63a7..d0ae834d787a 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_object.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.h
@@ -188,6 +188,24 @@ i915_gem_object_set_volatile(struct drm_i915_gem_object *obj)
 }
 
 static inline bool
+i915_gem_object_has_tiling_quirk(struct drm_i915_gem_object *obj)
+{
+	return test_bit(I915_TILING_QUIRK_BIT, &obj->flags);
+}
+
+static inline void
+i915_gem_object_set_tiling_quirk(struct drm_i915_gem_object *obj)
+{
+	set_bit(I915_TILING_QUIRK_BIT, &obj->flags);
+}
+
+static inline void
+i915_gem_object_clear_tiling_quirk(struct drm_i915_gem_object *obj)
+{
+	clear_bit(I915_TILING_QUIRK_BIT, &obj->flags);
+}
+
+static inline bool
 i915_gem_object_type_has(const struct drm_i915_gem_object *obj,
 			 unsigned long flags)
 {
@@ -201,6 +219,12 @@ i915_gem_object_has_struct_page(const struct drm_i915_gem_object *obj)
 }
 
 static inline bool
+i915_gem_object_has_iomem(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_type_has(obj, I915_GEM_OBJECT_HAS_IOMEM);
+}
+
+static inline bool
 i915_gem_object_is_shrinkable(const struct drm_i915_gem_object *obj)
 {
 	return i915_gem_object_type_has(obj, I915_GEM_OBJECT_IS_SHRINKABLE);
@@ -384,14 +408,6 @@ int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj);
 void i915_gem_object_truncate(struct drm_i915_gem_object *obj);
 void i915_gem_object_writeback(struct drm_i915_gem_object *obj);
 
-enum i915_map_type {
-	I915_MAP_WB = 0,
-	I915_MAP_WC,
-#define I915_MAP_OVERRIDE BIT(31)
-	I915_MAP_FORCE_WB = I915_MAP_WB | I915_MAP_OVERRIDE,
-	I915_MAP_FORCE_WC = I915_MAP_WC | I915_MAP_OVERRIDE,
-};
-
 /**
  * i915_gem_object_pin_map - return a contiguous mapping of the entire object
  * @obj: the object to map into kernel address space
@@ -435,10 +451,6 @@ static inline void i915_gem_object_unpin_map(struct drm_i915_gem_object *obj)
 
 void __i915_gem_object_release_map(struct drm_i915_gem_object *obj);
 
-void
-i915_gem_object_flush_write_domain(struct drm_i915_gem_object *obj,
-				   unsigned int flush_domains);
-
 int i915_gem_object_prepare_read(struct drm_i915_gem_object *obj,
 				 unsigned int *needs_clflush);
 int i915_gem_object_prepare_write(struct drm_i915_gem_object *obj,
@@ -486,7 +498,6 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
 				     u32 alignment,
 				     const struct i915_ggtt_view *view,
 				     unsigned int flags);
-void i915_gem_object_unpin_from_display_plane(struct i915_vma *vma);
 
 void i915_gem_object_make_unshrinkable(struct drm_i915_gem_object *obj);
 void i915_gem_object_make_shrinkable(struct drm_i915_gem_object *obj);
@@ -512,6 +523,9 @@ static inline void __start_cpu_write(struct drm_i915_gem_object *obj)
 		obj->cache_dirty = true;
 }
 
+void i915_gem_fence_wait_priority(struct dma_fence *fence,
+				  const struct i915_sched_attr *attr);
+
 int i915_gem_object_wait(struct drm_i915_gem_object *obj,
 			 unsigned int flags,
 			 long timeout);
@@ -540,4 +554,8 @@ i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
 		__i915_gem_object_invalidate_frontbuffer(obj, origin);
 }
 
+int i915_gem_object_read_from_page(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size);
+
+bool i915_gem_object_is_shmem(const struct drm_i915_gem_object *obj);
+
 #endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c b/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c
index aee7ad3cc3c6..d6dac21fce0b 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c
@@ -6,6 +6,7 @@
 #include "i915_drv.h"
 #include "gt/intel_context.h"
 #include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_buffer_pool.h"
 #include "gt/intel_ring.h"
@@ -34,7 +35,7 @@ struct i915_vma *intel_emit_vma_fill_blt(struct intel_context *ce,
 	count = div_u64(round_up(vma->size, block_size), block_size);
 	size = (1 + 8 * count) * sizeof(u32);
 	size = round_up(size, PAGE_SIZE);
-	pool = intel_gt_get_buffer_pool(ce->engine->gt, size);
+	pool = intel_gt_get_buffer_pool(ce->engine->gt, size, I915_MAP_WC);
 	if (IS_ERR(pool)) {
 		err = PTR_ERR(pool);
 		goto out_pm;
@@ -54,7 +55,7 @@ struct i915_vma *intel_emit_vma_fill_blt(struct intel_context *ce,
 	if (unlikely(err))
 		goto out_put;
 
-	cmd = i915_gem_object_pin_map(pool->obj, I915_MAP_WC);
+	cmd = i915_gem_object_pin_map(pool->obj, pool->type);
 	if (IS_ERR(cmd)) {
 		err = PTR_ERR(cmd);
 		goto out_unpin;
@@ -256,7 +257,7 @@ struct i915_vma *intel_emit_vma_copy_blt(struct intel_context *ce,
 	count = div_u64(round_up(dst->size, block_size), block_size);
 	size = (1 + 11 * count) * sizeof(u32);
 	size = round_up(size, PAGE_SIZE);
-	pool = intel_gt_get_buffer_pool(ce->engine->gt, size);
+	pool = intel_gt_get_buffer_pool(ce->engine->gt, size, I915_MAP_WC);
 	if (IS_ERR(pool)) {
 		err = PTR_ERR(pool);
 		goto out_pm;
@@ -276,7 +277,7 @@ struct i915_vma *intel_emit_vma_copy_blt(struct intel_context *ce,
 	if (unlikely(err))
 		goto out_put;
 
-	cmd = i915_gem_object_pin_map(pool->obj, I915_MAP_WC);
+	cmd = i915_gem_object_pin_map(pool->obj, pool->type);
 	if (IS_ERR(cmd)) {
 		err = PTR_ERR(cmd);
 		goto out_unpin;
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
index e2d9b7e1e152..0438e00d4ca7 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
@@ -67,6 +67,14 @@ struct drm_i915_gem_object_ops {
 	const char *name; /* friendly name for debug, e.g. lockdep classes */
 };
 
+enum i915_map_type {
+	I915_MAP_WB = 0,
+	I915_MAP_WC,
+#define I915_MAP_OVERRIDE BIT(31)
+	I915_MAP_FORCE_WB = I915_MAP_WB | I915_MAP_OVERRIDE,
+	I915_MAP_FORCE_WC = I915_MAP_WC | I915_MAP_OVERRIDE,
+};
+
 enum i915_mmap_type {
 	I915_MMAP_TYPE_GTT = 0,
 	I915_MMAP_TYPE_WC,
@@ -142,8 +150,6 @@ struct drm_i915_gem_object {
 	 */
 	struct list_head obj_link;
 
-	/** Stolen memory for this object, instead of being backed by shmem. */
-	struct drm_mm_node *stolen;
 	union {
 		struct rcu_head rcu;
 		struct llist_node freed;
@@ -167,6 +173,7 @@ struct drm_i915_gem_object {
 #define I915_BO_ALLOC_VOLATILE   BIT(1)
 #define I915_BO_ALLOC_FLAGS (I915_BO_ALLOC_CONTIGUOUS | I915_BO_ALLOC_VOLATILE)
 #define I915_BO_READONLY         BIT(2)
+#define I915_TILING_QUIRK_BIT    3 /* unknown swizzling; do not release! */
 
 	/*
 	 * Is the object to be mapped as read-only to the GPU
@@ -275,12 +282,6 @@ struct drm_i915_gem_object {
 		 * pages were last acquired.
 		 */
 		bool dirty:1;
-
-		/**
-		 * This is set if the object has been pinned due to unknown
-		 * swizzling.
-		 */
-		bool quirked:1;
 	} mm;
 
 	/** Record of address bit 17 of each page at last unbind. */
@@ -295,6 +296,8 @@ struct drm_i915_gem_object {
 			struct work_struct *work;
 		} userptr;
 
+		struct drm_mm_node *stolen;
+
 		unsigned long scratch;
 		u64 encode;
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
index e2c7b2a7895f..43028f3539a6 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_pages.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
@@ -16,6 +16,7 @@ void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj,
 {
 	struct drm_i915_private *i915 = to_i915(obj->base.dev);
 	unsigned long supported = INTEL_INFO(i915)->page_sizes;
+	bool shrinkable;
 	int i;
 
 	lockdep_assert_held(&obj->mm.lock);
@@ -38,13 +39,6 @@ void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj,
 
 	obj->mm.pages = pages;
 
-	if (i915_gem_object_is_tiled(obj) &&
-	    i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
-		GEM_BUG_ON(obj->mm.quirked);
-		__i915_gem_object_pin_pages(obj);
-		obj->mm.quirked = true;
-	}
-
 	GEM_BUG_ON(!sg_page_sizes);
 	obj->mm.page_sizes.phys = sg_page_sizes;
 
@@ -63,7 +57,16 @@ void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj,
 	}
 	GEM_BUG_ON(!HAS_PAGE_SIZES(i915, obj->mm.page_sizes.sg));
 
-	if (i915_gem_object_is_shrinkable(obj)) {
+	shrinkable = i915_gem_object_is_shrinkable(obj);
+
+	if (i915_gem_object_is_tiled(obj) &&
+	    i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
+		GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+		i915_gem_object_set_tiling_quirk(obj);
+		shrinkable = false;
+	}
+
+	if (shrinkable) {
 		struct list_head *list;
 		unsigned long flags;
 
@@ -238,7 +241,7 @@ unlock:
 
 /* The 'mapping' part of i915_gem_object_pin_map() below */
 static void *i915_gem_object_map_page(struct drm_i915_gem_object *obj,
-		enum i915_map_type type)
+				      enum i915_map_type type)
 {
 	unsigned long n_pages = obj->base.size >> PAGE_SHIFT, i;
 	struct page *stack[32], **pages = stack, *page;
@@ -281,7 +284,7 @@ static void *i915_gem_object_map_page(struct drm_i915_gem_object *obj,
 		/* Too big for stack -- allocate temporary array instead */
 		pages = kvmalloc_array(n_pages, sizeof(*pages), GFP_KERNEL);
 		if (!pages)
-			return NULL;
+			return ERR_PTR(-ENOMEM);
 	}
 
 	i = 0;
@@ -290,11 +293,12 @@ static void *i915_gem_object_map_page(struct drm_i915_gem_object *obj,
 	vaddr = vmap(pages, n_pages, 0, pgprot);
 	if (pages != stack)
 		kvfree(pages);
-	return vaddr;
+
+	return vaddr ?: ERR_PTR(-ENOMEM);
 }
 
 static void *i915_gem_object_map_pfn(struct drm_i915_gem_object *obj,
-		enum i915_map_type type)
+				     enum i915_map_type type)
 {
 	resource_size_t iomap = obj->mm.region->iomap.base -
 		obj->mm.region->region.start;
@@ -305,13 +309,13 @@ static void *i915_gem_object_map_pfn(struct drm_i915_gem_object *obj,
 	void *vaddr;
 
 	if (type != I915_MAP_WC)
-		return NULL;
+		return ERR_PTR(-ENODEV);
 
 	if (n_pfn > ARRAY_SIZE(stack)) {
 		/* Too big for stack -- allocate temporary array instead */
 		pfns = kvmalloc_array(n_pfn, sizeof(*pfns), GFP_KERNEL);
 		if (!pfns)
-			return NULL;
+			return ERR_PTR(-ENOMEM);
 	}
 
 	i = 0;
@@ -320,7 +324,8 @@ static void *i915_gem_object_map_pfn(struct drm_i915_gem_object *obj,
 	vaddr = vmap_pfn(pfns, n_pfn, pgprot_writecombine(PAGE_KERNEL_IO));
 	if (pfns != stack)
 		kvfree(pfns);
-	return vaddr;
+
+	return vaddr ?: ERR_PTR(-ENOMEM);
 }
 
 /* get, pin, and map the pages of the object into kernel space */
@@ -349,8 +354,10 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
 			GEM_BUG_ON(i915_gem_object_has_pinned_pages(obj));
 
 			err = ____i915_gem_object_get_pages(obj);
-			if (err)
-				goto err_unlock;
+			if (err) {
+				ptr = ERR_PTR(err);
+				goto out_unlock;
+			}
 
 			smp_mb__before_atomic();
 		}
@@ -362,7 +369,7 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
 	ptr = page_unpack_bits(obj->mm.mapping, &has_type);
 	if (ptr && has_type != type) {
 		if (pinned) {
-			err = -EBUSY;
+			ptr = ERR_PTR(-EBUSY);
 			goto err_unpin;
 		}
 
@@ -374,15 +381,13 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
 	if (!ptr) {
 		if (GEM_WARN_ON(type == I915_MAP_WC &&
 				!static_cpu_has(X86_FEATURE_PAT)))
-			ptr = NULL;
+			ptr = ERR_PTR(-ENODEV);
 		else if (i915_gem_object_has_struct_page(obj))
 			ptr = i915_gem_object_map_page(obj, type);
 		else
 			ptr = i915_gem_object_map_pfn(obj, type);
-		if (!ptr) {
-			err = -ENOMEM;
+		if (IS_ERR(ptr))
 			goto err_unpin;
-		}
 
 		obj->mm.mapping = page_pack_bits(ptr, type);
 	}
@@ -393,8 +398,6 @@ out_unlock:
 
 err_unpin:
 	atomic_dec(&obj->mm.pages_pin_count);
-err_unlock:
-	ptr = ERR_PTR(err);
 	goto out_unlock;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
index 3a4dfe2ef1da..3c0b157e2a35 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
@@ -213,7 +213,7 @@ int i915_gem_object_attach_phys(struct drm_i915_gem_object *obj, int align)
 	if (obj->ops == &i915_gem_phys_ops)
 		return 0;
 
-	if (obj->ops != &i915_gem_shmem_ops)
+	if (!i915_gem_object_is_shmem(obj))
 		return -EINVAL;
 
 	err = i915_gem_object_unbind(obj, I915_GEM_OBJECT_UNBIND_ACTIVE);
@@ -227,7 +227,7 @@ int i915_gem_object_attach_phys(struct drm_i915_gem_object *obj, int align)
 		goto err_unlock;
 	}
 
-	if (obj->mm.quirked) {
+	if (i915_gem_object_has_tiling_quirk(obj)) {
 		err = -EFAULT;
 		goto err_unlock;
 	}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pm.c b/drivers/gpu/drm/i915/gem/i915_gem_pm.c
index 40d3e40500fa..000e1cd8e920 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_pm.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pm.c
@@ -11,6 +11,13 @@
 
 #include "i915_drv.h"
 
+#if defined(CONFIG_X86)
+#include <asm/smp.h>
+#else
+#define wbinvd_on_all_cpus() \
+	pr_warn(DRIVER_NAME ": Missing cache flush in %s\n", __func__)
+#endif
+
 void i915_gem_suspend(struct drm_i915_private *i915)
 {
 	GEM_TRACE("%s\n", dev_name(i915->drm.dev));
@@ -32,13 +39,6 @@ void i915_gem_suspend(struct drm_i915_private *i915)
 	i915_gem_drain_freed_objects(i915);
 }
 
-static struct drm_i915_gem_object *first_mm_object(struct list_head *list)
-{
-	return list_first_entry_or_null(list,
-					struct drm_i915_gem_object,
-					mm.link);
-}
-
 void i915_gem_suspend_late(struct drm_i915_private *i915)
 {
 	struct drm_i915_gem_object *obj;
@@ -48,6 +48,7 @@ void i915_gem_suspend_late(struct drm_i915_private *i915)
 		NULL
 	}, **phase;
 	unsigned long flags;
+	bool flush = false;
 
 	/*
 	 * Neither the BIOS, ourselves or any other kernel
@@ -73,29 +74,56 @@ void i915_gem_suspend_late(struct drm_i915_private *i915)
 
 	spin_lock_irqsave(&i915->mm.obj_lock, flags);
 	for (phase = phases; *phase; phase++) {
-		LIST_HEAD(keep);
+		list_for_each_entry(obj, *phase, mm.link) {
+			if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
+				flush |= (obj->read_domains & I915_GEM_DOMAIN_CPU) == 0;
+			__start_cpu_write(obj); /* presume auto-hibernate */
+		}
+	}
+	spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+	if (flush)
+		wbinvd_on_all_cpus();
+}
 
-		while ((obj = first_mm_object(*phase))) {
-			list_move_tail(&obj->mm.link, &keep);
+int i915_gem_freeze(struct drm_i915_private *i915)
+{
+	/* Discard all purgeable objects, let userspace recover those as
+	 * required after resuming.
+	 */
+	i915_gem_shrink_all(i915);
 
-			/* Beware the background _i915_gem_free_objects */
-			if (!kref_get_unless_zero(&obj->base.refcount))
-				continue;
+	return 0;
+}
 
-			spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+int i915_gem_freeze_late(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	intel_wakeref_t wakeref;
 
-			i915_gem_object_lock(obj, NULL);
-			drm_WARN_ON(&i915->drm,
-			    i915_gem_object_set_to_gtt_domain(obj, false));
-			i915_gem_object_unlock(obj);
-			i915_gem_object_put(obj);
+	/*
+	 * Called just before we write the hibernation image.
+	 *
+	 * We need to update the domain tracking to reflect that the CPU
+	 * will be accessing all the pages to create and restore from the
+	 * hibernation, and so upon restoration those pages will be in the
+	 * CPU domain.
+	 *
+	 * To make sure the hibernation image contains the latest state,
+	 * we update that state just before writing out the image.
+	 *
+	 * To try and reduce the hibernation image, we manually shrink
+	 * the objects as well, see i915_gem_freeze()
+	 */
 
-			spin_lock_irqsave(&i915->mm.obj_lock, flags);
-		}
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		i915_gem_shrink(i915, -1UL, NULL, ~0);
+	i915_gem_drain_freed_objects(i915);
 
-		list_splice_tail(&keep, *phase);
-	}
-	spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+	wbinvd_on_all_cpus();
+	list_for_each_entry(obj, &i915->mm.shrink_list, mm.link)
+		__start_cpu_write(obj);
+
+	return 0;
 }
 
 void i915_gem_resume(struct drm_i915_private *i915)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pm.h b/drivers/gpu/drm/i915/gem/i915_gem_pm.h
index 26b78dbdc225..c9a66630e92e 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_pm.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pm.h
@@ -19,4 +19,7 @@ void i915_gem_idle_work_handler(struct work_struct *work);
 void i915_gem_suspend(struct drm_i915_private *i915);
 void i915_gem_suspend_late(struct drm_i915_private *i915);
 
+int i915_gem_freeze(struct drm_i915_private *i915);
+int i915_gem_freeze_late(struct drm_i915_private *i915);
+
 #endif /* __I915_GEM_PM_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_region.c b/drivers/gpu/drm/i915/gem/i915_gem_region.c
index 1515384d7e0e..3e3dad22a683 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_region.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_region.c
@@ -22,6 +22,7 @@ i915_gem_object_put_pages_buddy(struct drm_i915_gem_object *obj,
 int
 i915_gem_object_get_pages_buddy(struct drm_i915_gem_object *obj)
 {
+	const u64 max_segment = i915_sg_segment_size();
 	struct intel_memory_region *mem = obj->mm.region;
 	struct list_head *blocks = &obj->mm.blocks;
 	resource_size_t size = obj->base.size;
@@ -37,7 +38,7 @@ i915_gem_object_get_pages_buddy(struct drm_i915_gem_object *obj)
 	if (!st)
 		return -ENOMEM;
 
-	if (sg_alloc_table(st, size >> ilog2(mem->mm.chunk_size), GFP_KERNEL)) {
+	if (sg_alloc_table(st, size >> PAGE_SHIFT, GFP_KERNEL)) {
 		kfree(st);
 		return -ENOMEM;
 	}
@@ -64,27 +65,30 @@ i915_gem_object_get_pages_buddy(struct drm_i915_gem_object *obj)
 				   i915_buddy_block_size(&mem->mm, block));
 		offset = i915_buddy_block_offset(block);
 
-		GEM_BUG_ON(overflows_type(block_size, sg->length));
+		while (block_size) {
+			u64 len;
 
-		if (offset != prev_end ||
-		    add_overflows_t(typeof(sg->length), sg->length, block_size)) {
-			if (st->nents) {
-				sg_page_sizes |= sg->length;
-				sg = __sg_next(sg);
+			if (offset != prev_end || sg->length >= max_segment) {
+				if (st->nents) {
+					sg_page_sizes |= sg->length;
+					sg = __sg_next(sg);
+				}
+
+				sg_dma_address(sg) = mem->region.start + offset;
+				sg_dma_len(sg) = 0;
+				sg->length = 0;
+				st->nents++;
 			}
 
-			sg_dma_address(sg) = mem->region.start + offset;
-			sg_dma_len(sg) = block_size;
+			len = min(block_size, max_segment - sg->length);
+			sg->length += len;
+			sg_dma_len(sg) += len;
 
-			sg->length = block_size;
+			offset += len;
+			block_size -= len;
 
-			st->nents++;
-		} else {
-			sg->length += block_size;
-			sg_dma_len(sg) += block_size;
+			prev_end = offset;
 		}
-
-		prev_end = offset + block_size;
 	}
 
 	sg_page_sizes |= sg->length;
@@ -139,6 +143,7 @@ i915_gem_object_create_region(struct intel_memory_region *mem,
 			      unsigned int flags)
 {
 	struct drm_i915_gem_object *obj;
+	int err;
 
 	/*
 	 * NB: Our use of resource_size_t for the size stems from using struct
@@ -169,9 +174,18 @@ i915_gem_object_create_region(struct intel_memory_region *mem,
 	if (overflows_type(size, obj->base.size))
 		return ERR_PTR(-E2BIG);
 
-	obj = mem->ops->create_object(mem, size, flags);
-	if (!IS_ERR(obj))
-		trace_i915_gem_object_create(obj);
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
 
+	err = mem->ops->init_object(mem, obj, size, flags);
+	if (err)
+		goto err_object_free;
+
+	trace_i915_gem_object_create(obj);
 	return obj;
+
+err_object_free:
+	i915_gem_object_free(obj);
+	return ERR_PTR(err);
 }
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
index 75e8b71c18b9..cf83c208688c 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -464,26 +464,21 @@ static int __create_shmem(struct drm_i915_private *i915,
 	return 0;
 }
 
-static struct drm_i915_gem_object *
-create_shmem(struct intel_memory_region *mem,
-	     resource_size_t size,
-	     unsigned int flags)
+static int shmem_object_init(struct intel_memory_region *mem,
+			     struct drm_i915_gem_object *obj,
+			     resource_size_t size,
+			     unsigned int flags)
 {
 	static struct lock_class_key lock_class;
 	struct drm_i915_private *i915 = mem->i915;
-	struct drm_i915_gem_object *obj;
 	struct address_space *mapping;
 	unsigned int cache_level;
 	gfp_t mask;
 	int ret;
 
-	obj = i915_gem_object_alloc();
-	if (!obj)
-		return ERR_PTR(-ENOMEM);
-
 	ret = __create_shmem(i915, &obj->base, size);
 	if (ret)
-		goto fail;
+		return ret;
 
 	mask = GFP_HIGHUSER | __GFP_RECLAIMABLE;
 	if (IS_I965GM(i915) || IS_I965G(i915)) {
@@ -522,11 +517,7 @@ create_shmem(struct intel_memory_region *mem,
 
 	i915_gem_object_init_memory_region(obj, mem, 0);
 
-	return obj;
-
-fail:
-	i915_gem_object_free(obj);
-	return ERR_PTR(ret);
+	return 0;
 }
 
 struct drm_i915_gem_object *
@@ -611,7 +602,7 @@ static void release_shmem(struct intel_memory_region *mem)
 static const struct intel_memory_region_ops shmem_region_ops = {
 	.init = init_shmem,
 	.release = release_shmem,
-	.create_object = create_shmem,
+	.init_object = shmem_object_init,
 };
 
 struct intel_memory_region *i915_gem_shmem_setup(struct drm_i915_private *i915)
@@ -621,3 +612,8 @@ struct intel_memory_region *i915_gem_shmem_setup(struct drm_i915_private *i915)
 					  PAGE_SIZE, 0,
 					  &shmem_region_ops);
 }
+
+bool i915_gem_object_is_shmem(const struct drm_i915_gem_object *obj)
+{
+	return obj->ops == &i915_gem_shmem_ops;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c
index dc8f052a0ffe..c2dba1cd9532 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c
@@ -15,6 +15,7 @@
 
 #include "gt/intel_gt_requests.h"
 
+#include "dma_resv_utils.h"
 #include "i915_trace.h"
 
 static bool swap_available(void)
@@ -209,6 +210,8 @@ i915_gem_shrink(struct drm_i915_private *i915,
 				mutex_unlock(&obj->mm.lock);
 			}
 
+			dma_resv_prune(obj->base.resv);
+
 			scanned += obj->base.size >> PAGE_SHIFT;
 			i915_gem_object_put(obj);
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c
index 29bffc6afcc1..a1e197a6e999 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c
@@ -608,11 +608,10 @@ i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
 	struct drm_mm_node *stolen = fetch_and_zero(&obj->stolen);
 
 	GEM_BUG_ON(!stolen);
-
-	i915_gem_object_release_memory_region(obj);
-
 	i915_gem_stolen_remove_node(i915, stolen);
 	kfree(stolen);
+
+	i915_gem_object_release_memory_region(obj);
 }
 
 static const struct drm_i915_gem_object_ops i915_gem_object_stolen_ops = {
@@ -622,18 +621,13 @@ static const struct drm_i915_gem_object_ops i915_gem_object_stolen_ops = {
 	.release = i915_gem_object_release_stolen,
 };
 
-static struct drm_i915_gem_object *
-__i915_gem_object_create_stolen(struct intel_memory_region *mem,
-				struct drm_mm_node *stolen)
+static int __i915_gem_object_create_stolen(struct intel_memory_region *mem,
+					   struct drm_i915_gem_object *obj,
+					   struct drm_mm_node *stolen)
 {
 	static struct lock_class_key lock_class;
-	struct drm_i915_gem_object *obj;
 	unsigned int cache_level;
-	int err = -ENOMEM;
-
-	obj = i915_gem_object_alloc();
-	if (!obj)
-		goto err;
+	int err;
 
 	drm_gem_private_object_init(&mem->i915->drm, &obj->base, stolen->size);
 	i915_gem_object_init(obj, &i915_gem_object_stolen_ops, &lock_class);
@@ -645,55 +639,47 @@ __i915_gem_object_create_stolen(struct intel_memory_region *mem,
 
 	err = i915_gem_object_pin_pages(obj);
 	if (err)
-		goto cleanup;
+		return err;
 
 	i915_gem_object_init_memory_region(obj, mem, 0);
 
-	return obj;
-
-cleanup:
-	i915_gem_object_free(obj);
-err:
-	return ERR_PTR(err);
+	return 0;
 }
 
-static struct drm_i915_gem_object *
-_i915_gem_object_create_stolen(struct intel_memory_region *mem,
-			       resource_size_t size,
-			       unsigned int flags)
+static int _i915_gem_object_stolen_init(struct intel_memory_region *mem,
+					struct drm_i915_gem_object *obj,
+					resource_size_t size,
+					unsigned int flags)
 {
 	struct drm_i915_private *i915 = mem->i915;
-	struct drm_i915_gem_object *obj;
 	struct drm_mm_node *stolen;
 	int ret;
 
 	if (!drm_mm_initialized(&i915->mm.stolen))
-		return ERR_PTR(-ENODEV);
+		return -ENODEV;
 
 	if (size == 0)
-		return ERR_PTR(-EINVAL);
+		return -EINVAL;
 
 	stolen = kzalloc(sizeof(*stolen), GFP_KERNEL);
 	if (!stolen)
-		return ERR_PTR(-ENOMEM);
+		return -ENOMEM;
 
 	ret = i915_gem_stolen_insert_node(i915, stolen, size, 4096);
-	if (ret) {
-		obj = ERR_PTR(ret);
+	if (ret)
 		goto err_free;
-	}
 
-	obj = __i915_gem_object_create_stolen(mem, stolen);
-	if (IS_ERR(obj))
+	ret = __i915_gem_object_create_stolen(mem, obj, stolen);
+	if (ret)
 		goto err_remove;
 
-	return obj;
+	return 0;
 
 err_remove:
 	i915_gem_stolen_remove_node(i915, stolen);
 err_free:
 	kfree(stolen);
-	return obj;
+	return ret;
 }
 
 struct drm_i915_gem_object *
@@ -723,7 +709,7 @@ static void release_stolen(struct intel_memory_region *mem)
 static const struct intel_memory_region_ops i915_region_stolen_ops = {
 	.init = init_stolen,
 	.release = release_stolen,
-	.create_object = _i915_gem_object_create_stolen,
+	.init_object = _i915_gem_object_stolen_init,
 };
 
 struct intel_memory_region *i915_gem_stolen_setup(struct drm_i915_private *i915)
@@ -767,21 +753,32 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *i915,
 	mutex_lock(&i915->mm.stolen_lock);
 	ret = drm_mm_reserve_node(&i915->mm.stolen, stolen);
 	mutex_unlock(&i915->mm.stolen_lock);
-	if (ret) {
-		obj = ERR_PTR(ret);
+	if (ret)
 		goto err_free;
-	}
 
-	obj = __i915_gem_object_create_stolen(mem, stolen);
-	if (IS_ERR(obj))
+	obj = i915_gem_object_alloc();
+	if (!obj) {
+		ret = -ENOMEM;
 		goto err_stolen;
+	}
+
+	ret = __i915_gem_object_create_stolen(mem, obj, stolen);
+	if (ret)
+		goto err_object_free;
 
 	i915_gem_object_set_cache_coherency(obj, I915_CACHE_NONE);
 	return obj;
 
+err_object_free:
+	i915_gem_object_free(obj);
 err_stolen:
 	i915_gem_stolen_remove_node(i915, stolen);
 err_free:
 	kfree(stolen);
-	return obj;
+	return ERR_PTR(ret);
+}
+
+bool i915_gem_object_is_stolen(const struct drm_i915_gem_object *obj)
+{
+	return obj->ops == &i915_gem_object_stolen_ops;
 }
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_stolen.h b/drivers/gpu/drm/i915/gem/i915_gem_stolen.h
index 61e028063f9f..b03489706796 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_stolen.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_stolen.h
@@ -30,6 +30,8 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv
 					       resource_size_t stolen_offset,
 					       resource_size_t size);
 
+bool i915_gem_object_is_stolen(const struct drm_i915_gem_object *obj);
+
 #define I915_GEM_STOLEN_BIAS SZ_128K
 
 #endif /* __I915_GEM_STOLEN_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_tiling.c b/drivers/gpu/drm/i915/gem/i915_gem_tiling.c
index ffcaee74a249..d589d3d81085 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_tiling.c
@@ -270,14 +270,14 @@ i915_gem_object_set_tiling(struct drm_i915_gem_object *obj,
 	    obj->mm.madv == I915_MADV_WILLNEED &&
 	    i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
 		if (tiling == I915_TILING_NONE) {
-			GEM_BUG_ON(!obj->mm.quirked);
-			__i915_gem_object_unpin_pages(obj);
-			obj->mm.quirked = false;
+			GEM_BUG_ON(!i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_clear_tiling_quirk(obj);
+			i915_gem_object_make_shrinkable(obj);
 		}
 		if (!i915_gem_object_is_tiled(obj)) {
-			GEM_BUG_ON(obj->mm.quirked);
-			__i915_gem_object_pin_pages(obj);
-			obj->mm.quirked = true;
+			GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_make_unshrinkable(obj);
+			i915_gem_object_set_tiling_quirk(obj);
 		}
 	}
 	mutex_unlock(&obj->mm.lock);
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_wait.c b/drivers/gpu/drm/i915/gem/i915_gem_wait.c
index 8af55cd3e690..4b9856d5ba14 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_wait.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_wait.c
@@ -5,10 +5,12 @@
  */
 
 #include <linux/dma-fence-array.h>
+#include <linux/dma-fence-chain.h>
 #include <linux/jiffies.h>
 
 #include "gt/intel_engine.h"
 
+#include "dma_resv_utils.h"
 #include "i915_gem_ioctls.h"
 #include "i915_gem_object.h"
 
@@ -43,8 +45,7 @@ i915_gem_object_wait_reservation(struct dma_resv *resv,
 		unsigned int count, i;
 		int ret;
 
-		ret = dma_resv_get_fences_rcu(resv,
-							&excl, &count, &shared);
+		ret = dma_resv_get_fences_rcu(resv, &excl, &count, &shared);
 		if (ret)
 			return ret;
 
@@ -84,17 +85,14 @@ i915_gem_object_wait_reservation(struct dma_resv *resv,
 	 * Opportunistically prune the fences iff we know they have *all* been
 	 * signaled.
 	 */
-	if (prune_fences && dma_resv_trylock(resv)) {
-		if (dma_resv_test_signaled_rcu(resv, true))
-			dma_resv_add_excl_fence(resv, NULL);
-		dma_resv_unlock(resv);
-	}
+	if (prune_fences)
+		dma_resv_prune(resv);
 
 	return timeout;
 }
 
-static void __fence_set_priority(struct dma_fence *fence,
-				 const struct i915_sched_attr *attr)
+static void fence_set_priority(struct dma_fence *fence,
+			       const struct i915_sched_attr *attr)
 {
 	struct i915_request *rq;
 	struct intel_engine_cs *engine;
@@ -105,27 +103,47 @@ static void __fence_set_priority(struct dma_fence *fence,
 	rq = to_request(fence);
 	engine = rq->engine;
 
-	local_bh_disable();
 	rcu_read_lock(); /* RCU serialisation for set-wedged protection */
 	if (engine->schedule)
 		engine->schedule(rq, attr);
 	rcu_read_unlock();
-	local_bh_enable(); /* kick the tasklets if queues were reprioritised */
 }
 
-static void fence_set_priority(struct dma_fence *fence,
-			       const struct i915_sched_attr *attr)
+static inline bool __dma_fence_is_chain(const struct dma_fence *fence)
 {
+	return fence->ops == &dma_fence_chain_ops;
+}
+
+void i915_gem_fence_wait_priority(struct dma_fence *fence,
+				  const struct i915_sched_attr *attr)
+{
+	if (dma_fence_is_signaled(fence))
+		return;
+
+	local_bh_disable();
+
 	/* Recurse once into a fence-array */
 	if (dma_fence_is_array(fence)) {
 		struct dma_fence_array *array = to_dma_fence_array(fence);
 		int i;
 
 		for (i = 0; i < array->num_fences; i++)
-			__fence_set_priority(array->fences[i], attr);
+			fence_set_priority(array->fences[i], attr);
+	} else if (__dma_fence_is_chain(fence)) {
+		struct dma_fence *iter;
+
+		/* The chain is ordered; if we boost the last, we boost all */
+		dma_fence_chain_for_each(iter, fence) {
+			fence_set_priority(to_dma_fence_chain(iter)->fence,
+					   attr);
+			break;
+		}
+		dma_fence_put(iter);
 	} else {
-		__fence_set_priority(fence, attr);
+		fence_set_priority(fence, attr);
 	}
+
+	local_bh_enable(); /* kick the tasklets if queues were reprioritised */
 }
 
 int
@@ -141,12 +159,12 @@ i915_gem_object_wait_priority(struct drm_i915_gem_object *obj,
 		int ret;
 
 		ret = dma_resv_get_fences_rcu(obj->base.resv,
-							&excl, &count, &shared);
+					      &excl, &count, &shared);
 		if (ret)
 			return ret;
 
 		for (i = 0; i < count; i++) {
-			fence_set_priority(shared[i], attr);
+			i915_gem_fence_wait_priority(shared[i], attr);
 			dma_fence_put(shared[i]);
 		}
 
@@ -156,7 +174,7 @@ i915_gem_object_wait_priority(struct drm_i915_gem_object *obj,
 	}
 
 	if (excl) {
-		fence_set_priority(excl, attr);
+		i915_gem_fence_wait_priority(excl, attr);
 		dma_fence_put(excl);
 	}
 	return 0;
diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c b/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c
index a768ec61e966..2fb501a78a85 100644
--- a/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c
@@ -27,7 +27,7 @@ static void huge_free_pages(struct drm_i915_gem_object *obj,
 
 static int huge_get_pages(struct drm_i915_gem_object *obj)
 {
-#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY)
+#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_RETRY_MAYFAIL)
 	const unsigned long nreal = obj->scratch / PAGE_SIZE;
 	const unsigned long npages = obj->base.size / PAGE_SIZE;
 	struct scatterlist *sg, *src, *end;
diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
index 1f35e71429b4..aacf4856ccb4 100644
--- a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
@@ -368,6 +368,27 @@ static int igt_check_page_sizes(struct i915_vma *vma)
 		err = -EINVAL;
 	}
 
+	/*
+	 * The dma-api is like a box of chocolates when it comes to the
+	 * alignment of dma addresses, however for LMEM we have total control
+	 * and so can guarantee alignment, likewise when we allocate our blocks
+	 * they should appear in descending order, and if we know that we align
+	 * to the largest page size for the GTT address, we should be able to
+	 * assert that if we see 2M physical pages then we should also get 2M
+	 * GTT pages. If we don't then something might be wrong in our
+	 * construction of the backing pages.
+	 *
+	 * Maintaining alignment is required to utilise huge pages in the ppGGT.
+	 */
+	if (i915_gem_object_is_lmem(obj) &&
+	    IS_ALIGNED(vma->node.start, SZ_2M) &&
+	    vma->page_sizes.sg & SZ_2M &&
+	    vma->page_sizes.gtt < SZ_2M) {
+		pr_err("gtt pages mismatch for LMEM, expected 2M GTT pages, sg(%u), gtt(%u)\n",
+		       vma->page_sizes.sg, vma->page_sizes.gtt);
+		err = -EINVAL;
+	}
+
 	if (obj->mm.page_sizes.gtt) {
 		pr_err("obj->page_sizes.gtt(%u) should never be set\n",
 		       obj->mm.page_sizes.gtt);
@@ -1333,6 +1354,7 @@ static int igt_ppgtt_sanity_check(void *arg)
 		unsigned int flags;
 	} backends[] = {
 		{ igt_create_system, 0,                        },
+		{ igt_create_local,  0,                        },
 		{ igt_create_local,  I915_BO_ALLOC_CONTIGUOUS, },
 	};
 	struct {
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c
index 4e36d4897ea6..6a674a7994df 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c
@@ -20,13 +20,11 @@ static int __igt_client_fill(struct intel_engine_cs *engine)
 {
 	struct intel_context *ce = engine->kernel_context;
 	struct drm_i915_gem_object *obj;
-	struct rnd_state prng;
+	I915_RND_STATE(prng);
 	IGT_TIMEOUT(end);
 	u32 *vaddr;
 	int err = 0;
 
-	prandom_seed_state(&prng, i915_selftest.random_seed);
-
 	intel_engine_pm_get(engine);
 	do {
 		const u32 max_block_size = S16_MAX * PAGE_SIZE;
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
index 7049a6bbc03d..1117d2a44518 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
@@ -7,6 +7,7 @@
 #include <linux/prime_numbers.h>
 
 #include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_pm.h"
 #include "gt/intel_ring.h"
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c
index d27d87a678c8..d429c7643ff2 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c
@@ -7,6 +7,7 @@
 #include <linux/prime_numbers.h>
 
 #include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_pm.h"
 #include "gem/i915_gem_region.h"
diff --git a/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c b/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c
index e21b5023ca7d..d6783061bc72 100644
--- a/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c
+++ b/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c
@@ -9,6 +9,7 @@
 #include "gem/i915_gem_context.h"
 #include "gem/i915_gem_pm.h"
 #include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
 #include "i915_vma.h"
 #include "i915_drv.h"
diff --git a/drivers/gpu/drm/i915/gt/debugfs_gt_pm.c b/drivers/gpu/drm/i915/gt/debugfs_gt_pm.c
index 174a24553322..d4f4452ce5ed 100644
--- a/drivers/gpu/drm/i915/gt/debugfs_gt_pm.c
+++ b/drivers/gpu/drm/i915/gt/debugfs_gt_pm.c
@@ -11,6 +11,7 @@
 #include "i915_drv.h"
 #include "intel_gt.h"
 #include "intel_gt_clock_utils.h"
+#include "intel_gt_pm.h"
 #include "intel_llc.h"
 #include "intel_rc6.h"
 #include "intel_rps.h"
@@ -403,34 +404,34 @@ static int frequency_show(struct seq_file *m, void *unused)
 		seq_printf(m, "RPDECLIMIT: 0x%08x\n", rpdeclimit);
 		seq_printf(m, "RPNSWREQ: %dMHz\n", reqf);
 		seq_printf(m, "CAGF: %dMHz\n", cagf);
-		seq_printf(m, "RP CUR UP EI: %d (%dns)\n",
+		seq_printf(m, "RP CUR UP EI: %d (%lldns)\n",
 			   rpcurupei,
 			   intel_gt_pm_interval_to_ns(gt, rpcurupei));
-		seq_printf(m, "RP CUR UP: %d (%dns)\n",
+		seq_printf(m, "RP CUR UP: %d (%lldns)\n",
 			   rpcurup, intel_gt_pm_interval_to_ns(gt, rpcurup));
-		seq_printf(m, "RP PREV UP: %d (%dns)\n",
+		seq_printf(m, "RP PREV UP: %d (%lldns)\n",
 			   rpprevup, intel_gt_pm_interval_to_ns(gt, rpprevup));
 		seq_printf(m, "Up threshold: %d%%\n",
 			   rps->power.up_threshold);
-		seq_printf(m, "RP UP EI: %d (%dns)\n",
+		seq_printf(m, "RP UP EI: %d (%lldns)\n",
 			   rpupei, intel_gt_pm_interval_to_ns(gt, rpupei));
-		seq_printf(m, "RP UP THRESHOLD: %d (%dns)\n",
+		seq_printf(m, "RP UP THRESHOLD: %d (%lldns)\n",
 			   rpupt, intel_gt_pm_interval_to_ns(gt, rpupt));
 
-		seq_printf(m, "RP CUR DOWN EI: %d (%dns)\n",
+		seq_printf(m, "RP CUR DOWN EI: %d (%lldns)\n",
 			   rpcurdownei,
 			   intel_gt_pm_interval_to_ns(gt, rpcurdownei));
-		seq_printf(m, "RP CUR DOWN: %d (%dns)\n",
+		seq_printf(m, "RP CUR DOWN: %d (%lldns)\n",
 			   rpcurdown,
 			   intel_gt_pm_interval_to_ns(gt, rpcurdown));
-		seq_printf(m, "RP PREV DOWN: %d (%dns)\n",
+		seq_printf(m, "RP PREV DOWN: %d (%lldns)\n",
 			   rpprevdown,
 			   intel_gt_pm_interval_to_ns(gt, rpprevdown));
 		seq_printf(m, "Down threshold: %d%%\n",
 			   rps->power.down_threshold);
-		seq_printf(m, "RP DOWN EI: %d (%dns)\n",
+		seq_printf(m, "RP DOWN EI: %d (%lldns)\n",
 			   rpdownei, intel_gt_pm_interval_to_ns(gt, rpdownei));
-		seq_printf(m, "RP DOWN THRESHOLD: %d (%dns)\n",
+		seq_printf(m, "RP DOWN THRESHOLD: %d (%lldns)\n",
 			   rpdownt, intel_gt_pm_interval_to_ns(gt, rpdownt));
 
 		max_freq = (IS_GEN9_LP(i915) ? rp_state_cap >> 0 :
@@ -558,7 +559,9 @@ static int rps_boost_show(struct seq_file *m, void *data)
 
 	seq_printf(m, "RPS enabled? %s\n", yesno(intel_rps_is_enabled(rps)));
 	seq_printf(m, "RPS active? %s\n", yesno(intel_rps_is_active(rps)));
-	seq_printf(m, "GPU busy? %s\n", yesno(gt->awake));
+	seq_printf(m, "GPU busy? %s, %llums\n",
+		   yesno(gt->awake),
+		   ktime_to_ms(intel_gt_get_awake_time(gt)));
 	seq_printf(m, "Boosts outstanding? %d\n",
 		   atomic_read(&rps->num_waiters));
 	seq_printf(m, "Interactive? %d\n", READ_ONCE(rps->power.interactive));
@@ -575,7 +578,7 @@ static int rps_boost_show(struct seq_file *m, void *data)
 		   intel_gpu_freq(rps, rps->efficient_freq),
 		   intel_gpu_freq(rps, rps->boost_freq));
 
-	seq_printf(m, "Wait boosts: %d\n", atomic_read(&rps->boosts));
+	seq_printf(m, "Wait boosts: %d\n", READ_ONCE(rps->boosts));
 
 	if (INTEL_GEN(i915) >= 6 && intel_rps_is_active(rps)) {
 		struct intel_uncore *uncore = gt->uncore;
diff --git a/drivers/gpu/drm/i915/gt/gen6_ppgtt.c b/drivers/gpu/drm/i915/gt/gen6_ppgtt.c
index 680bd9442eb0..e08dff376339 100644
--- a/drivers/gpu/drm/i915/gt/gen6_ppgtt.c
+++ b/drivers/gpu/drm/i915/gt/gen6_ppgtt.c
@@ -12,9 +12,9 @@
 #include "intel_gt.h"
 
 /* Write pde (index) from the page directory @pd to the page table @pt */
-static inline void gen6_write_pde(const struct gen6_ppgtt *ppgtt,
-				  const unsigned int pde,
-				  const struct i915_page_table *pt)
+static void gen6_write_pde(const struct gen6_ppgtt *ppgtt,
+			   const unsigned int pde,
+			   const struct i915_page_table *pt)
 {
 	dma_addr_t addr = pt ? px_dma(pt) : px_dma(ppgtt->base.vm.scratch[1]);
 
@@ -27,8 +27,6 @@ void gen7_ppgtt_enable(struct intel_gt *gt)
 {
 	struct drm_i915_private *i915 = gt->i915;
 	struct intel_uncore *uncore = gt->uncore;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
 	u32 ecochk;
 
 	intel_uncore_rmw(uncore, GAC_ECO_BITS, 0, ECOBITS_PPGTT_CACHE64B);
@@ -41,13 +39,6 @@ void gen7_ppgtt_enable(struct intel_gt *gt)
 		ecochk &= ~ECOCHK_PPGTT_GFDT_IVB;
 	}
 	intel_uncore_write(uncore, GAM_ECOCHK, ecochk);
-
-	for_each_engine(engine, gt, id) {
-		/* GFX_MODE is per-ring on gen7+ */
-		ENGINE_WRITE(engine,
-			     RING_MODE_GEN7,
-			     _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-	}
 }
 
 void gen6_ppgtt_enable(struct intel_gt *gt)
diff --git a/drivers/gpu/drm/i915/gt/gen7_renderclear.c b/drivers/gpu/drm/i915/gt/gen7_renderclear.c
index d93d85cd3027..de575fdb033f 100644
--- a/drivers/gpu/drm/i915/gt/gen7_renderclear.c
+++ b/drivers/gpu/drm/i915/gt/gen7_renderclear.c
@@ -7,8 +7,6 @@
 #include "i915_drv.h"
 #include "intel_gpu_commands.h"
 
-#define MAX_URB_ENTRIES 64
-#define STATE_SIZE (4 * 1024)
 #define GT3_INLINE_DATA_DELAYS 0x1E00
 #define batch_advance(Y, CS) GEM_BUG_ON((Y)->end != (CS))
 
@@ -34,38 +32,59 @@ struct batch_chunk {
 };
 
 struct batch_vals {
-	u32 max_primitives;
-	u32 max_urb_entries;
-	u32 cmd_size;
-	u32 state_size;
+	u32 max_threads;
 	u32 state_start;
-	u32 batch_size;
+	u32 surface_start;
 	u32 surface_height;
 	u32 surface_width;
-	u32 scratch_size;
-	u32 max_size;
+	u32 size;
 };
 
+static int num_primitives(const struct batch_vals *bv)
+{
+	/*
+	 * We need to saturate the GPU with work in order to dispatch
+	 * a shader on every HW thread, and clear the thread-local registers.
+	 * In short, we have to dispatch work faster than the shaders can
+	 * run in order to fill the EU and occupy each HW thread.
+	 */
+	return bv->max_threads;
+}
+
 static void
 batch_get_defaults(struct drm_i915_private *i915, struct batch_vals *bv)
 {
 	if (IS_HASWELL(i915)) {
-		bv->max_primitives = 280;
-		bv->max_urb_entries = MAX_URB_ENTRIES;
+		switch (INTEL_INFO(i915)->gt) {
+		default:
+		case 1:
+			bv->max_threads = 70;
+			break;
+		case 2:
+			bv->max_threads = 140;
+			break;
+		case 3:
+			bv->max_threads = 280;
+			break;
+		}
 		bv->surface_height = 16 * 16;
 		bv->surface_width = 32 * 2 * 16;
 	} else {
-		bv->max_primitives = 128;
-		bv->max_urb_entries = MAX_URB_ENTRIES / 2;
+		switch (INTEL_INFO(i915)->gt) {
+		default:
+		case 1: /* including vlv */
+			bv->max_threads = 36;
+			break;
+		case 2:
+			bv->max_threads = 128;
+			break;
+		}
 		bv->surface_height = 16 * 8;
 		bv->surface_width = 32 * 16;
 	}
-	bv->cmd_size = bv->max_primitives * 4096;
-	bv->state_size = STATE_SIZE;
-	bv->state_start = bv->cmd_size;
-	bv->batch_size = bv->cmd_size + bv->state_size;
-	bv->scratch_size = bv->surface_height * bv->surface_width;
-	bv->max_size = bv->batch_size + bv->scratch_size;
+	bv->state_start = round_up(SZ_1K + num_primitives(bv) * 64, SZ_4K);
+	bv->surface_start = bv->state_start + SZ_4K;
+	bv->size = bv->surface_start + bv->surface_height * bv->surface_width;
 }
 
 static void batch_init(struct batch_chunk *bc,
@@ -155,7 +174,8 @@ static u32
 gen7_fill_binding_table(struct batch_chunk *state,
 			const struct batch_vals *bv)
 {
-	u32 surface_start = gen7_fill_surface_state(state, bv->batch_size, bv);
+	u32 surface_start =
+		gen7_fill_surface_state(state, bv->surface_start, bv);
 	u32 *cs = batch_alloc_items(state, 32, 8);
 	u32 offset = batch_offset(state, cs);
 
@@ -214,13 +234,13 @@ static void
 gen7_emit_state_base_address(struct batch_chunk *batch,
 			     u32 surface_state_base)
 {
-	u32 *cs = batch_alloc_items(batch, 0, 12);
+	u32 *cs = batch_alloc_items(batch, 0, 10);
 
-	*cs++ = STATE_BASE_ADDRESS | (12 - 2);
+	*cs++ = STATE_BASE_ADDRESS | (10 - 2);
 	/* general */
 	*cs++ = batch_addr(batch) | BASE_ADDRESS_MODIFY;
 	/* surface */
-	*cs++ = batch_addr(batch) | surface_state_base | BASE_ADDRESS_MODIFY;
+	*cs++ = (batch_addr(batch) + surface_state_base) | BASE_ADDRESS_MODIFY;
 	/* dynamic */
 	*cs++ = batch_addr(batch) | BASE_ADDRESS_MODIFY;
 	/* indirect */
@@ -233,8 +253,6 @@ gen7_emit_state_base_address(struct batch_chunk *batch,
 	*cs++ = BASE_ADDRESS_MODIFY;
 	*cs++ = 0;
 	*cs++ = BASE_ADDRESS_MODIFY;
-	*cs++ = 0;
-	*cs++ = 0;
 	batch_advance(batch, cs);
 }
 
@@ -244,8 +262,7 @@ gen7_emit_vfe_state(struct batch_chunk *batch,
 		    u32 urb_size, u32 curbe_size,
 		    u32 mode)
 {
-	u32 urb_entries = bv->max_urb_entries;
-	u32 threads = bv->max_primitives - 1;
+	u32 threads = bv->max_threads - 1;
 	u32 *cs = batch_alloc_items(batch, 32, 8);
 
 	*cs++ = MEDIA_VFE_STATE | (8 - 2);
@@ -254,7 +271,7 @@ gen7_emit_vfe_state(struct batch_chunk *batch,
 	*cs++ = 0;
 
 	/* number of threads & urb entries for GPGPU vs Media Mode */
-	*cs++ = threads << 16 | urb_entries << 8 | mode << 2;
+	*cs++ = threads << 16 | 1 << 8 | mode << 2;
 
 	*cs++ = 0;
 
@@ -293,17 +310,12 @@ gen7_emit_media_object(struct batch_chunk *batch,
 {
 	unsigned int x_offset = (media_object_index % 16) * 64;
 	unsigned int y_offset = (media_object_index / 16) * 16;
-	unsigned int inline_data_size;
-	unsigned int media_batch_size;
-	unsigned int i;
+	unsigned int pkt = 6 + 3;
 	u32 *cs;
 
-	inline_data_size = 112 * 8;
-	media_batch_size = inline_data_size + 6;
-
-	cs = batch_alloc_items(batch, 8, media_batch_size);
+	cs = batch_alloc_items(batch, 8, pkt);
 
-	*cs++ = MEDIA_OBJECT | (media_batch_size - 2);
+	*cs++ = MEDIA_OBJECT | (pkt - 2);
 
 	/* interface descriptor offset */
 	*cs++ = 0;
@@ -317,25 +329,46 @@ gen7_emit_media_object(struct batch_chunk *batch,
 	*cs++ = 0;
 
 	/* inline */
-	*cs++ = (y_offset << 16) | (x_offset);
+	*cs++ = y_offset << 16 | x_offset;
 	*cs++ = 0;
 	*cs++ = GT3_INLINE_DATA_DELAYS;
-	for (i = 3; i < inline_data_size; i++)
-		*cs++ = 0;
 
 	batch_advance(batch, cs);
 }
 
 static void gen7_emit_pipeline_flush(struct batch_chunk *batch)
 {
-	u32 *cs = batch_alloc_items(batch, 0, 5);
+	u32 *cs = batch_alloc_items(batch, 0, 4);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+		PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+		PIPE_CONTROL_DC_FLUSH_ENABLE |
+		PIPE_CONTROL_CS_STALL;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	batch_advance(batch, cs);
+}
+
+static void gen7_emit_pipeline_invalidate(struct batch_chunk *batch)
+{
+	u32 *cs = batch_alloc_items(batch, 0, 10);
+
+	/* ivb: Stall before STATE_CACHE_INVALIDATE */
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_STALL_AT_SCOREBOARD |
+		PIPE_CONTROL_CS_STALL;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
 
 	*cs++ = GFX_OP_PIPE_CONTROL(5);
-	*cs++ = PIPE_CONTROL_STATE_CACHE_INVALIDATE |
-		PIPE_CONTROL_GLOBAL_GTT_IVB;
+	*cs++ = PIPE_CONTROL_STATE_CACHE_INVALIDATE;
 	*cs++ = 0;
 	*cs++ = 0;
 	*cs++ = 0;
+
 	batch_advance(batch, cs);
 }
 
@@ -344,34 +377,48 @@ static void emit_batch(struct i915_vma * const vma,
 		       const struct batch_vals *bv)
 {
 	struct drm_i915_private *i915 = vma->vm->i915;
-	unsigned int desc_count = 64;
-	const u32 urb_size = 112;
+	const unsigned int desc_count = 1;
+	const unsigned int urb_size = 1;
 	struct batch_chunk cmds, state;
-	u32 interface_descriptor;
+	u32 descriptors;
 	unsigned int i;
 
-	batch_init(&cmds, vma, start, 0, bv->cmd_size);
-	batch_init(&state, vma, start, bv->state_start, bv->state_size);
+	batch_init(&cmds, vma, start, 0, bv->state_start);
+	batch_init(&state, vma, start, bv->state_start, SZ_4K);
 
-	interface_descriptor =
-		gen7_fill_interface_descriptor(&state, bv,
-					       IS_HASWELL(i915) ?
-					       &cb_kernel_hsw :
-					       &cb_kernel_ivb,
-					       desc_count);
+	descriptors = gen7_fill_interface_descriptor(&state, bv,
+						     IS_HASWELL(i915) ?
+						     &cb_kernel_hsw :
+						     &cb_kernel_ivb,
+						     desc_count);
+
+	/* Reset inherited context registers */
+	gen7_emit_pipeline_flush(&cmds);
+	gen7_emit_pipeline_invalidate(&cmds);
+	batch_add(&cmds, MI_LOAD_REGISTER_IMM(2));
+	batch_add(&cmds, i915_mmio_reg_offset(CACHE_MODE_0_GEN7));
+	batch_add(&cmds, 0xffff0000);
+	batch_add(&cmds, i915_mmio_reg_offset(CACHE_MODE_1));
+	batch_add(&cmds, 0xffff0000 | PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
+	gen7_emit_pipeline_invalidate(&cmds);
 	gen7_emit_pipeline_flush(&cmds);
+
+	/* Switch to the media pipeline and our base address */
+	gen7_emit_pipeline_invalidate(&cmds);
 	batch_add(&cmds, PIPELINE_SELECT | PIPELINE_SELECT_MEDIA);
 	batch_add(&cmds, MI_NOOP);
-	gen7_emit_state_base_address(&cmds, interface_descriptor);
+	gen7_emit_pipeline_invalidate(&cmds);
+
 	gen7_emit_pipeline_flush(&cmds);
+	gen7_emit_state_base_address(&cmds, descriptors);
+	gen7_emit_pipeline_invalidate(&cmds);
 
+	/* Set the clear-residual kernel state */
 	gen7_emit_vfe_state(&cmds, bv, urb_size - 1, 0, 0);
+	gen7_emit_interface_descriptor_load(&cmds, descriptors, desc_count);
 
-	gen7_emit_interface_descriptor_load(&cmds,
-					    interface_descriptor,
-					    desc_count);
-
-	for (i = 0; i < bv->max_primitives; i++)
+	/* Execute the kernel on all HW threads */
+	for (i = 0; i < num_primitives(bv); i++)
 		gen7_emit_media_object(&cmds, i);
 
 	batch_add(&cmds, MI_BATCH_BUFFER_END);
@@ -385,15 +432,15 @@ int gen7_setup_clear_gpr_bb(struct intel_engine_cs * const engine,
 
 	batch_get_defaults(engine->i915, &bv);
 	if (!vma)
-		return bv.max_size;
+		return bv.size;
 
-	GEM_BUG_ON(vma->obj->base.size < bv.max_size);
+	GEM_BUG_ON(vma->obj->base.size < bv.size);
 
 	batch = i915_gem_object_pin_map(vma->obj, I915_MAP_WC);
 	if (IS_ERR(batch))
 		return PTR_ERR(batch);
 
-	emit_batch(vma, memset(batch, 0, bv.max_size), &bv);
+	emit_batch(vma, memset(batch, 0, bv.size), &bv);
 
 	i915_gem_object_flush_map(vma->obj);
 	__i915_gem_object_release_map(vma->obj);
diff --git a/drivers/gpu/drm/i915/gt/gen8_engine_cs.c b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
new file mode 100644
index 000000000000..07ba524da90b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
@@ -0,0 +1,635 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#include "gen8_engine_cs.h"
+#include "i915_drv.h"
+#include "intel_lrc.h"
+#include "intel_gpu_commands.h"
+#include "intel_ring.h"
+
+int gen8_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	bool vf_flush_wa = false, dc_flush_wa = false;
+	u32 *cs, flags = 0;
+	int len;
+
+	flags |= PIPE_CONTROL_CS_STALL;
+
+	if (mode & EMIT_FLUSH) {
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+	}
+
+	if (mode & EMIT_INVALIDATE) {
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+		if (IS_GEN(rq->engine->i915, 9))
+			vf_flush_wa = true;
+
+		/* WaForGAMHang:kbl */
+		if (IS_KBL_GT_REVID(rq->engine->i915, 0, KBL_REVID_B0))
+			dc_flush_wa = true;
+	}
+
+	len = 6;
+
+	if (vf_flush_wa)
+		len += 6;
+
+	if (dc_flush_wa)
+		len += 12;
+
+	cs = intel_ring_begin(rq, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (vf_flush_wa)
+		cs = gen8_emit_pipe_control(cs, 0, 0);
+
+	if (dc_flush_wa)
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_DC_FLUSH_ENABLE,
+					    0);
+
+	cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+
+	if (dc_flush_wa)
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_CS_STALL, 0);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen8_emit_flush_xcs(struct i915_request *rq, u32 mode)
+{
+	u32 cmd, *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	cmd = MI_FLUSH_DW + 1;
+
+	/*
+	 * We always require a command barrier so that subsequent
+	 * commands, such as breadcrumb interrupts, are strictly ordered
+	 * wrt the contents of the write cache being flushed to memory
+	 * (and thus being coherent from the CPU).
+	 */
+	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
+
+	if (mode & EMIT_INVALIDATE) {
+		cmd |= MI_INVALIDATE_TLB;
+		if (rq->engine->class == VIDEO_DECODE_CLASS)
+			cmd |= MI_INVALIDATE_BSD;
+	}
+
+	*cs++ = cmd;
+	*cs++ = LRC_PPHWSP_SCRATCH_ADDR;
+	*cs++ = 0; /* upper addr */
+	*cs++ = 0; /* value */
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen11_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	if (mode & EMIT_FLUSH) {
+		u32 *cs;
+		u32 flags = 0;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		flags |= PIPE_CONTROL_TILE_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+
+		cs = intel_ring_begin(rq, 6);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+		intel_ring_advance(rq, cs);
+	}
+
+	if (mode & EMIT_INVALIDATE) {
+		u32 *cs;
+		u32 flags = 0;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		flags |= PIPE_CONTROL_COMMAND_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+
+		cs = intel_ring_begin(rq, 6);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+		intel_ring_advance(rq, cs);
+	}
+
+	return 0;
+}
+
+static u32 preparser_disable(bool state)
+{
+	return MI_ARB_CHECK | 1 << 8 | state;
+}
+
+static i915_reg_t aux_inv_reg(const struct intel_engine_cs *engine)
+{
+	static const i915_reg_t vd[] = {
+		GEN12_VD0_AUX_NV,
+		GEN12_VD1_AUX_NV,
+		GEN12_VD2_AUX_NV,
+		GEN12_VD3_AUX_NV,
+	};
+
+	static const i915_reg_t ve[] = {
+		GEN12_VE0_AUX_NV,
+		GEN12_VE1_AUX_NV,
+	};
+
+	if (engine->class == VIDEO_DECODE_CLASS)
+		return vd[engine->instance];
+
+	if (engine->class == VIDEO_ENHANCEMENT_CLASS)
+		return ve[engine->instance];
+
+	GEM_BUG_ON("unknown aux_inv reg\n");
+	return INVALID_MMIO_REG;
+}
+
+static u32 *gen12_emit_aux_table_inv(const i915_reg_t inv_reg, u32 *cs)
+{
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(inv_reg);
+	*cs++ = AUX_INV;
+	*cs++ = MI_NOOP;
+
+	return cs;
+}
+
+int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	if (mode & EMIT_FLUSH) {
+		u32 flags = 0;
+		u32 *cs;
+
+		flags |= PIPE_CONTROL_TILE_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_FLUSH_L3;
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		/* Wa_1409600907:tgl */
+		flags |= PIPE_CONTROL_DEPTH_STALL;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+		flags |= PIPE_CONTROL_QW_WRITE;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		cs = intel_ring_begin(rq, 6);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		cs = gen12_emit_pipe_control(cs,
+					     PIPE_CONTROL0_HDC_PIPELINE_FLUSH,
+					     flags, LRC_PPHWSP_SCRATCH_ADDR);
+		intel_ring_advance(rq, cs);
+	}
+
+	if (mode & EMIT_INVALIDATE) {
+		u32 flags = 0;
+		u32 *cs;
+
+		flags |= PIPE_CONTROL_COMMAND_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+		flags |= PIPE_CONTROL_QW_WRITE;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		cs = intel_ring_begin(rq, 8 + 4);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		/*
+		 * Prevent the pre-parser from skipping past the TLB
+		 * invalidate and loading a stale page for the batch
+		 * buffer / request payload.
+		 */
+		*cs++ = preparser_disable(true);
+
+		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+
+		/* hsdes: 1809175790 */
+		cs = gen12_emit_aux_table_inv(GEN12_GFX_CCS_AUX_NV, cs);
+
+		*cs++ = preparser_disable(false);
+		intel_ring_advance(rq, cs);
+	}
+
+	return 0;
+}
+
+int gen12_emit_flush_xcs(struct i915_request *rq, u32 mode)
+{
+	intel_engine_mask_t aux_inv = 0;
+	u32 cmd, *cs;
+
+	cmd = 4;
+	if (mode & EMIT_INVALIDATE)
+		cmd += 2;
+	if (mode & EMIT_INVALIDATE)
+		aux_inv = rq->engine->mask & ~BIT(BCS0);
+	if (aux_inv)
+		cmd += 2 * hweight8(aux_inv) + 2;
+
+	cs = intel_ring_begin(rq, cmd);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (mode & EMIT_INVALIDATE)
+		*cs++ = preparser_disable(true);
+
+	cmd = MI_FLUSH_DW + 1;
+
+	/*
+	 * We always require a command barrier so that subsequent
+	 * commands, such as breadcrumb interrupts, are strictly ordered
+	 * wrt the contents of the write cache being flushed to memory
+	 * (and thus being coherent from the CPU).
+	 */
+	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
+
+	if (mode & EMIT_INVALIDATE) {
+		cmd |= MI_INVALIDATE_TLB;
+		if (rq->engine->class == VIDEO_DECODE_CLASS)
+			cmd |= MI_INVALIDATE_BSD;
+	}
+
+	*cs++ = cmd;
+	*cs++ = LRC_PPHWSP_SCRATCH_ADDR;
+	*cs++ = 0; /* upper addr */
+	*cs++ = 0; /* value */
+
+	if (aux_inv) { /* hsdes: 1809175790 */
+		struct intel_engine_cs *engine;
+		unsigned int tmp;
+
+		*cs++ = MI_LOAD_REGISTER_IMM(hweight8(aux_inv));
+		for_each_engine_masked(engine, rq->engine->gt,
+				       aux_inv, tmp) {
+			*cs++ = i915_mmio_reg_offset(aux_inv_reg(engine));
+			*cs++ = AUX_INV;
+		}
+		*cs++ = MI_NOOP;
+	}
+
+	if (mode & EMIT_INVALIDATE)
+		*cs++ = preparser_disable(false);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static u32 preempt_address(struct intel_engine_cs *engine)
+{
+	return (i915_ggtt_offset(engine->status_page.vma) +
+		I915_GEM_HWS_PREEMPT_ADDR);
+}
+
+static u32 hwsp_offset(const struct i915_request *rq)
+{
+	const struct intel_timeline_cacheline *cl;
+
+	/* Before the request is executed, the timeline/cachline is fixed */
+
+	cl = rcu_dereference_protected(rq->hwsp_cacheline, 1);
+	if (cl)
+		return cl->ggtt_offset;
+
+	return rcu_dereference_protected(rq->timeline, 1)->hwsp_offset;
+}
+
+int gen8_emit_init_breadcrumb(struct i915_request *rq)
+{
+	u32 *cs;
+
+	GEM_BUG_ON(i915_request_has_initial_breadcrumb(rq));
+	if (!i915_request_timeline(rq)->has_initial_breadcrumb)
+		return 0;
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = hwsp_offset(rq);
+	*cs++ = 0;
+	*cs++ = rq->fence.seqno - 1;
+
+	/*
+	 * Check if we have been preempted before we even get started.
+	 *
+	 * After this point i915_request_started() reports true, even if
+	 * we get preempted and so are no longer running.
+	 *
+	 * i915_request_started() is used during preemption processing
+	 * to decide if the request is currently inside the user payload
+	 * or spinning on a kernel semaphore (or earlier). For no-preemption
+	 * requests, we do allow preemption on the semaphore before the user
+	 * payload, but do not allow preemption once the request is started.
+	 *
+	 * i915_request_started() is similarly used during GPU hangs to
+	 * determine if the user's payload was guilty, and if so, the
+	 * request is banned. Before the request is started, it is assumed
+	 * to be unharmed and an innocent victim of another's hang.
+	 */
+	*cs++ = MI_NOOP;
+	*cs++ = MI_ARB_CHECK;
+
+	intel_ring_advance(rq, cs);
+
+	/* Record the updated position of the request's payload */
+	rq->infix = intel_ring_offset(rq, cs);
+
+	__set_bit(I915_FENCE_FLAG_INITIAL_BREADCRUMB, &rq->fence.flags);
+
+	return 0;
+}
+
+int gen8_emit_bb_start_noarb(struct i915_request *rq,
+			     u64 offset, u32 len,
+			     const unsigned int flags)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/*
+	 * WaDisableCtxRestoreArbitration:bdw,chv
+	 *
+	 * We don't need to perform MI_ARB_ENABLE as often as we do (in
+	 * particular all the gen that do not need the w/a at all!), if we
+	 * took care to make sure that on every switch into this context
+	 * (both ordinary and for preemption) that arbitrartion was enabled
+	 * we would be fine.  However, for gen8 there is another w/a that
+	 * requires us to not preempt inside GPGPU execution, so we keep
+	 * arbitration disabled for gen8 batches. Arbitration will be
+	 * re-enabled before we close the request
+	 * (engine->emit_fini_breadcrumb).
+	 */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	/* FIXME(BDW+): Address space and security selectors. */
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen8_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       const unsigned int flags)
+{
+	u32 *cs;
+
+	if (unlikely(i915_request_has_nopreempt(rq)))
+		return gen8_emit_bb_start_noarb(rq, offset, len, flags);
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static void assert_request_valid(struct i915_request *rq)
+{
+	struct intel_ring *ring __maybe_unused = rq->ring;
+
+	/* Can we unwind this request without appearing to go forwards? */
+	GEM_BUG_ON(intel_ring_direction(ring, rq->wa_tail, rq->head) <= 0);
+}
+
+/*
+ * Reserve space for 2 NOOPs at the end of each request to be
+ * used as a workaround for not being allowed to do lite
+ * restore with HEAD==TAIL (WaIdleLiteRestore).
+ */
+static u32 *gen8_emit_wa_tail(struct i915_request *rq, u32 *cs)
+{
+	/* Ensure there's always at least one preemption point per-request. */
+	*cs++ = MI_ARB_CHECK;
+	*cs++ = MI_NOOP;
+	rq->wa_tail = intel_ring_offset(rq, cs);
+
+	/* Check that entire request is less than half the ring */
+	assert_request_valid(rq);
+
+	return cs;
+}
+
+static u32 *emit_preempt_busywait(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_ARB_CHECK; /* trigger IDLE->ACTIVE first */
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = preempt_address(rq->engine);
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+
+	return cs;
+}
+
+static __always_inline u32*
+gen8_emit_fini_breadcrumb_tail(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_USER_INTERRUPT;
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	if (intel_engine_has_semaphores(rq->engine))
+		cs = emit_preempt_busywait(rq, cs);
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return gen8_emit_wa_tail(rq, cs);
+}
+
+static u32 *emit_xcs_breadcrumb(struct i915_request *rq, u32 *cs)
+{
+	return gen8_emit_ggtt_write(cs, rq->fence.seqno, hwsp_offset(rq), 0);
+}
+
+u32 *gen8_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs)
+{
+	return gen8_emit_fini_breadcrumb_tail(rq, emit_xcs_breadcrumb(rq, cs));
+}
+
+u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	cs = gen8_emit_pipe_control(cs,
+				    PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+				    PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+				    PIPE_CONTROL_DC_FLUSH_ENABLE,
+				    0);
+
+	/* XXX flush+write+CS_STALL all in one upsets gem_concurrent_blt:kbl */
+	cs = gen8_emit_ggtt_write_rcs(cs,
+				      rq->fence.seqno,
+				      hwsp_offset(rq),
+				      PIPE_CONTROL_FLUSH_ENABLE |
+				      PIPE_CONTROL_CS_STALL);
+
+	return gen8_emit_fini_breadcrumb_tail(rq, cs);
+}
+
+u32 *gen11_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	cs = gen8_emit_ggtt_write_rcs(cs,
+				      rq->fence.seqno,
+				      hwsp_offset(rq),
+				      PIPE_CONTROL_CS_STALL |
+				      PIPE_CONTROL_TILE_CACHE_FLUSH |
+				      PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+				      PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+				      PIPE_CONTROL_DC_FLUSH_ENABLE |
+				      PIPE_CONTROL_FLUSH_ENABLE);
+
+	return gen8_emit_fini_breadcrumb_tail(rq, cs);
+}
+
+/*
+ * Note that the CS instruction pre-parser will not stall on the breadcrumb
+ * flush and will continue pre-fetching the instructions after it before the
+ * memory sync is completed. On pre-gen12 HW, the pre-parser will stop at
+ * BB_START/END instructions, so, even though we might pre-fetch the pre-amble
+ * of the next request before the memory has been flushed, we're guaranteed that
+ * we won't access the batch itself too early.
+ * However, on gen12+ the parser can pre-fetch across the BB_START/END commands,
+ * so, if the current request is modifying an instruction in the next request on
+ * the same intel_context, we might pre-fetch and then execute the pre-update
+ * instruction. To avoid this, the users of self-modifying code should either
+ * disable the parser around the code emitting the memory writes, via a new flag
+ * added to MI_ARB_CHECK, or emit the writes from a different intel_context. For
+ * the in-kernel use-cases we've opted to use a separate context, see
+ * reloc_gpu() as an example.
+ * All the above applies only to the instructions themselves. Non-inline data
+ * used by the instructions is not pre-fetched.
+ */
+
+static u32 *gen12_emit_preempt_busywait(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_ARB_CHECK; /* trigger IDLE->ACTIVE first */
+	*cs++ = MI_SEMAPHORE_WAIT_TOKEN |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = preempt_address(rq->engine);
+	*cs++ = 0;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static __always_inline u32*
+gen12_emit_fini_breadcrumb_tail(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_USER_INTERRUPT;
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	if (intel_engine_has_semaphores(rq->engine))
+		cs = gen12_emit_preempt_busywait(rq, cs);
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return gen8_emit_wa_tail(rq, cs);
+}
+
+u32 *gen12_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs)
+{
+	/* XXX Stalling flush before seqno write; post-sync not */
+	cs = emit_xcs_breadcrumb(rq, __gen8_emit_flush_dw(cs, 0, 0, 0));
+	return gen12_emit_fini_breadcrumb_tail(rq, cs);
+}
+
+u32 *gen12_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	cs = gen12_emit_ggtt_write_rcs(cs,
+				       rq->fence.seqno,
+				       hwsp_offset(rq),
+				       PIPE_CONTROL0_HDC_PIPELINE_FLUSH,
+				       PIPE_CONTROL_CS_STALL |
+				       PIPE_CONTROL_TILE_CACHE_FLUSH |
+				       PIPE_CONTROL_FLUSH_L3 |
+				       PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+				       PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+				       /* Wa_1409600907:tgl */
+				       PIPE_CONTROL_DEPTH_STALL |
+				       PIPE_CONTROL_DC_FLUSH_ENABLE |
+				       PIPE_CONTROL_FLUSH_ENABLE);
+
+	return gen12_emit_fini_breadcrumb_tail(rq, cs);
+}
diff --git a/drivers/gpu/drm/i915/gt/gen8_engine_cs.h b/drivers/gpu/drm/i915/gt/gen8_engine_cs.h
new file mode 100644
index 000000000000..cc6e21d3662a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_engine_cs.h
@@ -0,0 +1,127 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#ifndef __GEN8_ENGINE_CS_H__
+#define __GEN8_ENGINE_CS_H__
+
+#include <linux/string.h>
+#include <linux/types.h>
+
+#include "i915_gem.h" /* GEM_BUG_ON */
+
+#include "intel_gpu_commands.h"
+
+struct i915_request;
+
+int gen8_emit_flush_rcs(struct i915_request *rq, u32 mode);
+int gen11_emit_flush_rcs(struct i915_request *rq, u32 mode);
+int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode);
+
+int gen8_emit_flush_xcs(struct i915_request *rq, u32 mode);
+int gen12_emit_flush_xcs(struct i915_request *rq, u32 mode);
+
+int gen8_emit_init_breadcrumb(struct i915_request *rq);
+
+int gen8_emit_bb_start_noarb(struct i915_request *rq,
+			     u64 offset, u32 len,
+			     const unsigned int flags);
+int gen8_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       const unsigned int flags);
+
+u32 *gen8_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs);
+u32 *gen12_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs);
+
+u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+u32 *gen11_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+u32 *gen12_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+
+static inline u32 *
+__gen8_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u32 offset)
+{
+	memset(batch, 0, 6 * sizeof(u32));
+
+	batch[0] = GFX_OP_PIPE_CONTROL(6) | flags0;
+	batch[1] = flags1;
+	batch[2] = offset;
+
+	return batch + 6;
+}
+
+static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset)
+{
+	return __gen8_emit_pipe_control(batch, 0, flags, offset);
+}
+
+static inline u32 *gen12_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u32 offset)
+{
+	return __gen8_emit_pipe_control(batch, flags0, flags1, offset);
+}
+
+static inline u32 *
+__gen8_emit_write_rcs(u32 *cs, u32 value, u32 offset, u32 flags0, u32 flags1)
+{
+	*cs++ = GFX_OP_PIPE_CONTROL(6) | flags0;
+	*cs++ = flags1 | PIPE_CONTROL_QW_WRITE;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = value;
+	*cs++ = 0; /* We're thrashing one extra dword. */
+
+	return cs;
+}
+
+static inline u32*
+gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
+{
+	/* We're using qword write, offset should be aligned to 8 bytes. */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	return __gen8_emit_write_rcs(cs,
+				     value,
+				     gtt_offset,
+				     0,
+				     flags | PIPE_CONTROL_GLOBAL_GTT_IVB);
+}
+
+static inline u32*
+gen12_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags0, u32 flags1)
+{
+	/* We're using qword write, offset should be aligned to 8 bytes. */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	return __gen8_emit_write_rcs(cs,
+				     value,
+				     gtt_offset,
+				     flags0,
+				     flags1 | PIPE_CONTROL_GLOBAL_GTT_IVB);
+}
+
+static inline u32 *
+__gen8_emit_flush_dw(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
+{
+	*cs++ = (MI_FLUSH_DW + 1) | flags;
+	*cs++ = gtt_offset;
+	*cs++ = 0;
+	*cs++ = value;
+
+	return cs;
+}
+
+static inline u32 *
+gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
+{
+	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
+	GEM_BUG_ON(gtt_offset & (1 << 5));
+	/* Offset should be aligned to 8 bytes for both (QW/DW) write types */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	return __gen8_emit_flush_dw(cs,
+				    value,
+				    gtt_offset | MI_FLUSH_DW_USE_GTT,
+				    flags | MI_FLUSH_DW_OP_STOREDW);
+}
+
+#endif /* __GEN8_ENGINE_CS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/gen8_ppgtt.c b/drivers/gpu/drm/i915/gt/gen8_ppgtt.c
index a37c968ef8f7..755522ced60d 100644
--- a/drivers/gpu/drm/i915/gt/gen8_ppgtt.c
+++ b/drivers/gpu/drm/i915/gt/gen8_ppgtt.c
@@ -109,7 +109,7 @@ static void gen8_ppgtt_notify_vgt(struct i915_ppgtt *ppgtt, bool create)
 
 #define as_pd(x) container_of((x), typeof(struct i915_page_directory), pt)
 
-static inline unsigned int
+static unsigned int
 gen8_pd_range(u64 start, u64 end, int lvl, unsigned int *idx)
 {
 	const int shift = gen8_pd_shift(lvl);
@@ -125,7 +125,7 @@ gen8_pd_range(u64 start, u64 end, int lvl, unsigned int *idx)
 		return i915_pde_index(end, shift) - *idx;
 }
 
-static inline bool gen8_pd_contains(u64 start, u64 end, int lvl)
+static bool gen8_pd_contains(u64 start, u64 end, int lvl)
 {
 	const u64 mask = ~0ull << gen8_pd_shift(lvl + 1);
 
@@ -133,7 +133,7 @@ static inline bool gen8_pd_contains(u64 start, u64 end, int lvl)
 	return (start ^ end) & mask && (start & ~mask) == 0;
 }
 
-static inline unsigned int gen8_pt_count(u64 start, u64 end)
+static unsigned int gen8_pt_count(u64 start, u64 end)
 {
 	GEM_BUG_ON(start >= end);
 	if ((start ^ end) >> gen8_pd_shift(1))
@@ -142,14 +142,13 @@ static inline unsigned int gen8_pt_count(u64 start, u64 end)
 		return end - start;
 }
 
-static inline unsigned int
-gen8_pd_top_count(const struct i915_address_space *vm)
+static unsigned int gen8_pd_top_count(const struct i915_address_space *vm)
 {
 	unsigned int shift = __gen8_pte_shift(vm->top);
 	return (vm->total + (1ull << shift) - 1) >> shift;
 }
 
-static inline struct i915_page_directory *
+static struct i915_page_directory *
 gen8_pdp_for_page_index(struct i915_address_space * const vm, const u64 idx)
 {
 	struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(vm);
@@ -160,7 +159,7 @@ gen8_pdp_for_page_index(struct i915_address_space * const vm, const u64 idx)
 		return i915_pd_entry(ppgtt->pd, gen8_pd_index(idx, vm->top));
 }
 
-static inline struct i915_page_directory *
+static struct i915_page_directory *
 gen8_pdp_for_page_address(struct i915_address_space * const vm, const u64 addr)
 {
 	return gen8_pdp_for_page_index(vm, addr >> GEN8_PTE_SHIFT);
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
index a24cc1ff08a0..34a645d6babd 100644
--- a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
@@ -134,11 +134,6 @@ static bool remove_signaling_context(struct intel_breadcrumbs *b,
 	return true;
 }
 
-static inline bool __request_completed(const struct i915_request *rq)
-{
-	return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno);
-}
-
 __maybe_unused static bool
 check_signal_order(struct intel_context *ce, struct i915_request *rq)
 {
@@ -192,18 +187,6 @@ static void add_retire(struct intel_breadcrumbs *b, struct intel_timeline *tl)
 		intel_engine_add_retire(b->irq_engine, tl);
 }
 
-static bool __signal_request(struct i915_request *rq)
-{
-	GEM_BUG_ON(test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags));
-
-	if (!__dma_fence_signal(&rq->fence)) {
-		i915_request_put(rq);
-		return false;
-	}
-
-	return true;
-}
-
 static struct llist_node *
 slist_add(struct llist_node *node, struct llist_node *head)
 {
@@ -251,13 +234,14 @@ static void signal_irq_work(struct irq_work *work)
 		intel_breadcrumbs_disarm_irq(b);
 
 	rcu_read_lock();
+	atomic_inc(&b->signaler_active);
 	list_for_each_entry_rcu(ce, &b->signalers, signal_link) {
 		struct i915_request *rq;
 
 		list_for_each_entry_rcu(rq, &ce->signals, signal_link) {
 			bool release;
 
-			if (!__request_completed(rq))
+			if (!__i915_request_is_complete(rq))
 				break;
 
 			if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL,
@@ -273,17 +257,20 @@ static void signal_irq_work(struct irq_work *work)
 			list_del_rcu(&rq->signal_link);
 			release = remove_signaling_context(b, ce);
 			spin_unlock(&ce->signal_lock);
-
-			if (__signal_request(rq))
-				/* We own signal_node now, xfer to local list */
-				signal = slist_add(&rq->signal_node, signal);
-
 			if (release) {
-				add_retire(b, ce->timeline);
+				if (intel_timeline_is_last(ce->timeline, rq))
+					add_retire(b, ce->timeline);
 				intel_context_put(ce);
 			}
+
+			if (__dma_fence_signal(&rq->fence))
+				/* We own signal_node now, xfer to local list */
+				signal = slist_add(&rq->signal_node, signal);
+			else
+				i915_request_put(rq);
 		}
 	}
+	atomic_dec(&b->signaler_active);
 	rcu_read_unlock();
 
 	llist_for_each_safe(signal, sn, signal) {
@@ -342,17 +329,19 @@ void intel_breadcrumbs_reset(struct intel_breadcrumbs *b)
 	spin_unlock_irqrestore(&b->irq_lock, flags);
 }
 
-void intel_breadcrumbs_park(struct intel_breadcrumbs *b)
+void __intel_breadcrumbs_park(struct intel_breadcrumbs *b)
 {
-	/* Kick the work once more to drain the signalers */
+	if (!READ_ONCE(b->irq_armed))
+		return;
+
+	/* Kick the work once more to drain the signalers, and disarm the irq */
 	irq_work_sync(&b->irq_work);
-	while (unlikely(READ_ONCE(b->irq_armed))) {
+	while (READ_ONCE(b->irq_armed) && !atomic_read(&b->active)) {
 		local_irq_disable();
 		signal_irq_work(&b->irq_work);
 		local_irq_enable();
 		cond_resched();
 	}
-	GEM_BUG_ON(!list_empty(&b->signalers));
 }
 
 void intel_breadcrumbs_free(struct intel_breadcrumbs *b)
@@ -363,6 +352,17 @@ void intel_breadcrumbs_free(struct intel_breadcrumbs *b)
 	kfree(b);
 }
 
+static void irq_signal_request(struct i915_request *rq,
+			       struct intel_breadcrumbs *b)
+{
+	if (!__dma_fence_signal(&rq->fence))
+		return;
+
+	i915_request_get(rq);
+	if (llist_add(&rq->signal_node, &b->signaled_requests))
+		irq_work_queue(&b->irq_work);
+}
+
 static void insert_breadcrumb(struct i915_request *rq)
 {
 	struct intel_breadcrumbs *b = READ_ONCE(rq->engine)->breadcrumbs;
@@ -372,17 +372,13 @@ static void insert_breadcrumb(struct i915_request *rq)
 	if (test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags))
 		return;
 
-	i915_request_get(rq);
-
 	/*
 	 * If the request is already completed, we can transfer it
 	 * straight onto a signaled list, and queue the irq worker for
 	 * its signal completion.
 	 */
-	if (__request_completed(rq)) {
-		if (__signal_request(rq) &&
-		    llist_add(&rq->signal_node, &b->signaled_requests))
-			irq_work_queue(&b->irq_work);
+	if (__i915_request_is_complete(rq)) {
+		irq_signal_request(rq, b);
 		return;
 	}
 
@@ -413,6 +409,8 @@ static void insert_breadcrumb(struct i915_request *rq)
 				break;
 		}
 	}
+
+	i915_request_get(rq);
 	list_add_rcu(&rq->signal_link, pos);
 	GEM_BUG_ON(!check_signal_order(ce, rq));
 	GEM_BUG_ON(test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags));
@@ -453,22 +451,61 @@ bool i915_request_enable_breadcrumb(struct i915_request *rq)
 
 void i915_request_cancel_breadcrumb(struct i915_request *rq)
 {
+	struct intel_breadcrumbs *b = READ_ONCE(rq->engine)->breadcrumbs;
 	struct intel_context *ce = rq->context;
 	bool release;
 
-	if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags))
+	spin_lock(&ce->signal_lock);
+	if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)) {
+		spin_unlock(&ce->signal_lock);
 		return;
+	}
 
-	spin_lock(&ce->signal_lock);
 	list_del_rcu(&rq->signal_link);
-	release = remove_signaling_context(rq->engine->breadcrumbs, ce);
+	release = remove_signaling_context(b, ce);
 	spin_unlock(&ce->signal_lock);
 	if (release)
 		intel_context_put(ce);
 
+	if (__i915_request_is_complete(rq))
+		irq_signal_request(rq, b);
+
 	i915_request_put(rq);
 }
 
+void intel_context_remove_breadcrumbs(struct intel_context *ce,
+				      struct intel_breadcrumbs *b)
+{
+	struct i915_request *rq, *rn;
+	bool release = false;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ce->signal_lock, flags);
+
+	if (list_empty(&ce->signals))
+		goto unlock;
+
+	list_for_each_entry_safe(rq, rn, &ce->signals, signal_link) {
+		GEM_BUG_ON(!__i915_request_is_complete(rq));
+		if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL,
+					&rq->fence.flags))
+			continue;
+
+		list_del_rcu(&rq->signal_link);
+		irq_signal_request(rq, b);
+		i915_request_put(rq);
+	}
+	release = remove_signaling_context(b, ce);
+
+unlock:
+	spin_unlock_irqrestore(&ce->signal_lock, flags);
+	if (release)
+		intel_context_put(ce);
+
+	while (atomic_read(&b->signaler_active))
+		cpu_relax();
+}
+
 static void print_signals(struct intel_breadcrumbs *b, struct drm_printer *p)
 {
 	struct intel_context *ce;
@@ -481,8 +518,8 @@ static void print_signals(struct intel_breadcrumbs *b, struct drm_printer *p)
 		list_for_each_entry_rcu(rq, &ce->signals, signal_link)
 			drm_printf(p, "\t[%llx:%llx%s] @ %dms\n",
 				   rq->fence.context, rq->fence.seqno,
-				   i915_request_completed(rq) ? "!" :
-				   i915_request_started(rq) ? "*" :
+				   __i915_request_is_complete(rq) ? "!" :
+				   __i915_request_has_started(rq) ? "*" :
 				   "",
 				   jiffies_to_msecs(jiffies - rq->emitted_jiffies));
 	}
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h
index ed3d1deabfbd..3ce5ce270b04 100644
--- a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h
@@ -6,6 +6,7 @@
 #ifndef __INTEL_BREADCRUMBS__
 #define __INTEL_BREADCRUMBS__
 
+#include <linux/atomic.h>
 #include <linux/irq_work.h>
 
 #include "intel_engine_types.h"
@@ -19,7 +20,18 @@ intel_breadcrumbs_create(struct intel_engine_cs *irq_engine);
 void intel_breadcrumbs_free(struct intel_breadcrumbs *b);
 
 void intel_breadcrumbs_reset(struct intel_breadcrumbs *b);
-void intel_breadcrumbs_park(struct intel_breadcrumbs *b);
+void __intel_breadcrumbs_park(struct intel_breadcrumbs *b);
+
+static inline void intel_breadcrumbs_unpark(struct intel_breadcrumbs *b)
+{
+	atomic_inc(&b->active);
+}
+
+static inline void intel_breadcrumbs_park(struct intel_breadcrumbs *b)
+{
+	if (atomic_dec_and_test(&b->active))
+		__intel_breadcrumbs_park(b);
+}
 
 static inline void
 intel_engine_signal_breadcrumbs(struct intel_engine_cs *engine)
@@ -33,4 +45,7 @@ void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine,
 bool i915_request_enable_breadcrumb(struct i915_request *request);
 void i915_request_cancel_breadcrumb(struct i915_request *request);
 
+void intel_context_remove_breadcrumbs(struct intel_context *ce,
+				      struct intel_breadcrumbs *b);
+
 #endif /* __INTEL_BREADCRUMBS__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h b/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
index a74bb3062bd8..3a084ce8ff5e 100644
--- a/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
@@ -29,17 +29,20 @@
  * the overhead of waking that client is much preferred.
  */
 struct intel_breadcrumbs {
-	/* Not all breadcrumbs are attached to physical HW */
-	struct intel_engine_cs *irq_engine;
+	atomic_t active;
 
 	spinlock_t signalers_lock; /* protects the list of signalers */
 	struct list_head signalers;
 	struct llist_head signaled_requests;
+	atomic_t signaler_active;
 
 	spinlock_t irq_lock; /* protects the interrupt from hardirq context */
 	struct irq_work irq_work; /* for use from inside irq_lock */
 	unsigned int irq_enabled;
 	bool irq_armed;
+
+	/* Not all breadcrumbs are attached to physical HW */
+	struct intel_engine_cs *irq_engine;
 };
 
 #endif /* __INTEL_BREADCRUMBS_TYPES__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
index fda2eba81e22..d24ab6fa0ee5 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.h
+++ b/drivers/gpu/drm/i915/gt/intel_context.h
@@ -191,6 +191,11 @@ static inline bool intel_context_is_closed(const struct intel_context *ce)
 	return test_bit(CONTEXT_CLOSED_BIT, &ce->flags);
 }
 
+static inline bool intel_context_has_inflight(const struct intel_context *ce)
+{
+	return test_bit(COPS_HAS_INFLIGHT_BIT, &ce->ops->flags);
+}
+
 static inline bool intel_context_use_semaphores(const struct intel_context *ce)
 {
 	return test_bit(CONTEXT_USE_SEMAPHORES, &ce->flags);
@@ -248,16 +253,14 @@ intel_context_clear_nopreempt(struct intel_context *ce)
 
 static inline u64 intel_context_get_total_runtime_ns(struct intel_context *ce)
 {
-	const u32 period =
-		RUNTIME_INFO(ce->engine->i915)->cs_timestamp_period_ns;
+	const u32 period = ce->engine->gt->clock_period_ns;
 
 	return READ_ONCE(ce->runtime.total) * period;
 }
 
 static inline u64 intel_context_get_avg_runtime_ns(struct intel_context *ce)
 {
-	const u32 period =
-		RUNTIME_INFO(ce->engine->i915)->cs_timestamp_period_ns;
+	const u32 period = ce->engine->gt->clock_period_ns;
 
 	return mul_u32_u32(ewma_runtime_read(&ce->runtime.avg), period);
 }
diff --git a/drivers/gpu/drm/i915/gt/intel_context_sseu.c b/drivers/gpu/drm/i915/gt/intel_context_sseu.c
index b9c8163978a3..8dfd8f656aaa 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_sseu.c
+++ b/drivers/gpu/drm/i915/gt/intel_context_sseu.c
@@ -9,7 +9,6 @@
 #include "intel_engine_pm.h"
 #include "intel_gpu_commands.h"
 #include "intel_lrc.h"
-#include "intel_lrc_reg.h"
 #include "intel_ring.h"
 #include "intel_sseu.h"
 
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
index 52fa9c132746..e10d78601bbd 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -30,6 +30,10 @@ struct intel_context;
 struct intel_ring;
 
 struct intel_context_ops {
+	unsigned long flags;
+#define COPS_HAS_INFLIGHT_BIT 0
+#define COPS_HAS_INFLIGHT BIT(COPS_HAS_INFLIGHT_BIT)
+
 	int (*alloc)(struct intel_context *ce);
 
 	int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
@@ -58,8 +62,12 @@ struct intel_context {
 
 	struct intel_engine_cs *engine;
 	struct intel_engine_cs *inflight;
-#define intel_context_inflight(ce) ptr_mask_bits(READ_ONCE((ce)->inflight), 2)
-#define intel_context_inflight_count(ce) ptr_unmask_bits(READ_ONCE((ce)->inflight), 2)
+#define __intel_context_inflight(engine) ptr_mask_bits(engine, 3)
+#define __intel_context_inflight_count(engine) ptr_unmask_bits(engine, 3)
+#define intel_context_inflight(ce) \
+	__intel_context_inflight(READ_ONCE((ce)->inflight))
+#define intel_context_inflight_count(ce) \
+	__intel_context_inflight_count(READ_ONCE((ce)->inflight))
 
 	struct i915_address_space *vm;
 	struct i915_gem_context __rcu *gem_context;
@@ -81,12 +89,13 @@ struct intel_context {
 	unsigned long flags;
 #define CONTEXT_BARRIER_BIT		0
 #define CONTEXT_ALLOC_BIT		1
-#define CONTEXT_VALID_BIT		2
-#define CONTEXT_CLOSED_BIT		3
-#define CONTEXT_USE_SEMAPHORES		4
-#define CONTEXT_BANNED			5
-#define CONTEXT_FORCE_SINGLE_SUBMISSION	6
-#define CONTEXT_NOPREEMPT		7
+#define CONTEXT_INIT_BIT		2
+#define CONTEXT_VALID_BIT		3
+#define CONTEXT_CLOSED_BIT		4
+#define CONTEXT_USE_SEMAPHORES		5
+#define CONTEXT_BANNED			6
+#define CONTEXT_FORCE_SINGLE_SUBMISSION	7
+#define CONTEXT_NOPREEMPT		8
 
 	u32 *lrc_reg_state;
 	union {
diff --git a/drivers/gpu/drm/i915/gt/intel_engine.h b/drivers/gpu/drm/i915/gt/intel_engine.h
index 760fefdfe392..47ee8578e511 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine.h
+++ b/drivers/gpu/drm/i915/gt/intel_engine.h
@@ -15,7 +15,6 @@
 #include "i915_selftest.h"
 #include "gt/intel_timeline.h"
 #include "intel_engine_types.h"
-#include "intel_gpu_commands.h"
 #include "intel_workarounds.h"
 
 struct drm_printer;
@@ -223,91 +222,6 @@ void intel_engine_get_instdone(const struct intel_engine_cs *engine,
 
 void intel_engine_init_execlists(struct intel_engine_cs *engine);
 
-static inline u32 *__gen8_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u32 offset)
-{
-	memset(batch, 0, 6 * sizeof(u32));
-
-	batch[0] = GFX_OP_PIPE_CONTROL(6) | flags0;
-	batch[1] = flags1;
-	batch[2] = offset;
-
-	return batch + 6;
-}
-
-static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset)
-{
-	return __gen8_emit_pipe_control(batch, 0, flags, offset);
-}
-
-static inline u32 *gen12_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u32 offset)
-{
-	return __gen8_emit_pipe_control(batch, flags0, flags1, offset);
-}
-
-static inline u32 *
-__gen8_emit_write_rcs(u32 *cs, u32 value, u32 offset, u32 flags0, u32 flags1)
-{
-	*cs++ = GFX_OP_PIPE_CONTROL(6) | flags0;
-	*cs++ = flags1 | PIPE_CONTROL_QW_WRITE;
-	*cs++ = offset;
-	*cs++ = 0;
-	*cs++ = value;
-	*cs++ = 0; /* We're thrashing one extra dword. */
-
-	return cs;
-}
-
-static inline u32*
-gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
-{
-	/* We're using qword write, offset should be aligned to 8 bytes. */
-	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
-
-	return __gen8_emit_write_rcs(cs,
-				     value,
-				     gtt_offset,
-				     0,
-				     flags | PIPE_CONTROL_GLOBAL_GTT_IVB);
-}
-
-static inline u32*
-gen12_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags0, u32 flags1)
-{
-	/* We're using qword write, offset should be aligned to 8 bytes. */
-	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
-
-	return __gen8_emit_write_rcs(cs,
-				     value,
-				     gtt_offset,
-				     flags0,
-				     flags1 | PIPE_CONTROL_GLOBAL_GTT_IVB);
-}
-
-static inline u32 *
-__gen8_emit_flush_dw(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
-{
-	*cs++ = (MI_FLUSH_DW + 1) | flags;
-	*cs++ = gtt_offset;
-	*cs++ = 0;
-	*cs++ = value;
-
-	return cs;
-}
-
-static inline u32 *
-gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
-{
-	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
-	GEM_BUG_ON(gtt_offset & (1 << 5));
-	/* Offset should be aligned to 8 bytes for both (QW/DW) write types */
-	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
-
-	return __gen8_emit_flush_dw(cs,
-				    value,
-				    gtt_offset | MI_FLUSH_DW_USE_GTT,
-				    flags | MI_FLUSH_DW_OP_STOREDW);
-}
-
 static inline void __intel_engine_reset(struct intel_engine_cs *engine,
 					bool stalled)
 {
@@ -318,7 +232,12 @@ static inline void __intel_engine_reset(struct intel_engine_cs *engine,
 
 bool intel_engines_are_idle(struct intel_gt *gt);
 bool intel_engine_is_idle(struct intel_engine_cs *engine);
-void intel_engine_flush_submission(struct intel_engine_cs *engine);
+
+void __intel_engine_flush_submission(struct intel_engine_cs *engine, bool sync);
+static inline void intel_engine_flush_submission(struct intel_engine_cs *engine)
+{
+	__intel_engine_flush_submission(engine, true);
+}
 
 void intel_engines_reset_default_submission(struct intel_gt *gt);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
index 0b31670343f5..9cf555d6842b 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
@@ -33,12 +33,14 @@
 #include "intel_engine.h"
 #include "intel_engine_pm.h"
 #include "intel_engine_user.h"
+#include "intel_execlists_submission.h"
 #include "intel_gt.h"
 #include "intel_gt_requests.h"
 #include "intel_gt_pm.h"
-#include "intel_lrc.h"
+#include "intel_lrc_reg.h"
 #include "intel_reset.h"
 #include "intel_ring.h"
+#include "uc/intel_guc_submission.h"
 
 /* Haswell does have the CXT_SIZE register however it does not appear to be
  * valid. Now, docs explain in dwords what is in the context object. The full
@@ -340,7 +342,7 @@ static int intel_engine_setup(struct intel_gt *gt, enum intel_engine_id id)
 	engine->schedule = NULL;
 
 	ewma__engine_latency_init(&engine->latency);
-	seqlock_init(&engine->stats.lock);
+	seqcount_init(&engine->stats.lock);
 
 	ATOMIC_INIT_NOTIFIER_HEAD(&engine->context_status_notifier);
 
@@ -647,6 +649,8 @@ static int init_status_page(struct intel_engine_cs *engine)
 	void *vaddr;
 	int ret;
 
+	INIT_LIST_HEAD(&engine->status_page.timelines);
+
 	/*
 	 * Though the HWS register does support 36bit addresses, historically
 	 * we have had hangs and corruption reported due to wild writes if
@@ -709,9 +713,12 @@ static int engine_setup_common(struct intel_engine_cs *engine)
 		goto err_status;
 	}
 
+	err = intel_engine_init_cmd_parser(engine);
+	if (err)
+		goto err_cmd_parser;
+
 	intel_engine_init_active(engine, ENGINE_PHYSICAL);
 	intel_engine_init_execlists(engine);
-	intel_engine_init_cmd_parser(engine);
 	intel_engine_init__pm(engine);
 	intel_engine_init_retire(engine);
 
@@ -723,8 +730,13 @@ static int engine_setup_common(struct intel_engine_cs *engine)
 	intel_engine_init_whitelist(engine);
 	intel_engine_init_ctx_wa(engine);
 
+	if (INTEL_GEN(engine->i915) >= 12)
+		engine->flags |= I915_ENGINE_HAS_RELATIVE_MMIO;
+
 	return 0;
 
+err_cmd_parser:
+	intel_breadcrumbs_free(engine->breadcrumbs);
 err_status:
 	cleanup_status_page(engine);
 	return err;
@@ -829,6 +841,21 @@ create_pinned_context(struct intel_engine_cs *engine,
 	return ce;
 }
 
+static void destroy_pinned_context(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct i915_vma *hwsp = engine->status_page.vma;
+
+	GEM_BUG_ON(ce->timeline->hwsp_ggtt != hwsp);
+
+	mutex_lock(&hwsp->vm->mutex);
+	list_del(&ce->timeline->engine_link);
+	mutex_unlock(&hwsp->vm->mutex);
+
+	intel_context_unpin(ce);
+	intel_context_put(ce);
+}
+
 static struct intel_context *
 create_kernel_context(struct intel_engine_cs *engine)
 {
@@ -889,7 +916,9 @@ int intel_engines_init(struct intel_gt *gt)
 	enum intel_engine_id id;
 	int err;
 
-	if (HAS_EXECLISTS(gt->i915))
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		setup = intel_guc_submission_setup;
+	else if (HAS_EXECLISTS(gt->i915))
 		setup = intel_execlists_submission_setup;
 	else
 		setup = intel_ring_submission_setup;
@@ -925,7 +954,6 @@ void intel_engine_cleanup_common(struct intel_engine_cs *engine)
 	GEM_BUG_ON(!list_empty(&engine->active.requests));
 	tasklet_kill(&engine->execlists.tasklet); /* flush the callback */
 
-	cleanup_status_page(engine);
 	intel_breadcrumbs_free(engine->breadcrumbs);
 
 	intel_engine_fini_retire(engine);
@@ -934,11 +962,11 @@ void intel_engine_cleanup_common(struct intel_engine_cs *engine)
 	if (engine->default_state)
 		fput(engine->default_state);
 
-	if (engine->kernel_context) {
-		intel_context_unpin(engine->kernel_context);
-		intel_context_put(engine->kernel_context);
-	}
+	if (engine->kernel_context)
+		destroy_pinned_context(engine->kernel_context);
+
 	GEM_BUG_ON(!llist_empty(&engine->barrier_tasks));
+	cleanup_status_page(engine);
 
 	intel_wa_list_free(&engine->ctx_wa_list);
 	intel_wa_list_free(&engine->wa_list);
@@ -1002,32 +1030,50 @@ static unsigned long stop_timeout(const struct intel_engine_cs *engine)
 	return READ_ONCE(engine->props.stop_timeout_ms);
 }
 
-int intel_engine_stop_cs(struct intel_engine_cs *engine)
+static int __intel_engine_stop_cs(struct intel_engine_cs *engine,
+				  int fast_timeout_us,
+				  int slow_timeout_ms)
 {
 	struct intel_uncore *uncore = engine->uncore;
-	const u32 base = engine->mmio_base;
-	const i915_reg_t mode = RING_MI_MODE(base);
+	const i915_reg_t mode = RING_MI_MODE(engine->mmio_base);
 	int err;
 
+	intel_uncore_write_fw(uncore, mode, _MASKED_BIT_ENABLE(STOP_RING));
+	err = __intel_wait_for_register_fw(engine->uncore, mode,
+					   MODE_IDLE, MODE_IDLE,
+					   fast_timeout_us,
+					   slow_timeout_ms,
+					   NULL);
+
+	/* A final mmio read to let GPU writes be hopefully flushed to memory */
+	intel_uncore_posting_read_fw(uncore, mode);
+	return err;
+}
+
+int intel_engine_stop_cs(struct intel_engine_cs *engine)
+{
+	int err = 0;
+
 	if (INTEL_GEN(engine->i915) < 3)
 		return -ENODEV;
 
 	ENGINE_TRACE(engine, "\n");
+	if (__intel_engine_stop_cs(engine, 1000, stop_timeout(engine))) {
+		ENGINE_TRACE(engine,
+			     "timed out on STOP_RING -> IDLE; HEAD:%04x, TAIL:%04x\n",
+			     ENGINE_READ_FW(engine, RING_HEAD) & HEAD_ADDR,
+			     ENGINE_READ_FW(engine, RING_TAIL) & TAIL_ADDR);
 
-	intel_uncore_write_fw(uncore, mode, _MASKED_BIT_ENABLE(STOP_RING));
-
-	err = 0;
-	if (__intel_wait_for_register_fw(uncore,
-					 mode, MODE_IDLE, MODE_IDLE,
-					 1000, stop_timeout(engine),
-					 NULL)) {
-		ENGINE_TRACE(engine, "timed out on STOP_RING -> IDLE\n");
-		err = -ETIMEDOUT;
+		/*
+		 * Sometimes we observe that the idle flag is not
+		 * set even though the ring is empty. So double
+		 * check before giving up.
+		 */
+		if ((ENGINE_READ_FW(engine, RING_HEAD) & HEAD_ADDR) !=
+		    (ENGINE_READ_FW(engine, RING_TAIL) & TAIL_ADDR))
+			err = -ETIMEDOUT;
 	}
 
-	/* A final mmio read to let GPU writes be hopefully flushed to memory */
-	intel_uncore_posting_read_fw(uncore, mode);
-
 	return err;
 }
 
@@ -1189,17 +1235,13 @@ static bool ring_is_idle(struct intel_engine_cs *engine)
 	return idle;
 }
 
-void intel_engine_flush_submission(struct intel_engine_cs *engine)
+void __intel_engine_flush_submission(struct intel_engine_cs *engine, bool sync)
 {
 	struct tasklet_struct *t = &engine->execlists.tasklet;
 
 	if (!t->func)
 		return;
 
-	/* Synchronise and wait for the tasklet on another CPU */
-	tasklet_kill(t);
-
-	/* Having cancelled the tasklet, ensure that is run */
 	local_bh_disable();
 	if (tasklet_trylock(t)) {
 		/* Must wait for any GPU reset in progress. */
@@ -1208,6 +1250,10 @@ void intel_engine_flush_submission(struct intel_engine_cs *engine)
 		tasklet_unlock(t);
 	}
 	local_bh_enable();
+
+	/* Synchronise and wait for the tasklet on another CPU */
+	if (sync)
+		tasklet_unlock_wait(t);
 }
 
 /**
@@ -1273,8 +1319,12 @@ void intel_engines_reset_default_submission(struct intel_gt *gt)
 	struct intel_engine_cs *engine;
 	enum intel_engine_id id;
 
-	for_each_engine(engine, gt, id)
+	for_each_engine(engine, gt, id) {
+		if (engine->sanitize)
+			engine->sanitize(engine);
+
 		engine->set_default_submission(engine);
+	}
 }
 
 bool intel_engine_can_store_dword(struct intel_engine_cs *engine)
@@ -1294,44 +1344,6 @@ bool intel_engine_can_store_dword(struct intel_engine_cs *engine)
 	}
 }
 
-static int print_sched_attr(const struct i915_sched_attr *attr,
-			    char *buf, int x, int len)
-{
-	if (attr->priority == I915_PRIORITY_INVALID)
-		return x;
-
-	x += snprintf(buf + x, len - x,
-		      " prio=%d", attr->priority);
-
-	return x;
-}
-
-static void print_request(struct drm_printer *m,
-			  struct i915_request *rq,
-			  const char *prefix)
-{
-	const char *name = rq->fence.ops->get_timeline_name(&rq->fence);
-	char buf[80] = "";
-	int x = 0;
-
-	x = print_sched_attr(&rq->sched.attr, buf, x, sizeof(buf));
-
-	drm_printf(m, "%s %llx:%llx%s%s %s @ %dms: %s\n",
-		   prefix,
-		   rq->fence.context, rq->fence.seqno,
-		   i915_request_completed(rq) ? "!" :
-		   i915_request_started(rq) ? "*" :
-		   "",
-		   test_bit(DMA_FENCE_FLAG_SIGNALED_BIT,
-			    &rq->fence.flags) ? "+" :
-		   test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT,
-			    &rq->fence.flags) ? "-" :
-		   "",
-		   buf,
-		   jiffies_to_msecs(jiffies - rq->emitted_jiffies),
-		   name);
-}
-
 static struct intel_timeline *get_timeline(struct i915_request *rq)
 {
 	struct intel_timeline *tl;
@@ -1480,7 +1492,9 @@ static void intel_engine_print_registers(struct intel_engine_cs *engine,
 		drm_printf(m, "\tIPEHR: 0x%08x\n", ENGINE_READ(engine, IPEHR));
 	}
 
-	if (HAS_EXECLISTS(dev_priv)) {
+	if (intel_engine_in_guc_submission_mode(engine)) {
+		/* nothing to print yet */
+	} else if (HAS_EXECLISTS(dev_priv)) {
 		struct i915_request * const *port, *rq;
 		const u32 *hws =
 			&engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX];
@@ -1529,7 +1543,7 @@ static void intel_engine_print_registers(struct intel_engine_cs *engine,
 					intel_context_is_banned(rq->context) ? "*" : "");
 			len += print_ring(hdr + len, sizeof(hdr) - len, rq);
 			scnprintf(hdr + len, sizeof(hdr) - len, "rq: ");
-			print_request(m, rq, hdr);
+			i915_request_show(m, rq, hdr, 0);
 		}
 		for (port = execlists->pending; (rq = *port); port++) {
 			char hdr[160];
@@ -1543,7 +1557,7 @@ static void intel_engine_print_registers(struct intel_engine_cs *engine,
 					intel_context_is_banned(rq->context) ? "*" : "");
 			len += print_ring(hdr + len, sizeof(hdr) - len, rq);
 			scnprintf(hdr + len, sizeof(hdr) - len, "rq: ");
-			print_request(m, rq, hdr);
+			i915_request_show(m, rq, hdr, 0);
 		}
 		rcu_read_unlock();
 		execlists_active_unlock_bh(execlists);
@@ -1667,7 +1681,7 @@ void intel_engine_dump(struct intel_engine_cs *engine,
 			   ktime_to_ms(intel_engine_get_busy_time(engine,
 								  &dummy)));
 	drm_printf(m, "\tForcewake: %x domains, %d active\n",
-		   engine->fw_domain, atomic_read(&engine->fw_active));
+		   engine->fw_domain, READ_ONCE(engine->fw_active));
 
 	rcu_read_lock();
 	rq = READ_ONCE(engine->heartbeat.systole);
@@ -1687,7 +1701,7 @@ void intel_engine_dump(struct intel_engine_cs *engine,
 	if (rq) {
 		struct intel_timeline *tl = get_timeline(rq);
 
-		print_request(m, rq, "\t\tactive ");
+		i915_request_show(m, rq, "\t\tactive ", 0);
 
 		drm_printf(m, "\t\tring->start:  0x%08x\n",
 			   i915_ggtt_offset(rq->ring->vma));
@@ -1725,7 +1739,7 @@ void intel_engine_dump(struct intel_engine_cs *engine,
 		drm_printf(m, "\tDevice is asleep; skipping register dump\n");
 	}
 
-	intel_execlists_show_requests(engine, m, print_request, 8);
+	intel_execlists_show_requests(engine, m, i915_request_show, 8);
 
 	drm_printf(m, "HWSP:\n");
 	hexdump(m, engine->status_page.addr, PAGE_SIZE);
@@ -1745,7 +1759,7 @@ static ktime_t __intel_engine_get_busy_time(struct intel_engine_cs *engine,
 	 * add it to the total.
 	 */
 	*now = ktime_get();
-	if (atomic_read(&engine->stats.active))
+	if (READ_ONCE(engine->stats.active))
 		total = ktime_add(total, ktime_sub(*now, engine->stats.start));
 
 	return total;
@@ -1764,9 +1778,9 @@ ktime_t intel_engine_get_busy_time(struct intel_engine_cs *engine, ktime_t *now)
 	ktime_t total;
 
 	do {
-		seq = read_seqbegin(&engine->stats.lock);
+		seq = read_seqcount_begin(&engine->stats.lock);
 		total = __intel_engine_get_busy_time(engine, now);
-	} while (read_seqretry(&engine->stats.lock, seq));
+	} while (read_seqcount_retry(&engine->stats.lock, seq));
 
 	return total;
 }
@@ -1802,7 +1816,7 @@ intel_engine_find_active_request(struct intel_engine_cs *engine)
 		struct intel_timeline *tl = request->context->timeline;
 
 		list_for_each_entry_from_reverse(request, &tl->requests, link) {
-			if (i915_request_completed(request))
+			if (__i915_request_is_complete(request))
 				break;
 
 			active = request;
@@ -1813,10 +1827,10 @@ intel_engine_find_active_request(struct intel_engine_cs *engine)
 		return active;
 
 	list_for_each_entry(request, &engine->active.requests, sched.link) {
-		if (i915_request_completed(request))
+		if (__i915_request_is_complete(request))
 			continue;
 
-		if (!i915_request_started(request))
+		if (!__i915_request_has_started(request))
 			continue;
 
 		/* More than one preemptible request may match! */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
index 9060385cd69e..d7be2b9339f9 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
@@ -37,6 +37,18 @@ static bool next_heartbeat(struct intel_engine_cs *engine)
 	return true;
 }
 
+static struct i915_request *
+heartbeat_create(struct intel_context *ce, gfp_t gfp)
+{
+	struct i915_request *rq;
+
+	intel_context_enter(ce);
+	rq = __i915_request_create(ce, gfp);
+	intel_context_exit(ce);
+
+	return rq;
+}
+
 static void idle_pulse(struct intel_engine_cs *engine, struct i915_request *rq)
 {
 	engine->wakeref_serial = READ_ONCE(engine->serial) + 1;
@@ -45,6 +57,15 @@ static void idle_pulse(struct intel_engine_cs *engine, struct i915_request *rq)
 		engine->heartbeat.systole = i915_request_get(rq);
 }
 
+static void heartbeat_commit(struct i915_request *rq,
+			     const struct i915_sched_attr *attr)
+{
+	idle_pulse(rq->engine, rq);
+
+	__i915_request_commit(rq);
+	__i915_request_queue(rq, attr);
+}
+
 static void show_heartbeat(const struct i915_request *rq,
 			   struct intel_engine_cs *engine)
 {
@@ -139,16 +160,11 @@ static void heartbeat(struct work_struct *wrk)
 		goto out;
 	}
 
-	intel_context_enter(ce);
-	rq = __i915_request_create(ce, GFP_NOWAIT | __GFP_NOWARN);
-	intel_context_exit(ce);
+	rq = heartbeat_create(ce, GFP_NOWAIT | __GFP_NOWARN);
 	if (IS_ERR(rq))
 		goto unlock;
 
-	idle_pulse(engine, rq);
-
-	__i915_request_commit(rq);
-	__i915_request_queue(rq, &attr);
+	heartbeat_commit(rq, &attr);
 
 unlock:
 	mutex_unlock(&ce->timeline->mutex);
@@ -187,17 +203,13 @@ static int __intel_engine_pulse(struct intel_engine_cs *engine)
 	GEM_BUG_ON(!intel_engine_has_preemption(engine));
 	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
 
-	intel_context_enter(ce);
-	rq = __i915_request_create(ce, GFP_NOWAIT | __GFP_NOWARN);
-	intel_context_exit(ce);
+	rq = heartbeat_create(ce, GFP_NOWAIT | __GFP_NOWARN);
 	if (IS_ERR(rq))
 		return PTR_ERR(rq);
 
 	__set_bit(I915_FENCE_FLAG_SENTINEL, &rq->fence.flags);
-	idle_pulse(engine, rq);
 
-	__i915_request_commit(rq);
-	__i915_request_queue(rq, &attr);
+	heartbeat_commit(rq, &attr);
 	GEM_BUG_ON(rq->sched.attr.priority < I915_PRIORITY_BARRIER);
 
 	return 0;
@@ -273,8 +285,12 @@ int intel_engine_pulse(struct intel_engine_cs *engine)
 
 int intel_engine_flush_barriers(struct intel_engine_cs *engine)
 {
+	struct i915_sched_attr attr = {
+		.priority = I915_USER_PRIORITY(I915_PRIORITY_MIN),
+	};
+	struct intel_context *ce = engine->kernel_context;
 	struct i915_request *rq;
-	int err = 0;
+	int err;
 
 	if (llist_empty(&engine->barrier_tasks))
 		return 0;
@@ -282,15 +298,22 @@ int intel_engine_flush_barriers(struct intel_engine_cs *engine)
 	if (!intel_engine_pm_get_if_awake(engine))
 		return 0;
 
-	rq = i915_request_create(engine->kernel_context);
+	if (mutex_lock_interruptible(&ce->timeline->mutex)) {
+		err = -EINTR;
+		goto out_rpm;
+	}
+
+	rq = heartbeat_create(ce, GFP_KERNEL);
 	if (IS_ERR(rq)) {
 		err = PTR_ERR(rq);
-		goto out_rpm;
+		goto out_unlock;
 	}
 
-	idle_pulse(engine, rq);
-	i915_request_add(rq);
+	heartbeat_commit(rq, &attr);
 
+	err = 0;
+out_unlock:
+	mutex_unlock(&ce->timeline->mutex);
 out_rpm:
 	intel_engine_pm_put(engine);
 	return err;
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_pm.c b/drivers/gpu/drm/i915/gt/intel_engine_pm.c
index 499b09cb4acf..e67d09259dd0 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_pm.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_pm.c
@@ -60,18 +60,26 @@ static int __engine_unpark(struct intel_wakeref *wf)
 
 		/* Scrub the context image after our loss of control */
 		ce->ops->reset(ce);
+
+		CE_TRACE(ce, "reset { seqno:%x, *hwsp:%x, ring:%x }\n",
+			 ce->timeline->seqno,
+			 READ_ONCE(*ce->timeline->hwsp_seqno),
+			 ce->ring->emit);
+		GEM_BUG_ON(ce->timeline->seqno !=
+			   READ_ONCE(*ce->timeline->hwsp_seqno));
 	}
 
 	if (engine->unpark)
 		engine->unpark(engine);
 
+	intel_breadcrumbs_unpark(engine->breadcrumbs);
 	intel_engine_unpark_heartbeat(engine);
 	return 0;
 }
 
 #if IS_ENABLED(CONFIG_LOCKDEP)
 
-static inline unsigned long __timeline_mark_lock(struct intel_context *ce)
+static unsigned long __timeline_mark_lock(struct intel_context *ce)
 {
 	unsigned long flags;
 
@@ -81,8 +89,8 @@ static inline unsigned long __timeline_mark_lock(struct intel_context *ce)
 	return flags;
 }
 
-static inline void __timeline_mark_unlock(struct intel_context *ce,
-					  unsigned long flags)
+static void __timeline_mark_unlock(struct intel_context *ce,
+				   unsigned long flags)
 {
 	mutex_release(&ce->timeline->mutex.dep_map, _THIS_IP_);
 	local_irq_restore(flags);
@@ -90,13 +98,13 @@ static inline void __timeline_mark_unlock(struct intel_context *ce,
 
 #else
 
-static inline unsigned long __timeline_mark_lock(struct intel_context *ce)
+static unsigned long __timeline_mark_lock(struct intel_context *ce)
 {
 	return 0;
 }
 
-static inline void __timeline_mark_unlock(struct intel_context *ce,
-					  unsigned long flags)
+static void __timeline_mark_unlock(struct intel_context *ce,
+				   unsigned long flags)
 {
 }
 
@@ -136,7 +144,7 @@ __queue_and_release_pm(struct i915_request *rq,
 		list_add_tail(&tl->link, &timelines->active_list);
 
 	/* Hand the request over to HW and so engine_retire() */
-	__i915_request_queue(rq, NULL);
+	__i915_request_queue_bh(rq);
 
 	/* Let new submissions commence (and maybe retire this timeline) */
 	__intel_wakeref_defer_park(&engine->wakeref);
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_stats.h b/drivers/gpu/drm/i915/gt/intel_engine_stats.h
new file mode 100644
index 000000000000..24fbdd94351a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_stats.h
@@ -0,0 +1,60 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_ENGINE_STATS_H__
+#define __INTEL_ENGINE_STATS_H__
+
+#include <linux/atomic.h>
+#include <linux/ktime.h>
+#include <linux/seqlock.h>
+
+#include "i915_gem.h" /* GEM_BUG_ON */
+#include "intel_engine.h"
+
+static inline void intel_engine_context_in(struct intel_engine_cs *engine)
+{
+	unsigned long flags;
+
+	if (engine->stats.active) {
+		engine->stats.active++;
+		return;
+	}
+
+	/* The writer is serialised; but the pmu reader may be from hardirq */
+	local_irq_save(flags);
+	write_seqcount_begin(&engine->stats.lock);
+
+	engine->stats.start = ktime_get();
+	engine->stats.active++;
+
+	write_seqcount_end(&engine->stats.lock);
+	local_irq_restore(flags);
+
+	GEM_BUG_ON(!engine->stats.active);
+}
+
+static inline void intel_engine_context_out(struct intel_engine_cs *engine)
+{
+	unsigned long flags;
+
+	GEM_BUG_ON(!engine->stats.active);
+	if (engine->stats.active > 1) {
+		engine->stats.active--;
+		return;
+	}
+
+	local_irq_save(flags);
+	write_seqcount_begin(&engine->stats.lock);
+
+	engine->stats.active--;
+	engine->stats.total =
+		ktime_add(engine->stats.total,
+			  ktime_sub(ktime_get(), engine->stats.start));
+
+	write_seqcount_end(&engine->stats.lock);
+	local_irq_restore(flags);
+}
+
+#endif /* __INTEL_ENGINE_STATS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
index ee6312601c56..d2346b425547 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
@@ -68,6 +68,7 @@ typedef u8 intel_engine_mask_t;
 #define ALL_ENGINES ((intel_engine_mask_t)~0ul)
 
 struct intel_hw_status_page {
+	struct list_head timelines;
 	struct i915_vma *vma;
 	u32 *addr;
 };
@@ -183,7 +184,8 @@ struct intel_engine_execlists {
 	 * Reserve the upper 16b for tracking internal errors.
 	 */
 	u32 error_interrupt;
-#define ERROR_CSB BIT(31)
+#define ERROR_CSB	BIT(31)
+#define ERROR_PREEMPT	BIT(30)
 
 	/**
 	 * @reset_ccid: Active CCID [EXECLISTS_STATUS_HI] at the time of reset
@@ -237,16 +239,6 @@ struct intel_engine_execlists {
 	unsigned int port_mask;
 
 	/**
-	 * @switch_priority_hint: Second context priority.
-	 *
-	 * We submit multiple contexts to the HW simultaneously and would
-	 * like to occasionally switch between them to emulate timeslicing.
-	 * To know when timeslicing is suitable, we track the priority of
-	 * the context submitted second.
-	 */
-	int switch_priority_hint;
-
-	/**
 	 * @queue_priority_hint: Highest pending priority.
 	 *
 	 * When we add requests into the queue, or adjust the priority of
@@ -327,7 +319,7 @@ struct intel_engine_cs {
 	 * as possible.
 	 */
 	enum forcewake_domains fw_domain;
-	atomic_t fw_active;
+	unsigned int fw_active;
 
 	unsigned long context_tag;
 
@@ -524,12 +516,12 @@ struct intel_engine_cs {
 		/**
 		 * @active: Number of contexts currently scheduled in.
 		 */
-		atomic_t active;
+		unsigned int active;
 
 		/**
 		 * @lock: Lock protecting the below fields.
 		 */
-		seqlock_t lock;
+		seqcount_t lock;
 
 		/**
 		 * @total: Total time this engine was busy.
@@ -559,6 +551,8 @@ struct intel_engine_cs {
 		unsigned long stop_timeout_ms;
 		unsigned long timeslice_duration_ms;
 	} props, defaults;
+
+	I915_SELFTEST_DECLARE(struct fault_attr reset_timeout);
 };
 
 static inline bool
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
new file mode 100644
index 000000000000..ac1be7a632d3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -0,0 +1,3896 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+/**
+ * DOC: Logical Rings, Logical Ring Contexts and Execlists
+ *
+ * Motivation:
+ * GEN8 brings an expansion of the HW contexts: "Logical Ring Contexts".
+ * These expanded contexts enable a number of new abilities, especially
+ * "Execlists" (also implemented in this file).
+ *
+ * One of the main differences with the legacy HW contexts is that logical
+ * ring contexts incorporate many more things to the context's state, like
+ * PDPs or ringbuffer control registers:
+ *
+ * The reason why PDPs are included in the context is straightforward: as
+ * PPGTTs (per-process GTTs) are actually per-context, having the PDPs
+ * contained there mean you don't need to do a ppgtt->switch_mm yourself,
+ * instead, the GPU will do it for you on the context switch.
+ *
+ * But, what about the ringbuffer control registers (head, tail, etc..)?
+ * shouldn't we just need a set of those per engine command streamer? This is
+ * where the name "Logical Rings" starts to make sense: by virtualizing the
+ * rings, the engine cs shifts to a new "ring buffer" with every context
+ * switch. When you want to submit a workload to the GPU you: A) choose your
+ * context, B) find its appropriate virtualized ring, C) write commands to it
+ * and then, finally, D) tell the GPU to switch to that context.
+ *
+ * Instead of the legacy MI_SET_CONTEXT, the way you tell the GPU to switch
+ * to a contexts is via a context execution list, ergo "Execlists".
+ *
+ * LRC implementation:
+ * Regarding the creation of contexts, we have:
+ *
+ * - One global default context.
+ * - One local default context for each opened fd.
+ * - One local extra context for each context create ioctl call.
+ *
+ * Now that ringbuffers belong per-context (and not per-engine, like before)
+ * and that contexts are uniquely tied to a given engine (and not reusable,
+ * like before) we need:
+ *
+ * - One ringbuffer per-engine inside each context.
+ * - One backing object per-engine inside each context.
+ *
+ * The global default context starts its life with these new objects fully
+ * allocated and populated. The local default context for each opened fd is
+ * more complex, because we don't know at creation time which engine is going
+ * to use them. To handle this, we have implemented a deferred creation of LR
+ * contexts:
+ *
+ * The local context starts its life as a hollow or blank holder, that only
+ * gets populated for a given engine once we receive an execbuffer. If later
+ * on we receive another execbuffer ioctl for the same context but a different
+ * engine, we allocate/populate a new ringbuffer and context backing object and
+ * so on.
+ *
+ * Finally, regarding local contexts created using the ioctl call: as they are
+ * only allowed with the render ring, we can allocate & populate them right
+ * away (no need to defer anything, at least for now).
+ *
+ * Execlists implementation:
+ * Execlists are the new method by which, on gen8+ hardware, workloads are
+ * submitted for execution (as opposed to the legacy, ringbuffer-based, method).
+ * This method works as follows:
+ *
+ * When a request is committed, its commands (the BB start and any leading or
+ * trailing commands, like the seqno breadcrumbs) are placed in the ringbuffer
+ * for the appropriate context. The tail pointer in the hardware context is not
+ * updated at this time, but instead, kept by the driver in the ringbuffer
+ * structure. A structure representing this request is added to a request queue
+ * for the appropriate engine: this structure contains a copy of the context's
+ * tail after the request was written to the ring buffer and a pointer to the
+ * context itself.
+ *
+ * If the engine's request queue was empty before the request was added, the
+ * queue is processed immediately. Otherwise the queue will be processed during
+ * a context switch interrupt. In any case, elements on the queue will get sent
+ * (in pairs) to the GPU's ExecLists Submit Port (ELSP, for short) with a
+ * globally unique 20-bits submission ID.
+ *
+ * When execution of a request completes, the GPU updates the context status
+ * buffer with a context complete event and generates a context switch interrupt.
+ * During the interrupt handling, the driver examines the events in the buffer:
+ * for each context complete event, if the announced ID matches that on the head
+ * of the request queue, then that request is retired and removed from the queue.
+ *
+ * After processing, if any requests were retired and the queue is not empty
+ * then a new execution list can be submitted. The two requests at the front of
+ * the queue are next to be submitted but since a context may not occur twice in
+ * an execution list, if subsequent requests have the same ID as the first then
+ * the two requests must be combined. This is done simply by discarding requests
+ * at the head of the queue until either only one requests is left (in which case
+ * we use a NULL second context) or the first two requests have unique IDs.
+ *
+ * By always executing the first two requests in the queue the driver ensures
+ * that the GPU is kept as busy as possible. In the case where a single context
+ * completes but a second context is still executing, the request for this second
+ * context will be at the head of the queue when we remove the first one. This
+ * request will then be resubmitted along with a new request for a different context,
+ * which will cause the hardware to continue executing the second request and queue
+ * the new request (the GPU detects the condition of a context getting preempted
+ * with the same context and optimizes the context switch flow by not doing
+ * preemption, but just sampling the new tail pointer).
+ *
+ */
+#include <linux/interrupt.h>
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "i915_vgpu.h"
+#include "gen8_engine_cs.h"
+#include "intel_breadcrumbs.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_stats.h"
+#include "intel_execlists_submission.h"
+#include "intel_gt.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_requests.h"
+#include "intel_lrc.h"
+#include "intel_lrc_reg.h"
+#include "intel_mocs.h"
+#include "intel_reset.h"
+#include "intel_ring.h"
+#include "intel_workarounds.h"
+#include "shmem_utils.h"
+
+#define RING_EXECLIST_QFULL		(1 << 0x2)
+#define RING_EXECLIST1_VALID		(1 << 0x3)
+#define RING_EXECLIST0_VALID		(1 << 0x4)
+#define RING_EXECLIST_ACTIVE_STATUS	(3 << 0xE)
+#define RING_EXECLIST1_ACTIVE		(1 << 0x11)
+#define RING_EXECLIST0_ACTIVE		(1 << 0x12)
+
+#define GEN8_CTX_STATUS_IDLE_ACTIVE	(1 << 0)
+#define GEN8_CTX_STATUS_PREEMPTED	(1 << 1)
+#define GEN8_CTX_STATUS_ELEMENT_SWITCH	(1 << 2)
+#define GEN8_CTX_STATUS_ACTIVE_IDLE	(1 << 3)
+#define GEN8_CTX_STATUS_COMPLETE	(1 << 4)
+#define GEN8_CTX_STATUS_LITE_RESTORE	(1 << 15)
+
+#define GEN8_CTX_STATUS_COMPLETED_MASK \
+	 (GEN8_CTX_STATUS_COMPLETE | GEN8_CTX_STATUS_PREEMPTED)
+
+#define GEN12_CTX_STATUS_SWITCHED_TO_NEW_QUEUE	(0x1) /* lower csb dword */
+#define GEN12_CTX_SWITCH_DETAIL(csb_dw)	((csb_dw) & 0xF) /* upper csb dword */
+#define GEN12_CSB_SW_CTX_ID_MASK		GENMASK(25, 15)
+#define GEN12_IDLE_CTX_ID		0x7FF
+#define GEN12_CSB_CTX_VALID(csb_dw) \
+	(FIELD_GET(GEN12_CSB_SW_CTX_ID_MASK, csb_dw) != GEN12_IDLE_CTX_ID)
+
+/* Typical size of the average request (2 pipecontrols and a MI_BB) */
+#define EXECLISTS_REQUEST_SIZE 64 /* bytes */
+
+struct virtual_engine {
+	struct intel_engine_cs base;
+	struct intel_context context;
+	struct rcu_work rcu;
+
+	/*
+	 * We allow only a single request through the virtual engine at a time
+	 * (each request in the timeline waits for the completion fence of
+	 * the previous before being submitted). By restricting ourselves to
+	 * only submitting a single request, each request is placed on to a
+	 * physical to maximise load spreading (by virtue of the late greedy
+	 * scheduling -- each real engine takes the next available request
+	 * upon idling).
+	 */
+	struct i915_request *request;
+
+	/*
+	 * We keep a rbtree of available virtual engines inside each physical
+	 * engine, sorted by priority. Here we preallocate the nodes we need
+	 * for the virtual engine, indexed by physical_engine->id.
+	 */
+	struct ve_node {
+		struct rb_node rb;
+		int prio;
+	} nodes[I915_NUM_ENGINES];
+
+	/*
+	 * Keep track of bonded pairs -- restrictions upon on our selection
+	 * of physical engines any particular request may be submitted to.
+	 * If we receive a submit-fence from a master engine, we will only
+	 * use one of sibling_mask physical engines.
+	 */
+	struct ve_bond {
+		const struct intel_engine_cs *master;
+		intel_engine_mask_t sibling_mask;
+	} *bonds;
+	unsigned int num_bonds;
+
+	/* And finally, which physical engines this virtual engine maps onto. */
+	unsigned int num_siblings;
+	struct intel_engine_cs *siblings[];
+};
+
+static struct virtual_engine *to_virtual_engine(struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(!intel_engine_is_virtual(engine));
+	return container_of(engine, struct virtual_engine, base);
+}
+
+static struct i915_request *
+__active_request(const struct intel_timeline * const tl,
+		 struct i915_request *rq,
+		 int error)
+{
+	struct i915_request *active = rq;
+
+	list_for_each_entry_from_reverse(rq, &tl->requests, link) {
+		if (__i915_request_is_complete(rq))
+			break;
+
+		if (error) {
+			i915_request_set_error_once(rq, error);
+			__i915_request_skip(rq);
+		}
+		active = rq;
+	}
+
+	return active;
+}
+
+static struct i915_request *
+active_request(const struct intel_timeline * const tl, struct i915_request *rq)
+{
+	return __active_request(tl, rq, 0);
+}
+
+static void ring_set_paused(const struct intel_engine_cs *engine, int state)
+{
+	/*
+	 * We inspect HWS_PREEMPT with a semaphore inside
+	 * engine->emit_fini_breadcrumb. If the dword is true,
+	 * the ring is paused as the semaphore will busywait
+	 * until the dword is false.
+	 */
+	engine->status_page.addr[I915_GEM_HWS_PREEMPT] = state;
+	if (state)
+		wmb();
+}
+
+static struct i915_priolist *to_priolist(struct rb_node *rb)
+{
+	return rb_entry(rb, struct i915_priolist, node);
+}
+
+static int rq_prio(const struct i915_request *rq)
+{
+	return READ_ONCE(rq->sched.attr.priority);
+}
+
+static int effective_prio(const struct i915_request *rq)
+{
+	int prio = rq_prio(rq);
+
+	/*
+	 * If this request is special and must not be interrupted at any
+	 * cost, so be it. Note we are only checking the most recent request
+	 * in the context and so may be masking an earlier vip request. It
+	 * is hoped that under the conditions where nopreempt is used, this
+	 * will not matter (i.e. all requests to that context will be
+	 * nopreempt for as long as desired).
+	 */
+	if (i915_request_has_nopreempt(rq))
+		prio = I915_PRIORITY_UNPREEMPTABLE;
+
+	return prio;
+}
+
+static int queue_prio(const struct intel_engine_execlists *execlists)
+{
+	struct i915_priolist *p;
+	struct rb_node *rb;
+
+	rb = rb_first_cached(&execlists->queue);
+	if (!rb)
+		return INT_MIN;
+
+	/*
+	 * As the priolist[] are inverted, with the highest priority in [0],
+	 * we have to flip the index value to become priority.
+	 */
+	p = to_priolist(rb);
+	if (!I915_USER_PRIORITY_SHIFT)
+		return p->priority;
+
+	return ((p->priority + 1) << I915_USER_PRIORITY_SHIFT) - ffs(p->used);
+}
+
+static int virtual_prio(const struct intel_engine_execlists *el)
+{
+	struct rb_node *rb = rb_first_cached(&el->virtual);
+
+	return rb ? rb_entry(rb, struct ve_node, rb)->prio : INT_MIN;
+}
+
+static bool need_preempt(const struct intel_engine_cs *engine,
+			 const struct i915_request *rq)
+{
+	int last_prio;
+
+	if (!intel_engine_has_semaphores(engine))
+		return false;
+
+	/*
+	 * Check if the current priority hint merits a preemption attempt.
+	 *
+	 * We record the highest value priority we saw during rescheduling
+	 * prior to this dequeue, therefore we know that if it is strictly
+	 * less than the current tail of ESLP[0], we do not need to force
+	 * a preempt-to-idle cycle.
+	 *
+	 * However, the priority hint is a mere hint that we may need to
+	 * preempt. If that hint is stale or we may be trying to preempt
+	 * ourselves, ignore the request.
+	 *
+	 * More naturally we would write
+	 *      prio >= max(0, last);
+	 * except that we wish to prevent triggering preemption at the same
+	 * priority level: the task that is running should remain running
+	 * to preserve FIFO ordering of dependencies.
+	 */
+	last_prio = max(effective_prio(rq), I915_PRIORITY_NORMAL - 1);
+	if (engine->execlists.queue_priority_hint <= last_prio)
+		return false;
+
+	/*
+	 * Check against the first request in ELSP[1], it will, thanks to the
+	 * power of PI, be the highest priority of that context.
+	 */
+	if (!list_is_last(&rq->sched.link, &engine->active.requests) &&
+	    rq_prio(list_next_entry(rq, sched.link)) > last_prio)
+		return true;
+
+	/*
+	 * If the inflight context did not trigger the preemption, then maybe
+	 * it was the set of queued requests? Pick the highest priority in
+	 * the queue (the first active priolist) and see if it deserves to be
+	 * running instead of ELSP[0].
+	 *
+	 * The highest priority request in the queue can not be either
+	 * ELSP[0] or ELSP[1] as, thanks again to PI, if it was the same
+	 * context, it's priority would not exceed ELSP[0] aka last_prio.
+	 */
+	return max(virtual_prio(&engine->execlists),
+		   queue_prio(&engine->execlists)) > last_prio;
+}
+
+__maybe_unused static bool
+assert_priority_queue(const struct i915_request *prev,
+		      const struct i915_request *next)
+{
+	/*
+	 * Without preemption, the prev may refer to the still active element
+	 * which we refuse to let go.
+	 *
+	 * Even with preemption, there are times when we think it is better not
+	 * to preempt and leave an ostensibly lower priority request in flight.
+	 */
+	if (i915_request_is_active(prev))
+		return true;
+
+	return rq_prio(prev) >= rq_prio(next);
+}
+
+static struct i915_request *
+__unwind_incomplete_requests(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq, *rn, *active = NULL;
+	struct list_head *pl;
+	int prio = I915_PRIORITY_INVALID;
+
+	lockdep_assert_held(&engine->active.lock);
+
+	list_for_each_entry_safe_reverse(rq, rn,
+					 &engine->active.requests,
+					 sched.link) {
+		if (__i915_request_is_complete(rq)) {
+			list_del_init(&rq->sched.link);
+			continue;
+		}
+
+		__i915_request_unsubmit(rq);
+
+		GEM_BUG_ON(rq_prio(rq) == I915_PRIORITY_INVALID);
+		if (rq_prio(rq) != prio) {
+			prio = rq_prio(rq);
+			pl = i915_sched_lookup_priolist(engine, prio);
+		}
+		GEM_BUG_ON(RB_EMPTY_ROOT(&engine->execlists.queue.rb_root));
+
+		list_move(&rq->sched.link, pl);
+		set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+
+		/* Check in case we rollback so far we wrap [size/2] */
+		if (intel_ring_direction(rq->ring,
+					 rq->tail,
+					 rq->ring->tail + 8) > 0)
+			rq->context->lrc.desc |= CTX_DESC_FORCE_RESTORE;
+
+		active = rq;
+	}
+
+	return active;
+}
+
+struct i915_request *
+execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists)
+{
+	struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+
+	return __unwind_incomplete_requests(engine);
+}
+
+static void
+execlists_context_status_change(struct i915_request *rq, unsigned long status)
+{
+	/*
+	 * Only used when GVT-g is enabled now. When GVT-g is disabled,
+	 * The compiler should eliminate this function as dead-code.
+	 */
+	if (!IS_ENABLED(CONFIG_DRM_I915_GVT))
+		return;
+
+	atomic_notifier_call_chain(&rq->engine->context_status_notifier,
+				   status, rq);
+}
+
+static void reset_active(struct i915_request *rq,
+			 struct intel_engine_cs *engine)
+{
+	struct intel_context * const ce = rq->context;
+	u32 head;
+
+	/*
+	 * The executing context has been cancelled. We want to prevent
+	 * further execution along this context and propagate the error on
+	 * to anything depending on its results.
+	 *
+	 * In __i915_request_submit(), we apply the -EIO and remove the
+	 * requests' payloads for any banned requests. But first, we must
+	 * rewind the context back to the start of the incomplete request so
+	 * that we do not jump back into the middle of the batch.
+	 *
+	 * We preserve the breadcrumbs and semaphores of the incomplete
+	 * requests so that inter-timeline dependencies (i.e other timelines)
+	 * remain correctly ordered. And we defer to __i915_request_submit()
+	 * so that all asynchronous waits are correctly handled.
+	 */
+	ENGINE_TRACE(engine, "{ reset rq=%llx:%lld }\n",
+		     rq->fence.context, rq->fence.seqno);
+
+	/* On resubmission of the active request, payload will be scrubbed */
+	if (__i915_request_is_complete(rq))
+		head = rq->tail;
+	else
+		head = __active_request(ce->timeline, rq, -EIO)->head;
+	head = intel_ring_wrap(ce->ring, head);
+
+	/* Scrub the context image to prevent replaying the previous batch */
+	lrc_init_regs(ce, engine, true);
+
+	/* We've switched away, so this should be a no-op, but intent matters */
+	ce->lrc.lrca = lrc_update_regs(ce, engine, head);
+}
+
+static struct intel_engine_cs *
+__execlists_schedule_in(struct i915_request *rq)
+{
+	struct intel_engine_cs * const engine = rq->engine;
+	struct intel_context * const ce = rq->context;
+
+	intel_context_get(ce);
+
+	if (unlikely(intel_context_is_closed(ce) &&
+		     !intel_engine_has_heartbeat(engine)))
+		intel_context_set_banned(ce);
+
+	if (unlikely(intel_context_is_banned(ce)))
+		reset_active(rq, engine);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		lrc_check_regs(ce, engine, "before");
+
+	if (ce->tag) {
+		/* Use a fixed tag for OA and friends */
+		GEM_BUG_ON(ce->tag <= BITS_PER_LONG);
+		ce->lrc.ccid = ce->tag;
+	} else {
+		/* We don't need a strict matching tag, just different values */
+		unsigned int tag = __ffs(engine->context_tag);
+
+		GEM_BUG_ON(tag >= BITS_PER_LONG);
+		__clear_bit(tag, &engine->context_tag);
+		ce->lrc.ccid = (1 + tag) << (GEN11_SW_CTX_ID_SHIFT - 32);
+
+		BUILD_BUG_ON(BITS_PER_LONG > GEN12_MAX_CONTEXT_HW_ID);
+	}
+
+	ce->lrc.ccid |= engine->execlists.ccid;
+
+	__intel_gt_pm_get(engine->gt);
+	if (engine->fw_domain && !engine->fw_active++)
+		intel_uncore_forcewake_get(engine->uncore, engine->fw_domain);
+	execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_IN);
+	intel_engine_context_in(engine);
+
+	CE_TRACE(ce, "schedule-in, ccid:%x\n", ce->lrc.ccid);
+
+	return engine;
+}
+
+static void execlists_schedule_in(struct i915_request *rq, int idx)
+{
+	struct intel_context * const ce = rq->context;
+	struct intel_engine_cs *old;
+
+	GEM_BUG_ON(!intel_engine_pm_is_awake(rq->engine));
+	trace_i915_request_in(rq, idx);
+
+	old = ce->inflight;
+	if (!old)
+		old = __execlists_schedule_in(rq);
+	WRITE_ONCE(ce->inflight, ptr_inc(old));
+
+	GEM_BUG_ON(intel_context_inflight(ce) != rq->engine);
+}
+
+static void
+resubmit_virtual_request(struct i915_request *rq, struct virtual_engine *ve)
+{
+	struct intel_engine_cs *engine = rq->engine;
+
+	spin_lock_irq(&engine->active.lock);
+
+	clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+	WRITE_ONCE(rq->engine, &ve->base);
+	ve->base.submit_request(rq);
+
+	spin_unlock_irq(&engine->active.lock);
+}
+
+static void kick_siblings(struct i915_request *rq, struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+	struct intel_engine_cs *engine = rq->engine;
+
+	/*
+	 * After this point, the rq may be transferred to a new sibling, so
+	 * before we clear ce->inflight make sure that the context has been
+	 * removed from the b->signalers and furthermore we need to make sure
+	 * that the concurrent iterator in signal_irq_work is no longer
+	 * following ce->signal_link.
+	 */
+	if (!list_empty(&ce->signals))
+		intel_context_remove_breadcrumbs(ce, engine->breadcrumbs);
+
+	/*
+	 * This engine is now too busy to run this virtual request, so
+	 * see if we can find an alternative engine for it to execute on.
+	 * Once a request has become bonded to this engine, we treat it the
+	 * same as other native request.
+	 */
+	if (i915_request_in_priority_queue(rq) &&
+	    rq->execution_mask != engine->mask)
+		resubmit_virtual_request(rq, ve);
+
+	if (READ_ONCE(ve->request))
+		tasklet_hi_schedule(&ve->base.execlists.tasklet);
+}
+
+static void __execlists_schedule_out(struct i915_request * const rq,
+				     struct intel_context * const ce)
+{
+	struct intel_engine_cs * const engine = rq->engine;
+	unsigned int ccid;
+
+	/*
+	 * NB process_csb() is not under the engine->active.lock and hence
+	 * schedule_out can race with schedule_in meaning that we should
+	 * refrain from doing non-trivial work here.
+	 */
+
+	CE_TRACE(ce, "schedule-out, ccid:%x\n", ce->lrc.ccid);
+	GEM_BUG_ON(ce->inflight != engine);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		lrc_check_regs(ce, engine, "after");
+
+	/*
+	 * If we have just completed this context, the engine may now be
+	 * idle and we want to re-enter powersaving.
+	 */
+	if (intel_timeline_is_last(ce->timeline, rq) &&
+	    __i915_request_is_complete(rq))
+		intel_engine_add_retire(engine, ce->timeline);
+
+	ccid = ce->lrc.ccid;
+	ccid >>= GEN11_SW_CTX_ID_SHIFT - 32;
+	ccid &= GEN12_MAX_CONTEXT_HW_ID;
+	if (ccid < BITS_PER_LONG) {
+		GEM_BUG_ON(ccid == 0);
+		GEM_BUG_ON(test_bit(ccid - 1, &engine->context_tag));
+		__set_bit(ccid - 1, &engine->context_tag);
+	}
+
+	lrc_update_runtime(ce);
+	intel_engine_context_out(engine);
+	execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_OUT);
+	if (engine->fw_domain && !--engine->fw_active)
+		intel_uncore_forcewake_put(engine->uncore, engine->fw_domain);
+	intel_gt_pm_put_async(engine->gt);
+
+	/*
+	 * If this is part of a virtual engine, its next request may
+	 * have been blocked waiting for access to the active context.
+	 * We have to kick all the siblings again in case we need to
+	 * switch (e.g. the next request is not runnable on this
+	 * engine). Hopefully, we will already have submitted the next
+	 * request before the tasklet runs and do not need to rebuild
+	 * each virtual tree and kick everyone again.
+	 */
+	if (ce->engine != engine)
+		kick_siblings(rq, ce);
+
+	WRITE_ONCE(ce->inflight, NULL);
+	intel_context_put(ce);
+}
+
+static inline void execlists_schedule_out(struct i915_request *rq)
+{
+	struct intel_context * const ce = rq->context;
+
+	trace_i915_request_out(rq);
+
+	GEM_BUG_ON(!ce->inflight);
+	ce->inflight = ptr_dec(ce->inflight);
+	if (!__intel_context_inflight_count(ce->inflight))
+		__execlists_schedule_out(rq, ce);
+
+	i915_request_put(rq);
+}
+
+static u64 execlists_update_context(struct i915_request *rq)
+{
+	struct intel_context *ce = rq->context;
+	u64 desc = ce->lrc.desc;
+	u32 tail, prev;
+
+	/*
+	 * WaIdleLiteRestore:bdw,skl
+	 *
+	 * We should never submit the context with the same RING_TAIL twice
+	 * just in case we submit an empty ring, which confuses the HW.
+	 *
+	 * We append a couple of NOOPs (gen8_emit_wa_tail) after the end of
+	 * the normal request to be able to always advance the RING_TAIL on
+	 * subsequent resubmissions (for lite restore). Should that fail us,
+	 * and we try and submit the same tail again, force the context
+	 * reload.
+	 *
+	 * If we need to return to a preempted context, we need to skip the
+	 * lite-restore and force it to reload the RING_TAIL. Otherwise, the
+	 * HW has a tendency to ignore us rewinding the TAIL to the end of
+	 * an earlier request.
+	 */
+	GEM_BUG_ON(ce->lrc_reg_state[CTX_RING_TAIL] != rq->ring->tail);
+	prev = rq->ring->tail;
+	tail = intel_ring_set_tail(rq->ring, rq->tail);
+	if (unlikely(intel_ring_direction(rq->ring, tail, prev) <= 0))
+		desc |= CTX_DESC_FORCE_RESTORE;
+	ce->lrc_reg_state[CTX_RING_TAIL] = tail;
+	rq->tail = rq->wa_tail;
+
+	/*
+	 * Make sure the context image is complete before we submit it to HW.
+	 *
+	 * Ostensibly, writes (including the WCB) should be flushed prior to
+	 * an uncached write such as our mmio register access, the empirical
+	 * evidence (esp. on Braswell) suggests that the WC write into memory
+	 * may not be visible to the HW prior to the completion of the UC
+	 * register write and that we may begin execution from the context
+	 * before its image is complete leading to invalid PD chasing.
+	 */
+	wmb();
+
+	ce->lrc.desc &= ~CTX_DESC_FORCE_RESTORE;
+	return desc;
+}
+
+static void write_desc(struct intel_engine_execlists *execlists, u64 desc, u32 port)
+{
+	if (execlists->ctrl_reg) {
+		writel(lower_32_bits(desc), execlists->submit_reg + port * 2);
+		writel(upper_32_bits(desc), execlists->submit_reg + port * 2 + 1);
+	} else {
+		writel(upper_32_bits(desc), execlists->submit_reg);
+		writel(lower_32_bits(desc), execlists->submit_reg);
+	}
+}
+
+static __maybe_unused char *
+dump_port(char *buf, int buflen, const char *prefix, struct i915_request *rq)
+{
+	if (!rq)
+		return "";
+
+	snprintf(buf, buflen, "%sccid:%x %llx:%lld%s prio %d",
+		 prefix,
+		 rq->context->lrc.ccid,
+		 rq->fence.context, rq->fence.seqno,
+		 __i915_request_is_complete(rq) ? "!" :
+		 __i915_request_has_started(rq) ? "*" :
+		 "",
+		 rq_prio(rq));
+
+	return buf;
+}
+
+static __maybe_unused noinline void
+trace_ports(const struct intel_engine_execlists *execlists,
+	    const char *msg,
+	    struct i915_request * const *ports)
+{
+	const struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+	char __maybe_unused p0[40], p1[40];
+
+	if (!ports[0])
+		return;
+
+	ENGINE_TRACE(engine, "%s { %s%s }\n", msg,
+		     dump_port(p0, sizeof(p0), "", ports[0]),
+		     dump_port(p1, sizeof(p1), ", ", ports[1]));
+}
+
+static bool
+reset_in_progress(const struct intel_engine_execlists *execlists)
+{
+	return unlikely(!__tasklet_is_enabled(&execlists->tasklet));
+}
+
+static __maybe_unused noinline bool
+assert_pending_valid(const struct intel_engine_execlists *execlists,
+		     const char *msg)
+{
+	struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+	struct i915_request * const *port, *rq;
+	struct intel_context *ce = NULL;
+	bool sentinel = false;
+	u32 ccid = -1;
+
+	trace_ports(execlists, msg, execlists->pending);
+
+	/* We may be messing around with the lists during reset, lalala */
+	if (reset_in_progress(execlists))
+		return true;
+
+	if (!execlists->pending[0]) {
+		GEM_TRACE_ERR("%s: Nothing pending for promotion!\n",
+			      engine->name);
+		return false;
+	}
+
+	if (execlists->pending[execlists_num_ports(execlists)]) {
+		GEM_TRACE_ERR("%s: Excess pending[%d] for promotion!\n",
+			      engine->name, execlists_num_ports(execlists));
+		return false;
+	}
+
+	for (port = execlists->pending; (rq = *port); port++) {
+		unsigned long flags;
+		bool ok = true;
+
+		GEM_BUG_ON(!kref_read(&rq->fence.refcount));
+		GEM_BUG_ON(!i915_request_is_active(rq));
+
+		if (ce == rq->context) {
+			GEM_TRACE_ERR("%s: Dup context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+		ce = rq->context;
+
+		if (ccid == ce->lrc.ccid) {
+			GEM_TRACE_ERR("%s: Dup ccid:%x context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ccid, ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+		ccid = ce->lrc.ccid;
+
+		/*
+		 * Sentinels are supposed to be the last request so they flush
+		 * the current execution off the HW. Check that they are the only
+		 * request in the pending submission.
+		 */
+		if (sentinel) {
+			GEM_TRACE_ERR("%s: context:%llx after sentinel in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+		sentinel = i915_request_has_sentinel(rq);
+
+		/*
+		 * We want virtual requests to only be in the first slot so
+		 * that they are never stuck behind a hog and can be immediately
+		 * transferred onto the next idle engine.
+		 */
+		if (rq->execution_mask != engine->mask &&
+		    port != execlists->pending) {
+			GEM_TRACE_ERR("%s: virtual engine:%llx not in prime position[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+
+		/* Hold tightly onto the lock to prevent concurrent retires! */
+		if (!spin_trylock_irqsave(&rq->lock, flags))
+			continue;
+
+		if (__i915_request_is_complete(rq))
+			goto unlock;
+
+		if (i915_active_is_idle(&ce->active) &&
+		    !intel_context_is_barrier(ce)) {
+			GEM_TRACE_ERR("%s: Inactive context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			ok = false;
+			goto unlock;
+		}
+
+		if (!i915_vma_is_pinned(ce->state)) {
+			GEM_TRACE_ERR("%s: Unpinned context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			ok = false;
+			goto unlock;
+		}
+
+		if (!i915_vma_is_pinned(ce->ring->vma)) {
+			GEM_TRACE_ERR("%s: Unpinned ring:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			ok = false;
+			goto unlock;
+		}
+
+unlock:
+		spin_unlock_irqrestore(&rq->lock, flags);
+		if (!ok)
+			return false;
+	}
+
+	return ce;
+}
+
+static void execlists_submit_ports(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *execlists = &engine->execlists;
+	unsigned int n;
+
+	GEM_BUG_ON(!assert_pending_valid(execlists, "submit"));
+
+	/*
+	 * We can skip acquiring intel_runtime_pm_get() here as it was taken
+	 * on our behalf by the request (see i915_gem_mark_busy()) and it will
+	 * not be relinquished until the device is idle (see
+	 * i915_gem_idle_work_handler()). As a precaution, we make sure
+	 * that all ELSP are drained i.e. we have processed the CSB,
+	 * before allowing ourselves to idle and calling intel_runtime_pm_put().
+	 */
+	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+
+	/*
+	 * ELSQ note: the submit queue is not cleared after being submitted
+	 * to the HW so we need to make sure we always clean it up. This is
+	 * currently ensured by the fact that we always write the same number
+	 * of elsq entries, keep this in mind before changing the loop below.
+	 */
+	for (n = execlists_num_ports(execlists); n--; ) {
+		struct i915_request *rq = execlists->pending[n];
+
+		write_desc(execlists,
+			   rq ? execlists_update_context(rq) : 0,
+			   n);
+	}
+
+	/* we need to manually load the submit queue */
+	if (execlists->ctrl_reg)
+		writel(EL_CTRL_LOAD, execlists->ctrl_reg);
+}
+
+static bool ctx_single_port_submission(const struct intel_context *ce)
+{
+	return (IS_ENABLED(CONFIG_DRM_I915_GVT) &&
+		intel_context_force_single_submission(ce));
+}
+
+static bool can_merge_ctx(const struct intel_context *prev,
+			  const struct intel_context *next)
+{
+	if (prev != next)
+		return false;
+
+	if (ctx_single_port_submission(prev))
+		return false;
+
+	return true;
+}
+
+static unsigned long i915_request_flags(const struct i915_request *rq)
+{
+	return READ_ONCE(rq->fence.flags);
+}
+
+static bool can_merge_rq(const struct i915_request *prev,
+			 const struct i915_request *next)
+{
+	GEM_BUG_ON(prev == next);
+	GEM_BUG_ON(!assert_priority_queue(prev, next));
+
+	/*
+	 * We do not submit known completed requests. Therefore if the next
+	 * request is already completed, we can pretend to merge it in
+	 * with the previous context (and we will skip updating the ELSP
+	 * and tracking). Thus hopefully keeping the ELSP full with active
+	 * contexts, despite the best efforts of preempt-to-busy to confuse
+	 * us.
+	 */
+	if (__i915_request_is_complete(next))
+		return true;
+
+	if (unlikely((i915_request_flags(prev) ^ i915_request_flags(next)) &
+		     (BIT(I915_FENCE_FLAG_NOPREEMPT) |
+		      BIT(I915_FENCE_FLAG_SENTINEL))))
+		return false;
+
+	if (!can_merge_ctx(prev->context, next->context))
+		return false;
+
+	GEM_BUG_ON(i915_seqno_passed(prev->fence.seqno, next->fence.seqno));
+	return true;
+}
+
+static bool virtual_matches(const struct virtual_engine *ve,
+			    const struct i915_request *rq,
+			    const struct intel_engine_cs *engine)
+{
+	const struct intel_engine_cs *inflight;
+
+	if (!rq)
+		return false;
+
+	if (!(rq->execution_mask & engine->mask)) /* We peeked too soon! */
+		return false;
+
+	/*
+	 * We track when the HW has completed saving the context image
+	 * (i.e. when we have seen the final CS event switching out of
+	 * the context) and must not overwrite the context image before
+	 * then. This restricts us to only using the active engine
+	 * while the previous virtualized request is inflight (so
+	 * we reuse the register offsets). This is a very small
+	 * hystersis on the greedy seelction algorithm.
+	 */
+	inflight = intel_context_inflight(&ve->context);
+	if (inflight && inflight != engine)
+		return false;
+
+	return true;
+}
+
+static struct virtual_engine *
+first_virtual_engine(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *el = &engine->execlists;
+	struct rb_node *rb = rb_first_cached(&el->virtual);
+
+	while (rb) {
+		struct virtual_engine *ve =
+			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
+		struct i915_request *rq = READ_ONCE(ve->request);
+
+		/* lazily cleanup after another engine handled rq */
+		if (!rq || !virtual_matches(ve, rq, engine)) {
+			rb_erase_cached(rb, &el->virtual);
+			RB_CLEAR_NODE(rb);
+			rb = rb_first_cached(&el->virtual);
+			continue;
+		}
+
+		return ve;
+	}
+
+	return NULL;
+}
+
+static void virtual_xfer_context(struct virtual_engine *ve,
+				 struct intel_engine_cs *engine)
+{
+	unsigned int n;
+
+	if (likely(engine == ve->siblings[0]))
+		return;
+
+	GEM_BUG_ON(READ_ONCE(ve->context.inflight));
+	if (!intel_engine_has_relative_mmio(engine))
+		lrc_update_offsets(&ve->context, engine);
+
+	/*
+	 * Move the bound engine to the top of the list for
+	 * future execution. We then kick this tasklet first
+	 * before checking others, so that we preferentially
+	 * reuse this set of bound registers.
+	 */
+	for (n = 1; n < ve->num_siblings; n++) {
+		if (ve->siblings[n] == engine) {
+			swap(ve->siblings[n], ve->siblings[0]);
+			break;
+		}
+	}
+}
+
+static void defer_request(struct i915_request *rq, struct list_head * const pl)
+{
+	LIST_HEAD(list);
+
+	/*
+	 * We want to move the interrupted request to the back of
+	 * the round-robin list (i.e. its priority level), but
+	 * in doing so, we must then move all requests that were in
+	 * flight and were waiting for the interrupted request to
+	 * be run after it again.
+	 */
+	do {
+		struct i915_dependency *p;
+
+		GEM_BUG_ON(i915_request_is_active(rq));
+		list_move_tail(&rq->sched.link, pl);
+
+		for_each_waiter(p, rq) {
+			struct i915_request *w =
+				container_of(p->waiter, typeof(*w), sched);
+
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
+			/* Leave semaphores spinning on the other engines */
+			if (w->engine != rq->engine)
+				continue;
+
+			/* No waiter should start before its signaler */
+			GEM_BUG_ON(i915_request_has_initial_breadcrumb(w) &&
+				   __i915_request_has_started(w) &&
+				   !__i915_request_is_complete(rq));
+
+			GEM_BUG_ON(i915_request_is_active(w));
+			if (!i915_request_is_ready(w))
+				continue;
+
+			if (rq_prio(w) < rq_prio(rq))
+				continue;
+
+			GEM_BUG_ON(rq_prio(w) > rq_prio(rq));
+			list_move_tail(&w->sched.link, &list);
+		}
+
+		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
+	} while (rq);
+}
+
+static void defer_active(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+
+	rq = __unwind_incomplete_requests(engine);
+	if (!rq)
+		return;
+
+	defer_request(rq, i915_sched_lookup_priolist(engine, rq_prio(rq)));
+}
+
+static bool
+timeslice_yield(const struct intel_engine_execlists *el,
+		const struct i915_request *rq)
+{
+	/*
+	 * Once bitten, forever smitten!
+	 *
+	 * If the active context ever busy-waited on a semaphore,
+	 * it will be treated as a hog until the end of its timeslice (i.e.
+	 * until it is scheduled out and replaced by a new submission,
+	 * possibly even its own lite-restore). The HW only sends an interrupt
+	 * on the first miss, and we do know if that semaphore has been
+	 * signaled, or even if it is now stuck on another semaphore. Play
+	 * safe, yield if it might be stuck -- it will be given a fresh
+	 * timeslice in the near future.
+	 */
+	return rq->context->lrc.ccid == READ_ONCE(el->yield);
+}
+
+static bool needs_timeslice(const struct intel_engine_cs *engine,
+			    const struct i915_request *rq)
+{
+	if (!intel_engine_has_timeslices(engine))
+		return false;
+
+	/* If not currently active, or about to switch, wait for next event */
+	if (!rq || __i915_request_is_complete(rq))
+		return false;
+
+	/* We do not need to start the timeslice until after the ACK */
+	if (READ_ONCE(engine->execlists.pending[0]))
+		return false;
+
+	/* If ELSP[1] is occupied, always check to see if worth slicing */
+	if (!list_is_last_rcu(&rq->sched.link, &engine->active.requests)) {
+		ENGINE_TRACE(engine, "timeslice required for second inflight context\n");
+		return true;
+	}
+
+	/* Otherwise, ELSP[0] is by itself, but may be waiting in the queue */
+	if (!RB_EMPTY_ROOT(&engine->execlists.queue.rb_root)) {
+		ENGINE_TRACE(engine, "timeslice required for queue\n");
+		return true;
+	}
+
+	if (!RB_EMPTY_ROOT(&engine->execlists.virtual.rb_root)) {
+		ENGINE_TRACE(engine, "timeslice required for virtual\n");
+		return true;
+	}
+
+	return false;
+}
+
+static bool
+timeslice_expired(struct intel_engine_cs *engine, const struct i915_request *rq)
+{
+	const struct intel_engine_execlists *el = &engine->execlists;
+
+	if (i915_request_has_nopreempt(rq) && __i915_request_has_started(rq))
+		return false;
+
+	if (!needs_timeslice(engine, rq))
+		return false;
+
+	return timer_expired(&el->timer) || timeslice_yield(el, rq);
+}
+
+static unsigned long timeslice(const struct intel_engine_cs *engine)
+{
+	return READ_ONCE(engine->props.timeslice_duration_ms);
+}
+
+static void start_timeslice(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *el = &engine->execlists;
+	unsigned long duration;
+
+	/* Disable the timer if there is nothing to switch to */
+	duration = 0;
+	if (needs_timeslice(engine, *el->active)) {
+		/* Avoid continually prolonging an active timeslice */
+		if (timer_active(&el->timer)) {
+			/*
+			 * If we just submitted a new ELSP after an old
+			 * context, that context may have already consumed
+			 * its timeslice, so recheck.
+			 */
+			if (!timer_pending(&el->timer))
+				tasklet_hi_schedule(&el->tasklet);
+			return;
+		}
+
+		duration = timeslice(engine);
+	}
+
+	set_timer_ms(&el->timer, duration);
+}
+
+static void record_preemption(struct intel_engine_execlists *execlists)
+{
+	(void)I915_SELFTEST_ONLY(execlists->preempt_hang.count++);
+}
+
+static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
+					    const struct i915_request *rq)
+{
+	if (!rq)
+		return 0;
+
+	/* Force a fast reset for terminated contexts (ignoring sysfs!) */
+	if (unlikely(intel_context_is_banned(rq->context)))
+		return 1;
+
+	return READ_ONCE(engine->props.preempt_timeout_ms);
+}
+
+static void set_preempt_timeout(struct intel_engine_cs *engine,
+				const struct i915_request *rq)
+{
+	if (!intel_engine_has_preempt_reset(engine))
+		return;
+
+	set_timer_ms(&engine->execlists.preempt,
+		     active_preempt_timeout(engine, rq));
+}
+
+static bool completed(const struct i915_request *rq)
+{
+	if (i915_request_has_sentinel(rq))
+		return false;
+
+	return __i915_request_is_complete(rq);
+}
+
+static void execlists_dequeue(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct i915_request **port = execlists->pending;
+	struct i915_request ** const last_port = port + execlists->port_mask;
+	struct i915_request *last, * const *active;
+	struct virtual_engine *ve;
+	struct rb_node *rb;
+	bool submit = false;
+
+	/*
+	 * Hardware submission is through 2 ports. Conceptually each port
+	 * has a (RING_START, RING_HEAD, RING_TAIL) tuple. RING_START is
+	 * static for a context, and unique to each, so we only execute
+	 * requests belonging to a single context from each ring. RING_HEAD
+	 * is maintained by the CS in the context image, it marks the place
+	 * where it got up to last time, and through RING_TAIL we tell the CS
+	 * where we want to execute up to this time.
+	 *
+	 * In this list the requests are in order of execution. Consecutive
+	 * requests from the same context are adjacent in the ringbuffer. We
+	 * can combine these requests into a single RING_TAIL update:
+	 *
+	 *              RING_HEAD...req1...req2
+	 *                                    ^- RING_TAIL
+	 * since to execute req2 the CS must first execute req1.
+	 *
+	 * Our goal then is to point each port to the end of a consecutive
+	 * sequence of requests as being the most optimal (fewest wake ups
+	 * and context switches) submission.
+	 */
+
+	spin_lock(&engine->active.lock);
+
+	/*
+	 * If the queue is higher priority than the last
+	 * request in the currently active context, submit afresh.
+	 * We will resubmit again afterwards in case we need to split
+	 * the active context to interject the preemption request,
+	 * i.e. we will retrigger preemption following the ack in case
+	 * of trouble.
+	 *
+	 */
+	active = execlists->active;
+	while ((last = *active) && completed(last))
+		active++;
+
+	if (last) {
+		if (need_preempt(engine, last)) {
+			ENGINE_TRACE(engine,
+				     "preempting last=%llx:%lld, prio=%d, hint=%d\n",
+				     last->fence.context,
+				     last->fence.seqno,
+				     last->sched.attr.priority,
+				     execlists->queue_priority_hint);
+			record_preemption(execlists);
+
+			/*
+			 * Don't let the RING_HEAD advance past the breadcrumb
+			 * as we unwind (and until we resubmit) so that we do
+			 * not accidentally tell it to go backwards.
+			 */
+			ring_set_paused(engine, 1);
+
+			/*
+			 * Note that we have not stopped the GPU at this point,
+			 * so we are unwinding the incomplete requests as they
+			 * remain inflight and so by the time we do complete
+			 * the preemption, some of the unwound requests may
+			 * complete!
+			 */
+			__unwind_incomplete_requests(engine);
+
+			last = NULL;
+		} else if (timeslice_expired(engine, last)) {
+			ENGINE_TRACE(engine,
+				     "expired:%s last=%llx:%lld, prio=%d, hint=%d, yield?=%s\n",
+				     yesno(timer_expired(&execlists->timer)),
+				     last->fence.context, last->fence.seqno,
+				     rq_prio(last),
+				     execlists->queue_priority_hint,
+				     yesno(timeslice_yield(execlists, last)));
+
+			/*
+			 * Consume this timeslice; ensure we start a new one.
+			 *
+			 * The timeslice expired, and we will unwind the
+			 * running contexts and recompute the next ELSP.
+			 * If that submit will be the same pair of contexts
+			 * (due to dependency ordering), we will skip the
+			 * submission. If we don't cancel the timer now,
+			 * we will see that the timer has expired and
+			 * reschedule the tasklet; continually until the
+			 * next context switch or other preeemption event.
+			 *
+			 * Since we have decided to reschedule based on
+			 * consumption of this timeslice, if we submit the
+			 * same context again, grant it a full timeslice.
+			 */
+			cancel_timer(&execlists->timer);
+			ring_set_paused(engine, 1);
+			defer_active(engine);
+
+			/*
+			 * Unlike for preemption, if we rewind and continue
+			 * executing the same context as previously active,
+			 * the order of execution will remain the same and
+			 * the tail will only advance. We do not need to
+			 * force a full context restore, as a lite-restore
+			 * is sufficient to resample the monotonic TAIL.
+			 *
+			 * If we switch to any other context, similarly we
+			 * will not rewind TAIL of current context, and
+			 * normal save/restore will preserve state and allow
+			 * us to later continue executing the same request.
+			 */
+			last = NULL;
+		} else {
+			/*
+			 * Otherwise if we already have a request pending
+			 * for execution after the current one, we can
+			 * just wait until the next CS event before
+			 * queuing more. In either case we will force a
+			 * lite-restore preemption event, but if we wait
+			 * we hopefully coalesce several updates into a single
+			 * submission.
+			 */
+			if (active[1]) {
+				/*
+				 * Even if ELSP[1] is occupied and not worthy
+				 * of timeslices, our queue might be.
+				 */
+				spin_unlock(&engine->active.lock);
+				return;
+			}
+		}
+	}
+
+	/* XXX virtual is always taking precedence */
+	while ((ve = first_virtual_engine(engine))) {
+		struct i915_request *rq;
+
+		spin_lock(&ve->base.active.lock);
+
+		rq = ve->request;
+		if (unlikely(!virtual_matches(ve, rq, engine)))
+			goto unlock; /* lost the race to a sibling */
+
+		GEM_BUG_ON(rq->engine != &ve->base);
+		GEM_BUG_ON(rq->context != &ve->context);
+
+		if (unlikely(rq_prio(rq) < queue_prio(execlists))) {
+			spin_unlock(&ve->base.active.lock);
+			break;
+		}
+
+		if (last && !can_merge_rq(last, rq)) {
+			spin_unlock(&ve->base.active.lock);
+			spin_unlock(&engine->active.lock);
+			return; /* leave this for another sibling */
+		}
+
+		ENGINE_TRACE(engine,
+			     "virtual rq=%llx:%lld%s, new engine? %s\n",
+			     rq->fence.context,
+			     rq->fence.seqno,
+			     __i915_request_is_complete(rq) ? "!" :
+			     __i915_request_has_started(rq) ? "*" :
+			     "",
+			     yesno(engine != ve->siblings[0]));
+
+		WRITE_ONCE(ve->request, NULL);
+		WRITE_ONCE(ve->base.execlists.queue_priority_hint, INT_MIN);
+
+		rb = &ve->nodes[engine->id].rb;
+		rb_erase_cached(rb, &execlists->virtual);
+		RB_CLEAR_NODE(rb);
+
+		GEM_BUG_ON(!(rq->execution_mask & engine->mask));
+		WRITE_ONCE(rq->engine, engine);
+
+		if (__i915_request_submit(rq)) {
+			/*
+			 * Only after we confirm that we will submit
+			 * this request (i.e. it has not already
+			 * completed), do we want to update the context.
+			 *
+			 * This serves two purposes. It avoids
+			 * unnecessary work if we are resubmitting an
+			 * already completed request after timeslicing.
+			 * But more importantly, it prevents us altering
+			 * ve->siblings[] on an idle context, where
+			 * we may be using ve->siblings[] in
+			 * virtual_context_enter / virtual_context_exit.
+			 */
+			virtual_xfer_context(ve, engine);
+			GEM_BUG_ON(ve->siblings[0] != engine);
+
+			submit = true;
+			last = rq;
+		}
+
+		i915_request_put(rq);
+unlock:
+		spin_unlock(&ve->base.active.lock);
+
+		/*
+		 * Hmm, we have a bunch of virtual engine requests,
+		 * but the first one was already completed (thanks
+		 * preempt-to-busy!). Keep looking at the veng queue
+		 * until we have no more relevant requests (i.e.
+		 * the normal submit queue has higher priority).
+		 */
+		if (submit)
+			break;
+	}
+
+	while ((rb = rb_first_cached(&execlists->queue))) {
+		struct i915_priolist *p = to_priolist(rb);
+		struct i915_request *rq, *rn;
+		int i;
+
+		priolist_for_each_request_consume(rq, rn, p, i) {
+			bool merge = true;
+
+			/*
+			 * Can we combine this request with the current port?
+			 * It has to be the same context/ringbuffer and not
+			 * have any exceptions (e.g. GVT saying never to
+			 * combine contexts).
+			 *
+			 * If we can combine the requests, we can execute both
+			 * by updating the RING_TAIL to point to the end of the
+			 * second request, and so we never need to tell the
+			 * hardware about the first.
+			 */
+			if (last && !can_merge_rq(last, rq)) {
+				/*
+				 * If we are on the second port and cannot
+				 * combine this request with the last, then we
+				 * are done.
+				 */
+				if (port == last_port)
+					goto done;
+
+				/*
+				 * We must not populate both ELSP[] with the
+				 * same LRCA, i.e. we must submit 2 different
+				 * contexts if we submit 2 ELSP.
+				 */
+				if (last->context == rq->context)
+					goto done;
+
+				if (i915_request_has_sentinel(last))
+					goto done;
+
+				/*
+				 * We avoid submitting virtual requests into
+				 * the secondary ports so that we can migrate
+				 * the request immediately to another engine
+				 * rather than wait for the primary request.
+				 */
+				if (rq->execution_mask != engine->mask)
+					goto done;
+
+				/*
+				 * If GVT overrides us we only ever submit
+				 * port[0], leaving port[1] empty. Note that we
+				 * also have to be careful that we don't queue
+				 * the same context (even though a different
+				 * request) to the second port.
+				 */
+				if (ctx_single_port_submission(last->context) ||
+				    ctx_single_port_submission(rq->context))
+					goto done;
+
+				merge = false;
+			}
+
+			if (__i915_request_submit(rq)) {
+				if (!merge) {
+					*port++ = i915_request_get(last);
+					last = NULL;
+				}
+
+				GEM_BUG_ON(last &&
+					   !can_merge_ctx(last->context,
+							  rq->context));
+				GEM_BUG_ON(last &&
+					   i915_seqno_passed(last->fence.seqno,
+							     rq->fence.seqno));
+
+				submit = true;
+				last = rq;
+			}
+		}
+
+		rb_erase_cached(&p->node, &execlists->queue);
+		i915_priolist_free(p);
+	}
+done:
+	*port++ = i915_request_get(last);
+
+	/*
+	 * Here be a bit of magic! Or sleight-of-hand, whichever you prefer.
+	 *
+	 * We choose the priority hint such that if we add a request of greater
+	 * priority than this, we kick the submission tasklet to decide on
+	 * the right order of submitting the requests to hardware. We must
+	 * also be prepared to reorder requests as they are in-flight on the
+	 * HW. We derive the priority hint then as the first "hole" in
+	 * the HW submission ports and if there are no available slots,
+	 * the priority of the lowest executing request, i.e. last.
+	 *
+	 * When we do receive a higher priority request ready to run from the
+	 * user, see queue_request(), the priority hint is bumped to that
+	 * request triggering preemption on the next dequeue (or subsequent
+	 * interrupt for secondary ports).
+	 */
+	execlists->queue_priority_hint = queue_prio(execlists);
+	spin_unlock(&engine->active.lock);
+
+	/*
+	 * We can skip poking the HW if we ended up with exactly the same set
+	 * of requests as currently running, e.g. trying to timeslice a pair
+	 * of ordered contexts.
+	 */
+	if (submit &&
+	    memcmp(active,
+		   execlists->pending,
+		   (port - execlists->pending) * sizeof(*port))) {
+		*port = NULL;
+		while (port-- != execlists->pending)
+			execlists_schedule_in(*port, port - execlists->pending);
+
+		WRITE_ONCE(execlists->yield, -1);
+		set_preempt_timeout(engine, *active);
+		execlists_submit_ports(engine);
+	} else {
+		ring_set_paused(engine, 0);
+		while (port-- != execlists->pending)
+			i915_request_put(*port);
+		*execlists->pending = NULL;
+	}
+}
+
+static void execlists_dequeue_irq(struct intel_engine_cs *engine)
+{
+	local_irq_disable(); /* Suspend interrupts across request submission */
+	execlists_dequeue(engine);
+	local_irq_enable(); /* flush irq_work (e.g. breadcrumb enabling) */
+}
+
+static void clear_ports(struct i915_request **ports, int count)
+{
+	memset_p((void **)ports, NULL, count);
+}
+
+static void
+copy_ports(struct i915_request **dst, struct i915_request **src, int count)
+{
+	/* A memcpy_p() would be very useful here! */
+	while (count--)
+		WRITE_ONCE(*dst++, *src++); /* avoid write tearing */
+}
+
+static struct i915_request **
+cancel_port_requests(struct intel_engine_execlists * const execlists,
+		     struct i915_request **inactive)
+{
+	struct i915_request * const *port;
+
+	for (port = execlists->pending; *port; port++)
+		*inactive++ = *port;
+	clear_ports(execlists->pending, ARRAY_SIZE(execlists->pending));
+
+	/* Mark the end of active before we overwrite *active */
+	for (port = xchg(&execlists->active, execlists->pending); *port; port++)
+		*inactive++ = *port;
+	clear_ports(execlists->inflight, ARRAY_SIZE(execlists->inflight));
+
+	smp_wmb(); /* complete the seqlock for execlists_active() */
+	WRITE_ONCE(execlists->active, execlists->inflight);
+
+	/* Having cancelled all outstanding process_csb(), stop their timers */
+	GEM_BUG_ON(execlists->pending[0]);
+	cancel_timer(&execlists->timer);
+	cancel_timer(&execlists->preempt);
+
+	return inactive;
+}
+
+static void invalidate_csb_entries(const u64 *first, const u64 *last)
+{
+	clflush((void *)first);
+	clflush((void *)last);
+}
+
+/*
+ * Starting with Gen12, the status has a new format:
+ *
+ *     bit  0:     switched to new queue
+ *     bit  1:     reserved
+ *     bit  2:     semaphore wait mode (poll or signal), only valid when
+ *                 switch detail is set to "wait on semaphore"
+ *     bits 3-5:   engine class
+ *     bits 6-11:  engine instance
+ *     bits 12-14: reserved
+ *     bits 15-25: sw context id of the lrc the GT switched to
+ *     bits 26-31: sw counter of the lrc the GT switched to
+ *     bits 32-35: context switch detail
+ *                  - 0: ctx complete
+ *                  - 1: wait on sync flip
+ *                  - 2: wait on vblank
+ *                  - 3: wait on scanline
+ *                  - 4: wait on semaphore
+ *                  - 5: context preempted (not on SEMAPHORE_WAIT or
+ *                       WAIT_FOR_EVENT)
+ *     bit  36:    reserved
+ *     bits 37-43: wait detail (for switch detail 1 to 4)
+ *     bits 44-46: reserved
+ *     bits 47-57: sw context id of the lrc the GT switched away from
+ *     bits 58-63: sw counter of the lrc the GT switched away from
+ */
+static bool gen12_csb_parse(const u64 csb)
+{
+	bool ctx_away_valid = GEN12_CSB_CTX_VALID(upper_32_bits(csb));
+	bool new_queue =
+		lower_32_bits(csb) & GEN12_CTX_STATUS_SWITCHED_TO_NEW_QUEUE;
+
+	/*
+	 * The context switch detail is not guaranteed to be 5 when a preemption
+	 * occurs, so we can't just check for that. The check below works for
+	 * all the cases we care about, including preemptions of WAIT
+	 * instructions and lite-restore. Preempt-to-idle via the CTRL register
+	 * would require some extra handling, but we don't support that.
+	 */
+	if (!ctx_away_valid || new_queue) {
+		GEM_BUG_ON(!GEN12_CSB_CTX_VALID(lower_32_bits(csb)));
+		return true;
+	}
+
+	/*
+	 * switch detail = 5 is covered by the case above and we do not expect a
+	 * context switch on an unsuccessful wait instruction since we always
+	 * use polling mode.
+	 */
+	GEM_BUG_ON(GEN12_CTX_SWITCH_DETAIL(upper_32_bits(csb)));
+	return false;
+}
+
+static bool gen8_csb_parse(const u64 csb)
+{
+	return csb & (GEN8_CTX_STATUS_IDLE_ACTIVE | GEN8_CTX_STATUS_PREEMPTED);
+}
+
+static noinline u64
+wa_csb_read(const struct intel_engine_cs *engine, u64 * const csb)
+{
+	u64 entry;
+
+	/*
+	 * Reading from the HWSP has one particular advantage: we can detect
+	 * a stale entry. Since the write into HWSP is broken, we have no reason
+	 * to trust the HW at all, the mmio entry may equally be unordered, so
+	 * we prefer the path that is self-checking and as a last resort,
+	 * return the mmio value.
+	 *
+	 * tgl,dg1:HSDES#22011327657
+	 */
+	preempt_disable();
+	if (wait_for_atomic_us((entry = READ_ONCE(*csb)) != -1, 10)) {
+		int idx = csb - engine->execlists.csb_status;
+		int status;
+
+		status = GEN8_EXECLISTS_STATUS_BUF;
+		if (idx >= 6) {
+			status = GEN11_EXECLISTS_STATUS_BUF2;
+			idx -= 6;
+		}
+		status += sizeof(u64) * idx;
+
+		entry = intel_uncore_read64(engine->uncore,
+					    _MMIO(engine->mmio_base + status));
+	}
+	preempt_enable();
+
+	return entry;
+}
+
+static u64 csb_read(const struct intel_engine_cs *engine, u64 * const csb)
+{
+	u64 entry = READ_ONCE(*csb);
+
+	/*
+	 * Unfortunately, the GPU does not always serialise its write
+	 * of the CSB entries before its write of the CSB pointer, at least
+	 * from the perspective of the CPU, using what is known as a Global
+	 * Observation Point. We may read a new CSB tail pointer, but then
+	 * read the stale CSB entries, causing us to misinterpret the
+	 * context-switch events, and eventually declare the GPU hung.
+	 *
+	 * icl:HSDES#1806554093
+	 * tgl:HSDES#22011248461
+	 */
+	if (unlikely(entry == -1))
+		entry = wa_csb_read(engine, csb);
+
+	/* Consume this entry so that we can spot its future reuse. */
+	WRITE_ONCE(*csb, -1);
+
+	/* ELSP is an implicit wmb() before the GPU wraps and overwrites csb */
+	return entry;
+}
+
+static void new_timeslice(struct intel_engine_execlists *el)
+{
+	/* By cancelling, we will start afresh in start_timeslice() */
+	cancel_timer(&el->timer);
+}
+
+static struct i915_request **
+process_csb(struct intel_engine_cs *engine, struct i915_request **inactive)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	u64 * const buf = execlists->csb_status;
+	const u8 num_entries = execlists->csb_size;
+	struct i915_request **prev;
+	u8 head, tail;
+
+	/*
+	 * As we modify our execlists state tracking we require exclusive
+	 * access. Either we are inside the tasklet, or the tasklet is disabled
+	 * and we assume that is only inside the reset paths and so serialised.
+	 */
+	GEM_BUG_ON(!tasklet_is_locked(&execlists->tasklet) &&
+		   !reset_in_progress(execlists));
+	GEM_BUG_ON(!intel_engine_in_execlists_submission_mode(engine));
+
+	/*
+	 * Note that csb_write, csb_status may be either in HWSP or mmio.
+	 * When reading from the csb_write mmio register, we have to be
+	 * careful to only use the GEN8_CSB_WRITE_PTR portion, which is
+	 * the low 4bits. As it happens we know the next 4bits are always
+	 * zero and so we can simply masked off the low u8 of the register
+	 * and treat it identically to reading from the HWSP (without having
+	 * to use explicit shifting and masking, and probably bifurcating
+	 * the code to handle the legacy mmio read).
+	 */
+	head = execlists->csb_head;
+	tail = READ_ONCE(*execlists->csb_write);
+	if (unlikely(head == tail))
+		return inactive;
+
+	/*
+	 * We will consume all events from HW, or at least pretend to.
+	 *
+	 * The sequence of events from the HW is deterministic, and derived
+	 * from our writes to the ELSP, with a smidgen of variability for
+	 * the arrival of the asynchronous requests wrt to the inflight
+	 * execution. If the HW sends an event that does not correspond with
+	 * the one we are expecting, we have to abandon all hope as we lose
+	 * all tracking of what the engine is actually executing. We will
+	 * only detect we are out of sequence with the HW when we get an
+	 * 'impossible' event because we have already drained our own
+	 * preemption/promotion queue. If this occurs, we know that we likely
+	 * lost track of execution earlier and must unwind and restart, the
+	 * simplest way is by stop processing the event queue and force the
+	 * engine to reset.
+	 */
+	execlists->csb_head = tail;
+	ENGINE_TRACE(engine, "cs-irq head=%d, tail=%d\n", head, tail);
+
+	/*
+	 * Hopefully paired with a wmb() in HW!
+	 *
+	 * We must complete the read of the write pointer before any reads
+	 * from the CSB, so that we do not see stale values. Without an rmb
+	 * (lfence) the HW may speculatively perform the CSB[] reads *before*
+	 * we perform the READ_ONCE(*csb_write).
+	 */
+	rmb();
+
+	/* Remember who was last running under the timer */
+	prev = inactive;
+	*prev = NULL;
+
+	do {
+		bool promote;
+		u64 csb;
+
+		if (++head == num_entries)
+			head = 0;
+
+		/*
+		 * We are flying near dragons again.
+		 *
+		 * We hold a reference to the request in execlist_port[]
+		 * but no more than that. We are operating in softirq
+		 * context and so cannot hold any mutex or sleep. That
+		 * prevents us stopping the requests we are processing
+		 * in port[] from being retired simultaneously (the
+		 * breadcrumb will be complete before we see the
+		 * context-switch). As we only hold the reference to the
+		 * request, any pointer chasing underneath the request
+		 * is subject to a potential use-after-free. Thus we
+		 * store all of the bookkeeping within port[] as
+		 * required, and avoid using unguarded pointers beneath
+		 * request itself. The same applies to the atomic
+		 * status notifier.
+		 */
+
+		csb = csb_read(engine, buf + head);
+		ENGINE_TRACE(engine, "csb[%d]: status=0x%08x:0x%08x\n",
+			     head, upper_32_bits(csb), lower_32_bits(csb));
+
+		if (INTEL_GEN(engine->i915) >= 12)
+			promote = gen12_csb_parse(csb);
+		else
+			promote = gen8_csb_parse(csb);
+		if (promote) {
+			struct i915_request * const *old = execlists->active;
+
+			if (GEM_WARN_ON(!*execlists->pending)) {
+				execlists->error_interrupt |= ERROR_CSB;
+				break;
+			}
+
+			ring_set_paused(engine, 0);
+
+			/* Point active to the new ELSP; prevent overwriting */
+			WRITE_ONCE(execlists->active, execlists->pending);
+			smp_wmb(); /* notify execlists_active() */
+
+			/* cancel old inflight, prepare for switch */
+			trace_ports(execlists, "preempted", old);
+			while (*old)
+				*inactive++ = *old++;
+
+			/* switch pending to inflight */
+			GEM_BUG_ON(!assert_pending_valid(execlists, "promote"));
+			copy_ports(execlists->inflight,
+				   execlists->pending,
+				   execlists_num_ports(execlists));
+			smp_wmb(); /* complete the seqlock */
+			WRITE_ONCE(execlists->active, execlists->inflight);
+
+			/* XXX Magic delay for tgl */
+			ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
+
+			WRITE_ONCE(execlists->pending[0], NULL);
+		} else {
+			if (GEM_WARN_ON(!*execlists->active)) {
+				execlists->error_interrupt |= ERROR_CSB;
+				break;
+			}
+
+			/* port0 completed, advanced to port1 */
+			trace_ports(execlists, "completed", execlists->active);
+
+			/*
+			 * We rely on the hardware being strongly
+			 * ordered, that the breadcrumb write is
+			 * coherent (visible from the CPU) before the
+			 * user interrupt is processed. One might assume
+			 * that the breadcrumb write being before the
+			 * user interrupt and the CS event for the context
+			 * switch would therefore be before the CS event
+			 * itself...
+			 */
+			if (GEM_SHOW_DEBUG() &&
+			    !__i915_request_is_complete(*execlists->active)) {
+				struct i915_request *rq = *execlists->active;
+				const u32 *regs __maybe_unused =
+					rq->context->lrc_reg_state;
+
+				ENGINE_TRACE(engine,
+					     "context completed before request!\n");
+				ENGINE_TRACE(engine,
+					     "ring:{start:0x%08x, head:%04x, tail:%04x, ctl:%08x, mode:%08x}\n",
+					     ENGINE_READ(engine, RING_START),
+					     ENGINE_READ(engine, RING_HEAD) & HEAD_ADDR,
+					     ENGINE_READ(engine, RING_TAIL) & TAIL_ADDR,
+					     ENGINE_READ(engine, RING_CTL),
+					     ENGINE_READ(engine, RING_MI_MODE));
+				ENGINE_TRACE(engine,
+					     "rq:{start:%08x, head:%04x, tail:%04x, seqno:%llx:%d, hwsp:%d}, ",
+					     i915_ggtt_offset(rq->ring->vma),
+					     rq->head, rq->tail,
+					     rq->fence.context,
+					     lower_32_bits(rq->fence.seqno),
+					     hwsp_seqno(rq));
+				ENGINE_TRACE(engine,
+					     "ctx:{start:%08x, head:%04x, tail:%04x}, ",
+					     regs[CTX_RING_START],
+					     regs[CTX_RING_HEAD],
+					     regs[CTX_RING_TAIL]);
+			}
+
+			*inactive++ = *execlists->active++;
+
+			GEM_BUG_ON(execlists->active - execlists->inflight >
+				   execlists_num_ports(execlists));
+		}
+	} while (head != tail);
+
+	/*
+	 * Gen11 has proven to fail wrt global observation point between
+	 * entry and tail update, failing on the ordering and thus
+	 * we see an old entry in the context status buffer.
+	 *
+	 * Forcibly evict out entries for the next gpu csb update,
+	 * to increase the odds that we get a fresh entries with non
+	 * working hardware. The cost for doing so comes out mostly with
+	 * the wash as hardware, working or not, will need to do the
+	 * invalidation before.
+	 */
+	invalidate_csb_entries(&buf[0], &buf[num_entries - 1]);
+
+	/*
+	 * We assume that any event reflects a change in context flow
+	 * and merits a fresh timeslice. We reinstall the timer after
+	 * inspecting the queue to see if we need to resumbit.
+	 */
+	if (*prev != *execlists->active) /* elide lite-restores */
+		new_timeslice(execlists);
+
+	return inactive;
+}
+
+static void post_process_csb(struct i915_request **port,
+			     struct i915_request **last)
+{
+	while (port != last)
+		execlists_schedule_out(*port++);
+}
+
+static void __execlists_hold(struct i915_request *rq)
+{
+	LIST_HEAD(list);
+
+	do {
+		struct i915_dependency *p;
+
+		if (i915_request_is_active(rq))
+			__i915_request_unsubmit(rq);
+
+		clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+		list_move_tail(&rq->sched.link, &rq->engine->active.hold);
+		i915_request_set_hold(rq);
+		RQ_TRACE(rq, "on hold\n");
+
+		for_each_waiter(p, rq) {
+			struct i915_request *w =
+				container_of(p->waiter, typeof(*w), sched);
+
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
+			/* Leave semaphores spinning on the other engines */
+			if (w->engine != rq->engine)
+				continue;
+
+			if (!i915_request_is_ready(w))
+				continue;
+
+			if (__i915_request_is_complete(w))
+				continue;
+
+			if (i915_request_on_hold(w))
+				continue;
+
+			list_move_tail(&w->sched.link, &list);
+		}
+
+		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
+	} while (rq);
+}
+
+static bool execlists_hold(struct intel_engine_cs *engine,
+			   struct i915_request *rq)
+{
+	if (i915_request_on_hold(rq))
+		return false;
+
+	spin_lock_irq(&engine->active.lock);
+
+	if (__i915_request_is_complete(rq)) { /* too late! */
+		rq = NULL;
+		goto unlock;
+	}
+
+	/*
+	 * Transfer this request onto the hold queue to prevent it
+	 * being resumbitted to HW (and potentially completed) before we have
+	 * released it. Since we may have already submitted following
+	 * requests, we need to remove those as well.
+	 */
+	GEM_BUG_ON(i915_request_on_hold(rq));
+	GEM_BUG_ON(rq->engine != engine);
+	__execlists_hold(rq);
+	GEM_BUG_ON(list_empty(&engine->active.hold));
+
+unlock:
+	spin_unlock_irq(&engine->active.lock);
+	return rq;
+}
+
+static bool hold_request(const struct i915_request *rq)
+{
+	struct i915_dependency *p;
+	bool result = false;
+
+	/*
+	 * If one of our ancestors is on hold, we must also be on hold,
+	 * otherwise we will bypass it and execute before it.
+	 */
+	rcu_read_lock();
+	for_each_signaler(p, rq) {
+		const struct i915_request *s =
+			container_of(p->signaler, typeof(*s), sched);
+
+		if (s->engine != rq->engine)
+			continue;
+
+		result = i915_request_on_hold(s);
+		if (result)
+			break;
+	}
+	rcu_read_unlock();
+
+	return result;
+}
+
+static void __execlists_unhold(struct i915_request *rq)
+{
+	LIST_HEAD(list);
+
+	do {
+		struct i915_dependency *p;
+
+		RQ_TRACE(rq, "hold release\n");
+
+		GEM_BUG_ON(!i915_request_on_hold(rq));
+		GEM_BUG_ON(!i915_sw_fence_signaled(&rq->submit));
+
+		i915_request_clear_hold(rq);
+		list_move_tail(&rq->sched.link,
+			       i915_sched_lookup_priolist(rq->engine,
+							  rq_prio(rq)));
+		set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+
+		/* Also release any children on this engine that are ready */
+		for_each_waiter(p, rq) {
+			struct i915_request *w =
+				container_of(p->waiter, typeof(*w), sched);
+
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
+			/* Propagate any change in error status */
+			if (rq->fence.error)
+				i915_request_set_error_once(w, rq->fence.error);
+
+			if (w->engine != rq->engine)
+				continue;
+
+			if (!i915_request_on_hold(w))
+				continue;
+
+			/* Check that no other parents are also on hold */
+			if (hold_request(w))
+				continue;
+
+			list_move_tail(&w->sched.link, &list);
+		}
+
+		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
+	} while (rq);
+}
+
+static void execlists_unhold(struct intel_engine_cs *engine,
+			     struct i915_request *rq)
+{
+	spin_lock_irq(&engine->active.lock);
+
+	/*
+	 * Move this request back to the priority queue, and all of its
+	 * children and grandchildren that were suspended along with it.
+	 */
+	__execlists_unhold(rq);
+
+	if (rq_prio(rq) > engine->execlists.queue_priority_hint) {
+		engine->execlists.queue_priority_hint = rq_prio(rq);
+		tasklet_hi_schedule(&engine->execlists.tasklet);
+	}
+
+	spin_unlock_irq(&engine->active.lock);
+}
+
+struct execlists_capture {
+	struct work_struct work;
+	struct i915_request *rq;
+	struct i915_gpu_coredump *error;
+};
+
+static void execlists_capture_work(struct work_struct *work)
+{
+	struct execlists_capture *cap = container_of(work, typeof(*cap), work);
+	const gfp_t gfp = GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN;
+	struct intel_engine_cs *engine = cap->rq->engine;
+	struct intel_gt_coredump *gt = cap->error->gt;
+	struct intel_engine_capture_vma *vma;
+
+	/* Compress all the objects attached to the request, slow! */
+	vma = intel_engine_coredump_add_request(gt->engine, cap->rq, gfp);
+	if (vma) {
+		struct i915_vma_compress *compress =
+			i915_vma_capture_prepare(gt);
+
+		intel_engine_coredump_add_vma(gt->engine, vma, compress);
+		i915_vma_capture_finish(gt, compress);
+	}
+
+	gt->simulated = gt->engine->simulated;
+	cap->error->simulated = gt->simulated;
+
+	/* Publish the error state, and announce it to the world */
+	i915_error_state_store(cap->error);
+	i915_gpu_coredump_put(cap->error);
+
+	/* Return this request and all that depend upon it for signaling */
+	execlists_unhold(engine, cap->rq);
+	i915_request_put(cap->rq);
+
+	kfree(cap);
+}
+
+static struct execlists_capture *capture_regs(struct intel_engine_cs *engine)
+{
+	const gfp_t gfp = GFP_ATOMIC | __GFP_NOWARN;
+	struct execlists_capture *cap;
+
+	cap = kmalloc(sizeof(*cap), gfp);
+	if (!cap)
+		return NULL;
+
+	cap->error = i915_gpu_coredump_alloc(engine->i915, gfp);
+	if (!cap->error)
+		goto err_cap;
+
+	cap->error->gt = intel_gt_coredump_alloc(engine->gt, gfp);
+	if (!cap->error->gt)
+		goto err_gpu;
+
+	cap->error->gt->engine = intel_engine_coredump_alloc(engine, gfp);
+	if (!cap->error->gt->engine)
+		goto err_gt;
+
+	cap->error->gt->engine->hung = true;
+
+	return cap;
+
+err_gt:
+	kfree(cap->error->gt);
+err_gpu:
+	kfree(cap->error);
+err_cap:
+	kfree(cap);
+	return NULL;
+}
+
+static struct i915_request *
+active_context(struct intel_engine_cs *engine, u32 ccid)
+{
+	const struct intel_engine_execlists * const el = &engine->execlists;
+	struct i915_request * const *port, *rq;
+
+	/*
+	 * Use the most recent result from process_csb(), but just in case
+	 * we trigger an error (via interrupt) before the first CS event has
+	 * been written, peek at the next submission.
+	 */
+
+	for (port = el->active; (rq = *port); port++) {
+		if (rq->context->lrc.ccid == ccid) {
+			ENGINE_TRACE(engine,
+				     "ccid:%x found at active:%zd\n",
+				     ccid, port - el->active);
+			return rq;
+		}
+	}
+
+	for (port = el->pending; (rq = *port); port++) {
+		if (rq->context->lrc.ccid == ccid) {
+			ENGINE_TRACE(engine,
+				     "ccid:%x found at pending:%zd\n",
+				     ccid, port - el->pending);
+			return rq;
+		}
+	}
+
+	ENGINE_TRACE(engine, "ccid:%x not found\n", ccid);
+	return NULL;
+}
+
+static u32 active_ccid(struct intel_engine_cs *engine)
+{
+	return ENGINE_READ_FW(engine, RING_EXECLIST_STATUS_HI);
+}
+
+static void execlists_capture(struct intel_engine_cs *engine)
+{
+	struct execlists_capture *cap;
+
+	if (!IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR))
+		return;
+
+	/*
+	 * We need to _quickly_ capture the engine state before we reset.
+	 * We are inside an atomic section (softirq) here and we are delaying
+	 * the forced preemption event.
+	 */
+	cap = capture_regs(engine);
+	if (!cap)
+		return;
+
+	spin_lock_irq(&engine->active.lock);
+	cap->rq = active_context(engine, active_ccid(engine));
+	if (cap->rq) {
+		cap->rq = active_request(cap->rq->context->timeline, cap->rq);
+		cap->rq = i915_request_get_rcu(cap->rq);
+	}
+	spin_unlock_irq(&engine->active.lock);
+	if (!cap->rq)
+		goto err_free;
+
+	/*
+	 * Remove the request from the execlists queue, and take ownership
+	 * of the request. We pass it to our worker who will _slowly_ compress
+	 * all the pages the _user_ requested for debugging their batch, after
+	 * which we return it to the queue for signaling.
+	 *
+	 * By removing them from the execlists queue, we also remove the
+	 * requests from being processed by __unwind_incomplete_requests()
+	 * during the intel_engine_reset(), and so they will *not* be replayed
+	 * afterwards.
+	 *
+	 * Note that because we have not yet reset the engine at this point,
+	 * it is possible for the request that we have identified as being
+	 * guilty, did in fact complete and we will then hit an arbitration
+	 * point allowing the outstanding preemption to succeed. The likelihood
+	 * of that is very low (as capturing of the engine registers should be
+	 * fast enough to run inside an irq-off atomic section!), so we will
+	 * simply hold that request accountable for being non-preemptible
+	 * long enough to force the reset.
+	 */
+	if (!execlists_hold(engine, cap->rq))
+		goto err_rq;
+
+	INIT_WORK(&cap->work, execlists_capture_work);
+	schedule_work(&cap->work);
+	return;
+
+err_rq:
+	i915_request_put(cap->rq);
+err_free:
+	i915_gpu_coredump_put(cap->error);
+	kfree(cap);
+}
+
+static void execlists_reset(struct intel_engine_cs *engine, const char *msg)
+{
+	const unsigned int bit = I915_RESET_ENGINE + engine->id;
+	unsigned long *lock = &engine->gt->reset.flags;
+
+	if (!intel_has_reset_engine(engine->gt))
+		return;
+
+	if (test_and_set_bit(bit, lock))
+		return;
+
+	ENGINE_TRACE(engine, "reset for %s\n", msg);
+
+	/* Mark this tasklet as disabled to avoid waiting for it to complete */
+	tasklet_disable_nosync(&engine->execlists.tasklet);
+
+	ring_set_paused(engine, 1); /* Freeze the current request in place */
+	execlists_capture(engine);
+	intel_engine_reset(engine, msg);
+
+	tasklet_enable(&engine->execlists.tasklet);
+	clear_and_wake_up_bit(bit, lock);
+}
+
+static bool preempt_timeout(const struct intel_engine_cs *const engine)
+{
+	const struct timer_list *t = &engine->execlists.preempt;
+
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
+		return false;
+
+	if (!timer_expired(t))
+		return false;
+
+	return engine->execlists.pending[0];
+}
+
+/*
+ * Check the unread Context Status Buffers and manage the submission of new
+ * contexts to the ELSP accordingly.
+ */
+static void execlists_submission_tasklet(unsigned long data)
+{
+	struct intel_engine_cs * const engine = (struct intel_engine_cs *)data;
+	struct i915_request *post[2 * EXECLIST_MAX_PORTS];
+	struct i915_request **inactive;
+
+	rcu_read_lock();
+	inactive = process_csb(engine, post);
+	GEM_BUG_ON(inactive - post > ARRAY_SIZE(post));
+
+	if (unlikely(preempt_timeout(engine))) {
+		cancel_timer(&engine->execlists.preempt);
+		engine->execlists.error_interrupt |= ERROR_PREEMPT;
+	}
+
+	if (unlikely(READ_ONCE(engine->execlists.error_interrupt))) {
+		const char *msg;
+
+		/* Generate the error message in priority wrt to the user! */
+		if (engine->execlists.error_interrupt & GENMASK(15, 0))
+			msg = "CS error"; /* thrown by a user payload */
+		else if (engine->execlists.error_interrupt & ERROR_CSB)
+			msg = "invalid CSB event";
+		else if (engine->execlists.error_interrupt & ERROR_PREEMPT)
+			msg = "preemption time out";
+		else
+			msg = "internal error";
+
+		engine->execlists.error_interrupt = 0;
+		execlists_reset(engine, msg);
+	}
+
+	if (!engine->execlists.pending[0]) {
+		execlists_dequeue_irq(engine);
+		start_timeslice(engine);
+	}
+
+	post_process_csb(post, inactive);
+	rcu_read_unlock();
+}
+
+static void __execlists_kick(struct intel_engine_execlists *execlists)
+{
+	/* Kick the tasklet for some interrupt coalescing and reset handling */
+	tasklet_hi_schedule(&execlists->tasklet);
+}
+
+#define execlists_kick(t, member) \
+	__execlists_kick(container_of(t, struct intel_engine_execlists, member))
+
+static void execlists_timeslice(struct timer_list *timer)
+{
+	execlists_kick(timer, timer);
+}
+
+static void execlists_preempt(struct timer_list *timer)
+{
+	execlists_kick(timer, preempt);
+}
+
+static void queue_request(struct intel_engine_cs *engine,
+			  struct i915_request *rq)
+{
+	GEM_BUG_ON(!list_empty(&rq->sched.link));
+	list_add_tail(&rq->sched.link,
+		      i915_sched_lookup_priolist(engine, rq_prio(rq)));
+	set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+}
+
+static bool submit_queue(struct intel_engine_cs *engine,
+			 const struct i915_request *rq)
+{
+	struct intel_engine_execlists *execlists = &engine->execlists;
+
+	if (rq_prio(rq) <= execlists->queue_priority_hint)
+		return false;
+
+	execlists->queue_priority_hint = rq_prio(rq);
+	return true;
+}
+
+static bool ancestor_on_hold(const struct intel_engine_cs *engine,
+			     const struct i915_request *rq)
+{
+	GEM_BUG_ON(i915_request_on_hold(rq));
+	return !list_empty(&engine->active.hold) && hold_request(rq);
+}
+
+static void execlists_submit_request(struct i915_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&engine->active.lock, flags);
+
+	if (unlikely(ancestor_on_hold(engine, request))) {
+		RQ_TRACE(request, "ancestor on hold\n");
+		list_add_tail(&request->sched.link, &engine->active.hold);
+		i915_request_set_hold(request);
+	} else {
+		queue_request(engine, request);
+
+		GEM_BUG_ON(RB_EMPTY_ROOT(&engine->execlists.queue.rb_root));
+		GEM_BUG_ON(list_empty(&request->sched.link));
+
+		if (submit_queue(engine, request))
+			__execlists_kick(&engine->execlists);
+	}
+
+	spin_unlock_irqrestore(&engine->active.lock, flags);
+}
+
+static int execlists_context_pre_pin(struct intel_context *ce,
+				     struct i915_gem_ww_ctx *ww,
+				     void **vaddr)
+{
+	return lrc_pre_pin(ce, ce->engine, ww, vaddr);
+}
+
+static int execlists_context_pin(struct intel_context *ce, void *vaddr)
+{
+	return lrc_pin(ce, ce->engine, vaddr);
+}
+
+static int execlists_context_alloc(struct intel_context *ce)
+{
+	return lrc_alloc(ce, ce->engine);
+}
+
+static const struct intel_context_ops execlists_context_ops = {
+	.flags = COPS_HAS_INFLIGHT,
+
+	.alloc = execlists_context_alloc,
+
+	.pre_pin = execlists_context_pre_pin,
+	.pin = execlists_context_pin,
+	.unpin = lrc_unpin,
+	.post_unpin = lrc_post_unpin,
+
+	.enter = intel_context_enter_engine,
+	.exit = intel_context_exit_engine,
+
+	.reset = lrc_reset,
+	.destroy = lrc_destroy,
+};
+
+static int emit_pdps(struct i915_request *rq)
+{
+	const struct intel_engine_cs * const engine = rq->engine;
+	struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(rq->context->vm);
+	int err, i;
+	u32 *cs;
+
+	GEM_BUG_ON(intel_vgpu_active(rq->engine->i915));
+
+	/*
+	 * Beware ye of the dragons, this sequence is magic!
+	 *
+	 * Small changes to this sequence can cause anything from
+	 * GPU hangs to forcewake errors and machine lockups!
+	 */
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	/* Flush any residual operations from the context load */
+	err = engine->emit_flush(rq, EMIT_FLUSH);
+	if (err)
+		return err;
+
+	/* Magic required to prevent forcewake errors! */
+	err = engine->emit_flush(rq, EMIT_INVALIDATE);
+	if (err)
+		return err;
+
+	cs = intel_ring_begin(rq, 4 * GEN8_3LVL_PDPES + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Ensure the LRI have landed before we invalidate & continue */
+	*cs++ = MI_LOAD_REGISTER_IMM(2 * GEN8_3LVL_PDPES) | MI_LRI_FORCE_POSTED;
+	for (i = GEN8_3LVL_PDPES; i--; ) {
+		const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
+		u32 base = engine->mmio_base;
+
+		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, i));
+		*cs++ = upper_32_bits(pd_daddr);
+		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, i));
+		*cs++ = lower_32_bits(pd_daddr);
+	}
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	intel_ring_advance(rq, cs);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int execlists_request_alloc(struct i915_request *request)
+{
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_pinned(request->context));
+
+	/*
+	 * Flush enough space to reduce the likelihood of waiting after
+	 * we start building the request - in which case we will just
+	 * have to repeat work.
+	 */
+	request->reserved_space += EXECLISTS_REQUEST_SIZE;
+
+	/*
+	 * Note that after this point, we have committed to using
+	 * this request as it is being used to both track the
+	 * state of engine initialisation and liveness of the
+	 * golden renderstate above. Think twice before you try
+	 * to cancel/unwind this request now.
+	 */
+
+	if (!i915_vm_is_4lvl(request->context->vm)) {
+		ret = emit_pdps(request);
+		if (ret)
+			return ret;
+	}
+
+	/* Unconditionally invalidate GPU caches and TLBs. */
+	ret = request->engine->emit_flush(request, EMIT_INVALIDATE);
+	if (ret)
+		return ret;
+
+	request->reserved_space -= EXECLISTS_REQUEST_SIZE;
+	return 0;
+}
+
+static void reset_csb_pointers(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	const unsigned int reset_value = execlists->csb_size - 1;
+
+	ring_set_paused(engine, 0);
+
+	/*
+	 * Sometimes Icelake forgets to reset its pointers on a GPU reset.
+	 * Bludgeon them with a mmio update to be sure.
+	 */
+	ENGINE_WRITE(engine, RING_CONTEXT_STATUS_PTR,
+		     0xffff << 16 | reset_value << 8 | reset_value);
+	ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
+
+	/*
+	 * After a reset, the HW starts writing into CSB entry [0]. We
+	 * therefore have to set our HEAD pointer back one entry so that
+	 * the *first* entry we check is entry 0. To complicate this further,
+	 * as we don't wait for the first interrupt after reset, we have to
+	 * fake the HW write to point back to the last entry so that our
+	 * inline comparison of our cached head position against the last HW
+	 * write works even before the first interrupt.
+	 */
+	execlists->csb_head = reset_value;
+	WRITE_ONCE(*execlists->csb_write, reset_value);
+	wmb(); /* Make sure this is visible to HW (paranoia?) */
+
+	/* Check that the GPU does indeed update the CSB entries! */
+	memset(execlists->csb_status, -1, (reset_value + 1) * sizeof(u64));
+	invalidate_csb_entries(&execlists->csb_status[0],
+			       &execlists->csb_status[reset_value]);
+
+	/* Once more for luck and our trusty paranoia */
+	ENGINE_WRITE(engine, RING_CONTEXT_STATUS_PTR,
+		     0xffff << 16 | reset_value << 8 | reset_value);
+	ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
+
+	GEM_BUG_ON(READ_ONCE(*execlists->csb_write) != reset_value);
+}
+
+static void sanitize_hwsp(struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl;
+
+	list_for_each_entry(tl, &engine->status_page.timelines, engine_link)
+		intel_timeline_reset_seqno(tl);
+}
+
+static void execlists_sanitize(struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(execlists_active(&engine->execlists));
+
+	/*
+	 * Poison residual state on resume, in case the suspend didn't!
+	 *
+	 * We have to assume that across suspend/resume (or other loss
+	 * of control) that the contents of our pinned buffers has been
+	 * lost, replaced by garbage. Since this doesn't always happen,
+	 * let's poison such state so that we more quickly spot when
+	 * we falsely assume it has been preserved.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
+
+	reset_csb_pointers(engine);
+
+	/*
+	 * The kernel_context HWSP is stored in the status_page. As above,
+	 * that may be lost on resume/initialisation, and so we need to
+	 * reset the value in the HWSP.
+	 */
+	sanitize_hwsp(engine);
+
+	/* And scrub the dirty cachelines for the HWSP */
+	clflush_cache_range(engine->status_page.addr, PAGE_SIZE);
+}
+
+static void enable_error_interrupt(struct intel_engine_cs *engine)
+{
+	u32 status;
+
+	engine->execlists.error_interrupt = 0;
+	ENGINE_WRITE(engine, RING_EMR, ~0u);
+	ENGINE_WRITE(engine, RING_EIR, ~0u); /* clear all existing errors */
+
+	status = ENGINE_READ(engine, RING_ESR);
+	if (unlikely(status)) {
+		drm_err(&engine->i915->drm,
+			"engine '%s' resumed still in error: %08x\n",
+			engine->name, status);
+		__intel_gt_reset(engine->gt, engine->mask);
+	}
+
+	/*
+	 * On current gen8+, we have 2 signals to play with
+	 *
+	 * - I915_ERROR_INSTUCTION (bit 0)
+	 *
+	 *    Generate an error if the command parser encounters an invalid
+	 *    instruction
+	 *
+	 *    This is a fatal error.
+	 *
+	 * - CP_PRIV (bit 2)
+	 *
+	 *    Generate an error on privilege violation (where the CP replaces
+	 *    the instruction with a no-op). This also fires for writes into
+	 *    read-only scratch pages.
+	 *
+	 *    This is a non-fatal error, parsing continues.
+	 *
+	 * * there are a few others defined for odd HW that we do not use
+	 *
+	 * Since CP_PRIV fires for cases where we have chosen to ignore the
+	 * error (as the HW is validating and suppressing the mistakes), we
+	 * only unmask the instruction error bit.
+	 */
+	ENGINE_WRITE(engine, RING_EMR, ~I915_ERROR_INSTRUCTION);
+}
+
+static void enable_execlists(struct intel_engine_cs *engine)
+{
+	u32 mode;
+
+	assert_forcewakes_active(engine->uncore, FORCEWAKE_ALL);
+
+	intel_engine_set_hwsp_writemask(engine, ~0u); /* HWSTAM */
+
+	if (INTEL_GEN(engine->i915) >= 11)
+		mode = _MASKED_BIT_ENABLE(GEN11_GFX_DISABLE_LEGACY_MODE);
+	else
+		mode = _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE);
+	ENGINE_WRITE_FW(engine, RING_MODE_GEN7, mode);
+
+	ENGINE_WRITE_FW(engine, RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+
+	ENGINE_WRITE_FW(engine,
+			RING_HWS_PGA,
+			i915_ggtt_offset(engine->status_page.vma));
+	ENGINE_POSTING_READ(engine, RING_HWS_PGA);
+
+	enable_error_interrupt(engine);
+}
+
+static bool unexpected_starting_state(struct intel_engine_cs *engine)
+{
+	bool unexpected = false;
+
+	if (ENGINE_READ_FW(engine, RING_MI_MODE) & STOP_RING) {
+		drm_dbg(&engine->i915->drm,
+			"STOP_RING still set in RING_MI_MODE\n");
+		unexpected = true;
+	}
+
+	return unexpected;
+}
+
+static int execlists_resume(struct intel_engine_cs *engine)
+{
+	intel_mocs_init_engine(engine);
+
+	intel_breadcrumbs_reset(engine->breadcrumbs);
+
+	if (GEM_SHOW_DEBUG() && unexpected_starting_state(engine)) {
+		struct drm_printer p = drm_debug_printer(__func__);
+
+		intel_engine_dump(engine, &p, NULL);
+	}
+
+	enable_execlists(engine);
+
+	return 0;
+}
+
+static void execlists_reset_prepare(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+
+	ENGINE_TRACE(engine, "depth<-%d\n",
+		     atomic_read(&execlists->tasklet.count));
+
+	/*
+	 * Prevent request submission to the hardware until we have
+	 * completed the reset in i915_gem_reset_finish(). If a request
+	 * is completed by one engine, it may then queue a request
+	 * to a second via its execlists->tasklet *just* as we are
+	 * calling engine->resume() and also writing the ELSP.
+	 * Turning off the execlists->tasklet until the reset is over
+	 * prevents the race.
+	 */
+	__tasklet_disable_sync_once(&execlists->tasklet);
+	GEM_BUG_ON(!reset_in_progress(execlists));
+
+	/*
+	 * We stop engines, otherwise we might get failed reset and a
+	 * dead gpu (on elk). Also as modern gpu as kbl can suffer
+	 * from system hang if batchbuffer is progressing when
+	 * the reset is issued, regardless of READY_TO_RESET ack.
+	 * Thus assume it is best to stop engines on all gens
+	 * where we have a gpu reset.
+	 *
+	 * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES)
+	 *
+	 * FIXME: Wa for more modern gens needs to be validated
+	 */
+	ring_set_paused(engine, 1);
+	intel_engine_stop_cs(engine);
+
+	engine->execlists.reset_ccid = active_ccid(engine);
+}
+
+static struct i915_request **
+reset_csb(struct intel_engine_cs *engine, struct i915_request **inactive)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+
+	mb(); /* paranoia: read the CSB pointers from after the reset */
+	clflush(execlists->csb_write);
+	mb();
+
+	inactive = process_csb(engine, inactive); /* drain preemption events */
+
+	/* Following the reset, we need to reload the CSB read/write pointers */
+	reset_csb_pointers(engine);
+
+	return inactive;
+}
+
+static void
+execlists_reset_active(struct intel_engine_cs *engine, bool stalled)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	u32 head;
+
+	/*
+	 * Save the currently executing context, even if we completed
+	 * its request, it was still running at the time of the
+	 * reset and will have been clobbered.
+	 */
+	rq = active_context(engine, engine->execlists.reset_ccid);
+	if (!rq)
+		return;
+
+	ce = rq->context;
+	GEM_BUG_ON(!i915_vma_is_pinned(ce->state));
+
+	if (__i915_request_is_complete(rq)) {
+		/* Idle context; tidy up the ring so we can restart afresh */
+		head = intel_ring_wrap(ce->ring, rq->tail);
+		goto out_replay;
+	}
+
+	/* We still have requests in-flight; the engine should be active */
+	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+
+	/* Context has requests still in-flight; it should not be idle! */
+	GEM_BUG_ON(i915_active_is_idle(&ce->active));
+
+	rq = active_request(ce->timeline, rq);
+	head = intel_ring_wrap(ce->ring, rq->head);
+	GEM_BUG_ON(head == ce->ring->tail);
+
+	/*
+	 * If this request hasn't started yet, e.g. it is waiting on a
+	 * semaphore, we need to avoid skipping the request or else we
+	 * break the signaling chain. However, if the context is corrupt
+	 * the request will not restart and we will be stuck with a wedged
+	 * device. It is quite often the case that if we issue a reset
+	 * while the GPU is loading the context image, that the context
+	 * image becomes corrupt.
+	 *
+	 * Otherwise, if we have not started yet, the request should replay
+	 * perfectly and we do not need to flag the result as being erroneous.
+	 */
+	if (!__i915_request_has_started(rq))
+		goto out_replay;
+
+	/*
+	 * If the request was innocent, we leave the request in the ELSP
+	 * and will try to replay it on restarting. The context image may
+	 * have been corrupted by the reset, in which case we may have
+	 * to service a new GPU hang, but more likely we can continue on
+	 * without impact.
+	 *
+	 * If the request was guilty, we presume the context is corrupt
+	 * and have to at least restore the RING register in the context
+	 * image back to the expected values to skip over the guilty request.
+	 */
+	__i915_request_reset(rq, stalled);
+
+	/*
+	 * We want a simple context + ring to execute the breadcrumb update.
+	 * We cannot rely on the context being intact across the GPU hang,
+	 * so clear it and rebuild just what we need for the breadcrumb.
+	 * All pending requests for this context will be zapped, and any
+	 * future request will be after userspace has had the opportunity
+	 * to recreate its own state.
+	 */
+out_replay:
+	ENGINE_TRACE(engine, "replay {head:%04x, tail:%04x}\n",
+		     head, ce->ring->tail);
+	lrc_reset_regs(ce, engine);
+	ce->lrc.lrca = lrc_update_regs(ce, engine, head);
+}
+
+static void execlists_reset_csb(struct intel_engine_cs *engine, bool stalled)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct i915_request *post[2 * EXECLIST_MAX_PORTS];
+	struct i915_request **inactive;
+
+	rcu_read_lock();
+	inactive = reset_csb(engine, post);
+
+	execlists_reset_active(engine, true);
+
+	inactive = cancel_port_requests(execlists, inactive);
+	post_process_csb(post, inactive);
+	rcu_read_unlock();
+}
+
+static void execlists_reset_rewind(struct intel_engine_cs *engine, bool stalled)
+{
+	unsigned long flags;
+
+	ENGINE_TRACE(engine, "\n");
+
+	/* Process the csb, find the guilty context and throw away */
+	execlists_reset_csb(engine, stalled);
+
+	/* Push back any incomplete requests for replay after the reset. */
+	rcu_read_lock();
+	spin_lock_irqsave(&engine->active.lock, flags);
+	__unwind_incomplete_requests(engine);
+	spin_unlock_irqrestore(&engine->active.lock, flags);
+	rcu_read_unlock();
+}
+
+static void nop_submission_tasklet(unsigned long data)
+{
+	struct intel_engine_cs * const engine = (struct intel_engine_cs *)data;
+
+	/* The driver is wedged; don't process any more events. */
+	WRITE_ONCE(engine->execlists.queue_priority_hint, INT_MIN);
+}
+
+static void execlists_reset_cancel(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct i915_request *rq, *rn;
+	struct rb_node *rb;
+	unsigned long flags;
+
+	ENGINE_TRACE(engine, "\n");
+
+	/*
+	 * Before we call engine->cancel_requests(), we should have exclusive
+	 * access to the submission state. This is arranged for us by the
+	 * caller disabling the interrupt generation, the tasklet and other
+	 * threads that may then access the same state, giving us a free hand
+	 * to reset state. However, we still need to let lockdep be aware that
+	 * we know this state may be accessed in hardirq context, so we
+	 * disable the irq around this manipulation and we want to keep
+	 * the spinlock focused on its duties and not accidentally conflate
+	 * coverage to the submission's irq state. (Similarly, although we
+	 * shouldn't need to disable irq around the manipulation of the
+	 * submission's irq state, we also wish to remind ourselves that
+	 * it is irq state.)
+	 */
+	execlists_reset_csb(engine, true);
+
+	rcu_read_lock();
+	spin_lock_irqsave(&engine->active.lock, flags);
+
+	/* Mark all executing requests as skipped. */
+	list_for_each_entry(rq, &engine->active.requests, sched.link)
+		i915_request_mark_eio(rq);
+	intel_engine_signal_breadcrumbs(engine);
+
+	/* Flush the queued requests to the timeline list (for retiring). */
+	while ((rb = rb_first_cached(&execlists->queue))) {
+		struct i915_priolist *p = to_priolist(rb);
+		int i;
+
+		priolist_for_each_request_consume(rq, rn, p, i) {
+			i915_request_mark_eio(rq);
+			__i915_request_submit(rq);
+		}
+
+		rb_erase_cached(&p->node, &execlists->queue);
+		i915_priolist_free(p);
+	}
+
+	/* On-hold requests will be flushed to timeline upon their release */
+	list_for_each_entry(rq, &engine->active.hold, sched.link)
+		i915_request_mark_eio(rq);
+
+	/* Cancel all attached virtual engines */
+	while ((rb = rb_first_cached(&execlists->virtual))) {
+		struct virtual_engine *ve =
+			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
+
+		rb_erase_cached(rb, &execlists->virtual);
+		RB_CLEAR_NODE(rb);
+
+		spin_lock(&ve->base.active.lock);
+		rq = fetch_and_zero(&ve->request);
+		if (rq) {
+			i915_request_mark_eio(rq);
+
+			rq->engine = engine;
+			__i915_request_submit(rq);
+			i915_request_put(rq);
+
+			ve->base.execlists.queue_priority_hint = INT_MIN;
+		}
+		spin_unlock(&ve->base.active.lock);
+	}
+
+	/* Remaining _unready_ requests will be nop'ed when submitted */
+
+	execlists->queue_priority_hint = INT_MIN;
+	execlists->queue = RB_ROOT_CACHED;
+
+	GEM_BUG_ON(__tasklet_is_enabled(&execlists->tasklet));
+	execlists->tasklet.func = nop_submission_tasklet;
+
+	spin_unlock_irqrestore(&engine->active.lock, flags);
+	rcu_read_unlock();
+}
+
+static void execlists_reset_finish(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+
+	/*
+	 * After a GPU reset, we may have requests to replay. Do so now while
+	 * we still have the forcewake to be sure that the GPU is not allowed
+	 * to sleep before we restart and reload a context.
+	 *
+	 * If the GPU reset fails, the engine may still be alive with requests
+	 * inflight. We expect those to complete, or for the device to be
+	 * reset as the next level of recovery, and as a final resort we
+	 * will declare the device wedged.
+	 */
+	GEM_BUG_ON(!reset_in_progress(execlists));
+
+	/* And kick in case we missed a new request submission. */
+	if (__tasklet_enable(&execlists->tasklet))
+		__execlists_kick(execlists);
+
+	ENGINE_TRACE(engine, "depth->%d\n",
+		     atomic_read(&execlists->tasklet.count));
+}
+
+static void gen8_logical_ring_enable_irq(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR,
+		     ~(engine->irq_enable_mask | engine->irq_keep_mask));
+	ENGINE_POSTING_READ(engine, RING_IMR);
+}
+
+static void gen8_logical_ring_disable_irq(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR, ~engine->irq_keep_mask);
+}
+
+static void execlists_park(struct intel_engine_cs *engine)
+{
+	cancel_timer(&engine->execlists.timer);
+	cancel_timer(&engine->execlists.preempt);
+}
+
+static bool can_preempt(struct intel_engine_cs *engine)
+{
+	if (INTEL_GEN(engine->i915) > 8)
+		return true;
+
+	/* GPGPU on bdw requires extra w/a; not implemented */
+	return engine->class != RENDER_CLASS;
+}
+
+static void execlists_set_default_submission(struct intel_engine_cs *engine)
+{
+	engine->submit_request = execlists_submit_request;
+	engine->schedule = i915_schedule;
+	engine->execlists.tasklet.func = execlists_submission_tasklet;
+
+	engine->reset.prepare = execlists_reset_prepare;
+	engine->reset.rewind = execlists_reset_rewind;
+	engine->reset.cancel = execlists_reset_cancel;
+	engine->reset.finish = execlists_reset_finish;
+
+	engine->park = execlists_park;
+	engine->unpark = NULL;
+
+	engine->flags |= I915_ENGINE_SUPPORTS_STATS;
+	if (!intel_vgpu_active(engine->i915)) {
+		engine->flags |= I915_ENGINE_HAS_SEMAPHORES;
+		if (can_preempt(engine)) {
+			engine->flags |= I915_ENGINE_HAS_PREEMPTION;
+			if (IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+				engine->flags |= I915_ENGINE_HAS_TIMESLICES;
+		}
+	}
+
+	if (intel_engine_has_preemption(engine))
+		engine->emit_bb_start = gen8_emit_bb_start;
+	else
+		engine->emit_bb_start = gen8_emit_bb_start_noarb;
+}
+
+static void execlists_shutdown(struct intel_engine_cs *engine)
+{
+	/* Synchronise with residual timers and any softirq they raise */
+	del_timer_sync(&engine->execlists.timer);
+	del_timer_sync(&engine->execlists.preempt);
+	tasklet_kill(&engine->execlists.tasklet);
+}
+
+static void execlists_release(struct intel_engine_cs *engine)
+{
+	engine->sanitize = NULL; /* no longer in control, nothing to sanitize */
+
+	execlists_shutdown(engine);
+
+	intel_engine_cleanup_common(engine);
+	lrc_fini_wa_ctx(engine);
+}
+
+static void
+logical_ring_default_vfuncs(struct intel_engine_cs *engine)
+{
+	/* Default vfuncs which can be overriden by each engine. */
+
+	engine->resume = execlists_resume;
+
+	engine->cops = &execlists_context_ops;
+	engine->request_alloc = execlists_request_alloc;
+
+	engine->emit_flush = gen8_emit_flush_xcs;
+	engine->emit_init_breadcrumb = gen8_emit_init_breadcrumb;
+	engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_xcs;
+	if (INTEL_GEN(engine->i915) >= 12) {
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_xcs;
+		engine->emit_flush = gen12_emit_flush_xcs;
+	}
+	engine->set_default_submission = execlists_set_default_submission;
+
+	if (INTEL_GEN(engine->i915) < 11) {
+		engine->irq_enable = gen8_logical_ring_enable_irq;
+		engine->irq_disable = gen8_logical_ring_disable_irq;
+	} else {
+		/*
+		 * TODO: On Gen11 interrupt masks need to be clear
+		 * to allow C6 entry. Keep interrupts enabled at
+		 * and take the hit of generating extra interrupts
+		 * until a more refined solution exists.
+		 */
+	}
+}
+
+static void logical_ring_default_irqs(struct intel_engine_cs *engine)
+{
+	unsigned int shift = 0;
+
+	if (INTEL_GEN(engine->i915) < 11) {
+		const u8 irq_shifts[] = {
+			[RCS0]  = GEN8_RCS_IRQ_SHIFT,
+			[BCS0]  = GEN8_BCS_IRQ_SHIFT,
+			[VCS0]  = GEN8_VCS0_IRQ_SHIFT,
+			[VCS1]  = GEN8_VCS1_IRQ_SHIFT,
+			[VECS0] = GEN8_VECS_IRQ_SHIFT,
+		};
+
+		shift = irq_shifts[engine->id];
+	}
+
+	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift;
+	engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift;
+	engine->irq_keep_mask |= GT_CS_MASTER_ERROR_INTERRUPT << shift;
+	engine->irq_keep_mask |= GT_WAIT_SEMAPHORE_INTERRUPT << shift;
+}
+
+static void rcs_submission_override(struct intel_engine_cs *engine)
+{
+	switch (INTEL_GEN(engine->i915)) {
+	case 12:
+		engine->emit_flush = gen12_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_rcs;
+		break;
+	case 11:
+		engine->emit_flush = gen11_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen11_emit_fini_breadcrumb_rcs;
+		break;
+	default:
+		engine->emit_flush = gen8_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_rcs;
+		break;
+	}
+}
+
+int intel_execlists_submission_setup(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct drm_i915_private *i915 = engine->i915;
+	struct intel_uncore *uncore = engine->uncore;
+	u32 base = engine->mmio_base;
+
+	tasklet_init(&engine->execlists.tasklet,
+		     execlists_submission_tasklet, (unsigned long)engine);
+	timer_setup(&engine->execlists.timer, execlists_timeslice, 0);
+	timer_setup(&engine->execlists.preempt, execlists_preempt, 0);
+
+	logical_ring_default_vfuncs(engine);
+	logical_ring_default_irqs(engine);
+
+	if (engine->class == RENDER_CLASS)
+		rcs_submission_override(engine);
+
+	lrc_init_wa_ctx(engine);
+
+	if (HAS_LOGICAL_RING_ELSQ(i915)) {
+		execlists->submit_reg = uncore->regs +
+			i915_mmio_reg_offset(RING_EXECLIST_SQ_CONTENTS(base));
+		execlists->ctrl_reg = uncore->regs +
+			i915_mmio_reg_offset(RING_EXECLIST_CONTROL(base));
+	} else {
+		execlists->submit_reg = uncore->regs +
+			i915_mmio_reg_offset(RING_ELSP(base));
+	}
+
+	execlists->csb_status =
+		(u64 *)&engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX];
+
+	execlists->csb_write =
+		&engine->status_page.addr[intel_hws_csb_write_index(i915)];
+
+	if (INTEL_GEN(i915) < 11)
+		execlists->csb_size = GEN8_CSB_ENTRIES;
+	else
+		execlists->csb_size = GEN11_CSB_ENTRIES;
+
+	engine->context_tag = GENMASK(BITS_PER_LONG - 2, 0);
+	if (INTEL_GEN(engine->i915) >= 11) {
+		execlists->ccid |= engine->instance << (GEN11_ENGINE_INSTANCE_SHIFT - 32);
+		execlists->ccid |= engine->class << (GEN11_ENGINE_CLASS_SHIFT - 32);
+	}
+
+	/* Finally, take ownership and responsibility for cleanup! */
+	engine->sanitize = execlists_sanitize;
+	engine->release = execlists_release;
+
+	return 0;
+}
+
+static struct list_head *virtual_queue(struct virtual_engine *ve)
+{
+	return &ve->base.execlists.default_priolist.requests[0];
+}
+
+static void rcu_virtual_context_destroy(struct work_struct *wrk)
+{
+	struct virtual_engine *ve =
+		container_of(wrk, typeof(*ve), rcu.work);
+	unsigned int n;
+
+	GEM_BUG_ON(ve->context.inflight);
+
+	/* Preempt-to-busy may leave a stale request behind. */
+	if (unlikely(ve->request)) {
+		struct i915_request *old;
+
+		spin_lock_irq(&ve->base.active.lock);
+
+		old = fetch_and_zero(&ve->request);
+		if (old) {
+			GEM_BUG_ON(!__i915_request_is_complete(old));
+			__i915_request_submit(old);
+			i915_request_put(old);
+		}
+
+		spin_unlock_irq(&ve->base.active.lock);
+	}
+
+	/*
+	 * Flush the tasklet in case it is still running on another core.
+	 *
+	 * This needs to be done before we remove ourselves from the siblings'
+	 * rbtrees as in the case it is running in parallel, it may reinsert
+	 * the rb_node into a sibling.
+	 */
+	tasklet_kill(&ve->base.execlists.tasklet);
+
+	/* Decouple ourselves from the siblings, no more access allowed. */
+	for (n = 0; n < ve->num_siblings; n++) {
+		struct intel_engine_cs *sibling = ve->siblings[n];
+		struct rb_node *node = &ve->nodes[sibling->id].rb;
+
+		if (RB_EMPTY_NODE(node))
+			continue;
+
+		spin_lock_irq(&sibling->active.lock);
+
+		/* Detachment is lazily performed in the execlists tasklet */
+		if (!RB_EMPTY_NODE(node))
+			rb_erase_cached(node, &sibling->execlists.virtual);
+
+		spin_unlock_irq(&sibling->active.lock);
+	}
+	GEM_BUG_ON(__tasklet_is_scheduled(&ve->base.execlists.tasklet));
+	GEM_BUG_ON(!list_empty(virtual_queue(ve)));
+
+	lrc_fini(&ve->context);
+	intel_context_fini(&ve->context);
+
+	intel_breadcrumbs_free(ve->base.breadcrumbs);
+	intel_engine_free_request_pool(&ve->base);
+
+	kfree(ve->bonds);
+	kfree(ve);
+}
+
+static void virtual_context_destroy(struct kref *kref)
+{
+	struct virtual_engine *ve =
+		container_of(kref, typeof(*ve), context.ref);
+
+	GEM_BUG_ON(!list_empty(&ve->context.signals));
+
+	/*
+	 * When destroying the virtual engine, we have to be aware that
+	 * it may still be in use from an hardirq/softirq context causing
+	 * the resubmission of a completed request (background completion
+	 * due to preempt-to-busy). Before we can free the engine, we need
+	 * to flush the submission code and tasklets that are still potentially
+	 * accessing the engine. Flushing the tasklets requires process context,
+	 * and since we can guard the resubmit onto the engine with an RCU read
+	 * lock, we can delegate the free of the engine to an RCU worker.
+	 */
+	INIT_RCU_WORK(&ve->rcu, rcu_virtual_context_destroy);
+	queue_rcu_work(system_wq, &ve->rcu);
+}
+
+static void virtual_engine_initial_hint(struct virtual_engine *ve)
+{
+	int swp;
+
+	/*
+	 * Pick a random sibling on starting to help spread the load around.
+	 *
+	 * New contexts are typically created with exactly the same order
+	 * of siblings, and often started in batches. Due to the way we iterate
+	 * the array of sibling when submitting requests, sibling[0] is
+	 * prioritised for dequeuing. If we make sure that sibling[0] is fairly
+	 * randomised across the system, we also help spread the load by the
+	 * first engine we inspect being different each time.
+	 *
+	 * NB This does not force us to execute on this engine, it will just
+	 * typically be the first we inspect for submission.
+	 */
+	swp = prandom_u32_max(ve->num_siblings);
+	if (swp)
+		swap(ve->siblings[swp], ve->siblings[0]);
+}
+
+static int virtual_context_alloc(struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+
+	return lrc_alloc(ce, ve->siblings[0]);
+}
+
+static int virtual_context_pre_pin(struct intel_context *ce,
+				     struct i915_gem_ww_ctx *ww,
+				     void **vaddr)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+
+	/* Note: we must use a real engine class for setting up reg state */
+	return lrc_pre_pin(ce, ve->siblings[0], ww, vaddr);
+}
+
+static int virtual_context_pin(struct intel_context *ce, void *vaddr)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+
+	return lrc_pin(ce, ve->siblings[0], vaddr);
+}
+
+static void virtual_context_enter(struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+	unsigned int n;
+
+	for (n = 0; n < ve->num_siblings; n++)
+		intel_engine_pm_get(ve->siblings[n]);
+
+	intel_timeline_enter(ce->timeline);
+}
+
+static void virtual_context_exit(struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+	unsigned int n;
+
+	intel_timeline_exit(ce->timeline);
+
+	for (n = 0; n < ve->num_siblings; n++)
+		intel_engine_pm_put(ve->siblings[n]);
+}
+
+static const struct intel_context_ops virtual_context_ops = {
+	.flags = COPS_HAS_INFLIGHT,
+
+	.alloc = virtual_context_alloc,
+
+	.pre_pin = virtual_context_pre_pin,
+	.pin = virtual_context_pin,
+	.unpin = lrc_unpin,
+	.post_unpin = lrc_post_unpin,
+
+	.enter = virtual_context_enter,
+	.exit = virtual_context_exit,
+
+	.destroy = virtual_context_destroy,
+};
+
+static intel_engine_mask_t virtual_submission_mask(struct virtual_engine *ve)
+{
+	struct i915_request *rq;
+	intel_engine_mask_t mask;
+
+	rq = READ_ONCE(ve->request);
+	if (!rq)
+		return 0;
+
+	/* The rq is ready for submission; rq->execution_mask is now stable. */
+	mask = rq->execution_mask;
+	if (unlikely(!mask)) {
+		/* Invalid selection, submit to a random engine in error */
+		i915_request_set_error_once(rq, -ENODEV);
+		mask = ve->siblings[0]->mask;
+	}
+
+	ENGINE_TRACE(&ve->base, "rq=%llx:%lld, mask=%x, prio=%d\n",
+		     rq->fence.context, rq->fence.seqno,
+		     mask, ve->base.execlists.queue_priority_hint);
+
+	return mask;
+}
+
+static void virtual_submission_tasklet(unsigned long data)
+{
+	struct virtual_engine * const ve = (struct virtual_engine *)data;
+	const int prio = READ_ONCE(ve->base.execlists.queue_priority_hint);
+	intel_engine_mask_t mask;
+	unsigned int n;
+
+	rcu_read_lock();
+	mask = virtual_submission_mask(ve);
+	rcu_read_unlock();
+	if (unlikely(!mask))
+		return;
+
+	for (n = 0; n < ve->num_siblings; n++) {
+		struct intel_engine_cs *sibling = READ_ONCE(ve->siblings[n]);
+		struct ve_node * const node = &ve->nodes[sibling->id];
+		struct rb_node **parent, *rb;
+		bool first;
+
+		if (!READ_ONCE(ve->request))
+			break; /* already handled by a sibling's tasklet */
+
+		spin_lock_irq(&sibling->active.lock);
+
+		if (unlikely(!(mask & sibling->mask))) {
+			if (!RB_EMPTY_NODE(&node->rb)) {
+				rb_erase_cached(&node->rb,
+						&sibling->execlists.virtual);
+				RB_CLEAR_NODE(&node->rb);
+			}
+
+			goto unlock_engine;
+		}
+
+		if (unlikely(!RB_EMPTY_NODE(&node->rb))) {
+			/*
+			 * Cheat and avoid rebalancing the tree if we can
+			 * reuse this node in situ.
+			 */
+			first = rb_first_cached(&sibling->execlists.virtual) ==
+				&node->rb;
+			if (prio == node->prio || (prio > node->prio && first))
+				goto submit_engine;
+
+			rb_erase_cached(&node->rb, &sibling->execlists.virtual);
+		}
+
+		rb = NULL;
+		first = true;
+		parent = &sibling->execlists.virtual.rb_root.rb_node;
+		while (*parent) {
+			struct ve_node *other;
+
+			rb = *parent;
+			other = rb_entry(rb, typeof(*other), rb);
+			if (prio > other->prio) {
+				parent = &rb->rb_left;
+			} else {
+				parent = &rb->rb_right;
+				first = false;
+			}
+		}
+
+		rb_link_node(&node->rb, rb, parent);
+		rb_insert_color_cached(&node->rb,
+				       &sibling->execlists.virtual,
+				       first);
+
+submit_engine:
+		GEM_BUG_ON(RB_EMPTY_NODE(&node->rb));
+		node->prio = prio;
+		if (first && prio > sibling->execlists.queue_priority_hint)
+			tasklet_hi_schedule(&sibling->execlists.tasklet);
+
+unlock_engine:
+		spin_unlock_irq(&sibling->active.lock);
+
+		if (intel_context_inflight(&ve->context))
+			break;
+	}
+}
+
+static void virtual_submit_request(struct i915_request *rq)
+{
+	struct virtual_engine *ve = to_virtual_engine(rq->engine);
+	unsigned long flags;
+
+	ENGINE_TRACE(&ve->base, "rq=%llx:%lld\n",
+		     rq->fence.context,
+		     rq->fence.seqno);
+
+	GEM_BUG_ON(ve->base.submit_request != virtual_submit_request);
+
+	spin_lock_irqsave(&ve->base.active.lock, flags);
+
+	/* By the time we resubmit a request, it may be completed */
+	if (__i915_request_is_complete(rq)) {
+		__i915_request_submit(rq);
+		goto unlock;
+	}
+
+	if (ve->request) { /* background completion from preempt-to-busy */
+		GEM_BUG_ON(!__i915_request_is_complete(ve->request));
+		__i915_request_submit(ve->request);
+		i915_request_put(ve->request);
+	}
+
+	ve->base.execlists.queue_priority_hint = rq_prio(rq);
+	ve->request = i915_request_get(rq);
+
+	GEM_BUG_ON(!list_empty(virtual_queue(ve)));
+	list_move_tail(&rq->sched.link, virtual_queue(ve));
+
+	tasklet_hi_schedule(&ve->base.execlists.tasklet);
+
+unlock:
+	spin_unlock_irqrestore(&ve->base.active.lock, flags);
+}
+
+static struct ve_bond *
+virtual_find_bond(struct virtual_engine *ve,
+		  const struct intel_engine_cs *master)
+{
+	int i;
+
+	for (i = 0; i < ve->num_bonds; i++) {
+		if (ve->bonds[i].master == master)
+			return &ve->bonds[i];
+	}
+
+	return NULL;
+}
+
+static void
+virtual_bond_execute(struct i915_request *rq, struct dma_fence *signal)
+{
+	struct virtual_engine *ve = to_virtual_engine(rq->engine);
+	intel_engine_mask_t allowed, exec;
+	struct ve_bond *bond;
+
+	allowed = ~to_request(signal)->engine->mask;
+
+	bond = virtual_find_bond(ve, to_request(signal)->engine);
+	if (bond)
+		allowed &= bond->sibling_mask;
+
+	/* Restrict the bonded request to run on only the available engines */
+	exec = READ_ONCE(rq->execution_mask);
+	while (!try_cmpxchg(&rq->execution_mask, &exec, exec & allowed))
+		;
+
+	/* Prevent the master from being re-run on the bonded engines */
+	to_request(signal)->execution_mask &= ~allowed;
+}
+
+struct intel_context *
+intel_execlists_create_virtual(struct intel_engine_cs **siblings,
+			       unsigned int count)
+{
+	struct virtual_engine *ve;
+	unsigned int n;
+	int err;
+
+	if (count == 0)
+		return ERR_PTR(-EINVAL);
+
+	if (count == 1)
+		return intel_context_create(siblings[0]);
+
+	ve = kzalloc(struct_size(ve, siblings, count), GFP_KERNEL);
+	if (!ve)
+		return ERR_PTR(-ENOMEM);
+
+	ve->base.i915 = siblings[0]->i915;
+	ve->base.gt = siblings[0]->gt;
+	ve->base.uncore = siblings[0]->uncore;
+	ve->base.id = -1;
+
+	ve->base.class = OTHER_CLASS;
+	ve->base.uabi_class = I915_ENGINE_CLASS_INVALID;
+	ve->base.instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
+	ve->base.uabi_instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
+
+	/*
+	 * The decision on whether to submit a request using semaphores
+	 * depends on the saturated state of the engine. We only compute
+	 * this during HW submission of the request, and we need for this
+	 * state to be globally applied to all requests being submitted
+	 * to this engine. Virtual engines encompass more than one physical
+	 * engine and so we cannot accurately tell in advance if one of those
+	 * engines is already saturated and so cannot afford to use a semaphore
+	 * and be pessimized in priority for doing so -- if we are the only
+	 * context using semaphores after all other clients have stopped, we
+	 * will be starved on the saturated system. Such a global switch for
+	 * semaphores is less than ideal, but alas is the current compromise.
+	 */
+	ve->base.saturated = ALL_ENGINES;
+
+	snprintf(ve->base.name, sizeof(ve->base.name), "virtual");
+
+	intel_engine_init_active(&ve->base, ENGINE_VIRTUAL);
+	intel_engine_init_execlists(&ve->base);
+
+	ve->base.cops = &virtual_context_ops;
+	ve->base.request_alloc = execlists_request_alloc;
+
+	ve->base.schedule = i915_schedule;
+	ve->base.submit_request = virtual_submit_request;
+	ve->base.bond_execute = virtual_bond_execute;
+
+	INIT_LIST_HEAD(virtual_queue(ve));
+	ve->base.execlists.queue_priority_hint = INT_MIN;
+	tasklet_init(&ve->base.execlists.tasklet,
+		     virtual_submission_tasklet,
+		     (unsigned long)ve);
+
+	intel_context_init(&ve->context, &ve->base);
+
+	ve->base.breadcrumbs = intel_breadcrumbs_create(NULL);
+	if (!ve->base.breadcrumbs) {
+		err = -ENOMEM;
+		goto err_put;
+	}
+
+	for (n = 0; n < count; n++) {
+		struct intel_engine_cs *sibling = siblings[n];
+
+		GEM_BUG_ON(!is_power_of_2(sibling->mask));
+		if (sibling->mask & ve->base.mask) {
+			DRM_DEBUG("duplicate %s entry in load balancer\n",
+				  sibling->name);
+			err = -EINVAL;
+			goto err_put;
+		}
+
+		/*
+		 * The virtual engine implementation is tightly coupled to
+		 * the execlists backend -- we push out request directly
+		 * into a tree inside each physical engine. We could support
+		 * layering if we handle cloning of the requests and
+		 * submitting a copy into each backend.
+		 */
+		if (sibling->execlists.tasklet.func !=
+		    execlists_submission_tasklet) {
+			err = -ENODEV;
+			goto err_put;
+		}
+
+		GEM_BUG_ON(RB_EMPTY_NODE(&ve->nodes[sibling->id].rb));
+		RB_CLEAR_NODE(&ve->nodes[sibling->id].rb);
+
+		ve->siblings[ve->num_siblings++] = sibling;
+		ve->base.mask |= sibling->mask;
+
+		/*
+		 * All physical engines must be compatible for their emission
+		 * functions (as we build the instructions during request
+		 * construction and do not alter them before submission
+		 * on the physical engine). We use the engine class as a guide
+		 * here, although that could be refined.
+		 */
+		if (ve->base.class != OTHER_CLASS) {
+			if (ve->base.class != sibling->class) {
+				DRM_DEBUG("invalid mixing of engine class, sibling %d, already %d\n",
+					  sibling->class, ve->base.class);
+				err = -EINVAL;
+				goto err_put;
+			}
+			continue;
+		}
+
+		ve->base.class = sibling->class;
+		ve->base.uabi_class = sibling->uabi_class;
+		snprintf(ve->base.name, sizeof(ve->base.name),
+			 "v%dx%d", ve->base.class, count);
+		ve->base.context_size = sibling->context_size;
+
+		ve->base.emit_bb_start = sibling->emit_bb_start;
+		ve->base.emit_flush = sibling->emit_flush;
+		ve->base.emit_init_breadcrumb = sibling->emit_init_breadcrumb;
+		ve->base.emit_fini_breadcrumb = sibling->emit_fini_breadcrumb;
+		ve->base.emit_fini_breadcrumb_dw =
+			sibling->emit_fini_breadcrumb_dw;
+
+		ve->base.flags = sibling->flags;
+	}
+
+	ve->base.flags |= I915_ENGINE_IS_VIRTUAL;
+
+	virtual_engine_initial_hint(ve);
+	return &ve->context;
+
+err_put:
+	intel_context_put(&ve->context);
+	return ERR_PTR(err);
+}
+
+struct intel_context *
+intel_execlists_clone_virtual(struct intel_engine_cs *src)
+{
+	struct virtual_engine *se = to_virtual_engine(src);
+	struct intel_context *dst;
+
+	dst = intel_execlists_create_virtual(se->siblings,
+					     se->num_siblings);
+	if (IS_ERR(dst))
+		return dst;
+
+	if (se->num_bonds) {
+		struct virtual_engine *de = to_virtual_engine(dst->engine);
+
+		de->bonds = kmemdup(se->bonds,
+				    sizeof(*se->bonds) * se->num_bonds,
+				    GFP_KERNEL);
+		if (!de->bonds) {
+			intel_context_put(dst);
+			return ERR_PTR(-ENOMEM);
+		}
+
+		de->num_bonds = se->num_bonds;
+	}
+
+	return dst;
+}
+
+int intel_virtual_engine_attach_bond(struct intel_engine_cs *engine,
+				     const struct intel_engine_cs *master,
+				     const struct intel_engine_cs *sibling)
+{
+	struct virtual_engine *ve = to_virtual_engine(engine);
+	struct ve_bond *bond;
+	int n;
+
+	/* Sanity check the sibling is part of the virtual engine */
+	for (n = 0; n < ve->num_siblings; n++)
+		if (sibling == ve->siblings[n])
+			break;
+	if (n == ve->num_siblings)
+		return -EINVAL;
+
+	bond = virtual_find_bond(ve, master);
+	if (bond) {
+		bond->sibling_mask |= sibling->mask;
+		return 0;
+	}
+
+	bond = krealloc(ve->bonds,
+			sizeof(*bond) * (ve->num_bonds + 1),
+			GFP_KERNEL);
+	if (!bond)
+		return -ENOMEM;
+
+	bond[ve->num_bonds].master = master;
+	bond[ve->num_bonds].sibling_mask = sibling->mask;
+
+	ve->bonds = bond;
+	ve->num_bonds++;
+
+	return 0;
+}
+
+void intel_execlists_show_requests(struct intel_engine_cs *engine,
+				   struct drm_printer *m,
+				   void (*show_request)(struct drm_printer *m,
+							const struct i915_request *rq,
+							const char *prefix,
+							int indent),
+				   unsigned int max)
+{
+	const struct intel_engine_execlists *execlists = &engine->execlists;
+	struct i915_request *rq, *last;
+	unsigned long flags;
+	unsigned int count;
+	struct rb_node *rb;
+
+	spin_lock_irqsave(&engine->active.lock, flags);
+
+	last = NULL;
+	count = 0;
+	list_for_each_entry(rq, &engine->active.requests, sched.link) {
+		if (count++ < max - 1)
+			show_request(m, rq, "\t\t", 0);
+		else
+			last = rq;
+	}
+	if (last) {
+		if (count > max) {
+			drm_printf(m,
+				   "\t\t...skipping %d executing requests...\n",
+				   count - max);
+		}
+		show_request(m, last, "\t\t", 0);
+	}
+
+	if (execlists->queue_priority_hint != INT_MIN)
+		drm_printf(m, "\t\tQueue priority hint: %d\n",
+			   READ_ONCE(execlists->queue_priority_hint));
+
+	last = NULL;
+	count = 0;
+	for (rb = rb_first_cached(&execlists->queue); rb; rb = rb_next(rb)) {
+		struct i915_priolist *p = rb_entry(rb, typeof(*p), node);
+		int i;
+
+		priolist_for_each_request(rq, p, i) {
+			if (count++ < max - 1)
+				show_request(m, rq, "\t\t", 0);
+			else
+				last = rq;
+		}
+	}
+	if (last) {
+		if (count > max) {
+			drm_printf(m,
+				   "\t\t...skipping %d queued requests...\n",
+				   count - max);
+		}
+		show_request(m, last, "\t\t", 0);
+	}
+
+	last = NULL;
+	count = 0;
+	for (rb = rb_first_cached(&execlists->virtual); rb; rb = rb_next(rb)) {
+		struct virtual_engine *ve =
+			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
+		struct i915_request *rq = READ_ONCE(ve->request);
+
+		if (rq) {
+			if (count++ < max - 1)
+				show_request(m, rq, "\t\t", 0);
+			else
+				last = rq;
+		}
+	}
+	if (last) {
+		if (count > max) {
+			drm_printf(m,
+				   "\t\t...skipping %d virtual requests...\n",
+				   count - max);
+		}
+		show_request(m, last, "\t\t", 0);
+	}
+
+	spin_unlock_irqrestore(&engine->active.lock, flags);
+}
+
+bool
+intel_engine_in_execlists_submission_mode(const struct intel_engine_cs *engine)
+{
+	return engine->set_default_submission ==
+	       execlists_set_default_submission;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_execlists.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.h b/drivers/gpu/drm/i915/gt/intel_execlists_submission.h
new file mode 100644
index 000000000000..a8fd7adefd82
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.h
@@ -0,0 +1,47 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#ifndef __INTEL_EXECLISTS_SUBMISSION_H__
+#define __INTEL_EXECLISTS_SUBMISSION_H__
+
+#include <linux/types.h>
+
+struct drm_printer;
+
+struct i915_request;
+struct intel_context;
+struct intel_engine_cs;
+
+enum {
+	INTEL_CONTEXT_SCHEDULE_IN = 0,
+	INTEL_CONTEXT_SCHEDULE_OUT,
+	INTEL_CONTEXT_SCHEDULE_PREEMPTED,
+};
+
+int intel_execlists_submission_setup(struct intel_engine_cs *engine);
+
+void intel_execlists_show_requests(struct intel_engine_cs *engine,
+				   struct drm_printer *m,
+				   void (*show_request)(struct drm_printer *m,
+							const struct i915_request *rq,
+							const char *prefix,
+							int indent),
+				   unsigned int max);
+
+struct intel_context *
+intel_execlists_create_virtual(struct intel_engine_cs **siblings,
+			       unsigned int count);
+
+struct intel_context *
+intel_execlists_clone_virtual(struct intel_engine_cs *src);
+
+int intel_virtual_engine_attach_bond(struct intel_engine_cs *engine,
+				     const struct intel_engine_cs *master,
+				     const struct intel_engine_cs *sibling);
+
+bool
+intel_engine_in_execlists_submission_mode(const struct intel_engine_cs *engine);
+
+#endif /* __INTEL_EXECLISTS_SUBMISSION_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c
index cf94525be2c1..700588bc9d57 100644
--- a/drivers/gpu/drm/i915/gt/intel_ggtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c
@@ -101,7 +101,16 @@ static bool needs_idle_maps(struct drm_i915_private *i915)
 	 * Query intel_iommu to see if we need the workaround. Presumably that
 	 * was loaded first.
 	 */
-	return IS_GEN(i915, 5) && IS_MOBILE(i915) && intel_vtd_active();
+	if (!intel_vtd_active())
+		return false;
+
+	if (IS_GEN(i915, 5) && IS_MOBILE(i915))
+		return true;
+
+	if (IS_GEN(i915, 12))
+		return true; /* XXX DMAR fault reason 7 */
+
+	return false;
 }
 
 void i915_ggtt_suspend(struct i915_ggtt *ggtt)
@@ -526,16 +535,39 @@ static int init_ggtt(struct i915_ggtt *ggtt)
 
 	mutex_init(&ggtt->error_mutex);
 	if (ggtt->mappable_end) {
-		/* Reserve a mappable slot for our lockless error capture */
-		ret = drm_mm_insert_node_in_range(&ggtt->vm.mm,
-						  &ggtt->error_capture,
-						  PAGE_SIZE, 0,
-						  I915_COLOR_UNEVICTABLE,
-						  0, ggtt->mappable_end,
-						  DRM_MM_INSERT_LOW);
-		if (ret)
-			return ret;
+		/*
+		 * Reserve a mappable slot for our lockless error capture.
+		 *
+		 * We strongly prefer taking address 0x0 in order to protect
+		 * other critical buffers against accidental overwrites,
+		 * as writing to address 0 is a very common mistake.
+		 *
+		 * Since 0 may already be in use by the system (e.g. the BIOS
+		 * framebuffer), we let the reservation fail quietly and hope
+		 * 0 remains reserved always.
+		 *
+		 * If we fail to reserve 0, and then fail to find any space
+		 * for an error-capture, remain silent. We can afford not
+		 * to reserve an error_capture node as we have fallback
+		 * paths, and we trust that 0 will remain reserved. However,
+		 * the only likely reason for failure to insert is a driver
+		 * bug, which we expect to cause other failures...
+		 */
+		ggtt->error_capture.size = I915_GTT_PAGE_SIZE;
+		ggtt->error_capture.color = I915_COLOR_UNEVICTABLE;
+		if (drm_mm_reserve_node(&ggtt->vm.mm, &ggtt->error_capture))
+			drm_mm_insert_node_in_range(&ggtt->vm.mm,
+						    &ggtt->error_capture,
+						    ggtt->error_capture.size, 0,
+						    ggtt->error_capture.color,
+						    0, ggtt->mappable_end,
+						    DRM_MM_INSERT_LOW);
 	}
+	if (drm_mm_node_allocated(&ggtt->error_capture))
+		drm_dbg(&ggtt->vm.i915->drm,
+			"Reserved GGTT:[%llx, %llx] for use by error capture\n",
+			ggtt->error_capture.start,
+			ggtt->error_capture.start + ggtt->error_capture.size);
 
 	/*
 	 * The upper portion of the GuC address space has a sizeable hole
@@ -548,9 +580,9 @@ static int init_ggtt(struct i915_ggtt *ggtt)
 
 	/* Clear any non-preallocated blocks */
 	drm_mm_for_each_hole(entry, &ggtt->vm.mm, hole_start, hole_end) {
-		drm_dbg_kms(&ggtt->vm.i915->drm,
-			    "clearing unused GTT space: [%lx, %lx]\n",
-			    hole_start, hole_end);
+		drm_dbg(&ggtt->vm.i915->drm,
+			"clearing unused GTT space: [%lx, %lx]\n",
+			hole_start, hole_end);
 		ggtt->vm.clear_range(&ggtt->vm, hole_start,
 				     hole_end - hole_start);
 	}
@@ -1050,7 +1082,12 @@ static int i915_gmch_probe(struct i915_ggtt *ggtt)
 
 	ggtt->vm.alloc_pt_dma = alloc_pt_dma;
 
-	ggtt->do_idle_maps = needs_idle_maps(i915);
+	if (needs_idle_maps(i915)) {
+		drm_notice(&i915->drm,
+			   "Flushing DMA requests before IOMMU unmaps; performance may be degraded\n");
+		ggtt->do_idle_maps = true;
+	}
+
 	ggtt->vm.insert_page = i915_ggtt_insert_page;
 	ggtt->vm.insert_entries = i915_ggtt_insert_entries;
 	ggtt->vm.clear_range = i915_ggtt_clear_range;
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
index 7fb36b12fe7a..a357bb431815 100644
--- a/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
@@ -320,13 +320,31 @@ void i915_vma_revoke_fence(struct i915_vma *vma)
 		fence_write(fence);
 }
 
+static bool fence_is_active(const struct i915_fence_reg *fence)
+{
+	return fence->vma && i915_vma_is_active(fence->vma);
+}
+
 static struct i915_fence_reg *fence_find(struct i915_ggtt *ggtt)
 {
-	struct i915_fence_reg *fence;
+	struct i915_fence_reg *active = NULL;
+	struct i915_fence_reg *fence, *fn;
 
-	list_for_each_entry(fence, &ggtt->fence_list, link) {
+	list_for_each_entry_safe(fence, fn, &ggtt->fence_list, link) {
 		GEM_BUG_ON(fence->vma && fence->vma->fence != fence);
 
+		if (fence == active) /* now seen this fence twice */
+			active = ERR_PTR(-EAGAIN);
+
+		/* Prefer idle fences so we do not have to wait on the GPU */
+		if (active != ERR_PTR(-EAGAIN) && fence_is_active(fence)) {
+			if (!active)
+				active = fence;
+
+			list_move_tail(&fence->link, &ggtt->fence_list);
+			continue;
+		}
+
 		if (atomic_read(&fence->pin_count))
 			continue;
 
diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c
index 44f1d51e5ae5..d8e1ab412634 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt.c
@@ -46,6 +46,8 @@ void intel_gt_init_hw_early(struct intel_gt *gt, struct i915_ggtt *ggtt)
 
 int intel_gt_init_mmio(struct intel_gt *gt)
 {
+	intel_gt_init_clock_frequency(gt);
+
 	intel_uc_init_mmio(&gt->uc);
 	intel_sseu_info_init(gt);
 
@@ -546,8 +548,6 @@ int intel_gt_init(struct intel_gt *gt)
 	 */
 	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
 
-	intel_gt_init_clock_frequency(gt);
-
 	err = intel_gt_init_scratch(gt, IS_GEN(gt->i915, 2) ? SZ_256K : SZ_4K);
 	if (err)
 		goto out_fw;
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c
index 104cb30e8c13..06d84cf09570 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c
@@ -145,7 +145,8 @@ static void pool_retire(struct i915_active *ref)
 }
 
 static struct intel_gt_buffer_pool_node *
-node_create(struct intel_gt_buffer_pool *pool, size_t sz)
+node_create(struct intel_gt_buffer_pool *pool, size_t sz,
+	    enum i915_map_type type)
 {
 	struct intel_gt *gt = to_gt(pool);
 	struct intel_gt_buffer_pool_node *node;
@@ -169,12 +170,14 @@ node_create(struct intel_gt_buffer_pool *pool, size_t sz)
 
 	i915_gem_object_set_readonly(obj);
 
+	node->type = type;
 	node->obj = obj;
 	return node;
 }
 
 struct intel_gt_buffer_pool_node *
-intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size)
+intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size,
+			 enum i915_map_type type)
 {
 	struct intel_gt_buffer_pool *pool = &gt->buffer_pool;
 	struct intel_gt_buffer_pool_node *node;
@@ -191,6 +194,9 @@ intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size)
 		if (node->obj->base.size < size)
 			continue;
 
+		if (node->type != type)
+			continue;
+
 		age = READ_ONCE(node->age);
 		if (!age)
 			continue;
@@ -205,7 +211,7 @@ intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size)
 	rcu_read_unlock();
 
 	if (&node->link == list) {
-		node = node_create(pool, size);
+		node = node_create(pool, size, type);
 		if (IS_ERR(node))
 			return node;
 	}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h
index 42cbac003e8a..6068f8f1762e 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h
+++ b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h
@@ -15,7 +15,8 @@ struct intel_gt;
 struct i915_request;
 
 struct intel_gt_buffer_pool_node *
-intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size);
+intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size,
+			 enum i915_map_type type);
 
 static inline int
 intel_gt_buffer_pool_mark_active(struct intel_gt_buffer_pool_node *node,
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h
index bcf1658c9633..d8d82c890da8 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h
@@ -11,10 +11,9 @@
 #include <linux/spinlock.h>
 #include <linux/workqueue.h>
 
+#include "gem/i915_gem_object_types.h"
 #include "i915_active_types.h"
 
-struct drm_i915_gem_object;
-
 struct intel_gt_buffer_pool {
 	spinlock_t lock;
 	struct list_head cache_list[4];
@@ -31,6 +30,7 @@ struct intel_gt_buffer_pool_node {
 		struct rcu_head rcu;
 	};
 	unsigned long age;
+	enum i915_map_type type;
 };
 
 #endif /* INTEL_GT_BUFFER_POOL_TYPES_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c
index 999079686846..a4242ca8dcd7 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c
@@ -7,34 +7,146 @@
 #include "intel_gt.h"
 #include "intel_gt_clock_utils.h"
 
-#define MHZ_12   12000000 /* 12MHz (24MHz/2), 83.333ns */
-#define MHZ_12_5 12500000 /* 12.5MHz (25MHz/2), 80ns */
-#define MHZ_19_2 19200000 /* 19.2MHz, 52.083ns */
+static u32 read_reference_ts_freq(struct intel_uncore *uncore)
+{
+	u32 ts_override = intel_uncore_read(uncore, GEN9_TIMESTAMP_OVERRIDE);
+	u32 base_freq, frac_freq;
+
+	base_freq = ((ts_override & GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_MASK) >>
+		     GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_SHIFT) + 1;
+	base_freq *= 1000000;
+
+	frac_freq = ((ts_override &
+		      GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_MASK) >>
+		     GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_SHIFT);
+	frac_freq = 1000000 / (frac_freq + 1);
+
+	return base_freq + frac_freq;
+}
+
+static u32 gen10_get_crystal_clock_freq(struct intel_uncore *uncore,
+					u32 rpm_config_reg)
+{
+	u32 f19_2_mhz = 19200000;
+	u32 f24_mhz = 24000000;
+	u32 crystal_clock =
+		(rpm_config_reg & GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >>
+		GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT;
 
-static u32 read_clock_frequency(const struct intel_gt *gt)
+	switch (crystal_clock) {
+	case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ:
+		return f19_2_mhz;
+	case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ:
+		return f24_mhz;
+	default:
+		MISSING_CASE(crystal_clock);
+		return 0;
+	}
+}
+
+static u32 gen11_get_crystal_clock_freq(struct intel_uncore *uncore,
+					u32 rpm_config_reg)
 {
-	if (INTEL_GEN(gt->i915) >= 11) {
-		u32 config;
-
-		config = intel_uncore_read(gt->uncore, RPM_CONFIG0);
-		config &= GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK;
-		config >>= GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT;
-
-		switch (config) {
-		case 0: return MHZ_12;
-		case 1:
-		case 2: return MHZ_19_2;
-		default:
-		case 3: return MHZ_12_5;
+	u32 f19_2_mhz = 19200000;
+	u32 f24_mhz = 24000000;
+	u32 f25_mhz = 25000000;
+	u32 f38_4_mhz = 38400000;
+	u32 crystal_clock =
+		(rpm_config_reg & GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >>
+		GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT;
+
+	switch (crystal_clock) {
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ:
+		return f24_mhz;
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ:
+		return f19_2_mhz;
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ:
+		return f38_4_mhz;
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ:
+		return f25_mhz;
+	default:
+		MISSING_CASE(crystal_clock);
+		return 0;
+	}
+}
+
+static u32 read_clock_frequency(struct intel_uncore *uncore)
+{
+	u32 f12_5_mhz = 12500000;
+	u32 f19_2_mhz = 19200000;
+	u32 f24_mhz = 24000000;
+
+	if (INTEL_GEN(uncore->i915) <= 4) {
+		/*
+		 * PRMs say:
+		 *
+		 *     "The value in this register increments once every 16
+		 *      hclks." (through the “Clocking Configuration�
+		 *      (“CLKCFG�) MCHBAR register)
+		 */
+		return RUNTIME_INFO(uncore->i915)->rawclk_freq * 1000 / 16;
+	} else if (INTEL_GEN(uncore->i915) <= 8) {
+		/*
+		 * PRMs say:
+		 *
+		 *     "The PCU TSC counts 10ns increments; this timestamp
+		 *      reflects bits 38:3 of the TSC (i.e. 80ns granularity,
+		 *      rolling over every 1.5 hours).
+		 */
+		return f12_5_mhz;
+	} else if (INTEL_GEN(uncore->i915) <= 9) {
+		u32 ctc_reg = intel_uncore_read(uncore, CTC_MODE);
+		u32 freq = 0;
+
+		if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) {
+			freq = read_reference_ts_freq(uncore);
+		} else {
+			freq = IS_GEN9_LP(uncore->i915) ? f19_2_mhz : f24_mhz;
+
+			/*
+			 * Now figure out how the command stream's timestamp
+			 * register increments from this frequency (it might
+			 * increment only every few clock cycle).
+			 */
+			freq >>= 3 - ((ctc_reg & CTC_SHIFT_PARAMETER_MASK) >>
+				      CTC_SHIFT_PARAMETER_SHIFT);
 		}
-	} else if (INTEL_GEN(gt->i915) >= 9) {
-		if (IS_GEN9_LP(gt->i915))
-			return MHZ_19_2;
-		else
-			return MHZ_12;
-	} else {
-		return MHZ_12_5;
+
+		return freq;
+	} else if (INTEL_GEN(uncore->i915) <= 12) {
+		u32 ctc_reg = intel_uncore_read(uncore, CTC_MODE);
+		u32 freq = 0;
+
+		/*
+		 * First figure out the reference frequency. There are 2 ways
+		 * we can compute the frequency, either through the
+		 * TIMESTAMP_OVERRIDE register or through RPM_CONFIG. CTC_MODE
+		 * tells us which one we should use.
+		 */
+		if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) {
+			freq = read_reference_ts_freq(uncore);
+		} else {
+			u32 c0 = intel_uncore_read(uncore, RPM_CONFIG0);
+
+			if (INTEL_GEN(uncore->i915) <= 10)
+				freq = gen10_get_crystal_clock_freq(uncore, c0);
+			else
+				freq = gen11_get_crystal_clock_freq(uncore, c0);
+
+			/*
+			 * Now figure out how the command stream's timestamp
+			 * register increments from this frequency (it might
+			 * increment only every few clock cycle).
+			 */
+			freq >>= 3 - ((c0 & GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK) >>
+				      GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT);
+		}
+
+		return freq;
 	}
+
+	MISSING_CASE("Unknown gen, unable to read command streamer timestamp frequency\n");
+	return 0;
 }
 
 void intel_gt_init_clock_frequency(struct intel_gt *gt)
@@ -43,20 +155,27 @@ void intel_gt_init_clock_frequency(struct intel_gt *gt)
 	 * Note that on gen11+, the clock frequency may be reconfigured.
 	 * We do not, and we assume nobody else does.
 	 */
-	gt->clock_frequency = read_clock_frequency(gt);
+	gt->clock_frequency = read_clock_frequency(gt->uncore);
+	if (gt->clock_frequency)
+		gt->clock_period_ns = intel_gt_clock_interval_to_ns(gt, 1);
+
 	GT_TRACE(gt,
-		 "Using clock frequency: %dkHz\n",
-		 gt->clock_frequency / 1000);
+		 "Using clock frequency: %dkHz, period: %dns, wrap: %lldms\n",
+		 gt->clock_frequency / 1000,
+		 gt->clock_period_ns,
+		 div_u64(mul_u32_u32(gt->clock_period_ns, S32_MAX),
+			 USEC_PER_SEC));
+
 }
 
 #if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
 void intel_gt_check_clock_frequency(const struct intel_gt *gt)
 {
-	if (gt->clock_frequency != read_clock_frequency(gt)) {
+	if (gt->clock_frequency != read_clock_frequency(gt->uncore)) {
 		dev_err(gt->i915->drm.dev,
 			"GT clock frequency changed, was %uHz, now %uHz!\n",
 			gt->clock_frequency,
-			read_clock_frequency(gt));
+			read_clock_frequency(gt->uncore));
 	}
 }
 #endif
@@ -66,26 +185,24 @@ static u64 div_u64_roundup(u64 nom, u32 den)
 	return div_u64(nom + den - 1, den);
 }
 
-u32 intel_gt_clock_interval_to_ns(const struct intel_gt *gt, u32 count)
+u64 intel_gt_clock_interval_to_ns(const struct intel_gt *gt, u64 count)
 {
-	return div_u64_roundup(mul_u32_u32(count, 1000 * 1000 * 1000),
-			       gt->clock_frequency);
+	return div_u64_roundup(count * NSEC_PER_SEC, gt->clock_frequency);
 }
 
-u32 intel_gt_pm_interval_to_ns(const struct intel_gt *gt, u32 count)
+u64 intel_gt_pm_interval_to_ns(const struct intel_gt *gt, u64 count)
 {
 	return intel_gt_clock_interval_to_ns(gt, 16 * count);
 }
 
-u32 intel_gt_ns_to_clock_interval(const struct intel_gt *gt, u32 ns)
+u64 intel_gt_ns_to_clock_interval(const struct intel_gt *gt, u64 ns)
 {
-	return div_u64_roundup(mul_u32_u32(gt->clock_frequency, ns),
-			       1000 * 1000 * 1000);
+	return div_u64_roundup(gt->clock_frequency * ns, NSEC_PER_SEC);
 }
 
-u32 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u32 ns)
+u64 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u64 ns)
 {
-	u32 val;
+	u64 val;
 
 	/*
 	 * Make these a multiple of magic 25 to avoid SNB (eg. Dell XPS
@@ -94,9 +211,9 @@ u32 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u32 ns)
 	 * EI/thresholds are "bad", leading to a very sluggish or even
 	 * frozen machine.
 	 */
-	val = DIV_ROUND_UP(intel_gt_ns_to_clock_interval(gt, ns), 16);
+	val = div_u64_roundup(intel_gt_ns_to_clock_interval(gt, ns), 16);
 	if (IS_GEN(gt->i915, 6))
-		val = roundup(val, 25);
+		val = div_u64_roundup(val, 25) * 25;
 
 	return val;
 }
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h
index f793c89f2cbd..8b03e97a85df 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h
+++ b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h
@@ -18,10 +18,10 @@ void intel_gt_check_clock_frequency(const struct intel_gt *gt);
 static inline void intel_gt_check_clock_frequency(const struct intel_gt *gt) {}
 #endif
 
-u32 intel_gt_clock_interval_to_ns(const struct intel_gt *gt, u32 count);
-u32 intel_gt_pm_interval_to_ns(const struct intel_gt *gt, u32 count);
+u64 intel_gt_clock_interval_to_ns(const struct intel_gt *gt, u64 count);
+u64 intel_gt_pm_interval_to_ns(const struct intel_gt *gt, u64 count);
 
-u32 intel_gt_ns_to_clock_interval(const struct intel_gt *gt, u32 ns);
-u32 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u32 ns);
+u64 intel_gt_ns_to_clock_interval(const struct intel_gt *gt, u64 ns);
+u64 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u64 ns);
 
 #endif /* __INTEL_GT_CLOCK_UTILS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_irq.c b/drivers/gpu/drm/i915/gt/intel_gt_irq.c
index 257063a57101..9830342aa6f4 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_irq.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt_irq.c
@@ -11,6 +11,7 @@
 #include "intel_breadcrumbs.h"
 #include "intel_gt.h"
 #include "intel_gt_irq.h"
+#include "intel_lrc_reg.h"
 #include "intel_uncore.h"
 #include "intel_rps.h"
 
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm.c b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
index 274aa0dd7050..c94e8ac884eb 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_pm.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
@@ -39,6 +39,28 @@ static void user_forcewake(struct intel_gt *gt, bool suspend)
 	intel_gt_pm_put(gt);
 }
 
+static void runtime_begin(struct intel_gt *gt)
+{
+	local_irq_disable();
+	write_seqcount_begin(&gt->stats.lock);
+	gt->stats.start = ktime_get();
+	gt->stats.active = true;
+	write_seqcount_end(&gt->stats.lock);
+	local_irq_enable();
+}
+
+static void runtime_end(struct intel_gt *gt)
+{
+	local_irq_disable();
+	write_seqcount_begin(&gt->stats.lock);
+	gt->stats.active = false;
+	gt->stats.total =
+		ktime_add(gt->stats.total,
+			  ktime_sub(ktime_get(), gt->stats.start));
+	write_seqcount_end(&gt->stats.lock);
+	local_irq_enable();
+}
+
 static int __gt_unpark(struct intel_wakeref *wf)
 {
 	struct intel_gt *gt = container_of(wf, typeof(*gt), wakeref);
@@ -67,6 +89,7 @@ static int __gt_unpark(struct intel_wakeref *wf)
 	i915_pmu_gt_unparked(i915);
 
 	intel_gt_unpark_requests(gt);
+	runtime_begin(gt);
 
 	return 0;
 }
@@ -79,6 +102,7 @@ static int __gt_park(struct intel_wakeref *wf)
 
 	GT_TRACE(gt, "\n");
 
+	runtime_end(gt);
 	intel_gt_park_requests(gt);
 
 	i915_vma_parked(gt);
@@ -106,6 +130,7 @@ static const struct intel_wakeref_ops wf_ops = {
 void intel_gt_pm_init_early(struct intel_gt *gt)
 {
 	intel_wakeref_init(&gt->wakeref, gt->uncore->rpm, &wf_ops);
+	seqcount_mutex_init(&gt->stats.lock, &gt->wakeref.mutex);
 }
 
 void intel_gt_pm_init(struct intel_gt *gt)
@@ -339,6 +364,30 @@ int intel_gt_runtime_resume(struct intel_gt *gt)
 	return intel_uc_runtime_resume(&gt->uc);
 }
 
+static ktime_t __intel_gt_get_awake_time(const struct intel_gt *gt)
+{
+	ktime_t total = gt->stats.total;
+
+	if (gt->stats.active)
+		total = ktime_add(total,
+				  ktime_sub(ktime_get(), gt->stats.start));
+
+	return total;
+}
+
+ktime_t intel_gt_get_awake_time(const struct intel_gt *gt)
+{
+	unsigned int seq;
+	ktime_t total;
+
+	do {
+		seq = read_seqcount_begin(&gt->stats.lock);
+		total = __intel_gt_get_awake_time(gt);
+	} while (read_seqcount_retry(&gt->stats.lock, seq));
+
+	return total;
+}
+
 #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
 #include "selftest_gt_pm.c"
 #endif
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm.h b/drivers/gpu/drm/i915/gt/intel_gt_pm.h
index 60f0e2fbe55c..63846a856e7e 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_pm.h
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm.h
@@ -58,6 +58,8 @@ int intel_gt_resume(struct intel_gt *gt);
 void intel_gt_runtime_suspend(struct intel_gt *gt);
 int intel_gt_runtime_resume(struct intel_gt *gt);
 
+ktime_t intel_gt_get_awake_time(const struct intel_gt *gt);
+
 static inline bool is_mock_gt(const struct intel_gt *gt)
 {
 	return I915_SELFTEST_ONLY(gt->awake == -ENODEV);
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.c b/drivers/gpu/drm/i915/gt/intel_gt_requests.c
index 66fcbf9d0fdd..dc06c78c9eeb 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_requests.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.c
@@ -135,13 +135,8 @@ long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout)
 	struct intel_gt_timelines *timelines = &gt->timelines;
 	struct intel_timeline *tl, *tn;
 	unsigned long active_count = 0;
-	bool interruptible;
 	LIST_HEAD(free);
 
-	interruptible = true;
-	if (unlikely(timeout < 0))
-		timeout = -timeout, interruptible = false;
-
 	flush_submission(gt, timeout); /* kick the ksoftirqd tasklets */
 	spin_lock(&timelines->lock);
 	list_for_each_entry_safe(tl, tn, &timelines->active_list, link) {
@@ -163,7 +158,7 @@ long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout)
 				mutex_unlock(&tl->mutex);
 
 				timeout = dma_fence_wait_timeout(fence,
-								 interruptible,
+								 true,
 								 timeout);
 				dma_fence_put(fence);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_types.h b/drivers/gpu/drm/i915/gt/intel_gt_types.h
index 6d39a4a11bf3..a83d3e18254d 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_gt_types.h
@@ -75,6 +75,7 @@ struct intel_gt {
 	intel_wakeref_t awake;
 
 	u32 clock_frequency;
+	u32 clock_period_ns;
 
 	struct intel_llc llc;
 	struct intel_rc6 rc6;
@@ -87,6 +88,30 @@ struct intel_gt {
 
 	u32 pm_guc_events;
 
+	struct {
+		bool active;
+
+		/**
+		 * @lock: Lock protecting the below fields.
+		 */
+		seqcount_mutex_t lock;
+
+		/**
+		 * @total: Total time this engine was busy.
+		 *
+		 * Accumulated time not counting the most recent block in cases
+		 * where engine is currently busy (active > 0).
+		 */
+		ktime_t total;
+
+		/**
+		 * @start: Timestamp of the last idle to active transition.
+		 *
+		 * Idle is defined as active == 0, active is active > 0.
+		 */
+		ktime_t start;
+	} stats;
+
 	struct intel_engine_cs *engine[I915_NUM_ENGINES];
 	struct intel_engine_cs *engine_class[MAX_ENGINE_CLASS + 1]
 					    [MAX_ENGINE_INSTANCE + 1];
diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.c b/drivers/gpu/drm/i915/gt/intel_gtt.c
index 7bfe9072be9a..04aa6601e984 100644
--- a/drivers/gpu/drm/i915/gt/intel_gtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_gtt.c
@@ -422,6 +422,35 @@ void setup_private_pat(struct intel_uncore *uncore)
 		bdw_setup_private_ppat(uncore);
 }
 
+struct i915_vma *
+__vm_create_scratch_for_read(struct i915_address_space *vm, unsigned long size)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_internal(vm->i915, PAGE_ALIGN(size));
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	i915_gem_object_set_cache_coherency(obj, I915_CACHING_CACHED);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_vma_pin(vma, 0, 0,
+			   i915_vma_is_ggtt(vma) ? PIN_GLOBAL : PIN_USER);
+	if (err) {
+		i915_vma_put(vma);
+		return ERR_PTR(err);
+	}
+
+	return vma;
+}
+
 #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
 #include "selftests/mock_gtt.c"
 #endif
diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.h b/drivers/gpu/drm/i915/gt/intel_gtt.h
index 8a33940a71f3..29c10fde8ce3 100644
--- a/drivers/gpu/drm/i915/gt/intel_gtt.h
+++ b/drivers/gpu/drm/i915/gt/intel_gtt.h
@@ -573,6 +573,9 @@ int i915_vm_pin_pt_stash(struct i915_address_space *vm,
 void i915_vm_free_pt_stash(struct i915_address_space *vm,
 			   struct i915_vm_pt_stash *stash);
 
+struct i915_vma *
+__vm_create_scratch_for_read(struct i915_address_space *vm, unsigned long size);
+
 static inline struct sgt_dma {
 	struct scatterlist *sg;
 	dma_addr_t dma, max;
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index 7614a3d24fca..94f485b591af 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -1,599 +1,23 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright © 2014 Intel Corporation
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
- * IN THE SOFTWARE.
- *
- * Authors:
- *    Ben Widawsky <ben@bwidawsk.net>
- *    Michel Thierry <michel.thierry@intel.com>
- *    Thomas Daniel <thomas.daniel@intel.com>
- *    Oscar Mateo <oscar.mateo@intel.com>
- *
- */
-
-/**
- * DOC: Logical Rings, Logical Ring Contexts and Execlists
- *
- * Motivation:
- * GEN8 brings an expansion of the HW contexts: "Logical Ring Contexts".
- * These expanded contexts enable a number of new abilities, especially
- * "Execlists" (also implemented in this file).
- *
- * One of the main differences with the legacy HW contexts is that logical
- * ring contexts incorporate many more things to the context's state, like
- * PDPs or ringbuffer control registers:
- *
- * The reason why PDPs are included in the context is straightforward: as
- * PPGTTs (per-process GTTs) are actually per-context, having the PDPs
- * contained there mean you don't need to do a ppgtt->switch_mm yourself,
- * instead, the GPU will do it for you on the context switch.
- *
- * But, what about the ringbuffer control registers (head, tail, etc..)?
- * shouldn't we just need a set of those per engine command streamer? This is
- * where the name "Logical Rings" starts to make sense: by virtualizing the
- * rings, the engine cs shifts to a new "ring buffer" with every context
- * switch. When you want to submit a workload to the GPU you: A) choose your
- * context, B) find its appropriate virtualized ring, C) write commands to it
- * and then, finally, D) tell the GPU to switch to that context.
- *
- * Instead of the legacy MI_SET_CONTEXT, the way you tell the GPU to switch
- * to a contexts is via a context execution list, ergo "Execlists".
- *
- * LRC implementation:
- * Regarding the creation of contexts, we have:
- *
- * - One global default context.
- * - One local default context for each opened fd.
- * - One local extra context for each context create ioctl call.
- *
- * Now that ringbuffers belong per-context (and not per-engine, like before)
- * and that contexts are uniquely tied to a given engine (and not reusable,
- * like before) we need:
- *
- * - One ringbuffer per-engine inside each context.
- * - One backing object per-engine inside each context.
- *
- * The global default context starts its life with these new objects fully
- * allocated and populated. The local default context for each opened fd is
- * more complex, because we don't know at creation time which engine is going
- * to use them. To handle this, we have implemented a deferred creation of LR
- * contexts:
- *
- * The local context starts its life as a hollow or blank holder, that only
- * gets populated for a given engine once we receive an execbuffer. If later
- * on we receive another execbuffer ioctl for the same context but a different
- * engine, we allocate/populate a new ringbuffer and context backing object and
- * so on.
- *
- * Finally, regarding local contexts created using the ioctl call: as they are
- * only allowed with the render ring, we can allocate & populate them right
- * away (no need to defer anything, at least for now).
- *
- * Execlists implementation:
- * Execlists are the new method by which, on gen8+ hardware, workloads are
- * submitted for execution (as opposed to the legacy, ringbuffer-based, method).
- * This method works as follows:
- *
- * When a request is committed, its commands (the BB start and any leading or
- * trailing commands, like the seqno breadcrumbs) are placed in the ringbuffer
- * for the appropriate context. The tail pointer in the hardware context is not
- * updated at this time, but instead, kept by the driver in the ringbuffer
- * structure. A structure representing this request is added to a request queue
- * for the appropriate engine: this structure contains a copy of the context's
- * tail after the request was written to the ring buffer and a pointer to the
- * context itself.
- *
- * If the engine's request queue was empty before the request was added, the
- * queue is processed immediately. Otherwise the queue will be processed during
- * a context switch interrupt. In any case, elements on the queue will get sent
- * (in pairs) to the GPU's ExecLists Submit Port (ELSP, for short) with a
- * globally unique 20-bits submission ID.
- *
- * When execution of a request completes, the GPU updates the context status
- * buffer with a context complete event and generates a context switch interrupt.
- * During the interrupt handling, the driver examines the events in the buffer:
- * for each context complete event, if the announced ID matches that on the head
- * of the request queue, then that request is retired and removed from the queue.
- *
- * After processing, if any requests were retired and the queue is not empty
- * then a new execution list can be submitted. The two requests at the front of
- * the queue are next to be submitted but since a context may not occur twice in
- * an execution list, if subsequent requests have the same ID as the first then
- * the two requests must be combined. This is done simply by discarding requests
- * at the head of the queue until either only one requests is left (in which case
- * we use a NULL second context) or the first two requests have unique IDs.
- *
- * By always executing the first two requests in the queue the driver ensures
- * that the GPU is kept as busy as possible. In the case where a single context
- * completes but a second context is still executing, the request for this second
- * context will be at the head of the queue when we remove the first one. This
- * request will then be resubmitted along with a new request for a different context,
- * which will cause the hardware to continue executing the second request and queue
- * the new request (the GPU detects the condition of a context getting preempted
- * with the same context and optimizes the context switch flow by not doing
- * preemption, but just sampling the new tail pointer).
- *
  */
-#include <linux/interrupt.h>
 
+#include "gen8_engine_cs.h"
 #include "i915_drv.h"
 #include "i915_perf.h"
-#include "i915_trace.h"
-#include "i915_vgpu.h"
-#include "intel_breadcrumbs.h"
-#include "intel_context.h"
-#include "intel_engine_pm.h"
+#include "intel_engine.h"
+#include "intel_gpu_commands.h"
 #include "intel_gt.h"
-#include "intel_gt_pm.h"
-#include "intel_gt_requests.h"
+#include "intel_lrc.h"
 #include "intel_lrc_reg.h"
-#include "intel_mocs.h"
-#include "intel_reset.h"
 #include "intel_ring.h"
-#include "intel_workarounds.h"
 #include "shmem_utils.h"
 
-#define RING_EXECLIST_QFULL		(1 << 0x2)
-#define RING_EXECLIST1_VALID		(1 << 0x3)
-#define RING_EXECLIST0_VALID		(1 << 0x4)
-#define RING_EXECLIST_ACTIVE_STATUS	(3 << 0xE)
-#define RING_EXECLIST1_ACTIVE		(1 << 0x11)
-#define RING_EXECLIST0_ACTIVE		(1 << 0x12)
-
-#define GEN8_CTX_STATUS_IDLE_ACTIVE	(1 << 0)
-#define GEN8_CTX_STATUS_PREEMPTED	(1 << 1)
-#define GEN8_CTX_STATUS_ELEMENT_SWITCH	(1 << 2)
-#define GEN8_CTX_STATUS_ACTIVE_IDLE	(1 << 3)
-#define GEN8_CTX_STATUS_COMPLETE	(1 << 4)
-#define GEN8_CTX_STATUS_LITE_RESTORE	(1 << 15)
-
-#define GEN8_CTX_STATUS_COMPLETED_MASK \
-	 (GEN8_CTX_STATUS_COMPLETE | GEN8_CTX_STATUS_PREEMPTED)
-
-#define CTX_DESC_FORCE_RESTORE BIT_ULL(2)
-
-#define GEN12_CTX_STATUS_SWITCHED_TO_NEW_QUEUE	(0x1) /* lower csb dword */
-#define GEN12_CTX_SWITCH_DETAIL(csb_dw)	((csb_dw) & 0xF) /* upper csb dword */
-#define GEN12_CSB_SW_CTX_ID_MASK		GENMASK(25, 15)
-#define GEN12_IDLE_CTX_ID		0x7FF
-#define GEN12_CSB_CTX_VALID(csb_dw) \
-	(FIELD_GET(GEN12_CSB_SW_CTX_ID_MASK, csb_dw) != GEN12_IDLE_CTX_ID)
-
-/* Typical size of the average request (2 pipecontrols and a MI_BB) */
-#define EXECLISTS_REQUEST_SIZE 64 /* bytes */
-
-struct virtual_engine {
-	struct intel_engine_cs base;
-	struct intel_context context;
-	struct rcu_work rcu;
-
-	/*
-	 * We allow only a single request through the virtual engine at a time
-	 * (each request in the timeline waits for the completion fence of
-	 * the previous before being submitted). By restricting ourselves to
-	 * only submitting a single request, each request is placed on to a
-	 * physical to maximise load spreading (by virtue of the late greedy
-	 * scheduling -- each real engine takes the next available request
-	 * upon idling).
-	 */
-	struct i915_request *request;
-
-	/*
-	 * We keep a rbtree of available virtual engines inside each physical
-	 * engine, sorted by priority. Here we preallocate the nodes we need
-	 * for the virtual engine, indexed by physical_engine->id.
-	 */
-	struct ve_node {
-		struct rb_node rb;
-		int prio;
-	} nodes[I915_NUM_ENGINES];
-
-	/*
-	 * Keep track of bonded pairs -- restrictions upon on our selection
-	 * of physical engines any particular request may be submitted to.
-	 * If we receive a submit-fence from a master engine, we will only
-	 * use one of sibling_mask physical engines.
-	 */
-	struct ve_bond {
-		const struct intel_engine_cs *master;
-		intel_engine_mask_t sibling_mask;
-	} *bonds;
-	unsigned int num_bonds;
-
-	/* And finally, which physical engines this virtual engine maps onto. */
-	unsigned int num_siblings;
-	struct intel_engine_cs *siblings[];
-};
-
-static struct virtual_engine *to_virtual_engine(struct intel_engine_cs *engine)
-{
-	GEM_BUG_ON(!intel_engine_is_virtual(engine));
-	return container_of(engine, struct virtual_engine, base);
-}
-
-static int __execlists_context_alloc(struct intel_context *ce,
-				     struct intel_engine_cs *engine);
-
-static void execlists_init_reg_state(u32 *reg_state,
-				     const struct intel_context *ce,
-				     const struct intel_engine_cs *engine,
-				     const struct intel_ring *ring,
-				     bool close);
-static void
-__execlists_update_reg_state(const struct intel_context *ce,
-			     const struct intel_engine_cs *engine,
-			     u32 head);
-
-static int lrc_ring_mi_mode(const struct intel_engine_cs *engine)
-{
-	if (INTEL_GEN(engine->i915) >= 12)
-		return 0x60;
-	else if (INTEL_GEN(engine->i915) >= 9)
-		return 0x54;
-	else if (engine->class == RENDER_CLASS)
-		return 0x58;
-	else
-		return -1;
-}
-
-static int lrc_ring_gpr0(const struct intel_engine_cs *engine)
-{
-	if (INTEL_GEN(engine->i915) >= 12)
-		return 0x74;
-	else if (INTEL_GEN(engine->i915) >= 9)
-		return 0x68;
-	else if (engine->class == RENDER_CLASS)
-		return 0xd8;
-	else
-		return -1;
-}
-
-static int lrc_ring_wa_bb_per_ctx(const struct intel_engine_cs *engine)
-{
-	if (INTEL_GEN(engine->i915) >= 12)
-		return 0x12;
-	else if (INTEL_GEN(engine->i915) >= 9 || engine->class == RENDER_CLASS)
-		return 0x18;
-	else
-		return -1;
-}
-
-static int lrc_ring_indirect_ptr(const struct intel_engine_cs *engine)
-{
-	int x;
-
-	x = lrc_ring_wa_bb_per_ctx(engine);
-	if (x < 0)
-		return x;
-
-	return x + 2;
-}
-
-static int lrc_ring_indirect_offset(const struct intel_engine_cs *engine)
-{
-	int x;
-
-	x = lrc_ring_indirect_ptr(engine);
-	if (x < 0)
-		return x;
-
-	return x + 2;
-}
-
-static int lrc_ring_cmd_buf_cctl(const struct intel_engine_cs *engine)
-{
-	if (engine->class != RENDER_CLASS)
-		return -1;
-
-	if (INTEL_GEN(engine->i915) >= 12)
-		return 0xb6;
-	else if (INTEL_GEN(engine->i915) >= 11)
-		return 0xaa;
-	else
-		return -1;
-}
-
-static u32
-lrc_ring_indirect_offset_default(const struct intel_engine_cs *engine)
-{
-	switch (INTEL_GEN(engine->i915)) {
-	default:
-		MISSING_CASE(INTEL_GEN(engine->i915));
-		fallthrough;
-	case 12:
-		return GEN12_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
-	case 11:
-		return GEN11_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
-	case 10:
-		return GEN10_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
-	case 9:
-		return GEN9_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
-	case 8:
-		return GEN8_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
-	}
-}
-
-static void
-lrc_ring_setup_indirect_ctx(u32 *regs,
-			    const struct intel_engine_cs *engine,
-			    u32 ctx_bb_ggtt_addr,
-			    u32 size)
-{
-	GEM_BUG_ON(!size);
-	GEM_BUG_ON(!IS_ALIGNED(size, CACHELINE_BYTES));
-	GEM_BUG_ON(lrc_ring_indirect_ptr(engine) == -1);
-	regs[lrc_ring_indirect_ptr(engine) + 1] =
-		ctx_bb_ggtt_addr | (size / CACHELINE_BYTES);
-
-	GEM_BUG_ON(lrc_ring_indirect_offset(engine) == -1);
-	regs[lrc_ring_indirect_offset(engine) + 1] =
-		lrc_ring_indirect_offset_default(engine) << 6;
-}
-
-static u32 intel_context_get_runtime(const struct intel_context *ce)
-{
-	/*
-	 * We can use either ppHWSP[16] which is recorded before the context
-	 * switch (and so excludes the cost of context switches) or use the
-	 * value from the context image itself, which is saved/restored earlier
-	 * and so includes the cost of the save.
-	 */
-	return READ_ONCE(ce->lrc_reg_state[CTX_TIMESTAMP]);
-}
-
-static void mark_eio(struct i915_request *rq)
-{
-	if (i915_request_completed(rq))
-		return;
-
-	GEM_BUG_ON(i915_request_signaled(rq));
-
-	i915_request_set_error_once(rq, -EIO);
-	i915_request_mark_complete(rq);
-}
-
-static struct i915_request *
-active_request(const struct intel_timeline * const tl, struct i915_request *rq)
-{
-	struct i915_request *active = rq;
-
-	rcu_read_lock();
-	list_for_each_entry_continue_reverse(rq, &tl->requests, link) {
-		if (i915_request_completed(rq))
-			break;
-
-		active = rq;
-	}
-	rcu_read_unlock();
-
-	return active;
-}
-
-static inline u32 intel_hws_preempt_address(struct intel_engine_cs *engine)
-{
-	return (i915_ggtt_offset(engine->status_page.vma) +
-		I915_GEM_HWS_PREEMPT_ADDR);
-}
-
-static inline void
-ring_set_paused(const struct intel_engine_cs *engine, int state)
-{
-	/*
-	 * We inspect HWS_PREEMPT with a semaphore inside
-	 * engine->emit_fini_breadcrumb. If the dword is true,
-	 * the ring is paused as the semaphore will busywait
-	 * until the dword is false.
-	 */
-	engine->status_page.addr[I915_GEM_HWS_PREEMPT] = state;
-	if (state)
-		wmb();
-}
-
-static inline struct i915_priolist *to_priolist(struct rb_node *rb)
-{
-	return rb_entry(rb, struct i915_priolist, node);
-}
-
-static inline int rq_prio(const struct i915_request *rq)
-{
-	return READ_ONCE(rq->sched.attr.priority);
-}
-
-static int effective_prio(const struct i915_request *rq)
-{
-	int prio = rq_prio(rq);
-
-	/*
-	 * If this request is special and must not be interrupted at any
-	 * cost, so be it. Note we are only checking the most recent request
-	 * in the context and so may be masking an earlier vip request. It
-	 * is hoped that under the conditions where nopreempt is used, this
-	 * will not matter (i.e. all requests to that context will be
-	 * nopreempt for as long as desired).
-	 */
-	if (i915_request_has_nopreempt(rq))
-		prio = I915_PRIORITY_UNPREEMPTABLE;
-
-	return prio;
-}
-
-static int queue_prio(const struct intel_engine_execlists *execlists)
-{
-	struct i915_priolist *p;
-	struct rb_node *rb;
-
-	rb = rb_first_cached(&execlists->queue);
-	if (!rb)
-		return INT_MIN;
-
-	/*
-	 * As the priolist[] are inverted, with the highest priority in [0],
-	 * we have to flip the index value to become priority.
-	 */
-	p = to_priolist(rb);
-	if (!I915_USER_PRIORITY_SHIFT)
-		return p->priority;
-
-	return ((p->priority + 1) << I915_USER_PRIORITY_SHIFT) - ffs(p->used);
-}
-
-static inline bool need_preempt(const struct intel_engine_cs *engine,
-				const struct i915_request *rq,
-				struct rb_node *rb)
-{
-	int last_prio;
-
-	if (!intel_engine_has_semaphores(engine))
-		return false;
-
-	/*
-	 * Check if the current priority hint merits a preemption attempt.
-	 *
-	 * We record the highest value priority we saw during rescheduling
-	 * prior to this dequeue, therefore we know that if it is strictly
-	 * less than the current tail of ESLP[0], we do not need to force
-	 * a preempt-to-idle cycle.
-	 *
-	 * However, the priority hint is a mere hint that we may need to
-	 * preempt. If that hint is stale or we may be trying to preempt
-	 * ourselves, ignore the request.
-	 *
-	 * More naturally we would write
-	 *      prio >= max(0, last);
-	 * except that we wish to prevent triggering preemption at the same
-	 * priority level: the task that is running should remain running
-	 * to preserve FIFO ordering of dependencies.
-	 */
-	last_prio = max(effective_prio(rq), I915_PRIORITY_NORMAL - 1);
-	if (engine->execlists.queue_priority_hint <= last_prio)
-		return false;
-
-	/*
-	 * Check against the first request in ELSP[1], it will, thanks to the
-	 * power of PI, be the highest priority of that context.
-	 */
-	if (!list_is_last(&rq->sched.link, &engine->active.requests) &&
-	    rq_prio(list_next_entry(rq, sched.link)) > last_prio)
-		return true;
-
-	if (rb) {
-		struct virtual_engine *ve =
-			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
-		bool preempt = false;
-
-		if (engine == ve->siblings[0]) { /* only preempt one sibling */
-			struct i915_request *next;
-
-			rcu_read_lock();
-			next = READ_ONCE(ve->request);
-			if (next)
-				preempt = rq_prio(next) > last_prio;
-			rcu_read_unlock();
-		}
-
-		if (preempt)
-			return preempt;
-	}
-
-	/*
-	 * If the inflight context did not trigger the preemption, then maybe
-	 * it was the set of queued requests? Pick the highest priority in
-	 * the queue (the first active priolist) and see if it deserves to be
-	 * running instead of ELSP[0].
-	 *
-	 * The highest priority request in the queue can not be either
-	 * ELSP[0] or ELSP[1] as, thanks again to PI, if it was the same
-	 * context, it's priority would not exceed ELSP[0] aka last_prio.
-	 */
-	return queue_prio(&engine->execlists) > last_prio;
-}
-
-__maybe_unused static inline bool
-assert_priority_queue(const struct i915_request *prev,
-		      const struct i915_request *next)
-{
-	/*
-	 * Without preemption, the prev may refer to the still active element
-	 * which we refuse to let go.
-	 *
-	 * Even with preemption, there are times when we think it is better not
-	 * to preempt and leave an ostensibly lower priority request in flight.
-	 */
-	if (i915_request_is_active(prev))
-		return true;
-
-	return rq_prio(prev) >= rq_prio(next);
-}
-
-/*
- * The context descriptor encodes various attributes of a context,
- * including its GTT address and some flags. Because it's fairly
- * expensive to calculate, we'll just do it once and cache the result,
- * which remains valid until the context is unpinned.
- *
- * This is what a descriptor looks like, from LSB to MSB::
- *
- *      bits  0-11:    flags, GEN8_CTX_* (cached in ctx->desc_template)
- *      bits 12-31:    LRCA, GTT address of (the HWSP of) this context
- *      bits 32-52:    ctx ID, a globally unique tag (highest bit used by GuC)
- *      bits 53-54:    mbz, reserved for use by hardware
- *      bits 55-63:    group ID, currently unused and set to 0
- *
- * Starting from Gen11, the upper dword of the descriptor has a new format:
- *
- *      bits 32-36:    reserved
- *      bits 37-47:    SW context ID
- *      bits 48:53:    engine instance
- *      bit 54:        mbz, reserved for use by hardware
- *      bits 55-60:    SW counter
- *      bits 61-63:    engine class
- *
- * engine info, SW context ID and SW counter need to form a unique number
- * (Context ID) per lrc.
- */
-static u32
-lrc_descriptor(struct intel_context *ce, struct intel_engine_cs *engine)
-{
-	u32 desc;
-
-	desc = INTEL_LEGACY_32B_CONTEXT;
-	if (i915_vm_is_4lvl(ce->vm))
-		desc = INTEL_LEGACY_64B_CONTEXT;
-	desc <<= GEN8_CTX_ADDRESSING_MODE_SHIFT;
-
-	desc |= GEN8_CTX_VALID | GEN8_CTX_PRIVILEGE;
-	if (IS_GEN(engine->i915, 8))
-		desc |= GEN8_CTX_L3LLC_COHERENT;
-
-	return i915_ggtt_offset(ce->state) | desc;
-}
-
-static inline unsigned int dword_in_page(void *addr)
-{
-	return offset_in_page(addr) / sizeof(u32);
-}
-
 static void set_offsets(u32 *regs,
 			const u8 *data,
 			const struct intel_engine_cs *engine,
-			bool clear)
+			bool close)
 #define NOP(x) (BIT(7) | (x))
 #define LRI(count, flags) ((flags) << 6 | (count) | BUILD_BUG_ON_ZERO(count >= BIT(6)))
 #define POSTED BIT(0)
@@ -601,7 +25,7 @@ static void set_offsets(u32 *regs,
 #define REG16(x) \
 	(((x) >> 9) | BIT(7) | BUILD_BUG_ON_ZERO(x >= 0x10000)), \
 	(((x) >> 2) & 0x7f)
-#define END(total_state_size) 0, (total_state_size)
+#define END 0
 {
 	const u32 base = engine->mmio_base;
 
@@ -610,8 +34,6 @@ static void set_offsets(u32 *regs,
 
 		if (*data & BIT(7)) { /* skip */
 			count = *data++ & ~BIT(7);
-			if (clear)
-				memset32(regs, MI_NOOP, count);
 			regs += count;
 			continue;
 		}
@@ -639,19 +61,11 @@ static void set_offsets(u32 *regs,
 			} while (v & BIT(7));
 
 			regs[0] = base + (offset << 2);
-			if (clear)
-				regs[1] = 0;
 			regs += 2;
 		} while (--count);
 	}
 
-	if (clear) {
-		u8 count = *++data;
-
-		/* Clear past the tail for HW access */
-		GEM_BUG_ON(dword_in_page(regs) > count);
-		memset32(regs, MI_NOOP, count - dword_in_page(regs));
-
+	if (close) {
 		/* Close the batch; used mainly by live_lrc_layout() */
 		*regs = MI_BATCH_BUFFER_END;
 		if (INTEL_GEN(engine->i915) >= 10)
@@ -691,7 +105,7 @@ static const u8 gen8_xcs_offsets[] = {
 	REG16(0x200),
 	REG(0x028),
 
-	END(80)
+	END
 };
 
 static const u8 gen9_xcs_offsets[] = {
@@ -775,7 +189,7 @@ static const u8 gen9_xcs_offsets[] = {
 	REG16(0x67c),
 	REG(0x068),
 
-	END(176)
+	END
 };
 
 static const u8 gen12_xcs_offsets[] = {
@@ -807,7 +221,7 @@ static const u8 gen12_xcs_offsets[] = {
 	REG16(0x274),
 	REG16(0x270),
 
-	END(80)
+	END
 };
 
 static const u8 gen8_rcs_offsets[] = {
@@ -844,7 +258,7 @@ static const u8 gen8_rcs_offsets[] = {
 	LRI(1, 0),
 	REG(0x0c8),
 
-	END(80)
+	END
 };
 
 static const u8 gen9_rcs_offsets[] = {
@@ -928,7 +342,7 @@ static const u8 gen9_rcs_offsets[] = {
 	REG16(0x67c),
 	REG(0x68),
 
-	END(176)
+	END
 };
 
 static const u8 gen11_rcs_offsets[] = {
@@ -969,7 +383,7 @@ static const u8 gen11_rcs_offsets[] = {
 	LRI(1, 0),
 	REG(0x0c8),
 
-	END(80)
+	END
 };
 
 static const u8 gen12_rcs_offsets[] = {
@@ -1065,7 +479,7 @@ static const u8 gen12_rcs_offsets[] = {
 	REG(0x084),
 	NOP(1),
 
-	END(192)
+	END
 };
 
 #undef END
@@ -1104,2284 +518,440 @@ static const u8 *reg_offsets(const struct intel_engine_cs *engine)
 	}
 }
 
-static struct i915_request *
-__unwind_incomplete_requests(struct intel_engine_cs *engine)
-{
-	struct i915_request *rq, *rn, *active = NULL;
-	struct list_head *pl;
-	int prio = I915_PRIORITY_INVALID;
-
-	lockdep_assert_held(&engine->active.lock);
-
-	list_for_each_entry_safe_reverse(rq, rn,
-					 &engine->active.requests,
-					 sched.link) {
-		if (i915_request_completed(rq))
-			continue; /* XXX */
-
-		__i915_request_unsubmit(rq);
-
-		/*
-		 * Push the request back into the queue for later resubmission.
-		 * If this request is not native to this physical engine (i.e.
-		 * it came from a virtual source), push it back onto the virtual
-		 * engine so that it can be moved across onto another physical
-		 * engine as load dictates.
-		 */
-		if (likely(rq->execution_mask == engine->mask)) {
-			GEM_BUG_ON(rq_prio(rq) == I915_PRIORITY_INVALID);
-			if (rq_prio(rq) != prio) {
-				prio = rq_prio(rq);
-				pl = i915_sched_lookup_priolist(engine, prio);
-			}
-			GEM_BUG_ON(RB_EMPTY_ROOT(&engine->execlists.queue.rb_root));
-
-			list_move(&rq->sched.link, pl);
-			set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
-
-			/* Check in case we rollback so far we wrap [size/2] */
-			if (intel_ring_direction(rq->ring,
-						 rq->tail,
-						 rq->ring->tail + 8) > 0)
-				rq->context->lrc.desc |= CTX_DESC_FORCE_RESTORE;
-
-			active = rq;
-		} else {
-			struct intel_engine_cs *owner = rq->context->engine;
-
-			WRITE_ONCE(rq->engine, owner);
-			owner->submit_request(rq);
-			active = NULL;
-		}
-	}
-
-	return active;
-}
-
-struct i915_request *
-execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists)
-{
-	struct intel_engine_cs *engine =
-		container_of(execlists, typeof(*engine), execlists);
-
-	return __unwind_incomplete_requests(engine);
-}
-
-static inline void
-execlists_context_status_change(struct i915_request *rq, unsigned long status)
+static int lrc_ring_mi_mode(const struct intel_engine_cs *engine)
 {
-	/*
-	 * Only used when GVT-g is enabled now. When GVT-g is disabled,
-	 * The compiler should eliminate this function as dead-code.
-	 */
-	if (!IS_ENABLED(CONFIG_DRM_I915_GVT))
-		return;
-
-	atomic_notifier_call_chain(&rq->engine->context_status_notifier,
-				   status, rq);
+	if (INTEL_GEN(engine->i915) >= 12)
+		return 0x60;
+	else if (INTEL_GEN(engine->i915) >= 9)
+		return 0x54;
+	else if (engine->class == RENDER_CLASS)
+		return 0x58;
+	else
+		return -1;
 }
 
-static void intel_engine_context_in(struct intel_engine_cs *engine)
+static int lrc_ring_gpr0(const struct intel_engine_cs *engine)
 {
-	unsigned long flags;
-
-	if (atomic_add_unless(&engine->stats.active, 1, 0))
-		return;
-
-	write_seqlock_irqsave(&engine->stats.lock, flags);
-	if (!atomic_add_unless(&engine->stats.active, 1, 0)) {
-		engine->stats.start = ktime_get();
-		atomic_inc(&engine->stats.active);
-	}
-	write_sequnlock_irqrestore(&engine->stats.lock, flags);
+	if (INTEL_GEN(engine->i915) >= 12)
+		return 0x74;
+	else if (INTEL_GEN(engine->i915) >= 9)
+		return 0x68;
+	else if (engine->class == RENDER_CLASS)
+		return 0xd8;
+	else
+		return -1;
 }
 
-static void intel_engine_context_out(struct intel_engine_cs *engine)
+static int lrc_ring_wa_bb_per_ctx(const struct intel_engine_cs *engine)
 {
-	unsigned long flags;
-
-	GEM_BUG_ON(!atomic_read(&engine->stats.active));
-
-	if (atomic_add_unless(&engine->stats.active, -1, 1))
-		return;
-
-	write_seqlock_irqsave(&engine->stats.lock, flags);
-	if (atomic_dec_and_test(&engine->stats.active)) {
-		engine->stats.total =
-			ktime_add(engine->stats.total,
-				  ktime_sub(ktime_get(), engine->stats.start));
-	}
-	write_sequnlock_irqrestore(&engine->stats.lock, flags);
+	if (INTEL_GEN(engine->i915) >= 12)
+		return 0x12;
+	else if (INTEL_GEN(engine->i915) >= 9 || engine->class == RENDER_CLASS)
+		return 0x18;
+	else
+		return -1;
 }
 
-static void
-execlists_check_context(const struct intel_context *ce,
-			const struct intel_engine_cs *engine,
-			const char *when)
+static int lrc_ring_indirect_ptr(const struct intel_engine_cs *engine)
 {
-	const struct intel_ring *ring = ce->ring;
-	u32 *regs = ce->lrc_reg_state;
-	bool valid = true;
 	int x;
 
-	if (regs[CTX_RING_START] != i915_ggtt_offset(ring->vma)) {
-		pr_err("%s: context submitted with incorrect RING_START [%08x], expected %08x\n",
-		       engine->name,
-		       regs[CTX_RING_START],
-		       i915_ggtt_offset(ring->vma));
-		regs[CTX_RING_START] = i915_ggtt_offset(ring->vma);
-		valid = false;
-	}
-
-	if ((regs[CTX_RING_CTL] & ~(RING_WAIT | RING_WAIT_SEMAPHORE)) !=
-	    (RING_CTL_SIZE(ring->size) | RING_VALID)) {
-		pr_err("%s: context submitted with incorrect RING_CTL [%08x], expected %08x\n",
-		       engine->name,
-		       regs[CTX_RING_CTL],
-		       (u32)(RING_CTL_SIZE(ring->size) | RING_VALID));
-		regs[CTX_RING_CTL] = RING_CTL_SIZE(ring->size) | RING_VALID;
-		valid = false;
-	}
-
-	x = lrc_ring_mi_mode(engine);
-	if (x != -1 && regs[x + 1] & (regs[x + 1] >> 16) & STOP_RING) {
-		pr_err("%s: context submitted with STOP_RING [%08x] in RING_MI_MODE\n",
-		       engine->name, regs[x + 1]);
-		regs[x + 1] &= ~STOP_RING;
-		regs[x + 1] |= STOP_RING << 16;
-		valid = false;
-	}
+	x = lrc_ring_wa_bb_per_ctx(engine);
+	if (x < 0)
+		return x;
 
-	WARN_ONCE(!valid, "Invalid lrc state found %s submission\n", when);
+	return x + 2;
 }
 
-static void restore_default_state(struct intel_context *ce,
-				  struct intel_engine_cs *engine)
+static int lrc_ring_indirect_offset(const struct intel_engine_cs *engine)
 {
-	u32 *regs;
+	int x;
 
-	regs = memset(ce->lrc_reg_state, 0, engine->context_size - PAGE_SIZE);
-	execlists_init_reg_state(regs, ce, engine, ce->ring, true);
+	x = lrc_ring_indirect_ptr(engine);
+	if (x < 0)
+		return x;
 
-	ce->runtime.last = intel_context_get_runtime(ce);
+	return x + 2;
 }
 
-static void reset_active(struct i915_request *rq,
-			 struct intel_engine_cs *engine)
+static int lrc_ring_cmd_buf_cctl(const struct intel_engine_cs *engine)
 {
-	struct intel_context * const ce = rq->context;
-	u32 head;
-
-	/*
-	 * The executing context has been cancelled. We want to prevent
-	 * further execution along this context and propagate the error on
-	 * to anything depending on its results.
-	 *
-	 * In __i915_request_submit(), we apply the -EIO and remove the
-	 * requests' payloads for any banned requests. But first, we must
-	 * rewind the context back to the start of the incomplete request so
-	 * that we do not jump back into the middle of the batch.
-	 *
-	 * We preserve the breadcrumbs and semaphores of the incomplete
-	 * requests so that inter-timeline dependencies (i.e other timelines)
-	 * remain correctly ordered. And we defer to __i915_request_submit()
-	 * so that all asynchronous waits are correctly handled.
-	 */
-	ENGINE_TRACE(engine, "{ rq=%llx:%lld }\n",
-		     rq->fence.context, rq->fence.seqno);
+	if (engine->class != RENDER_CLASS)
+		return -1;
 
-	/* On resubmission of the active request, payload will be scrubbed */
-	if (i915_request_completed(rq))
-		head = rq->tail;
+	if (INTEL_GEN(engine->i915) >= 12)
+		return 0xb6;
+	else if (INTEL_GEN(engine->i915) >= 11)
+		return 0xaa;
 	else
-		head = active_request(ce->timeline, rq)->head;
-	head = intel_ring_wrap(ce->ring, head);
-
-	/* Scrub the context image to prevent replaying the previous batch */
-	restore_default_state(ce, engine);
-	__execlists_update_reg_state(ce, engine, head);
-
-	/* We've switched away, so this should be a no-op, but intent matters */
-	ce->lrc.desc |= CTX_DESC_FORCE_RESTORE;
-}
-
-static void st_update_runtime_underflow(struct intel_context *ce, s32 dt)
-{
-#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
-	ce->runtime.num_underflow += dt < 0;
-	ce->runtime.max_underflow = max_t(u32, ce->runtime.max_underflow, -dt);
-#endif
+		return -1;
 }
 
-static void intel_context_update_runtime(struct intel_context *ce)
+static u32
+lrc_ring_indirect_offset_default(const struct intel_engine_cs *engine)
 {
-	u32 old;
-	s32 dt;
-
-	if (intel_context_is_barrier(ce))
-		return;
-
-	old = ce->runtime.last;
-	ce->runtime.last = intel_context_get_runtime(ce);
-	dt = ce->runtime.last - old;
-
-	if (unlikely(dt <= 0)) {
-		CE_TRACE(ce, "runtime underflow: last=%u, new=%u, delta=%d\n",
-			 old, ce->runtime.last, dt);
-		st_update_runtime_underflow(ce, dt);
-		return;
+	switch (INTEL_GEN(engine->i915)) {
+	default:
+		MISSING_CASE(INTEL_GEN(engine->i915));
+		fallthrough;
+	case 12:
+		return GEN12_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 11:
+		return GEN11_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 10:
+		return GEN10_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 9:
+		return GEN9_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 8:
+		return GEN8_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
 	}
-
-	ewma_runtime_add(&ce->runtime.avg, dt);
-	ce->runtime.total += dt;
 }
 
-static inline struct intel_engine_cs *
-__execlists_schedule_in(struct i915_request *rq)
+static void
+lrc_setup_indirect_ctx(u32 *regs,
+		       const struct intel_engine_cs *engine,
+		       u32 ctx_bb_ggtt_addr,
+		       u32 size)
 {
-	struct intel_engine_cs * const engine = rq->engine;
-	struct intel_context * const ce = rq->context;
-
-	intel_context_get(ce);
-
-	if (unlikely(intel_context_is_banned(ce)))
-		reset_active(rq, engine);
-
-	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		execlists_check_context(ce, engine, "before");
-
-	if (ce->tag) {
-		/* Use a fixed tag for OA and friends */
-		GEM_BUG_ON(ce->tag <= BITS_PER_LONG);
-		ce->lrc.ccid = ce->tag;
-	} else {
-		/* We don't need a strict matching tag, just different values */
-		unsigned int tag = ffs(READ_ONCE(engine->context_tag));
-
-		GEM_BUG_ON(tag == 0 || tag >= BITS_PER_LONG);
-		clear_bit(tag - 1, &engine->context_tag);
-		ce->lrc.ccid = tag << (GEN11_SW_CTX_ID_SHIFT - 32);
-
-		BUILD_BUG_ON(BITS_PER_LONG > GEN12_MAX_CONTEXT_HW_ID);
-	}
-
-	ce->lrc.ccid |= engine->execlists.ccid;
-
-	__intel_gt_pm_get(engine->gt);
-	if (engine->fw_domain && !atomic_fetch_inc(&engine->fw_active))
-		intel_uncore_forcewake_get(engine->uncore, engine->fw_domain);
-	execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_IN);
-	intel_engine_context_in(engine);
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, CACHELINE_BYTES));
+	GEM_BUG_ON(lrc_ring_indirect_ptr(engine) == -1);
+	regs[lrc_ring_indirect_ptr(engine) + 1] =
+		ctx_bb_ggtt_addr | (size / CACHELINE_BYTES);
 
-	return engine;
+	GEM_BUG_ON(lrc_ring_indirect_offset(engine) == -1);
+	regs[lrc_ring_indirect_offset(engine) + 1] =
+		lrc_ring_indirect_offset_default(engine) << 6;
 }
 
-static inline struct i915_request *
-execlists_schedule_in(struct i915_request *rq, int idx)
+static void init_common_regs(u32 * const regs,
+			     const struct intel_context *ce,
+			     const struct intel_engine_cs *engine,
+			     bool inhibit)
 {
-	struct intel_context * const ce = rq->context;
-	struct intel_engine_cs *old;
-
-	GEM_BUG_ON(!intel_engine_pm_is_awake(rq->engine));
-	trace_i915_request_in(rq, idx);
-
-	old = READ_ONCE(ce->inflight);
-	do {
-		if (!old) {
-			WRITE_ONCE(ce->inflight, __execlists_schedule_in(rq));
-			break;
-		}
-	} while (!try_cmpxchg(&ce->inflight, &old, ptr_inc(old)));
-
-	GEM_BUG_ON(intel_context_inflight(ce) != rq->engine);
-	return i915_request_get(rq);
-}
+	u32 ctl;
 
-static void kick_siblings(struct i915_request *rq, struct intel_context *ce)
-{
-	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
-	struct i915_request *next = READ_ONCE(ve->request);
+	ctl = _MASKED_BIT_ENABLE(CTX_CTRL_INHIBIT_SYN_CTX_SWITCH);
+	ctl |= _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT);
+	if (inhibit)
+		ctl |= CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT;
+	if (INTEL_GEN(engine->i915) < 11)
+		ctl |= _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT |
+					   CTX_CTRL_RS_CTX_ENABLE);
+	regs[CTX_CONTEXT_CONTROL] = ctl;
 
-	if (next == rq || (next && next->execution_mask & ~rq->execution_mask))
-		tasklet_hi_schedule(&ve->base.execlists.tasklet);
+	regs[CTX_TIMESTAMP] = ce->runtime.last;
 }
 
-static inline void
-__execlists_schedule_out(struct i915_request *rq,
-			 struct intel_engine_cs * const engine,
-			 unsigned int ccid)
+static void init_wa_bb_regs(u32 * const regs,
+			    const struct intel_engine_cs *engine)
 {
-	struct intel_context * const ce = rq->context;
-
-	/*
-	 * NB process_csb() is not under the engine->active.lock and hence
-	 * schedule_out can race with schedule_in meaning that we should
-	 * refrain from doing non-trivial work here.
-	 */
+	const struct i915_ctx_workarounds * const wa_ctx = &engine->wa_ctx;
 
-	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		execlists_check_context(ce, engine, "after");
+	if (wa_ctx->per_ctx.size) {
+		const u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
 
-	/*
-	 * If we have just completed this context, the engine may now be
-	 * idle and we want to re-enter powersaving.
-	 */
-	if (list_is_last_rcu(&rq->link, &ce->timeline->requests) &&
-	    i915_request_completed(rq))
-		intel_engine_add_retire(engine, ce->timeline);
-
-	ccid >>= GEN11_SW_CTX_ID_SHIFT - 32;
-	ccid &= GEN12_MAX_CONTEXT_HW_ID;
-	if (ccid < BITS_PER_LONG) {
-		GEM_BUG_ON(ccid == 0);
-		GEM_BUG_ON(test_bit(ccid - 1, &engine->context_tag));
-		set_bit(ccid - 1, &engine->context_tag);
+		GEM_BUG_ON(lrc_ring_wa_bb_per_ctx(engine) == -1);
+		regs[lrc_ring_wa_bb_per_ctx(engine) + 1] =
+			(ggtt_offset + wa_ctx->per_ctx.offset) | 0x01;
 	}
 
-	intel_context_update_runtime(ce);
-	intel_engine_context_out(engine);
-	execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_OUT);
-	if (engine->fw_domain && !atomic_dec_return(&engine->fw_active))
-		intel_uncore_forcewake_put(engine->uncore, engine->fw_domain);
-	intel_gt_pm_put_async(engine->gt);
-
-	/*
-	 * If this is part of a virtual engine, its next request may
-	 * have been blocked waiting for access to the active context.
-	 * We have to kick all the siblings again in case we need to
-	 * switch (e.g. the next request is not runnable on this
-	 * engine). Hopefully, we will already have submitted the next
-	 * request before the tasklet runs and do not need to rebuild
-	 * each virtual tree and kick everyone again.
-	 */
-	if (ce->engine != engine)
-		kick_siblings(rq, ce);
-
-	intel_context_put(ce);
-}
-
-static inline void
-execlists_schedule_out(struct i915_request *rq)
-{
-	struct intel_context * const ce = rq->context;
-	struct intel_engine_cs *cur, *old;
-	u32 ccid;
-
-	trace_i915_request_out(rq);
-
-	ccid = rq->context->lrc.ccid;
-	old = READ_ONCE(ce->inflight);
-	do
-		cur = ptr_unmask_bits(old, 2) ? ptr_dec(old) : NULL;
-	while (!try_cmpxchg(&ce->inflight, &old, cur));
-	if (!cur)
-		__execlists_schedule_out(rq, old, ccid);
-
-	i915_request_put(rq);
-}
-
-static u64 execlists_update_context(struct i915_request *rq)
-{
-	struct intel_context *ce = rq->context;
-	u64 desc = ce->lrc.desc;
-	u32 tail, prev;
-
-	/*
-	 * WaIdleLiteRestore:bdw,skl
-	 *
-	 * We should never submit the context with the same RING_TAIL twice
-	 * just in case we submit an empty ring, which confuses the HW.
-	 *
-	 * We append a couple of NOOPs (gen8_emit_wa_tail) after the end of
-	 * the normal request to be able to always advance the RING_TAIL on
-	 * subsequent resubmissions (for lite restore). Should that fail us,
-	 * and we try and submit the same tail again, force the context
-	 * reload.
-	 *
-	 * If we need to return to a preempted context, we need to skip the
-	 * lite-restore and force it to reload the RING_TAIL. Otherwise, the
-	 * HW has a tendency to ignore us rewinding the TAIL to the end of
-	 * an earlier request.
-	 */
-	GEM_BUG_ON(ce->lrc_reg_state[CTX_RING_TAIL] != rq->ring->tail);
-	prev = rq->ring->tail;
-	tail = intel_ring_set_tail(rq->ring, rq->tail);
-	if (unlikely(intel_ring_direction(rq->ring, tail, prev) <= 0))
-		desc |= CTX_DESC_FORCE_RESTORE;
-	ce->lrc_reg_state[CTX_RING_TAIL] = tail;
-	rq->tail = rq->wa_tail;
-
-	/*
-	 * Make sure the context image is complete before we submit it to HW.
-	 *
-	 * Ostensibly, writes (including the WCB) should be flushed prior to
-	 * an uncached write such as our mmio register access, the empirical
-	 * evidence (esp. on Braswell) suggests that the WC write into memory
-	 * may not be visible to the HW prior to the completion of the UC
-	 * register write and that we may begin execution from the context
-	 * before its image is complete leading to invalid PD chasing.
-	 */
-	wmb();
-
-	ce->lrc.desc &= ~CTX_DESC_FORCE_RESTORE;
-	return desc;
-}
-
-static inline void write_desc(struct intel_engine_execlists *execlists, u64 desc, u32 port)
-{
-	if (execlists->ctrl_reg) {
-		writel(lower_32_bits(desc), execlists->submit_reg + port * 2);
-		writel(upper_32_bits(desc), execlists->submit_reg + port * 2 + 1);
-	} else {
-		writel(upper_32_bits(desc), execlists->submit_reg);
-		writel(lower_32_bits(desc), execlists->submit_reg);
+	if (wa_ctx->indirect_ctx.size) {
+		lrc_setup_indirect_ctx(regs, engine,
+				       i915_ggtt_offset(wa_ctx->vma) +
+				       wa_ctx->indirect_ctx.offset,
+				       wa_ctx->indirect_ctx.size);
 	}
 }
 
-static __maybe_unused char *
-dump_port(char *buf, int buflen, const char *prefix, struct i915_request *rq)
+static void init_ppgtt_regs(u32 *regs, const struct i915_ppgtt *ppgtt)
 {
-	if (!rq)
-		return "";
-
-	snprintf(buf, buflen, "%sccid:%x %llx:%lld%s prio %d",
-		 prefix,
-		 rq->context->lrc.ccid,
-		 rq->fence.context, rq->fence.seqno,
-		 i915_request_completed(rq) ? "!" :
-		 i915_request_started(rq) ? "*" :
-		 "",
-		 rq_prio(rq));
-
-	return buf;
-}
-
-static __maybe_unused void
-trace_ports(const struct intel_engine_execlists *execlists,
-	    const char *msg,
-	    struct i915_request * const *ports)
-{
-	const struct intel_engine_cs *engine =
-		container_of(execlists, typeof(*engine), execlists);
-	char __maybe_unused p0[40], p1[40];
-
-	if (!ports[0])
-		return;
-
-	ENGINE_TRACE(engine, "%s { %s%s }\n", msg,
-		     dump_port(p0, sizeof(p0), "", ports[0]),
-		     dump_port(p1, sizeof(p1), ", ", ports[1]));
-}
-
-static inline bool
-reset_in_progress(const struct intel_engine_execlists *execlists)
-{
-	return unlikely(!__tasklet_is_enabled(&execlists->tasklet));
-}
-
-static __maybe_unused bool
-assert_pending_valid(const struct intel_engine_execlists *execlists,
-		     const char *msg)
-{
-	struct intel_engine_cs *engine =
-		container_of(execlists, typeof(*engine), execlists);
-	struct i915_request * const *port, *rq;
-	struct intel_context *ce = NULL;
-	bool sentinel = false;
-	u32 ccid = -1;
-
-	trace_ports(execlists, msg, execlists->pending);
-
-	/* We may be messing around with the lists during reset, lalala */
-	if (reset_in_progress(execlists))
-		return true;
-
-	if (!execlists->pending[0]) {
-		GEM_TRACE_ERR("%s: Nothing pending for promotion!\n",
-			      engine->name);
-		return false;
-	}
-
-	if (execlists->pending[execlists_num_ports(execlists)]) {
-		GEM_TRACE_ERR("%s: Excess pending[%d] for promotion!\n",
-			      engine->name, execlists_num_ports(execlists));
-		return false;
-	}
-
-	for (port = execlists->pending; (rq = *port); port++) {
-		unsigned long flags;
-		bool ok = true;
-
-		GEM_BUG_ON(!kref_read(&rq->fence.refcount));
-		GEM_BUG_ON(!i915_request_is_active(rq));
-
-		if (ce == rq->context) {
-			GEM_TRACE_ERR("%s: Dup context:%llx in pending[%zd]\n",
-				      engine->name,
-				      ce->timeline->fence_context,
-				      port - execlists->pending);
-			return false;
-		}
-		ce = rq->context;
-
-		if (ccid == ce->lrc.ccid) {
-			GEM_TRACE_ERR("%s: Dup ccid:%x context:%llx in pending[%zd]\n",
-				      engine->name,
-				      ccid, ce->timeline->fence_context,
-				      port - execlists->pending);
-			return false;
-		}
-		ccid = ce->lrc.ccid;
-
-		/*
-		 * Sentinels are supposed to be the last request so they flush
-		 * the current execution off the HW. Check that they are the only
-		 * request in the pending submission.
+	if (i915_vm_is_4lvl(&ppgtt->vm)) {
+		/* 64b PPGTT (48bit canonical)
+		 * PDP0_DESCRIPTOR contains the base address to PML4 and
+		 * other PDP Descriptors are ignored.
 		 */
-		if (sentinel) {
-			GEM_TRACE_ERR("%s: context:%llx after sentinel in pending[%zd]\n",
-				      engine->name,
-				      ce->timeline->fence_context,
-				      port - execlists->pending);
-			return false;
-		}
-		sentinel = i915_request_has_sentinel(rq);
-
-		/* Hold tightly onto the lock to prevent concurrent retires! */
-		if (!spin_trylock_irqsave(&rq->lock, flags))
-			continue;
-
-		if (i915_request_completed(rq))
-			goto unlock;
-
-		if (i915_active_is_idle(&ce->active) &&
-		    !intel_context_is_barrier(ce)) {
-			GEM_TRACE_ERR("%s: Inactive context:%llx in pending[%zd]\n",
-				      engine->name,
-				      ce->timeline->fence_context,
-				      port - execlists->pending);
-			ok = false;
-			goto unlock;
-		}
-
-		if (!i915_vma_is_pinned(ce->state)) {
-			GEM_TRACE_ERR("%s: Unpinned context:%llx in pending[%zd]\n",
-				      engine->name,
-				      ce->timeline->fence_context,
-				      port - execlists->pending);
-			ok = false;
-			goto unlock;
-		}
-
-		if (!i915_vma_is_pinned(ce->ring->vma)) {
-			GEM_TRACE_ERR("%s: Unpinned ring:%llx in pending[%zd]\n",
-				      engine->name,
-				      ce->timeline->fence_context,
-				      port - execlists->pending);
-			ok = false;
-			goto unlock;
-		}
-
-unlock:
-		spin_unlock_irqrestore(&rq->lock, flags);
-		if (!ok)
-			return false;
-	}
-
-	return ce;
-}
-
-static void execlists_submit_ports(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists *execlists = &engine->execlists;
-	unsigned int n;
-
-	GEM_BUG_ON(!assert_pending_valid(execlists, "submit"));
-
-	/*
-	 * We can skip acquiring intel_runtime_pm_get() here as it was taken
-	 * on our behalf by the request (see i915_gem_mark_busy()) and it will
-	 * not be relinquished until the device is idle (see
-	 * i915_gem_idle_work_handler()). As a precaution, we make sure
-	 * that all ELSP are drained i.e. we have processed the CSB,
-	 * before allowing ourselves to idle and calling intel_runtime_pm_put().
-	 */
-	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
-
-	/*
-	 * ELSQ note: the submit queue is not cleared after being submitted
-	 * to the HW so we need to make sure we always clean it up. This is
-	 * currently ensured by the fact that we always write the same number
-	 * of elsq entries, keep this in mind before changing the loop below.
-	 */
-	for (n = execlists_num_ports(execlists); n--; ) {
-		struct i915_request *rq = execlists->pending[n];
-
-		write_desc(execlists,
-			   rq ? execlists_update_context(rq) : 0,
-			   n);
-	}
-
-	/* we need to manually load the submit queue */
-	if (execlists->ctrl_reg)
-		writel(EL_CTRL_LOAD, execlists->ctrl_reg);
-}
-
-static bool ctx_single_port_submission(const struct intel_context *ce)
-{
-	return (IS_ENABLED(CONFIG_DRM_I915_GVT) &&
-		intel_context_force_single_submission(ce));
-}
-
-static bool can_merge_ctx(const struct intel_context *prev,
-			  const struct intel_context *next)
-{
-	if (prev != next)
-		return false;
-
-	if (ctx_single_port_submission(prev))
-		return false;
-
-	return true;
-}
-
-static unsigned long i915_request_flags(const struct i915_request *rq)
-{
-	return READ_ONCE(rq->fence.flags);
-}
-
-static bool can_merge_rq(const struct i915_request *prev,
-			 const struct i915_request *next)
-{
-	GEM_BUG_ON(prev == next);
-	GEM_BUG_ON(!assert_priority_queue(prev, next));
-
-	/*
-	 * We do not submit known completed requests. Therefore if the next
-	 * request is already completed, we can pretend to merge it in
-	 * with the previous context (and we will skip updating the ELSP
-	 * and tracking). Thus hopefully keeping the ELSP full with active
-	 * contexts, despite the best efforts of preempt-to-busy to confuse
-	 * us.
-	 */
-	if (i915_request_completed(next))
-		return true;
-
-	if (unlikely((i915_request_flags(prev) ^ i915_request_flags(next)) &
-		     (BIT(I915_FENCE_FLAG_NOPREEMPT) |
-		      BIT(I915_FENCE_FLAG_SENTINEL))))
-		return false;
-
-	if (!can_merge_ctx(prev->context, next->context))
-		return false;
-
-	GEM_BUG_ON(i915_seqno_passed(prev->fence.seqno, next->fence.seqno));
-	return true;
-}
-
-static void virtual_update_register_offsets(u32 *regs,
-					    struct intel_engine_cs *engine)
-{
-	set_offsets(regs, reg_offsets(engine), engine, false);
-}
-
-static bool virtual_matches(const struct virtual_engine *ve,
-			    const struct i915_request *rq,
-			    const struct intel_engine_cs *engine)
-{
-	const struct intel_engine_cs *inflight;
-
-	if (!(rq->execution_mask & engine->mask)) /* We peeked too soon! */
-		return false;
-
-	/*
-	 * We track when the HW has completed saving the context image
-	 * (i.e. when we have seen the final CS event switching out of
-	 * the context) and must not overwrite the context image before
-	 * then. This restricts us to only using the active engine
-	 * while the previous virtualized request is inflight (so
-	 * we reuse the register offsets). This is a very small
-	 * hystersis on the greedy seelction algorithm.
-	 */
-	inflight = intel_context_inflight(&ve->context);
-	if (inflight && inflight != engine)
-		return false;
-
-	return true;
-}
-
-static void virtual_xfer_context(struct virtual_engine *ve,
-				 struct intel_engine_cs *engine)
-{
-	unsigned int n;
-
-	if (likely(engine == ve->siblings[0]))
-		return;
-
-	GEM_BUG_ON(READ_ONCE(ve->context.inflight));
-	if (!intel_engine_has_relative_mmio(engine))
-		virtual_update_register_offsets(ve->context.lrc_reg_state,
-						engine);
-
-	/*
-	 * Move the bound engine to the top of the list for
-	 * future execution. We then kick this tasklet first
-	 * before checking others, so that we preferentially
-	 * reuse this set of bound registers.
-	 */
-	for (n = 1; n < ve->num_siblings; n++) {
-		if (ve->siblings[n] == engine) {
-			swap(ve->siblings[n], ve->siblings[0]);
-			break;
-		}
+		ASSIGN_CTX_PML4(ppgtt, regs);
+	} else {
+		ASSIGN_CTX_PDP(ppgtt, regs, 3);
+		ASSIGN_CTX_PDP(ppgtt, regs, 2);
+		ASSIGN_CTX_PDP(ppgtt, regs, 1);
+		ASSIGN_CTX_PDP(ppgtt, regs, 0);
 	}
 }
 
-#define for_each_waiter(p__, rq__) \
-	list_for_each_entry_lockless(p__, \
-				     &(rq__)->sched.waiters_list, \
-				     wait_link)
-
-#define for_each_signaler(p__, rq__) \
-	list_for_each_entry_rcu(p__, \
-				&(rq__)->sched.signalers_list, \
-				signal_link)
-
-static void defer_request(struct i915_request *rq, struct list_head * const pl)
-{
-	LIST_HEAD(list);
-
-	/*
-	 * We want to move the interrupted request to the back of
-	 * the round-robin list (i.e. its priority level), but
-	 * in doing so, we must then move all requests that were in
-	 * flight and were waiting for the interrupted request to
-	 * be run after it again.
-	 */
-	do {
-		struct i915_dependency *p;
-
-		GEM_BUG_ON(i915_request_is_active(rq));
-		list_move_tail(&rq->sched.link, pl);
-
-		for_each_waiter(p, rq) {
-			struct i915_request *w =
-				container_of(p->waiter, typeof(*w), sched);
-
-			if (p->flags & I915_DEPENDENCY_WEAK)
-				continue;
-
-			/* Leave semaphores spinning on the other engines */
-			if (w->engine != rq->engine)
-				continue;
-
-			/* No waiter should start before its signaler */
-			GEM_BUG_ON(i915_request_has_initial_breadcrumb(w) &&
-				   i915_request_started(w) &&
-				   !i915_request_completed(rq));
-
-			GEM_BUG_ON(i915_request_is_active(w));
-			if (!i915_request_is_ready(w))
-				continue;
-
-			if (rq_prio(w) < rq_prio(rq))
-				continue;
-
-			GEM_BUG_ON(rq_prio(w) > rq_prio(rq));
-			list_move_tail(&w->sched.link, &list);
-		}
-
-		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
-	} while (rq);
-}
-
-static void defer_active(struct intel_engine_cs *engine)
+static struct i915_ppgtt *vm_alias(struct i915_address_space *vm)
 {
-	struct i915_request *rq;
-
-	rq = __unwind_incomplete_requests(engine);
-	if (!rq)
-		return;
-
-	defer_request(rq, i915_sched_lookup_priolist(engine, rq_prio(rq)));
+	if (i915_is_ggtt(vm))
+		return i915_vm_to_ggtt(vm)->alias;
+	else
+		return i915_vm_to_ppgtt(vm);
 }
 
-static bool
-need_timeslice(const struct intel_engine_cs *engine,
-	       const struct i915_request *rq,
-	       const struct rb_node *rb)
+static void __reset_stop_ring(u32 *regs, const struct intel_engine_cs *engine)
 {
-	int hint;
-
-	if (!intel_engine_has_timeslices(engine))
-		return false;
-
-	hint = engine->execlists.queue_priority_hint;
-
-	if (rb) {
-		const struct virtual_engine *ve =
-			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
-		const struct intel_engine_cs *inflight =
-			intel_context_inflight(&ve->context);
-
-		if (!inflight || inflight == engine) {
-			struct i915_request *next;
+	int x;
 
-			rcu_read_lock();
-			next = READ_ONCE(ve->request);
-			if (next)
-				hint = max(hint, rq_prio(next));
-			rcu_read_unlock();
-		}
+	x = lrc_ring_mi_mode(engine);
+	if (x != -1) {
+		regs[x + 1] &= ~STOP_RING;
+		regs[x + 1] |= STOP_RING << 16;
 	}
-
-	if (!list_is_last(&rq->sched.link, &engine->active.requests))
-		hint = max(hint, rq_prio(list_next_entry(rq, sched.link)));
-
-	GEM_BUG_ON(hint >= I915_PRIORITY_UNPREEMPTABLE);
-	return hint >= effective_prio(rq);
 }
 
-static bool
-timeslice_yield(const struct intel_engine_execlists *el,
-		const struct i915_request *rq)
+static void __lrc_init_regs(u32 *regs,
+			    const struct intel_context *ce,
+			    const struct intel_engine_cs *engine,
+			    bool inhibit)
 {
 	/*
-	 * Once bitten, forever smitten!
+	 * A context is actually a big batch buffer with several
+	 * MI_LOAD_REGISTER_IMM commands followed by (reg, value) pairs. The
+	 * values we are setting here are only for the first context restore:
+	 * on a subsequent save, the GPU will recreate this batchbuffer with new
+	 * values (including all the missing MI_LOAD_REGISTER_IMM commands that
+	 * we are not initializing here).
 	 *
-	 * If the active context ever busy-waited on a semaphore,
-	 * it will be treated as a hog until the end of its timeslice (i.e.
-	 * until it is scheduled out and replaced by a new submission,
-	 * possibly even its own lite-restore). The HW only sends an interrupt
-	 * on the first miss, and we do know if that semaphore has been
-	 * signaled, or even if it is now stuck on another semaphore. Play
-	 * safe, yield if it might be stuck -- it will be given a fresh
-	 * timeslice in the near future.
+	 * Must keep consistent with virtual_update_register_offsets().
 	 */
-	return rq->context->lrc.ccid == READ_ONCE(el->yield);
-}
-
-static bool
-timeslice_expired(const struct intel_engine_execlists *el,
-		  const struct i915_request *rq)
-{
-	return timer_expired(&el->timer) || timeslice_yield(el, rq);
-}
-
-static int
-switch_prio(struct intel_engine_cs *engine, const struct i915_request *rq)
-{
-	if (list_is_last(&rq->sched.link, &engine->active.requests))
-		return engine->execlists.queue_priority_hint;
-
-	return rq_prio(list_next_entry(rq, sched.link));
-}
-
-static inline unsigned long
-timeslice(const struct intel_engine_cs *engine)
-{
-	return READ_ONCE(engine->props.timeslice_duration_ms);
-}
 
-static unsigned long active_timeslice(const struct intel_engine_cs *engine)
-{
-	const struct intel_engine_execlists *execlists = &engine->execlists;
-	const struct i915_request *rq = *execlists->active;
-
-	if (!rq || i915_request_completed(rq))
-		return 0;
-
-	if (READ_ONCE(execlists->switch_priority_hint) < effective_prio(rq))
-		return 0;
-
-	return timeslice(engine);
-}
+	if (inhibit)
+		memset(regs, 0, PAGE_SIZE);
 
-static void set_timeslice(struct intel_engine_cs *engine)
-{
-	unsigned long duration;
+	set_offsets(regs, reg_offsets(engine), engine, inhibit);
 
-	if (!intel_engine_has_timeslices(engine))
-		return;
+	init_common_regs(regs, ce, engine, inhibit);
+	init_ppgtt_regs(regs, vm_alias(ce->vm));
 
-	duration = active_timeslice(engine);
-	ENGINE_TRACE(engine, "bump timeslicing, interval:%lu", duration);
+	init_wa_bb_regs(regs, engine);
 
-	set_timer_ms(&engine->execlists.timer, duration);
+	__reset_stop_ring(regs, engine);
 }
 
-static void start_timeslice(struct intel_engine_cs *engine, int prio)
+void lrc_init_regs(const struct intel_context *ce,
+		   const struct intel_engine_cs *engine,
+		   bool inhibit)
 {
-	struct intel_engine_execlists *execlists = &engine->execlists;
-	unsigned long duration;
-
-	if (!intel_engine_has_timeslices(engine))
-		return;
-
-	WRITE_ONCE(execlists->switch_priority_hint, prio);
-	if (prio == INT_MIN)
-		return;
-
-	if (timer_pending(&execlists->timer))
-		return;
-
-	duration = timeslice(engine);
-	ENGINE_TRACE(engine,
-		     "start timeslicing, prio:%d, interval:%lu",
-		     prio, duration);
-
-	set_timer_ms(&execlists->timer, duration);
+	__lrc_init_regs(ce->lrc_reg_state, ce, engine, inhibit);
 }
 
-static void record_preemption(struct intel_engine_execlists *execlists)
+void lrc_reset_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine)
 {
-	(void)I915_SELFTEST_ONLY(execlists->preempt_hang.count++);
+	__reset_stop_ring(ce->lrc_reg_state, engine);
 }
 
-static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
-					    const struct i915_request *rq)
+static void
+set_redzone(void *vaddr, const struct intel_engine_cs *engine)
 {
-	if (!rq)
-		return 0;
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		return;
 
-	/* Force a fast reset for terminated contexts (ignoring sysfs!) */
-	if (unlikely(intel_context_is_banned(rq->context)))
-		return 1;
+	vaddr += engine->context_size;
 
-	return READ_ONCE(engine->props.preempt_timeout_ms);
+	memset(vaddr, CONTEXT_REDZONE, I915_GTT_PAGE_SIZE);
 }
 
-static void set_preempt_timeout(struct intel_engine_cs *engine,
-				const struct i915_request *rq)
+static void
+check_redzone(const void *vaddr, const struct intel_engine_cs *engine)
 {
-	if (!intel_engine_has_preempt_reset(engine))
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
 		return;
 
-	set_timer_ms(&engine->execlists.preempt,
-		     active_preempt_timeout(engine, rq));
-}
-
-static inline void clear_ports(struct i915_request **ports, int count)
-{
-	memset_p((void **)ports, NULL, count);
-}
+	vaddr += engine->context_size;
 
-static inline void
-copy_ports(struct i915_request **dst, struct i915_request **src, int count)
-{
-	/* A memcpy_p() would be very useful here! */
-	while (count--)
-		WRITE_ONCE(*dst++, *src++); /* avoid write tearing */
+	if (memchr_inv(vaddr, CONTEXT_REDZONE, I915_GTT_PAGE_SIZE))
+		drm_err_once(&engine->i915->drm,
+			     "%s context redzone overwritten!\n",
+			     engine->name);
 }
 
-static void execlists_dequeue(struct intel_engine_cs *engine)
+void lrc_init_state(struct intel_context *ce,
+		    struct intel_engine_cs *engine,
+		    void *state)
 {
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	struct i915_request **port = execlists->pending;
-	struct i915_request ** const last_port = port + execlists->port_mask;
-	struct i915_request * const *active;
-	struct i915_request *last;
-	struct rb_node *rb;
-	bool submit = false;
-
-	/*
-	 * Hardware submission is through 2 ports. Conceptually each port
-	 * has a (RING_START, RING_HEAD, RING_TAIL) tuple. RING_START is
-	 * static for a context, and unique to each, so we only execute
-	 * requests belonging to a single context from each ring. RING_HEAD
-	 * is maintained by the CS in the context image, it marks the place
-	 * where it got up to last time, and through RING_TAIL we tell the CS
-	 * where we want to execute up to this time.
-	 *
-	 * In this list the requests are in order of execution. Consecutive
-	 * requests from the same context are adjacent in the ringbuffer. We
-	 * can combine these requests into a single RING_TAIL update:
-	 *
-	 *              RING_HEAD...req1...req2
-	 *                                    ^- RING_TAIL
-	 * since to execute req2 the CS must first execute req1.
-	 *
-	 * Our goal then is to point each port to the end of a consecutive
-	 * sequence of requests as being the most optimal (fewest wake ups
-	 * and context switches) submission.
-	 */
-
-	for (rb = rb_first_cached(&execlists->virtual); rb; ) {
-		struct virtual_engine *ve =
-			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
-		struct i915_request *rq = READ_ONCE(ve->request);
-
-		if (!rq) { /* lazily cleanup after another engine handled rq */
-			rb_erase_cached(rb, &execlists->virtual);
-			RB_CLEAR_NODE(rb);
-			rb = rb_first_cached(&execlists->virtual);
-			continue;
-		}
-
-		if (!virtual_matches(ve, rq, engine)) {
-			rb = rb_next(rb);
-			continue;
-		}
-
-		break;
-	}
-
-	/*
-	 * If the queue is higher priority than the last
-	 * request in the currently active context, submit afresh.
-	 * We will resubmit again afterwards in case we need to split
-	 * the active context to interject the preemption request,
-	 * i.e. we will retrigger preemption following the ack in case
-	 * of trouble.
-	 */
-	active = READ_ONCE(execlists->active);
-
-	/*
-	 * In theory we can skip over completed contexts that have not
-	 * yet been processed by events (as those events are in flight):
-	 *
-	 * while ((last = *active) && i915_request_completed(last))
-	 *	active++;
-	 *
-	 * However, the GPU cannot handle this as it will ultimately
-	 * find itself trying to jump back into a context it has just
-	 * completed and barf.
-	 */
-
-	if ((last = *active)) {
-		if (need_preempt(engine, last, rb)) {
-			if (i915_request_completed(last)) {
-				tasklet_hi_schedule(&execlists->tasklet);
-				return;
-			}
-
-			ENGINE_TRACE(engine,
-				     "preempting last=%llx:%lld, prio=%d, hint=%d\n",
-				     last->fence.context,
-				     last->fence.seqno,
-				     last->sched.attr.priority,
-				     execlists->queue_priority_hint);
-			record_preemption(execlists);
-
-			/*
-			 * Don't let the RING_HEAD advance past the breadcrumb
-			 * as we unwind (and until we resubmit) so that we do
-			 * not accidentally tell it to go backwards.
-			 */
-			ring_set_paused(engine, 1);
-
-			/*
-			 * Note that we have not stopped the GPU at this point,
-			 * so we are unwinding the incomplete requests as they
-			 * remain inflight and so by the time we do complete
-			 * the preemption, some of the unwound requests may
-			 * complete!
-			 */
-			__unwind_incomplete_requests(engine);
-
-			last = NULL;
-		} else if (need_timeslice(engine, last, rb) &&
-			   timeslice_expired(execlists, last)) {
-			if (i915_request_completed(last)) {
-				tasklet_hi_schedule(&execlists->tasklet);
-				return;
-			}
-
-			ENGINE_TRACE(engine,
-				     "expired last=%llx:%lld, prio=%d, hint=%d, yield?=%s\n",
-				     last->fence.context,
-				     last->fence.seqno,
-				     last->sched.attr.priority,
-				     execlists->queue_priority_hint,
-				     yesno(timeslice_yield(execlists, last)));
-
-			ring_set_paused(engine, 1);
-			defer_active(engine);
-
-			/*
-			 * Unlike for preemption, if we rewind and continue
-			 * executing the same context as previously active,
-			 * the order of execution will remain the same and
-			 * the tail will only advance. We do not need to
-			 * force a full context restore, as a lite-restore
-			 * is sufficient to resample the monotonic TAIL.
-			 *
-			 * If we switch to any other context, similarly we
-			 * will not rewind TAIL of current context, and
-			 * normal save/restore will preserve state and allow
-			 * us to later continue executing the same request.
-			 */
-			last = NULL;
-		} else {
-			/*
-			 * Otherwise if we already have a request pending
-			 * for execution after the current one, we can
-			 * just wait until the next CS event before
-			 * queuing more. In either case we will force a
-			 * lite-restore preemption event, but if we wait
-			 * we hopefully coalesce several updates into a single
-			 * submission.
-			 */
-			if (!list_is_last(&last->sched.link,
-					  &engine->active.requests)) {
-				/*
-				 * Even if ELSP[1] is occupied and not worthy
-				 * of timeslices, our queue might be.
-				 */
-				start_timeslice(engine, queue_prio(execlists));
-				return;
-			}
-		}
-	}
-
-	while (rb) { /* XXX virtual is always taking precedence */
-		struct virtual_engine *ve =
-			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
-		struct i915_request *rq;
-
-		spin_lock(&ve->base.active.lock);
-
-		rq = ve->request;
-		if (unlikely(!rq)) { /* lost the race to a sibling */
-			spin_unlock(&ve->base.active.lock);
-			rb_erase_cached(rb, &execlists->virtual);
-			RB_CLEAR_NODE(rb);
-			rb = rb_first_cached(&execlists->virtual);
-			continue;
-		}
+	bool inhibit = true;
 
-		GEM_BUG_ON(rq != ve->request);
-		GEM_BUG_ON(rq->engine != &ve->base);
-		GEM_BUG_ON(rq->context != &ve->context);
-
-		if (rq_prio(rq) >= queue_prio(execlists)) {
-			if (!virtual_matches(ve, rq, engine)) {
-				spin_unlock(&ve->base.active.lock);
-				rb = rb_next(rb);
-				continue;
-			}
-
-			if (last && !can_merge_rq(last, rq)) {
-				spin_unlock(&ve->base.active.lock);
-				start_timeslice(engine, rq_prio(rq));
-				return; /* leave this for another sibling */
-			}
-
-			ENGINE_TRACE(engine,
-				     "virtual rq=%llx:%lld%s, new engine? %s\n",
-				     rq->fence.context,
-				     rq->fence.seqno,
-				     i915_request_completed(rq) ? "!" :
-				     i915_request_started(rq) ? "*" :
-				     "",
-				     yesno(engine != ve->siblings[0]));
-
-			WRITE_ONCE(ve->request, NULL);
-			WRITE_ONCE(ve->base.execlists.queue_priority_hint,
-				   INT_MIN);
-			rb_erase_cached(rb, &execlists->virtual);
-			RB_CLEAR_NODE(rb);
-
-			GEM_BUG_ON(!(rq->execution_mask & engine->mask));
-			WRITE_ONCE(rq->engine, engine);
-
-			if (__i915_request_submit(rq)) {
-				/*
-				 * Only after we confirm that we will submit
-				 * this request (i.e. it has not already
-				 * completed), do we want to update the context.
-				 *
-				 * This serves two purposes. It avoids
-				 * unnecessary work if we are resubmitting an
-				 * already completed request after timeslicing.
-				 * But more importantly, it prevents us altering
-				 * ve->siblings[] on an idle context, where
-				 * we may be using ve->siblings[] in
-				 * virtual_context_enter / virtual_context_exit.
-				 */
-				virtual_xfer_context(ve, engine);
-				GEM_BUG_ON(ve->siblings[0] != engine);
-
-				submit = true;
-				last = rq;
-			}
-			i915_request_put(rq);
-
-			/*
-			 * Hmm, we have a bunch of virtual engine requests,
-			 * but the first one was already completed (thanks
-			 * preempt-to-busy!). Keep looking at the veng queue
-			 * until we have no more relevant requests (i.e.
-			 * the normal submit queue has higher priority).
-			 */
-			if (!submit) {
-				spin_unlock(&ve->base.active.lock);
-				rb = rb_first_cached(&execlists->virtual);
-				continue;
-			}
-		}
+	set_redzone(state, engine);
 
-		spin_unlock(&ve->base.active.lock);
-		break;
+	if (engine->default_state) {
+		shmem_read(engine->default_state, 0,
+			   state, engine->context_size);
+		__set_bit(CONTEXT_VALID_BIT, &ce->flags);
+		inhibit = false;
 	}
 
-	while ((rb = rb_first_cached(&execlists->queue))) {
-		struct i915_priolist *p = to_priolist(rb);
-		struct i915_request *rq, *rn;
-		int i;
-
-		priolist_for_each_request_consume(rq, rn, p, i) {
-			bool merge = true;
-
-			/*
-			 * Can we combine this request with the current port?
-			 * It has to be the same context/ringbuffer and not
-			 * have any exceptions (e.g. GVT saying never to
-			 * combine contexts).
-			 *
-			 * If we can combine the requests, we can execute both
-			 * by updating the RING_TAIL to point to the end of the
-			 * second request, and so we never need to tell the
-			 * hardware about the first.
-			 */
-			if (last && !can_merge_rq(last, rq)) {
-				/*
-				 * If we are on the second port and cannot
-				 * combine this request with the last, then we
-				 * are done.
-				 */
-				if (port == last_port)
-					goto done;
-
-				/*
-				 * We must not populate both ELSP[] with the
-				 * same LRCA, i.e. we must submit 2 different
-				 * contexts if we submit 2 ELSP.
-				 */
-				if (last->context == rq->context)
-					goto done;
-
-				if (i915_request_has_sentinel(last))
-					goto done;
-
-				/*
-				 * If GVT overrides us we only ever submit
-				 * port[0], leaving port[1] empty. Note that we
-				 * also have to be careful that we don't queue
-				 * the same context (even though a different
-				 * request) to the second port.
-				 */
-				if (ctx_single_port_submission(last->context) ||
-				    ctx_single_port_submission(rq->context))
-					goto done;
-
-				merge = false;
-			}
-
-			if (__i915_request_submit(rq)) {
-				if (!merge) {
-					*port = execlists_schedule_in(last, port - execlists->pending);
-					port++;
-					last = NULL;
-				}
-
-				GEM_BUG_ON(last &&
-					   !can_merge_ctx(last->context,
-							  rq->context));
-				GEM_BUG_ON(last &&
-					   i915_seqno_passed(last->fence.seqno,
-							     rq->fence.seqno));
-
-				submit = true;
-				last = rq;
-			}
-		}
-
-		rb_erase_cached(&p->node, &execlists->queue);
-		i915_priolist_free(p);
-	}
+	/* Clear the ppHWSP (inc. per-context counters) */
+	memset(state, 0, PAGE_SIZE);
 
-done:
 	/*
-	 * Here be a bit of magic! Or sleight-of-hand, whichever you prefer.
-	 *
-	 * We choose the priority hint such that if we add a request of greater
-	 * priority than this, we kick the submission tasklet to decide on
-	 * the right order of submitting the requests to hardware. We must
-	 * also be prepared to reorder requests as they are in-flight on the
-	 * HW. We derive the priority hint then as the first "hole" in
-	 * the HW submission ports and if there are no available slots,
-	 * the priority of the lowest executing request, i.e. last.
-	 *
-	 * When we do receive a higher priority request ready to run from the
-	 * user, see queue_request(), the priority hint is bumped to that
-	 * request triggering preemption on the next dequeue (or subsequent
-	 * interrupt for secondary ports).
+	 * The second page of the context object contains some registers which
+	 * must be set up prior to the first execution.
 	 */
-	execlists->queue_priority_hint = queue_prio(execlists);
-
-	if (submit) {
-		*port = execlists_schedule_in(last, port - execlists->pending);
-		execlists->switch_priority_hint =
-			switch_prio(engine, *execlists->pending);
-
-		/*
-		 * Skip if we ended up with exactly the same set of requests,
-		 * e.g. trying to timeslice a pair of ordered contexts
-		 */
-		if (!memcmp(active, execlists->pending,
-			    (port - execlists->pending + 1) * sizeof(*port))) {
-			do
-				execlists_schedule_out(fetch_and_zero(port));
-			while (port-- != execlists->pending);
-
-			goto skip_submit;
-		}
-		clear_ports(port + 1, last_port - port);
-
-		WRITE_ONCE(execlists->yield, -1);
-		set_preempt_timeout(engine, *active);
-		execlists_submit_ports(engine);
-	} else {
-		start_timeslice(engine, execlists->queue_priority_hint);
-skip_submit:
-		ring_set_paused(engine, 0);
-	}
+	__lrc_init_regs(state + LRC_STATE_OFFSET, ce, engine, inhibit);
 }
 
-static void
-cancel_port_requests(struct intel_engine_execlists * const execlists)
+static struct i915_vma *
+__lrc_alloc_state(struct intel_context *ce, struct intel_engine_cs *engine)
 {
-	struct i915_request * const *port;
-
-	for (port = execlists->pending; *port; port++)
-		execlists_schedule_out(*port);
-	clear_ports(execlists->pending, ARRAY_SIZE(execlists->pending));
-
-	/* Mark the end of active before we overwrite *active */
-	for (port = xchg(&execlists->active, execlists->pending); *port; port++)
-		execlists_schedule_out(*port);
-	clear_ports(execlists->inflight, ARRAY_SIZE(execlists->inflight));
-
-	smp_wmb(); /* complete the seqlock for execlists_active() */
-	WRITE_ONCE(execlists->active, execlists->inflight);
-}
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 context_size;
 
-static inline void
-invalidate_csb_entries(const u64 *first, const u64 *last)
-{
-	clflush((void *)first);
-	clflush((void *)last);
-}
+	context_size = round_up(engine->context_size, I915_GTT_PAGE_SIZE);
 
-/*
- * Starting with Gen12, the status has a new format:
- *
- *     bit  0:     switched to new queue
- *     bit  1:     reserved
- *     bit  2:     semaphore wait mode (poll or signal), only valid when
- *                 switch detail is set to "wait on semaphore"
- *     bits 3-5:   engine class
- *     bits 6-11:  engine instance
- *     bits 12-14: reserved
- *     bits 15-25: sw context id of the lrc the GT switched to
- *     bits 26-31: sw counter of the lrc the GT switched to
- *     bits 32-35: context switch detail
- *                  - 0: ctx complete
- *                  - 1: wait on sync flip
- *                  - 2: wait on vblank
- *                  - 3: wait on scanline
- *                  - 4: wait on semaphore
- *                  - 5: context preempted (not on SEMAPHORE_WAIT or
- *                       WAIT_FOR_EVENT)
- *     bit  36:    reserved
- *     bits 37-43: wait detail (for switch detail 1 to 4)
- *     bits 44-46: reserved
- *     bits 47-57: sw context id of the lrc the GT switched away from
- *     bits 58-63: sw counter of the lrc the GT switched away from
- */
-static inline bool gen12_csb_parse(const u64 csb)
-{
-	bool ctx_away_valid = GEN12_CSB_CTX_VALID(upper_32_bits(csb));
-	bool new_queue =
-		lower_32_bits(csb) & GEN12_CTX_STATUS_SWITCHED_TO_NEW_QUEUE;
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		context_size += I915_GTT_PAGE_SIZE; /* for redzone */
 
-	/*
-	 * The context switch detail is not guaranteed to be 5 when a preemption
-	 * occurs, so we can't just check for that. The check below works for
-	 * all the cases we care about, including preemptions of WAIT
-	 * instructions and lite-restore. Preempt-to-idle via the CTRL register
-	 * would require some extra handling, but we don't support that.
-	 */
-	if (!ctx_away_valid || new_queue) {
-		GEM_BUG_ON(!GEN12_CSB_CTX_VALID(lower_32_bits(csb)));
-		return true;
+	if (INTEL_GEN(engine->i915) == 12) {
+		ce->wa_bb_page = context_size / PAGE_SIZE;
+		context_size += PAGE_SIZE;
 	}
 
-	/*
-	 * switch detail = 5 is covered by the case above and we do not expect a
-	 * context switch on an unsuccessful wait instruction since we always
-	 * use polling mode.
-	 */
-	GEM_BUG_ON(GEN12_CTX_SWITCH_DETAIL(upper_32_bits(csb)));
-	return false;
-}
-
-static inline bool gen8_csb_parse(const u64 csb)
-{
-	return csb & (GEN8_CTX_STATUS_IDLE_ACTIVE | GEN8_CTX_STATUS_PREEMPTED);
-}
-
-static noinline u64
-wa_csb_read(const struct intel_engine_cs *engine, u64 * const csb)
-{
-	u64 entry;
-
-	/*
-	 * Reading from the HWSP has one particular advantage: we can detect
-	 * a stale entry. Since the write into HWSP is broken, we have no reason
-	 * to trust the HW at all, the mmio entry may equally be unordered, so
-	 * we prefer the path that is self-checking and as a last resort,
-	 * return the mmio value.
-	 *
-	 * tgl,dg1:HSDES#22011327657
-	 */
-	preempt_disable();
-	if (wait_for_atomic_us((entry = READ_ONCE(*csb)) != -1, 10)) {
-		int idx = csb - engine->execlists.csb_status;
-		int status;
-
-		status = GEN8_EXECLISTS_STATUS_BUF;
-		if (idx >= 6) {
-			status = GEN11_EXECLISTS_STATUS_BUF2;
-			idx -= 6;
-		}
-		status += sizeof(u64) * idx;
+	obj = i915_gem_object_create_shmem(engine->i915, context_size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
 
-		entry = intel_uncore_read64(engine->uncore,
-					    _MMIO(engine->mmio_base + status));
+	vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
 	}
-	preempt_enable();
-
-	return entry;
-}
 
-static inline u64
-csb_read(const struct intel_engine_cs *engine, u64 * const csb)
-{
-	u64 entry = READ_ONCE(*csb);
-
-	/*
-	 * Unfortunately, the GPU does not always serialise its write
-	 * of the CSB entries before its write of the CSB pointer, at least
-	 * from the perspective of the CPU, using what is known as a Global
-	 * Observation Point. We may read a new CSB tail pointer, but then
-	 * read the stale CSB entries, causing us to misinterpret the
-	 * context-switch events, and eventually declare the GPU hung.
-	 *
-	 * icl:HSDES#1806554093
-	 * tgl:HSDES#22011248461
-	 */
-	if (unlikely(entry == -1))
-		entry = wa_csb_read(engine, csb);
-
-	/* Consume this entry so that we can spot its future reuse. */
-	WRITE_ONCE(*csb, -1);
-
-	/* ELSP is an implicit wmb() before the GPU wraps and overwrites csb */
-	return entry;
+	return vma;
 }
 
-static void process_csb(struct intel_engine_cs *engine)
+static struct intel_timeline *
+pinned_timeline(struct intel_context *ce, struct intel_engine_cs *engine)
 {
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	u64 * const buf = execlists->csb_status;
-	const u8 num_entries = execlists->csb_size;
-	u8 head, tail;
-
-	/*
-	 * As we modify our execlists state tracking we require exclusive
-	 * access. Either we are inside the tasklet, or the tasklet is disabled
-	 * and we assume that is only inside the reset paths and so serialised.
-	 */
-	GEM_BUG_ON(!tasklet_is_locked(&execlists->tasklet) &&
-		   !reset_in_progress(execlists));
-	GEM_BUG_ON(!intel_engine_in_execlists_submission_mode(engine));
-
-	/*
-	 * Note that csb_write, csb_status may be either in HWSP or mmio.
-	 * When reading from the csb_write mmio register, we have to be
-	 * careful to only use the GEN8_CSB_WRITE_PTR portion, which is
-	 * the low 4bits. As it happens we know the next 4bits are always
-	 * zero and so we can simply masked off the low u8 of the register
-	 * and treat it identically to reading from the HWSP (without having
-	 * to use explicit shifting and masking, and probably bifurcating
-	 * the code to handle the legacy mmio read).
-	 */
-	head = execlists->csb_head;
-	tail = READ_ONCE(*execlists->csb_write);
-	if (unlikely(head == tail))
-		return;
-
-	/*
-	 * We will consume all events from HW, or at least pretend to.
-	 *
-	 * The sequence of events from the HW is deterministic, and derived
-	 * from our writes to the ELSP, with a smidgen of variability for
-	 * the arrival of the asynchronous requests wrt to the inflight
-	 * execution. If the HW sends an event that does not correspond with
-	 * the one we are expecting, we have to abandon all hope as we lose
-	 * all tracking of what the engine is actually executing. We will
-	 * only detect we are out of sequence with the HW when we get an
-	 * 'impossible' event because we have already drained our own
-	 * preemption/promotion queue. If this occurs, we know that we likely
-	 * lost track of execution earlier and must unwind and restart, the
-	 * simplest way is by stop processing the event queue and force the
-	 * engine to reset.
-	 */
-	execlists->csb_head = tail;
-	ENGINE_TRACE(engine, "cs-irq head=%d, tail=%d\n", head, tail);
-
-	/*
-	 * Hopefully paired with a wmb() in HW!
-	 *
-	 * We must complete the read of the write pointer before any reads
-	 * from the CSB, so that we do not see stale values. Without an rmb
-	 * (lfence) the HW may speculatively perform the CSB[] reads *before*
-	 * we perform the READ_ONCE(*csb_write).
-	 */
-	rmb();
-	do {
-		bool promote;
-		u64 csb;
-
-		if (++head == num_entries)
-			head = 0;
-
-		/*
-		 * We are flying near dragons again.
-		 *
-		 * We hold a reference to the request in execlist_port[]
-		 * but no more than that. We are operating in softirq
-		 * context and so cannot hold any mutex or sleep. That
-		 * prevents us stopping the requests we are processing
-		 * in port[] from being retired simultaneously (the
-		 * breadcrumb will be complete before we see the
-		 * context-switch). As we only hold the reference to the
-		 * request, any pointer chasing underneath the request
-		 * is subject to a potential use-after-free. Thus we
-		 * store all of the bookkeeping within port[] as
-		 * required, and avoid using unguarded pointers beneath
-		 * request itself. The same applies to the atomic
-		 * status notifier.
-		 */
-
-		csb = csb_read(engine, buf + head);
-		ENGINE_TRACE(engine, "csb[%d]: status=0x%08x:0x%08x\n",
-			     head, upper_32_bits(csb), lower_32_bits(csb));
-
-		if (INTEL_GEN(engine->i915) >= 12)
-			promote = gen12_csb_parse(csb);
-		else
-			promote = gen8_csb_parse(csb);
-		if (promote) {
-			struct i915_request * const *old = execlists->active;
-
-			if (GEM_WARN_ON(!*execlists->pending)) {
-				execlists->error_interrupt |= ERROR_CSB;
-				break;
-			}
-
-			ring_set_paused(engine, 0);
-
-			/* Point active to the new ELSP; prevent overwriting */
-			WRITE_ONCE(execlists->active, execlists->pending);
-			smp_wmb(); /* notify execlists_active() */
-
-			/* cancel old inflight, prepare for switch */
-			trace_ports(execlists, "preempted", old);
-			while (*old)
-				execlists_schedule_out(*old++);
-
-			/* switch pending to inflight */
-			GEM_BUG_ON(!assert_pending_valid(execlists, "promote"));
-			copy_ports(execlists->inflight,
-				   execlists->pending,
-				   execlists_num_ports(execlists));
-			smp_wmb(); /* complete the seqlock */
-			WRITE_ONCE(execlists->active, execlists->inflight);
-
-			/* XXX Magic delay for tgl */
-			ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
-
-			WRITE_ONCE(execlists->pending[0], NULL);
-		} else {
-			if (GEM_WARN_ON(!*execlists->active)) {
-				execlists->error_interrupt |= ERROR_CSB;
-				break;
-			}
-
-			/* port0 completed, advanced to port1 */
-			trace_ports(execlists, "completed", execlists->active);
-
-			/*
-			 * We rely on the hardware being strongly
-			 * ordered, that the breadcrumb write is
-			 * coherent (visible from the CPU) before the
-			 * user interrupt is processed. One might assume
-			 * that the breadcrumb write being before the
-			 * user interrupt and the CS event for the context
-			 * switch would therefore be before the CS event
-			 * itself...
-			 */
-			if (GEM_SHOW_DEBUG() &&
-			    !i915_request_completed(*execlists->active)) {
-				struct i915_request *rq = *execlists->active;
-				const u32 *regs __maybe_unused =
-					rq->context->lrc_reg_state;
-
-				ENGINE_TRACE(engine,
-					     "context completed before request!\n");
-				ENGINE_TRACE(engine,
-					     "ring:{start:0x%08x, head:%04x, tail:%04x, ctl:%08x, mode:%08x}\n",
-					     ENGINE_READ(engine, RING_START),
-					     ENGINE_READ(engine, RING_HEAD) & HEAD_ADDR,
-					     ENGINE_READ(engine, RING_TAIL) & TAIL_ADDR,
-					     ENGINE_READ(engine, RING_CTL),
-					     ENGINE_READ(engine, RING_MI_MODE));
-				ENGINE_TRACE(engine,
-					     "rq:{start:%08x, head:%04x, tail:%04x, seqno:%llx:%d, hwsp:%d}, ",
-					     i915_ggtt_offset(rq->ring->vma),
-					     rq->head, rq->tail,
-					     rq->fence.context,
-					     lower_32_bits(rq->fence.seqno),
-					     hwsp_seqno(rq));
-				ENGINE_TRACE(engine,
-					     "ctx:{start:%08x, head:%04x, tail:%04x}, ",
-					     regs[CTX_RING_START],
-					     regs[CTX_RING_HEAD],
-					     regs[CTX_RING_TAIL]);
-			}
-
-			execlists_schedule_out(*execlists->active++);
-
-			GEM_BUG_ON(execlists->active - execlists->inflight >
-				   execlists_num_ports(execlists));
-		}
-	} while (head != tail);
-
-	set_timeslice(engine);
-
-	/*
-	 * Gen11 has proven to fail wrt global observation point between
-	 * entry and tail update, failing on the ordering and thus
-	 * we see an old entry in the context status buffer.
-	 *
-	 * Forcibly evict out entries for the next gpu csb update,
-	 * to increase the odds that we get a fresh entries with non
-	 * working hardware. The cost for doing so comes out mostly with
-	 * the wash as hardware, working or not, will need to do the
-	 * invalidation before.
-	 */
-	invalidate_csb_entries(&buf[0], &buf[num_entries - 1]);
-}
+	struct intel_timeline *tl = fetch_and_zero(&ce->timeline);
 
-static void __execlists_submission_tasklet(struct intel_engine_cs *const engine)
-{
-	lockdep_assert_held(&engine->active.lock);
-	if (!READ_ONCE(engine->execlists.pending[0])) {
-		rcu_read_lock(); /* protect peeking at execlists->active */
-		execlists_dequeue(engine);
-		rcu_read_unlock();
-	}
+	return intel_timeline_create_from_engine(engine, page_unmask_bits(tl));
 }
 
-static void __execlists_hold(struct i915_request *rq)
+int lrc_alloc(struct intel_context *ce, struct intel_engine_cs *engine)
 {
-	LIST_HEAD(list);
-
-	do {
-		struct i915_dependency *p;
-
-		if (i915_request_is_active(rq))
-			__i915_request_unsubmit(rq);
-
-		clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
-		list_move_tail(&rq->sched.link, &rq->engine->active.hold);
-		i915_request_set_hold(rq);
-		RQ_TRACE(rq, "on hold\n");
-
-		for_each_waiter(p, rq) {
-			struct i915_request *w =
-				container_of(p->waiter, typeof(*w), sched);
-
-			/* Leave semaphores spinning on the other engines */
-			if (w->engine != rq->engine)
-				continue;
-
-			if (!i915_request_is_ready(w))
-				continue;
-
-			if (i915_request_completed(w))
-				continue;
-
-			if (i915_request_on_hold(w))
-				continue;
-
-			list_move_tail(&w->sched.link, &list);
-		}
-
-		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
-	} while (rq);
-}
+	struct intel_ring *ring;
+	struct i915_vma *vma;
+	int err;
 
-static bool execlists_hold(struct intel_engine_cs *engine,
-			   struct i915_request *rq)
-{
-	if (i915_request_on_hold(rq))
-		return false;
+	GEM_BUG_ON(ce->state);
 
-	spin_lock_irq(&engine->active.lock);
+	vma = __lrc_alloc_state(ce, engine);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
 
-	if (i915_request_completed(rq)) { /* too late! */
-		rq = NULL;
-		goto unlock;
+	ring = intel_engine_create_ring(engine, (unsigned long)ce->ring);
+	if (IS_ERR(ring)) {
+		err = PTR_ERR(ring);
+		goto err_vma;
 	}
 
-	if (rq->engine != engine) { /* preempted virtual engine */
-		struct virtual_engine *ve = to_virtual_engine(rq->engine);
-
-		/*
-		 * intel_context_inflight() is only protected by virtue
-		 * of process_csb() being called only by the tasklet (or
-		 * directly from inside reset while the tasklet is suspended).
-		 * Assert that neither of those are allowed to run while we
-		 * poke at the request queues.
-		 */
-		GEM_BUG_ON(!reset_in_progress(&engine->execlists));
+	if (!page_mask_bits(ce->timeline)) {
+		struct intel_timeline *tl;
 
 		/*
-		 * An unsubmitted request along a virtual engine will
-		 * remain on the active (this) engine until we are able
-		 * to process the context switch away (and so mark the
-		 * context as no longer in flight). That cannot have happened
-		 * yet, otherwise we would not be hanging!
+		 * Use the static global HWSP for the kernel context, and
+		 * a dynamically allocated cacheline for everyone else.
 		 */
-		spin_lock(&ve->base.active.lock);
-		GEM_BUG_ON(intel_context_inflight(rq->context) != engine);
-		GEM_BUG_ON(ve->request != rq);
-		ve->request = NULL;
-		spin_unlock(&ve->base.active.lock);
-		i915_request_put(rq);
-
-		rq->engine = engine;
-	}
-
-	/*
-	 * Transfer this request onto the hold queue to prevent it
-	 * being resumbitted to HW (and potentially completed) before we have
-	 * released it. Since we may have already submitted following
-	 * requests, we need to remove those as well.
-	 */
-	GEM_BUG_ON(i915_request_on_hold(rq));
-	GEM_BUG_ON(rq->engine != engine);
-	__execlists_hold(rq);
-	GEM_BUG_ON(list_empty(&engine->active.hold));
-
-unlock:
-	spin_unlock_irq(&engine->active.lock);
-	return rq;
-}
-
-static bool hold_request(const struct i915_request *rq)
-{
-	struct i915_dependency *p;
-	bool result = false;
-
-	/*
-	 * If one of our ancestors is on hold, we must also be on hold,
-	 * otherwise we will bypass it and execute before it.
-	 */
-	rcu_read_lock();
-	for_each_signaler(p, rq) {
-		const struct i915_request *s =
-			container_of(p->signaler, typeof(*s), sched);
-
-		if (s->engine != rq->engine)
-			continue;
-
-		result = i915_request_on_hold(s);
-		if (result)
-			break;
-	}
-	rcu_read_unlock();
-
-	return result;
-}
-
-static void __execlists_unhold(struct i915_request *rq)
-{
-	LIST_HEAD(list);
-
-	do {
-		struct i915_dependency *p;
-
-		RQ_TRACE(rq, "hold release\n");
-
-		GEM_BUG_ON(!i915_request_on_hold(rq));
-		GEM_BUG_ON(!i915_sw_fence_signaled(&rq->submit));
-
-		i915_request_clear_hold(rq);
-		list_move_tail(&rq->sched.link,
-			       i915_sched_lookup_priolist(rq->engine,
-							  rq_prio(rq)));
-		set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
-
-		/* Also release any children on this engine that are ready */
-		for_each_waiter(p, rq) {
-			struct i915_request *w =
-				container_of(p->waiter, typeof(*w), sched);
-
-			/* Propagate any change in error status */
-			if (rq->fence.error)
-				i915_request_set_error_once(w, rq->fence.error);
-
-			if (w->engine != rq->engine)
-				continue;
-
-			if (!i915_request_on_hold(w))
-				continue;
-
-			/* Check that no other parents are also on hold */
-			if (hold_request(w))
-				continue;
-
-			list_move_tail(&w->sched.link, &list);
-		}
-
-		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
-	} while (rq);
-}
-
-static void execlists_unhold(struct intel_engine_cs *engine,
-			     struct i915_request *rq)
-{
-	spin_lock_irq(&engine->active.lock);
-
-	/*
-	 * Move this request back to the priority queue, and all of its
-	 * children and grandchildren that were suspended along with it.
-	 */
-	__execlists_unhold(rq);
-
-	if (rq_prio(rq) > engine->execlists.queue_priority_hint) {
-		engine->execlists.queue_priority_hint = rq_prio(rq);
-		tasklet_hi_schedule(&engine->execlists.tasklet);
-	}
-
-	spin_unlock_irq(&engine->active.lock);
-}
-
-struct execlists_capture {
-	struct work_struct work;
-	struct i915_request *rq;
-	struct i915_gpu_coredump *error;
-};
-
-static void execlists_capture_work(struct work_struct *work)
-{
-	struct execlists_capture *cap = container_of(work, typeof(*cap), work);
-	const gfp_t gfp = GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN;
-	struct intel_engine_cs *engine = cap->rq->engine;
-	struct intel_gt_coredump *gt = cap->error->gt;
-	struct intel_engine_capture_vma *vma;
-
-	/* Compress all the objects attached to the request, slow! */
-	vma = intel_engine_coredump_add_request(gt->engine, cap->rq, gfp);
-	if (vma) {
-		struct i915_vma_compress *compress =
-			i915_vma_capture_prepare(gt);
-
-		intel_engine_coredump_add_vma(gt->engine, vma, compress);
-		i915_vma_capture_finish(gt, compress);
-	}
-
-	gt->simulated = gt->engine->simulated;
-	cap->error->simulated = gt->simulated;
-
-	/* Publish the error state, and announce it to the world */
-	i915_error_state_store(cap->error);
-	i915_gpu_coredump_put(cap->error);
-
-	/* Return this request and all that depend upon it for signaling */
-	execlists_unhold(engine, cap->rq);
-	i915_request_put(cap->rq);
-
-	kfree(cap);
-}
-
-static struct execlists_capture *capture_regs(struct intel_engine_cs *engine)
-{
-	const gfp_t gfp = GFP_ATOMIC | __GFP_NOWARN;
-	struct execlists_capture *cap;
-
-	cap = kmalloc(sizeof(*cap), gfp);
-	if (!cap)
-		return NULL;
-
-	cap->error = i915_gpu_coredump_alloc(engine->i915, gfp);
-	if (!cap->error)
-		goto err_cap;
-
-	cap->error->gt = intel_gt_coredump_alloc(engine->gt, gfp);
-	if (!cap->error->gt)
-		goto err_gpu;
-
-	cap->error->gt->engine = intel_engine_coredump_alloc(engine, gfp);
-	if (!cap->error->gt->engine)
-		goto err_gt;
-
-	cap->error->gt->engine->hung = true;
-
-	return cap;
-
-err_gt:
-	kfree(cap->error->gt);
-err_gpu:
-	kfree(cap->error);
-err_cap:
-	kfree(cap);
-	return NULL;
-}
-
-static struct i915_request *
-active_context(struct intel_engine_cs *engine, u32 ccid)
-{
-	const struct intel_engine_execlists * const el = &engine->execlists;
-	struct i915_request * const *port, *rq;
-
-	/*
-	 * Use the most recent result from process_csb(), but just in case
-	 * we trigger an error (via interrupt) before the first CS event has
-	 * been written, peek at the next submission.
-	 */
-
-	for (port = el->active; (rq = *port); port++) {
-		if (rq->context->lrc.ccid == ccid) {
-			ENGINE_TRACE(engine,
-				     "ccid found at active:%zd\n",
-				     port - el->active);
-			return rq;
-		}
-	}
-
-	for (port = el->pending; (rq = *port); port++) {
-		if (rq->context->lrc.ccid == ccid) {
-			ENGINE_TRACE(engine,
-				     "ccid found at pending:%zd\n",
-				     port - el->pending);
-			return rq;
+		if (unlikely(ce->timeline))
+			tl = pinned_timeline(ce, engine);
+		else
+			tl = intel_timeline_create(engine->gt);
+		if (IS_ERR(tl)) {
+			err = PTR_ERR(tl);
+			goto err_ring;
 		}
-	}
-
-	ENGINE_TRACE(engine, "ccid:%x not found\n", ccid);
-	return NULL;
-}
-
-static u32 active_ccid(struct intel_engine_cs *engine)
-{
-	return ENGINE_READ_FW(engine, RING_EXECLIST_STATUS_HI);
-}
-
-static void execlists_capture(struct intel_engine_cs *engine)
-{
-	struct execlists_capture *cap;
-
-	if (!IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR))
-		return;
-
-	/*
-	 * We need to _quickly_ capture the engine state before we reset.
-	 * We are inside an atomic section (softirq) here and we are delaying
-	 * the forced preemption event.
-	 */
-	cap = capture_regs(engine);
-	if (!cap)
-		return;
 
-	spin_lock_irq(&engine->active.lock);
-	cap->rq = active_context(engine, active_ccid(engine));
-	if (cap->rq) {
-		cap->rq = active_request(cap->rq->context->timeline, cap->rq);
-		cap->rq = i915_request_get_rcu(cap->rq);
+		ce->timeline = tl;
 	}
-	spin_unlock_irq(&engine->active.lock);
-	if (!cap->rq)
-		goto err_free;
-
-	/*
-	 * Remove the request from the execlists queue, and take ownership
-	 * of the request. We pass it to our worker who will _slowly_ compress
-	 * all the pages the _user_ requested for debugging their batch, after
-	 * which we return it to the queue for signaling.
-	 *
-	 * By removing them from the execlists queue, we also remove the
-	 * requests from being processed by __unwind_incomplete_requests()
-	 * during the intel_engine_reset(), and so they will *not* be replayed
-	 * afterwards.
-	 *
-	 * Note that because we have not yet reset the engine at this point,
-	 * it is possible for the request that we have identified as being
-	 * guilty, did in fact complete and we will then hit an arbitration
-	 * point allowing the outstanding preemption to succeed. The likelihood
-	 * of that is very low (as capturing of the engine registers should be
-	 * fast enough to run inside an irq-off atomic section!), so we will
-	 * simply hold that request accountable for being non-preemptible
-	 * long enough to force the reset.
-	 */
-	if (!execlists_hold(engine, cap->rq))
-		goto err_rq;
-
-	INIT_WORK(&cap->work, execlists_capture_work);
-	schedule_work(&cap->work);
-	return;
-
-err_rq:
-	i915_request_put(cap->rq);
-err_free:
-	i915_gpu_coredump_put(cap->error);
-	kfree(cap);
-}
-
-static void execlists_reset(struct intel_engine_cs *engine, const char *msg)
-{
-	const unsigned int bit = I915_RESET_ENGINE + engine->id;
-	unsigned long *lock = &engine->gt->reset.flags;
 
-	if (!intel_has_reset_engine(engine->gt))
-		return;
-
-	if (test_and_set_bit(bit, lock))
-		return;
-
-	ENGINE_TRACE(engine, "reset for %s\n", msg);
-
-	/* Mark this tasklet as disabled to avoid waiting for it to complete */
-	tasklet_disable_nosync(&engine->execlists.tasklet);
+	ce->ring = ring;
+	ce->state = vma;
 
-	ring_set_paused(engine, 1); /* Freeze the current request in place */
-	execlists_capture(engine);
-	intel_engine_reset(engine, msg);
+	return 0;
 
-	tasklet_enable(&engine->execlists.tasklet);
-	clear_and_wake_up_bit(bit, lock);
+err_ring:
+	intel_ring_put(ring);
+err_vma:
+	i915_vma_put(vma);
+	return err;
 }
 
-static bool preempt_timeout(const struct intel_engine_cs *const engine)
+void lrc_reset(struct intel_context *ce)
 {
-	const struct timer_list *t = &engine->execlists.preempt;
-
-	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
-		return false;
+	GEM_BUG_ON(!intel_context_is_pinned(ce));
 
-	if (!timer_expired(t))
-		return false;
+	intel_ring_reset(ce->ring, ce->ring->emit);
 
-	return READ_ONCE(engine->execlists.pending[0]);
+	/* Scrub away the garbage */
+	lrc_init_regs(ce, ce->engine, true);
+	ce->lrc.lrca = lrc_update_regs(ce, ce->engine, ce->ring->tail);
 }
 
-/*
- * Check the unread Context Status Buffers and manage the submission of new
- * contexts to the ELSP accordingly.
- */
-static void execlists_submission_tasklet(unsigned long data)
+int
+lrc_pre_pin(struct intel_context *ce,
+	    struct intel_engine_cs *engine,
+	    struct i915_gem_ww_ctx *ww,
+	    void **vaddr)
 {
-	struct intel_engine_cs * const engine = (struct intel_engine_cs *)data;
-	bool timeout = preempt_timeout(engine);
-
-	process_csb(engine);
-
-	if (unlikely(READ_ONCE(engine->execlists.error_interrupt))) {
-		const char *msg;
-
-		/* Generate the error message in priority wrt to the user! */
-		if (engine->execlists.error_interrupt & GENMASK(15, 0))
-			msg = "CS error"; /* thrown by a user payload */
-		else if (engine->execlists.error_interrupt & ERROR_CSB)
-			msg = "invalid CSB event";
-		else
-			msg = "internal error";
-
-		engine->execlists.error_interrupt = 0;
-		execlists_reset(engine, msg);
-	}
-
-	if (!READ_ONCE(engine->execlists.pending[0]) || timeout) {
-		unsigned long flags;
+	GEM_BUG_ON(!ce->state);
+	GEM_BUG_ON(!i915_vma_is_pinned(ce->state));
 
-		spin_lock_irqsave(&engine->active.lock, flags);
-		__execlists_submission_tasklet(engine);
-		spin_unlock_irqrestore(&engine->active.lock, flags);
+	*vaddr = i915_gem_object_pin_map(ce->state->obj,
+					 i915_coherent_map_type(ce->engine->i915) |
+					 I915_MAP_OVERRIDE);
 
-		/* Recheck after serialising with direct-submission */
-		if (unlikely(timeout && preempt_timeout(engine))) {
-			cancel_timer(&engine->execlists.preempt);
-			execlists_reset(engine, "preemption time out");
-		}
-	}
+	return PTR_ERR_OR_ZERO(*vaddr);
 }
 
-static void __execlists_kick(struct intel_engine_execlists *execlists)
+int
+lrc_pin(struct intel_context *ce,
+	struct intel_engine_cs *engine,
+	void *vaddr)
 {
-	/* Kick the tasklet for some interrupt coalescing and reset handling */
-	tasklet_hi_schedule(&execlists->tasklet);
-}
+	ce->lrc_reg_state = vaddr + LRC_STATE_OFFSET;
 
-#define execlists_kick(t, member) \
-	__execlists_kick(container_of(t, struct intel_engine_execlists, member))
+	if (!__test_and_set_bit(CONTEXT_INIT_BIT, &ce->flags))
+		lrc_init_state(ce, engine, vaddr);
 
-static void execlists_timeslice(struct timer_list *timer)
-{
-	execlists_kick(timer, timer);
-}
-
-static void execlists_preempt(struct timer_list *timer)
-{
-	execlists_kick(timer, preempt);
+	ce->lrc.lrca = lrc_update_regs(ce, engine, ce->ring->tail);
+	return 0;
 }
 
-static void queue_request(struct intel_engine_cs *engine,
-			  struct i915_request *rq)
+void lrc_unpin(struct intel_context *ce)
 {
-	GEM_BUG_ON(!list_empty(&rq->sched.link));
-	list_add_tail(&rq->sched.link,
-		      i915_sched_lookup_priolist(engine, rq_prio(rq)));
-	set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+	check_redzone((void *)ce->lrc_reg_state - LRC_STATE_OFFSET,
+		      ce->engine);
 }
 
-static void __submit_queue_imm(struct intel_engine_cs *engine)
+void lrc_post_unpin(struct intel_context *ce)
 {
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-
-	if (reset_in_progress(execlists))
-		return; /* defer until we restart the engine following reset */
-
-	__execlists_submission_tasklet(engine);
+	i915_gem_object_unpin_map(ce->state->obj);
 }
 
-static void submit_queue(struct intel_engine_cs *engine,
-			 const struct i915_request *rq)
+void lrc_fini(struct intel_context *ce)
 {
-	struct intel_engine_execlists *execlists = &engine->execlists;
-
-	if (rq_prio(rq) <= execlists->queue_priority_hint)
+	if (!ce->state)
 		return;
 
-	execlists->queue_priority_hint = rq_prio(rq);
-	__submit_queue_imm(engine);
+	intel_ring_put(fetch_and_zero(&ce->ring));
+	i915_vma_put(fetch_and_zero(&ce->state));
 }
 
-static bool ancestor_on_hold(const struct intel_engine_cs *engine,
-			     const struct i915_request *rq)
-{
-	GEM_BUG_ON(i915_request_on_hold(rq));
-	return !list_empty(&engine->active.hold) && hold_request(rq);
-}
-
-static void flush_csb(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists *el = &engine->execlists;
-
-	if (READ_ONCE(el->pending[0]) && tasklet_trylock(&el->tasklet)) {
-		if (!reset_in_progress(el))
-			process_csb(engine);
-		tasklet_unlock(&el->tasklet);
-	}
-}
-
-static void execlists_submit_request(struct i915_request *request)
-{
-	struct intel_engine_cs *engine = request->engine;
-	unsigned long flags;
-
-	/* Hopefully we clear execlists->pending[] to let us through */
-	flush_csb(engine);
-
-	/* Will be called from irq-context when using foreign fences. */
-	spin_lock_irqsave(&engine->active.lock, flags);
-
-	if (unlikely(ancestor_on_hold(engine, request))) {
-		RQ_TRACE(request, "ancestor on hold\n");
-		list_add_tail(&request->sched.link, &engine->active.hold);
-		i915_request_set_hold(request);
-	} else {
-		queue_request(engine, request);
-
-		GEM_BUG_ON(RB_EMPTY_ROOT(&engine->execlists.queue.rb_root));
-		GEM_BUG_ON(list_empty(&request->sched.link));
-
-		submit_queue(engine, request);
-	}
-
-	spin_unlock_irqrestore(&engine->active.lock, flags);
-}
-
-static void __execlists_context_fini(struct intel_context *ce)
-{
-	intel_ring_put(ce->ring);
-	i915_vma_put(ce->state);
-}
-
-static void execlists_context_destroy(struct kref *kref)
+void lrc_destroy(struct kref *kref)
 {
 	struct intel_context *ce = container_of(kref, typeof(*ce), ref);
 
 	GEM_BUG_ON(!i915_active_is_idle(&ce->active));
 	GEM_BUG_ON(intel_context_is_pinned(ce));
 
-	if (ce->state)
-		__execlists_context_fini(ce);
+	lrc_fini(ce);
 
 	intel_context_fini(ce);
 	intel_context_free(ce);
 }
 
-static void
-set_redzone(void *vaddr, const struct intel_engine_cs *engine)
-{
-	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		return;
-
-	vaddr += engine->context_size;
-
-	memset(vaddr, CONTEXT_REDZONE, I915_GTT_PAGE_SIZE);
-}
-
-static void
-check_redzone(const void *vaddr, const struct intel_engine_cs *engine)
-{
-	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		return;
-
-	vaddr += engine->context_size;
-
-	if (memchr_inv(vaddr, CONTEXT_REDZONE, I915_GTT_PAGE_SIZE))
-		drm_err_once(&engine->i915->drm,
-			     "%s context redzone overwritten!\n",
-			     engine->name);
-}
-
-static void execlists_context_unpin(struct intel_context *ce)
-{
-	check_redzone((void *)ce->lrc_reg_state - LRC_STATE_OFFSET,
-		      ce->engine);
-}
-
-static void execlists_context_post_unpin(struct intel_context *ce)
-{
-	i915_gem_object_unpin_map(ce->state->obj);
-}
-
 static u32 *
 gen12_emit_timestamp_wa(const struct intel_context *ce, u32 *cs)
 {
@@ -3465,7 +1035,7 @@ gen12_emit_indirect_ctx_xcs(const struct intel_context *ce, u32 *cs)
 	return cs;
 }
 
-static inline u32 context_wa_bb_offset(const struct intel_context *ce)
+static u32 context_wa_bb_offset(const struct intel_context *ce)
 {
 	return PAGE_SIZE * ce->wa_bb_page;
 }
@@ -3496,16 +1066,57 @@ setup_indirect_ctx_bb(const struct intel_context *ce,
 	while ((unsigned long)cs % CACHELINE_BYTES)
 		*cs++ = MI_NOOP;
 
-	lrc_ring_setup_indirect_ctx(ce->lrc_reg_state, engine,
-				    i915_ggtt_offset(ce->state) +
-				    context_wa_bb_offset(ce),
-				    (cs - start) * sizeof(*cs));
+	lrc_setup_indirect_ctx(ce->lrc_reg_state, engine,
+			       i915_ggtt_offset(ce->state) +
+			       context_wa_bb_offset(ce),
+			       (cs - start) * sizeof(*cs));
 }
 
-static void
-__execlists_update_reg_state(const struct intel_context *ce,
-			     const struct intel_engine_cs *engine,
-			     u32 head)
+/*
+ * The context descriptor encodes various attributes of a context,
+ * including its GTT address and some flags. Because it's fairly
+ * expensive to calculate, we'll just do it once and cache the result,
+ * which remains valid until the context is unpinned.
+ *
+ * This is what a descriptor looks like, from LSB to MSB::
+ *
+ *      bits  0-11:    flags, GEN8_CTX_* (cached in ctx->desc_template)
+ *      bits 12-31:    LRCA, GTT address of (the HWSP of) this context
+ *      bits 32-52:    ctx ID, a globally unique tag (highest bit used by GuC)
+ *      bits 53-54:    mbz, reserved for use by hardware
+ *      bits 55-63:    group ID, currently unused and set to 0
+ *
+ * Starting from Gen11, the upper dword of the descriptor has a new format:
+ *
+ *      bits 32-36:    reserved
+ *      bits 37-47:    SW context ID
+ *      bits 48:53:    engine instance
+ *      bit 54:        mbz, reserved for use by hardware
+ *      bits 55-60:    SW counter
+ *      bits 61-63:    engine class
+ *
+ * engine info, SW context ID and SW counter need to form a unique number
+ * (Context ID) per lrc.
+ */
+static u32 lrc_descriptor(const struct intel_context *ce)
+{
+	u32 desc;
+
+	desc = INTEL_LEGACY_32B_CONTEXT;
+	if (i915_vm_is_4lvl(ce->vm))
+		desc = INTEL_LEGACY_64B_CONTEXT;
+	desc <<= GEN8_CTX_ADDRESSING_MODE_SHIFT;
+
+	desc |= GEN8_CTX_VALID | GEN8_CTX_PRIVILEGE;
+	if (IS_GEN(ce->vm->i915, 8))
+		desc |= GEN8_CTX_L3LLC_COHERENT;
+
+	return i915_ggtt_offset(ce->state) | desc;
+}
+
+u32 lrc_update_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    u32 head)
 {
 	struct intel_ring *ring = ce->ring;
 	u32 *regs = ce->lrc_reg_state;
@@ -3537,205 +1148,54 @@ __execlists_update_reg_state(const struct intel_context *ce,
 		GEM_BUG_ON(engine->wa_ctx.indirect_ctx.size);
 		setup_indirect_ctx_bb(ce, engine, fn);
 	}
-}
 
-static int
-execlists_context_pre_pin(struct intel_context *ce,
-			  struct i915_gem_ww_ctx *ww, void **vaddr)
-{
-	GEM_BUG_ON(!ce->state);
-	GEM_BUG_ON(!i915_vma_is_pinned(ce->state));
-
-	*vaddr = i915_gem_object_pin_map(ce->state->obj,
-					i915_coherent_map_type(ce->engine->i915) |
-					I915_MAP_OVERRIDE);
-
-	return PTR_ERR_OR_ZERO(*vaddr);
-}
-
-static int
-__execlists_context_pin(struct intel_context *ce,
-			struct intel_engine_cs *engine,
-			void *vaddr)
-{
-	ce->lrc.lrca = lrc_descriptor(ce, engine) | CTX_DESC_FORCE_RESTORE;
-	ce->lrc_reg_state = vaddr + LRC_STATE_OFFSET;
-	__execlists_update_reg_state(ce, engine, ce->ring->tail);
-
-	return 0;
-}
-
-static int execlists_context_pin(struct intel_context *ce, void *vaddr)
-{
-	return __execlists_context_pin(ce, ce->engine, vaddr);
+	return lrc_descriptor(ce) | CTX_DESC_FORCE_RESTORE;
 }
 
-static int execlists_context_alloc(struct intel_context *ce)
+void lrc_update_offsets(struct intel_context *ce,
+			struct intel_engine_cs *engine)
 {
-	return __execlists_context_alloc(ce, ce->engine);
+	set_offsets(ce->lrc_reg_state, reg_offsets(engine), engine, false);
 }
 
-static void execlists_context_reset(struct intel_context *ce)
+void lrc_check_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    const char *when)
 {
-	CE_TRACE(ce, "reset\n");
-	GEM_BUG_ON(!intel_context_is_pinned(ce));
-
-	intel_ring_reset(ce->ring, ce->ring->emit);
-
-	/* Scrub away the garbage */
-	execlists_init_reg_state(ce->lrc_reg_state,
-				 ce, ce->engine, ce->ring, true);
-	__execlists_update_reg_state(ce, ce->engine, ce->ring->tail);
-
-	ce->lrc.desc |= CTX_DESC_FORCE_RESTORE;
-}
-
-static const struct intel_context_ops execlists_context_ops = {
-	.alloc = execlists_context_alloc,
-
-	.pre_pin = execlists_context_pre_pin,
-	.pin = execlists_context_pin,
-	.unpin = execlists_context_unpin,
-	.post_unpin = execlists_context_post_unpin,
-
-	.enter = intel_context_enter_engine,
-	.exit = intel_context_exit_engine,
-
-	.reset = execlists_context_reset,
-	.destroy = execlists_context_destroy,
-};
-
-static u32 hwsp_offset(const struct i915_request *rq)
-{
-	const struct intel_timeline_cacheline *cl;
-
-	/* Before the request is executed, the timeline/cachline is fixed */
-
-	cl = rcu_dereference_protected(rq->hwsp_cacheline, 1);
-	if (cl)
-		return cl->ggtt_offset;
-
-	return rcu_dereference_protected(rq->timeline, 1)->hwsp_offset;
-}
-
-static int gen8_emit_init_breadcrumb(struct i915_request *rq)
-{
-	u32 *cs;
-
-	GEM_BUG_ON(i915_request_has_initial_breadcrumb(rq));
-	if (!i915_request_timeline(rq)->has_initial_breadcrumb)
-		return 0;
-
-	cs = intel_ring_begin(rq, 6);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	/*
-	 * Check if we have been preempted before we even get started.
-	 *
-	 * After this point i915_request_started() reports true, even if
-	 * we get preempted and so are no longer running.
-	 */
-	*cs++ = MI_ARB_CHECK;
-	*cs++ = MI_NOOP;
-
-	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-	*cs++ = hwsp_offset(rq);
-	*cs++ = 0;
-	*cs++ = rq->fence.seqno - 1;
-
-	intel_ring_advance(rq, cs);
-
-	/* Record the updated position of the request's payload */
-	rq->infix = intel_ring_offset(rq, cs);
-
-	__set_bit(I915_FENCE_FLAG_INITIAL_BREADCRUMB, &rq->fence.flags);
-
-	return 0;
-}
-
-static int emit_pdps(struct i915_request *rq)
-{
-	const struct intel_engine_cs * const engine = rq->engine;
-	struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(rq->context->vm);
-	int err, i;
-	u32 *cs;
-
-	GEM_BUG_ON(intel_vgpu_active(rq->engine->i915));
-
-	/*
-	 * Beware ye of the dragons, this sequence is magic!
-	 *
-	 * Small changes to this sequence can cause anything from
-	 * GPU hangs to forcewake errors and machine lockups!
-	 */
-
-	/* Flush any residual operations from the context load */
-	err = engine->emit_flush(rq, EMIT_FLUSH);
-	if (err)
-		return err;
+	const struct intel_ring *ring = ce->ring;
+	u32 *regs = ce->lrc_reg_state;
+	bool valid = true;
+	int x;
 
-	/* Magic required to prevent forcewake errors! */
-	err = engine->emit_flush(rq, EMIT_INVALIDATE);
-	if (err)
-		return err;
-
-	cs = intel_ring_begin(rq, 4 * GEN8_3LVL_PDPES + 2);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	/* Ensure the LRI have landed before we invalidate & continue */
-	*cs++ = MI_LOAD_REGISTER_IMM(2 * GEN8_3LVL_PDPES) | MI_LRI_FORCE_POSTED;
-	for (i = GEN8_3LVL_PDPES; i--; ) {
-		const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
-		u32 base = engine->mmio_base;
-
-		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, i));
-		*cs++ = upper_32_bits(pd_daddr);
-		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, i));
-		*cs++ = lower_32_bits(pd_daddr);
+	if (regs[CTX_RING_START] != i915_ggtt_offset(ring->vma)) {
+		pr_err("%s: context submitted with incorrect RING_START [%08x], expected %08x\n",
+		       engine->name,
+		       regs[CTX_RING_START],
+		       i915_ggtt_offset(ring->vma));
+		regs[CTX_RING_START] = i915_ggtt_offset(ring->vma);
+		valid = false;
 	}
-	*cs++ = MI_NOOP;
 
-	intel_ring_advance(rq, cs);
-
-	return 0;
-}
-
-static int execlists_request_alloc(struct i915_request *request)
-{
-	int ret;
-
-	GEM_BUG_ON(!intel_context_is_pinned(request->context));
-
-	/*
-	 * Flush enough space to reduce the likelihood of waiting after
-	 * we start building the request - in which case we will just
-	 * have to repeat work.
-	 */
-	request->reserved_space += EXECLISTS_REQUEST_SIZE;
-
-	/*
-	 * Note that after this point, we have committed to using
-	 * this request as it is being used to both track the
-	 * state of engine initialisation and liveness of the
-	 * golden renderstate above. Think twice before you try
-	 * to cancel/unwind this request now.
-	 */
-
-	if (!i915_vm_is_4lvl(request->context->vm)) {
-		ret = emit_pdps(request);
-		if (ret)
-			return ret;
+	if ((regs[CTX_RING_CTL] & ~(RING_WAIT | RING_WAIT_SEMAPHORE)) !=
+	    (RING_CTL_SIZE(ring->size) | RING_VALID)) {
+		pr_err("%s: context submitted with incorrect RING_CTL [%08x], expected %08x\n",
+		       engine->name,
+		       regs[CTX_RING_CTL],
+		       (u32)(RING_CTL_SIZE(ring->size) | RING_VALID));
+		regs[CTX_RING_CTL] = RING_CTL_SIZE(ring->size) | RING_VALID;
+		valid = false;
 	}
 
-	/* Unconditionally invalidate GPU caches and TLBs. */
-	ret = request->engine->emit_flush(request, EMIT_INVALIDATE);
-	if (ret)
-		return ret;
+	x = lrc_ring_mi_mode(engine);
+	if (x != -1 && regs[x + 1] & (regs[x + 1] >> 16) & STOP_RING) {
+		pr_err("%s: context submitted with STOP_RING [%08x] in RING_MI_MODE\n",
+		       engine->name, regs[x + 1]);
+		regs[x + 1] &= ~STOP_RING;
+		regs[x + 1] |= STOP_RING << 16;
+		valid = false;
+	}
 
-	request->reserved_space -= EXECLISTS_REQUEST_SIZE;
-	return 0;
+	WARN_ONCE(!valid, "Invalid lrc state found %s submission\n", when);
 }
 
 /*
@@ -3955,7 +1415,7 @@ gen10_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
 	return batch;
 }
 
-#define CTX_WA_BB_OBJ_SIZE (PAGE_SIZE)
+#define CTX_WA_BB_SIZE (PAGE_SIZE)
 
 static int lrc_setup_wa_ctx(struct intel_engine_cs *engine)
 {
@@ -3963,7 +1423,7 @@ static int lrc_setup_wa_ctx(struct intel_engine_cs *engine)
 	struct i915_vma *vma;
 	int err;
 
-	obj = i915_gem_object_create_shmem(engine->i915, CTX_WA_BB_OBJ_SIZE);
+	obj = i915_gem_object_create_shmem(engine->i915, CTX_WA_BB_SIZE);
 	if (IS_ERR(obj))
 		return PTR_ERR(obj);
 
@@ -3985,30 +1445,34 @@ err:
 	return err;
 }
 
-static void lrc_destroy_wa_ctx(struct intel_engine_cs *engine)
+void lrc_fini_wa_ctx(struct intel_engine_cs *engine)
 {
 	i915_vma_unpin_and_release(&engine->wa_ctx.vma, 0);
+
+	/* Called on error unwind, clear all flags to prevent further use */
+	memset(&engine->wa_ctx, 0, sizeof(engine->wa_ctx));
 }
 
 typedef u32 *(*wa_bb_func_t)(struct intel_engine_cs *engine, u32 *batch);
 
-static int intel_init_workaround_bb(struct intel_engine_cs *engine)
+void lrc_init_wa_ctx(struct intel_engine_cs *engine)
 {
 	struct i915_ctx_workarounds *wa_ctx = &engine->wa_ctx;
-	struct i915_wa_ctx_bb *wa_bb[2] = { &wa_ctx->indirect_ctx,
-					    &wa_ctx->per_ctx };
-	wa_bb_func_t wa_bb_fn[2];
+	struct i915_wa_ctx_bb *wa_bb[] = {
+		&wa_ctx->indirect_ctx, &wa_ctx->per_ctx
+	};
+	wa_bb_func_t wa_bb_fn[ARRAY_SIZE(wa_bb)];
 	void *batch, *batch_ptr;
 	unsigned int i;
-	int ret;
+	int err;
 
 	if (engine->class != RENDER_CLASS)
-		return 0;
+		return;
 
 	switch (INTEL_GEN(engine->i915)) {
 	case 12:
 	case 11:
-		return 0;
+		return;
 	case 10:
 		wa_bb_fn[0] = gen10_init_indirectctx_bb;
 		wa_bb_fn[1] = NULL;
@@ -4023,14 +1487,20 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine)
 		break;
 	default:
 		MISSING_CASE(INTEL_GEN(engine->i915));
-		return 0;
+		return;
 	}
 
-	ret = lrc_setup_wa_ctx(engine);
-	if (ret) {
-		drm_dbg(&engine->i915->drm,
-			"Failed to setup context WA page: %d\n", ret);
-		return ret;
+	err = lrc_setup_wa_ctx(engine);
+	if (err) {
+		/*
+		 * We continue even if we fail to initialize WA batch
+		 * because we only expect rare glitches but nothing
+		 * critical to prevent us from using GPU
+		 */
+		drm_err(&engine->i915->drm,
+			"Ignoring context switch w/a allocation error:%d\n",
+			err);
+		return;
 	}
 
 	batch = i915_gem_object_pin_map(wa_ctx->vma->obj, I915_MAP_WB);
@@ -4045,2104 +1515,52 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine)
 		wa_bb[i]->offset = batch_ptr - batch;
 		if (GEM_DEBUG_WARN_ON(!IS_ALIGNED(wa_bb[i]->offset,
 						  CACHELINE_BYTES))) {
-			ret = -EINVAL;
+			err = -EINVAL;
 			break;
 		}
 		if (wa_bb_fn[i])
 			batch_ptr = wa_bb_fn[i](engine, batch_ptr);
 		wa_bb[i]->size = batch_ptr - (batch + wa_bb[i]->offset);
 	}
-	GEM_BUG_ON(batch_ptr - batch > CTX_WA_BB_OBJ_SIZE);
+	GEM_BUG_ON(batch_ptr - batch > CTX_WA_BB_SIZE);
 
 	__i915_gem_object_flush_map(wa_ctx->vma->obj, 0, batch_ptr - batch);
 	__i915_gem_object_release_map(wa_ctx->vma->obj);
-	if (ret)
-		lrc_destroy_wa_ctx(engine);
-
-	return ret;
-}
-
-static void reset_csb_pointers(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	const unsigned int reset_value = execlists->csb_size - 1;
-
-	ring_set_paused(engine, 0);
-
-	/*
-	 * Sometimes Icelake forgets to reset its pointers on a GPU reset.
-	 * Bludgeon them with a mmio update to be sure.
-	 */
-	ENGINE_WRITE(engine, RING_CONTEXT_STATUS_PTR,
-		     0xffff << 16 | reset_value << 8 | reset_value);
-	ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
-
-	/*
-	 * After a reset, the HW starts writing into CSB entry [0]. We
-	 * therefore have to set our HEAD pointer back one entry so that
-	 * the *first* entry we check is entry 0. To complicate this further,
-	 * as we don't wait for the first interrupt after reset, we have to
-	 * fake the HW write to point back to the last entry so that our
-	 * inline comparison of our cached head position against the last HW
-	 * write works even before the first interrupt.
-	 */
-	execlists->csb_head = reset_value;
-	WRITE_ONCE(*execlists->csb_write, reset_value);
-	wmb(); /* Make sure this is visible to HW (paranoia?) */
-
-	/* Check that the GPU does indeed update the CSB entries! */
-	memset(execlists->csb_status, -1, (reset_value + 1) * sizeof(u64));
-	invalidate_csb_entries(&execlists->csb_status[0],
-			       &execlists->csb_status[reset_value]);
-
-	/* Once more for luck and our trusty paranoia */
-	ENGINE_WRITE(engine, RING_CONTEXT_STATUS_PTR,
-		     0xffff << 16 | reset_value << 8 | reset_value);
-	ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
-
-	GEM_BUG_ON(READ_ONCE(*execlists->csb_write) != reset_value);
-}
-
-static void execlists_sanitize(struct intel_engine_cs *engine)
-{
-	GEM_BUG_ON(execlists_active(&engine->execlists));
-
-	/*
-	 * Poison residual state on resume, in case the suspend didn't!
-	 *
-	 * We have to assume that across suspend/resume (or other loss
-	 * of control) that the contents of our pinned buffers has been
-	 * lost, replaced by garbage. Since this doesn't always happen,
-	 * let's poison such state so that we more quickly spot when
-	 * we falsely assume it has been preserved.
-	 */
-	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
-
-	reset_csb_pointers(engine);
-
-	/*
-	 * The kernel_context HWSP is stored in the status_page. As above,
-	 * that may be lost on resume/initialisation, and so we need to
-	 * reset the value in the HWSP.
-	 */
-	intel_timeline_reset_seqno(engine->kernel_context->timeline);
-
-	/* And scrub the dirty cachelines for the HWSP */
-	clflush_cache_range(engine->status_page.addr, PAGE_SIZE);
-}
-
-static void enable_error_interrupt(struct intel_engine_cs *engine)
-{
-	u32 status;
-
-	engine->execlists.error_interrupt = 0;
-	ENGINE_WRITE(engine, RING_EMR, ~0u);
-	ENGINE_WRITE(engine, RING_EIR, ~0u); /* clear all existing errors */
-
-	status = ENGINE_READ(engine, RING_ESR);
-	if (unlikely(status)) {
-		drm_err(&engine->i915->drm,
-			"engine '%s' resumed still in error: %08x\n",
-			engine->name, status);
-		__intel_gt_reset(engine->gt, engine->mask);
-	}
-
-	/*
-	 * On current gen8+, we have 2 signals to play with
-	 *
-	 * - I915_ERROR_INSTUCTION (bit 0)
-	 *
-	 *    Generate an error if the command parser encounters an invalid
-	 *    instruction
-	 *
-	 *    This is a fatal error.
-	 *
-	 * - CP_PRIV (bit 2)
-	 *
-	 *    Generate an error on privilege violation (where the CP replaces
-	 *    the instruction with a no-op). This also fires for writes into
-	 *    read-only scratch pages.
-	 *
-	 *    This is a non-fatal error, parsing continues.
-	 *
-	 * * there are a few others defined for odd HW that we do not use
-	 *
-	 * Since CP_PRIV fires for cases where we have chosen to ignore the
-	 * error (as the HW is validating and suppressing the mistakes), we
-	 * only unmask the instruction error bit.
-	 */
-	ENGINE_WRITE(engine, RING_EMR, ~I915_ERROR_INSTRUCTION);
-}
-
-static void enable_execlists(struct intel_engine_cs *engine)
-{
-	u32 mode;
-
-	assert_forcewakes_active(engine->uncore, FORCEWAKE_ALL);
-
-	intel_engine_set_hwsp_writemask(engine, ~0u); /* HWSTAM */
-
-	if (INTEL_GEN(engine->i915) >= 11)
-		mode = _MASKED_BIT_ENABLE(GEN11_GFX_DISABLE_LEGACY_MODE);
-	else
-		mode = _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE);
-	ENGINE_WRITE_FW(engine, RING_MODE_GEN7, mode);
-
-	ENGINE_WRITE_FW(engine, RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
-
-	ENGINE_WRITE_FW(engine,
-			RING_HWS_PGA,
-			i915_ggtt_offset(engine->status_page.vma));
-	ENGINE_POSTING_READ(engine, RING_HWS_PGA);
-
-	enable_error_interrupt(engine);
-
-	engine->context_tag = GENMASK(BITS_PER_LONG - 2, 0);
-}
-
-static bool unexpected_starting_state(struct intel_engine_cs *engine)
-{
-	bool unexpected = false;
-
-	if (ENGINE_READ_FW(engine, RING_MI_MODE) & STOP_RING) {
-		drm_dbg(&engine->i915->drm,
-			"STOP_RING still set in RING_MI_MODE\n");
-		unexpected = true;
-	}
-
-	return unexpected;
-}
-
-static int execlists_resume(struct intel_engine_cs *engine)
-{
-	intel_mocs_init_engine(engine);
-
-	intel_breadcrumbs_reset(engine->breadcrumbs);
-
-	if (GEM_SHOW_DEBUG() && unexpected_starting_state(engine)) {
-		struct drm_printer p = drm_debug_printer(__func__);
-
-		intel_engine_dump(engine, &p, NULL);
-	}
-
-	enable_execlists(engine);
-
-	return 0;
-}
 
-static void execlists_reset_prepare(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	unsigned long flags;
-
-	ENGINE_TRACE(engine, "depth<-%d\n",
-		     atomic_read(&execlists->tasklet.count));
-
-	/*
-	 * Prevent request submission to the hardware until we have
-	 * completed the reset in i915_gem_reset_finish(). If a request
-	 * is completed by one engine, it may then queue a request
-	 * to a second via its execlists->tasklet *just* as we are
-	 * calling engine->resume() and also writing the ELSP.
-	 * Turning off the execlists->tasklet until the reset is over
-	 * prevents the race.
-	 */
-	__tasklet_disable_sync_once(&execlists->tasklet);
-	GEM_BUG_ON(!reset_in_progress(execlists));
-
-	/* And flush any current direct submission. */
-	spin_lock_irqsave(&engine->active.lock, flags);
-	spin_unlock_irqrestore(&engine->active.lock, flags);
-
-	/*
-	 * We stop engines, otherwise we might get failed reset and a
-	 * dead gpu (on elk). Also as modern gpu as kbl can suffer
-	 * from system hang if batchbuffer is progressing when
-	 * the reset is issued, regardless of READY_TO_RESET ack.
-	 * Thus assume it is best to stop engines on all gens
-	 * where we have a gpu reset.
-	 *
-	 * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES)
-	 *
-	 * FIXME: Wa for more modern gens needs to be validated
-	 */
-	ring_set_paused(engine, 1);
-	intel_engine_stop_cs(engine);
-
-	engine->execlists.reset_ccid = active_ccid(engine);
-}
-
-static void __reset_stop_ring(u32 *regs, const struct intel_engine_cs *engine)
-{
-	int x;
-
-	x = lrc_ring_mi_mode(engine);
-	if (x != -1) {
-		regs[x + 1] &= ~STOP_RING;
-		regs[x + 1] |= STOP_RING << 16;
-	}
-}
-
-static void __execlists_reset_reg_state(const struct intel_context *ce,
-					const struct intel_engine_cs *engine)
-{
-	u32 *regs = ce->lrc_reg_state;
-
-	__reset_stop_ring(regs, engine);
+	/* Verify that we can handle failure to setup the wa_ctx */
+	if (err || i915_inject_probe_error(engine->i915, -ENODEV))
+		lrc_fini_wa_ctx(engine);
 }
 
-static void __execlists_reset(struct intel_engine_cs *engine, bool stalled)
-{
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	struct intel_context *ce;
-	struct i915_request *rq;
-	u32 head;
-
-	mb(); /* paranoia: read the CSB pointers from after the reset */
-	clflush(execlists->csb_write);
-	mb();
-
-	process_csb(engine); /* drain preemption events */
-
-	/* Following the reset, we need to reload the CSB read/write pointers */
-	reset_csb_pointers(engine);
-
-	/*
-	 * Save the currently executing context, even if we completed
-	 * its request, it was still running at the time of the
-	 * reset and will have been clobbered.
-	 */
-	rq = active_context(engine, engine->execlists.reset_ccid);
-	if (!rq)
-		goto unwind;
-
-	ce = rq->context;
-	GEM_BUG_ON(!i915_vma_is_pinned(ce->state));
-
-	if (i915_request_completed(rq)) {
-		/* Idle context; tidy up the ring so we can restart afresh */
-		head = intel_ring_wrap(ce->ring, rq->tail);
-		goto out_replay;
-	}
-
-	/* We still have requests in-flight; the engine should be active */
-	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
-
-	/* Context has requests still in-flight; it should not be idle! */
-	GEM_BUG_ON(i915_active_is_idle(&ce->active));
-
-	rq = active_request(ce->timeline, rq);
-	head = intel_ring_wrap(ce->ring, rq->head);
-	GEM_BUG_ON(head == ce->ring->tail);
-
-	/*
-	 * If this request hasn't started yet, e.g. it is waiting on a
-	 * semaphore, we need to avoid skipping the request or else we
-	 * break the signaling chain. However, if the context is corrupt
-	 * the request will not restart and we will be stuck with a wedged
-	 * device. It is quite often the case that if we issue a reset
-	 * while the GPU is loading the context image, that the context
-	 * image becomes corrupt.
-	 *
-	 * Otherwise, if we have not started yet, the request should replay
-	 * perfectly and we do not need to flag the result as being erroneous.
-	 */
-	if (!i915_request_started(rq))
-		goto out_replay;
-
-	/*
-	 * If the request was innocent, we leave the request in the ELSP
-	 * and will try to replay it on restarting. The context image may
-	 * have been corrupted by the reset, in which case we may have
-	 * to service a new GPU hang, but more likely we can continue on
-	 * without impact.
-	 *
-	 * If the request was guilty, we presume the context is corrupt
-	 * and have to at least restore the RING register in the context
-	 * image back to the expected values to skip over the guilty request.
-	 */
-	__i915_request_reset(rq, stalled);
-
-	/*
-	 * We want a simple context + ring to execute the breadcrumb update.
-	 * We cannot rely on the context being intact across the GPU hang,
-	 * so clear it and rebuild just what we need for the breadcrumb.
-	 * All pending requests for this context will be zapped, and any
-	 * future request will be after userspace has had the opportunity
-	 * to recreate its own state.
-	 */
-out_replay:
-	ENGINE_TRACE(engine, "replay {head:%04x, tail:%04x}\n",
-		     head, ce->ring->tail);
-	__execlists_reset_reg_state(ce, engine);
-	__execlists_update_reg_state(ce, engine, head);
-	ce->lrc.desc |= CTX_DESC_FORCE_RESTORE; /* paranoid: GPU was reset! */
-
-unwind:
-	/* Push back any incomplete requests for replay after the reset. */
-	cancel_port_requests(execlists);
-	__unwind_incomplete_requests(engine);
-}
-
-static void execlists_reset_rewind(struct intel_engine_cs *engine, bool stalled)
-{
-	unsigned long flags;
-
-	ENGINE_TRACE(engine, "\n");
-
-	spin_lock_irqsave(&engine->active.lock, flags);
-
-	__execlists_reset(engine, stalled);
-
-	spin_unlock_irqrestore(&engine->active.lock, flags);
-}
-
-static void nop_submission_tasklet(unsigned long data)
-{
-	struct intel_engine_cs * const engine = (struct intel_engine_cs *)data;
-
-	/* The driver is wedged; don't process any more events. */
-	WRITE_ONCE(engine->execlists.queue_priority_hint, INT_MIN);
-}
-
-static void execlists_reset_cancel(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	struct i915_request *rq, *rn;
-	struct rb_node *rb;
-	unsigned long flags;
-
-	ENGINE_TRACE(engine, "\n");
-
-	/*
-	 * Before we call engine->cancel_requests(), we should have exclusive
-	 * access to the submission state. This is arranged for us by the
-	 * caller disabling the interrupt generation, the tasklet and other
-	 * threads that may then access the same state, giving us a free hand
-	 * to reset state. However, we still need to let lockdep be aware that
-	 * we know this state may be accessed in hardirq context, so we
-	 * disable the irq around this manipulation and we want to keep
-	 * the spinlock focused on its duties and not accidentally conflate
-	 * coverage to the submission's irq state. (Similarly, although we
-	 * shouldn't need to disable irq around the manipulation of the
-	 * submission's irq state, we also wish to remind ourselves that
-	 * it is irq state.)
-	 */
-	spin_lock_irqsave(&engine->active.lock, flags);
-
-	__execlists_reset(engine, true);
-
-	/* Mark all executing requests as skipped. */
-	list_for_each_entry(rq, &engine->active.requests, sched.link)
-		mark_eio(rq);
-	intel_engine_signal_breadcrumbs(engine);
-
-	/* Flush the queued requests to the timeline list (for retiring). */
-	while ((rb = rb_first_cached(&execlists->queue))) {
-		struct i915_priolist *p = to_priolist(rb);
-		int i;
-
-		priolist_for_each_request_consume(rq, rn, p, i) {
-			mark_eio(rq);
-			__i915_request_submit(rq);
-		}
-
-		rb_erase_cached(&p->node, &execlists->queue);
-		i915_priolist_free(p);
-	}
-
-	/* On-hold requests will be flushed to timeline upon their release */
-	list_for_each_entry(rq, &engine->active.hold, sched.link)
-		mark_eio(rq);
-
-	/* Cancel all attached virtual engines */
-	while ((rb = rb_first_cached(&execlists->virtual))) {
-		struct virtual_engine *ve =
-			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
-
-		rb_erase_cached(rb, &execlists->virtual);
-		RB_CLEAR_NODE(rb);
-
-		spin_lock(&ve->base.active.lock);
-		rq = fetch_and_zero(&ve->request);
-		if (rq) {
-			mark_eio(rq);
-
-			rq->engine = engine;
-			__i915_request_submit(rq);
-			i915_request_put(rq);
-
-			ve->base.execlists.queue_priority_hint = INT_MIN;
-		}
-		spin_unlock(&ve->base.active.lock);
-	}
-
-	/* Remaining _unready_ requests will be nop'ed when submitted */
-
-	execlists->queue_priority_hint = INT_MIN;
-	execlists->queue = RB_ROOT_CACHED;
-
-	GEM_BUG_ON(__tasklet_is_enabled(&execlists->tasklet));
-	execlists->tasklet.func = nop_submission_tasklet;
-
-	spin_unlock_irqrestore(&engine->active.lock, flags);
-}
-
-static void execlists_reset_finish(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-
-	/*
-	 * After a GPU reset, we may have requests to replay. Do so now while
-	 * we still have the forcewake to be sure that the GPU is not allowed
-	 * to sleep before we restart and reload a context.
-	 */
-	GEM_BUG_ON(!reset_in_progress(execlists));
-	if (!RB_EMPTY_ROOT(&execlists->queue.rb_root))
-		execlists->tasklet.func(execlists->tasklet.data);
-
-	if (__tasklet_enable(&execlists->tasklet))
-		/* And kick in case we missed a new request submission. */
-		tasklet_hi_schedule(&execlists->tasklet);
-	ENGINE_TRACE(engine, "depth->%d\n",
-		     atomic_read(&execlists->tasklet.count));
-}
-
-static int gen8_emit_bb_start_noarb(struct i915_request *rq,
-				    u64 offset, u32 len,
-				    const unsigned int flags)
-{
-	u32 *cs;
-
-	cs = intel_ring_begin(rq, 4);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	/*
-	 * WaDisableCtxRestoreArbitration:bdw,chv
-	 *
-	 * We don't need to perform MI_ARB_ENABLE as often as we do (in
-	 * particular all the gen that do not need the w/a at all!), if we
-	 * took care to make sure that on every switch into this context
-	 * (both ordinary and for preemption) that arbitrartion was enabled
-	 * we would be fine.  However, for gen8 there is another w/a that
-	 * requires us to not preempt inside GPGPU execution, so we keep
-	 * arbitration disabled for gen8 batches. Arbitration will be
-	 * re-enabled before we close the request
-	 * (engine->emit_fini_breadcrumb).
-	 */
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
-
-	/* FIXME(BDW+): Address space and security selectors. */
-	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
-		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
-	*cs++ = lower_32_bits(offset);
-	*cs++ = upper_32_bits(offset);
-
-	intel_ring_advance(rq, cs);
-
-	return 0;
-}
-
-static int gen8_emit_bb_start(struct i915_request *rq,
-			      u64 offset, u32 len,
-			      const unsigned int flags)
-{
-	u32 *cs;
-
-	cs = intel_ring_begin(rq, 6);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
-
-	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
-		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
-	*cs++ = lower_32_bits(offset);
-	*cs++ = upper_32_bits(offset);
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
-	*cs++ = MI_NOOP;
-
-	intel_ring_advance(rq, cs);
-
-	return 0;
-}
-
-static void gen8_logical_ring_enable_irq(struct intel_engine_cs *engine)
-{
-	ENGINE_WRITE(engine, RING_IMR,
-		     ~(engine->irq_enable_mask | engine->irq_keep_mask));
-	ENGINE_POSTING_READ(engine, RING_IMR);
-}
-
-static void gen8_logical_ring_disable_irq(struct intel_engine_cs *engine)
-{
-	ENGINE_WRITE(engine, RING_IMR, ~engine->irq_keep_mask);
-}
-
-static int gen8_emit_flush(struct i915_request *request, u32 mode)
-{
-	u32 cmd, *cs;
-
-	cs = intel_ring_begin(request, 4);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	cmd = MI_FLUSH_DW + 1;
-
-	/* We always require a command barrier so that subsequent
-	 * commands, such as breadcrumb interrupts, are strictly ordered
-	 * wrt the contents of the write cache being flushed to memory
-	 * (and thus being coherent from the CPU).
-	 */
-	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
-
-	if (mode & EMIT_INVALIDATE) {
-		cmd |= MI_INVALIDATE_TLB;
-		if (request->engine->class == VIDEO_DECODE_CLASS)
-			cmd |= MI_INVALIDATE_BSD;
-	}
-
-	*cs++ = cmd;
-	*cs++ = LRC_PPHWSP_SCRATCH_ADDR;
-	*cs++ = 0; /* upper addr */
-	*cs++ = 0; /* value */
-	intel_ring_advance(request, cs);
-
-	return 0;
-}
-
-static int gen8_emit_flush_render(struct i915_request *request,
-				  u32 mode)
-{
-	bool vf_flush_wa = false, dc_flush_wa = false;
-	u32 *cs, flags = 0;
-	int len;
-
-	flags |= PIPE_CONTROL_CS_STALL;
-
-	if (mode & EMIT_FLUSH) {
-		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
-		flags |= PIPE_CONTROL_FLUSH_ENABLE;
-	}
-
-	if (mode & EMIT_INVALIDATE) {
-		flags |= PIPE_CONTROL_TLB_INVALIDATE;
-		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_QW_WRITE;
-		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
-
-		/*
-		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
-		 * pipe control.
-		 */
-		if (IS_GEN(request->engine->i915, 9))
-			vf_flush_wa = true;
-
-		/* WaForGAMHang:kbl */
-		if (IS_KBL_GT_REVID(request->engine->i915, 0, KBL_REVID_B0))
-			dc_flush_wa = true;
-	}
-
-	len = 6;
-
-	if (vf_flush_wa)
-		len += 6;
-
-	if (dc_flush_wa)
-		len += 12;
-
-	cs = intel_ring_begin(request, len);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	if (vf_flush_wa)
-		cs = gen8_emit_pipe_control(cs, 0, 0);
-
-	if (dc_flush_wa)
-		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_DC_FLUSH_ENABLE,
-					    0);
-
-	cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
-
-	if (dc_flush_wa)
-		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_CS_STALL, 0);
-
-	intel_ring_advance(request, cs);
-
-	return 0;
-}
-
-static int gen11_emit_flush_render(struct i915_request *request,
-				   u32 mode)
-{
-	if (mode & EMIT_FLUSH) {
-		u32 *cs;
-		u32 flags = 0;
-
-		flags |= PIPE_CONTROL_CS_STALL;
-
-		flags |= PIPE_CONTROL_TILE_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
-		flags |= PIPE_CONTROL_FLUSH_ENABLE;
-		flags |= PIPE_CONTROL_QW_WRITE;
-		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
-
-		cs = intel_ring_begin(request, 6);
-		if (IS_ERR(cs))
-			return PTR_ERR(cs);
-
-		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
-		intel_ring_advance(request, cs);
-	}
-
-	if (mode & EMIT_INVALIDATE) {
-		u32 *cs;
-		u32 flags = 0;
-
-		flags |= PIPE_CONTROL_CS_STALL;
-
-		flags |= PIPE_CONTROL_COMMAND_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_TLB_INVALIDATE;
-		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_QW_WRITE;
-		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
-
-		cs = intel_ring_begin(request, 6);
-		if (IS_ERR(cs))
-			return PTR_ERR(cs);
-
-		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
-		intel_ring_advance(request, cs);
-	}
-
-	return 0;
-}
-
-static u32 preparser_disable(bool state)
-{
-	return MI_ARB_CHECK | 1 << 8 | state;
-}
-
-static i915_reg_t aux_inv_reg(const struct intel_engine_cs *engine)
-{
-	static const i915_reg_t vd[] = {
-		GEN12_VD0_AUX_NV,
-		GEN12_VD1_AUX_NV,
-		GEN12_VD2_AUX_NV,
-		GEN12_VD3_AUX_NV,
-	};
-
-	static const i915_reg_t ve[] = {
-		GEN12_VE0_AUX_NV,
-		GEN12_VE1_AUX_NV,
-	};
-
-	if (engine->class == VIDEO_DECODE_CLASS)
-		return vd[engine->instance];
-
-	if (engine->class == VIDEO_ENHANCEMENT_CLASS)
-		return ve[engine->instance];
-
-	GEM_BUG_ON("unknown aux_inv_reg\n");
-
-	return INVALID_MMIO_REG;
-}
-
-static u32 *
-gen12_emit_aux_table_inv(const i915_reg_t inv_reg, u32 *cs)
-{
-	*cs++ = MI_LOAD_REGISTER_IMM(1);
-	*cs++ = i915_mmio_reg_offset(inv_reg);
-	*cs++ = AUX_INV;
-	*cs++ = MI_NOOP;
-
-	return cs;
-}
-
-static int gen12_emit_flush_render(struct i915_request *request,
-				   u32 mode)
-{
-	if (mode & EMIT_FLUSH) {
-		u32 flags = 0;
-		u32 *cs;
-
-		flags |= PIPE_CONTROL_TILE_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_FLUSH_L3;
-		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
-		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
-		/* Wa_1409600907:tgl */
-		flags |= PIPE_CONTROL_DEPTH_STALL;
-		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
-		flags |= PIPE_CONTROL_FLUSH_ENABLE;
-
-		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
-		flags |= PIPE_CONTROL_QW_WRITE;
-
-		flags |= PIPE_CONTROL_CS_STALL;
-
-		cs = intel_ring_begin(request, 6);
-		if (IS_ERR(cs))
-			return PTR_ERR(cs);
-
-		cs = gen12_emit_pipe_control(cs,
-					     PIPE_CONTROL0_HDC_PIPELINE_FLUSH,
-					     flags, LRC_PPHWSP_SCRATCH_ADDR);
-		intel_ring_advance(request, cs);
-	}
-
-	if (mode & EMIT_INVALIDATE) {
-		u32 flags = 0;
-		u32 *cs;
-
-		flags |= PIPE_CONTROL_COMMAND_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_TLB_INVALIDATE;
-		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
-		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
-
-		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
-		flags |= PIPE_CONTROL_QW_WRITE;
-
-		flags |= PIPE_CONTROL_CS_STALL;
-
-		cs = intel_ring_begin(request, 8 + 4);
-		if (IS_ERR(cs))
-			return PTR_ERR(cs);
-
-		/*
-		 * Prevent the pre-parser from skipping past the TLB
-		 * invalidate and loading a stale page for the batch
-		 * buffer / request payload.
-		 */
-		*cs++ = preparser_disable(true);
-
-		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
-
-		/* hsdes: 1809175790 */
-		cs = gen12_emit_aux_table_inv(GEN12_GFX_CCS_AUX_NV, cs);
-
-		*cs++ = preparser_disable(false);
-		intel_ring_advance(request, cs);
-	}
-
-	return 0;
-}
-
-static int gen12_emit_flush(struct i915_request *request, u32 mode)
-{
-	intel_engine_mask_t aux_inv = 0;
-	u32 cmd, *cs;
-
-	cmd = 4;
-	if (mode & EMIT_INVALIDATE)
-		cmd += 2;
-	if (mode & EMIT_INVALIDATE)
-		aux_inv = request->engine->mask & ~BIT(BCS0);
-	if (aux_inv)
-		cmd += 2 * hweight8(aux_inv) + 2;
-
-	cs = intel_ring_begin(request, cmd);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	if (mode & EMIT_INVALIDATE)
-		*cs++ = preparser_disable(true);
-
-	cmd = MI_FLUSH_DW + 1;
-
-	/* We always require a command barrier so that subsequent
-	 * commands, such as breadcrumb interrupts, are strictly ordered
-	 * wrt the contents of the write cache being flushed to memory
-	 * (and thus being coherent from the CPU).
-	 */
-	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
-
-	if (mode & EMIT_INVALIDATE) {
-		cmd |= MI_INVALIDATE_TLB;
-		if (request->engine->class == VIDEO_DECODE_CLASS)
-			cmd |= MI_INVALIDATE_BSD;
-	}
-
-	*cs++ = cmd;
-	*cs++ = LRC_PPHWSP_SCRATCH_ADDR;
-	*cs++ = 0; /* upper addr */
-	*cs++ = 0; /* value */
-
-	if (aux_inv) { /* hsdes: 1809175790 */
-		struct intel_engine_cs *engine;
-		unsigned int tmp;
-
-		*cs++ = MI_LOAD_REGISTER_IMM(hweight8(aux_inv));
-		for_each_engine_masked(engine, request->engine->gt,
-				       aux_inv, tmp) {
-			*cs++ = i915_mmio_reg_offset(aux_inv_reg(engine));
-			*cs++ = AUX_INV;
-		}
-		*cs++ = MI_NOOP;
-	}
-
-	if (mode & EMIT_INVALIDATE)
-		*cs++ = preparser_disable(false);
-
-	intel_ring_advance(request, cs);
-
-	return 0;
-}
-
-static void assert_request_valid(struct i915_request *rq)
-{
-	struct intel_ring *ring __maybe_unused = rq->ring;
-
-	/* Can we unwind this request without appearing to go forwards? */
-	GEM_BUG_ON(intel_ring_direction(ring, rq->wa_tail, rq->head) <= 0);
-}
-
-/*
- * Reserve space for 2 NOOPs at the end of each request to be
- * used as a workaround for not being allowed to do lite
- * restore with HEAD==TAIL (WaIdleLiteRestore).
- */
-static u32 *gen8_emit_wa_tail(struct i915_request *request, u32 *cs)
-{
-	/* Ensure there's always at least one preemption point per-request. */
-	*cs++ = MI_ARB_CHECK;
-	*cs++ = MI_NOOP;
-	request->wa_tail = intel_ring_offset(request, cs);
-
-	/* Check that entire request is less than half the ring */
-	assert_request_valid(request);
-
-	return cs;
-}
-
-static u32 *emit_preempt_busywait(struct i915_request *request, u32 *cs)
-{
-	*cs++ = MI_SEMAPHORE_WAIT |
-		MI_SEMAPHORE_GLOBAL_GTT |
-		MI_SEMAPHORE_POLL |
-		MI_SEMAPHORE_SAD_EQ_SDD;
-	*cs++ = 0;
-	*cs++ = intel_hws_preempt_address(request->engine);
-	*cs++ = 0;
-
-	return cs;
-}
-
-static __always_inline u32*
-gen8_emit_fini_breadcrumb_tail(struct i915_request *request, u32 *cs)
-{
-	*cs++ = MI_USER_INTERRUPT;
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
-	if (intel_engine_has_semaphores(request->engine))
-		cs = emit_preempt_busywait(request, cs);
-
-	request->tail = intel_ring_offset(request, cs);
-	assert_ring_tail_valid(request->ring, request->tail);
-
-	return gen8_emit_wa_tail(request, cs);
-}
-
-static u32 *emit_xcs_breadcrumb(struct i915_request *rq, u32 *cs)
-{
-	return gen8_emit_ggtt_write(cs, rq->fence.seqno, hwsp_offset(rq), 0);
-}
-
-static u32 *gen8_emit_fini_breadcrumb(struct i915_request *rq, u32 *cs)
-{
-	return gen8_emit_fini_breadcrumb_tail(rq, emit_xcs_breadcrumb(rq, cs));
-}
-
-static u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs)
-{
-	cs = gen8_emit_pipe_control(cs,
-				    PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
-				    PIPE_CONTROL_DEPTH_CACHE_FLUSH |
-				    PIPE_CONTROL_DC_FLUSH_ENABLE,
-				    0);
-
-	/* XXX flush+write+CS_STALL all in one upsets gem_concurrent_blt:kbl */
-	cs = gen8_emit_ggtt_write_rcs(cs,
-				      request->fence.seqno,
-				      hwsp_offset(request),
-				      PIPE_CONTROL_FLUSH_ENABLE |
-				      PIPE_CONTROL_CS_STALL);
-
-	return gen8_emit_fini_breadcrumb_tail(request, cs);
-}
-
-static u32 *
-gen11_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs)
-{
-	cs = gen8_emit_ggtt_write_rcs(cs,
-				      request->fence.seqno,
-				      hwsp_offset(request),
-				      PIPE_CONTROL_CS_STALL |
-				      PIPE_CONTROL_TILE_CACHE_FLUSH |
-				      PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
-				      PIPE_CONTROL_DEPTH_CACHE_FLUSH |
-				      PIPE_CONTROL_DC_FLUSH_ENABLE |
-				      PIPE_CONTROL_FLUSH_ENABLE);
-
-	return gen8_emit_fini_breadcrumb_tail(request, cs);
-}
-
-/*
- * Note that the CS instruction pre-parser will not stall on the breadcrumb
- * flush and will continue pre-fetching the instructions after it before the
- * memory sync is completed. On pre-gen12 HW, the pre-parser will stop at
- * BB_START/END instructions, so, even though we might pre-fetch the pre-amble
- * of the next request before the memory has been flushed, we're guaranteed that
- * we won't access the batch itself too early.
- * However, on gen12+ the parser can pre-fetch across the BB_START/END commands,
- * so, if the current request is modifying an instruction in the next request on
- * the same intel_context, we might pre-fetch and then execute the pre-update
- * instruction. To avoid this, the users of self-modifying code should either
- * disable the parser around the code emitting the memory writes, via a new flag
- * added to MI_ARB_CHECK, or emit the writes from a different intel_context. For
- * the in-kernel use-cases we've opted to use a separate context, see
- * reloc_gpu() as an example.
- * All the above applies only to the instructions themselves. Non-inline data
- * used by the instructions is not pre-fetched.
- */
-
-static u32 *gen12_emit_preempt_busywait(struct i915_request *request, u32 *cs)
-{
-	*cs++ = MI_SEMAPHORE_WAIT_TOKEN |
-		MI_SEMAPHORE_GLOBAL_GTT |
-		MI_SEMAPHORE_POLL |
-		MI_SEMAPHORE_SAD_EQ_SDD;
-	*cs++ = 0;
-	*cs++ = intel_hws_preempt_address(request->engine);
-	*cs++ = 0;
-	*cs++ = 0;
-	*cs++ = MI_NOOP;
-
-	return cs;
-}
-
-static __always_inline u32*
-gen12_emit_fini_breadcrumb_tail(struct i915_request *request, u32 *cs)
-{
-	*cs++ = MI_USER_INTERRUPT;
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
-	if (intel_engine_has_semaphores(request->engine))
-		cs = gen12_emit_preempt_busywait(request, cs);
-
-	request->tail = intel_ring_offset(request, cs);
-	assert_ring_tail_valid(request->ring, request->tail);
-
-	return gen8_emit_wa_tail(request, cs);
-}
-
-static u32 *gen12_emit_fini_breadcrumb(struct i915_request *rq, u32 *cs)
-{
-	/* XXX Stalling flush before seqno write; post-sync not */
-	cs = emit_xcs_breadcrumb(rq, __gen8_emit_flush_dw(cs, 0, 0, 0));
-	return gen12_emit_fini_breadcrumb_tail(rq, cs);
-}
-
-static u32 *
-gen12_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs)
-{
-	cs = gen12_emit_ggtt_write_rcs(cs,
-				       request->fence.seqno,
-				       hwsp_offset(request),
-				       PIPE_CONTROL0_HDC_PIPELINE_FLUSH,
-				       PIPE_CONTROL_CS_STALL |
-				       PIPE_CONTROL_TILE_CACHE_FLUSH |
-				       PIPE_CONTROL_FLUSH_L3 |
-				       PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
-				       PIPE_CONTROL_DEPTH_CACHE_FLUSH |
-				       /* Wa_1409600907:tgl */
-				       PIPE_CONTROL_DEPTH_STALL |
-				       PIPE_CONTROL_DC_FLUSH_ENABLE |
-				       PIPE_CONTROL_FLUSH_ENABLE);
-
-	return gen12_emit_fini_breadcrumb_tail(request, cs);
-}
-
-static void execlists_park(struct intel_engine_cs *engine)
-{
-	cancel_timer(&engine->execlists.timer);
-	cancel_timer(&engine->execlists.preempt);
-}
-
-void intel_execlists_set_default_submission(struct intel_engine_cs *engine)
-{
-	engine->submit_request = execlists_submit_request;
-	engine->schedule = i915_schedule;
-	engine->execlists.tasklet.func = execlists_submission_tasklet;
-
-	engine->reset.prepare = execlists_reset_prepare;
-	engine->reset.rewind = execlists_reset_rewind;
-	engine->reset.cancel = execlists_reset_cancel;
-	engine->reset.finish = execlists_reset_finish;
-
-	engine->park = execlists_park;
-	engine->unpark = NULL;
-
-	engine->flags |= I915_ENGINE_SUPPORTS_STATS;
-	if (!intel_vgpu_active(engine->i915)) {
-		engine->flags |= I915_ENGINE_HAS_SEMAPHORES;
-		if (HAS_LOGICAL_RING_PREEMPTION(engine->i915)) {
-			engine->flags |= I915_ENGINE_HAS_PREEMPTION;
-			if (IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
-				engine->flags |= I915_ENGINE_HAS_TIMESLICES;
-		}
-	}
-
-	if (INTEL_GEN(engine->i915) >= 12)
-		engine->flags |= I915_ENGINE_HAS_RELATIVE_MMIO;
-
-	if (intel_engine_has_preemption(engine))
-		engine->emit_bb_start = gen8_emit_bb_start;
-	else
-		engine->emit_bb_start = gen8_emit_bb_start_noarb;
-}
-
-static void execlists_shutdown(struct intel_engine_cs *engine)
-{
-	/* Synchronise with residual timers and any softirq they raise */
-	del_timer_sync(&engine->execlists.timer);
-	del_timer_sync(&engine->execlists.preempt);
-	tasklet_kill(&engine->execlists.tasklet);
-}
-
-static void execlists_release(struct intel_engine_cs *engine)
-{
-	engine->sanitize = NULL; /* no longer in control, nothing to sanitize */
-
-	execlists_shutdown(engine);
-
-	intel_engine_cleanup_common(engine);
-	lrc_destroy_wa_ctx(engine);
-}
-
-static void
-logical_ring_default_vfuncs(struct intel_engine_cs *engine)
-{
-	/* Default vfuncs which can be overriden by each engine. */
-
-	engine->resume = execlists_resume;
-
-	engine->cops = &execlists_context_ops;
-	engine->request_alloc = execlists_request_alloc;
-
-	engine->emit_flush = gen8_emit_flush;
-	engine->emit_init_breadcrumb = gen8_emit_init_breadcrumb;
-	engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb;
-	if (INTEL_GEN(engine->i915) >= 12) {
-		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb;
-		engine->emit_flush = gen12_emit_flush;
-	}
-	engine->set_default_submission = intel_execlists_set_default_submission;
-
-	if (INTEL_GEN(engine->i915) < 11) {
-		engine->irq_enable = gen8_logical_ring_enable_irq;
-		engine->irq_disable = gen8_logical_ring_disable_irq;
-	} else {
-		/*
-		 * TODO: On Gen11 interrupt masks need to be clear
-		 * to allow C6 entry. Keep interrupts enabled at
-		 * and take the hit of generating extra interrupts
-		 * until a more refined solution exists.
-		 */
-	}
-}
-
-static inline void
-logical_ring_default_irqs(struct intel_engine_cs *engine)
-{
-	unsigned int shift = 0;
-
-	if (INTEL_GEN(engine->i915) < 11) {
-		const u8 irq_shifts[] = {
-			[RCS0]  = GEN8_RCS_IRQ_SHIFT,
-			[BCS0]  = GEN8_BCS_IRQ_SHIFT,
-			[VCS0]  = GEN8_VCS0_IRQ_SHIFT,
-			[VCS1]  = GEN8_VCS1_IRQ_SHIFT,
-			[VECS0] = GEN8_VECS_IRQ_SHIFT,
-		};
-
-		shift = irq_shifts[engine->id];
-	}
-
-	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift;
-	engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift;
-	engine->irq_keep_mask |= GT_CS_MASTER_ERROR_INTERRUPT << shift;
-	engine->irq_keep_mask |= GT_WAIT_SEMAPHORE_INTERRUPT << shift;
-}
-
-static void rcs_submission_override(struct intel_engine_cs *engine)
-{
-	switch (INTEL_GEN(engine->i915)) {
-	case 12:
-		engine->emit_flush = gen12_emit_flush_render;
-		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_rcs;
-		break;
-	case 11:
-		engine->emit_flush = gen11_emit_flush_render;
-		engine->emit_fini_breadcrumb = gen11_emit_fini_breadcrumb_rcs;
-		break;
-	default:
-		engine->emit_flush = gen8_emit_flush_render;
-		engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_rcs;
-		break;
-	}
-}
-
-int intel_execlists_submission_setup(struct intel_engine_cs *engine)
-{
-	struct intel_engine_execlists * const execlists = &engine->execlists;
-	struct drm_i915_private *i915 = engine->i915;
-	struct intel_uncore *uncore = engine->uncore;
-	u32 base = engine->mmio_base;
-
-	tasklet_init(&engine->execlists.tasklet,
-		     execlists_submission_tasklet, (unsigned long)engine);
-	timer_setup(&engine->execlists.timer, execlists_timeslice, 0);
-	timer_setup(&engine->execlists.preempt, execlists_preempt, 0);
-
-	logical_ring_default_vfuncs(engine);
-	logical_ring_default_irqs(engine);
-
-	if (engine->class == RENDER_CLASS)
-		rcs_submission_override(engine);
-
-	if (intel_init_workaround_bb(engine))
-		/*
-		 * We continue even if we fail to initialize WA batch
-		 * because we only expect rare glitches but nothing
-		 * critical to prevent us from using GPU
-		 */
-		drm_err(&i915->drm, "WA batch buffer initialization failed\n");
-
-	if (HAS_LOGICAL_RING_ELSQ(i915)) {
-		execlists->submit_reg = uncore->regs +
-			i915_mmio_reg_offset(RING_EXECLIST_SQ_CONTENTS(base));
-		execlists->ctrl_reg = uncore->regs +
-			i915_mmio_reg_offset(RING_EXECLIST_CONTROL(base));
-	} else {
-		execlists->submit_reg = uncore->regs +
-			i915_mmio_reg_offset(RING_ELSP(base));
-	}
-
-	execlists->csb_status =
-		(u64 *)&engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX];
-
-	execlists->csb_write =
-		&engine->status_page.addr[intel_hws_csb_write_index(i915)];
-
-	if (INTEL_GEN(i915) < 11)
-		execlists->csb_size = GEN8_CSB_ENTRIES;
-	else
-		execlists->csb_size = GEN11_CSB_ENTRIES;
-
-	if (INTEL_GEN(engine->i915) >= 11) {
-		execlists->ccid |= engine->instance << (GEN11_ENGINE_INSTANCE_SHIFT - 32);
-		execlists->ccid |= engine->class << (GEN11_ENGINE_CLASS_SHIFT - 32);
-	}
-
-	/* Finally, take ownership and responsibility for cleanup! */
-	engine->sanitize = execlists_sanitize;
-	engine->release = execlists_release;
-
-	return 0;
-}
-
-static void init_common_reg_state(u32 * const regs,
-				  const struct intel_engine_cs *engine,
-				  const struct intel_ring *ring,
-				  bool inhibit)
-{
-	u32 ctl;
-
-	ctl = _MASKED_BIT_ENABLE(CTX_CTRL_INHIBIT_SYN_CTX_SWITCH);
-	ctl |= _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT);
-	if (inhibit)
-		ctl |= CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT;
-	if (INTEL_GEN(engine->i915) < 11)
-		ctl |= _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT |
-					   CTX_CTRL_RS_CTX_ENABLE);
-	regs[CTX_CONTEXT_CONTROL] = ctl;
-
-	regs[CTX_RING_CTL] = RING_CTL_SIZE(ring->size) | RING_VALID;
-	regs[CTX_TIMESTAMP] = 0;
-}
-
-static void init_wa_bb_reg_state(u32 * const regs,
-				 const struct intel_engine_cs *engine)
-{
-	const struct i915_ctx_workarounds * const wa_ctx = &engine->wa_ctx;
-
-	if (wa_ctx->per_ctx.size) {
-		const u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
-
-		GEM_BUG_ON(lrc_ring_wa_bb_per_ctx(engine) == -1);
-		regs[lrc_ring_wa_bb_per_ctx(engine) + 1] =
-			(ggtt_offset + wa_ctx->per_ctx.offset) | 0x01;
-	}
-
-	if (wa_ctx->indirect_ctx.size) {
-		lrc_ring_setup_indirect_ctx(regs, engine,
-					    i915_ggtt_offset(wa_ctx->vma) +
-					    wa_ctx->indirect_ctx.offset,
-					    wa_ctx->indirect_ctx.size);
-	}
-}
-
-static void init_ppgtt_reg_state(u32 *regs, const struct i915_ppgtt *ppgtt)
-{
-	if (i915_vm_is_4lvl(&ppgtt->vm)) {
-		/* 64b PPGTT (48bit canonical)
-		 * PDP0_DESCRIPTOR contains the base address to PML4 and
-		 * other PDP Descriptors are ignored.
-		 */
-		ASSIGN_CTX_PML4(ppgtt, regs);
-	} else {
-		ASSIGN_CTX_PDP(ppgtt, regs, 3);
-		ASSIGN_CTX_PDP(ppgtt, regs, 2);
-		ASSIGN_CTX_PDP(ppgtt, regs, 1);
-		ASSIGN_CTX_PDP(ppgtt, regs, 0);
-	}
-}
-
-static struct i915_ppgtt *vm_alias(struct i915_address_space *vm)
-{
-	if (i915_is_ggtt(vm))
-		return i915_vm_to_ggtt(vm)->alias;
-	else
-		return i915_vm_to_ppgtt(vm);
-}
-
-static void execlists_init_reg_state(u32 *regs,
-				     const struct intel_context *ce,
-				     const struct intel_engine_cs *engine,
-				     const struct intel_ring *ring,
-				     bool inhibit)
-{
-	/*
-	 * A context is actually a big batch buffer with several
-	 * MI_LOAD_REGISTER_IMM commands followed by (reg, value) pairs. The
-	 * values we are setting here are only for the first context restore:
-	 * on a subsequent save, the GPU will recreate this batchbuffer with new
-	 * values (including all the missing MI_LOAD_REGISTER_IMM commands that
-	 * we are not initializing here).
-	 *
-	 * Must keep consistent with virtual_update_register_offsets().
-	 */
-	set_offsets(regs, reg_offsets(engine), engine, inhibit);
-
-	init_common_reg_state(regs, engine, ring, inhibit);
-	init_ppgtt_reg_state(regs, vm_alias(ce->vm));
-
-	init_wa_bb_reg_state(regs, engine);
-
-	__reset_stop_ring(regs, engine);
-}
-
-static int
-populate_lr_context(struct intel_context *ce,
-		    struct drm_i915_gem_object *ctx_obj,
-		    struct intel_engine_cs *engine,
-		    struct intel_ring *ring)
-{
-	bool inhibit = true;
-	void *vaddr;
-
-	vaddr = i915_gem_object_pin_map(ctx_obj, I915_MAP_WB);
-	if (IS_ERR(vaddr)) {
-		drm_dbg(&engine->i915->drm, "Could not map object pages!\n");
-		return PTR_ERR(vaddr);
-	}
-
-	set_redzone(vaddr, engine);
-
-	if (engine->default_state) {
-		shmem_read(engine->default_state, 0,
-			   vaddr, engine->context_size);
-		__set_bit(CONTEXT_VALID_BIT, &ce->flags);
-		inhibit = false;
-	}
-
-	/* Clear the ppHWSP (inc. per-context counters) */
-	memset(vaddr, 0, PAGE_SIZE);
-
-	/*
-	 * The second page of the context object contains some registers which
-	 * must be set up prior to the first execution.
-	 */
-	execlists_init_reg_state(vaddr + LRC_STATE_OFFSET,
-				 ce, engine, ring, inhibit);
-
-	__i915_gem_object_flush_map(ctx_obj, 0, engine->context_size);
-	i915_gem_object_unpin_map(ctx_obj);
-	return 0;
-}
-
-static struct intel_timeline *pinned_timeline(struct intel_context *ce)
-{
-	struct intel_timeline *tl = fetch_and_zero(&ce->timeline);
-
-	return intel_timeline_create_from_engine(ce->engine,
-						 page_unmask_bits(tl));
-}
-
-static int __execlists_context_alloc(struct intel_context *ce,
-				     struct intel_engine_cs *engine)
-{
-	struct drm_i915_gem_object *ctx_obj;
-	struct intel_ring *ring;
-	struct i915_vma *vma;
-	u32 context_size;
-	int ret;
-
-	GEM_BUG_ON(ce->state);
-	context_size = round_up(engine->context_size, I915_GTT_PAGE_SIZE);
-
-	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		context_size += I915_GTT_PAGE_SIZE; /* for redzone */
-
-	if (INTEL_GEN(engine->i915) == 12) {
-		ce->wa_bb_page = context_size / PAGE_SIZE;
-		context_size += PAGE_SIZE;
-	}
-
-	ctx_obj = i915_gem_object_create_shmem(engine->i915, context_size);
-	if (IS_ERR(ctx_obj))
-		return PTR_ERR(ctx_obj);
-
-	vma = i915_vma_instance(ctx_obj, &engine->gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		ret = PTR_ERR(vma);
-		goto error_deref_obj;
-	}
-
-	if (!page_mask_bits(ce->timeline)) {
-		struct intel_timeline *tl;
-
-		/*
-		 * Use the static global HWSP for the kernel context, and
-		 * a dynamically allocated cacheline for everyone else.
-		 */
-		if (unlikely(ce->timeline))
-			tl = pinned_timeline(ce);
-		else
-			tl = intel_timeline_create(engine->gt);
-		if (IS_ERR(tl)) {
-			ret = PTR_ERR(tl);
-			goto error_deref_obj;
-		}
-
-		ce->timeline = tl;
-	}
-
-	ring = intel_engine_create_ring(engine, (unsigned long)ce->ring);
-	if (IS_ERR(ring)) {
-		ret = PTR_ERR(ring);
-		goto error_deref_obj;
-	}
-
-	ret = populate_lr_context(ce, ctx_obj, engine, ring);
-	if (ret) {
-		drm_dbg(&engine->i915->drm,
-			"Failed to populate LRC: %d\n", ret);
-		goto error_ring_free;
-	}
-
-	ce->ring = ring;
-	ce->state = vma;
-
-	return 0;
-
-error_ring_free:
-	intel_ring_put(ring);
-error_deref_obj:
-	i915_gem_object_put(ctx_obj);
-	return ret;
-}
-
-static struct list_head *virtual_queue(struct virtual_engine *ve)
-{
-	return &ve->base.execlists.default_priolist.requests[0];
-}
-
-static void rcu_virtual_context_destroy(struct work_struct *wrk)
-{
-	struct virtual_engine *ve =
-		container_of(wrk, typeof(*ve), rcu.work);
-	unsigned int n;
-
-	GEM_BUG_ON(ve->context.inflight);
-
-	/* Preempt-to-busy may leave a stale request behind. */
-	if (unlikely(ve->request)) {
-		struct i915_request *old;
-
-		spin_lock_irq(&ve->base.active.lock);
-
-		old = fetch_and_zero(&ve->request);
-		if (old) {
-			GEM_BUG_ON(!i915_request_completed(old));
-			__i915_request_submit(old);
-			i915_request_put(old);
-		}
-
-		spin_unlock_irq(&ve->base.active.lock);
-	}
-
-	/*
-	 * Flush the tasklet in case it is still running on another core.
-	 *
-	 * This needs to be done before we remove ourselves from the siblings'
-	 * rbtrees as in the case it is running in parallel, it may reinsert
-	 * the rb_node into a sibling.
-	 */
-	tasklet_kill(&ve->base.execlists.tasklet);
-
-	/* Decouple ourselves from the siblings, no more access allowed. */
-	for (n = 0; n < ve->num_siblings; n++) {
-		struct intel_engine_cs *sibling = ve->siblings[n];
-		struct rb_node *node = &ve->nodes[sibling->id].rb;
-
-		if (RB_EMPTY_NODE(node))
-			continue;
-
-		spin_lock_irq(&sibling->active.lock);
-
-		/* Detachment is lazily performed in the execlists tasklet */
-		if (!RB_EMPTY_NODE(node))
-			rb_erase_cached(node, &sibling->execlists.virtual);
-
-		spin_unlock_irq(&sibling->active.lock);
-	}
-	GEM_BUG_ON(__tasklet_is_scheduled(&ve->base.execlists.tasklet));
-	GEM_BUG_ON(!list_empty(virtual_queue(ve)));
-
-	if (ve->context.state)
-		__execlists_context_fini(&ve->context);
-	intel_context_fini(&ve->context);
-
-	intel_breadcrumbs_free(ve->base.breadcrumbs);
-	intel_engine_free_request_pool(&ve->base);
-
-	kfree(ve->bonds);
-	kfree(ve);
-}
-
-static void virtual_context_destroy(struct kref *kref)
-{
-	struct virtual_engine *ve =
-		container_of(kref, typeof(*ve), context.ref);
-
-	GEM_BUG_ON(!list_empty(&ve->context.signals));
-
-	/*
-	 * When destroying the virtual engine, we have to be aware that
-	 * it may still be in use from an hardirq/softirq context causing
-	 * the resubmission of a completed request (background completion
-	 * due to preempt-to-busy). Before we can free the engine, we need
-	 * to flush the submission code and tasklets that are still potentially
-	 * accessing the engine. Flushing the tasklets requires process context,
-	 * and since we can guard the resubmit onto the engine with an RCU read
-	 * lock, we can delegate the free of the engine to an RCU worker.
-	 */
-	INIT_RCU_WORK(&ve->rcu, rcu_virtual_context_destroy);
-	queue_rcu_work(system_wq, &ve->rcu);
-}
-
-static void virtual_engine_initial_hint(struct virtual_engine *ve)
-{
-	int swp;
-
-	/*
-	 * Pick a random sibling on starting to help spread the load around.
-	 *
-	 * New contexts are typically created with exactly the same order
-	 * of siblings, and often started in batches. Due to the way we iterate
-	 * the array of sibling when submitting requests, sibling[0] is
-	 * prioritised for dequeuing. If we make sure that sibling[0] is fairly
-	 * randomised across the system, we also help spread the load by the
-	 * first engine we inspect being different each time.
-	 *
-	 * NB This does not force us to execute on this engine, it will just
-	 * typically be the first we inspect for submission.
-	 */
-	swp = prandom_u32_max(ve->num_siblings);
-	if (swp)
-		swap(ve->siblings[swp], ve->siblings[0]);
-}
-
-static int virtual_context_alloc(struct intel_context *ce)
-{
-	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
-
-	return __execlists_context_alloc(ce, ve->siblings[0]);
-}
-
-static int virtual_context_pin(struct intel_context *ce, void *vaddr)
-{
-	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
-
-	/* Note: we must use a real engine class for setting up reg state */
-	return __execlists_context_pin(ce, ve->siblings[0], vaddr);
-}
-
-static void virtual_context_enter(struct intel_context *ce)
-{
-	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
-	unsigned int n;
-
-	for (n = 0; n < ve->num_siblings; n++)
-		intel_engine_pm_get(ve->siblings[n]);
-
-	intel_timeline_enter(ce->timeline);
-}
-
-static void virtual_context_exit(struct intel_context *ce)
+static void st_update_runtime_underflow(struct intel_context *ce, s32 dt)
 {
-	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
-	unsigned int n;
-
-	intel_timeline_exit(ce->timeline);
-
-	for (n = 0; n < ve->num_siblings; n++)
-		intel_engine_pm_put(ve->siblings[n]);
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+	ce->runtime.num_underflow++;
+	ce->runtime.max_underflow = max_t(u32, ce->runtime.max_underflow, -dt);
+#endif
 }
 
-static const struct intel_context_ops virtual_context_ops = {
-	.alloc = virtual_context_alloc,
-
-	.pre_pin = execlists_context_pre_pin,
-	.pin = virtual_context_pin,
-	.unpin = execlists_context_unpin,
-	.post_unpin = execlists_context_post_unpin,
-
-	.enter = virtual_context_enter,
-	.exit = virtual_context_exit,
-
-	.destroy = virtual_context_destroy,
-};
-
-static intel_engine_mask_t virtual_submission_mask(struct virtual_engine *ve)
+void lrc_update_runtime(struct intel_context *ce)
 {
-	struct i915_request *rq;
-	intel_engine_mask_t mask;
-
-	rq = READ_ONCE(ve->request);
-	if (!rq)
-		return 0;
-
-	/* The rq is ready for submission; rq->execution_mask is now stable. */
-	mask = rq->execution_mask;
-	if (unlikely(!mask)) {
-		/* Invalid selection, submit to a random engine in error */
-		i915_request_set_error_once(rq, -ENODEV);
-		mask = ve->siblings[0]->mask;
-	}
-
-	ENGINE_TRACE(&ve->base, "rq=%llx:%lld, mask=%x, prio=%d\n",
-		     rq->fence.context, rq->fence.seqno,
-		     mask, ve->base.execlists.queue_priority_hint);
-
-	return mask;
-}
+	u32 old;
+	s32 dt;
 
-static void virtual_submission_tasklet(unsigned long data)
-{
-	struct virtual_engine * const ve = (struct virtual_engine *)data;
-	const int prio = READ_ONCE(ve->base.execlists.queue_priority_hint);
-	intel_engine_mask_t mask;
-	unsigned int n;
-
-	rcu_read_lock();
-	mask = virtual_submission_mask(ve);
-	rcu_read_unlock();
-	if (unlikely(!mask))
+	if (intel_context_is_barrier(ce))
 		return;
 
-	local_irq_disable();
-	for (n = 0; n < ve->num_siblings; n++) {
-		struct intel_engine_cs *sibling = READ_ONCE(ve->siblings[n]);
-		struct ve_node * const node = &ve->nodes[sibling->id];
-		struct rb_node **parent, *rb;
-		bool first;
-
-		if (!READ_ONCE(ve->request))
-			break; /* already handled by a sibling's tasklet */
-
-		if (unlikely(!(mask & sibling->mask))) {
-			if (!RB_EMPTY_NODE(&node->rb)) {
-				spin_lock(&sibling->active.lock);
-				rb_erase_cached(&node->rb,
-						&sibling->execlists.virtual);
-				RB_CLEAR_NODE(&node->rb);
-				spin_unlock(&sibling->active.lock);
-			}
-			continue;
-		}
-
-		spin_lock(&sibling->active.lock);
-
-		if (!RB_EMPTY_NODE(&node->rb)) {
-			/*
-			 * Cheat and avoid rebalancing the tree if we can
-			 * reuse this node in situ.
-			 */
-			first = rb_first_cached(&sibling->execlists.virtual) ==
-				&node->rb;
-			if (prio == node->prio || (prio > node->prio && first))
-				goto submit_engine;
-
-			rb_erase_cached(&node->rb, &sibling->execlists.virtual);
-		}
-
-		rb = NULL;
-		first = true;
-		parent = &sibling->execlists.virtual.rb_root.rb_node;
-		while (*parent) {
-			struct ve_node *other;
-
-			rb = *parent;
-			other = rb_entry(rb, typeof(*other), rb);
-			if (prio > other->prio) {
-				parent = &rb->rb_left;
-			} else {
-				parent = &rb->rb_right;
-				first = false;
-			}
-		}
-
-		rb_link_node(&node->rb, rb, parent);
-		rb_insert_color_cached(&node->rb,
-				       &sibling->execlists.virtual,
-				       first);
-
-submit_engine:
-		GEM_BUG_ON(RB_EMPTY_NODE(&node->rb));
-		node->prio = prio;
-		if (first && prio > sibling->execlists.queue_priority_hint)
-			tasklet_hi_schedule(&sibling->execlists.tasklet);
-
-		spin_unlock(&sibling->active.lock);
-	}
-	local_irq_enable();
-}
-
-static void virtual_submit_request(struct i915_request *rq)
-{
-	struct virtual_engine *ve = to_virtual_engine(rq->engine);
-	struct i915_request *old;
-	unsigned long flags;
-
-	ENGINE_TRACE(&ve->base, "rq=%llx:%lld\n",
-		     rq->fence.context,
-		     rq->fence.seqno);
-
-	GEM_BUG_ON(ve->base.submit_request != virtual_submit_request);
-
-	spin_lock_irqsave(&ve->base.active.lock, flags);
-
-	old = ve->request;
-	if (old) { /* background completion event from preempt-to-busy */
-		GEM_BUG_ON(!i915_request_completed(old));
-		__i915_request_submit(old);
-		i915_request_put(old);
-	}
-
-	if (i915_request_completed(rq)) {
-		__i915_request_submit(rq);
-
-		ve->base.execlists.queue_priority_hint = INT_MIN;
-		ve->request = NULL;
-	} else {
-		ve->base.execlists.queue_priority_hint = rq_prio(rq);
-		ve->request = i915_request_get(rq);
-
-		GEM_BUG_ON(!list_empty(virtual_queue(ve)));
-		list_move_tail(&rq->sched.link, virtual_queue(ve));
-
-		tasklet_hi_schedule(&ve->base.execlists.tasklet);
-	}
-
-	spin_unlock_irqrestore(&ve->base.active.lock, flags);
-}
-
-static struct ve_bond *
-virtual_find_bond(struct virtual_engine *ve,
-		  const struct intel_engine_cs *master)
-{
-	int i;
-
-	for (i = 0; i < ve->num_bonds; i++) {
-		if (ve->bonds[i].master == master)
-			return &ve->bonds[i];
-	}
-
-	return NULL;
-}
-
-static void
-virtual_bond_execute(struct i915_request *rq, struct dma_fence *signal)
-{
-	struct virtual_engine *ve = to_virtual_engine(rq->engine);
-	intel_engine_mask_t allowed, exec;
-	struct ve_bond *bond;
-
-	allowed = ~to_request(signal)->engine->mask;
-
-	bond = virtual_find_bond(ve, to_request(signal)->engine);
-	if (bond)
-		allowed &= bond->sibling_mask;
-
-	/* Restrict the bonded request to run on only the available engines */
-	exec = READ_ONCE(rq->execution_mask);
-	while (!try_cmpxchg(&rq->execution_mask, &exec, exec & allowed))
-		;
-
-	/* Prevent the master from being re-run on the bonded engines */
-	to_request(signal)->execution_mask &= ~allowed;
-}
-
-struct intel_context *
-intel_execlists_create_virtual(struct intel_engine_cs **siblings,
-			       unsigned int count)
-{
-	struct virtual_engine *ve;
-	unsigned int n;
-	int err;
-
-	if (count == 0)
-		return ERR_PTR(-EINVAL);
-
-	if (count == 1)
-		return intel_context_create(siblings[0]);
-
-	ve = kzalloc(struct_size(ve, siblings, count), GFP_KERNEL);
-	if (!ve)
-		return ERR_PTR(-ENOMEM);
-
-	ve->base.i915 = siblings[0]->i915;
-	ve->base.gt = siblings[0]->gt;
-	ve->base.uncore = siblings[0]->uncore;
-	ve->base.id = -1;
-
-	ve->base.class = OTHER_CLASS;
-	ve->base.uabi_class = I915_ENGINE_CLASS_INVALID;
-	ve->base.instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
-	ve->base.uabi_instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
-
-	/*
-	 * The decision on whether to submit a request using semaphores
-	 * depends on the saturated state of the engine. We only compute
-	 * this during HW submission of the request, and we need for this
-	 * state to be globally applied to all requests being submitted
-	 * to this engine. Virtual engines encompass more than one physical
-	 * engine and so we cannot accurately tell in advance if one of those
-	 * engines is already saturated and so cannot afford to use a semaphore
-	 * and be pessimized in priority for doing so -- if we are the only
-	 * context using semaphores after all other clients have stopped, we
-	 * will be starved on the saturated system. Such a global switch for
-	 * semaphores is less than ideal, but alas is the current compromise.
-	 */
-	ve->base.saturated = ALL_ENGINES;
-
-	snprintf(ve->base.name, sizeof(ve->base.name), "virtual");
-
-	intel_engine_init_active(&ve->base, ENGINE_VIRTUAL);
-	intel_engine_init_execlists(&ve->base);
-
-	ve->base.cops = &virtual_context_ops;
-	ve->base.request_alloc = execlists_request_alloc;
-
-	ve->base.schedule = i915_schedule;
-	ve->base.submit_request = virtual_submit_request;
-	ve->base.bond_execute = virtual_bond_execute;
-
-	INIT_LIST_HEAD(virtual_queue(ve));
-	ve->base.execlists.queue_priority_hint = INT_MIN;
-	tasklet_init(&ve->base.execlists.tasklet,
-		     virtual_submission_tasklet,
-		     (unsigned long)ve);
-
-	intel_context_init(&ve->context, &ve->base);
-
-	ve->base.breadcrumbs = intel_breadcrumbs_create(NULL);
-	if (!ve->base.breadcrumbs) {
-		err = -ENOMEM;
-		goto err_put;
-	}
-
-	for (n = 0; n < count; n++) {
-		struct intel_engine_cs *sibling = siblings[n];
-
-		GEM_BUG_ON(!is_power_of_2(sibling->mask));
-		if (sibling->mask & ve->base.mask) {
-			DRM_DEBUG("duplicate %s entry in load balancer\n",
-				  sibling->name);
-			err = -EINVAL;
-			goto err_put;
-		}
-
-		/*
-		 * The virtual engine implementation is tightly coupled to
-		 * the execlists backend -- we push out request directly
-		 * into a tree inside each physical engine. We could support
-		 * layering if we handle cloning of the requests and
-		 * submitting a copy into each backend.
-		 */
-		if (sibling->execlists.tasklet.func !=
-		    execlists_submission_tasklet) {
-			err = -ENODEV;
-			goto err_put;
-		}
-
-		GEM_BUG_ON(RB_EMPTY_NODE(&ve->nodes[sibling->id].rb));
-		RB_CLEAR_NODE(&ve->nodes[sibling->id].rb);
-
-		ve->siblings[ve->num_siblings++] = sibling;
-		ve->base.mask |= sibling->mask;
-
-		/*
-		 * All physical engines must be compatible for their emission
-		 * functions (as we build the instructions during request
-		 * construction and do not alter them before submission
-		 * on the physical engine). We use the engine class as a guide
-		 * here, although that could be refined.
-		 */
-		if (ve->base.class != OTHER_CLASS) {
-			if (ve->base.class != sibling->class) {
-				DRM_DEBUG("invalid mixing of engine class, sibling %d, already %d\n",
-					  sibling->class, ve->base.class);
-				err = -EINVAL;
-				goto err_put;
-			}
-			continue;
-		}
-
-		ve->base.class = sibling->class;
-		ve->base.uabi_class = sibling->uabi_class;
-		snprintf(ve->base.name, sizeof(ve->base.name),
-			 "v%dx%d", ve->base.class, count);
-		ve->base.context_size = sibling->context_size;
-
-		ve->base.emit_bb_start = sibling->emit_bb_start;
-		ve->base.emit_flush = sibling->emit_flush;
-		ve->base.emit_init_breadcrumb = sibling->emit_init_breadcrumb;
-		ve->base.emit_fini_breadcrumb = sibling->emit_fini_breadcrumb;
-		ve->base.emit_fini_breadcrumb_dw =
-			sibling->emit_fini_breadcrumb_dw;
-
-		ve->base.flags = sibling->flags;
-	}
-
-	ve->base.flags |= I915_ENGINE_IS_VIRTUAL;
-
-	virtual_engine_initial_hint(ve);
-	return &ve->context;
-
-err_put:
-	intel_context_put(&ve->context);
-	return ERR_PTR(err);
-}
-
-struct intel_context *
-intel_execlists_clone_virtual(struct intel_engine_cs *src)
-{
-	struct virtual_engine *se = to_virtual_engine(src);
-	struct intel_context *dst;
-
-	dst = intel_execlists_create_virtual(se->siblings,
-					     se->num_siblings);
-	if (IS_ERR(dst))
-		return dst;
-
-	if (se->num_bonds) {
-		struct virtual_engine *de = to_virtual_engine(dst->engine);
-
-		de->bonds = kmemdup(se->bonds,
-				    sizeof(*se->bonds) * se->num_bonds,
-				    GFP_KERNEL);
-		if (!de->bonds) {
-			intel_context_put(dst);
-			return ERR_PTR(-ENOMEM);
-		}
-
-		de->num_bonds = se->num_bonds;
-	}
-
-	return dst;
-}
-
-int intel_virtual_engine_attach_bond(struct intel_engine_cs *engine,
-				     const struct intel_engine_cs *master,
-				     const struct intel_engine_cs *sibling)
-{
-	struct virtual_engine *ve = to_virtual_engine(engine);
-	struct ve_bond *bond;
-	int n;
-
-	/* Sanity check the sibling is part of the virtual engine */
-	for (n = 0; n < ve->num_siblings; n++)
-		if (sibling == ve->siblings[n])
-			break;
-	if (n == ve->num_siblings)
-		return -EINVAL;
-
-	bond = virtual_find_bond(ve, master);
-	if (bond) {
-		bond->sibling_mask |= sibling->mask;
-		return 0;
-	}
-
-	bond = krealloc(ve->bonds,
-			sizeof(*bond) * (ve->num_bonds + 1),
-			GFP_KERNEL);
-	if (!bond)
-		return -ENOMEM;
-
-	bond[ve->num_bonds].master = master;
-	bond[ve->num_bonds].sibling_mask = sibling->mask;
-
-	ve->bonds = bond;
-	ve->num_bonds++;
-
-	return 0;
-}
-
-void intel_execlists_show_requests(struct intel_engine_cs *engine,
-				   struct drm_printer *m,
-				   void (*show_request)(struct drm_printer *m,
-							struct i915_request *rq,
-							const char *prefix),
-				   unsigned int max)
-{
-	const struct intel_engine_execlists *execlists = &engine->execlists;
-	struct i915_request *rq, *last;
-	unsigned long flags;
-	unsigned int count;
-	struct rb_node *rb;
-
-	spin_lock_irqsave(&engine->active.lock, flags);
-
-	last = NULL;
-	count = 0;
-	list_for_each_entry(rq, &engine->active.requests, sched.link) {
-		if (count++ < max - 1)
-			show_request(m, rq, "\t\tE ");
-		else
-			last = rq;
-	}
-	if (last) {
-		if (count > max) {
-			drm_printf(m,
-				   "\t\t...skipping %d executing requests...\n",
-				   count - max);
-		}
-		show_request(m, last, "\t\tE ");
-	}
-
-	if (execlists->switch_priority_hint != INT_MIN)
-		drm_printf(m, "\t\tSwitch priority hint: %d\n",
-			   READ_ONCE(execlists->switch_priority_hint));
-	if (execlists->queue_priority_hint != INT_MIN)
-		drm_printf(m, "\t\tQueue priority hint: %d\n",
-			   READ_ONCE(execlists->queue_priority_hint));
-
-	last = NULL;
-	count = 0;
-	for (rb = rb_first_cached(&execlists->queue); rb; rb = rb_next(rb)) {
-		struct i915_priolist *p = rb_entry(rb, typeof(*p), node);
-		int i;
-
-		priolist_for_each_request(rq, p, i) {
-			if (count++ < max - 1)
-				show_request(m, rq, "\t\tQ ");
-			else
-				last = rq;
-		}
-	}
-	if (last) {
-		if (count > max) {
-			drm_printf(m,
-				   "\t\t...skipping %d queued requests...\n",
-				   count - max);
-		}
-		show_request(m, last, "\t\tQ ");
-	}
+	old = ce->runtime.last;
+	ce->runtime.last = lrc_get_runtime(ce);
+	dt = ce->runtime.last - old;
 
-	last = NULL;
-	count = 0;
-	for (rb = rb_first_cached(&execlists->virtual); rb; rb = rb_next(rb)) {
-		struct virtual_engine *ve =
-			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
-		struct i915_request *rq = READ_ONCE(ve->request);
-
-		if (rq) {
-			if (count++ < max - 1)
-				show_request(m, rq, "\t\tV ");
-			else
-				last = rq;
-		}
-	}
-	if (last) {
-		if (count > max) {
-			drm_printf(m,
-				   "\t\t...skipping %d virtual requests...\n",
-				   count - max);
-		}
-		show_request(m, last, "\t\tV ");
+	if (unlikely(dt < 0)) {
+		CE_TRACE(ce, "runtime underflow: last=%u, new=%u, delta=%d\n",
+			 old, ce->runtime.last, dt);
+		st_update_runtime_underflow(ce, dt);
+		return;
 	}
 
-	spin_unlock_irqrestore(&engine->active.lock, flags);
-}
-
-void intel_lr_context_reset(struct intel_engine_cs *engine,
-			    struct intel_context *ce,
-			    u32 head,
-			    bool scrub)
-{
-	GEM_BUG_ON(!intel_context_is_pinned(ce));
-
-	/*
-	 * We want a simple context + ring to execute the breadcrumb update.
-	 * We cannot rely on the context being intact across the GPU hang,
-	 * so clear it and rebuild just what we need for the breadcrumb.
-	 * All pending requests for this context will be zapped, and any
-	 * future request will be after userspace has had the opportunity
-	 * to recreate its own state.
-	 */
-	if (scrub)
-		restore_default_state(ce, engine);
-
-	/* Rerun the request; its payload has been neutered (if guilty). */
-	__execlists_update_reg_state(ce, engine, head);
-}
-
-bool
-intel_engine_in_execlists_submission_mode(const struct intel_engine_cs *engine)
-{
-	return engine->set_default_submission ==
-	       intel_execlists_set_default_submission;
+	ewma_runtime_add(&ce->runtime.avg, dt);
+	ce->runtime.total += dt;
 }
 
 #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.h b/drivers/gpu/drm/i915/gt/intel_lrc.h
index c2d287f25497..7f697845c4cf 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.h
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.h
@@ -1,90 +1,20 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright © 2014 Intel Corporation
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
  */
 
-#ifndef _INTEL_LRC_H_
-#define _INTEL_LRC_H_
+#ifndef __INTEL_LRC_H__
+#define __INTEL_LRC_H__
 
 #include <linux/types.h>
 
-struct drm_printer;
+#include "intel_context.h"
+#include "intel_lrc_reg.h"
 
-struct drm_i915_private;
-struct i915_gem_context;
-struct i915_request;
-struct intel_context;
+struct drm_i915_gem_object;
 struct intel_engine_cs;
+struct intel_ring;
 
-/* Execlists regs */
-#define RING_ELSP(base)				_MMIO((base) + 0x230)
-#define RING_EXECLIST_STATUS_LO(base)		_MMIO((base) + 0x234)
-#define RING_EXECLIST_STATUS_HI(base)		_MMIO((base) + 0x234 + 4)
-#define RING_CONTEXT_CONTROL(base)		_MMIO((base) + 0x244)
-#define	  CTX_CTRL_INHIBIT_SYN_CTX_SWITCH	(1 << 3)
-#define	  CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT	(1 << 0)
-#define   CTX_CTRL_RS_CTX_ENABLE		(1 << 1)
-#define	  CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT	(1 << 2)
-#define	  GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE	(1 << 8)
-#define RING_CONTEXT_STATUS_PTR(base)		_MMIO((base) + 0x3a0)
-#define RING_EXECLIST_SQ_CONTENTS(base)		_MMIO((base) + 0x510)
-#define RING_EXECLIST_CONTROL(base)		_MMIO((base) + 0x550)
-
-#define	  EL_CTRL_LOAD				(1 << 0)
-
-/* The docs specify that the write pointer wraps around after 5h, "After status
- * is written out to the last available status QW at offset 5h, this pointer
- * wraps to 0."
- *
- * Therefore, one must infer than even though there are 3 bits available, 6 and
- * 7 appear to be * reserved.
- */
-#define GEN8_CSB_ENTRIES 6
-#define GEN8_CSB_PTR_MASK 0x7
-#define GEN8_CSB_READ_PTR_MASK (GEN8_CSB_PTR_MASK << 8)
-#define GEN8_CSB_WRITE_PTR_MASK (GEN8_CSB_PTR_MASK << 0)
-
-#define GEN11_CSB_ENTRIES 12
-#define GEN11_CSB_PTR_MASK 0xf
-#define GEN11_CSB_READ_PTR_MASK (GEN11_CSB_PTR_MASK << 8)
-#define GEN11_CSB_WRITE_PTR_MASK (GEN11_CSB_PTR_MASK << 0)
-
-#define MAX_CONTEXT_HW_ID (1<<21) /* exclusive */
-#define MAX_GUC_CONTEXT_HW_ID (1 << 20) /* exclusive */
-#define GEN11_MAX_CONTEXT_HW_ID (1<<11) /* exclusive */
-/* in Gen12 ID 0x7FF is reserved to indicate idle */
-#define GEN12_MAX_CONTEXT_HW_ID	(GEN11_MAX_CONTEXT_HW_ID - 1)
-
-enum {
-	INTEL_CONTEXT_SCHEDULE_IN = 0,
-	INTEL_CONTEXT_SCHEDULE_OUT,
-	INTEL_CONTEXT_SCHEDULE_PREEMPTED,
-};
-
-/* Logical Rings */
-void intel_logical_ring_cleanup(struct intel_engine_cs *engine);
-
-int intel_execlists_submission_setup(struct intel_engine_cs *engine);
-
-/* Logical Ring Contexts */
 /* At the start of the context image is its per-process HWS page */
 #define LRC_PPHWSP_PN	(0)
 #define LRC_PPHWSP_SZ	(1)
@@ -96,32 +26,57 @@ int intel_execlists_submission_setup(struct intel_engine_cs *engine);
 #define LRC_PPHWSP_SCRATCH		0x34
 #define LRC_PPHWSP_SCRATCH_ADDR		(LRC_PPHWSP_SCRATCH * sizeof(u32))
 
-void intel_execlists_set_default_submission(struct intel_engine_cs *engine);
-
-void intel_lr_context_reset(struct intel_engine_cs *engine,
-			    struct intel_context *ce,
-			    u32 head,
-			    bool scrub);
-
-void intel_execlists_show_requests(struct intel_engine_cs *engine,
-				   struct drm_printer *m,
-				   void (*show_request)(struct drm_printer *m,
-							struct i915_request *rq,
-							const char *prefix),
-				   unsigned int max);
-
-struct intel_context *
-intel_execlists_create_virtual(struct intel_engine_cs **siblings,
-			       unsigned int count);
-
-struct intel_context *
-intel_execlists_clone_virtual(struct intel_engine_cs *src);
-
-int intel_virtual_engine_attach_bond(struct intel_engine_cs *engine,
-				     const struct intel_engine_cs *master,
-				     const struct intel_engine_cs *sibling);
-
-bool
-intel_engine_in_execlists_submission_mode(const struct intel_engine_cs *engine);
-
-#endif /* _INTEL_LRC_H_ */
+void lrc_init_wa_ctx(struct intel_engine_cs *engine);
+void lrc_fini_wa_ctx(struct intel_engine_cs *engine);
+
+int lrc_alloc(struct intel_context *ce,
+	      struct intel_engine_cs *engine);
+void lrc_reset(struct intel_context *ce);
+void lrc_fini(struct intel_context *ce);
+void lrc_destroy(struct kref *kref);
+
+int
+lrc_pre_pin(struct intel_context *ce,
+	    struct intel_engine_cs *engine,
+	    struct i915_gem_ww_ctx *ww,
+	    void **vaddr);
+int
+lrc_pin(struct intel_context *ce,
+	struct intel_engine_cs *engine,
+	void *vaddr);
+void lrc_unpin(struct intel_context *ce);
+void lrc_post_unpin(struct intel_context *ce);
+
+void lrc_init_state(struct intel_context *ce,
+		    struct intel_engine_cs *engine,
+		    void *state);
+
+void lrc_init_regs(const struct intel_context *ce,
+		   const struct intel_engine_cs *engine,
+		   bool clear);
+void lrc_reset_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine);
+
+u32 lrc_update_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    u32 head);
+void lrc_update_offsets(struct intel_context *ce,
+			struct intel_engine_cs *engine);
+
+void lrc_check_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    const char *when);
+
+void lrc_update_runtime(struct intel_context *ce);
+static inline u32 lrc_get_runtime(const struct intel_context *ce)
+{
+	/*
+	 * We can use either ppHWSP[16] which is recorded before the context
+	 * switch (and so excludes the cost of context switches) or use the
+	 * value from the context image itself, which is saved/restored earlier
+	 * and so includes the cost of the save.
+	 */
+	return READ_ONCE(ce->lrc_reg_state[CTX_TIMESTAMP]);
+}
+
+#endif /* __INTEL_LRC_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
index 1b51f7b9a5c3..65fe76738335 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
+++ b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
@@ -9,6 +9,8 @@
 
 #include <linux/types.h>
 
+#define CTX_DESC_FORCE_RESTORE BIT_ULL(2)
+
 /* GEN8 to GEN12 Reg State Context */
 #define CTX_CONTEXT_CONTROL		(0x02 + 1)
 #define CTX_RING_HEAD			(0x04 + 1)
@@ -52,4 +54,43 @@
 #define GEN8_EXECLISTS_STATUS_BUF 0x370
 #define GEN11_EXECLISTS_STATUS_BUF2 0x3c0
 
+/* Execlists regs */
+#define RING_ELSP(base)				_MMIO((base) + 0x230)
+#define RING_EXECLIST_STATUS_LO(base)		_MMIO((base) + 0x234)
+#define RING_EXECLIST_STATUS_HI(base)		_MMIO((base) + 0x234 + 4)
+#define RING_CONTEXT_CONTROL(base)		_MMIO((base) + 0x244)
+#define	  CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT	REG_BIT(0)
+#define   CTX_CTRL_RS_CTX_ENABLE		REG_BIT(1)
+#define	  CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT	REG_BIT(2)
+#define	  CTX_CTRL_INHIBIT_SYN_CTX_SWITCH	REG_BIT(3)
+#define	  GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE	REG_BIT(8)
+#define RING_CONTEXT_STATUS_PTR(base)		_MMIO((base) + 0x3a0)
+#define RING_EXECLIST_SQ_CONTENTS(base)		_MMIO((base) + 0x510)
+#define RING_EXECLIST_CONTROL(base)		_MMIO((base) + 0x550)
+#define	  EL_CTRL_LOAD				REG_BIT(0)
+
+/*
+ * The docs specify that the write pointer wraps around after 5h, "After status
+ * is written out to the last available status QW at offset 5h, this pointer
+ * wraps to 0."
+ *
+ * Therefore, one must infer than even though there are 3 bits available, 6 and
+ * 7 appear to be * reserved.
+ */
+#define GEN8_CSB_ENTRIES 6
+#define GEN8_CSB_PTR_MASK 0x7
+#define GEN8_CSB_READ_PTR_MASK	(GEN8_CSB_PTR_MASK << 8)
+#define GEN8_CSB_WRITE_PTR_MASK	(GEN8_CSB_PTR_MASK << 0)
+
+#define GEN11_CSB_ENTRIES 12
+#define GEN11_CSB_PTR_MASK 0xf
+#define GEN11_CSB_READ_PTR_MASK		(GEN11_CSB_PTR_MASK << 8)
+#define GEN11_CSB_WRITE_PTR_MASK	(GEN11_CSB_PTR_MASK << 0)
+
+#define MAX_CONTEXT_HW_ID	(1 << 21) /* exclusive */
+#define MAX_GUC_CONTEXT_HW_ID	(1 << 20) /* exclusive */
+#define GEN11_MAX_CONTEXT_HW_ID	(1 << 11) /* exclusive */
+/* in Gen12 ID 0x7FF is reserved to indicate idle */
+#define GEN12_MAX_CONTEXT_HW_ID	(GEN11_MAX_CONTEXT_HW_ID - 1)
+
 #endif /* _INTEL_LRC_REG_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_mocs.c b/drivers/gpu/drm/i915/gt/intel_mocs.c
index ab6870242e18..8acb84960cd0 100644
--- a/drivers/gpu/drm/i915/gt/intel_mocs.c
+++ b/drivers/gpu/drm/i915/gt/intel_mocs.c
@@ -24,8 +24,8 @@
 
 #include "intel_engine.h"
 #include "intel_gt.h"
+#include "intel_lrc_reg.h"
 #include "intel_mocs.h"
-#include "intel_lrc.h"
 #include "intel_ring.h"
 
 /* structures required */
@@ -472,7 +472,7 @@ static u16 get_entry_l3cc(const struct drm_i915_mocs_table *table,
 	return table->table[I915_MOCS_PTE].l3cc_value;
 }
 
-static inline u32 l3cc_combine(u16 low, u16 high)
+static u32 l3cc_combine(u16 low, u16 high)
 {
 	return low | (u32)high << 16;
 }
diff --git a/drivers/gpu/drm/i915/gt/intel_ppgtt.c b/drivers/gpu/drm/i915/gt/intel_ppgtt.c
index 46d9aceda64c..96b85a10ef33 100644
--- a/drivers/gpu/drm/i915/gt/intel_ppgtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_ppgtt.c
@@ -80,7 +80,7 @@ void free_px(struct i915_address_space *vm, struct i915_page_table *pt, int lvl)
 	kfree(pt);
 }
 
-static inline void
+static void
 write_dma_entry(struct drm_i915_gem_object * const pdma,
 		const unsigned short idx,
 		const u64 encoded_entry)
diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.c b/drivers/gpu/drm/i915/gt/intel_rc6.c
index d7b8e4457fc2..35504c97f11d 100644
--- a/drivers/gpu/drm/i915/gt/intel_rc6.c
+++ b/drivers/gpu/drm/i915/gt/intel_rc6.c
@@ -49,7 +49,7 @@ static struct drm_i915_private *rc6_to_i915(struct intel_rc6 *rc)
 	return rc6_to_gt(rc)->i915;
 }
 
-static inline void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val)
+static void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val)
 {
 	intel_uncore_write_fw(uncore, reg, val);
 }
diff --git a/drivers/gpu/drm/i915/intel_region_lmem.c b/drivers/gpu/drm/i915/gt/intel_region_lmem.c
index 40d8f1a95df6..60393ce5614d 100644
--- a/drivers/gpu/drm/i915/intel_region_lmem.c
+++ b/drivers/gpu/drm/i915/gt/intel_region_lmem.c
@@ -95,10 +95,10 @@ region_lmem_init(struct intel_memory_region *mem)
 	return ret;
 }
 
-const struct intel_memory_region_ops intel_region_lmem_ops = {
+static const struct intel_memory_region_ops intel_region_lmem_ops = {
 	.init = region_lmem_init,
 	.release = region_lmem_release,
-	.create_object = __i915_gem_lmem_object_create,
+	.init_object = __i915_gem_lmem_object_init,
 };
 
 struct intel_memory_region *
diff --git a/drivers/gpu/drm/i915/intel_region_lmem.h b/drivers/gpu/drm/i915/gt/intel_region_lmem.h
index 213def7c7b8a..8ea43e538dab 100644
--- a/drivers/gpu/drm/i915/intel_region_lmem.h
+++ b/drivers/gpu/drm/i915/gt/intel_region_lmem.h
@@ -8,8 +8,6 @@
 
 struct drm_i915_private;
 
-extern const struct intel_memory_region_ops intel_region_lmem_ops;
-
 struct intel_memory_region *
 intel_setup_fake_lmem(struct drm_i915_private *i915);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_renderstate.c b/drivers/gpu/drm/i915/gt/intel_renderstate.c
index ea2a77c7b469..ca816ba22197 100644
--- a/drivers/gpu/drm/i915/gt/intel_renderstate.c
+++ b/drivers/gpu/drm/i915/gt/intel_renderstate.c
@@ -27,7 +27,8 @@
 
 #include "i915_drv.h"
 #include "intel_renderstate.h"
-#include "gt/intel_context.h"
+#include "intel_context.h"
+#include "intel_gpu_commands.h"
 #include "intel_ring.h"
 
 static const struct intel_renderstate_rodata *
diff --git a/drivers/gpu/drm/i915/gt/intel_reset.c b/drivers/gpu/drm/i915/gt/intel_reset.c
index 3654c955e6be..61410cd62927 100644
--- a/drivers/gpu/drm/i915/gt/intel_reset.c
+++ b/drivers/gpu/drm/i915/gt/intel_reset.c
@@ -40,20 +40,19 @@ static void rmw_clear_fw(struct intel_uncore *uncore, i915_reg_t reg, u32 clr)
 	intel_uncore_rmw_fw(uncore, reg, clr, 0);
 }
 
-static void engine_skip_context(struct i915_request *rq)
+static void skip_context(struct i915_request *rq)
 {
-	struct intel_engine_cs *engine = rq->engine;
 	struct intel_context *hung_ctx = rq->context;
 
-	if (!i915_request_is_active(rq))
-		return;
+	list_for_each_entry_from_rcu(rq, &hung_ctx->timeline->requests, link) {
+		if (!i915_request_is_active(rq))
+			return;
 
-	lockdep_assert_held(&engine->active.lock);
-	list_for_each_entry_continue(rq, &engine->active.requests, sched.link)
 		if (rq->context == hung_ctx) {
 			i915_request_set_error_once(rq, -EIO);
 			__i915_request_skip(rq);
 		}
+	}
 }
 
 static void client_mark_guilty(struct i915_gem_context *ctx, bool banned)
@@ -152,15 +151,14 @@ static void mark_innocent(struct i915_request *rq)
 void __i915_request_reset(struct i915_request *rq, bool guilty)
 {
 	RQ_TRACE(rq, "guilty? %s\n", yesno(guilty));
-
-	GEM_BUG_ON(i915_request_completed(rq));
+	GEM_BUG_ON(__i915_request_is_complete(rq));
 
 	rcu_read_lock(); /* protect the GEM context */
 	if (guilty) {
 		i915_request_set_error_once(rq, -EIO);
 		__i915_request_skip(rq);
 		if (mark_guilty(rq))
-			engine_skip_context(rq);
+			skip_context(rq);
 	} else {
 		i915_request_set_error_once(rq, -EAGAIN);
 		mark_innocent(rq);
@@ -231,7 +229,7 @@ static int g4x_do_reset(struct intel_gt *gt,
 			      GRDOM_MEDIA | GRDOM_RESET_ENABLE);
 	ret =  wait_for_atomic(g4x_reset_complete(pdev), 50);
 	if (ret) {
-		drm_dbg(&gt->i915->drm, "Wait for media reset failed\n");
+		GT_TRACE(gt, "Wait for media reset failed\n");
 		goto out;
 	}
 
@@ -239,7 +237,7 @@ static int g4x_do_reset(struct intel_gt *gt,
 			      GRDOM_RENDER | GRDOM_RESET_ENABLE);
 	ret =  wait_for_atomic(g4x_reset_complete(pdev), 50);
 	if (ret) {
-		drm_dbg(&gt->i915->drm, "Wait for render reset failed\n");
+		GT_TRACE(gt, "Wait for render reset failed\n");
 		goto out;
 	}
 
@@ -265,7 +263,7 @@ static int ilk_do_reset(struct intel_gt *gt, intel_engine_mask_t engine_mask,
 					   5000, 0,
 					   NULL);
 	if (ret) {
-		drm_dbg(&gt->i915->drm, "Wait for render reset failed\n");
+		GT_TRACE(gt, "Wait for render reset failed\n");
 		goto out;
 	}
 
@@ -276,7 +274,7 @@ static int ilk_do_reset(struct intel_gt *gt, intel_engine_mask_t engine_mask,
 					   5000, 0,
 					   NULL);
 	if (ret) {
-		drm_dbg(&gt->i915->drm, "Wait for media reset failed\n");
+		GT_TRACE(gt, "Wait for media reset failed\n");
 		goto out;
 	}
 
@@ -305,9 +303,9 @@ static int gen6_hw_domain_reset(struct intel_gt *gt, u32 hw_domain_mask)
 					   500, 0,
 					   NULL);
 	if (err)
-		drm_dbg(&gt->i915->drm,
-			"Wait for 0x%08x engines reset failed\n",
-			hw_domain_mask);
+		GT_TRACE(gt,
+			 "Wait for 0x%08x engines reset failed\n",
+			 hw_domain_mask);
 
 	return err;
 }
@@ -407,8 +405,7 @@ static int gen11_lock_sfc(struct intel_engine_cs *engine, u32 *hw_mask)
 		return 0;
 
 	if (ret) {
-		drm_dbg(&engine->i915->drm,
-			"Wait for SFC forced lock ack failed\n");
+		ENGINE_TRACE(engine, "Wait for SFC forced lock ack failed\n");
 		return ret;
 	}
 
@@ -499,6 +496,9 @@ static int gen8_engine_reset_prepare(struct intel_engine_cs *engine)
 	u32 request, mask, ack;
 	int ret;
 
+	if (I915_SELFTEST_ONLY(should_fail(&engine->reset_timeout, 1)))
+		return -ETIMEDOUT;
+
 	ack = intel_uncore_read_fw(uncore, reg);
 	if (ack & RESET_CTL_CAT_ERROR) {
 		/*
@@ -754,8 +754,10 @@ static int gt_reset(struct intel_gt *gt, intel_engine_mask_t stalled_mask)
 	if (err)
 		return err;
 
+	local_bh_disable();
 	for_each_engine(engine, gt, id)
 		__intel_engine_reset(engine, stalled_mask & engine->mask);
+	local_bh_enable();
 
 	intel_ggtt_restore_fences(gt->ggtt);
 
@@ -833,9 +835,11 @@ static void __intel_gt_set_wedged(struct intel_gt *gt)
 	set_bit(I915_WEDGED, &gt->reset.flags);
 
 	/* Mark all executing requests as skipped */
+	local_bh_disable();
 	for_each_engine(engine, gt, id)
 		if (engine->reset.cancel)
 			engine->reset.cancel(engine);
+	local_bh_enable();
 
 	reset_finish(gt, awake);
 
@@ -1105,25 +1109,12 @@ error:
 	goto finish;
 }
 
-static inline int intel_gt_reset_engine(struct intel_engine_cs *engine)
+static int intel_gt_reset_engine(struct intel_engine_cs *engine)
 {
 	return __intel_gt_reset(engine->gt, engine->mask);
 }
 
-/**
- * intel_engine_reset - reset GPU engine to recover from a hang
- * @engine: engine to reset
- * @msg: reason for GPU reset; or NULL for no drm_notice()
- *
- * Reset a specific GPU engine. Useful if a hang is detected.
- * Returns zero on successful reset or otherwise an error code.
- *
- * Procedure is:
- *  - identifies the request that caused the hang and it is dropped
- *  - reset engine (which will force the engine to idle)
- *  - re-init/configure engine
- */
-int intel_engine_reset(struct intel_engine_cs *engine, const char *msg)
+int __intel_engine_reset_bh(struct intel_engine_cs *engine, const char *msg)
 {
 	struct intel_gt *gt = engine->gt;
 	bool uses_guc = intel_engine_in_guc_submission_mode(engine);
@@ -1148,8 +1139,7 @@ int intel_engine_reset(struct intel_engine_cs *engine, const char *msg)
 		ret = intel_guc_reset_engine(&engine->gt->uc.guc, engine);
 	if (ret) {
 		/* If we fail here, we expect to fallback to a global reset */
-		drm_dbg(&gt->i915->drm, "%sFailed to reset %s, ret=%d\n",
-			uses_guc ? "GuC " : "", engine->name, ret);
+		ENGINE_TRACE(engine, "Failed to reset, err: %d\n", ret);
 		goto out;
 	}
 
@@ -1174,6 +1164,30 @@ out:
 	return ret;
 }
 
+/**
+ * intel_engine_reset - reset GPU engine to recover from a hang
+ * @engine: engine to reset
+ * @msg: reason for GPU reset; or NULL for no drm_notice()
+ *
+ * Reset a specific GPU engine. Useful if a hang is detected.
+ * Returns zero on successful reset or otherwise an error code.
+ *
+ * Procedure is:
+ *  - identifies the request that caused the hang and it is dropped
+ *  - reset engine (which will force the engine to idle)
+ *  - re-init/configure engine
+ */
+int intel_engine_reset(struct intel_engine_cs *engine, const char *msg)
+{
+	int err;
+
+	local_bh_disable();
+	err = __intel_engine_reset_bh(engine, msg);
+	local_bh_enable();
+
+	return err;
+}
+
 static void intel_gt_reset_global(struct intel_gt *gt,
 				  u32 engine_mask,
 				  const char *reason)
@@ -1186,7 +1200,7 @@ static void intel_gt_reset_global(struct intel_gt *gt,
 
 	kobject_uevent_env(kobj, KOBJ_CHANGE, error_event);
 
-	drm_dbg(&gt->i915->drm, "resetting chip, engines=%x\n", engine_mask);
+	GT_TRACE(gt, "resetting chip, engines=%x\n", engine_mask);
 	kobject_uevent_env(kobj, KOBJ_CHANGE, reset_event);
 
 	/* Use a watchdog to ensure that our reset completes */
@@ -1260,18 +1274,20 @@ void intel_gt_handle_error(struct intel_gt *gt,
 	 * single reset fails.
 	 */
 	if (intel_has_reset_engine(gt) && !intel_gt_is_wedged(gt)) {
+		local_bh_disable();
 		for_each_engine_masked(engine, gt, engine_mask, tmp) {
 			BUILD_BUG_ON(I915_RESET_MODESET >= I915_RESET_ENGINE);
 			if (test_and_set_bit(I915_RESET_ENGINE + engine->id,
 					     &gt->reset.flags))
 				continue;
 
-			if (intel_engine_reset(engine, msg) == 0)
+			if (__intel_engine_reset_bh(engine, msg) == 0)
 				engine_mask &= ~engine->mask;
 
 			clear_and_wake_up_bit(I915_RESET_ENGINE + engine->id,
 					      &gt->reset.flags);
 		}
+		local_bh_enable();
 	}
 
 	if (!engine_mask)
@@ -1380,6 +1396,17 @@ void intel_gt_init_reset(struct intel_gt *gt)
 	mutex_init(&gt->reset.mutex);
 	init_srcu_struct(&gt->reset.backoff_srcu);
 
+	/*
+	 * While undesirable to wait inside the shrinker, complain anyway.
+	 *
+	 * If we have to wait during shrinking, we guarantee forward progress
+	 * by forcing the reset. Therefore during the reset we must not
+	 * re-enter the shrinker. By declaring that we take the reset mutex
+	 * within the shrinker, we forbid ourselves from performing any
+	 * fs-reclaim or taking related locks during reset.
+	 */
+	i915_gem_shrinker_taints_mutex(gt->i915, &gt->reset.mutex);
+
 	/* no GPU until we are ready! */
 	__set_bit(I915_WEDGED, &gt->reset.flags);
 }
diff --git a/drivers/gpu/drm/i915/gt/intel_reset.h b/drivers/gpu/drm/i915/gt/intel_reset.h
index a0eec7c11c0c..7dbf5cc8a333 100644
--- a/drivers/gpu/drm/i915/gt/intel_reset.h
+++ b/drivers/gpu/drm/i915/gt/intel_reset.h
@@ -34,6 +34,8 @@ void intel_gt_reset(struct intel_gt *gt,
 		    const char *reason);
 int intel_engine_reset(struct intel_engine_cs *engine,
 		       const char *reason);
+int __intel_engine_reset_bh(struct intel_engine_cs *engine,
+			    const char *reason);
 
 void __i915_request_reset(struct i915_request *rq, bool guilty);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_ring.c b/drivers/gpu/drm/i915/gt/intel_ring.c
index 4034a4bac7f0..78d1360caa0f 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring.c
@@ -5,9 +5,11 @@
  */
 
 #include "gem/i915_gem_object.h"
+
 #include "i915_drv.h"
 #include "i915_vma.h"
 #include "intel_engine.h"
+#include "intel_gpu_commands.h"
 #include "intel_ring.h"
 #include "intel_timeline.h"
 
@@ -40,7 +42,7 @@ int intel_ring_pin(struct intel_ring *ring, struct i915_gem_ww_ctx *ww)
 	/* Ring wraparound at offset 0 sometimes hangs. No idea why. */
 	flags = PIN_OFFSET_BIAS | i915_ggtt_pin_bias(vma);
 
-	if (vma->obj->stolen)
+	if (i915_gem_object_is_stolen(vma->obj))
 		flags |= PIN_MAPPABLE;
 	else
 		flags |= PIN_HIGH;
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
index a41b43f445b8..4984ff565424 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -32,6 +32,7 @@
 #include "gen6_ppgtt.h"
 #include "gen7_renderclear.h"
 #include "i915_drv.h"
+#include "i915_mitigations.h"
 #include "intel_breadcrumbs.h"
 #include "intel_context.h"
 #include "intel_gt.h"
@@ -121,31 +122,27 @@ static void set_hwsp(struct intel_engine_cs *engine, u32 offset)
 		hwsp = RING_HWS_PGA(engine->mmio_base);
 	}
 
-	intel_uncore_write(engine->uncore, hwsp, offset);
-	intel_uncore_posting_read(engine->uncore, hwsp);
+	intel_uncore_write_fw(engine->uncore, hwsp, offset);
+	intel_uncore_posting_read_fw(engine->uncore, hwsp);
 }
 
 static void flush_cs_tlb(struct intel_engine_cs *engine)
 {
-	struct drm_i915_private *dev_priv = engine->i915;
-
-	if (!IS_GEN_RANGE(dev_priv, 6, 7))
+	if (!IS_GEN_RANGE(engine->i915, 6, 7))
 		return;
 
 	/* ring should be idle before issuing a sync flush*/
-	drm_WARN_ON(&dev_priv->drm,
-		    (ENGINE_READ(engine, RING_MI_MODE) & MODE_IDLE) == 0);
-
-	ENGINE_WRITE(engine, RING_INSTPM,
-		     _MASKED_BIT_ENABLE(INSTPM_TLB_INVALIDATE |
-					INSTPM_SYNC_FLUSH));
-	if (intel_wait_for_register(engine->uncore,
-				    RING_INSTPM(engine->mmio_base),
-				    INSTPM_SYNC_FLUSH, 0,
-				    1000))
-		drm_err(&dev_priv->drm,
-			"%s: wait for SyncFlush to complete for TLB invalidation timed out\n",
-			engine->name);
+	GEM_DEBUG_WARN_ON((ENGINE_READ(engine, RING_MI_MODE) & MODE_IDLE) == 0);
+
+	ENGINE_WRITE_FW(engine, RING_INSTPM,
+			_MASKED_BIT_ENABLE(INSTPM_TLB_INVALIDATE |
+					   INSTPM_SYNC_FLUSH));
+	if (__intel_wait_for_register_fw(engine->uncore,
+					 RING_INSTPM(engine->mmio_base),
+					 INSTPM_SYNC_FLUSH, 0,
+					 2000, 0, NULL))
+		ENGINE_TRACE(engine,
+			     "wait for SyncFlush to complete for TLB invalidation timed out\n");
 }
 
 static void ring_setup_status_page(struct intel_engine_cs *engine)
@@ -156,44 +153,6 @@ static void ring_setup_status_page(struct intel_engine_cs *engine)
 	flush_cs_tlb(engine);
 }
 
-static bool stop_ring(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-
-	if (INTEL_GEN(dev_priv) > 2) {
-		ENGINE_WRITE(engine,
-			     RING_MI_MODE, _MASKED_BIT_ENABLE(STOP_RING));
-		if (intel_wait_for_register(engine->uncore,
-					    RING_MI_MODE(engine->mmio_base),
-					    MODE_IDLE,
-					    MODE_IDLE,
-					    1000)) {
-			drm_err(&dev_priv->drm,
-				"%s : timed out trying to stop ring\n",
-				engine->name);
-
-			/*
-			 * Sometimes we observe that the idle flag is not
-			 * set even though the ring is empty. So double
-			 * check before giving up.
-			 */
-			if (ENGINE_READ(engine, RING_HEAD) !=
-			    ENGINE_READ(engine, RING_TAIL))
-				return false;
-		}
-	}
-
-	ENGINE_WRITE(engine, RING_HEAD, ENGINE_READ(engine, RING_TAIL));
-
-	ENGINE_WRITE(engine, RING_HEAD, 0);
-	ENGINE_WRITE(engine, RING_TAIL, 0);
-
-	/* The ring must be empty before it is disabled */
-	ENGINE_WRITE(engine, RING_CTL, 0);
-
-	return (ENGINE_READ(engine, RING_HEAD) & HEAD_ADDR) == 0;
-}
-
 static struct i915_address_space *vm_alias(struct i915_address_space *vm)
 {
 	if (i915_is_ggtt(vm))
@@ -211,9 +170,16 @@ static void set_pp_dir(struct intel_engine_cs *engine)
 {
 	struct i915_address_space *vm = vm_alias(engine->gt->vm);
 
-	if (vm) {
-		ENGINE_WRITE(engine, RING_PP_DIR_DCLV, PP_DIR_DCLV_2G);
-		ENGINE_WRITE(engine, RING_PP_DIR_BASE, pp_dir(vm));
+	if (!vm)
+		return;
+
+	ENGINE_WRITE_FW(engine, RING_PP_DIR_DCLV, PP_DIR_DCLV_2G);
+	ENGINE_WRITE_FW(engine, RING_PP_DIR_BASE, pp_dir(vm));
+
+	if (INTEL_GEN(engine->i915) >= 7) {
+		ENGINE_WRITE_FW(engine,
+				RING_MODE_GEN7,
+				_MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
 	}
 }
 
@@ -221,38 +187,10 @@ static int xcs_resume(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
 	struct intel_ring *ring = engine->legacy.ring;
-	int ret = 0;
 
 	ENGINE_TRACE(engine, "ring:{HEAD:%04x, TAIL:%04x}\n",
 		     ring->head, ring->tail);
 
-	intel_uncore_forcewake_get(engine->uncore, FORCEWAKE_ALL);
-
-	/* WaClearRingBufHeadRegAtInit:ctg,elk */
-	if (!stop_ring(engine)) {
-		/* G45 ring initialization often fails to reset head to zero */
-		drm_dbg(&dev_priv->drm, "%s head not reset to zero "
-			"ctl %08x head %08x tail %08x start %08x\n",
-			engine->name,
-			ENGINE_READ(engine, RING_CTL),
-			ENGINE_READ(engine, RING_HEAD),
-			ENGINE_READ(engine, RING_TAIL),
-			ENGINE_READ(engine, RING_START));
-
-		if (!stop_ring(engine)) {
-			drm_err(&dev_priv->drm,
-				"failed to set %s head to zero "
-				"ctl %08x head %08x tail %08x start %08x\n",
-				engine->name,
-				ENGINE_READ(engine, RING_CTL),
-				ENGINE_READ(engine, RING_HEAD),
-				ENGINE_READ(engine, RING_TAIL),
-				ENGINE_READ(engine, RING_START));
-			ret = -EIO;
-			goto out;
-		}
-	}
-
 	if (HWS_NEEDS_PHYSICAL(dev_priv))
 		ring_setup_phys_status_page(engine);
 	else
@@ -269,7 +207,7 @@ static int xcs_resume(struct intel_engine_cs *engine)
 	 * also enforces ordering), otherwise the hw might lose the new ring
 	 * register values.
 	 */
-	ENGINE_WRITE(engine, RING_START, i915_ggtt_offset(ring->vma));
+	ENGINE_WRITE_FW(engine, RING_START, i915_ggtt_offset(ring->vma));
 
 	/* Check that the ring offsets point within the ring! */
 	GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->head));
@@ -279,53 +217,98 @@ static int xcs_resume(struct intel_engine_cs *engine)
 	set_pp_dir(engine);
 
 	/* First wake the ring up to an empty/idle ring */
-	ENGINE_WRITE(engine, RING_HEAD, ring->head);
-	ENGINE_WRITE(engine, RING_TAIL, ring->head);
+	ENGINE_WRITE_FW(engine, RING_HEAD, ring->head);
+	ENGINE_WRITE_FW(engine, RING_TAIL, ring->head);
 	ENGINE_POSTING_READ(engine, RING_TAIL);
 
-	ENGINE_WRITE(engine, RING_CTL, RING_CTL_SIZE(ring->size) | RING_VALID);
+	ENGINE_WRITE_FW(engine, RING_CTL,
+			RING_CTL_SIZE(ring->size) | RING_VALID);
 
 	/* If the head is still not zero, the ring is dead */
-	if (intel_wait_for_register(engine->uncore,
-				    RING_CTL(engine->mmio_base),
-				    RING_VALID, RING_VALID,
-				    50)) {
-		drm_err(&dev_priv->drm, "%s initialization failed "
-			  "ctl %08x (valid? %d) head %08x [%08x] tail %08x [%08x] start %08x [expected %08x]\n",
-			  engine->name,
-			  ENGINE_READ(engine, RING_CTL),
-			  ENGINE_READ(engine, RING_CTL) & RING_VALID,
-			  ENGINE_READ(engine, RING_HEAD), ring->head,
-			  ENGINE_READ(engine, RING_TAIL), ring->tail,
-			  ENGINE_READ(engine, RING_START),
-			  i915_ggtt_offset(ring->vma));
-		ret = -EIO;
-		goto out;
+	if (__intel_wait_for_register_fw(engine->uncore,
+					 RING_CTL(engine->mmio_base),
+					 RING_VALID, RING_VALID,
+					 5000, 0, NULL)) {
+		drm_err(&dev_priv->drm,
+			"%s initialization failed; "
+			"ctl %08x (valid? %d) head %08x [%08x] tail %08x [%08x] start %08x [expected %08x]\n",
+			engine->name,
+			ENGINE_READ(engine, RING_CTL),
+			ENGINE_READ(engine, RING_CTL) & RING_VALID,
+			ENGINE_READ(engine, RING_HEAD), ring->head,
+			ENGINE_READ(engine, RING_TAIL), ring->tail,
+			ENGINE_READ(engine, RING_START),
+			i915_ggtt_offset(ring->vma));
+		return -EIO;
 	}
 
 	if (INTEL_GEN(dev_priv) > 2)
-		ENGINE_WRITE(engine,
-			     RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+		ENGINE_WRITE_FW(engine,
+				RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
 
 	/* Now awake, let it get started */
 	if (ring->tail != ring->head) {
-		ENGINE_WRITE(engine, RING_TAIL, ring->tail);
+		ENGINE_WRITE_FW(engine, RING_TAIL, ring->tail);
 		ENGINE_POSTING_READ(engine, RING_TAIL);
 	}
 
 	/* Papering over lost _interrupts_ immediately following the restart */
 	intel_engine_signal_breadcrumbs(engine);
-out:
-	intel_uncore_forcewake_put(engine->uncore, FORCEWAKE_ALL);
+	return 0;
+}
 
-	return ret;
+static void sanitize_hwsp(struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl;
+
+	list_for_each_entry(tl, &engine->status_page.timelines, engine_link)
+		intel_timeline_reset_seqno(tl);
 }
 
-static void reset_prepare(struct intel_engine_cs *engine)
+static void xcs_sanitize(struct intel_engine_cs *engine)
+{
+	/*
+	 * Poison residual state on resume, in case the suspend didn't!
+	 *
+	 * We have to assume that across suspend/resume (or other loss
+	 * of control) that the contents of our pinned buffers has been
+	 * lost, replaced by garbage. Since this doesn't always happen,
+	 * let's poison such state so that we more quickly spot when
+	 * we falsely assume it has been preserved.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
+
+	/*
+	 * The kernel_context HWSP is stored in the status_page. As above,
+	 * that may be lost on resume/initialisation, and so we need to
+	 * reset the value in the HWSP.
+	 */
+	sanitize_hwsp(engine);
+
+	/* And scrub the dirty cachelines for the HWSP */
+	clflush_cache_range(engine->status_page.addr, PAGE_SIZE);
+}
+
+static bool stop_ring(struct intel_engine_cs *engine)
 {
-	struct intel_uncore *uncore = engine->uncore;
-	const u32 base = engine->mmio_base;
+	/* Empty the ring by skipping to the end */
+	ENGINE_WRITE_FW(engine, RING_HEAD, ENGINE_READ_FW(engine, RING_TAIL));
+	ENGINE_POSTING_READ(engine, RING_HEAD);
+
+	/* The ring must be empty before it is disabled */
+	ENGINE_WRITE_FW(engine, RING_CTL, 0);
+	ENGINE_POSTING_READ(engine, RING_CTL);
+
+	/* Then reset the disabled ring */
+	ENGINE_WRITE_FW(engine, RING_HEAD, 0);
+	ENGINE_WRITE_FW(engine, RING_TAIL, 0);
+
+	return (ENGINE_READ_FW(engine, RING_HEAD) & HEAD_ADDR) == 0;
+}
 
+static void reset_prepare(struct intel_engine_cs *engine)
+{
 	/*
 	 * We stop engines, otherwise we might get failed reset and a
 	 * dead gpu (on elk). Also as modern gpu as kbl can suffer
@@ -337,30 +320,35 @@ static void reset_prepare(struct intel_engine_cs *engine)
 	 * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES)
 	 *
 	 * WaMediaResetMainRingCleanup:ctg,elk (presumably)
+	 * WaClearRingBufHeadRegAtInit:ctg,elk
 	 *
 	 * FIXME: Wa for more modern gens needs to be validated
 	 */
 	ENGINE_TRACE(engine, "\n");
+	intel_engine_stop_cs(engine);
 
-	if (intel_engine_stop_cs(engine))
-		ENGINE_TRACE(engine, "timed out on STOP_RING\n");
-
-	intel_uncore_write_fw(uncore,
-			      RING_HEAD(base),
-			      intel_uncore_read_fw(uncore, RING_TAIL(base)));
-	intel_uncore_posting_read_fw(uncore, RING_HEAD(base)); /* paranoia */
-
-	intel_uncore_write_fw(uncore, RING_HEAD(base), 0);
-	intel_uncore_write_fw(uncore, RING_TAIL(base), 0);
-	intel_uncore_posting_read_fw(uncore, RING_TAIL(base));
-
-	/* The ring must be empty before it is disabled */
-	intel_uncore_write_fw(uncore, RING_CTL(base), 0);
+	if (!stop_ring(engine)) {
+		/* G45 ring initialization often fails to reset head to zero */
+		drm_dbg(&engine->i915->drm,
+			"%s head not reset to zero "
+			"ctl %08x head %08x tail %08x start %08x\n",
+			engine->name,
+			ENGINE_READ_FW(engine, RING_CTL),
+			ENGINE_READ_FW(engine, RING_HEAD),
+			ENGINE_READ_FW(engine, RING_TAIL),
+			ENGINE_READ_FW(engine, RING_START));
+	}
 
-	/* Check acts as a post */
-	if (intel_uncore_read_fw(uncore, RING_HEAD(base)))
-		ENGINE_TRACE(engine, "ring head [%x] not parked\n",
-			     intel_uncore_read_fw(uncore, RING_HEAD(base)));
+	if (!stop_ring(engine)) {
+		drm_err(&engine->i915->drm,
+			"failed to set %s head to zero "
+			"ctl %08x head %08x tail %08x start %08x\n",
+			engine->name,
+			ENGINE_READ_FW(engine, RING_CTL),
+			ENGINE_READ_FW(engine, RING_HEAD),
+			ENGINE_READ_FW(engine, RING_TAIL),
+			ENGINE_READ_FW(engine, RING_START));
+	}
 }
 
 static void reset_rewind(struct intel_engine_cs *engine, bool stalled)
@@ -371,12 +359,14 @@ static void reset_rewind(struct intel_engine_cs *engine, bool stalled)
 
 	rq = NULL;
 	spin_lock_irqsave(&engine->active.lock, flags);
+	rcu_read_lock();
 	list_for_each_entry(pos, &engine->active.requests, sched.link) {
-		if (!i915_request_completed(pos)) {
+		if (!__i915_request_is_complete(pos)) {
 			rq = pos;
 			break;
 		}
 	}
+	rcu_read_unlock();
 
 	/*
 	 * The guilty request will get skipped on a hung engine.
@@ -440,10 +430,8 @@ static void reset_cancel(struct intel_engine_cs *engine)
 	spin_lock_irqsave(&engine->active.lock, flags);
 
 	/* Mark all submitted requests as skipped. */
-	list_for_each_entry(request, &engine->active.requests, sched.link) {
-		i915_request_set_error_once(request, -EIO);
-		i915_request_mark_complete(request);
-	}
+	list_for_each_entry(request, &engine->active.requests, sched.link)
+		i915_request_mark_eio(request);
 	intel_engine_signal_breadcrumbs(engine);
 
 	/* Remaining _unready_ requests will be nop'ed when submitted */
@@ -602,6 +590,7 @@ static int ring_context_pin(struct intel_context *ce, void *unused)
 static void ring_context_reset(struct intel_context *ce)
 {
 	intel_ring_reset(ce->ring, ce->ring->emit);
+	clear_bit(CONTEXT_VALID_BIT, &ce->flags);
 }
 
 static const struct intel_context_ops ring_context_ops = {
@@ -653,9 +642,9 @@ static int load_pd_dir(struct i915_request *rq,
 	return rq->engine->emit_flush(rq, EMIT_FLUSH);
 }
 
-static inline int mi_set_context(struct i915_request *rq,
-				 struct intel_context *ce,
-				 u32 flags)
+static int mi_set_context(struct i915_request *rq,
+			  struct intel_context *ce,
+			  u32 flags)
 {
 	struct intel_engine_cs *engine = rq->engine;
 	struct drm_i915_private *i915 = engine->i915;
@@ -886,7 +875,8 @@ static int switch_context(struct i915_request *rq)
 	GEM_BUG_ON(HAS_EXECLISTS(engine->i915));
 
 	if (engine->wa_ctx.vma && ce != engine->kernel_context) {
-		if (engine->wa_ctx.vma->private != ce) {
+		if (engine->wa_ctx.vma->private != ce &&
+		    i915_mitigate_clear_residuals()) {
 			ret = clear_residuals(rq);
 			if (ret)
 				return ret;
@@ -1069,6 +1059,8 @@ static void setup_common(struct intel_engine_cs *engine)
 	setup_irq(engine);
 
 	engine->resume = xcs_resume;
+	engine->sanitize = xcs_sanitize;
+
 	engine->reset.prepare = reset_prepare;
 	engine->reset.rewind = reset_rewind;
 	engine->reset.cancel = reset_cancel;
@@ -1290,7 +1282,7 @@ int intel_ring_submission_setup(struct intel_engine_cs *engine)
 
 	GEM_BUG_ON(timeline->hwsp_ggtt != engine->status_page.vma);
 
-	if (IS_HASWELL(engine->i915) && engine->class == RENDER_CLASS) {
+	if (IS_GEN(engine->i915, 7) && engine->class == RENDER_CLASS) {
 		err = gen7_ctx_switch_bb_init(engine);
 		if (err)
 			goto err_ring_unpin;
diff --git a/drivers/gpu/drm/i915/gt/intel_rps.c b/drivers/gpu/drm/i915/gt/intel_rps.c
index b629eeb14002..ee5835c29c03 100644
--- a/drivers/gpu/drm/i915/gt/intel_rps.c
+++ b/drivers/gpu/drm/i915/gt/intel_rps.c
@@ -43,7 +43,7 @@ static u32 rps_pm_sanitize_mask(struct intel_rps *rps, u32 mask)
 	return mask & ~rps->pm_intrmsk_mbz;
 }
 
-static inline void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val)
+static void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val)
 {
 	intel_uncore_write_fw(uncore, reg, val);
 }
@@ -400,7 +400,7 @@ static unsigned int gen5_invert_freq(struct intel_rps *rps,
 	return val;
 }
 
-static bool gen5_rps_set(struct intel_rps *rps, u8 val)
+static int __gen5_rps_set(struct intel_rps *rps, u8 val)
 {
 	struct intel_uncore *uncore = rps_to_uncore(rps);
 	u16 rgvswctl;
@@ -410,7 +410,7 @@ static bool gen5_rps_set(struct intel_rps *rps, u8 val)
 	rgvswctl = intel_uncore_read16(uncore, MEMSWCTL);
 	if (rgvswctl & MEMCTL_CMD_STS) {
 		DRM_DEBUG("gpu busy, RCS change rejected\n");
-		return false; /* still busy with another command */
+		return -EBUSY; /* still busy with another command */
 	}
 
 	/* Invert the frequency bin into an ips delay */
@@ -426,7 +426,18 @@ static bool gen5_rps_set(struct intel_rps *rps, u8 val)
 	rgvswctl |= MEMCTL_CMD_STS;
 	intel_uncore_write16(uncore, MEMSWCTL, rgvswctl);
 
-	return true;
+	return 0;
+}
+
+static int gen5_rps_set(struct intel_rps *rps, u8 val)
+{
+	int err;
+
+	spin_lock_irq(&mchdev_lock);
+	err = __gen5_rps_set(rps, val);
+	spin_unlock_irq(&mchdev_lock);
+
+	return err;
 }
 
 static unsigned long intel_pxfreq(u32 vidfreq)
@@ -557,7 +568,7 @@ static bool gen5_rps_enable(struct intel_rps *rps)
 			"stuck trying to change perf mode\n");
 	mdelay(1);
 
-	gen5_rps_set(rps, rps->cur_freq);
+	__gen5_rps_set(rps, rps->cur_freq);
 
 	rps->ips.last_count1 = intel_uncore_read(uncore, DMIEC);
 	rps->ips.last_count1 += intel_uncore_read(uncore, DDREC);
@@ -599,7 +610,7 @@ static void gen5_rps_disable(struct intel_rps *rps)
 	intel_uncore_write(uncore, MEMINTRSTS, MEMINT_EVAL_CHG);
 
 	/* Go back to the starting frequency */
-	gen5_rps_set(rps, rps->idle_freq);
+	__gen5_rps_set(rps, rps->idle_freq);
 	mdelay(1);
 	rgvswctl |= MEMCTL_CMD_STS;
 	intel_uncore_write(uncore, MEMSWCTL, rgvswctl);
@@ -797,20 +808,19 @@ static int rps_set(struct intel_rps *rps, u8 val, bool update)
 	struct drm_i915_private *i915 = rps_to_i915(rps);
 	int err;
 
-	if (INTEL_GEN(i915) < 6)
-		return 0;
-
 	if (val == rps->last_freq)
 		return 0;
 
 	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
 		err = vlv_rps_set(rps, val);
-	else
+	else if (INTEL_GEN(i915) >= 6)
 		err = gen6_rps_set(rps, val);
+	else
+		err = gen5_rps_set(rps, val);
 	if (err)
 		return err;
 
-	if (update)
+	if (update && INTEL_GEN(i915) >= 6)
 		gen6_rps_set_thresholds(rps, val);
 	rps->last_freq = val;
 
@@ -852,6 +862,8 @@ void intel_rps_park(struct intel_rps *rps)
 {
 	int adj;
 
+	GEM_BUG_ON(atomic_read(&rps->num_waiters));
+
 	if (!intel_rps_clear_active(rps))
 		return;
 
@@ -907,28 +919,27 @@ void intel_rps_park(struct intel_rps *rps)
 
 void intel_rps_boost(struct i915_request *rq)
 {
-	struct intel_rps *rps = &READ_ONCE(rq->engine)->gt->rps;
-	unsigned long flags;
-
-	if (i915_request_signaled(rq) || !intel_rps_is_active(rps))
+	if (i915_request_signaled(rq) || i915_request_has_waitboost(rq))
 		return;
 
 	/* Serializes with i915_request_retire() */
-	spin_lock_irqsave(&rq->lock, flags);
-	if (!i915_request_has_waitboost(rq) &&
-	    !dma_fence_is_signaled_locked(&rq->fence)) {
-		set_bit(I915_FENCE_FLAG_BOOST, &rq->fence.flags);
+	if (!test_and_set_bit(I915_FENCE_FLAG_BOOST, &rq->fence.flags)) {
+		struct intel_rps *rps = &READ_ONCE(rq->engine)->gt->rps;
+
+		if (atomic_fetch_inc(&rps->num_waiters))
+			return;
+
+		if (!intel_rps_is_active(rps))
+			return;
 
 		GT_TRACE(rps_to_gt(rps), "boost fence:%llx:%llx\n",
 			 rq->fence.context, rq->fence.seqno);
 
-		if (!atomic_fetch_inc(&rps->num_waiters) &&
-		    READ_ONCE(rps->cur_freq) < rps->boost_freq)
+		if (READ_ONCE(rps->cur_freq) < rps->boost_freq)
 			schedule_work(&rps->work);
 
-		atomic_inc(&rps->boosts);
+		WRITE_ONCE(rps->boosts, rps->boosts + 1); /* debug only */
 	}
-	spin_unlock_irqrestore(&rq->lock, flags);
 }
 
 int intel_rps_set(struct intel_rps *rps, u8 val)
@@ -1798,7 +1809,7 @@ void gen5_rps_irq_handler(struct intel_rps *rps)
 			 rps->min_freq_softlimit,
 			 rps->max_freq_softlimit);
 
-	if (new_freq != rps->cur_freq && gen5_rps_set(rps, new_freq))
+	if (new_freq != rps->cur_freq && !__gen5_rps_set(rps, new_freq))
 		rps->cur_freq = new_freq;
 
 	spin_unlock(&mchdev_lock);
@@ -2109,7 +2120,7 @@ bool i915_gpu_turbo_disable(void)
 
 	spin_lock_irq(&mchdev_lock);
 	rps->max_freq_softlimit = rps->min_freq;
-	ret = gen5_rps_set(&i915->gt.rps, rps->min_freq);
+	ret = !__gen5_rps_set(&i915->gt.rps, rps->min_freq);
 	spin_unlock_irq(&mchdev_lock);
 
 	drm_dev_put(&i915->drm);
diff --git a/drivers/gpu/drm/i915/gt/intel_rps_types.h b/drivers/gpu/drm/i915/gt/intel_rps_types.h
index 38083f0402d9..029fe13cf303 100644
--- a/drivers/gpu/drm/i915/gt/intel_rps_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_rps_types.h
@@ -93,7 +93,7 @@ struct intel_rps {
 	} power;
 
 	atomic_t num_waiters;
-	atomic_t boosts;
+	unsigned int boosts;
 
 	/* manual wa residency calculations */
 	struct intel_rps_ei ei;
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.c b/drivers/gpu/drm/i915/gt/intel_timeline.c
index 7ea94d201fe6..037b0e3ccbed 100644
--- a/drivers/gpu/drm/i915/gt/intel_timeline.c
+++ b/drivers/gpu/drm/i915/gt/intel_timeline.c
@@ -126,6 +126,10 @@ static void __rcu_cacheline_free(struct rcu_head *rcu)
 	struct intel_timeline_cacheline *cl =
 		container_of(rcu, typeof(*cl), rcu);
 
+	/* Must wait until after all *rq->hwsp are complete before removing */
+	i915_gem_object_unpin_map(cl->hwsp->vma->obj);
+	__idle_hwsp_free(cl->hwsp, ptr_unmask_bits(cl->vaddr, CACHELINE_BITS));
+
 	i915_active_fini(&cl->active);
 	kfree(cl);
 }
@@ -133,11 +137,6 @@ static void __rcu_cacheline_free(struct rcu_head *rcu)
 static void __idle_cacheline_free(struct intel_timeline_cacheline *cl)
 {
 	GEM_BUG_ON(!i915_active_is_idle(&cl->active));
-
-	i915_gem_object_unpin_map(cl->hwsp->vma->obj);
-	i915_vma_put(cl->hwsp->vma);
-	__idle_hwsp_free(cl->hwsp, ptr_unmask_bits(cl->vaddr, CACHELINE_BITS));
-
 	call_rcu(&cl->rcu, __rcu_cacheline_free);
 }
 
@@ -179,7 +178,6 @@ cacheline_alloc(struct intel_timeline_hwsp *hwsp, unsigned int cacheline)
 		return ERR_CAST(vaddr);
 	}
 
-	i915_vma_get(hwsp->vma);
 	cl->hwsp = hwsp;
 	cl->vaddr = page_pack_bits(vaddr, cacheline);
 
@@ -321,6 +319,25 @@ __intel_timeline_create(struct intel_gt *gt,
 	return timeline;
 }
 
+struct intel_timeline *
+intel_timeline_create_from_engine(struct intel_engine_cs *engine,
+				  unsigned int offset)
+{
+	struct i915_vma *hwsp = engine->status_page.vma;
+	struct intel_timeline *tl;
+
+	tl = __intel_timeline_create(engine->gt, hwsp, offset);
+	if (IS_ERR(tl))
+		return tl;
+
+	/* Borrow a nearby lock; we only create these timelines during init */
+	mutex_lock(&hwsp->vm->mutex);
+	list_add_tail(&tl->engine_link, &engine->status_page.timelines);
+	mutex_unlock(&hwsp->vm->mutex);
+
+	return tl;
+}
+
 void __intel_timeline_pin(struct intel_timeline *tl)
 {
 	GEM_BUG_ON(!atomic_read(&tl->pin_count));
@@ -565,11 +582,11 @@ int intel_timeline_read_hwsp(struct i915_request *from,
 
 	rcu_read_lock();
 	cl = rcu_dereference(from->hwsp_cacheline);
-	if (i915_request_completed(from)) /* confirm cacheline is valid */
+	if (i915_request_signaled(from)) /* confirm cacheline is valid */
 		goto unlock;
 	if (unlikely(!i915_active_acquire_if_busy(&cl->active)))
 		goto unlock; /* seqno wrapped and completed! */
-	if (unlikely(i915_request_completed(from)))
+	if (unlikely(__i915_request_is_complete(from)))
 		goto release;
 	rcu_read_unlock();
 
@@ -617,6 +634,86 @@ void intel_gt_fini_timelines(struct intel_gt *gt)
 	GEM_BUG_ON(!list_empty(&timelines->hwsp_free_list));
 }
 
+void intel_gt_show_timelines(struct intel_gt *gt,
+			     struct drm_printer *m,
+			     void (*show_request)(struct drm_printer *m,
+						  const struct i915_request *rq,
+						  const char *prefix,
+						  int indent))
+{
+	struct intel_gt_timelines *timelines = &gt->timelines;
+	struct intel_timeline *tl, *tn;
+	LIST_HEAD(free);
+
+	spin_lock(&timelines->lock);
+	list_for_each_entry_safe(tl, tn, &timelines->active_list, link) {
+		unsigned long count, ready, inflight;
+		struct i915_request *rq, *rn;
+		struct dma_fence *fence;
+
+		if (!mutex_trylock(&tl->mutex)) {
+			drm_printf(m, "Timeline %llx: busy; skipping\n",
+				   tl->fence_context);
+			continue;
+		}
+
+		intel_timeline_get(tl);
+		GEM_BUG_ON(!atomic_read(&tl->active_count));
+		atomic_inc(&tl->active_count); /* pin the list element */
+		spin_unlock(&timelines->lock);
+
+		count = 0;
+		ready = 0;
+		inflight = 0;
+		list_for_each_entry_safe(rq, rn, &tl->requests, link) {
+			if (i915_request_completed(rq))
+				continue;
+
+			count++;
+			if (i915_request_is_ready(rq))
+				ready++;
+			if (i915_request_is_active(rq))
+				inflight++;
+		}
+
+		drm_printf(m, "Timeline %llx: { ", tl->fence_context);
+		drm_printf(m, "count: %lu, ready: %lu, inflight: %lu",
+			   count, ready, inflight);
+		drm_printf(m, ", seqno: { current: %d, last: %d }",
+			   *tl->hwsp_seqno, tl->seqno);
+		fence = i915_active_fence_get(&tl->last_request);
+		if (fence) {
+			drm_printf(m, ", engine: %s",
+				   to_request(fence)->engine->name);
+			dma_fence_put(fence);
+		}
+		drm_printf(m, " }\n");
+
+		if (show_request) {
+			list_for_each_entry_safe(rq, rn, &tl->requests, link)
+				show_request(m, rq, "", 2);
+		}
+
+		mutex_unlock(&tl->mutex);
+		spin_lock(&timelines->lock);
+
+		/* Resume list iteration after reacquiring spinlock */
+		list_safe_reset_next(tl, tn, link);
+		if (atomic_dec_and_test(&tl->active_count))
+			list_del(&tl->link);
+
+		/* Defer the final release to after the spinlock */
+		if (refcount_dec_and_test(&tl->kref.refcount)) {
+			GEM_BUG_ON(atomic_read(&tl->active_count));
+			list_add(&tl->link, &free);
+		}
+	}
+	spin_unlock(&timelines->lock);
+
+	list_for_each_entry_safe(tl, tn, &free, link)
+		__intel_timeline_free(&tl->kref);
+}
+
 #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
 #include "gt/selftests/mock_timeline.c"
 #include "gt/selftest_timeline.c"
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.h b/drivers/gpu/drm/i915/gt/intel_timeline.h
index 9882cd911d8e..dcdee692a80e 100644
--- a/drivers/gpu/drm/i915/gt/intel_timeline.h
+++ b/drivers/gpu/drm/i915/gt/intel_timeline.h
@@ -31,6 +31,8 @@
 #include "i915_syncmap.h"
 #include "intel_timeline_types.h"
 
+struct drm_printer;
+
 struct intel_timeline *
 __intel_timeline_create(struct intel_gt *gt,
 			struct i915_vma *global_hwsp,
@@ -42,14 +44,9 @@ intel_timeline_create(struct intel_gt *gt)
 	return __intel_timeline_create(gt, NULL, 0);
 }
 
-static inline struct intel_timeline *
+struct intel_timeline *
 intel_timeline_create_from_engine(struct intel_engine_cs *engine,
-				  unsigned int offset)
-{
-	return __intel_timeline_create(engine->gt,
-				       engine->status_page.vma,
-				       offset);
-}
+				  unsigned int offset);
 
 static inline struct intel_timeline *
 intel_timeline_get(struct intel_timeline *timeline)
@@ -106,4 +103,18 @@ int intel_timeline_read_hwsp(struct i915_request *from,
 void intel_gt_init_timelines(struct intel_gt *gt);
 void intel_gt_fini_timelines(struct intel_gt *gt);
 
+void intel_gt_show_timelines(struct intel_gt *gt,
+			     struct drm_printer *m,
+			     void (*show_request)(struct drm_printer *m,
+						  const struct i915_request *rq,
+						  const char *prefix,
+						  int indent));
+
+static inline bool
+intel_timeline_is_last(const struct intel_timeline *tl,
+		       const struct i915_request *rq)
+{
+	return list_is_last_rcu(&rq->link, &tl->requests);
+}
+
 #endif
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline_types.h b/drivers/gpu/drm/i915/gt/intel_timeline_types.h
index 4474f487f589..e360f50706bf 100644
--- a/drivers/gpu/drm/i915/gt/intel_timeline_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_timeline_types.h
@@ -84,6 +84,8 @@ struct intel_timeline {
 	struct list_head link;
 	struct intel_gt *gt;
 
+	struct list_head engine_link;
+
 	struct kref kref;
 	struct rcu_head rcu;
 };
diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c
index adc9a8ea410a..ec366cf9ef56 100644
--- a/drivers/gpu/drm/i915/gt/intel_workarounds.c
+++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c
@@ -7,6 +7,7 @@
 #include "i915_drv.h"
 #include "intel_context.h"
 #include "intel_engine_pm.h"
+#include "intel_gpu_commands.h"
 #include "intel_gt.h"
 #include "intel_ring.h"
 #include "intel_workarounds.h"
@@ -194,7 +195,7 @@ static void wa_add(struct i915_wa_list *wal, i915_reg_t reg,
 }
 
 static void
-wa_write_masked_or(struct i915_wa_list *wal, i915_reg_t reg, u32 clear, u32 set)
+wa_write_clr_set(struct i915_wa_list *wal, i915_reg_t reg, u32 clear, u32 set)
 {
 	wa_add(wal, reg, clear, set, clear);
 }
@@ -202,21 +203,32 @@ wa_write_masked_or(struct i915_wa_list *wal, i915_reg_t reg, u32 clear, u32 set)
 static void
 wa_write(struct i915_wa_list *wal, i915_reg_t reg, u32 set)
 {
-	wa_write_masked_or(wal, reg, ~0, set);
+	wa_write_clr_set(wal, reg, ~0, set);
 }
 
 static void
 wa_write_or(struct i915_wa_list *wal, i915_reg_t reg, u32 set)
 {
-	wa_write_masked_or(wal, reg, set, set);
+	wa_write_clr_set(wal, reg, set, set);
 }
 
 static void
 wa_write_clr(struct i915_wa_list *wal, i915_reg_t reg, u32 clr)
 {
-	wa_write_masked_or(wal, reg, clr, 0);
+	wa_write_clr_set(wal, reg, clr, 0);
 }
 
+/*
+ * WA operations on "masked register". A masked register has the upper 16 bits
+ * documented as "masked" in b-spec. Its purpose is to allow writing to just a
+ * portion of the register without a rmw: you simply write in the upper 16 bits
+ * the mask of bits you are going to modify.
+ *
+ * The wa_masked_* family of functions already does the necessary operations to
+ * calculate the mask based on the parameters passed, so user only has to
+ * provide the lower 16 bits of that register.
+ */
+
 static void
 wa_masked_en(struct i915_wa_list *wal, i915_reg_t reg, u32 val)
 {
@@ -229,38 +241,36 @@ wa_masked_dis(struct i915_wa_list *wal, i915_reg_t reg, u32 val)
 	wa_add(wal, reg, 0, _MASKED_BIT_DISABLE(val), val);
 }
 
-#define WA_SET_BIT_MASKED(addr, mask) \
-	wa_masked_en(wal, (addr), (mask))
-
-#define WA_CLR_BIT_MASKED(addr, mask) \
-	wa_masked_dis(wal, (addr), (mask))
-
-#define WA_SET_FIELD_MASKED(addr, mask, value) \
-	wa_write_masked_or(wal, (addr), 0, _MASKED_FIELD((mask), (value)))
+static void
+wa_masked_field_set(struct i915_wa_list *wal, i915_reg_t reg,
+		    u32 mask, u32 val)
+{
+	wa_add(wal, reg, 0, _MASKED_FIELD(mask, val), mask);
+}
 
 static void gen6_ctx_workarounds_init(struct intel_engine_cs *engine,
 				      struct i915_wa_list *wal)
 {
-	WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING);
+	wa_masked_en(wal, INSTPM, INSTPM_FORCE_ORDERING);
 }
 
 static void gen7_ctx_workarounds_init(struct intel_engine_cs *engine,
 				      struct i915_wa_list *wal)
 {
-	WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING);
+	wa_masked_en(wal, INSTPM, INSTPM_FORCE_ORDERING);
 }
 
 static void gen8_ctx_workarounds_init(struct intel_engine_cs *engine,
 				      struct i915_wa_list *wal)
 {
-	WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING);
+	wa_masked_en(wal, INSTPM, INSTPM_FORCE_ORDERING);
 
 	/* WaDisableAsyncFlipPerfMode:bdw,chv */
-	WA_SET_BIT_MASKED(MI_MODE, ASYNC_FLIP_PERF_DISABLE);
+	wa_masked_en(wal, MI_MODE, ASYNC_FLIP_PERF_DISABLE);
 
 	/* WaDisablePartialInstShootdown:bdw,chv */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
-			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
+	wa_masked_en(wal, GEN8_ROW_CHICKEN,
+		     PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
 
 	/* Use Force Non-Coherent whenever executing a 3D context. This is a
 	 * workaround for for a possible hang in the unlikely event a TLB
@@ -268,9 +278,9 @@ static void gen8_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 */
 	/* WaForceEnableNonCoherent:bdw,chv */
 	/* WaHdcDisableFetchWhenMasked:bdw,chv */
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  HDC_DONOT_FETCH_MEM_WHEN_MASKED |
-			  HDC_FORCE_NON_COHERENT);
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     HDC_DONOT_FETCH_MEM_WHEN_MASKED |
+		     HDC_FORCE_NON_COHERENT);
 
 	/* From the Haswell PRM, Command Reference: Registers, CACHE_MODE_0:
 	 * "The Hierarchical Z RAW Stall Optimization allows non-overlapping
@@ -280,10 +290,10 @@ static void gen8_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 *
 	 * This optimization is off by default for BDW and CHV; turn it on.
 	 */
-	WA_CLR_BIT_MASKED(CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE);
+	wa_masked_dis(wal, CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE);
 
 	/* Wa4x4STCOptimizationDisable:bdw,chv */
-	WA_SET_BIT_MASKED(CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE);
+	wa_masked_en(wal, CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE);
 
 	/*
 	 * BSpec recommends 8x4 when MSAA is used,
@@ -293,7 +303,7 @@ static void gen8_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 * disable bit, which we don't touch here, but it's good
 	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
 	 */
-	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
+	wa_masked_field_set(wal, GEN7_GT_MODE,
 			    GEN6_WIZ_HASHING_MASK,
 			    GEN6_WIZ_HASHING_16x4);
 }
@@ -306,24 +316,24 @@ static void bdw_ctx_workarounds_init(struct intel_engine_cs *engine,
 	gen8_ctx_workarounds_init(engine, wal);
 
 	/* WaDisableThreadStallDopClockGating:bdw (pre-production) */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
+	wa_masked_en(wal, GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
 
 	/* WaDisableDopClockGating:bdw
 	 *
 	 * Also see the related UCGTCL1 write in bdw_init_clock_gating()
 	 * to disable EUTC clock gating.
 	 */
-	WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2,
-			  DOP_CLOCK_GATING_DISABLE);
+	wa_masked_en(wal, GEN7_ROW_CHICKEN2,
+		     DOP_CLOCK_GATING_DISABLE);
 
-	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
-			  GEN8_SAMPLER_POWER_BYPASS_DIS);
+	wa_masked_en(wal, HALF_SLICE_CHICKEN3,
+		     GEN8_SAMPLER_POWER_BYPASS_DIS);
 
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  /* WaForceContextSaveRestoreNonCoherent:bdw */
-			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
-			  /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */
-			  (IS_BDW_GT3(i915) ? HDC_FENCE_DEST_SLM_DISABLE : 0));
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     /* WaForceContextSaveRestoreNonCoherent:bdw */
+		     HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
+		     /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */
+		     (IS_BDW_GT3(i915) ? HDC_FENCE_DEST_SLM_DISABLE : 0));
 }
 
 static void chv_ctx_workarounds_init(struct intel_engine_cs *engine,
@@ -332,10 +342,10 @@ static void chv_ctx_workarounds_init(struct intel_engine_cs *engine,
 	gen8_ctx_workarounds_init(engine, wal);
 
 	/* WaDisableThreadStallDopClockGating:chv */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
+	wa_masked_en(wal, GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
 
 	/* Improve HiZ throughput on CHV. */
-	WA_SET_BIT_MASKED(HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X);
+	wa_masked_en(wal, HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X);
 }
 
 static void gen9_ctx_workarounds_init(struct intel_engine_cs *engine,
@@ -349,38 +359,38 @@ static void gen9_ctx_workarounds_init(struct intel_engine_cs *engine,
 		 * Must match Display Engine. See
 		 * WaCompressedResourceDisplayNewHashMode.
 		 */
-		WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-				  GEN9_PBE_COMPRESSED_HASH_SELECTION);
-		WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7,
-				  GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR);
+		wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+			     GEN9_PBE_COMPRESSED_HASH_SELECTION);
+		wa_masked_en(wal, GEN9_HALF_SLICE_CHICKEN7,
+			     GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR);
 	}
 
 	/* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */
 	/* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
-			  FLOW_CONTROL_ENABLE |
-			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
+	wa_masked_en(wal, GEN8_ROW_CHICKEN,
+		     FLOW_CONTROL_ENABLE |
+		     PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
 
 	/* WaEnableYV12BugFixInHalfSliceChicken7:skl,bxt,kbl,glk,cfl */
 	/* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl,cfl */
-	WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7,
-			  GEN9_ENABLE_YV12_BUGFIX |
-			  GEN9_ENABLE_GPGPU_PREEMPTION);
+	wa_masked_en(wal, GEN9_HALF_SLICE_CHICKEN7,
+		     GEN9_ENABLE_YV12_BUGFIX |
+		     GEN9_ENABLE_GPGPU_PREEMPTION);
 
 	/* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk,cfl */
 	/* WaDisablePartialResolveInVc:skl,bxt,kbl,cfl */
-	WA_SET_BIT_MASKED(CACHE_MODE_1,
-			  GEN8_4x4_STC_OPTIMIZATION_DISABLE |
-			  GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE);
+	wa_masked_en(wal, CACHE_MODE_1,
+		     GEN8_4x4_STC_OPTIMIZATION_DISABLE |
+		     GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE);
 
 	/* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk,cfl */
-	WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5,
-			  GEN9_CCS_TLB_PREFETCH_ENABLE);
+	wa_masked_dis(wal, GEN9_HALF_SLICE_CHICKEN5,
+		      GEN9_CCS_TLB_PREFETCH_ENABLE);
 
 	/* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl,cfl */
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
-			  HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE);
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
+		     HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE);
 
 	/* WaForceEnableNonCoherent and WaDisableHDCInvalidation are
 	 * both tied to WaForceContextSaveRestoreNonCoherent
@@ -396,19 +406,19 @@ static void gen9_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 */
 
 	/* WaForceEnableNonCoherent:skl,bxt,kbl,cfl */
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  HDC_FORCE_NON_COHERENT);
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     HDC_FORCE_NON_COHERENT);
 
 	/* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl,cfl */
 	if (IS_SKYLAKE(i915) ||
 	    IS_KABYLAKE(i915) ||
 	    IS_COFFEELAKE(i915) ||
 	    IS_COMETLAKE(i915))
-		WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
-				  GEN8_SAMPLER_POWER_BYPASS_DIS);
+		wa_masked_en(wal, HALF_SLICE_CHICKEN3,
+			     GEN8_SAMPLER_POWER_BYPASS_DIS);
 
 	/* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk,cfl */
-	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
+	wa_masked_en(wal, HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
 
 	/*
 	 * Supporting preemption with fine-granularity requires changes in the
@@ -422,16 +432,16 @@ static void gen9_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 */
 
 	/* WaDisable3DMidCmdPreemption:skl,bxt,glk,cfl,[cnl] */
-	WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL);
+	wa_masked_dis(wal, GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL);
 
 	/* WaDisableGPGPUMidCmdPreemption:skl,bxt,blk,cfl,[cnl] */
-	WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1,
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
 			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
 			    GEN9_PREEMPT_GPGPU_COMMAND_LEVEL);
 
 	/* WaClearHIZ_WM_CHICKEN3:bxt,glk */
 	if (IS_GEN9_LP(i915))
-		WA_SET_BIT_MASKED(GEN9_WM_CHICKEN3, GEN9_FACTOR_IN_CLR_VAL_HIZ);
+		wa_masked_en(wal, GEN9_WM_CHICKEN3, GEN9_FACTOR_IN_CLR_VAL_HIZ);
 }
 
 static void skl_tune_iz_hashing(struct intel_engine_cs *engine,
@@ -465,7 +475,7 @@ static void skl_tune_iz_hashing(struct intel_engine_cs *engine,
 		return;
 
 	/* Tune IZ hashing. See intel_device_info_runtime_init() */
-	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
+	wa_masked_field_set(wal, GEN7_GT_MODE,
 			    GEN9_IZ_HASHING_MASK(2) |
 			    GEN9_IZ_HASHING_MASK(1) |
 			    GEN9_IZ_HASHING_MASK(0),
@@ -487,12 +497,12 @@ static void bxt_ctx_workarounds_init(struct intel_engine_cs *engine,
 	gen9_ctx_workarounds_init(engine, wal);
 
 	/* WaDisableThreadStallDopClockGating:bxt */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
-			  STALL_DOP_GATING_DISABLE);
+	wa_masked_en(wal, GEN8_ROW_CHICKEN,
+		     STALL_DOP_GATING_DISABLE);
 
 	/* WaToEnableHwFixForPushConstHWBug:bxt */
-	WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-			  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
 }
 
 static void kbl_ctx_workarounds_init(struct intel_engine_cs *engine,
@@ -504,12 +514,12 @@ static void kbl_ctx_workarounds_init(struct intel_engine_cs *engine,
 
 	/* WaToEnableHwFixForPushConstHWBug:kbl */
 	if (IS_KBL_GT_REVID(i915, KBL_REVID_C0, REVID_FOREVER))
-		WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-				  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+		wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+			     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
 
 	/* WaDisableSbeCacheDispatchPortSharing:kbl */
-	WA_SET_BIT_MASKED(GEN7_HALF_SLICE_CHICKEN1,
-			  GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
+	wa_masked_en(wal, GEN7_HALF_SLICE_CHICKEN1,
+		     GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
 }
 
 static void glk_ctx_workarounds_init(struct intel_engine_cs *engine,
@@ -518,8 +528,8 @@ static void glk_ctx_workarounds_init(struct intel_engine_cs *engine,
 	gen9_ctx_workarounds_init(engine, wal);
 
 	/* WaToEnableHwFixForPushConstHWBug:glk */
-	WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-			  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
 }
 
 static void cfl_ctx_workarounds_init(struct intel_engine_cs *engine,
@@ -528,41 +538,41 @@ static void cfl_ctx_workarounds_init(struct intel_engine_cs *engine,
 	gen9_ctx_workarounds_init(engine, wal);
 
 	/* WaToEnableHwFixForPushConstHWBug:cfl */
-	WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-			  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
 
 	/* WaDisableSbeCacheDispatchPortSharing:cfl */
-	WA_SET_BIT_MASKED(GEN7_HALF_SLICE_CHICKEN1,
-			  GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
+	wa_masked_en(wal, GEN7_HALF_SLICE_CHICKEN1,
+		     GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
 }
 
 static void cnl_ctx_workarounds_init(struct intel_engine_cs *engine,
 				     struct i915_wa_list *wal)
 {
 	/* WaForceContextSaveRestoreNonCoherent:cnl */
-	WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0,
-			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT);
+	wa_masked_en(wal, CNL_HDC_CHICKEN0,
+		     HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT);
 
 	/* WaDisableReplayBufferBankArbitrationOptimization:cnl */
-	WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-			  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
 
 	/* WaPushConstantDereferenceHoldDisable:cnl */
-	WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE);
+	wa_masked_en(wal, GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE);
 
 	/* FtrEnableFastAnisoL1BankingFix:cnl */
-	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX);
+	wa_masked_en(wal, HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX);
 
 	/* WaDisable3DMidCmdPreemption:cnl */
-	WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL);
+	wa_masked_dis(wal, GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL);
 
 	/* WaDisableGPGPUMidCmdPreemption:cnl */
-	WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1,
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
 			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
 			    GEN9_PREEMPT_GPGPU_COMMAND_LEVEL);
 
 	/* WaDisableEarlyEOT:cnl */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, DISABLE_EARLY_EOT);
+	wa_masked_en(wal, GEN8_ROW_CHICKEN, DISABLE_EARLY_EOT);
 }
 
 static void icl_ctx_workarounds_init(struct intel_engine_cs *engine,
@@ -580,8 +590,8 @@ static void icl_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 * Formerly known as WaPushConstantDereferenceHoldDisable
 	 */
 	if (IS_ICL_REVID(i915, ICL_REVID_A0, ICL_REVID_B0))
-		WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2,
-				  PUSH_CONSTANT_DEREF_DISABLE);
+		wa_masked_en(wal, GEN7_ROW_CHICKEN2,
+			     PUSH_CONSTANT_DEREF_DISABLE);
 
 	/* WaForceEnableNonCoherent:icl
 	 * This is not the same workaround as in early Gen9 platforms, where
@@ -590,40 +600,40 @@ static void icl_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 * (the register is whitelisted in hardware now, so UMDs can opt in
 	 * for coherency if they have a good reason).
 	 */
-	WA_SET_BIT_MASKED(ICL_HDC_MODE, HDC_FORCE_NON_COHERENT);
+	wa_masked_en(wal, ICL_HDC_MODE, HDC_FORCE_NON_COHERENT);
 
 	/* Wa_2006611047:icl (pre-prod)
 	 * Formerly known as WaDisableImprovedTdlClkGating
 	 */
 	if (IS_ICL_REVID(i915, ICL_REVID_A0, ICL_REVID_A0))
-		WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2,
-				  GEN11_TDL_CLOCK_GATING_FIX_DISABLE);
+		wa_masked_en(wal, GEN7_ROW_CHICKEN2,
+			     GEN11_TDL_CLOCK_GATING_FIX_DISABLE);
 
 	/* Wa_2006665173:icl (pre-prod) */
 	if (IS_ICL_REVID(i915, ICL_REVID_A0, ICL_REVID_A0))
-		WA_SET_BIT_MASKED(GEN11_COMMON_SLICE_CHICKEN3,
-				  GEN11_BLEND_EMB_FIX_DISABLE_IN_RCC);
+		wa_masked_en(wal, GEN11_COMMON_SLICE_CHICKEN3,
+			     GEN11_BLEND_EMB_FIX_DISABLE_IN_RCC);
 
 	/* WaEnableFloatBlendOptimization:icl */
-	wa_write_masked_or(wal,
-			   GEN10_CACHE_MODE_SS,
-			   0, /* write-only, so skip validation */
-			   _MASKED_BIT_ENABLE(FLOAT_BLEND_OPTIMIZATION_ENABLE));
+	wa_write_clr_set(wal,
+			 GEN10_CACHE_MODE_SS,
+			 0, /* write-only, so skip validation */
+			 _MASKED_BIT_ENABLE(FLOAT_BLEND_OPTIMIZATION_ENABLE));
 
 	/* WaDisableGPGPUMidThreadPreemption:icl */
-	WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1,
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
 			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
 			    GEN9_PREEMPT_GPGPU_THREAD_GROUP_LEVEL);
 
 	/* allow headerless messages for preemptible GPGPU context */
-	WA_SET_BIT_MASKED(GEN10_SAMPLER_MODE,
-			  GEN11_SAMPLER_ENABLE_HEADLESS_MSG);
+	wa_masked_en(wal, GEN10_SAMPLER_MODE,
+		     GEN11_SAMPLER_ENABLE_HEADLESS_MSG);
 
 	/* Wa_1604278689:icl,ehl */
 	wa_write(wal, IVB_FBC_RT_BASE, 0xFFFFFFFF & ~ILK_FBC_RT_VALID);
-	wa_write_masked_or(wal, IVB_FBC_RT_BASE_UPPER,
-			   0, /* write-only register; skip validation */
-			   0xFFFFFFFF);
+	wa_write_clr_set(wal, IVB_FBC_RT_BASE_UPPER,
+			 0, /* write-only register; skip validation */
+			 0xFFFFFFFF);
 
 	/* Wa_1406306137:icl,ehl */
 	wa_masked_en(wal, GEN9_ROW_CHICKEN4, GEN11_DIS_PICK_2ND_EU);
@@ -643,11 +653,11 @@ static void gen12_ctx_workarounds_init(struct intel_engine_cs *engine,
 	 * Wa_14010443199:rkl
 	 * Wa_14010698770:rkl
 	 */
-	WA_SET_BIT_MASKED(GEN11_COMMON_SLICE_CHICKEN3,
-			  GEN12_DISABLE_CPS_AWARE_COLOR_PIPE);
+	wa_masked_en(wal, GEN11_COMMON_SLICE_CHICKEN3,
+		     GEN12_DISABLE_CPS_AWARE_COLOR_PIPE);
 
 	/* WaDisableGPGPUMidThreadPreemption:gen12 */
-	WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1,
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
 			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
 			    GEN9_PREEMPT_GPGPU_THREAD_GROUP_LEVEL);
 }
@@ -680,12 +690,22 @@ static void dg1_ctx_workarounds_init(struct intel_engine_cs *engine,
 	gen12_ctx_workarounds_init(engine, wal);
 
 	/* Wa_1409044764 */
-	WA_CLR_BIT_MASKED(GEN11_COMMON_SLICE_CHICKEN3,
-			  DG1_FLOAT_POINT_BLEND_OPT_STRICT_MODE_EN);
+	wa_masked_dis(wal, GEN11_COMMON_SLICE_CHICKEN3,
+		      DG1_FLOAT_POINT_BLEND_OPT_STRICT_MODE_EN);
 
 	/* Wa_22010493298 */
-	WA_SET_BIT_MASKED(HIZ_CHICKEN,
-			  DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE);
+	wa_masked_en(wal, HIZ_CHICKEN,
+		     DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE);
+
+	/*
+	 * Wa_16011163337
+	 *
+	 * Like in tgl_ctx_workarounds_init(), read verification is ignored due
+	 * to Wa_1608008084.
+	 */
+	wa_add(wal,
+	       FF_MODE2,
+	       FF_MODE2_GS_TIMER_MASK, FF_MODE2_GS_TIMER_224, 0);
 }
 
 static void
@@ -804,57 +824,11 @@ ilk_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 static void
 snb_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 {
-	/* WaDisableHiZPlanesWhenMSAAEnabled:snb */
-	wa_masked_en(wal,
-		     _3D_CHICKEN,
-		     _3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB);
-
-	/* WaDisable_RenderCache_OperationalFlush:snb */
-	wa_masked_dis(wal, CACHE_MODE_0, RC_OP_FLUSH_ENABLE);
-
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	wa_add(wal,
-	       GEN6_GT_MODE, 0,
-	       _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4),
-	       GEN6_WIZ_HASHING_16x4);
-
-	wa_masked_dis(wal, CACHE_MODE_0, CM0_STC_EVICT_DISABLE_LRA_SNB);
-
-	wa_masked_en(wal,
-		     _3D_CHICKEN3,
-		     /* WaStripsFansDisableFastClipPerformanceFix:snb */
-		     _3D_CHICKEN3_SF_DISABLE_FASTCLIP_CULL |
-		     /*
-		      * Bspec says:
-		      * "This bit must be set if 3DSTATE_CLIP clip mode is set
-		      * to normal and 3DSTATE_SF number of SF output attributes
-		      * is more than 16."
-		      */
-		   _3D_CHICKEN3_SF_DISABLE_PIPELINED_ATTR_FETCH);
 }
 
 static void
 ivb_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 {
-	/* WaDisableEarlyCull:ivb */
-	wa_masked_en(wal, _3D_CHICKEN3, _3D_CHICKEN_SF_DISABLE_OBJEND_CULL);
-
-	/* WaDisablePSDDualDispatchEnable:ivb */
-	if (IS_IVB_GT1(i915))
-		wa_masked_en(wal,
-			     GEN7_HALF_SLICE_CHICKEN1,
-			     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
-
-	/* WaDisable_RenderCache_OperationalFlush:ivb */
-	wa_masked_dis(wal, CACHE_MODE_0_GEN7, RC_OP_FLUSH_ENABLE);
-
 	/* Apply the WaDisableRHWOOptimizationForRenderHang:ivb workaround. */
 	wa_masked_dis(wal,
 		      GEN7_COMMON_SLICE_CHICKEN1,
@@ -866,91 +840,15 @@ ivb_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 
 	/* WaForceL3Serialization:ivb */
 	wa_write_clr(wal, GEN7_L3SQCREG4, L3SQ_URB_READ_CAM_MATCH_DISABLE);
-
-	/*
-	 * WaVSThreadDispatchOverride:ivb,vlv
-	 *
-	 * This actually overrides the dispatch
-	 * mode for all thread types.
-	 */
-	wa_write_masked_or(wal, GEN7_FF_THREAD_MODE,
-			   GEN7_FF_SCHED_MASK,
-			   GEN7_FF_TS_SCHED_HW |
-			   GEN7_FF_VS_SCHED_HW |
-			   GEN7_FF_DS_SCHED_HW);
-
-	if (0) { /* causes HiZ corruption on ivb:gt1 */
-		/* enable HiZ Raw Stall Optimization */
-		wa_masked_dis(wal, CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE);
-	}
-
-	/* WaDisable4x2SubspanOptimization:ivb */
-	wa_masked_en(wal, CACHE_MODE_1, PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
-
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	wa_add(wal, GEN7_GT_MODE, 0,
-	       _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4),
-	       GEN6_WIZ_HASHING_16x4);
 }
 
 static void
 vlv_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 {
-	/* WaDisableEarlyCull:vlv */
-	wa_masked_en(wal, _3D_CHICKEN3, _3D_CHICKEN_SF_DISABLE_OBJEND_CULL);
-
-	/* WaPsdDispatchEnable:vlv */
-	/* WaDisablePSDDualDispatchEnable:vlv */
-	wa_masked_en(wal,
-		     GEN7_HALF_SLICE_CHICKEN1,
-		     GEN7_MAX_PS_THREAD_DEP |
-		     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
-
-	/* WaDisable_RenderCache_OperationalFlush:vlv */
-	wa_masked_dis(wal, CACHE_MODE_0_GEN7, RC_OP_FLUSH_ENABLE);
-
 	/* WaForceL3Serialization:vlv */
 	wa_write_clr(wal, GEN7_L3SQCREG4, L3SQ_URB_READ_CAM_MATCH_DISABLE);
 
 	/*
-	 * WaVSThreadDispatchOverride:ivb,vlv
-	 *
-	 * This actually overrides the dispatch
-	 * mode for all thread types.
-	 */
-	wa_write_masked_or(wal,
-			   GEN7_FF_THREAD_MODE,
-			   GEN7_FF_SCHED_MASK,
-			   GEN7_FF_TS_SCHED_HW |
-			   GEN7_FF_VS_SCHED_HW |
-			   GEN7_FF_DS_SCHED_HW);
-
-	/*
-	 * BSpec says this must be set, even though
-	 * WaDisable4x2SubspanOptimization isn't listed for VLV.
-	 */
-	wa_masked_en(wal, CACHE_MODE_1, PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
-
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	wa_add(wal, GEN7_GT_MODE, 0,
-	       _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4),
-	       GEN6_WIZ_HASHING_16x4);
-
-	/*
 	 * WaIncreaseL3CreditsForVLVB0:vlv
 	 * This is the hardware default actually.
 	 */
@@ -970,31 +868,6 @@ hsw_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 
 	/* WaVSRefCountFullforceMissDisable:hsw */
 	wa_write_clr(wal, GEN7_FF_THREAD_MODE, GEN7_FF_VS_REF_CNT_FFME);
-
-	wa_masked_dis(wal,
-		      CACHE_MODE_0_GEN7,
-		      /* WaDisable_RenderCache_OperationalFlush:hsw */
-		      RC_OP_FLUSH_ENABLE |
-		      /* enable HiZ Raw Stall Optimization */
-		      HIZ_RAW_STALL_OPT_DISABLE);
-
-	/* WaDisable4x2SubspanOptimization:hsw */
-	wa_masked_en(wal, CACHE_MODE_1, PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
-
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	wa_add(wal, GEN7_GT_MODE, 0,
-	       _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4),
-	       GEN6_WIZ_HASHING_16x4);
-
-	/* WaSampleCChickenBitEnable:hsw */
-	wa_masked_en(wal, HALF_SLICE_CHICKEN3, HSW_SAMPLE_C_PERFORMANCE);
 }
 
 static void
@@ -1164,7 +1037,7 @@ wa_init_mcr(struct drm_i915_private *i915, struct i915_wa_list *wal)
 
 	drm_dbg(&i915->drm, "MCR slice/subslice = %x\n", mcr);
 
-	wa_write_masked_or(wal, GEN8_MCR_SELECTOR, mcr_mask, mcr);
+	wa_write_clr_set(wal, GEN8_MCR_SELECTOR, mcr_mask, mcr);
 }
 
 static void
@@ -1189,10 +1062,10 @@ icl_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 		    GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
 
 	/* WaModifyGamTlbPartitioning:icl */
-	wa_write_masked_or(wal,
-			   GEN11_GACB_PERF_CTRL,
-			   GEN11_HASH_CTRL_MASK,
-			   GEN11_HASH_CTRL_BIT0 | GEN11_HASH_CTRL_BIT4);
+	wa_write_clr_set(wal,
+			 GEN11_GACB_PERF_CTRL,
+			 GEN11_HASH_CTRL_MASK,
+			 GEN11_HASH_CTRL_BIT0 | GEN11_HASH_CTRL_BIT4);
 
 	/* Wa_1405766107:icl
 	 * Formerly known as WaCL2SFHalfMaxAlloc
@@ -1260,6 +1133,11 @@ tgl_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal)
 		wa_write_or(wal,
 			    SLICE_UNIT_LEVEL_CLKGATE,
 			    L3_CLKGATE_DIS | L3_CR2X_CLKGATE_DIS);
+
+	/* Wa_1408615072:tgl[a0] */
+	if (IS_TGL_UY_GT_REVID(i915, TGL_REVID_A0, TGL_REVID_A0))
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE2,
+			    VSUNIT_CLKGATE_DIS_TGL);
 }
 
 static void
@@ -1358,9 +1236,9 @@ static bool
 wa_verify(const struct i915_wa *wa, u32 cur, const char *name, const char *from)
 {
 	if ((cur ^ wa->set) & wa->read) {
-		DRM_ERROR("%s workaround lost on %s! (%x=%x/%x, expected %x)\n",
+		DRM_ERROR("%s workaround lost on %s! (reg[%x]=0x%x, relevant bits were 0x%x vs expected 0x%x)\n",
 			  name, from, i915_mmio_reg_offset(wa->reg),
-			  cur, cur & wa->read, wa->set);
+			  cur, cur & wa->read, wa->set & wa->read);
 
 		return false;
 	}
@@ -1425,7 +1303,8 @@ bool intel_gt_verify_workarounds(struct intel_gt *gt, const char *from)
 	return wa_list_verify(gt->uncore, &gt->i915->gt_wa_list, from);
 }
 
-static inline bool is_nonpriv_flags_valid(u32 flags)
+__maybe_unused
+static bool is_nonpriv_flags_valid(u32 flags)
 {
 	/* Check only valid flag bits are set */
 	if (flags & ~RING_FORCE_TO_NONPRIV_MASK_VALID)
@@ -1752,10 +1631,6 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 		wa_write_or(wal,
 			    GEN7_SARCHKMD,
 			    GEN7_DISABLE_SAMPLER_PREFETCH);
-
-		/* Wa_1408615072:tgl */
-		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE2,
-			    VSUNIT_CLKGATE_DIS_TGL);
 	}
 
 	if (IS_DG1(i915) || IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) {
@@ -1770,6 +1645,19 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 		 */
 		wa_write_or(wal, GEN7_FF_THREAD_MODE,
 			    GEN12_FF_TESSELATION_DOP_GATE_DISABLE);
+
+		/*
+		 * Wa_1606700617:tgl,dg1
+		 * Wa_22010271021:tgl,rkl,dg1
+		 */
+		wa_masked_en(wal,
+			     GEN9_CS_DEBUG_MODE1,
+			     FF_DOP_CLOCK_GATE_DISABLE);
+
+		/* Wa_1406941453:tgl,rkl,dg1 */
+		wa_masked_en(wal,
+			     GEN10_SAMPLER_MODE,
+			     ENABLE_SMALLPL);
 	}
 
 	if (IS_DG1_REVID(i915, DG1_REVID_A0, DG1_REVID_A0) ||
@@ -1798,21 +1686,6 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 			     GEN6_RC_SLEEP_PSMI_CONTROL,
 			     GEN12_WAIT_FOR_EVENT_POWER_DOWN_DISABLE |
 			     GEN8_RC_SEMA_IDLE_MSG_DISABLE);
-
-		/*
-		 * Wa_1606700617:tgl
-		 * Wa_22010271021:tgl,rkl
-		 */
-		wa_masked_en(wal,
-			     GEN9_CS_DEBUG_MODE1,
-			     FF_DOP_CLOCK_GATE_DISABLE);
-	}
-
-	if (IS_GEN(i915, 12)) {
-		/* Wa_1406941453:gen12 */
-		wa_masked_en(wal,
-			     GEN10_SAMPLER_MODE,
-			     ENABLE_SMALLPL);
 	}
 
 	if (IS_GEN(i915, 11)) {
@@ -1838,14 +1711,14 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 		 * Wa_1604223664:icl
 		 * Formerly known as WaL3BankAddressHashing
 		 */
-		wa_write_masked_or(wal,
-				   GEN8_GARBCNTL,
-				   GEN11_HASH_CTRL_EXCL_MASK,
-				   GEN11_HASH_CTRL_EXCL_BIT0);
-		wa_write_masked_or(wal,
-				   GEN11_GLBLINVL,
-				   GEN11_BANK_HASH_ADDR_EXCL_MASK,
-				   GEN11_BANK_HASH_ADDR_EXCL_BIT0);
+		wa_write_clr_set(wal,
+				 GEN8_GARBCNTL,
+				 GEN11_HASH_CTRL_EXCL_MASK,
+				 GEN11_HASH_CTRL_EXCL_BIT0);
+		wa_write_clr_set(wal,
+				 GEN11_GLBLINVL,
+				 GEN11_BANK_HASH_ADDR_EXCL_MASK,
+				 GEN11_BANK_HASH_ADDR_EXCL_BIT0);
 
 		/*
 		 * Wa_1405733216:icl
@@ -1874,10 +1747,10 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 			    GEN7_DISABLE_SAMPLER_PREFETCH);
 
 		/* Wa_1409178092:icl */
-		wa_write_masked_or(wal,
-				   GEN11_SCRATCH2,
-				   GEN11_COHERENT_PARTIAL_WRITE_MERGE_ENABLE,
-				   0);
+		wa_write_clr_set(wal,
+				 GEN11_SCRATCH2,
+				 GEN11_COHERENT_PARTIAL_WRITE_MERGE_ENABLE,
+				 0);
 
 		/* WaEnable32PlaneMode:icl */
 		wa_masked_en(wal, GEN9_CSFE_CHICKEN1_RCS,
@@ -1951,24 +1824,132 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 
 		/* WaProgramL3SqcReg1DefaultForPerf:bxt,glk */
 		if (IS_GEN9_LP(i915))
-			wa_write_masked_or(wal,
-					   GEN8_L3SQCREG1,
-					   L3_PRIO_CREDITS_MASK,
-					   L3_GENERAL_PRIO_CREDITS(62) |
-					   L3_HIGH_PRIO_CREDITS(2));
+			wa_write_clr_set(wal,
+					 GEN8_L3SQCREG1,
+					 L3_PRIO_CREDITS_MASK,
+					 L3_GENERAL_PRIO_CREDITS(62) |
+					 L3_HIGH_PRIO_CREDITS(2));
 
 		/* WaOCLCoherentLineFlush:skl,bxt,kbl,cfl */
 		wa_write_or(wal,
 			    GEN8_L3SQCREG4,
 			    GEN8_LQSC_FLUSH_COHERENT_LINES);
+
+		/* Disable atomics in L3 to prevent unrecoverable hangs */
+		wa_write_clr_set(wal, GEN9_SCRATCH_LNCF1,
+				 GEN9_LNCF_NONIA_COHERENT_ATOMICS_ENABLE, 0);
+		wa_write_clr_set(wal, GEN8_L3SQCREG4,
+				 GEN8_LQSQ_NONIA_COHERENT_ATOMICS_ENABLE, 0);
+		wa_write_clr_set(wal, GEN9_SCRATCH1,
+				 EVICTION_PERF_FIX_ENABLE, 0);
+	}
+
+	if (IS_HASWELL(i915)) {
+		/* WaSampleCChickenBitEnable:hsw */
+		wa_masked_en(wal,
+			     HALF_SLICE_CHICKEN3, HSW_SAMPLE_C_PERFORMANCE);
+
+		wa_masked_dis(wal,
+			      CACHE_MODE_0_GEN7,
+			      /* enable HiZ Raw Stall Optimization */
+			      HIZ_RAW_STALL_OPT_DISABLE);
+
+		/* WaDisable4x2SubspanOptimization:hsw */
+		wa_masked_en(wal, CACHE_MODE_1, PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
+	}
+
+	if (IS_VALLEYVIEW(i915)) {
+		/* WaDisableEarlyCull:vlv */
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     _3D_CHICKEN_SF_DISABLE_OBJEND_CULL);
+
+		/*
+		 * WaVSThreadDispatchOverride:ivb,vlv
+		 *
+		 * This actually overrides the dispatch
+		 * mode for all thread types.
+		 */
+		wa_write_clr_set(wal,
+				 GEN7_FF_THREAD_MODE,
+				 GEN7_FF_SCHED_MASK,
+				 GEN7_FF_TS_SCHED_HW |
+				 GEN7_FF_VS_SCHED_HW |
+				 GEN7_FF_DS_SCHED_HW);
+
+		/* WaPsdDispatchEnable:vlv */
+		/* WaDisablePSDDualDispatchEnable:vlv */
+		wa_masked_en(wal,
+			     GEN7_HALF_SLICE_CHICKEN1,
+			     GEN7_MAX_PS_THREAD_DEP |
+			     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
 	}
 
-	if (IS_GEN(i915, 7))
+	if (IS_IVYBRIDGE(i915)) {
+		/* WaDisableEarlyCull:ivb */
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     _3D_CHICKEN_SF_DISABLE_OBJEND_CULL);
+
+		if (0) { /* causes HiZ corruption on ivb:gt1 */
+			/* enable HiZ Raw Stall Optimization */
+			wa_masked_dis(wal,
+				      CACHE_MODE_0_GEN7,
+				      HIZ_RAW_STALL_OPT_DISABLE);
+		}
+
+		/*
+		 * WaVSThreadDispatchOverride:ivb,vlv
+		 *
+		 * This actually overrides the dispatch
+		 * mode for all thread types.
+		 */
+		wa_write_clr_set(wal,
+				 GEN7_FF_THREAD_MODE,
+				 GEN7_FF_SCHED_MASK,
+				 GEN7_FF_TS_SCHED_HW |
+				 GEN7_FF_VS_SCHED_HW |
+				 GEN7_FF_DS_SCHED_HW);
+
+		/* WaDisablePSDDualDispatchEnable:ivb */
+		if (IS_IVB_GT1(i915))
+			wa_masked_en(wal,
+				     GEN7_HALF_SLICE_CHICKEN1,
+				     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
+	}
+
+	if (IS_GEN(i915, 7)) {
 		/* WaBCSVCSTlbInvalidationMode:ivb,vlv,hsw */
 		wa_masked_en(wal,
 			     GFX_MODE_GEN7,
 			     GFX_TLB_INVALIDATE_EXPLICIT | GFX_REPLAY_MODE);
 
+		/* WaDisable_RenderCache_OperationalFlush:ivb,vlv,hsw */
+		wa_masked_dis(wal, CACHE_MODE_0_GEN7, RC_OP_FLUSH_ENABLE);
+
+		/*
+		 * BSpec says this must be set, even though
+		 * WaDisable4x2SubspanOptimization:ivb,hsw
+		 * WaDisable4x2SubspanOptimization isn't listed for VLV.
+		 */
+		wa_masked_en(wal,
+			     CACHE_MODE_1,
+			     PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
+
+		/*
+		 * BSpec recommends 8x4 when MSAA is used,
+		 * however in practice 16x4 seems fastest.
+		 *
+		 * Note that PS/WM thread counts depend on the WIZ hashing
+		 * disable bit, which we don't touch here, but it's good
+		 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+		 */
+		wa_add(wal, GEN7_GT_MODE, 0,
+		       _MASKED_FIELD(GEN6_WIZ_HASHING_MASK,
+				     GEN6_WIZ_HASHING_16x4),
+		       GEN6_WIZ_HASHING_16x4);
+	}
+
 	if (IS_GEN_RANGE(i915, 6, 7))
 		/*
 		 * We need to disable the AsyncFlip performance optimisations in
@@ -1991,6 +1972,39 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
 			     GFX_MODE,
 			     GFX_TLB_INVALIDATE_EXPLICIT);
 
+		/* WaDisableHiZPlanesWhenMSAAEnabled:snb */
+		wa_masked_en(wal,
+			     _3D_CHICKEN,
+			     _3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB);
+
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     /* WaStripsFansDisableFastClipPerformanceFix:snb */
+			     _3D_CHICKEN3_SF_DISABLE_FASTCLIP_CULL |
+			     /*
+			      * Bspec says:
+			      * "This bit must be set if 3DSTATE_CLIP clip mode is set
+			      * to normal and 3DSTATE_SF number of SF output attributes
+			      * is more than 16."
+			      */
+			     _3D_CHICKEN3_SF_DISABLE_PIPELINED_ATTR_FETCH);
+
+		/*
+		 * BSpec recommends 8x4 when MSAA is used,
+		 * however in practice 16x4 seems fastest.
+		 *
+		 * Note that PS/WM thread counts depend on the WIZ hashing
+		 * disable bit, which we don't touch here, but it's good
+		 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+		 */
+		wa_add(wal,
+		       GEN6_GT_MODE, 0,
+		       _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4),
+		       GEN6_WIZ_HASHING_16x4);
+
+		/* WaDisable_RenderCache_OperationalFlush:snb */
+		wa_masked_dis(wal, CACHE_MODE_0, RC_OP_FLUSH_ENABLE);
+
 		/*
 		 * From the Sandybridge PRM, volume 1 part 3, page 24:
 		 * "If this bit is set, STCunit will have LRA as replacement
@@ -2067,39 +2081,6 @@ void intel_engine_apply_workarounds(struct intel_engine_cs *engine)
 	wa_list_apply(engine->uncore, &engine->wa_list);
 }
 
-static struct i915_vma *
-create_scratch(struct i915_address_space *vm, int count)
-{
-	struct drm_i915_gem_object *obj;
-	struct i915_vma *vma;
-	unsigned int size;
-	int err;
-
-	size = round_up(count * sizeof(u32), PAGE_SIZE);
-	obj = i915_gem_object_create_internal(vm->i915, size);
-	if (IS_ERR(obj))
-		return ERR_CAST(obj);
-
-	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
-
-	vma = i915_vma_instance(obj, vm, NULL);
-	if (IS_ERR(vma)) {
-		err = PTR_ERR(vma);
-		goto err_obj;
-	}
-
-	err = i915_vma_pin(vma, 0, 0,
-			   i915_vma_is_ggtt(vma) ? PIN_GLOBAL : PIN_USER);
-	if (err)
-		goto err_obj;
-
-	return vma;
-
-err_obj:
-	i915_gem_object_put(obj);
-	return ERR_PTR(err);
-}
-
 struct mcr_range {
 	u32 start;
 	u32 end;
@@ -2202,7 +2183,8 @@ static int engine_wa_list_verify(struct intel_context *ce,
 	if (!wal->count)
 		return 0;
 
-	vma = create_scratch(&ce->engine->gt->ggtt->vm, wal->count);
+	vma = __vm_create_scratch_for_read(&ce->engine->gt->ggtt->vm,
+					   wal->count * sizeof(u32));
 	if (IS_ERR(vma))
 		return PTR_ERR(vma);
 
diff --git a/drivers/gpu/drm/i915/gt/mock_engine.c b/drivers/gpu/drm/i915/gt/mock_engine.c
index 2f830017c51d..4b4f03b70df7 100644
--- a/drivers/gpu/drm/i915/gt/mock_engine.c
+++ b/drivers/gpu/drm/i915/gt/mock_engine.c
@@ -245,17 +245,6 @@ static void mock_reset_rewind(struct intel_engine_cs *engine, bool stalled)
 	GEM_BUG_ON(stalled);
 }
 
-static void mark_eio(struct i915_request *rq)
-{
-	if (i915_request_completed(rq))
-		return;
-
-	GEM_BUG_ON(i915_request_signaled(rq));
-
-	i915_request_set_error_once(rq, -EIO);
-	i915_request_mark_complete(rq);
-}
-
 static void mock_reset_cancel(struct intel_engine_cs *engine)
 {
 	struct mock_engine *mock =
@@ -269,12 +258,12 @@ static void mock_reset_cancel(struct intel_engine_cs *engine)
 
 	/* Mark all submitted requests as skipped. */
 	list_for_each_entry(rq, &engine->active.requests, sched.link)
-		mark_eio(rq);
+		i915_request_mark_eio(rq);
 	intel_engine_signal_breadcrumbs(engine);
 
 	/* Cancel and submit all pending requests. */
 	list_for_each_entry(rq, &mock->hw_queue, mock.link) {
-		mark_eio(rq);
+		i915_request_mark_eio(rq);
 		__i915_request_submit(rq);
 	}
 	INIT_LIST_HEAD(&mock->hw_queue);
diff --git a/drivers/gpu/drm/i915/gt/selftest_context.c b/drivers/gpu/drm/i915/gt/selftest_context.c
index 1f4020e906a8..db738d400168 100644
--- a/drivers/gpu/drm/i915/gt/selftest_context.c
+++ b/drivers/gpu/drm/i915/gt/selftest_context.c
@@ -25,7 +25,7 @@ static int request_sync(struct i915_request *rq)
 	/* Opencode i915_request_add() so we can keep the timeline locked. */
 	__i915_request_commit(rq);
 	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
-	__i915_request_queue(rq, NULL);
+	__i915_request_queue_bh(rq);
 
 	timeout = i915_request_wait(rq, 0, HZ / 10);
 	if (timeout < 0)
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_cs.c b/drivers/gpu/drm/i915/gt/selftest_engine_cs.c
index 729c3c7b11e2..439c8984f5fa 100644
--- a/drivers/gpu/drm/i915/gt/selftest_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_cs.c
@@ -6,6 +6,7 @@
 
 #include <linux/sort.h>
 
+#include "intel_gpu_commands.h"
 #include "intel_gt_pm.h"
 #include "intel_rps.h"
 
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c
index b88aa35ad75b..223ab88f7e57 100644
--- a/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c
@@ -197,6 +197,7 @@ static int cmp_u32(const void *_a, const void *_b)
 
 static int __live_heartbeat_fast(struct intel_engine_cs *engine)
 {
+	const unsigned int error_threshold = max(20000u, jiffies_to_usecs(6));
 	struct intel_context *ce;
 	struct i915_request *rq;
 	ktime_t t0, t1;
@@ -254,12 +255,18 @@ static int __live_heartbeat_fast(struct intel_engine_cs *engine)
 		times[0],
 		times[ARRAY_SIZE(times) - 1]);
 
-	/* Min work delay is 2 * 2 (worst), +1 for scheduling, +1 for slack */
-	if (times[ARRAY_SIZE(times) / 2] > jiffies_to_usecs(6)) {
+	/*
+	 * Ideally, the upper bound on min work delay would be something like
+	 * 2 * 2 (worst), +1 for scheduling, +1 for slack. In practice, we
+	 * are, even with system_wq_highpri, at the mercy of the CPU scheduler
+	 * and may be stuck behind some slow work for many millisecond. Such
+	 * as our very own display workers.
+	 */
+	if (times[ARRAY_SIZE(times) / 2] > error_threshold) {
 		pr_err("%s: Heartbeat delay was %uus, expected less than %dus\n",
 		       engine->name,
 		       times[ARRAY_SIZE(times) / 2],
-		       jiffies_to_usecs(6));
+		       error_threshold);
 		err = -EINVAL;
 	}
 
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_pm.c b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
index b08fc5390e8a..c3d965279fc3 100644
--- a/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
@@ -4,13 +4,215 @@
  * Copyright © 2018 Intel Corporation
  */
 
+#include <linux/sort.h>
+
 #include "i915_selftest.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt_clock_utils.h"
 #include "selftest_engine.h"
 #include "selftest_engine_heartbeat.h"
 #include "selftests/igt_atomic.h"
 #include "selftests/igt_flush_test.h"
 #include "selftests/igt_spinner.h"
 
+#define COUNT 5
+
+static int cmp_u64(const void *A, const void *B)
+{
+	const u64 *a = A, *b = B;
+
+	return *a - *b;
+}
+
+static u64 trifilter(u64 *a)
+{
+	sort(a, COUNT, sizeof(*a), cmp_u64, NULL);
+	return (a[1] + 2 * a[2] + a[3]) >> 2;
+}
+
+static u32 *emit_wait(u32 *cs, u32 offset, int op, u32 value)
+{
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		op;
+	*cs++ = value;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *emit_store(u32 *cs, u32 offset, u32 value)
+{
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = value;
+
+	return cs;
+}
+
+static u32 *emit_srm(u32 *cs, i915_reg_t reg, u32 offset)
+{
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(reg);
+	*cs++ = offset;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static void write_semaphore(u32 *x, u32 value)
+{
+	WRITE_ONCE(*x, value);
+	wmb();
+}
+
+static int __measure_timestamps(struct intel_context *ce,
+				u64 *dt, u64 *d_ring, u64 *d_ctx)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	u32 *sema = memset32(engine->status_page.addr + 1000, 0, 5);
+	u32 offset = i915_ggtt_offset(engine->status_page.vma);
+	struct i915_request *rq;
+	u32 *cs;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 28);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	/* Signal & wait for start */
+	cs = emit_store(cs, offset + 4008, 1);
+	cs = emit_wait(cs, offset + 4008, MI_SEMAPHORE_SAD_NEQ_SDD, 1);
+
+	cs = emit_srm(cs, RING_TIMESTAMP(engine->mmio_base), offset + 4000);
+	cs = emit_srm(cs, RING_CTX_TIMESTAMP(engine->mmio_base), offset + 4004);
+
+	/* Busy wait */
+	cs = emit_wait(cs, offset + 4008, MI_SEMAPHORE_SAD_EQ_SDD, 1);
+
+	cs = emit_srm(cs, RING_TIMESTAMP(engine->mmio_base), offset + 4016);
+	cs = emit_srm(cs, RING_CTX_TIMESTAMP(engine->mmio_base), offset + 4012);
+
+	intel_ring_advance(rq, cs);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	intel_engine_flush_submission(engine);
+
+	/* Wait for the request to start executing, that then waits for us */
+	while (READ_ONCE(sema[2]) == 0)
+		cpu_relax();
+
+	/* Run the request for a 100us, sampling timestamps before/after */
+	preempt_disable();
+	*dt = local_clock();
+	write_semaphore(&sema[2], 0);
+	udelay(100);
+	*dt = local_clock() - *dt;
+	write_semaphore(&sema[2], 1);
+	preempt_enable();
+
+	if (i915_request_wait(rq, 0, HZ / 2) < 0) {
+		i915_request_put(rq);
+		return -ETIME;
+	}
+	i915_request_put(rq);
+
+	pr_debug("%s CTX_TIMESTAMP: [%x, %x], RING_TIMESTAMP: [%x, %x]\n",
+		 engine->name, sema[1], sema[3], sema[0], sema[4]);
+
+	*d_ctx = sema[3] - sema[1];
+	*d_ring = sema[4] - sema[0];
+	return 0;
+}
+
+static int __live_engine_timestamps(struct intel_engine_cs *engine)
+{
+	u64 s_ring[COUNT], s_ctx[COUNT], st[COUNT], d_ring, d_ctx, dt;
+	struct intel_context *ce;
+	int i, err = 0;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	for (i = 0; i < COUNT; i++) {
+		err = __measure_timestamps(ce, &st[i], &s_ring[i], &s_ctx[i]);
+		if (err)
+			break;
+	}
+	intel_context_put(ce);
+	if (err)
+		return err;
+
+	dt = trifilter(st);
+	d_ring = trifilter(s_ring);
+	d_ctx = trifilter(s_ctx);
+
+	pr_info("%s elapsed:%lldns, CTX_TIMESTAMP:%lldns, RING_TIMESTAMP:%lldns\n",
+		engine->name, dt,
+		intel_gt_clock_interval_to_ns(engine->gt, d_ctx),
+		intel_gt_clock_interval_to_ns(engine->gt, d_ring));
+
+	d_ring = intel_gt_clock_interval_to_ns(engine->gt, d_ring);
+	if (3 * dt > 4 * d_ring || 4 * dt < 3 * d_ring) {
+		pr_err("%s Mismatch between ring timestamp and walltime!\n",
+		       engine->name);
+		return -EINVAL;
+	}
+
+	d_ring = trifilter(s_ring);
+	d_ctx = trifilter(s_ctx);
+
+	d_ctx *= engine->gt->clock_frequency;
+	if (IS_ICELAKE(engine->i915))
+		d_ring *= 12500000; /* Fixed 80ns for icl ctx timestamp? */
+	else
+		d_ring *= engine->gt->clock_frequency;
+
+	if (3 * d_ctx > 4 * d_ring || 4 * d_ctx < 3 * d_ring) {
+		pr_err("%s Mismatch between ring and context timestamps!\n",
+		       engine->name);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int live_engine_timestamps(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Check that CS_TIMESTAMP / CTX_TIMESTAMP are in sync, i.e. share
+	 * the same CS clock.
+	 */
+
+	if (INTEL_GEN(gt->i915) < 8)
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		int err;
+
+		st_engine_heartbeat_disable(engine);
+		err = __live_engine_timestamps(engine);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
 static int live_engine_busy_stats(void *arg)
 {
 	struct intel_gt *gt = arg;
@@ -177,6 +379,7 @@ static int live_engine_pm(void *arg)
 int live_engine_pm_selftests(struct intel_gt *gt)
 {
 	static const struct i915_subtest tests[] = {
+		SUBTEST(live_engine_timestamps),
 		SUBTEST(live_engine_busy_stats),
 		SUBTEST(live_engine_pm),
 	};
diff --git a/drivers/gpu/drm/i915/gt/selftest_execlists.c b/drivers/gpu/drm/i915/gt/selftest_execlists.c
new file mode 100644
index 000000000000..264b5ebdb021
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_execlists.c
@@ -0,0 +1,4741 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_reset.h"
+#include "gt/selftest_engine_heartbeat.h"
+
+#include "i915_selftest.h"
+#include "selftests/i915_random.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_live_test.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/lib_sw_fence.h"
+
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+
+#define CS_GPR(engine, n) ((engine)->mmio_base + 0x600 + (n) * 4)
+#define NUM_GPR 16
+#define NUM_GPR_DW (NUM_GPR * 2) /* each GPR is 2 dwords */
+
+static bool is_active(struct i915_request *rq)
+{
+	if (i915_request_is_active(rq))
+		return true;
+
+	if (i915_request_on_hold(rq))
+		return true;
+
+	if (i915_request_has_initial_breadcrumb(rq) && i915_request_started(rq))
+		return true;
+
+	return false;
+}
+
+static int wait_for_submit(struct intel_engine_cs *engine,
+			   struct i915_request *rq,
+			   unsigned long timeout)
+{
+	/* Ignore our own attempts to suppress excess tasklets */
+	tasklet_hi_schedule(&engine->execlists.tasklet);
+
+	timeout += jiffies;
+	do {
+		bool done = time_after(jiffies, timeout);
+
+		if (i915_request_completed(rq)) /* that was quick! */
+			return 0;
+
+		/* Wait until the HW has acknowleged the submission (or err) */
+		intel_engine_flush_submission(engine);
+		if (!READ_ONCE(engine->execlists.pending[0]) && is_active(rq))
+			return 0;
+
+		if (done)
+			return -ETIME;
+
+		cond_resched();
+	} while (1);
+}
+
+static int wait_for_reset(struct intel_engine_cs *engine,
+			  struct i915_request *rq,
+			  unsigned long timeout)
+{
+	timeout += jiffies;
+
+	do {
+		cond_resched();
+		intel_engine_flush_submission(engine);
+
+		if (READ_ONCE(engine->execlists.pending[0]))
+			continue;
+
+		if (i915_request_completed(rq))
+			break;
+
+		if (READ_ONCE(rq->fence.error))
+			break;
+	} while (time_before(jiffies, timeout));
+
+	flush_scheduled_work();
+
+	if (rq->fence.error != -EIO) {
+		pr_err("%s: hanging request %llx:%lld not reset\n",
+		       engine->name,
+		       rq->fence.context,
+		       rq->fence.seqno);
+		return -EINVAL;
+	}
+
+	/* Give the request a jiffie to complete after flushing the worker */
+	if (i915_request_wait(rq, 0,
+			      max(0l, (long)(timeout - jiffies)) + 1) < 0) {
+		pr_err("%s: hanging request %llx:%lld did not complete\n",
+		       engine->name,
+		       rq->fence.context,
+		       rq->fence.seqno);
+		return -ETIME;
+	}
+
+	return 0;
+}
+
+static int live_sanitycheck(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	if (!HAS_LOGICAL_RING_CONTEXTS(gt->i915))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_ctx;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			GEM_TRACE("spinner failed to start\n");
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto out_ctx;
+		}
+
+		igt_spinner_end(&spin);
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			goto out_ctx;
+		}
+
+out_ctx:
+		intel_context_put(ce);
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_unlite_restore(struct intel_gt *gt, int prio)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = -ENOMEM;
+
+	/*
+	 * Check that we can correctly context switch between 2 instances
+	 * on the same engine from the same parent context.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return err;
+
+	err = 0;
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce[2] = {};
+		struct i915_request *rq[2];
+		struct igt_live_test t;
+		int n;
+
+		if (prio && !intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+		st_engine_heartbeat_disable(engine);
+
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			struct intel_context *tmp;
+
+			tmp = intel_context_create(engine);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				goto err_ce;
+			}
+
+			err = intel_context_pin(tmp);
+			if (err) {
+				intel_context_put(tmp);
+				goto err_ce;
+			}
+
+			/*
+			 * Setup the pair of contexts such that if we
+			 * lite-restore using the RING_TAIL from ce[1] it
+			 * will execute garbage from ce[0]->ring.
+			 */
+			memset(tmp->ring->vaddr,
+			       POISON_INUSE, /* IPEHR: 0x5a5a5a5a [hung!] */
+			       tmp->ring->vma->size);
+
+			ce[n] = tmp;
+		}
+		GEM_BUG_ON(!ce[1]->ring->size);
+		intel_ring_reset(ce[1]->ring, ce[1]->ring->size / 2);
+		lrc_update_regs(ce[1], engine, ce[1]->ring->head);
+
+		rq[0] = igt_spinner_create_request(&spin, ce[0], MI_ARB_CHECK);
+		if (IS_ERR(rq[0])) {
+			err = PTR_ERR(rq[0]);
+			goto err_ce;
+		}
+
+		i915_request_get(rq[0]);
+		i915_request_add(rq[0]);
+		GEM_BUG_ON(rq[0]->postfix > ce[1]->ring->emit);
+
+		if (!igt_wait_for_spinner(&spin, rq[0])) {
+			i915_request_put(rq[0]);
+			goto err_ce;
+		}
+
+		rq[1] = i915_request_create(ce[1]);
+		if (IS_ERR(rq[1])) {
+			err = PTR_ERR(rq[1]);
+			i915_request_put(rq[0]);
+			goto err_ce;
+		}
+
+		if (!prio) {
+			/*
+			 * Ensure we do the switch to ce[1] on completion.
+			 *
+			 * rq[0] is already submitted, so this should reduce
+			 * to a no-op (a wait on a request on the same engine
+			 * uses the submit fence, not the completion fence),
+			 * but it will install a dependency on rq[1] for rq[0]
+			 * that will prevent the pair being reordered by
+			 * timeslicing.
+			 */
+			i915_request_await_dma_fence(rq[1], &rq[0]->fence);
+		}
+
+		i915_request_get(rq[1]);
+		i915_request_add(rq[1]);
+		GEM_BUG_ON(rq[1]->postfix <= rq[0]->postfix);
+		i915_request_put(rq[0]);
+
+		if (prio) {
+			struct i915_sched_attr attr = {
+				.priority = prio,
+			};
+
+			/* Alternatively preempt the spinner with ce[1] */
+			engine->schedule(rq[1], &attr);
+		}
+
+		/* And switch back to ce[0] for good measure */
+		rq[0] = i915_request_create(ce[0]);
+		if (IS_ERR(rq[0])) {
+			err = PTR_ERR(rq[0]);
+			i915_request_put(rq[1]);
+			goto err_ce;
+		}
+
+		i915_request_await_dma_fence(rq[0], &rq[1]->fence);
+		i915_request_get(rq[0]);
+		i915_request_add(rq[0]);
+		GEM_BUG_ON(rq[0]->postfix > rq[1]->postfix);
+		i915_request_put(rq[1]);
+		i915_request_put(rq[0]);
+
+err_ce:
+		intel_engine_flush_submission(engine);
+		igt_spinner_end(&spin);
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			if (IS_ERR_OR_NULL(ce[n]))
+				break;
+
+			intel_context_unpin(ce[n]);
+			intel_context_put(ce[n]);
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_unlite_switch(void *arg)
+{
+	return live_unlite_restore(arg, 0);
+}
+
+static int live_unlite_preempt(void *arg)
+{
+	return live_unlite_restore(arg, I915_USER_PRIORITY(I915_PRIORITY_MAX));
+}
+
+static int live_unlite_ring(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct igt_spinner spin;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Setup a preemption event that will cause almost the entire ring
+	 * to be unwound, potentially fooling our intel_ring_direction()
+	 * into emitting a forward lite-restore instead of the rollback.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce[2] = {};
+		struct i915_request *rq;
+		struct igt_live_test t;
+		int n;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+		st_engine_heartbeat_disable(engine);
+
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			struct intel_context *tmp;
+
+			tmp = intel_context_create(engine);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				goto err_ce;
+			}
+
+			err = intel_context_pin(tmp);
+			if (err) {
+				intel_context_put(tmp);
+				goto err_ce;
+			}
+
+			memset32(tmp->ring->vaddr,
+				 0xdeadbeef, /* trigger a hang if executed */
+				 tmp->ring->vma->size / sizeof(u32));
+
+			ce[n] = tmp;
+		}
+
+		/* Create max prio spinner, followed by N low prio nops */
+		rq = igt_spinner_create_request(&spin, ce[0], MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ce;
+		}
+
+		i915_request_get(rq);
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			intel_gt_set_wedged(gt);
+			i915_request_put(rq);
+			err = -ETIME;
+			goto err_ce;
+		}
+
+		/* Fill the ring, until we will cause a wrap */
+		n = 0;
+		while (intel_ring_direction(ce[0]->ring,
+					    rq->wa_tail,
+					    ce[0]->ring->tail) <= 0) {
+			struct i915_request *tmp;
+
+			tmp = intel_context_create_request(ce[0]);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				i915_request_put(rq);
+				goto err_ce;
+			}
+
+			i915_request_add(tmp);
+			intel_engine_flush_submission(engine);
+			n++;
+		}
+		intel_engine_flush_submission(engine);
+		pr_debug("%s: Filled ring with %d nop tails {size:%x, tail:%x, emit:%x, rq.tail:%x}\n",
+			 engine->name, n,
+			 ce[0]->ring->size,
+			 ce[0]->ring->tail,
+			 ce[0]->ring->emit,
+			 rq->tail);
+		GEM_BUG_ON(intel_ring_direction(ce[0]->ring,
+						rq->tail,
+						ce[0]->ring->tail) <= 0);
+		i915_request_put(rq);
+
+		/* Create a second ring to preempt the first ring after rq[0] */
+		rq = intel_context_create_request(ce[1]);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ce;
+		}
+
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		err = wait_for_submit(engine, rq, HZ / 2);
+		i915_request_put(rq);
+		if (err) {
+			pr_err("%s: preemption request was not submitted\n",
+			       engine->name);
+			err = -ETIME;
+		}
+
+		pr_debug("%s: ring[0]:{ tail:%x, emit:%x }, ring[1]:{ tail:%x, emit:%x }\n",
+			 engine->name,
+			 ce[0]->ring->tail, ce[0]->ring->emit,
+			 ce[1]->ring->tail, ce[1]->ring->emit);
+
+err_ce:
+		intel_engine_flush_submission(engine);
+		igt_spinner_end(&spin);
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			if (IS_ERR_OR_NULL(ce[n]))
+				break;
+
+			intel_context_unpin(ce[n]);
+			intel_context_put(ce[n]);
+		}
+		st_engine_heartbeat_enable(engine);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_pin_rewind(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * We have to be careful not to trust intel_ring too much, for example
+	 * ring->head is updated upon retire which is out of sync with pinning
+	 * the context. Thus we cannot use ring->head to set CTX_RING_HEAD,
+	 * or else we risk writing an older, stale value.
+	 *
+	 * To simulate this, let's apply a bit of deliberate sabotague.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+		struct intel_ring *ring;
+		struct igt_live_test t;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		err = intel_context_pin(ce);
+		if (err) {
+			intel_context_put(ce);
+			break;
+		}
+
+		/* Keep the context awake while we play games */
+		err = i915_active_acquire(&ce->active);
+		if (err) {
+			intel_context_unpin(ce);
+			intel_context_put(ce);
+			break;
+		}
+		ring = ce->ring;
+
+		/* Poison the ring, and offset the next request from HEAD */
+		memset32(ring->vaddr, STACK_MAGIC, ring->size / sizeof(u32));
+		ring->emit = ring->size / 2;
+		ring->tail = ring->emit;
+		GEM_BUG_ON(ring->head);
+
+		intel_context_unpin(ce);
+
+		/* Submit a simple nop request */
+		GEM_BUG_ON(intel_context_is_pinned(ce));
+		rq = intel_context_create_request(ce);
+		i915_active_release(&ce->active); /* e.g. async retire */
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+		GEM_BUG_ON(!rq->head);
+		i915_request_add(rq);
+
+		/* Expect not to hang! */
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			break;
+		}
+	}
+
+	return err;
+}
+
+static int live_hold_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * In order to support offline error capture for fast preempt reset,
+	 * we need to decouple the guilty request and ensure that it and its
+	 * descendents are not executed while the capture is in progress.
+	 */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		st_engine_heartbeat_disable(engine);
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			intel_gt_set_wedged(gt);
+			err = -ETIME;
+			goto out;
+		}
+
+		/* We have our request executing, now remove it and reset */
+
+		local_bh_disable();
+		if (test_and_set_bit(I915_RESET_ENGINE + id,
+				     &gt->reset.flags)) {
+			local_bh_enable();
+			intel_gt_set_wedged(gt);
+			err = -EBUSY;
+			goto out;
+		}
+		tasklet_disable(&engine->execlists.tasklet);
+
+		engine->execlists.tasklet.func(engine->execlists.tasklet.data);
+		GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
+
+		i915_request_get(rq);
+		execlists_hold(engine, rq);
+		GEM_BUG_ON(!i915_request_on_hold(rq));
+
+		__intel_engine_reset_bh(engine, NULL);
+		GEM_BUG_ON(rq->fence.error != -EIO);
+
+		tasklet_enable(&engine->execlists.tasklet);
+		clear_and_wake_up_bit(I915_RESET_ENGINE + id,
+				      &gt->reset.flags);
+		local_bh_enable();
+
+		/* Check that we do not resubmit the held request */
+		if (!i915_request_wait(rq, 0, HZ / 5)) {
+			pr_err("%s: on hold request completed!\n",
+			       engine->name);
+			i915_request_put(rq);
+			err = -EIO;
+			goto out;
+		}
+		GEM_BUG_ON(!i915_request_on_hold(rq));
+
+		/* But is resubmitted on release */
+		execlists_unhold(engine, rq);
+		if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+			pr_err("%s: held request did not complete!\n",
+			       engine->name);
+			intel_gt_set_wedged(gt);
+			err = -ETIME;
+		}
+		i915_request_put(rq);
+
+out:
+		st_engine_heartbeat_enable(engine);
+		intel_context_put(ce);
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static const char *error_repr(int err)
+{
+	return err ? "bad" : "good";
+}
+
+static int live_error_interrupt(void *arg)
+{
+	static const struct error_phase {
+		enum { GOOD = 0, BAD = -EIO } error[2];
+	} phases[] = {
+		{ { BAD,  GOOD } },
+		{ { BAD,  BAD  } },
+		{ { BAD,  GOOD } },
+		{ { GOOD, GOOD } }, /* sentinel */
+	};
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * We hook up the CS_MASTER_ERROR_INTERRUPT to have forewarning
+	 * of invalid commands in user batches that will cause a GPU hang.
+	 * This is a faster mechanism than using hangcheck/heartbeats, but
+	 * only detects problems the HW knows about -- it will not warn when
+	 * we kill the HW!
+	 *
+	 * To verify our detection and reset, we throw some invalid commands
+	 * at the HW and wait for the interrupt.
+	 */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		const struct error_phase *p;
+		int err = 0;
+
+		st_engine_heartbeat_disable(engine);
+
+		for (p = phases; p->error[0] != GOOD; p++) {
+			struct i915_request *client[ARRAY_SIZE(phases->error)];
+			u32 *cs;
+			int i;
+
+			memset(client, 0, sizeof(*client));
+			for (i = 0; i < ARRAY_SIZE(client); i++) {
+				struct intel_context *ce;
+				struct i915_request *rq;
+
+				ce = intel_context_create(engine);
+				if (IS_ERR(ce)) {
+					err = PTR_ERR(ce);
+					goto out;
+				}
+
+				rq = intel_context_create_request(ce);
+				intel_context_put(ce);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					goto out;
+				}
+
+				if (rq->engine->emit_init_breadcrumb) {
+					err = rq->engine->emit_init_breadcrumb(rq);
+					if (err) {
+						i915_request_add(rq);
+						goto out;
+					}
+				}
+
+				cs = intel_ring_begin(rq, 2);
+				if (IS_ERR(cs)) {
+					i915_request_add(rq);
+					err = PTR_ERR(cs);
+					goto out;
+				}
+
+				if (p->error[i]) {
+					*cs++ = 0xdeadbeef;
+					*cs++ = 0xdeadbeef;
+				} else {
+					*cs++ = MI_NOOP;
+					*cs++ = MI_NOOP;
+				}
+
+				client[i] = i915_request_get(rq);
+				i915_request_add(rq);
+			}
+
+			err = wait_for_submit(engine, client[0], HZ / 2);
+			if (err) {
+				pr_err("%s: first request did not start within time!\n",
+				       engine->name);
+				err = -ETIME;
+				goto out;
+			}
+
+			for (i = 0; i < ARRAY_SIZE(client); i++) {
+				if (i915_request_wait(client[i], 0, HZ / 5) < 0)
+					pr_debug("%s: %s request incomplete!\n",
+						 engine->name,
+						 error_repr(p->error[i]));
+
+				if (!i915_request_started(client[i])) {
+					pr_err("%s: %s request not started!\n",
+					       engine->name,
+					       error_repr(p->error[i]));
+					err = -ETIME;
+					goto out;
+				}
+
+				/* Kick the tasklet to process the error */
+				intel_engine_flush_submission(engine);
+				if (client[i]->fence.error != p->error[i]) {
+					pr_err("%s: %s request (%s) with wrong error code: %d\n",
+					       engine->name,
+					       error_repr(p->error[i]),
+					       i915_request_completed(client[i]) ? "completed" : "running",
+					       client[i]->fence.error);
+					err = -EINVAL;
+					goto out;
+				}
+			}
+
+out:
+			for (i = 0; i < ARRAY_SIZE(client); i++)
+				if (client[i])
+					i915_request_put(client[i]);
+			if (err) {
+				pr_err("%s: failed at phase[%zd] { %d, %d }\n",
+				       engine->name, p - phases,
+				       p->error[0], p->error[1]);
+				break;
+			}
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (err) {
+			intel_gt_set_wedged(gt);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static int
+emit_semaphore_chain(struct i915_request *rq, struct i915_vma *vma, int idx)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 10);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_NEQ_SDD;
+	*cs++ = 0;
+	*cs++ = i915_ggtt_offset(vma) + 4 * idx;
+	*cs++ = 0;
+
+	if (idx > 0) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = i915_ggtt_offset(vma) + 4 * (idx - 1);
+		*cs++ = 0;
+		*cs++ = 1;
+	} else {
+		*cs++ = MI_NOOP;
+		*cs++ = MI_NOOP;
+		*cs++ = MI_NOOP;
+		*cs++ = MI_NOOP;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	intel_ring_advance(rq, cs);
+	return 0;
+}
+
+static struct i915_request *
+semaphore_queue(struct intel_engine_cs *engine, struct i915_vma *vma, int idx)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		goto out_ce;
+
+	err = 0;
+	if (rq->engine->emit_init_breadcrumb)
+		err = rq->engine->emit_init_breadcrumb(rq);
+	if (err == 0)
+		err = emit_semaphore_chain(rq, vma, idx);
+	if (err == 0)
+		i915_request_get(rq);
+	i915_request_add(rq);
+	if (err)
+		rq = ERR_PTR(err);
+
+out_ce:
+	intel_context_put(ce);
+	return rq;
+}
+
+static int
+release_queue(struct intel_engine_cs *engine,
+	      struct i915_vma *vma,
+	      int idx, int prio)
+{
+	struct i915_sched_attr attr = {
+		.priority = prio,
+	};
+	struct i915_request *rq;
+	u32 *cs;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = i915_ggtt_offset(vma) + 4 * (idx - 1);
+	*cs++ = 0;
+	*cs++ = 1;
+
+	intel_ring_advance(rq, cs);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	local_bh_disable();
+	engine->schedule(rq, &attr);
+	local_bh_enable(); /* kick tasklet */
+
+	i915_request_put(rq);
+
+	return 0;
+}
+
+static int
+slice_semaphore_queue(struct intel_engine_cs *outer,
+		      struct i915_vma *vma,
+		      int count)
+{
+	struct intel_engine_cs *engine;
+	struct i915_request *head;
+	enum intel_engine_id id;
+	int err, i, n = 0;
+
+	head = semaphore_queue(outer, vma, n++);
+	if (IS_ERR(head))
+		return PTR_ERR(head);
+
+	for_each_engine(engine, outer->gt, id) {
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		for (i = 0; i < count; i++) {
+			struct i915_request *rq;
+
+			rq = semaphore_queue(engine, vma, n++);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto out;
+			}
+
+			i915_request_put(rq);
+		}
+	}
+
+	err = release_queue(outer, vma, n, I915_PRIORITY_BARRIER);
+	if (err)
+		goto out;
+
+	if (i915_request_wait(head, 0,
+			      2 * outer->gt->info.num_engines * (count + 2) * (count + 3)) < 0) {
+		pr_err("%s: Failed to slice along semaphore chain of length (%d, %d)!\n",
+		       outer->name, count, n);
+		GEM_TRACE_DUMP();
+		intel_gt_set_wedged(outer->gt);
+		err = -EIO;
+	}
+
+out:
+	i915_request_put(head);
+	return err;
+}
+
+static int live_timeslice_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct drm_i915_gem_object *obj;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct i915_vma *vma;
+	void *vaddr;
+	int err = 0;
+
+	/*
+	 * If a request takes too long, we would like to give other users
+	 * a fair go on the GPU. In particular, users may create batches
+	 * that wait upon external input, where that input may even be
+	 * supplied by another GPU job. To avoid blocking forever, we
+	 * need to preempt the current task and replace it with another
+	 * ready task.
+	 */
+	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+		return 0;
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_map;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_pin;
+
+	for_each_engine(engine, gt, id) {
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		memset(vaddr, 0, PAGE_SIZE);
+
+		st_engine_heartbeat_disable(engine);
+		err = slice_semaphore_queue(engine, vma, 5);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			goto err_pin;
+
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			goto err_pin;
+		}
+	}
+
+err_pin:
+	i915_vma_unpin(vma);
+err_map:
+	i915_gem_object_unpin_map(obj);
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static struct i915_request *
+create_rewinder(struct intel_context *ce,
+		struct i915_request *wait,
+		void *slot, int idx)
+{
+	const u32 offset =
+		i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(slot);
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	if (wait) {
+		err = i915_request_await_dma_fence(rq, &wait->fence);
+		if (err)
+			goto err;
+	}
+
+	cs = intel_ring_begin(rq, 14);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_GTE_SDD;
+	*cs++ = idx;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(rq->engine->mmio_base));
+	*cs++ = offset + idx * sizeof(u32);
+	*cs++ = 0;
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = idx + 1;
+
+	intel_ring_advance(rq, cs);
+
+	rq->sched.attr.priority = I915_PRIORITY_MASK;
+	err = 0;
+err:
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (err) {
+		i915_request_put(rq);
+		return ERR_PTR(err);
+	}
+
+	return rq;
+}
+
+static int live_timeslice_rewind(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * The usual presumption on timeslice expiration is that we replace
+	 * the active context with another. However, given a chain of
+	 * dependencies we may end up with replacing the context with itself,
+	 * but only a few of those requests, forcing us to rewind the
+	 * RING_TAIL of the original request.
+	 */
+	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		enum { A1, A2, B1 };
+		enum { X = 1, Z, Y };
+		struct i915_request *rq[3] = {};
+		struct intel_context *ce;
+		unsigned long timeslice;
+		int i, err = 0;
+		u32 *slot;
+
+		if (!intel_engine_has_timeslices(engine))
+			continue;
+
+		/*
+		 * A:rq1 -- semaphore wait, timestamp X
+		 * A:rq2 -- write timestamp Y
+		 *
+		 * B:rq1 [await A:rq1] -- write timestamp Z
+		 *
+		 * Force timeslice, release semaphore.
+		 *
+		 * Expect execution/evaluation order XZY
+		 */
+
+		st_engine_heartbeat_disable(engine);
+		timeslice = xchg(&engine->props.timeslice_duration_ms, 1);
+
+		slot = memset32(engine->status_page.addr + 1000, 0, 4);
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto err;
+		}
+
+		rq[A1] = create_rewinder(ce, NULL, slot, X);
+		if (IS_ERR(rq[A1])) {
+			intel_context_put(ce);
+			goto err;
+		}
+
+		rq[A2] = create_rewinder(ce, NULL, slot, Y);
+		intel_context_put(ce);
+		if (IS_ERR(rq[A2]))
+			goto err;
+
+		err = wait_for_submit(engine, rq[A2], HZ / 2);
+		if (err) {
+			pr_err("%s: failed to submit first context\n",
+			       engine->name);
+			goto err;
+		}
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto err;
+		}
+
+		rq[B1] = create_rewinder(ce, rq[A1], slot, Z);
+		intel_context_put(ce);
+		if (IS_ERR(rq[2]))
+			goto err;
+
+		err = wait_for_submit(engine, rq[B1], HZ / 2);
+		if (err) {
+			pr_err("%s: failed to submit second context\n",
+			       engine->name);
+			goto err;
+		}
+
+		/* ELSP[] = { { A:rq1, A:rq2 }, { B:rq1 } } */
+		ENGINE_TRACE(engine, "forcing tasklet for rewind\n");
+		while (i915_request_is_active(rq[A2])) { /* semaphore yield! */
+			/* Wait for the timeslice to kick in */
+			del_timer(&engine->execlists.timer);
+			tasklet_hi_schedule(&engine->execlists.tasklet);
+			intel_engine_flush_submission(engine);
+		}
+		/* -> ELSP[] = { { A:rq1 }, { B:rq1 } } */
+		GEM_BUG_ON(!i915_request_is_active(rq[A1]));
+		GEM_BUG_ON(!i915_request_is_active(rq[B1]));
+		GEM_BUG_ON(i915_request_is_active(rq[A2]));
+
+		/* Release the hounds! */
+		slot[0] = 1;
+		wmb(); /* "pairs" with GPU; paranoid kick of internal CPU$ */
+
+		for (i = 1; i <= 3; i++) {
+			unsigned long timeout = jiffies + HZ / 2;
+
+			while (!READ_ONCE(slot[i]) &&
+			       time_before(jiffies, timeout))
+				;
+
+			if (!time_before(jiffies, timeout)) {
+				pr_err("%s: rq[%d] timed out\n",
+				       engine->name, i - 1);
+				err = -ETIME;
+				goto err;
+			}
+
+			pr_debug("%s: slot[%d]:%x\n", engine->name, i, slot[i]);
+		}
+
+		/* XZY: XZ < XY */
+		if (slot[Z] - slot[X] >= slot[Y] - slot[X]) {
+			pr_err("%s: timeslicing did not run context B [%u] before A [%u]!\n",
+			       engine->name,
+			       slot[Z] - slot[X],
+			       slot[Y] - slot[X]);
+			err = -EINVAL;
+		}
+
+err:
+		memset32(&slot[0], -1, 4);
+		wmb();
+
+		engine->props.timeslice_duration_ms = timeslice;
+		st_engine_heartbeat_enable(engine);
+		for (i = 0; i < 3; i++)
+			i915_request_put(rq[i]);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static struct i915_request *nop_request(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return rq;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	return rq;
+}
+
+static long slice_timeout(struct intel_engine_cs *engine)
+{
+	long timeout;
+
+	/* Enough time for a timeslice to kick in, and kick out */
+	timeout = 2 * msecs_to_jiffies_timeout(timeslice(engine));
+
+	/* Enough time for the nop request to complete */
+	timeout += HZ / 5;
+
+	return timeout + 1;
+}
+
+static int live_timeslice_queue(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct drm_i915_gem_object *obj;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct i915_vma *vma;
+	void *vaddr;
+	int err = 0;
+
+	/*
+	 * Make sure that even if ELSP[0] and ELSP[1] are filled with
+	 * timeslicing between them disabled, we *do* enable timeslicing
+	 * if the queue demands it. (Normally, we do not submit if
+	 * ELSP[1] is already occupied, so must rely on timeslicing to
+	 * eject ELSP[0] in favour of the queue.)
+	 */
+	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+		return 0;
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_map;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_pin;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_sched_attr attr = {
+			.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX),
+		};
+		struct i915_request *rq, *nop;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+		memset(vaddr, 0, PAGE_SIZE);
+
+		/* ELSP[0]: semaphore wait */
+		rq = semaphore_queue(engine, vma, 0);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_heartbeat;
+		}
+		engine->schedule(rq, &attr);
+		err = wait_for_submit(engine, rq, HZ / 2);
+		if (err) {
+			pr_err("%s: Timed out trying to submit semaphores\n",
+			       engine->name);
+			goto err_rq;
+		}
+
+		/* ELSP[1]: nop request */
+		nop = nop_request(engine);
+		if (IS_ERR(nop)) {
+			err = PTR_ERR(nop);
+			goto err_rq;
+		}
+		err = wait_for_submit(engine, nop, HZ / 2);
+		i915_request_put(nop);
+		if (err) {
+			pr_err("%s: Timed out trying to submit nop\n",
+			       engine->name);
+			goto err_rq;
+		}
+
+		GEM_BUG_ON(i915_request_completed(rq));
+		GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
+
+		/* Queue: semaphore signal, matching priority as semaphore */
+		err = release_queue(engine, vma, 1, effective_prio(rq));
+		if (err)
+			goto err_rq;
+
+		/* Wait until we ack the release_queue and start timeslicing */
+		do {
+			cond_resched();
+			intel_engine_flush_submission(engine);
+		} while (READ_ONCE(engine->execlists.pending[0]));
+
+		/* Timeslice every jiffy, so within 2 we should signal */
+		if (i915_request_wait(rq, 0, slice_timeout(engine)) < 0) {
+			struct drm_printer p =
+				drm_info_printer(gt->i915->drm.dev);
+
+			pr_err("%s: Failed to timeslice into queue\n",
+			       engine->name);
+			intel_engine_dump(engine, &p,
+					  "%s\n", engine->name);
+
+			memset(vaddr, 0xff, PAGE_SIZE);
+			err = -EIO;
+		}
+err_rq:
+		i915_request_put(rq);
+err_heartbeat:
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+	}
+
+err_pin:
+	i915_vma_unpin(vma);
+err_map:
+	i915_gem_object_unpin_map(obj);
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int live_timeslice_nopreempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * We should not timeslice into a request that is marked with
+	 * I915_REQUEST_NOPREEMPT.
+	 */
+	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+		unsigned long timeslice;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		st_engine_heartbeat_disable(engine);
+		timeslice = xchg(&engine->props.timeslice_duration_ms, 1);
+
+		/* Create an unpreemptible spinner */
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_heartbeat;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			i915_request_put(rq);
+			err = -ETIME;
+			goto out_spin;
+		}
+
+		set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq->fence.flags);
+		i915_request_put(rq);
+
+		/* Followed by a maximum priority barrier (heartbeat) */
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out_spin;
+		}
+
+		rq = intel_context_create_request(ce);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_spin;
+		}
+
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		/*
+		 * Wait until the barrier is in ELSP, and we know timeslicing
+		 * will have been activated.
+		 */
+		if (wait_for_submit(engine, rq, HZ / 2)) {
+			i915_request_put(rq);
+			err = -ETIME;
+			goto out_spin;
+		}
+
+		/*
+		 * Since the ELSP[0] request is unpreemptible, it should not
+		 * allow the maximum priority barrier through. Wait long
+		 * enough to see if it is timesliced in by mistake.
+		 */
+		if (i915_request_wait(rq, 0, slice_timeout(engine)) >= 0) {
+			pr_err("%s: I915_PRIORITY_BARRIER request completed, bypassing no-preempt request\n",
+			       engine->name);
+			err = -EINVAL;
+		}
+		i915_request_put(rq);
+
+out_spin:
+		igt_spinner_end(&spin);
+out_heartbeat:
+		xchg(&engine->props.timeslice_duration_ms, timeslice);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			break;
+		}
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_busywait_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct intel_engine_cs *engine;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+	u32 *map;
+
+	/*
+	 * Verify that even without HAS_LOGICAL_RING_PREEMPTION, we can
+	 * preempt the busywaits used to synchronise between rings.
+	 */
+
+	ctx_hi = kernel_context(gt->i915);
+	if (!ctx_hi)
+		return -ENOMEM;
+	ctx_hi->sched.priority =
+		I915_USER_PRIORITY(I915_CONTEXT_MAX_USER_PRIORITY);
+
+	ctx_lo = kernel_context(gt->i915);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+	ctx_lo->sched.priority =
+		I915_USER_PRIORITY(I915_CONTEXT_MIN_USER_PRIORITY);
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto err_ctx_lo;
+	}
+
+	map = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(map)) {
+		err = PTR_ERR(map);
+		goto err_obj;
+	}
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_map;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_map;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_vma;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *lo, *hi;
+		struct igt_live_test t;
+		u32 *cs;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_vma;
+		}
+
+		/*
+		 * We create two requests. The low priority request
+		 * busywaits on a semaphore (inside the ringbuffer where
+		 * is should be preemptible) and the high priority requests
+		 * uses a MI_STORE_DWORD_IMM to update the semaphore value
+		 * allowing the first request to complete. If preemption
+		 * fails, we hang instead.
+		 */
+
+		lo = igt_request_alloc(ctx_lo, engine);
+		if (IS_ERR(lo)) {
+			err = PTR_ERR(lo);
+			goto err_vma;
+		}
+
+		cs = intel_ring_begin(lo, 8);
+		if (IS_ERR(cs)) {
+			err = PTR_ERR(cs);
+			i915_request_add(lo);
+			goto err_vma;
+		}
+
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = i915_ggtt_offset(vma);
+		*cs++ = 0;
+		*cs++ = 1;
+
+		/* XXX Do we need a flush + invalidate here? */
+
+		*cs++ = MI_SEMAPHORE_WAIT |
+			MI_SEMAPHORE_GLOBAL_GTT |
+			MI_SEMAPHORE_POLL |
+			MI_SEMAPHORE_SAD_EQ_SDD;
+		*cs++ = 0;
+		*cs++ = i915_ggtt_offset(vma);
+		*cs++ = 0;
+
+		intel_ring_advance(lo, cs);
+
+		i915_request_get(lo);
+		i915_request_add(lo);
+
+		if (wait_for(READ_ONCE(*map), 10)) {
+			i915_request_put(lo);
+			err = -ETIMEDOUT;
+			goto err_vma;
+		}
+
+		/* Low priority request should be busywaiting now */
+		if (i915_request_wait(lo, 0, 1) != -ETIME) {
+			i915_request_put(lo);
+			pr_err("%s: Busywaiting request did not!\n",
+			       engine->name);
+			err = -EIO;
+			goto err_vma;
+		}
+
+		hi = igt_request_alloc(ctx_hi, engine);
+		if (IS_ERR(hi)) {
+			err = PTR_ERR(hi);
+			i915_request_put(lo);
+			goto err_vma;
+		}
+
+		cs = intel_ring_begin(hi, 4);
+		if (IS_ERR(cs)) {
+			err = PTR_ERR(cs);
+			i915_request_add(hi);
+			i915_request_put(lo);
+			goto err_vma;
+		}
+
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = i915_ggtt_offset(vma);
+		*cs++ = 0;
+		*cs++ = 0;
+
+		intel_ring_advance(hi, cs);
+		i915_request_add(hi);
+
+		if (i915_request_wait(lo, 0, HZ / 5) < 0) {
+			struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+			pr_err("%s: Failed to preempt semaphore busywait!\n",
+			       engine->name);
+
+			intel_engine_dump(engine, &p, "%s\n", engine->name);
+			GEM_TRACE_DUMP();
+
+			i915_request_put(lo);
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_vma;
+		}
+		GEM_BUG_ON(READ_ONCE(*map));
+		i915_request_put(lo);
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_vma;
+		}
+	}
+
+	err = 0;
+err_vma:
+	i915_vma_unpin(vma);
+err_map:
+	i915_gem_object_unpin_map(obj);
+err_obj:
+	i915_gem_object_put(obj);
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+	return err;
+}
+
+static struct i915_request *
+spinner_create_request(struct igt_spinner *spin,
+		       struct i915_gem_context *ctx,
+		       struct intel_engine_cs *engine,
+		       u32 arb)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+
+	ce = i915_gem_context_get_engine(ctx, engine->legacy_idx);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	rq = igt_spinner_create_request(spin, ce, arb);
+	intel_context_put(ce);
+	return rq;
+}
+
+static int live_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct igt_spinner spin_hi, spin_lo;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	if (igt_spinner_init(&spin_hi, gt))
+		return -ENOMEM;
+
+	if (igt_spinner_init(&spin_lo, gt))
+		goto err_spin_hi;
+
+	ctx_hi = kernel_context(gt->i915);
+	if (!ctx_hi)
+		goto err_spin_lo;
+	ctx_hi->sched.priority =
+		I915_USER_PRIORITY(I915_CONTEXT_MAX_USER_PRIORITY);
+
+	ctx_lo = kernel_context(gt->i915);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+	ctx_lo->sched.priority =
+		I915_USER_PRIORITY(I915_CONTEXT_MIN_USER_PRIORITY);
+
+	for_each_engine(engine, gt, id) {
+		struct igt_live_test t;
+		struct i915_request *rq;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+
+		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ctx_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_lo, rq)) {
+			GEM_TRACE("lo spinner failed to start\n");
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+
+		rq = spinner_create_request(&spin_hi, ctx_hi, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			igt_spinner_end(&spin_lo);
+			err = PTR_ERR(rq);
+			goto err_ctx_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_hi, rq)) {
+			GEM_TRACE("hi spinner failed to start\n");
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+
+		igt_spinner_end(&spin_hi);
+		igt_spinner_end(&spin_lo);
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+	}
+
+	err = 0;
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+err_spin_lo:
+	igt_spinner_fini(&spin_lo);
+err_spin_hi:
+	igt_spinner_fini(&spin_hi);
+	return err;
+}
+
+static int live_late_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct igt_spinner spin_hi, spin_lo;
+	struct intel_engine_cs *engine;
+	struct i915_sched_attr attr = {};
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	if (igt_spinner_init(&spin_hi, gt))
+		return -ENOMEM;
+
+	if (igt_spinner_init(&spin_lo, gt))
+		goto err_spin_hi;
+
+	ctx_hi = kernel_context(gt->i915);
+	if (!ctx_hi)
+		goto err_spin_lo;
+
+	ctx_lo = kernel_context(gt->i915);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+
+	/* Make sure ctx_lo stays before ctx_hi until we trigger preemption. */
+	ctx_lo->sched.priority = I915_USER_PRIORITY(1);
+
+	for_each_engine(engine, gt, id) {
+		struct igt_live_test t;
+		struct i915_request *rq;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+
+		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ctx_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_lo, rq)) {
+			pr_err("First context failed to start\n");
+			goto err_wedged;
+		}
+
+		rq = spinner_create_request(&spin_hi, ctx_hi, engine,
+					    MI_NOOP);
+		if (IS_ERR(rq)) {
+			igt_spinner_end(&spin_lo);
+			err = PTR_ERR(rq);
+			goto err_ctx_lo;
+		}
+
+		i915_request_add(rq);
+		if (igt_wait_for_spinner(&spin_hi, rq)) {
+			pr_err("Second context overtook first?\n");
+			goto err_wedged;
+		}
+
+		attr.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX);
+		engine->schedule(rq, &attr);
+
+		if (!igt_wait_for_spinner(&spin_hi, rq)) {
+			pr_err("High priority context failed to preempt the low priority context\n");
+			GEM_TRACE_DUMP();
+			goto err_wedged;
+		}
+
+		igt_spinner_end(&spin_hi);
+		igt_spinner_end(&spin_lo);
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+	}
+
+	err = 0;
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+err_spin_lo:
+	igt_spinner_fini(&spin_lo);
+err_spin_hi:
+	igt_spinner_fini(&spin_hi);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&spin_hi);
+	igt_spinner_end(&spin_lo);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_ctx_lo;
+}
+
+struct preempt_client {
+	struct igt_spinner spin;
+	struct i915_gem_context *ctx;
+};
+
+static int preempt_client_init(struct intel_gt *gt, struct preempt_client *c)
+{
+	c->ctx = kernel_context(gt->i915);
+	if (!c->ctx)
+		return -ENOMEM;
+
+	if (igt_spinner_init(&c->spin, gt))
+		goto err_ctx;
+
+	return 0;
+
+err_ctx:
+	kernel_context_close(c->ctx);
+	return -ENOMEM;
+}
+
+static void preempt_client_fini(struct preempt_client *c)
+{
+	igt_spinner_fini(&c->spin);
+	kernel_context_close(c->ctx);
+}
+
+static int live_nopreempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct preempt_client a, b;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Verify that we can disable preemption for an individual request
+	 * that may be being observed and not want to be interrupted.
+	 */
+
+	if (preempt_client_init(gt, &a))
+		return -ENOMEM;
+	if (preempt_client_init(gt, &b))
+		goto err_client_a;
+	b.ctx->sched.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX);
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq_a, *rq_b;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		engine->execlists.preempt_hang.count = 0;
+
+		rq_a = spinner_create_request(&a.spin,
+					      a.ctx, engine,
+					      MI_ARB_CHECK);
+		if (IS_ERR(rq_a)) {
+			err = PTR_ERR(rq_a);
+			goto err_client_b;
+		}
+
+		/* Low priority client, but unpreemptable! */
+		__set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq_a->fence.flags);
+
+		i915_request_add(rq_a);
+		if (!igt_wait_for_spinner(&a.spin, rq_a)) {
+			pr_err("First client failed to start\n");
+			goto err_wedged;
+		}
+
+		rq_b = spinner_create_request(&b.spin,
+					      b.ctx, engine,
+					      MI_ARB_CHECK);
+		if (IS_ERR(rq_b)) {
+			err = PTR_ERR(rq_b);
+			goto err_client_b;
+		}
+
+		i915_request_add(rq_b);
+
+		/* B is much more important than A! (But A is unpreemptable.) */
+		GEM_BUG_ON(rq_prio(rq_b) <= rq_prio(rq_a));
+
+		/* Wait long enough for preemption and timeslicing */
+		if (igt_wait_for_spinner(&b.spin, rq_b)) {
+			pr_err("Second client started too early!\n");
+			goto err_wedged;
+		}
+
+		igt_spinner_end(&a.spin);
+
+		if (!igt_wait_for_spinner(&b.spin, rq_b)) {
+			pr_err("Second client failed to start\n");
+			goto err_wedged;
+		}
+
+		igt_spinner_end(&b.spin);
+
+		if (engine->execlists.preempt_hang.count) {
+			pr_err("Preemption recorded x%d; should have been suppressed!\n",
+			       engine->execlists.preempt_hang.count);
+			err = -EINVAL;
+			goto err_wedged;
+		}
+
+		if (igt_flush_test(gt->i915))
+			goto err_wedged;
+	}
+
+	err = 0;
+err_client_b:
+	preempt_client_fini(&b);
+err_client_a:
+	preempt_client_fini(&a);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&b.spin);
+	igt_spinner_end(&a.spin);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_client_b;
+}
+
+struct live_preempt_cancel {
+	struct intel_engine_cs *engine;
+	struct preempt_client a, b;
+};
+
+static int __cancel_active0(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq;
+	struct igt_live_test t;
+	int err;
+
+	/* Preempt cancel of ELSP0 */
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	if (igt_live_test_begin(&t, arg->engine->i915,
+				__func__, arg->engine->name))
+		return -EIO;
+
+	rq = spinner_create_request(&arg->a.spin,
+				    arg->a.ctx, arg->engine,
+				    MI_ARB_CHECK);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	clear_bit(CONTEXT_BANNED, &rq->context->flags);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
+		err = -EIO;
+		goto out;
+	}
+
+	intel_context_set_banned(rq->context);
+	err = intel_engine_pulse(arg->engine);
+	if (err)
+		goto out;
+
+	err = wait_for_reset(arg->engine, rq, HZ / 2);
+	if (err) {
+		pr_err("Cancelled inflight0 request did not reset\n");
+		goto out;
+	}
+
+out:
+	i915_request_put(rq);
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int __cancel_active1(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq[2] = {};
+	struct igt_live_test t;
+	int err;
+
+	/* Preempt cancel of ELSP1 */
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	if (igt_live_test_begin(&t, arg->engine->i915,
+				__func__, arg->engine->name))
+		return -EIO;
+
+	rq[0] = spinner_create_request(&arg->a.spin,
+				       arg->a.ctx, arg->engine,
+				       MI_NOOP); /* no preemption */
+	if (IS_ERR(rq[0]))
+		return PTR_ERR(rq[0]);
+
+	clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
+	i915_request_get(rq[0]);
+	i915_request_add(rq[0]);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
+		err = -EIO;
+		goto out;
+	}
+
+	rq[1] = spinner_create_request(&arg->b.spin,
+				       arg->b.ctx, arg->engine,
+				       MI_ARB_CHECK);
+	if (IS_ERR(rq[1])) {
+		err = PTR_ERR(rq[1]);
+		goto out;
+	}
+
+	clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
+	i915_request_get(rq[1]);
+	err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
+	i915_request_add(rq[1]);
+	if (err)
+		goto out;
+
+	intel_context_set_banned(rq[1]->context);
+	err = intel_engine_pulse(arg->engine);
+	if (err)
+		goto out;
+
+	igt_spinner_end(&arg->a.spin);
+	err = wait_for_reset(arg->engine, rq[1], HZ / 2);
+	if (err)
+		goto out;
+
+	if (rq[0]->fence.error != 0) {
+		pr_err("Normal inflight0 request did not complete\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (rq[1]->fence.error != -EIO) {
+		pr_err("Cancelled inflight1 request did not report -EIO\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+out:
+	i915_request_put(rq[1]);
+	i915_request_put(rq[0]);
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int __cancel_queued(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq[3] = {};
+	struct igt_live_test t;
+	int err;
+
+	/* Full ELSP and one in the wings */
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	if (igt_live_test_begin(&t, arg->engine->i915,
+				__func__, arg->engine->name))
+		return -EIO;
+
+	rq[0] = spinner_create_request(&arg->a.spin,
+				       arg->a.ctx, arg->engine,
+				       MI_ARB_CHECK);
+	if (IS_ERR(rq[0]))
+		return PTR_ERR(rq[0]);
+
+	clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
+	i915_request_get(rq[0]);
+	i915_request_add(rq[0]);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
+		err = -EIO;
+		goto out;
+	}
+
+	rq[1] = igt_request_alloc(arg->b.ctx, arg->engine);
+	if (IS_ERR(rq[1])) {
+		err = PTR_ERR(rq[1]);
+		goto out;
+	}
+
+	clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
+	i915_request_get(rq[1]);
+	err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
+	i915_request_add(rq[1]);
+	if (err)
+		goto out;
+
+	rq[2] = spinner_create_request(&arg->b.spin,
+				       arg->a.ctx, arg->engine,
+				       MI_ARB_CHECK);
+	if (IS_ERR(rq[2])) {
+		err = PTR_ERR(rq[2]);
+		goto out;
+	}
+
+	i915_request_get(rq[2]);
+	err = i915_request_await_dma_fence(rq[2], &rq[1]->fence);
+	i915_request_add(rq[2]);
+	if (err)
+		goto out;
+
+	intel_context_set_banned(rq[2]->context);
+	err = intel_engine_pulse(arg->engine);
+	if (err)
+		goto out;
+
+	err = wait_for_reset(arg->engine, rq[2], HZ / 2);
+	if (err)
+		goto out;
+
+	if (rq[0]->fence.error != -EIO) {
+		pr_err("Cancelled inflight0 request did not report -EIO\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (rq[1]->fence.error != 0) {
+		pr_err("Normal inflight1 request did not complete\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (rq[2]->fence.error != -EIO) {
+		pr_err("Cancelled queued request did not report -EIO\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+out:
+	i915_request_put(rq[2]);
+	i915_request_put(rq[1]);
+	i915_request_put(rq[0]);
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int __cancel_hostile(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq;
+	int err;
+
+	/* Preempt cancel non-preemptible spinner in ELSP0 */
+	if (!IS_ACTIVE(CONFIG_DRM_I915_PREEMPT_TIMEOUT))
+		return 0;
+
+	if (!intel_has_reset_engine(arg->engine->gt))
+		return 0;
+
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	rq = spinner_create_request(&arg->a.spin,
+				    arg->a.ctx, arg->engine,
+				    MI_NOOP); /* preemption disabled */
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	clear_bit(CONTEXT_BANNED, &rq->context->flags);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
+		err = -EIO;
+		goto out;
+	}
+
+	intel_context_set_banned(rq->context);
+	err = intel_engine_pulse(arg->engine); /* force reset */
+	if (err)
+		goto out;
+
+	err = wait_for_reset(arg->engine, rq, HZ / 2);
+	if (err) {
+		pr_err("Cancelled inflight0 request did not reset\n");
+		goto out;
+	}
+
+out:
+	i915_request_put(rq);
+	if (igt_flush_test(arg->engine->i915))
+		err = -EIO;
+	return err;
+}
+
+static void force_reset_timeout(struct intel_engine_cs *engine)
+{
+	engine->reset_timeout.probability = 999;
+	atomic_set(&engine->reset_timeout.times, -1);
+}
+
+static void cancel_reset_timeout(struct intel_engine_cs *engine)
+{
+	memset(&engine->reset_timeout, 0, sizeof(engine->reset_timeout));
+}
+
+static int __cancel_fail(struct live_preempt_cancel *arg)
+{
+	struct intel_engine_cs *engine = arg->engine;
+	struct i915_request *rq;
+	int err;
+
+	if (!IS_ACTIVE(CONFIG_DRM_I915_PREEMPT_TIMEOUT))
+		return 0;
+
+	if (!intel_has_reset_engine(engine->gt))
+		return 0;
+
+	GEM_TRACE("%s(%s)\n", __func__, engine->name);
+	rq = spinner_create_request(&arg->a.spin,
+				    arg->a.ctx, engine,
+				    MI_NOOP); /* preemption disabled */
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	clear_bit(CONTEXT_BANNED, &rq->context->flags);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
+		err = -EIO;
+		goto out;
+	}
+
+	intel_context_set_banned(rq->context);
+
+	err = intel_engine_pulse(engine);
+	if (err)
+		goto out;
+
+	force_reset_timeout(engine);
+
+	/* force preempt reset [failure] */
+	while (!engine->execlists.pending[0])
+		intel_engine_flush_submission(engine);
+	del_timer_sync(&engine->execlists.preempt);
+	intel_engine_flush_submission(engine);
+
+	cancel_reset_timeout(engine);
+
+	/* after failure, require heartbeats to reset device */
+	intel_engine_set_heartbeat(engine, 1);
+	err = wait_for_reset(engine, rq, HZ / 2);
+	intel_engine_set_heartbeat(engine,
+				   engine->defaults.heartbeat_interval_ms);
+	if (err) {
+		pr_err("Cancelled inflight0 request did not reset\n");
+		goto out;
+	}
+
+out:
+	i915_request_put(rq);
+	if (igt_flush_test(engine->i915))
+		err = -EIO;
+	return err;
+}
+
+static int live_preempt_cancel(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct live_preempt_cancel data;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * To cancel an inflight context, we need to first remove it from the
+	 * GPU. That sounds like preemption! Plus a little bit of bookkeeping.
+	 */
+
+	if (preempt_client_init(gt, &data.a))
+		return -ENOMEM;
+	if (preempt_client_init(gt, &data.b))
+		goto err_client_a;
+
+	for_each_engine(data.engine, gt, id) {
+		if (!intel_engine_has_preemption(data.engine))
+			continue;
+
+		err = __cancel_active0(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_active1(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_queued(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_hostile(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_fail(&data);
+		if (err)
+			goto err_wedged;
+	}
+
+	err = 0;
+err_client_b:
+	preempt_client_fini(&data.b);
+err_client_a:
+	preempt_client_fini(&data.a);
+	return err;
+
+err_wedged:
+	GEM_TRACE_DUMP();
+	igt_spinner_end(&data.b.spin);
+	igt_spinner_end(&data.a.spin);
+	intel_gt_set_wedged(gt);
+	goto err_client_b;
+}
+
+static int live_suppress_self_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct i915_sched_attr attr = {
+		.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX)
+	};
+	struct preempt_client a, b;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Verify that if a preemption request does not cause a change in
+	 * the current execution order, the preempt-to-idle injection is
+	 * skipped and that we do not accidentally apply it after the CS
+	 * completion event.
+	 */
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0; /* presume black blox */
+
+	if (intel_vgpu_active(gt->i915))
+		return 0; /* GVT forces single port & request submission */
+
+	if (preempt_client_init(gt, &a))
+		return -ENOMEM;
+	if (preempt_client_init(gt, &b))
+		goto err_client_a;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq_a, *rq_b;
+		int depth;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_flush_test(gt->i915))
+			goto err_wedged;
+
+		st_engine_heartbeat_disable(engine);
+		engine->execlists.preempt_hang.count = 0;
+
+		rq_a = spinner_create_request(&a.spin,
+					      a.ctx, engine,
+					      MI_NOOP);
+		if (IS_ERR(rq_a)) {
+			err = PTR_ERR(rq_a);
+			st_engine_heartbeat_enable(engine);
+			goto err_client_b;
+		}
+
+		i915_request_add(rq_a);
+		if (!igt_wait_for_spinner(&a.spin, rq_a)) {
+			pr_err("First client failed to start\n");
+			st_engine_heartbeat_enable(engine);
+			goto err_wedged;
+		}
+
+		/* Keep postponing the timer to avoid premature slicing */
+		mod_timer(&engine->execlists.timer, jiffies + HZ);
+		for (depth = 0; depth < 8; depth++) {
+			rq_b = spinner_create_request(&b.spin,
+						      b.ctx, engine,
+						      MI_NOOP);
+			if (IS_ERR(rq_b)) {
+				err = PTR_ERR(rq_b);
+				st_engine_heartbeat_enable(engine);
+				goto err_client_b;
+			}
+			i915_request_add(rq_b);
+
+			GEM_BUG_ON(i915_request_completed(rq_a));
+			engine->schedule(rq_a, &attr);
+			igt_spinner_end(&a.spin);
+
+			if (!igt_wait_for_spinner(&b.spin, rq_b)) {
+				pr_err("Second client failed to start\n");
+				st_engine_heartbeat_enable(engine);
+				goto err_wedged;
+			}
+
+			swap(a, b);
+			rq_a = rq_b;
+		}
+		igt_spinner_end(&a.spin);
+
+		if (engine->execlists.preempt_hang.count) {
+			pr_err("Preemption on %s recorded x%d, depth %d; should have been suppressed!\n",
+			       engine->name,
+			       engine->execlists.preempt_hang.count,
+			       depth);
+			st_engine_heartbeat_enable(engine);
+			err = -EINVAL;
+			goto err_client_b;
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (igt_flush_test(gt->i915))
+			goto err_wedged;
+	}
+
+	err = 0;
+err_client_b:
+	preempt_client_fini(&b);
+err_client_a:
+	preempt_client_fini(&a);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&b.spin);
+	igt_spinner_end(&a.spin);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_client_b;
+}
+
+static int live_chain_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct preempt_client hi, lo;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Build a chain AB...BA between two contexts (A, B) and request
+	 * preemption of the last request. It should then complete before
+	 * the previously submitted spinner in B.
+	 */
+
+	if (preempt_client_init(gt, &hi))
+		return -ENOMEM;
+
+	if (preempt_client_init(gt, &lo))
+		goto err_client_hi;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_sched_attr attr = {
+			.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX),
+		};
+		struct igt_live_test t;
+		struct i915_request *rq;
+		int ring_size, count, i;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		rq = spinner_create_request(&lo.spin,
+					    lo.ctx, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq))
+			goto err_wedged;
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		ring_size = rq->wa_tail - rq->head;
+		if (ring_size < 0)
+			ring_size += rq->ring->size;
+		ring_size = rq->ring->size / ring_size;
+		pr_debug("%s(%s): Using maximum of %d requests\n",
+			 __func__, engine->name, ring_size);
+
+		igt_spinner_end(&lo.spin);
+		if (i915_request_wait(rq, 0, HZ / 2) < 0) {
+			pr_err("Timed out waiting to flush %s\n", engine->name);
+			i915_request_put(rq);
+			goto err_wedged;
+		}
+		i915_request_put(rq);
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_wedged;
+		}
+
+		for_each_prime_number_from(count, 1, ring_size) {
+			rq = spinner_create_request(&hi.spin,
+						    hi.ctx, engine,
+						    MI_ARB_CHECK);
+			if (IS_ERR(rq))
+				goto err_wedged;
+			i915_request_add(rq);
+			if (!igt_wait_for_spinner(&hi.spin, rq))
+				goto err_wedged;
+
+			rq = spinner_create_request(&lo.spin,
+						    lo.ctx, engine,
+						    MI_ARB_CHECK);
+			if (IS_ERR(rq))
+				goto err_wedged;
+			i915_request_add(rq);
+
+			for (i = 0; i < count; i++) {
+				rq = igt_request_alloc(lo.ctx, engine);
+				if (IS_ERR(rq))
+					goto err_wedged;
+				i915_request_add(rq);
+			}
+
+			rq = igt_request_alloc(hi.ctx, engine);
+			if (IS_ERR(rq))
+				goto err_wedged;
+
+			i915_request_get(rq);
+			i915_request_add(rq);
+			engine->schedule(rq, &attr);
+
+			igt_spinner_end(&hi.spin);
+			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+				struct drm_printer p =
+					drm_info_printer(gt->i915->drm.dev);
+
+				pr_err("Failed to preempt over chain of %d\n",
+				       count);
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+				i915_request_put(rq);
+				goto err_wedged;
+			}
+			igt_spinner_end(&lo.spin);
+			i915_request_put(rq);
+
+			rq = igt_request_alloc(lo.ctx, engine);
+			if (IS_ERR(rq))
+				goto err_wedged;
+
+			i915_request_get(rq);
+			i915_request_add(rq);
+
+			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+				struct drm_printer p =
+					drm_info_printer(gt->i915->drm.dev);
+
+				pr_err("Failed to flush low priority chain of %d requests\n",
+				       count);
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+
+				i915_request_put(rq);
+				goto err_wedged;
+			}
+			i915_request_put(rq);
+		}
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_wedged;
+		}
+	}
+
+	err = 0;
+err_client_lo:
+	preempt_client_fini(&lo);
+err_client_hi:
+	preempt_client_fini(&hi);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&hi.spin);
+	igt_spinner_end(&lo.spin);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_client_lo;
+}
+
+static int create_gang(struct intel_engine_cs *engine,
+		       struct i915_request **prev)
+{
+	struct drm_i915_gem_object *obj;
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	obj = i915_gem_object_create_internal(engine->i915, 4096);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto err_ce;
+	}
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err_obj;
+
+	cs = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_obj;
+	}
+
+	/* Semaphore target: spin until zero */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = lower_32_bits(vma->node.start);
+	*cs++ = upper_32_bits(vma->node.start);
+
+	if (*prev) {
+		u64 offset = (*prev)->batch->node.start;
+
+		/* Terminate the spinner in the next lower priority batch. */
+		*cs++ = MI_STORE_DWORD_IMM_GEN4;
+		*cs++ = lower_32_bits(offset);
+		*cs++ = upper_32_bits(offset);
+		*cs++ = 0;
+	}
+
+	*cs++ = MI_BATCH_BUFFER_END;
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_obj;
+	}
+
+	rq->batch = i915_vma_get(vma);
+	i915_request_get(rq);
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, false);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, 0);
+	if (!err)
+		err = rq->engine->emit_bb_start(rq,
+						vma->node.start,
+						PAGE_SIZE, 0);
+	i915_vma_unlock(vma);
+	i915_request_add(rq);
+	if (err)
+		goto err_rq;
+
+	i915_gem_object_put(obj);
+	intel_context_put(ce);
+
+	rq->mock.link.next = &(*prev)->mock.link;
+	*prev = rq;
+	return 0;
+
+err_rq:
+	i915_vma_put(rq->batch);
+	i915_request_put(rq);
+err_obj:
+	i915_gem_object_put(obj);
+err_ce:
+	intel_context_put(ce);
+	return err;
+}
+
+static int __live_preempt_ring(struct intel_engine_cs *engine,
+			       struct igt_spinner *spin,
+			       int queue_sz, int ring_sz)
+{
+	struct intel_context *ce[2] = {};
+	struct i915_request *rq;
+	struct igt_live_test t;
+	int err = 0;
+	int n;
+
+	if (igt_live_test_begin(&t, engine->i915, __func__, engine->name))
+		return -EIO;
+
+	for (n = 0; n < ARRAY_SIZE(ce); n++) {
+		struct intel_context *tmp;
+
+		tmp = intel_context_create(engine);
+		if (IS_ERR(tmp)) {
+			err = PTR_ERR(tmp);
+			goto err_ce;
+		}
+
+		tmp->ring = __intel_context_ring_size(ring_sz);
+
+		err = intel_context_pin(tmp);
+		if (err) {
+			intel_context_put(tmp);
+			goto err_ce;
+		}
+
+		memset32(tmp->ring->vaddr,
+			 0xdeadbeef, /* trigger a hang if executed */
+			 tmp->ring->vma->size / sizeof(u32));
+
+		ce[n] = tmp;
+	}
+
+	rq = igt_spinner_create_request(spin, ce[0], MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_ce;
+	}
+
+	i915_request_get(rq);
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	i915_request_add(rq);
+
+	if (!igt_wait_for_spinner(spin, rq)) {
+		intel_gt_set_wedged(engine->gt);
+		i915_request_put(rq);
+		err = -ETIME;
+		goto err_ce;
+	}
+
+	/* Fill the ring, until we will cause a wrap */
+	n = 0;
+	while (ce[0]->ring->tail - rq->wa_tail <= queue_sz) {
+		struct i915_request *tmp;
+
+		tmp = intel_context_create_request(ce[0]);
+		if (IS_ERR(tmp)) {
+			err = PTR_ERR(tmp);
+			i915_request_put(rq);
+			goto err_ce;
+		}
+
+		i915_request_add(tmp);
+		intel_engine_flush_submission(engine);
+		n++;
+	}
+	intel_engine_flush_submission(engine);
+	pr_debug("%s: Filled %d with %d nop tails {size:%x, tail:%x, emit:%x, rq.tail:%x}\n",
+		 engine->name, queue_sz, n,
+		 ce[0]->ring->size,
+		 ce[0]->ring->tail,
+		 ce[0]->ring->emit,
+		 rq->tail);
+	i915_request_put(rq);
+
+	/* Create a second request to preempt the first ring */
+	rq = intel_context_create_request(ce[1]);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_ce;
+	}
+
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	err = wait_for_submit(engine, rq, HZ / 2);
+	i915_request_put(rq);
+	if (err) {
+		pr_err("%s: preemption request was not submited\n",
+		       engine->name);
+		err = -ETIME;
+	}
+
+	pr_debug("%s: ring[0]:{ tail:%x, emit:%x }, ring[1]:{ tail:%x, emit:%x }\n",
+		 engine->name,
+		 ce[0]->ring->tail, ce[0]->ring->emit,
+		 ce[1]->ring->tail, ce[1]->ring->emit);
+
+err_ce:
+	intel_engine_flush_submission(engine);
+	igt_spinner_end(spin);
+	for (n = 0; n < ARRAY_SIZE(ce); n++) {
+		if (IS_ERR_OR_NULL(ce[n]))
+			break;
+
+		intel_context_unpin(ce[n]);
+		intel_context_put(ce[n]);
+	}
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int live_preempt_ring(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct igt_spinner spin;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check that we rollback large chunks of a ring in order to do a
+	 * preemption event. Similar to live_unlite_ring, but looking at
+	 * ring size rather than the impact of intel_ring_direction().
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		int n;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+
+		for (n = 0; n <= 3; n++) {
+			err = __live_preempt_ring(engine, &spin,
+						  n * SZ_4K / 4, SZ_4K);
+			if (err)
+				break;
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_preempt_gang(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Build as long a chain of preempters as we can, with each
+	 * request higher priority than the last. Once we are ready, we release
+	 * the last batch which then precolates down the chain, each releasing
+	 * the next oldest in turn. The intent is to simply push as hard as we
+	 * can with the number of preemptions, trying to exceed narrow HW
+	 * limits. At a minimum, we insist that we can sort all the user
+	 * high priority levels into execution order.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq = NULL;
+		struct igt_live_test t;
+		IGT_TIMEOUT(end_time);
+		int prio = 0;
+		int err = 0;
+		u32 *cs;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name))
+			return -EIO;
+
+		do {
+			struct i915_sched_attr attr = {
+				.priority = I915_USER_PRIORITY(prio++),
+			};
+
+			err = create_gang(engine, &rq);
+			if (err)
+				break;
+
+			/* Submit each spinner at increasing priority */
+			engine->schedule(rq, &attr);
+		} while (prio <= I915_PRIORITY_MAX &&
+			 !__igt_timeout(end_time, NULL));
+		pr_debug("%s: Preempt chain of %d requests\n",
+			 engine->name, prio);
+
+		/*
+		 * Such that the last spinner is the highest priority and
+		 * should execute first. When that spinner completes,
+		 * it will terminate the next lowest spinner until there
+		 * are no more spinners and the gang is complete.
+		 */
+		cs = i915_gem_object_pin_map(rq->batch->obj, I915_MAP_WC);
+		if (!IS_ERR(cs)) {
+			*cs = 0;
+			i915_gem_object_unpin_map(rq->batch->obj);
+		} else {
+			err = PTR_ERR(cs);
+			intel_gt_set_wedged(gt);
+		}
+
+		while (rq) { /* wait for each rq from highest to lowest prio */
+			struct i915_request *n = list_next_entry(rq, mock.link);
+
+			if (err == 0 && i915_request_wait(rq, 0, HZ / 5) < 0) {
+				struct drm_printer p =
+					drm_info_printer(engine->i915->drm.dev);
+
+				pr_err("Failed to flush chain of %d requests, at %d\n",
+				       prio, rq_prio(rq) >> I915_USER_PRIORITY_SHIFT);
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+
+				err = -ETIME;
+			}
+
+			i915_vma_put(rq->batch);
+			i915_request_put(rq);
+			rq = n;
+		}
+
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static struct i915_vma *
+create_gpr_user(struct intel_engine_cs *engine,
+		struct i915_vma *result,
+		unsigned int offset)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+	int i;
+
+	obj = i915_gem_object_create_internal(engine->i915, 4096);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, result->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err) {
+		i915_vma_put(vma);
+		return ERR_PTR(err);
+	}
+
+	cs = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		i915_vma_put(vma);
+		return ERR_CAST(cs);
+	}
+
+	/* All GPR are clear for new contexts. We use GPR(0) as a constant */
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = CS_GPR(engine, 0);
+	*cs++ = 1;
+
+	for (i = 1; i < NUM_GPR; i++) {
+		u64 addr;
+
+		/*
+		 * Perform: GPR[i]++
+		 *
+		 * As we read and write into the context saved GPR[i], if
+		 * we restart this batch buffer from an earlier point, we
+		 * will repeat the increment and store a value > 1.
+		 */
+		*cs++ = MI_MATH(4);
+		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(i));
+		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(0));
+		*cs++ = MI_MATH_ADD;
+		*cs++ = MI_MATH_STORE(MI_MATH_REG(i), MI_MATH_REG_ACCU);
+
+		addr = result->node.start + offset + i * sizeof(*cs);
+		*cs++ = MI_STORE_REGISTER_MEM_GEN8;
+		*cs++ = CS_GPR(engine, 2 * i);
+		*cs++ = lower_32_bits(addr);
+		*cs++ = upper_32_bits(addr);
+
+		*cs++ = MI_SEMAPHORE_WAIT |
+			MI_SEMAPHORE_POLL |
+			MI_SEMAPHORE_SAD_GTE_SDD;
+		*cs++ = i;
+		*cs++ = lower_32_bits(result->node.start);
+		*cs++ = upper_32_bits(result->node.start);
+	}
+
+	*cs++ = MI_BATCH_BUFFER_END;
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	return vma;
+}
+
+static struct i915_vma *create_global(struct intel_gt *gt, size_t sz)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_internal(gt->i915, sz);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_ggtt_pin(vma, NULL, 0, 0);
+	if (err) {
+		i915_vma_put(vma);
+		return ERR_PTR(err);
+	}
+
+	return vma;
+}
+
+static struct i915_request *
+create_gpr_client(struct intel_engine_cs *engine,
+		  struct i915_vma *global,
+		  unsigned int offset)
+{
+	struct i915_vma *batch, *vma;
+	struct intel_context *ce;
+	struct i915_request *rq;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	vma = i915_vma_instance(global->obj, ce->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_ce;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto out_ce;
+
+	batch = create_gpr_user(engine, vma, offset);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto out_vma;
+	}
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_batch;
+	}
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, false);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, 0);
+	i915_vma_unlock(vma);
+
+	i915_vma_lock(batch);
+	if (!err)
+		err = i915_request_await_object(rq, batch->obj, false);
+	if (!err)
+		err = i915_vma_move_to_active(batch, rq, 0);
+	if (!err)
+		err = rq->engine->emit_bb_start(rq,
+						batch->node.start,
+						PAGE_SIZE, 0);
+	i915_vma_unlock(batch);
+	i915_vma_unpin(batch);
+
+	if (!err)
+		i915_request_get(rq);
+	i915_request_add(rq);
+
+out_batch:
+	i915_vma_put(batch);
+out_vma:
+	i915_vma_unpin(vma);
+out_ce:
+	intel_context_put(ce);
+	return err ? ERR_PTR(err) : rq;
+}
+
+static int preempt_user(struct intel_engine_cs *engine,
+			struct i915_vma *global,
+			int id)
+{
+	struct i915_sched_attr attr = {
+		.priority = I915_PRIORITY_MAX
+	};
+	struct i915_request *rq;
+	int err = 0;
+	u32 *cs;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = i915_ggtt_offset(global);
+	*cs++ = 0;
+	*cs++ = id;
+
+	intel_ring_advance(rq, cs);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	engine->schedule(rq, &attr);
+
+	if (i915_request_wait(rq, 0, HZ / 2) < 0)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static int live_preempt_user(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct i915_vma *global;
+	enum intel_engine_id id;
+	u32 *result;
+	int err = 0;
+
+	/*
+	 * In our other tests, we look at preemption in carefully
+	 * controlled conditions in the ringbuffer. Since most of the
+	 * time is spent in user batches, most of our preemptions naturally
+	 * occur there. We want to verify that when we preempt inside a batch
+	 * we continue on from the current instruction and do not roll back
+	 * to the start, or another earlier arbitration point.
+	 *
+	 * To verify this, we create a batch which is a mixture of
+	 * MI_MATH (gpr++) MI_SRM (gpr) and preemption points. Then with
+	 * a few preempting contexts thrown into the mix, we look for any
+	 * repeated instructions (which show up as incorrect values).
+	 */
+
+	global = create_global(gt, 4096);
+	if (IS_ERR(global))
+		return PTR_ERR(global);
+
+	result = i915_gem_object_pin_map(global->obj, I915_MAP_WC);
+	if (IS_ERR(result)) {
+		i915_vma_unpin_and_release(&global, 0);
+		return PTR_ERR(result);
+	}
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *client[3] = {};
+		struct igt_live_test t;
+		int i;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (IS_GEN(gt->i915, 8) && engine->class != RENDER_CLASS)
+			continue; /* we need per-context GPR */
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+
+		memset(result, 0, 4096);
+
+		for (i = 0; i < ARRAY_SIZE(client); i++) {
+			struct i915_request *rq;
+
+			rq = create_gpr_client(engine, global,
+					       NUM_GPR * i * sizeof(u32));
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto end_test;
+			}
+
+			client[i] = rq;
+		}
+
+		/* Continuously preempt the set of 3 running contexts */
+		for (i = 1; i <= NUM_GPR; i++) {
+			err = preempt_user(engine, global, i);
+			if (err)
+				goto end_test;
+		}
+
+		if (READ_ONCE(result[0]) != NUM_GPR) {
+			pr_err("%s: Failed to release semaphore\n",
+			       engine->name);
+			err = -EIO;
+			goto end_test;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(client); i++) {
+			int gpr;
+
+			if (i915_request_wait(client[i], 0, HZ / 2) < 0) {
+				err = -ETIME;
+				goto end_test;
+			}
+
+			for (gpr = 1; gpr < NUM_GPR; gpr++) {
+				if (result[NUM_GPR * i + gpr] != 1) {
+					pr_err("%s: Invalid result, client %d, gpr %d, result: %d\n",
+					       engine->name,
+					       i, gpr, result[NUM_GPR * i + gpr]);
+					err = -EINVAL;
+					goto end_test;
+				}
+			}
+		}
+
+end_test:
+		for (i = 0; i < ARRAY_SIZE(client); i++) {
+			if (!client[i])
+				break;
+
+			i915_request_put(client[i]);
+		}
+
+		/* Flush the semaphores on error */
+		smp_store_mb(result[0], -1);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	i915_vma_unpin_and_release(&global, I915_VMA_RELEASE_MAP);
+	return err;
+}
+
+static int live_preempt_timeout(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct igt_spinner spin_lo;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Check that we force preemption to occur by cancelling the previous
+	 * context if it refuses to yield the GPU.
+	 */
+	if (!IS_ACTIVE(CONFIG_DRM_I915_PREEMPT_TIMEOUT))
+		return 0;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (igt_spinner_init(&spin_lo, gt))
+		return -ENOMEM;
+
+	ctx_hi = kernel_context(gt->i915);
+	if (!ctx_hi)
+		goto err_spin_lo;
+	ctx_hi->sched.priority =
+		I915_USER_PRIORITY(I915_CONTEXT_MAX_USER_PRIORITY);
+
+	ctx_lo = kernel_context(gt->i915);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+	ctx_lo->sched.priority =
+		I915_USER_PRIORITY(I915_CONTEXT_MIN_USER_PRIORITY);
+
+	for_each_engine(engine, gt, id) {
+		unsigned long saved_timeout;
+		struct i915_request *rq;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
+					    MI_NOOP); /* preemption disabled */
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ctx_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_lo, rq)) {
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_ctx_lo;
+		}
+
+		rq = igt_request_alloc(ctx_hi, engine);
+		if (IS_ERR(rq)) {
+			igt_spinner_end(&spin_lo);
+			err = PTR_ERR(rq);
+			goto err_ctx_lo;
+		}
+
+		/* Flush the previous CS ack before changing timeouts */
+		while (READ_ONCE(engine->execlists.pending[0]))
+			cpu_relax();
+
+		saved_timeout = engine->props.preempt_timeout_ms;
+		engine->props.preempt_timeout_ms = 1; /* in ms, -> 1 jiffie */
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		intel_engine_flush_submission(engine);
+		engine->props.preempt_timeout_ms = saved_timeout;
+
+		if (i915_request_wait(rq, 0, HZ / 10) < 0) {
+			intel_gt_set_wedged(gt);
+			i915_request_put(rq);
+			err = -ETIME;
+			goto err_ctx_lo;
+		}
+
+		igt_spinner_end(&spin_lo);
+		i915_request_put(rq);
+	}
+
+	err = 0;
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+err_spin_lo:
+	igt_spinner_fini(&spin_lo);
+	return err;
+}
+
+static int random_range(struct rnd_state *rnd, int min, int max)
+{
+	return i915_prandom_u32_max_state(max - min, rnd) + min;
+}
+
+static int random_priority(struct rnd_state *rnd)
+{
+	return random_range(rnd, I915_PRIORITY_MIN, I915_PRIORITY_MAX);
+}
+
+struct preempt_smoke {
+	struct intel_gt *gt;
+	struct i915_gem_context **contexts;
+	struct intel_engine_cs *engine;
+	struct drm_i915_gem_object *batch;
+	unsigned int ncontext;
+	struct rnd_state prng;
+	unsigned long count;
+};
+
+static struct i915_gem_context *smoke_context(struct preempt_smoke *smoke)
+{
+	return smoke->contexts[i915_prandom_u32_max_state(smoke->ncontext,
+							  &smoke->prng)];
+}
+
+static int smoke_submit(struct preempt_smoke *smoke,
+			struct i915_gem_context *ctx, int prio,
+			struct drm_i915_gem_object *batch)
+{
+	struct i915_request *rq;
+	struct i915_vma *vma = NULL;
+	int err = 0;
+
+	if (batch) {
+		struct i915_address_space *vm;
+
+		vm = i915_gem_context_get_vm_rcu(ctx);
+		vma = i915_vma_instance(batch, vm, NULL);
+		i915_vm_put(vm);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+
+		err = i915_vma_pin(vma, 0, 0, PIN_USER);
+		if (err)
+			return err;
+	}
+
+	ctx->sched.priority = prio;
+
+	rq = igt_request_alloc(ctx, smoke->engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto unpin;
+	}
+
+	if (vma) {
+		i915_vma_lock(vma);
+		err = i915_request_await_object(rq, vma->obj, false);
+		if (!err)
+			err = i915_vma_move_to_active(vma, rq, 0);
+		if (!err)
+			err = rq->engine->emit_bb_start(rq,
+							vma->node.start,
+							PAGE_SIZE, 0);
+		i915_vma_unlock(vma);
+	}
+
+	i915_request_add(rq);
+
+unpin:
+	if (vma)
+		i915_vma_unpin(vma);
+
+	return err;
+}
+
+static int smoke_crescendo_thread(void *arg)
+{
+	struct preempt_smoke *smoke = arg;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+
+	count = 0;
+	do {
+		struct i915_gem_context *ctx = smoke_context(smoke);
+		int err;
+
+		err = smoke_submit(smoke,
+				   ctx, count % I915_PRIORITY_MAX,
+				   smoke->batch);
+		if (err)
+			return err;
+
+		count++;
+	} while (count < smoke->ncontext && !__igt_timeout(end_time, NULL));
+
+	smoke->count = count;
+	return 0;
+}
+
+static int smoke_crescendo(struct preempt_smoke *smoke, unsigned int flags)
+#define BATCH BIT(0)
+{
+	struct task_struct *tsk[I915_NUM_ENGINES] = {};
+	struct preempt_smoke arg[I915_NUM_ENGINES];
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long count;
+	int err = 0;
+
+	for_each_engine(engine, smoke->gt, id) {
+		arg[id] = *smoke;
+		arg[id].engine = engine;
+		if (!(flags & BATCH))
+			arg[id].batch = NULL;
+		arg[id].count = 0;
+
+		tsk[id] = kthread_run(smoke_crescendo_thread, &arg,
+				      "igt/smoke:%d", id);
+		if (IS_ERR(tsk[id])) {
+			err = PTR_ERR(tsk[id]);
+			break;
+		}
+		get_task_struct(tsk[id]);
+	}
+
+	yield(); /* start all threads before we kthread_stop() */
+
+	count = 0;
+	for_each_engine(engine, smoke->gt, id) {
+		int status;
+
+		if (IS_ERR_OR_NULL(tsk[id]))
+			continue;
+
+		status = kthread_stop(tsk[id]);
+		if (status && !err)
+			err = status;
+
+		count += arg[id].count;
+
+		put_task_struct(tsk[id]);
+	}
+
+	pr_info("Submitted %lu crescendo:%x requests across %d engines and %d contexts\n",
+		count, flags, smoke->gt->info.num_engines, smoke->ncontext);
+	return 0;
+}
+
+static int smoke_random(struct preempt_smoke *smoke, unsigned int flags)
+{
+	enum intel_engine_id id;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+
+	count = 0;
+	do {
+		for_each_engine(smoke->engine, smoke->gt, id) {
+			struct i915_gem_context *ctx = smoke_context(smoke);
+			int err;
+
+			err = smoke_submit(smoke,
+					   ctx, random_priority(&smoke->prng),
+					   flags & BATCH ? smoke->batch : NULL);
+			if (err)
+				return err;
+
+			count++;
+		}
+	} while (count < smoke->ncontext && !__igt_timeout(end_time, NULL));
+
+	pr_info("Submitted %lu random:%x requests across %d engines and %d contexts\n",
+		count, flags, smoke->gt->info.num_engines, smoke->ncontext);
+	return 0;
+}
+
+static int live_preempt_smoke(void *arg)
+{
+	struct preempt_smoke smoke = {
+		.gt = arg,
+		.prng = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed),
+		.ncontext = 256,
+	};
+	const unsigned int phase[] = { 0, BATCH };
+	struct igt_live_test t;
+	int err = -ENOMEM;
+	u32 *cs;
+	int n;
+
+	smoke.contexts = kmalloc_array(smoke.ncontext,
+				       sizeof(*smoke.contexts),
+				       GFP_KERNEL);
+	if (!smoke.contexts)
+		return -ENOMEM;
+
+	smoke.batch =
+		i915_gem_object_create_internal(smoke.gt->i915, PAGE_SIZE);
+	if (IS_ERR(smoke.batch)) {
+		err = PTR_ERR(smoke.batch);
+		goto err_free;
+	}
+
+	cs = i915_gem_object_pin_map(smoke.batch, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_batch;
+	}
+	for (n = 0; n < PAGE_SIZE / sizeof(*cs) - 1; n++)
+		cs[n] = MI_ARB_CHECK;
+	cs[n] = MI_BATCH_BUFFER_END;
+	i915_gem_object_flush_map(smoke.batch);
+	i915_gem_object_unpin_map(smoke.batch);
+
+	if (igt_live_test_begin(&t, smoke.gt->i915, __func__, "all")) {
+		err = -EIO;
+		goto err_batch;
+	}
+
+	for (n = 0; n < smoke.ncontext; n++) {
+		smoke.contexts[n] = kernel_context(smoke.gt->i915);
+		if (!smoke.contexts[n])
+			goto err_ctx;
+	}
+
+	for (n = 0; n < ARRAY_SIZE(phase); n++) {
+		err = smoke_crescendo(&smoke, phase[n]);
+		if (err)
+			goto err_ctx;
+
+		err = smoke_random(&smoke, phase[n]);
+		if (err)
+			goto err_ctx;
+	}
+
+err_ctx:
+	if (igt_live_test_end(&t))
+		err = -EIO;
+
+	for (n = 0; n < smoke.ncontext; n++) {
+		if (!smoke.contexts[n])
+			break;
+		kernel_context_close(smoke.contexts[n]);
+	}
+
+err_batch:
+	i915_gem_object_put(smoke.batch);
+err_free:
+	kfree(smoke.contexts);
+
+	return err;
+}
+
+static int nop_virtual_engine(struct intel_gt *gt,
+			      struct intel_engine_cs **siblings,
+			      unsigned int nsibling,
+			      unsigned int nctx,
+			      unsigned int flags)
+#define CHAIN BIT(0)
+{
+	IGT_TIMEOUT(end_time);
+	struct i915_request *request[16] = {};
+	struct intel_context *ve[16];
+	unsigned long n, prime, nc;
+	struct igt_live_test t;
+	ktime_t times[2] = {};
+	int err;
+
+	GEM_BUG_ON(!nctx || nctx > ARRAY_SIZE(ve));
+
+	for (n = 0; n < nctx; n++) {
+		ve[n] = intel_execlists_create_virtual(siblings, nsibling);
+		if (IS_ERR(ve[n])) {
+			err = PTR_ERR(ve[n]);
+			nctx = n;
+			goto out;
+		}
+
+		err = intel_context_pin(ve[n]);
+		if (err) {
+			intel_context_put(ve[n]);
+			nctx = n;
+			goto out;
+		}
+	}
+
+	err = igt_live_test_begin(&t, gt->i915, __func__, ve[0]->engine->name);
+	if (err)
+		goto out;
+
+	for_each_prime_number_from(prime, 1, 8192) {
+		times[1] = ktime_get_raw();
+
+		if (flags & CHAIN) {
+			for (nc = 0; nc < nctx; nc++) {
+				for (n = 0; n < prime; n++) {
+					struct i915_request *rq;
+
+					rq = i915_request_create(ve[nc]);
+					if (IS_ERR(rq)) {
+						err = PTR_ERR(rq);
+						goto out;
+					}
+
+					if (request[nc])
+						i915_request_put(request[nc]);
+					request[nc] = i915_request_get(rq);
+					i915_request_add(rq);
+				}
+			}
+		} else {
+			for (n = 0; n < prime; n++) {
+				for (nc = 0; nc < nctx; nc++) {
+					struct i915_request *rq;
+
+					rq = i915_request_create(ve[nc]);
+					if (IS_ERR(rq)) {
+						err = PTR_ERR(rq);
+						goto out;
+					}
+
+					if (request[nc])
+						i915_request_put(request[nc]);
+					request[nc] = i915_request_get(rq);
+					i915_request_add(rq);
+				}
+			}
+		}
+
+		for (nc = 0; nc < nctx; nc++) {
+			if (i915_request_wait(request[nc], 0, HZ / 10) < 0) {
+				pr_err("%s(%s): wait for %llx:%lld timed out\n",
+				       __func__, ve[0]->engine->name,
+				       request[nc]->fence.context,
+				       request[nc]->fence.seqno);
+
+				GEM_TRACE("%s(%s) failed at request %llx:%lld\n",
+					  __func__, ve[0]->engine->name,
+					  request[nc]->fence.context,
+					  request[nc]->fence.seqno);
+				GEM_TRACE_DUMP();
+				intel_gt_set_wedged(gt);
+				break;
+			}
+		}
+
+		times[1] = ktime_sub(ktime_get_raw(), times[1]);
+		if (prime == 1)
+			times[0] = times[1];
+
+		for (nc = 0; nc < nctx; nc++) {
+			i915_request_put(request[nc]);
+			request[nc] = NULL;
+		}
+
+		if (__igt_timeout(end_time, NULL))
+			break;
+	}
+
+	err = igt_live_test_end(&t);
+	if (err)
+		goto out;
+
+	pr_info("Requestx%d latencies on %s: 1 = %lluns, %lu = %lluns\n",
+		nctx, ve[0]->engine->name, ktime_to_ns(times[0]),
+		prime, div64_u64(ktime_to_ns(times[1]), prime));
+
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	for (nc = 0; nc < nctx; nc++) {
+		i915_request_put(request[nc]);
+		intel_context_unpin(ve[nc]);
+		intel_context_put(ve[nc]);
+	}
+	return err;
+}
+
+static unsigned int
+__select_siblings(struct intel_gt *gt,
+		  unsigned int class,
+		  struct intel_engine_cs **siblings,
+		  bool (*filter)(const struct intel_engine_cs *))
+{
+	unsigned int n = 0;
+	unsigned int inst;
+
+	for (inst = 0; inst <= MAX_ENGINE_INSTANCE; inst++) {
+		if (!gt->engine_class[class][inst])
+			continue;
+
+		if (filter && !filter(gt->engine_class[class][inst]))
+			continue;
+
+		siblings[n++] = gt->engine_class[class][inst];
+	}
+
+	return n;
+}
+
+static unsigned int
+select_siblings(struct intel_gt *gt,
+		unsigned int class,
+		struct intel_engine_cs **siblings)
+{
+	return __select_siblings(gt, class, siblings, NULL);
+}
+
+static int live_virtual_engine(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		err = nop_virtual_engine(gt, &engine, 1, 1, 0);
+		if (err) {
+			pr_err("Failed to wrap engine %s: err=%d\n",
+			       engine->name, err);
+			return err;
+		}
+	}
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		int nsibling, n;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		for (n = 1; n <= nsibling + 1; n++) {
+			err = nop_virtual_engine(gt, siblings, nsibling,
+						 n, 0);
+			if (err)
+				return err;
+		}
+
+		err = nop_virtual_engine(gt, siblings, nsibling, n, CHAIN);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int mask_virtual_engine(struct intel_gt *gt,
+			       struct intel_engine_cs **siblings,
+			       unsigned int nsibling)
+{
+	struct i915_request *request[MAX_ENGINE_INSTANCE + 1];
+	struct intel_context *ve;
+	struct igt_live_test t;
+	unsigned int n;
+	int err;
+
+	/*
+	 * Check that by setting the execution mask on a request, we can
+	 * restrict it to our desired engine within the virtual engine.
+	 */
+
+	ve = intel_execlists_create_virtual(siblings, nsibling);
+	if (IS_ERR(ve)) {
+		err = PTR_ERR(ve);
+		goto out_close;
+	}
+
+	err = intel_context_pin(ve);
+	if (err)
+		goto out_put;
+
+	err = igt_live_test_begin(&t, gt->i915, __func__, ve->engine->name);
+	if (err)
+		goto out_unpin;
+
+	for (n = 0; n < nsibling; n++) {
+		request[n] = i915_request_create(ve);
+		if (IS_ERR(request[n])) {
+			err = PTR_ERR(request[n]);
+			nsibling = n;
+			goto out;
+		}
+
+		/* Reverse order as it's more likely to be unnatural */
+		request[n]->execution_mask = siblings[nsibling - n - 1]->mask;
+
+		i915_request_get(request[n]);
+		i915_request_add(request[n]);
+	}
+
+	for (n = 0; n < nsibling; n++) {
+		if (i915_request_wait(request[n], 0, HZ / 10) < 0) {
+			pr_err("%s(%s): wait for %llx:%lld timed out\n",
+			       __func__, ve->engine->name,
+			       request[n]->fence.context,
+			       request[n]->fence.seqno);
+
+			GEM_TRACE("%s(%s) failed at request %llx:%lld\n",
+				  __func__, ve->engine->name,
+				  request[n]->fence.context,
+				  request[n]->fence.seqno);
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto out;
+		}
+
+		if (request[n]->engine != siblings[nsibling - n - 1]) {
+			pr_err("Executed on wrong sibling '%s', expected '%s'\n",
+			       request[n]->engine->name,
+			       siblings[nsibling - n - 1]->name);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	err = igt_live_test_end(&t);
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	for (n = 0; n < nsibling; n++)
+		i915_request_put(request[n]);
+
+out_unpin:
+	intel_context_unpin(ve);
+out_put:
+	intel_context_put(ve);
+out_close:
+	return err;
+}
+
+static int live_virtual_mask(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		unsigned int nsibling;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		err = mask_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int slicein_virtual_engine(struct intel_gt *gt,
+				  struct intel_engine_cs **siblings,
+				  unsigned int nsibling)
+{
+	const long timeout = slice_timeout(siblings[0]);
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct igt_spinner spin;
+	unsigned int n;
+	int err = 0;
+
+	/*
+	 * Virtual requests must take part in timeslicing on the target engines.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for (n = 0; n < nsibling; n++) {
+		ce = intel_context_create(siblings[n]);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		i915_request_add(rq);
+	}
+
+	ce = intel_execlists_create_virtual(siblings, nsibling);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out;
+	}
+
+	rq = intel_context_create_request(ce);
+	intel_context_put(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, timeout) < 0) {
+		GEM_TRACE_ERR("%s(%s) failed to slice in virtual request\n",
+			      __func__, rq->engine->name);
+		GEM_TRACE_DUMP();
+		intel_gt_set_wedged(gt);
+		err = -EIO;
+	}
+	i915_request_put(rq);
+
+out:
+	igt_spinner_end(&spin);
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int sliceout_virtual_engine(struct intel_gt *gt,
+				   struct intel_engine_cs **siblings,
+				   unsigned int nsibling)
+{
+	const long timeout = slice_timeout(siblings[0]);
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct igt_spinner spin;
+	unsigned int n;
+	int err = 0;
+
+	/*
+	 * Virtual requests must allow others a fair timeslice.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	/* XXX We do not handle oversubscription and fairness with normal rq */
+	for (n = 0; n < nsibling; n++) {
+		ce = intel_execlists_create_virtual(siblings, nsibling);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		i915_request_add(rq);
+	}
+
+	for (n = 0; !err && n < nsibling; n++) {
+		ce = intel_context_create(siblings[n]);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		rq = intel_context_create_request(ce);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+		if (i915_request_wait(rq, 0, timeout) < 0) {
+			GEM_TRACE_ERR("%s(%s) failed to slice out virtual request\n",
+				      __func__, siblings[n]->name);
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+		}
+		i915_request_put(rq);
+	}
+
+out:
+	igt_spinner_end(&spin);
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_virtual_slice(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		unsigned int nsibling;
+
+		nsibling = __select_siblings(gt, class, siblings,
+					     intel_engine_has_timeslices);
+		if (nsibling < 2)
+			continue;
+
+		err = slicein_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+
+		err = sliceout_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int preserved_virtual_engine(struct intel_gt *gt,
+				    struct intel_engine_cs **siblings,
+				    unsigned int nsibling)
+{
+	struct i915_request *last = NULL;
+	struct intel_context *ve;
+	struct i915_vma *scratch;
+	struct igt_live_test t;
+	unsigned int n;
+	int err = 0;
+	u32 *cs;
+
+	scratch = __vm_create_scratch_for_read(&siblings[0]->gt->ggtt->vm,
+					       PAGE_SIZE);
+	if (IS_ERR(scratch))
+		return PTR_ERR(scratch);
+
+	err = i915_vma_sync(scratch);
+	if (err)
+		goto out_scratch;
+
+	ve = intel_execlists_create_virtual(siblings, nsibling);
+	if (IS_ERR(ve)) {
+		err = PTR_ERR(ve);
+		goto out_scratch;
+	}
+
+	err = intel_context_pin(ve);
+	if (err)
+		goto out_put;
+
+	err = igt_live_test_begin(&t, gt->i915, __func__, ve->engine->name);
+	if (err)
+		goto out_unpin;
+
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		struct intel_engine_cs *engine = siblings[n % nsibling];
+		struct i915_request *rq;
+
+		rq = i915_request_create(ve);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_end;
+		}
+
+		i915_request_put(last);
+		last = i915_request_get(rq);
+
+		cs = intel_ring_begin(rq, 8);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto out_end;
+		}
+
+		*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+		*cs++ = CS_GPR(engine, n);
+		*cs++ = i915_ggtt_offset(scratch) + n * sizeof(u32);
+		*cs++ = 0;
+
+		*cs++ = MI_LOAD_REGISTER_IMM(1);
+		*cs++ = CS_GPR(engine, (n + 1) % NUM_GPR_DW);
+		*cs++ = n + 1;
+
+		*cs++ = MI_NOOP;
+		intel_ring_advance(rq, cs);
+
+		/* Restrict this request to run on a particular engine */
+		rq->execution_mask = engine->mask;
+		i915_request_add(rq);
+	}
+
+	if (i915_request_wait(last, 0, HZ / 5) < 0) {
+		err = -ETIME;
+		goto out_end;
+	}
+
+	cs = i915_gem_object_pin_map(scratch->obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto out_end;
+	}
+
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		if (cs[n] != n) {
+			pr_err("Incorrect value[%d] found for GPR[%d]\n",
+			       cs[n], n);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	i915_gem_object_unpin_map(scratch->obj);
+
+out_end:
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	i915_request_put(last);
+out_unpin:
+	intel_context_unpin(ve);
+out_put:
+	intel_context_put(ve);
+out_scratch:
+	i915_vma_unpin_and_release(&scratch, 0);
+	return err;
+}
+
+static int live_virtual_preserved(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+
+	/*
+	 * Check that the context image retains non-privileged (user) registers
+	 * from one engine to the next. For this we check that the CS_GPR
+	 * are preserved.
+	 */
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	/* As we use CS_GPR we cannot run before they existed on all engines. */
+	if (INTEL_GEN(gt->i915) < 9)
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		int nsibling, err;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		err = preserved_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int bond_virtual_engine(struct intel_gt *gt,
+			       unsigned int class,
+			       struct intel_engine_cs **siblings,
+			       unsigned int nsibling,
+			       unsigned int flags)
+#define BOND_SCHEDULE BIT(0)
+{
+	struct intel_engine_cs *master;
+	struct i915_request *rq[16];
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	unsigned long n;
+	int err;
+
+	/*
+	 * A set of bonded requests is intended to be run concurrently
+	 * across a number of engines. We use one request per-engine
+	 * and a magic fence to schedule each of the bonded requests
+	 * at the same time. A consequence of our current scheduler is that
+	 * we only move requests to the HW ready queue when the request
+	 * becomes ready, that is when all of its prerequisite fences have
+	 * been signaled. As one of those fences is the master submit fence,
+	 * there is a delay on all secondary fences as the HW may be
+	 * currently busy. Equally, as all the requests are independent,
+	 * they may have other fences that delay individual request
+	 * submission to HW. Ergo, we do not guarantee that all requests are
+	 * immediately submitted to HW at the same time, just that if the
+	 * rules are abided by, they are ready at the same time as the
+	 * first is submitted. Userspace can embed semaphores in its batch
+	 * to ensure parallel execution of its phases as it requires.
+	 * Though naturally it gets requested that perhaps the scheduler should
+	 * take care of parallel execution, even across preemption events on
+	 * different HW. (The proper answer is of course "lalalala".)
+	 *
+	 * With the submit-fence, we have identified three possible phases
+	 * of synchronisation depending on the master fence: queued (not
+	 * ready), executing, and signaled. The first two are quite simple
+	 * and checked below. However, the signaled master fence handling is
+	 * contentious. Currently we do not distinguish between a signaled
+	 * fence and an expired fence, as once signaled it does not convey
+	 * any information about the previous execution. It may even be freed
+	 * and hence checking later it may not exist at all. Ergo we currently
+	 * do not apply the bonding constraint for an already signaled fence,
+	 * as our expectation is that it should not constrain the secondaries
+	 * and is outside of the scope of the bonded request API (i.e. all
+	 * userspace requests are meant to be running in parallel). As
+	 * it imposes no constraint, and is effectively a no-op, we do not
+	 * check below as normal execution flows are checked extensively above.
+	 *
+	 * XXX Is the degenerate handling of signaled submit fences the
+	 * expected behaviour for userpace?
+	 */
+
+	GEM_BUG_ON(nsibling >= ARRAY_SIZE(rq) - 1);
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	err = 0;
+	rq[0] = ERR_PTR(-ENOMEM);
+	for_each_engine(master, gt, id) {
+		struct i915_sw_fence fence = {};
+		struct intel_context *ce;
+
+		if (master->class == class)
+			continue;
+
+		ce = intel_context_create(master);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		memset_p((void *)rq, ERR_PTR(-EINVAL), ARRAY_SIZE(rq));
+
+		rq[0] = igt_spinner_create_request(&spin, ce, MI_NOOP);
+		intel_context_put(ce);
+		if (IS_ERR(rq[0])) {
+			err = PTR_ERR(rq[0]);
+			goto out;
+		}
+		i915_request_get(rq[0]);
+
+		if (flags & BOND_SCHEDULE) {
+			onstack_fence_init(&fence);
+			err = i915_sw_fence_await_sw_fence_gfp(&rq[0]->submit,
+							       &fence,
+							       GFP_KERNEL);
+		}
+
+		i915_request_add(rq[0]);
+		if (err < 0)
+			goto out;
+
+		if (!(flags & BOND_SCHEDULE) &&
+		    !igt_wait_for_spinner(&spin, rq[0])) {
+			err = -EIO;
+			goto out;
+		}
+
+		for (n = 0; n < nsibling; n++) {
+			struct intel_context *ve;
+
+			ve = intel_execlists_create_virtual(siblings, nsibling);
+			if (IS_ERR(ve)) {
+				err = PTR_ERR(ve);
+				onstack_fence_fini(&fence);
+				goto out;
+			}
+
+			err = intel_virtual_engine_attach_bond(ve->engine,
+							       master,
+							       siblings[n]);
+			if (err) {
+				intel_context_put(ve);
+				onstack_fence_fini(&fence);
+				goto out;
+			}
+
+			err = intel_context_pin(ve);
+			intel_context_put(ve);
+			if (err) {
+				onstack_fence_fini(&fence);
+				goto out;
+			}
+
+			rq[n + 1] = i915_request_create(ve);
+			intel_context_unpin(ve);
+			if (IS_ERR(rq[n + 1])) {
+				err = PTR_ERR(rq[n + 1]);
+				onstack_fence_fini(&fence);
+				goto out;
+			}
+			i915_request_get(rq[n + 1]);
+
+			err = i915_request_await_execution(rq[n + 1],
+							   &rq[0]->fence,
+							   ve->engine->bond_execute);
+			i915_request_add(rq[n + 1]);
+			if (err < 0) {
+				onstack_fence_fini(&fence);
+				goto out;
+			}
+		}
+		onstack_fence_fini(&fence);
+		intel_engine_flush_submission(master);
+		igt_spinner_end(&spin);
+
+		if (i915_request_wait(rq[0], 0, HZ / 10) < 0) {
+			pr_err("Master request did not execute (on %s)!\n",
+			       rq[0]->engine->name);
+			err = -EIO;
+			goto out;
+		}
+
+		for (n = 0; n < nsibling; n++) {
+			if (i915_request_wait(rq[n + 1], 0,
+					      MAX_SCHEDULE_TIMEOUT) < 0) {
+				err = -EIO;
+				goto out;
+			}
+
+			if (rq[n + 1]->engine != siblings[n]) {
+				pr_err("Bonded request did not execute on target engine: expected %s, used %s; master was %s\n",
+				       siblings[n]->name,
+				       rq[n + 1]->engine->name,
+				       rq[0]->engine->name);
+				err = -EINVAL;
+				goto out;
+			}
+		}
+
+		for (n = 0; !IS_ERR(rq[n]); n++)
+			i915_request_put(rq[n]);
+		rq[0] = ERR_PTR(-ENOMEM);
+	}
+
+out:
+	for (n = 0; !IS_ERR(rq[n]); n++)
+		i915_request_put(rq[n]);
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_virtual_bond(void *arg)
+{
+	static const struct phase {
+		const char *name;
+		unsigned int flags;
+	} phases[] = {
+		{ "", 0 },
+		{ "schedule", BOND_SCHEDULE },
+		{ },
+	};
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		const struct phase *p;
+		int nsibling;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		for (p = phases; p->name; p++) {
+			err = bond_virtual_engine(gt,
+						  class, siblings, nsibling,
+						  p->flags);
+			if (err) {
+				pr_err("%s(%s): failed class=%d, nsibling=%d, err=%d\n",
+				       __func__, p->name, class, nsibling, err);
+				return err;
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int reset_virtual_engine(struct intel_gt *gt,
+				struct intel_engine_cs **siblings,
+				unsigned int nsibling)
+{
+	struct intel_engine_cs *engine;
+	struct intel_context *ve;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	unsigned int n;
+	int err = 0;
+
+	/*
+	 * In order to support offline error capture for fast preempt reset,
+	 * we need to decouple the guilty request and ensure that it and its
+	 * descendents are not executed while the capture is in progress.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	ve = intel_execlists_create_virtual(siblings, nsibling);
+	if (IS_ERR(ve)) {
+		err = PTR_ERR(ve);
+		goto out_spin;
+	}
+
+	for (n = 0; n < nsibling; n++)
+		st_engine_heartbeat_disable(siblings[n]);
+
+	rq = igt_spinner_create_request(&spin, ve, MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_heartbeat;
+	}
+	i915_request_add(rq);
+
+	if (!igt_wait_for_spinner(&spin, rq)) {
+		intel_gt_set_wedged(gt);
+		err = -ETIME;
+		goto out_heartbeat;
+	}
+
+	engine = rq->engine;
+	GEM_BUG_ON(engine == ve->engine);
+
+	/* Take ownership of the reset and tasklet */
+	local_bh_disable();
+	if (test_and_set_bit(I915_RESET_ENGINE + engine->id,
+			     &gt->reset.flags)) {
+		local_bh_enable();
+		intel_gt_set_wedged(gt);
+		err = -EBUSY;
+		goto out_heartbeat;
+	}
+	tasklet_disable(&engine->execlists.tasklet);
+
+	engine->execlists.tasklet.func(engine->execlists.tasklet.data);
+	GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
+
+	/* Fake a preemption event; failed of course */
+	spin_lock_irq(&engine->active.lock);
+	__unwind_incomplete_requests(engine);
+	spin_unlock_irq(&engine->active.lock);
+	GEM_BUG_ON(rq->engine != engine);
+
+	/* Reset the engine while keeping our active request on hold */
+	execlists_hold(engine, rq);
+	GEM_BUG_ON(!i915_request_on_hold(rq));
+
+	__intel_engine_reset_bh(engine, NULL);
+	GEM_BUG_ON(rq->fence.error != -EIO);
+
+	/* Release our grasp on the engine, letting CS flow again */
+	tasklet_enable(&engine->execlists.tasklet);
+	clear_and_wake_up_bit(I915_RESET_ENGINE + engine->id, &gt->reset.flags);
+	local_bh_enable();
+
+	/* Check that we do not resubmit the held request */
+	i915_request_get(rq);
+	if (!i915_request_wait(rq, 0, HZ / 5)) {
+		pr_err("%s: on hold request completed!\n",
+		       engine->name);
+		intel_gt_set_wedged(gt);
+		err = -EIO;
+		goto out_rq;
+	}
+	GEM_BUG_ON(!i915_request_on_hold(rq));
+
+	/* But is resubmitted on release */
+	execlists_unhold(engine, rq);
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		pr_err("%s: held request did not complete!\n",
+		       engine->name);
+		intel_gt_set_wedged(gt);
+		err = -ETIME;
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_heartbeat:
+	for (n = 0; n < nsibling; n++)
+		st_engine_heartbeat_enable(siblings[n]);
+
+	intel_context_put(ve);
+out_spin:
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_virtual_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+
+	/*
+	 * Check that we handle a reset event within a virtual engine.
+	 * Only the physical engine is reset, but we have to check the flow
+	 * of the virtual requests around the reset, and make sure it is not
+	 * forgotten.
+	 */
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		int nsibling, err;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		err = reset_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int intel_execlists_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_sanitycheck),
+		SUBTEST(live_unlite_switch),
+		SUBTEST(live_unlite_preempt),
+		SUBTEST(live_unlite_ring),
+		SUBTEST(live_pin_rewind),
+		SUBTEST(live_hold_reset),
+		SUBTEST(live_error_interrupt),
+		SUBTEST(live_timeslice_preempt),
+		SUBTEST(live_timeslice_rewind),
+		SUBTEST(live_timeslice_queue),
+		SUBTEST(live_timeslice_nopreempt),
+		SUBTEST(live_busywait_preempt),
+		SUBTEST(live_preempt),
+		SUBTEST(live_late_preempt),
+		SUBTEST(live_nopreempt),
+		SUBTEST(live_preempt_cancel),
+		SUBTEST(live_suppress_self_preempt),
+		SUBTEST(live_chain_preempt),
+		SUBTEST(live_preempt_ring),
+		SUBTEST(live_preempt_gang),
+		SUBTEST(live_preempt_timeout),
+		SUBTEST(live_preempt_user),
+		SUBTEST(live_preempt_smoke),
+		SUBTEST(live_virtual_engine),
+		SUBTEST(live_virtual_mask),
+		SUBTEST(live_virtual_preserved),
+		SUBTEST(live_virtual_slice),
+		SUBTEST(live_virtual_bond),
+		SUBTEST(live_virtual_reset),
+	};
+
+	if (!HAS_EXECLISTS(i915))
+		return 0;
+
+	if (intel_gt_is_wedged(&i915->gt))
+		return 0;
+
+	return intel_gt_live_subtests(tests, &i915->gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_gt_pm.c b/drivers/gpu/drm/i915/gt/selftest_gt_pm.c
index 6180a47c1b51..5d911f724ebe 100644
--- a/drivers/gpu/drm/i915/gt/selftest_gt_pm.c
+++ b/drivers/gpu/drm/i915/gt/selftest_gt_pm.c
@@ -71,7 +71,7 @@ static int live_gt_clocks(void *arg)
 	enum intel_engine_id id;
 	int err = 0;
 
-	if (!RUNTIME_INFO(gt->i915)->cs_timestamp_frequency_hz) { /* unknown */
+	if (!gt->clock_frequency) { /* unknown */
 		pr_info("CS_TIMESTAMP frequency unknown\n");
 		return 0;
 	}
@@ -112,12 +112,12 @@ static int live_gt_clocks(void *arg)
 
 		measure_clocks(engine, &cycles, &dt);
 
-		time = i915_cs_timestamp_ticks_to_ns(engine->i915, cycles);
-		expected = i915_cs_timestamp_ns_to_ticks(engine->i915, dt);
+		time = intel_gt_clock_interval_to_ns(engine->gt, cycles);
+		expected = intel_gt_ns_to_clock_interval(engine->gt, dt);
 
 		pr_info("%s: TIMESTAMP %d cycles [%lldns] in %lldns [%d cycles], using CS clock frequency of %uKHz\n",
 			engine->name, cycles, time, dt, expected,
-			RUNTIME_INFO(engine->i915)->cs_timestamp_frequency_hz / 1000);
+			engine->gt->clock_frequency / 1000);
 
 		if (9 * time < 8 * dt || 8 * time > 9 * dt) {
 			pr_err("%s: CS ticks did not match walltime!\n",
diff --git a/drivers/gpu/drm/i915/gt/selftest_hangcheck.c b/drivers/gpu/drm/i915/gt/selftest_hangcheck.c
index fb5ebf930ab2..463bb6a700c8 100644
--- a/drivers/gpu/drm/i915/gt/selftest_hangcheck.c
+++ b/drivers/gpu/drm/i915/gt/selftest_hangcheck.c
@@ -506,7 +506,8 @@ static int igt_reset_nop_engine(void *arg)
 			}
 			err = intel_engine_reset(engine, NULL);
 			if (err) {
-				pr_err("i915_reset_engine failed\n");
+				pr_err("intel_engine_reset(%s) failed, err:%d\n",
+				       engine->name, err);
 				break;
 			}
 
@@ -539,6 +540,149 @@ static int igt_reset_nop_engine(void *arg)
 	return 0;
 }
 
+static void force_reset_timeout(struct intel_engine_cs *engine)
+{
+	engine->reset_timeout.probability = 999;
+	atomic_set(&engine->reset_timeout.times, -1);
+}
+
+static void cancel_reset_timeout(struct intel_engine_cs *engine)
+{
+	memset(&engine->reset_timeout, 0, sizeof(engine->reset_timeout));
+}
+
+static int igt_reset_fail_engine(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* Check that we can recover from engine-reset failues */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		unsigned int count;
+		struct intel_context *ce;
+		IGT_TIMEOUT(end_time);
+		int err;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce))
+			return PTR_ERR(ce);
+
+		st_engine_heartbeat_disable(engine);
+		set_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+
+		force_reset_timeout(engine);
+		err = intel_engine_reset(engine, NULL);
+		cancel_reset_timeout(engine);
+		if (err == 0) /* timeouts only generated on gen8+ */
+			goto skip;
+
+		count = 0;
+		do {
+			struct i915_request *last = NULL;
+			int i;
+
+			if (!wait_for_idle(engine)) {
+				pr_err("%s failed to idle before reset\n",
+				       engine->name);
+				err = -EIO;
+				break;
+			}
+
+			for (i = 0; i < count % 15; i++) {
+				struct i915_request *rq;
+
+				rq = intel_context_create_request(ce);
+				if (IS_ERR(rq)) {
+					struct drm_printer p =
+						drm_info_printer(gt->i915->drm.dev);
+					intel_engine_dump(engine, &p,
+							  "%s(%s): failed to submit request\n",
+							  __func__,
+							  engine->name);
+
+					GEM_TRACE("%s(%s): failed to submit request\n",
+						  __func__,
+						  engine->name);
+					GEM_TRACE_DUMP();
+
+					intel_gt_set_wedged(gt);
+					if (last)
+						i915_request_put(last);
+
+					err = PTR_ERR(rq);
+					goto out;
+				}
+
+				if (last)
+					i915_request_put(last);
+				last = i915_request_get(rq);
+				i915_request_add(rq);
+			}
+
+			if (count & 1) {
+				err = intel_engine_reset(engine, NULL);
+				if (err) {
+					GEM_TRACE_ERR("intel_engine_reset(%s) failed, err:%d\n",
+						      engine->name, err);
+					GEM_TRACE_DUMP();
+					i915_request_put(last);
+					break;
+				}
+			} else {
+				force_reset_timeout(engine);
+				err = intel_engine_reset(engine, NULL);
+				cancel_reset_timeout(engine);
+				if (err != -ETIMEDOUT) {
+					pr_err("intel_engine_reset(%s) did not fail, err:%d\n",
+					       engine->name, err);
+					i915_request_put(last);
+					break;
+				}
+			}
+
+			err = 0;
+			if (last) {
+				if (i915_request_wait(last, 0, HZ / 2) < 0) {
+					struct drm_printer p =
+						drm_info_printer(gt->i915->drm.dev);
+
+					intel_engine_dump(engine, &p,
+							  "%s(%s): failed to complete request\n",
+							  __func__,
+							  engine->name);
+
+					GEM_TRACE("%s(%s): failed to complete request\n",
+						  __func__,
+						  engine->name);
+					GEM_TRACE_DUMP();
+
+					err = -EIO;
+				}
+				i915_request_put(last);
+			}
+			count++;
+		} while (err == 0 && time_before(jiffies, end_time));
+out:
+		pr_info("%s(%s): %d resets\n", __func__, engine->name, count);
+skip:
+		clear_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+		st_engine_heartbeat_enable(engine);
+		intel_context_put(ce);
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
 static int __igt_reset_engine(struct intel_gt *gt, bool active)
 {
 	struct i915_gpu_error *global = &gt->i915->gpu_error;
@@ -560,6 +704,7 @@ static int __igt_reset_engine(struct intel_gt *gt, bool active)
 
 	for_each_engine(engine, gt, id) {
 		unsigned int reset_count, reset_engine_count;
+		unsigned long count;
 		IGT_TIMEOUT(end_time);
 
 		if (active && !intel_engine_can_store_dword(engine))
@@ -577,6 +722,7 @@ static int __igt_reset_engine(struct intel_gt *gt, bool active)
 
 		st_engine_heartbeat_disable(engine);
 		set_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+		count = 0;
 		do {
 			if (active) {
 				struct i915_request *rq;
@@ -608,7 +754,8 @@ static int __igt_reset_engine(struct intel_gt *gt, bool active)
 
 			err = intel_engine_reset(engine, NULL);
 			if (err) {
-				pr_err("i915_reset_engine failed\n");
+				pr_err("intel_engine_reset(%s) failed, err:%d\n",
+				       engine->name, err);
 				break;
 			}
 
@@ -625,9 +772,13 @@ static int __igt_reset_engine(struct intel_gt *gt, bool active)
 				err = -EINVAL;
 				break;
 			}
+
+			count++;
 		} while (time_before(jiffies, end_time));
 		clear_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
 		st_engine_heartbeat_enable(engine);
+		pr_info("%s: Completed %lu %s resets\n",
+			engine->name, count, active ? "active" : "idle");
 
 		if (err)
 			break;
@@ -1478,7 +1629,8 @@ static int igt_reset_queue(void *arg)
 			prev = rq;
 			count++;
 		} while (time_before(jiffies, end_time));
-		pr_info("%s: Completed %d resets\n", engine->name, count);
+		pr_info("%s: Completed %d queued resets\n",
+			engine->name, count);
 
 		*h.batch = MI_BATCH_BUFFER_END;
 		intel_gt_chipset_flush(engine->gt);
@@ -1575,13 +1727,21 @@ static int __igt_atomic_reset_engine(struct intel_engine_cs *engine,
 	GEM_TRACE("i915_reset_engine(%s:%s) under %s\n",
 		  engine->name, mode, p->name);
 
-	tasklet_disable(t);
+	if (t->func)
+		tasklet_disable(t);
+	if (strcmp(p->name, "softirq"))
+		local_bh_disable();
 	p->critical_section_begin();
 
-	err = intel_engine_reset(engine, NULL);
+	err = __intel_engine_reset_bh(engine, NULL);
 
 	p->critical_section_end();
-	tasklet_enable(t);
+	if (strcmp(p->name, "softirq"))
+		local_bh_enable();
+	if (t->func) {
+		tasklet_enable(t);
+		tasklet_hi_schedule(t);
+	}
 
 	if (err)
 		pr_err("i915_reset_engine(%s:%s) failed under %s\n",
@@ -1687,6 +1847,7 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915)
 		SUBTEST(igt_reset_nop_engine),
 		SUBTEST(igt_reset_idle_engine),
 		SUBTEST(igt_reset_active_engine),
+		SUBTEST(igt_reset_fail_engine),
 		SUBTEST(igt_reset_engines),
 		SUBTEST(igt_reset_engines_atomic),
 		SUBTEST(igt_reset_queue),
diff --git a/drivers/gpu/drm/i915/gt/selftest_lrc.c b/drivers/gpu/drm/i915/gt/selftest_lrc.c
index 95d41c01d0e0..920979a89413 100644
--- a/drivers/gpu/drm/i915/gt/selftest_lrc.c
+++ b/drivers/gpu/drm/i915/gt/selftest_lrc.c
@@ -1,22 +1,22 @@
+// SPDX-License-Identifier: MIT
 /*
- * SPDX-License-Identifier: MIT
- *
  * Copyright © 2018 Intel Corporation
  */
 
 #include <linux/prime_numbers.h>
 
-#include "gem/i915_gem_pm.h"
-#include "gt/intel_engine_heartbeat.h"
-#include "gt/intel_reset.h"
-#include "gt/selftest_engine_heartbeat.h"
-
 #include "i915_selftest.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_reset.h"
+#include "intel_ring.h"
+#include "selftest_engine_heartbeat.h"
 #include "selftests/i915_random.h"
 #include "selftests/igt_flush_test.h"
 #include "selftests/igt_live_test.h"
 #include "selftests/igt_spinner.h"
 #include "selftests/lib_sw_fence.h"
+#include "shmem_utils.h"
 
 #include "gem/selftests/igt_gem_utils.h"
 #include "gem/selftests/mock_context.h"
@@ -27,29 +27,7 @@
 
 static struct i915_vma *create_scratch(struct intel_gt *gt)
 {
-	struct drm_i915_gem_object *obj;
-	struct i915_vma *vma;
-	int err;
-
-	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
-	if (IS_ERR(obj))
-		return ERR_CAST(obj);
-
-	i915_gem_object_set_cache_coherency(obj, I915_CACHING_CACHED);
-
-	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		i915_gem_object_put(obj);
-		return vma;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
-	if (err) {
-		i915_gem_object_put(obj);
-		return ERR_PTR(err);
-	}
-
-	return vma;
+	return __vm_create_scratch_for_read(&gt->ggtt->vm, PAGE_SIZE);
 }
 
 static bool is_active(struct i915_request *rq)
@@ -70,6 +48,9 @@ static int wait_for_submit(struct intel_engine_cs *engine,
 			   struct i915_request *rq,
 			   unsigned long timeout)
 {
+	/* Ignore our own attempts to suppress excess tasklets */
+	tasklet_hi_schedule(&engine->execlists.tasklet);
+
 	timeout += jiffies;
 	do {
 		bool done = time_after(jiffies, timeout);
@@ -89,4623 +70,6 @@ static int wait_for_submit(struct intel_engine_cs *engine,
 	} while (1);
 }
 
-static int wait_for_reset(struct intel_engine_cs *engine,
-			  struct i915_request *rq,
-			  unsigned long timeout)
-{
-	timeout += jiffies;
-
-	do {
-		cond_resched();
-		intel_engine_flush_submission(engine);
-
-		if (READ_ONCE(engine->execlists.pending[0]))
-			continue;
-
-		if (i915_request_completed(rq))
-			break;
-
-		if (READ_ONCE(rq->fence.error))
-			break;
-	} while (time_before(jiffies, timeout));
-
-	flush_scheduled_work();
-
-	if (rq->fence.error != -EIO) {
-		pr_err("%s: hanging request %llx:%lld not reset\n",
-		       engine->name,
-		       rq->fence.context,
-		       rq->fence.seqno);
-		return -EINVAL;
-	}
-
-	/* Give the request a jiffie to complete after flushing the worker */
-	if (i915_request_wait(rq, 0,
-			      max(0l, (long)(timeout - jiffies)) + 1) < 0) {
-		pr_err("%s: hanging request %llx:%lld did not complete\n",
-		       engine->name,
-		       rq->fence.context,
-		       rq->fence.seqno);
-		return -ETIME;
-	}
-
-	return 0;
-}
-
-static int live_sanitycheck(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	struct igt_spinner spin;
-	int err = 0;
-
-	if (!HAS_LOGICAL_RING_CONTEXTS(gt->i915))
-		return 0;
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	for_each_engine(engine, gt, id) {
-		struct intel_context *ce;
-		struct i915_request *rq;
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			break;
-		}
-
-		rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out_ctx;
-		}
-
-		i915_request_add(rq);
-		if (!igt_wait_for_spinner(&spin, rq)) {
-			GEM_TRACE("spinner failed to start\n");
-			GEM_TRACE_DUMP();
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-			goto out_ctx;
-		}
-
-		igt_spinner_end(&spin);
-		if (igt_flush_test(gt->i915)) {
-			err = -EIO;
-			goto out_ctx;
-		}
-
-out_ctx:
-		intel_context_put(ce);
-		if (err)
-			break;
-	}
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_unlite_restore(struct intel_gt *gt, int prio)
-{
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	struct igt_spinner spin;
-	int err = -ENOMEM;
-
-	/*
-	 * Check that we can correctly context switch between 2 instances
-	 * on the same engine from the same parent context.
-	 */
-
-	if (igt_spinner_init(&spin, gt))
-		return err;
-
-	err = 0;
-	for_each_engine(engine, gt, id) {
-		struct intel_context *ce[2] = {};
-		struct i915_request *rq[2];
-		struct igt_live_test t;
-		int n;
-
-		if (prio && !intel_engine_has_preemption(engine))
-			continue;
-
-		if (!intel_engine_can_store_dword(engine))
-			continue;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			break;
-		}
-		st_engine_heartbeat_disable(engine);
-
-		for (n = 0; n < ARRAY_SIZE(ce); n++) {
-			struct intel_context *tmp;
-
-			tmp = intel_context_create(engine);
-			if (IS_ERR(tmp)) {
-				err = PTR_ERR(tmp);
-				goto err_ce;
-			}
-
-			err = intel_context_pin(tmp);
-			if (err) {
-				intel_context_put(tmp);
-				goto err_ce;
-			}
-
-			/*
-			 * Setup the pair of contexts such that if we
-			 * lite-restore using the RING_TAIL from ce[1] it
-			 * will execute garbage from ce[0]->ring.
-			 */
-			memset(tmp->ring->vaddr,
-			       POISON_INUSE, /* IPEHR: 0x5a5a5a5a [hung!] */
-			       tmp->ring->vma->size);
-
-			ce[n] = tmp;
-		}
-		GEM_BUG_ON(!ce[1]->ring->size);
-		intel_ring_reset(ce[1]->ring, ce[1]->ring->size / 2);
-		__execlists_update_reg_state(ce[1], engine, ce[1]->ring->head);
-
-		rq[0] = igt_spinner_create_request(&spin, ce[0], MI_ARB_CHECK);
-		if (IS_ERR(rq[0])) {
-			err = PTR_ERR(rq[0]);
-			goto err_ce;
-		}
-
-		i915_request_get(rq[0]);
-		i915_request_add(rq[0]);
-		GEM_BUG_ON(rq[0]->postfix > ce[1]->ring->emit);
-
-		if (!igt_wait_for_spinner(&spin, rq[0])) {
-			i915_request_put(rq[0]);
-			goto err_ce;
-		}
-
-		rq[1] = i915_request_create(ce[1]);
-		if (IS_ERR(rq[1])) {
-			err = PTR_ERR(rq[1]);
-			i915_request_put(rq[0]);
-			goto err_ce;
-		}
-
-		if (!prio) {
-			/*
-			 * Ensure we do the switch to ce[1] on completion.
-			 *
-			 * rq[0] is already submitted, so this should reduce
-			 * to a no-op (a wait on a request on the same engine
-			 * uses the submit fence, not the completion fence),
-			 * but it will install a dependency on rq[1] for rq[0]
-			 * that will prevent the pair being reordered by
-			 * timeslicing.
-			 */
-			i915_request_await_dma_fence(rq[1], &rq[0]->fence);
-		}
-
-		i915_request_get(rq[1]);
-		i915_request_add(rq[1]);
-		GEM_BUG_ON(rq[1]->postfix <= rq[0]->postfix);
-		i915_request_put(rq[0]);
-
-		if (prio) {
-			struct i915_sched_attr attr = {
-				.priority = prio,
-			};
-
-			/* Alternatively preempt the spinner with ce[1] */
-			engine->schedule(rq[1], &attr);
-		}
-
-		/* And switch back to ce[0] for good measure */
-		rq[0] = i915_request_create(ce[0]);
-		if (IS_ERR(rq[0])) {
-			err = PTR_ERR(rq[0]);
-			i915_request_put(rq[1]);
-			goto err_ce;
-		}
-
-		i915_request_await_dma_fence(rq[0], &rq[1]->fence);
-		i915_request_get(rq[0]);
-		i915_request_add(rq[0]);
-		GEM_BUG_ON(rq[0]->postfix > rq[1]->postfix);
-		i915_request_put(rq[1]);
-		i915_request_put(rq[0]);
-
-err_ce:
-		intel_engine_flush_submission(engine);
-		igt_spinner_end(&spin);
-		for (n = 0; n < ARRAY_SIZE(ce); n++) {
-			if (IS_ERR_OR_NULL(ce[n]))
-				break;
-
-			intel_context_unpin(ce[n]);
-			intel_context_put(ce[n]);
-		}
-
-		st_engine_heartbeat_enable(engine);
-		if (igt_live_test_end(&t))
-			err = -EIO;
-		if (err)
-			break;
-	}
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_unlite_switch(void *arg)
-{
-	return live_unlite_restore(arg, 0);
-}
-
-static int live_unlite_preempt(void *arg)
-{
-	return live_unlite_restore(arg, I915_USER_PRIORITY(I915_PRIORITY_MAX));
-}
-
-static int live_unlite_ring(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	struct igt_spinner spin;
-	enum intel_engine_id id;
-	int err = 0;
-
-	/*
-	 * Setup a preemption event that will cause almost the entire ring
-	 * to be unwound, potentially fooling our intel_ring_direction()
-	 * into emitting a forward lite-restore instead of the rollback.
-	 */
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	for_each_engine(engine, gt, id) {
-		struct intel_context *ce[2] = {};
-		struct i915_request *rq;
-		struct igt_live_test t;
-		int n;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (!intel_engine_can_store_dword(engine))
-			continue;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			break;
-		}
-		st_engine_heartbeat_disable(engine);
-
-		for (n = 0; n < ARRAY_SIZE(ce); n++) {
-			struct intel_context *tmp;
-
-			tmp = intel_context_create(engine);
-			if (IS_ERR(tmp)) {
-				err = PTR_ERR(tmp);
-				goto err_ce;
-			}
-
-			err = intel_context_pin(tmp);
-			if (err) {
-				intel_context_put(tmp);
-				goto err_ce;
-			}
-
-			memset32(tmp->ring->vaddr,
-				 0xdeadbeef, /* trigger a hang if executed */
-				 tmp->ring->vma->size / sizeof(u32));
-
-			ce[n] = tmp;
-		}
-
-		/* Create max prio spinner, followed by N low prio nops */
-		rq = igt_spinner_create_request(&spin, ce[0], MI_ARB_CHECK);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto err_ce;
-		}
-
-		i915_request_get(rq);
-		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
-		i915_request_add(rq);
-
-		if (!igt_wait_for_spinner(&spin, rq)) {
-			intel_gt_set_wedged(gt);
-			i915_request_put(rq);
-			err = -ETIME;
-			goto err_ce;
-		}
-
-		/* Fill the ring, until we will cause a wrap */
-		n = 0;
-		while (intel_ring_direction(ce[0]->ring,
-					    rq->wa_tail,
-					    ce[0]->ring->tail) <= 0) {
-			struct i915_request *tmp;
-
-			tmp = intel_context_create_request(ce[0]);
-			if (IS_ERR(tmp)) {
-				err = PTR_ERR(tmp);
-				i915_request_put(rq);
-				goto err_ce;
-			}
-
-			i915_request_add(tmp);
-			intel_engine_flush_submission(engine);
-			n++;
-		}
-		intel_engine_flush_submission(engine);
-		pr_debug("%s: Filled ring with %d nop tails {size:%x, tail:%x, emit:%x, rq.tail:%x}\n",
-			 engine->name, n,
-			 ce[0]->ring->size,
-			 ce[0]->ring->tail,
-			 ce[0]->ring->emit,
-			 rq->tail);
-		GEM_BUG_ON(intel_ring_direction(ce[0]->ring,
-						rq->tail,
-						ce[0]->ring->tail) <= 0);
-		i915_request_put(rq);
-
-		/* Create a second ring to preempt the first ring after rq[0] */
-		rq = intel_context_create_request(ce[1]);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto err_ce;
-		}
-
-		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
-		i915_request_get(rq);
-		i915_request_add(rq);
-
-		err = wait_for_submit(engine, rq, HZ / 2);
-		i915_request_put(rq);
-		if (err) {
-			pr_err("%s: preemption request was not submitted\n",
-			       engine->name);
-			err = -ETIME;
-		}
-
-		pr_debug("%s: ring[0]:{ tail:%x, emit:%x }, ring[1]:{ tail:%x, emit:%x }\n",
-			 engine->name,
-			 ce[0]->ring->tail, ce[0]->ring->emit,
-			 ce[1]->ring->tail, ce[1]->ring->emit);
-
-err_ce:
-		intel_engine_flush_submission(engine);
-		igt_spinner_end(&spin);
-		for (n = 0; n < ARRAY_SIZE(ce); n++) {
-			if (IS_ERR_OR_NULL(ce[n]))
-				break;
-
-			intel_context_unpin(ce[n]);
-			intel_context_put(ce[n]);
-		}
-		st_engine_heartbeat_enable(engine);
-		if (igt_live_test_end(&t))
-			err = -EIO;
-		if (err)
-			break;
-	}
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_pin_rewind(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	int err = 0;
-
-	/*
-	 * We have to be careful not to trust intel_ring too much, for example
-	 * ring->head is updated upon retire which is out of sync with pinning
-	 * the context. Thus we cannot use ring->head to set CTX_RING_HEAD,
-	 * or else we risk writing an older, stale value.
-	 *
-	 * To simulate this, let's apply a bit of deliberate sabotague.
-	 */
-
-	for_each_engine(engine, gt, id) {
-		struct intel_context *ce;
-		struct i915_request *rq;
-		struct intel_ring *ring;
-		struct igt_live_test t;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			break;
-		}
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			break;
-		}
-
-		err = intel_context_pin(ce);
-		if (err) {
-			intel_context_put(ce);
-			break;
-		}
-
-		/* Keep the context awake while we play games */
-		err = i915_active_acquire(&ce->active);
-		if (err) {
-			intel_context_unpin(ce);
-			intel_context_put(ce);
-			break;
-		}
-		ring = ce->ring;
-
-		/* Poison the ring, and offset the next request from HEAD */
-		memset32(ring->vaddr, STACK_MAGIC, ring->size / sizeof(u32));
-		ring->emit = ring->size / 2;
-		ring->tail = ring->emit;
-		GEM_BUG_ON(ring->head);
-
-		intel_context_unpin(ce);
-
-		/* Submit a simple nop request */
-		GEM_BUG_ON(intel_context_is_pinned(ce));
-		rq = intel_context_create_request(ce);
-		i915_active_release(&ce->active); /* e.g. async retire */
-		intel_context_put(ce);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			break;
-		}
-		GEM_BUG_ON(!rq->head);
-		i915_request_add(rq);
-
-		/* Expect not to hang! */
-		if (igt_live_test_end(&t)) {
-			err = -EIO;
-			break;
-		}
-	}
-
-	return err;
-}
-
-static int live_hold_reset(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	struct igt_spinner spin;
-	int err = 0;
-
-	/*
-	 * In order to support offline error capture for fast preempt reset,
-	 * we need to decouple the guilty request and ensure that it and its
-	 * descendents are not executed while the capture is in progress.
-	 */
-
-	if (!intel_has_reset_engine(gt))
-		return 0;
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	for_each_engine(engine, gt, id) {
-		struct intel_context *ce;
-		struct i915_request *rq;
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			break;
-		}
-
-		st_engine_heartbeat_disable(engine);
-
-		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out;
-		}
-		i915_request_add(rq);
-
-		if (!igt_wait_for_spinner(&spin, rq)) {
-			intel_gt_set_wedged(gt);
-			err = -ETIME;
-			goto out;
-		}
-
-		/* We have our request executing, now remove it and reset */
-
-		if (test_and_set_bit(I915_RESET_ENGINE + id,
-				     &gt->reset.flags)) {
-			intel_gt_set_wedged(gt);
-			err = -EBUSY;
-			goto out;
-		}
-		tasklet_disable(&engine->execlists.tasklet);
-
-		engine->execlists.tasklet.func(engine->execlists.tasklet.data);
-		GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
-
-		i915_request_get(rq);
-		execlists_hold(engine, rq);
-		GEM_BUG_ON(!i915_request_on_hold(rq));
-
-		intel_engine_reset(engine, NULL);
-		GEM_BUG_ON(rq->fence.error != -EIO);
-
-		tasklet_enable(&engine->execlists.tasklet);
-		clear_and_wake_up_bit(I915_RESET_ENGINE + id,
-				      &gt->reset.flags);
-
-		/* Check that we do not resubmit the held request */
-		if (!i915_request_wait(rq, 0, HZ / 5)) {
-			pr_err("%s: on hold request completed!\n",
-			       engine->name);
-			i915_request_put(rq);
-			err = -EIO;
-			goto out;
-		}
-		GEM_BUG_ON(!i915_request_on_hold(rq));
-
-		/* But is resubmitted on release */
-		execlists_unhold(engine, rq);
-		if (i915_request_wait(rq, 0, HZ / 5) < 0) {
-			pr_err("%s: held request did not complete!\n",
-			       engine->name);
-			intel_gt_set_wedged(gt);
-			err = -ETIME;
-		}
-		i915_request_put(rq);
-
-out:
-		st_engine_heartbeat_enable(engine);
-		intel_context_put(ce);
-		if (err)
-			break;
-	}
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static const char *error_repr(int err)
-{
-	return err ? "bad" : "good";
-}
-
-static int live_error_interrupt(void *arg)
-{
-	static const struct error_phase {
-		enum { GOOD = 0, BAD = -EIO } error[2];
-	} phases[] = {
-		{ { BAD,  GOOD } },
-		{ { BAD,  BAD  } },
-		{ { BAD,  GOOD } },
-		{ { GOOD, GOOD } }, /* sentinel */
-	};
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-
-	/*
-	 * We hook up the CS_MASTER_ERROR_INTERRUPT to have forewarning
-	 * of invalid commands in user batches that will cause a GPU hang.
-	 * This is a faster mechanism than using hangcheck/heartbeats, but
-	 * only detects problems the HW knows about -- it will not warn when
-	 * we kill the HW!
-	 *
-	 * To verify our detection and reset, we throw some invalid commands
-	 * at the HW and wait for the interrupt.
-	 */
-
-	if (!intel_has_reset_engine(gt))
-		return 0;
-
-	for_each_engine(engine, gt, id) {
-		const struct error_phase *p;
-		int err = 0;
-
-		st_engine_heartbeat_disable(engine);
-
-		for (p = phases; p->error[0] != GOOD; p++) {
-			struct i915_request *client[ARRAY_SIZE(phases->error)];
-			u32 *cs;
-			int i;
-
-			memset(client, 0, sizeof(*client));
-			for (i = 0; i < ARRAY_SIZE(client); i++) {
-				struct intel_context *ce;
-				struct i915_request *rq;
-
-				ce = intel_context_create(engine);
-				if (IS_ERR(ce)) {
-					err = PTR_ERR(ce);
-					goto out;
-				}
-
-				rq = intel_context_create_request(ce);
-				intel_context_put(ce);
-				if (IS_ERR(rq)) {
-					err = PTR_ERR(rq);
-					goto out;
-				}
-
-				if (rq->engine->emit_init_breadcrumb) {
-					err = rq->engine->emit_init_breadcrumb(rq);
-					if (err) {
-						i915_request_add(rq);
-						goto out;
-					}
-				}
-
-				cs = intel_ring_begin(rq, 2);
-				if (IS_ERR(cs)) {
-					i915_request_add(rq);
-					err = PTR_ERR(cs);
-					goto out;
-				}
-
-				if (p->error[i]) {
-					*cs++ = 0xdeadbeef;
-					*cs++ = 0xdeadbeef;
-				} else {
-					*cs++ = MI_NOOP;
-					*cs++ = MI_NOOP;
-				}
-
-				client[i] = i915_request_get(rq);
-				i915_request_add(rq);
-			}
-
-			err = wait_for_submit(engine, client[0], HZ / 2);
-			if (err) {
-				pr_err("%s: first request did not start within time!\n",
-				       engine->name);
-				err = -ETIME;
-				goto out;
-			}
-
-			for (i = 0; i < ARRAY_SIZE(client); i++) {
-				if (i915_request_wait(client[i], 0, HZ / 5) < 0)
-					pr_debug("%s: %s request incomplete!\n",
-						 engine->name,
-						 error_repr(p->error[i]));
-
-				if (!i915_request_started(client[i])) {
-					pr_err("%s: %s request not started!\n",
-					       engine->name,
-					       error_repr(p->error[i]));
-					err = -ETIME;
-					goto out;
-				}
-
-				/* Kick the tasklet to process the error */
-				intel_engine_flush_submission(engine);
-				if (client[i]->fence.error != p->error[i]) {
-					pr_err("%s: %s request (%s) with wrong error code: %d\n",
-					       engine->name,
-					       error_repr(p->error[i]),
-					       i915_request_completed(client[i]) ? "completed" : "running",
-					       client[i]->fence.error);
-					err = -EINVAL;
-					goto out;
-				}
-			}
-
-out:
-			for (i = 0; i < ARRAY_SIZE(client); i++)
-				if (client[i])
-					i915_request_put(client[i]);
-			if (err) {
-				pr_err("%s: failed at phase[%zd] { %d, %d }\n",
-				       engine->name, p - phases,
-				       p->error[0], p->error[1]);
-				break;
-			}
-		}
-
-		st_engine_heartbeat_enable(engine);
-		if (err) {
-			intel_gt_set_wedged(gt);
-			return err;
-		}
-	}
-
-	return 0;
-}
-
-static int
-emit_semaphore_chain(struct i915_request *rq, struct i915_vma *vma, int idx)
-{
-	u32 *cs;
-
-	cs = intel_ring_begin(rq, 10);
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
-
-	*cs++ = MI_SEMAPHORE_WAIT |
-		MI_SEMAPHORE_GLOBAL_GTT |
-		MI_SEMAPHORE_POLL |
-		MI_SEMAPHORE_SAD_NEQ_SDD;
-	*cs++ = 0;
-	*cs++ = i915_ggtt_offset(vma) + 4 * idx;
-	*cs++ = 0;
-
-	if (idx > 0) {
-		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-		*cs++ = i915_ggtt_offset(vma) + 4 * (idx - 1);
-		*cs++ = 0;
-		*cs++ = 1;
-	} else {
-		*cs++ = MI_NOOP;
-		*cs++ = MI_NOOP;
-		*cs++ = MI_NOOP;
-		*cs++ = MI_NOOP;
-	}
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
-
-	intel_ring_advance(rq, cs);
-	return 0;
-}
-
-static struct i915_request *
-semaphore_queue(struct intel_engine_cs *engine, struct i915_vma *vma, int idx)
-{
-	struct intel_context *ce;
-	struct i915_request *rq;
-	int err;
-
-	ce = intel_context_create(engine);
-	if (IS_ERR(ce))
-		return ERR_CAST(ce);
-
-	rq = intel_context_create_request(ce);
-	if (IS_ERR(rq))
-		goto out_ce;
-
-	err = 0;
-	if (rq->engine->emit_init_breadcrumb)
-		err = rq->engine->emit_init_breadcrumb(rq);
-	if (err == 0)
-		err = emit_semaphore_chain(rq, vma, idx);
-	if (err == 0)
-		i915_request_get(rq);
-	i915_request_add(rq);
-	if (err)
-		rq = ERR_PTR(err);
-
-out_ce:
-	intel_context_put(ce);
-	return rq;
-}
-
-static int
-release_queue(struct intel_engine_cs *engine,
-	      struct i915_vma *vma,
-	      int idx, int prio)
-{
-	struct i915_sched_attr attr = {
-		.priority = prio,
-	};
-	struct i915_request *rq;
-	u32 *cs;
-
-	rq = intel_engine_create_kernel_request(engine);
-	if (IS_ERR(rq))
-		return PTR_ERR(rq);
-
-	cs = intel_ring_begin(rq, 4);
-	if (IS_ERR(cs)) {
-		i915_request_add(rq);
-		return PTR_ERR(cs);
-	}
-
-	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-	*cs++ = i915_ggtt_offset(vma) + 4 * (idx - 1);
-	*cs++ = 0;
-	*cs++ = 1;
-
-	intel_ring_advance(rq, cs);
-
-	i915_request_get(rq);
-	i915_request_add(rq);
-
-	local_bh_disable();
-	engine->schedule(rq, &attr);
-	local_bh_enable(); /* kick tasklet */
-
-	i915_request_put(rq);
-
-	return 0;
-}
-
-static int
-slice_semaphore_queue(struct intel_engine_cs *outer,
-		      struct i915_vma *vma,
-		      int count)
-{
-	struct intel_engine_cs *engine;
-	struct i915_request *head;
-	enum intel_engine_id id;
-	int err, i, n = 0;
-
-	head = semaphore_queue(outer, vma, n++);
-	if (IS_ERR(head))
-		return PTR_ERR(head);
-
-	for_each_engine(engine, outer->gt, id) {
-		for (i = 0; i < count; i++) {
-			struct i915_request *rq;
-
-			rq = semaphore_queue(engine, vma, n++);
-			if (IS_ERR(rq)) {
-				err = PTR_ERR(rq);
-				goto out;
-			}
-
-			i915_request_put(rq);
-		}
-	}
-
-	err = release_queue(outer, vma, n, I915_PRIORITY_BARRIER);
-	if (err)
-		goto out;
-
-	if (i915_request_wait(head, 0,
-			      2 * outer->gt->info.num_engines * (count + 2) * (count + 3)) < 0) {
-		pr_err("Failed to slice along semaphore chain of length (%d, %d)!\n",
-		       count, n);
-		GEM_TRACE_DUMP();
-		intel_gt_set_wedged(outer->gt);
-		err = -EIO;
-	}
-
-out:
-	i915_request_put(head);
-	return err;
-}
-
-static int live_timeslice_preempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct drm_i915_gem_object *obj;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	struct i915_vma *vma;
-	void *vaddr;
-	int err = 0;
-
-	/*
-	 * If a request takes too long, we would like to give other users
-	 * a fair go on the GPU. In particular, users may create batches
-	 * that wait upon external input, where that input may even be
-	 * supplied by another GPU job. To avoid blocking forever, we
-	 * need to preempt the current task and replace it with another
-	 * ready task.
-	 */
-	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
-		return 0;
-
-	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
-	if (IS_ERR(obj))
-		return PTR_ERR(obj);
-
-	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		err = PTR_ERR(vma);
-		goto err_obj;
-	}
-
-	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WC);
-	if (IS_ERR(vaddr)) {
-		err = PTR_ERR(vaddr);
-		goto err_obj;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
-	if (err)
-		goto err_map;
-
-	err = i915_vma_sync(vma);
-	if (err)
-		goto err_pin;
-
-	for_each_engine(engine, gt, id) {
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		memset(vaddr, 0, PAGE_SIZE);
-
-		st_engine_heartbeat_disable(engine);
-		err = slice_semaphore_queue(engine, vma, 5);
-		st_engine_heartbeat_enable(engine);
-		if (err)
-			goto err_pin;
-
-		if (igt_flush_test(gt->i915)) {
-			err = -EIO;
-			goto err_pin;
-		}
-	}
-
-err_pin:
-	i915_vma_unpin(vma);
-err_map:
-	i915_gem_object_unpin_map(obj);
-err_obj:
-	i915_gem_object_put(obj);
-	return err;
-}
-
-static struct i915_request *
-create_rewinder(struct intel_context *ce,
-		struct i915_request *wait,
-		void *slot, int idx)
-{
-	const u32 offset =
-		i915_ggtt_offset(ce->engine->status_page.vma) +
-		offset_in_page(slot);
-	struct i915_request *rq;
-	u32 *cs;
-	int err;
-
-	rq = intel_context_create_request(ce);
-	if (IS_ERR(rq))
-		return rq;
-
-	if (wait) {
-		err = i915_request_await_dma_fence(rq, &wait->fence);
-		if (err)
-			goto err;
-	}
-
-	cs = intel_ring_begin(rq, 14);
-	if (IS_ERR(cs)) {
-		err = PTR_ERR(cs);
-		goto err;
-	}
-
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
-	*cs++ = MI_NOOP;
-
-	*cs++ = MI_SEMAPHORE_WAIT |
-		MI_SEMAPHORE_GLOBAL_GTT |
-		MI_SEMAPHORE_POLL |
-		MI_SEMAPHORE_SAD_GTE_SDD;
-	*cs++ = idx;
-	*cs++ = offset;
-	*cs++ = 0;
-
-	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
-	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(rq->engine->mmio_base));
-	*cs++ = offset + idx * sizeof(u32);
-	*cs++ = 0;
-
-	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-	*cs++ = offset;
-	*cs++ = 0;
-	*cs++ = idx + 1;
-
-	intel_ring_advance(rq, cs);
-
-	rq->sched.attr.priority = I915_PRIORITY_MASK;
-	err = 0;
-err:
-	i915_request_get(rq);
-	i915_request_add(rq);
-	if (err) {
-		i915_request_put(rq);
-		return ERR_PTR(err);
-	}
-
-	return rq;
-}
-
-static int live_timeslice_rewind(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-
-	/*
-	 * The usual presumption on timeslice expiration is that we replace
-	 * the active context with another. However, given a chain of
-	 * dependencies we may end up with replacing the context with itself,
-	 * but only a few of those requests, forcing us to rewind the
-	 * RING_TAIL of the original request.
-	 */
-	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
-		return 0;
-
-	for_each_engine(engine, gt, id) {
-		enum { A1, A2, B1 };
-		enum { X = 1, Z, Y };
-		struct i915_request *rq[3] = {};
-		struct intel_context *ce;
-		unsigned long timeslice;
-		int i, err = 0;
-		u32 *slot;
-
-		if (!intel_engine_has_timeslices(engine))
-			continue;
-
-		/*
-		 * A:rq1 -- semaphore wait, timestamp X
-		 * A:rq2 -- write timestamp Y
-		 *
-		 * B:rq1 [await A:rq1] -- write timestamp Z
-		 *
-		 * Force timeslice, release semaphore.
-		 *
-		 * Expect execution/evaluation order XZY
-		 */
-
-		st_engine_heartbeat_disable(engine);
-		timeslice = xchg(&engine->props.timeslice_duration_ms, 1);
-
-		slot = memset32(engine->status_page.addr + 1000, 0, 4);
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto err;
-		}
-
-		rq[A1] = create_rewinder(ce, NULL, slot, X);
-		if (IS_ERR(rq[A1])) {
-			intel_context_put(ce);
-			goto err;
-		}
-
-		rq[A2] = create_rewinder(ce, NULL, slot, Y);
-		intel_context_put(ce);
-		if (IS_ERR(rq[A2]))
-			goto err;
-
-		err = wait_for_submit(engine, rq[A2], HZ / 2);
-		if (err) {
-			pr_err("%s: failed to submit first context\n",
-			       engine->name);
-			goto err;
-		}
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto err;
-		}
-
-		rq[B1] = create_rewinder(ce, rq[A1], slot, Z);
-		intel_context_put(ce);
-		if (IS_ERR(rq[2]))
-			goto err;
-
-		err = wait_for_submit(engine, rq[B1], HZ / 2);
-		if (err) {
-			pr_err("%s: failed to submit second context\n",
-			       engine->name);
-			goto err;
-		}
-
-		/* ELSP[] = { { A:rq1, A:rq2 }, { B:rq1 } } */
-		ENGINE_TRACE(engine, "forcing tasklet for rewind\n");
-		if (i915_request_is_active(rq[A2])) { /* semaphore yielded! */
-			/* Wait for the timeslice to kick in */
-			del_timer(&engine->execlists.timer);
-			tasklet_hi_schedule(&engine->execlists.tasklet);
-			intel_engine_flush_submission(engine);
-		}
-		/* -> ELSP[] = { { A:rq1 }, { B:rq1 } } */
-		GEM_BUG_ON(!i915_request_is_active(rq[A1]));
-		GEM_BUG_ON(!i915_request_is_active(rq[B1]));
-		GEM_BUG_ON(i915_request_is_active(rq[A2]));
-
-		/* Release the hounds! */
-		slot[0] = 1;
-		wmb(); /* "pairs" with GPU; paranoid kick of internal CPU$ */
-
-		for (i = 1; i <= 3; i++) {
-			unsigned long timeout = jiffies + HZ / 2;
-
-			while (!READ_ONCE(slot[i]) &&
-			       time_before(jiffies, timeout))
-				;
-
-			if (!time_before(jiffies, timeout)) {
-				pr_err("%s: rq[%d] timed out\n",
-				       engine->name, i - 1);
-				err = -ETIME;
-				goto err;
-			}
-
-			pr_debug("%s: slot[%d]:%x\n", engine->name, i, slot[i]);
-		}
-
-		/* XZY: XZ < XY */
-		if (slot[Z] - slot[X] >= slot[Y] - slot[X]) {
-			pr_err("%s: timeslicing did not run context B [%u] before A [%u]!\n",
-			       engine->name,
-			       slot[Z] - slot[X],
-			       slot[Y] - slot[X]);
-			err = -EINVAL;
-		}
-
-err:
-		memset32(&slot[0], -1, 4);
-		wmb();
-
-		engine->props.timeslice_duration_ms = timeslice;
-		st_engine_heartbeat_enable(engine);
-		for (i = 0; i < 3; i++)
-			i915_request_put(rq[i]);
-		if (igt_flush_test(gt->i915))
-			err = -EIO;
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-static struct i915_request *nop_request(struct intel_engine_cs *engine)
-{
-	struct i915_request *rq;
-
-	rq = intel_engine_create_kernel_request(engine);
-	if (IS_ERR(rq))
-		return rq;
-
-	i915_request_get(rq);
-	i915_request_add(rq);
-
-	return rq;
-}
-
-static long slice_timeout(struct intel_engine_cs *engine)
-{
-	long timeout;
-
-	/* Enough time for a timeslice to kick in, and kick out */
-	timeout = 2 * msecs_to_jiffies_timeout(timeslice(engine));
-
-	/* Enough time for the nop request to complete */
-	timeout += HZ / 5;
-
-	return timeout + 1;
-}
-
-static int live_timeslice_queue(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct drm_i915_gem_object *obj;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	struct i915_vma *vma;
-	void *vaddr;
-	int err = 0;
-
-	/*
-	 * Make sure that even if ELSP[0] and ELSP[1] are filled with
-	 * timeslicing between them disabled, we *do* enable timeslicing
-	 * if the queue demands it. (Normally, we do not submit if
-	 * ELSP[1] is already occupied, so must rely on timeslicing to
-	 * eject ELSP[0] in favour of the queue.)
-	 */
-	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
-		return 0;
-
-	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
-	if (IS_ERR(obj))
-		return PTR_ERR(obj);
-
-	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		err = PTR_ERR(vma);
-		goto err_obj;
-	}
-
-	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WC);
-	if (IS_ERR(vaddr)) {
-		err = PTR_ERR(vaddr);
-		goto err_obj;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
-	if (err)
-		goto err_map;
-
-	err = i915_vma_sync(vma);
-	if (err)
-		goto err_pin;
-
-	for_each_engine(engine, gt, id) {
-		struct i915_sched_attr attr = {
-			.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX),
-		};
-		struct i915_request *rq, *nop;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		st_engine_heartbeat_disable(engine);
-		memset(vaddr, 0, PAGE_SIZE);
-
-		/* ELSP[0]: semaphore wait */
-		rq = semaphore_queue(engine, vma, 0);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto err_heartbeat;
-		}
-		engine->schedule(rq, &attr);
-		err = wait_for_submit(engine, rq, HZ / 2);
-		if (err) {
-			pr_err("%s: Timed out trying to submit semaphores\n",
-			       engine->name);
-			goto err_rq;
-		}
-
-		/* ELSP[1]: nop request */
-		nop = nop_request(engine);
-		if (IS_ERR(nop)) {
-			err = PTR_ERR(nop);
-			goto err_rq;
-		}
-		err = wait_for_submit(engine, nop, HZ / 2);
-		i915_request_put(nop);
-		if (err) {
-			pr_err("%s: Timed out trying to submit nop\n",
-			       engine->name);
-			goto err_rq;
-		}
-
-		GEM_BUG_ON(i915_request_completed(rq));
-		GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
-
-		/* Queue: semaphore signal, matching priority as semaphore */
-		err = release_queue(engine, vma, 1, effective_prio(rq));
-		if (err)
-			goto err_rq;
-
-		/* Wait until we ack the release_queue and start timeslicing */
-		do {
-			cond_resched();
-			intel_engine_flush_submission(engine);
-		} while (READ_ONCE(engine->execlists.pending[0]));
-
-		/* Timeslice every jiffy, so within 2 we should signal */
-		if (i915_request_wait(rq, 0, slice_timeout(engine)) < 0) {
-			struct drm_printer p =
-				drm_info_printer(gt->i915->drm.dev);
-
-			pr_err("%s: Failed to timeslice into queue\n",
-			       engine->name);
-			intel_engine_dump(engine, &p,
-					  "%s\n", engine->name);
-
-			memset(vaddr, 0xff, PAGE_SIZE);
-			err = -EIO;
-		}
-err_rq:
-		i915_request_put(rq);
-err_heartbeat:
-		st_engine_heartbeat_enable(engine);
-		if (err)
-			break;
-	}
-
-err_pin:
-	i915_vma_unpin(vma);
-err_map:
-	i915_gem_object_unpin_map(obj);
-err_obj:
-	i915_gem_object_put(obj);
-	return err;
-}
-
-static int live_timeslice_nopreempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	struct igt_spinner spin;
-	int err = 0;
-
-	/*
-	 * We should not timeslice into a request that is marked with
-	 * I915_REQUEST_NOPREEMPT.
-	 */
-	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
-		return 0;
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	for_each_engine(engine, gt, id) {
-		struct intel_context *ce;
-		struct i915_request *rq;
-		unsigned long timeslice;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			break;
-		}
-
-		st_engine_heartbeat_disable(engine);
-		timeslice = xchg(&engine->props.timeslice_duration_ms, 1);
-
-		/* Create an unpreemptible spinner */
-
-		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
-		intel_context_put(ce);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out_heartbeat;
-		}
-
-		i915_request_get(rq);
-		i915_request_add(rq);
-
-		if (!igt_wait_for_spinner(&spin, rq)) {
-			i915_request_put(rq);
-			err = -ETIME;
-			goto out_spin;
-		}
-
-		set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq->fence.flags);
-		i915_request_put(rq);
-
-		/* Followed by a maximum priority barrier (heartbeat) */
-
-		ce = intel_context_create(engine);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto out_spin;
-		}
-
-		rq = intel_context_create_request(ce);
-		intel_context_put(ce);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out_spin;
-		}
-
-		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
-		i915_request_get(rq);
-		i915_request_add(rq);
-
-		/*
-		 * Wait until the barrier is in ELSP, and we know timeslicing
-		 * will have been activated.
-		 */
-		if (wait_for_submit(engine, rq, HZ / 2)) {
-			i915_request_put(rq);
-			err = -ETIME;
-			goto out_spin;
-		}
-
-		/*
-		 * Since the ELSP[0] request is unpreemptible, it should not
-		 * allow the maximum priority barrier through. Wait long
-		 * enough to see if it is timesliced in by mistake.
-		 */
-		if (i915_request_wait(rq, 0, slice_timeout(engine)) >= 0) {
-			pr_err("%s: I915_PRIORITY_BARRIER request completed, bypassing no-preempt request\n",
-			       engine->name);
-			err = -EINVAL;
-		}
-		i915_request_put(rq);
-
-out_spin:
-		igt_spinner_end(&spin);
-out_heartbeat:
-		xchg(&engine->props.timeslice_duration_ms, timeslice);
-		st_engine_heartbeat_enable(engine);
-		if (err)
-			break;
-
-		if (igt_flush_test(gt->i915)) {
-			err = -EIO;
-			break;
-		}
-	}
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_busywait_preempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct i915_gem_context *ctx_hi, *ctx_lo;
-	struct intel_engine_cs *engine;
-	struct drm_i915_gem_object *obj;
-	struct i915_vma *vma;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-	u32 *map;
-
-	/*
-	 * Verify that even without HAS_LOGICAL_RING_PREEMPTION, we can
-	 * preempt the busywaits used to synchronise between rings.
-	 */
-
-	ctx_hi = kernel_context(gt->i915);
-	if (!ctx_hi)
-		return -ENOMEM;
-	ctx_hi->sched.priority =
-		I915_USER_PRIORITY(I915_CONTEXT_MAX_USER_PRIORITY);
-
-	ctx_lo = kernel_context(gt->i915);
-	if (!ctx_lo)
-		goto err_ctx_hi;
-	ctx_lo->sched.priority =
-		I915_USER_PRIORITY(I915_CONTEXT_MIN_USER_PRIORITY);
-
-	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
-	if (IS_ERR(obj)) {
-		err = PTR_ERR(obj);
-		goto err_ctx_lo;
-	}
-
-	map = i915_gem_object_pin_map(obj, I915_MAP_WC);
-	if (IS_ERR(map)) {
-		err = PTR_ERR(map);
-		goto err_obj;
-	}
-
-	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		err = PTR_ERR(vma);
-		goto err_map;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
-	if (err)
-		goto err_map;
-
-	err = i915_vma_sync(vma);
-	if (err)
-		goto err_vma;
-
-	for_each_engine(engine, gt, id) {
-		struct i915_request *lo, *hi;
-		struct igt_live_test t;
-		u32 *cs;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (!intel_engine_can_store_dword(engine))
-			continue;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			goto err_vma;
-		}
-
-		/*
-		 * We create two requests. The low priority request
-		 * busywaits on a semaphore (inside the ringbuffer where
-		 * is should be preemptible) and the high priority requests
-		 * uses a MI_STORE_DWORD_IMM to update the semaphore value
-		 * allowing the first request to complete. If preemption
-		 * fails, we hang instead.
-		 */
-
-		lo = igt_request_alloc(ctx_lo, engine);
-		if (IS_ERR(lo)) {
-			err = PTR_ERR(lo);
-			goto err_vma;
-		}
-
-		cs = intel_ring_begin(lo, 8);
-		if (IS_ERR(cs)) {
-			err = PTR_ERR(cs);
-			i915_request_add(lo);
-			goto err_vma;
-		}
-
-		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-		*cs++ = i915_ggtt_offset(vma);
-		*cs++ = 0;
-		*cs++ = 1;
-
-		/* XXX Do we need a flush + invalidate here? */
-
-		*cs++ = MI_SEMAPHORE_WAIT |
-			MI_SEMAPHORE_GLOBAL_GTT |
-			MI_SEMAPHORE_POLL |
-			MI_SEMAPHORE_SAD_EQ_SDD;
-		*cs++ = 0;
-		*cs++ = i915_ggtt_offset(vma);
-		*cs++ = 0;
-
-		intel_ring_advance(lo, cs);
-
-		i915_request_get(lo);
-		i915_request_add(lo);
-
-		if (wait_for(READ_ONCE(*map), 10)) {
-			i915_request_put(lo);
-			err = -ETIMEDOUT;
-			goto err_vma;
-		}
-
-		/* Low priority request should be busywaiting now */
-		if (i915_request_wait(lo, 0, 1) != -ETIME) {
-			i915_request_put(lo);
-			pr_err("%s: Busywaiting request did not!\n",
-			       engine->name);
-			err = -EIO;
-			goto err_vma;
-		}
-
-		hi = igt_request_alloc(ctx_hi, engine);
-		if (IS_ERR(hi)) {
-			err = PTR_ERR(hi);
-			i915_request_put(lo);
-			goto err_vma;
-		}
-
-		cs = intel_ring_begin(hi, 4);
-		if (IS_ERR(cs)) {
-			err = PTR_ERR(cs);
-			i915_request_add(hi);
-			i915_request_put(lo);
-			goto err_vma;
-		}
-
-		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-		*cs++ = i915_ggtt_offset(vma);
-		*cs++ = 0;
-		*cs++ = 0;
-
-		intel_ring_advance(hi, cs);
-		i915_request_add(hi);
-
-		if (i915_request_wait(lo, 0, HZ / 5) < 0) {
-			struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
-
-			pr_err("%s: Failed to preempt semaphore busywait!\n",
-			       engine->name);
-
-			intel_engine_dump(engine, &p, "%s\n", engine->name);
-			GEM_TRACE_DUMP();
-
-			i915_request_put(lo);
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-			goto err_vma;
-		}
-		GEM_BUG_ON(READ_ONCE(*map));
-		i915_request_put(lo);
-
-		if (igt_live_test_end(&t)) {
-			err = -EIO;
-			goto err_vma;
-		}
-	}
-
-	err = 0;
-err_vma:
-	i915_vma_unpin(vma);
-err_map:
-	i915_gem_object_unpin_map(obj);
-err_obj:
-	i915_gem_object_put(obj);
-err_ctx_lo:
-	kernel_context_close(ctx_lo);
-err_ctx_hi:
-	kernel_context_close(ctx_hi);
-	return err;
-}
-
-static struct i915_request *
-spinner_create_request(struct igt_spinner *spin,
-		       struct i915_gem_context *ctx,
-		       struct intel_engine_cs *engine,
-		       u32 arb)
-{
-	struct intel_context *ce;
-	struct i915_request *rq;
-
-	ce = i915_gem_context_get_engine(ctx, engine->legacy_idx);
-	if (IS_ERR(ce))
-		return ERR_CAST(ce);
-
-	rq = igt_spinner_create_request(spin, ce, arb);
-	intel_context_put(ce);
-	return rq;
-}
-
-static int live_preempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct i915_gem_context *ctx_hi, *ctx_lo;
-	struct igt_spinner spin_hi, spin_lo;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (!(gt->i915->caps.scheduler & I915_SCHEDULER_CAP_PREEMPTION))
-		pr_err("Logical preemption supported, but not exposed\n");
-
-	if (igt_spinner_init(&spin_hi, gt))
-		return -ENOMEM;
-
-	if (igt_spinner_init(&spin_lo, gt))
-		goto err_spin_hi;
-
-	ctx_hi = kernel_context(gt->i915);
-	if (!ctx_hi)
-		goto err_spin_lo;
-	ctx_hi->sched.priority =
-		I915_USER_PRIORITY(I915_CONTEXT_MAX_USER_PRIORITY);
-
-	ctx_lo = kernel_context(gt->i915);
-	if (!ctx_lo)
-		goto err_ctx_hi;
-	ctx_lo->sched.priority =
-		I915_USER_PRIORITY(I915_CONTEXT_MIN_USER_PRIORITY);
-
-	for_each_engine(engine, gt, id) {
-		struct igt_live_test t;
-		struct i915_request *rq;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-
-		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
-					    MI_ARB_CHECK);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto err_ctx_lo;
-		}
-
-		i915_request_add(rq);
-		if (!igt_wait_for_spinner(&spin_lo, rq)) {
-			GEM_TRACE("lo spinner failed to start\n");
-			GEM_TRACE_DUMP();
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-
-		rq = spinner_create_request(&spin_hi, ctx_hi, engine,
-					    MI_ARB_CHECK);
-		if (IS_ERR(rq)) {
-			igt_spinner_end(&spin_lo);
-			err = PTR_ERR(rq);
-			goto err_ctx_lo;
-		}
-
-		i915_request_add(rq);
-		if (!igt_wait_for_spinner(&spin_hi, rq)) {
-			GEM_TRACE("hi spinner failed to start\n");
-			GEM_TRACE_DUMP();
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-
-		igt_spinner_end(&spin_hi);
-		igt_spinner_end(&spin_lo);
-
-		if (igt_live_test_end(&t)) {
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-	}
-
-	err = 0;
-err_ctx_lo:
-	kernel_context_close(ctx_lo);
-err_ctx_hi:
-	kernel_context_close(ctx_hi);
-err_spin_lo:
-	igt_spinner_fini(&spin_lo);
-err_spin_hi:
-	igt_spinner_fini(&spin_hi);
-	return err;
-}
-
-static int live_late_preempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct i915_gem_context *ctx_hi, *ctx_lo;
-	struct igt_spinner spin_hi, spin_lo;
-	struct intel_engine_cs *engine;
-	struct i915_sched_attr attr = {};
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (igt_spinner_init(&spin_hi, gt))
-		return -ENOMEM;
-
-	if (igt_spinner_init(&spin_lo, gt))
-		goto err_spin_hi;
-
-	ctx_hi = kernel_context(gt->i915);
-	if (!ctx_hi)
-		goto err_spin_lo;
-
-	ctx_lo = kernel_context(gt->i915);
-	if (!ctx_lo)
-		goto err_ctx_hi;
-
-	/* Make sure ctx_lo stays before ctx_hi until we trigger preemption. */
-	ctx_lo->sched.priority = I915_USER_PRIORITY(1);
-
-	for_each_engine(engine, gt, id) {
-		struct igt_live_test t;
-		struct i915_request *rq;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-
-		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
-					    MI_ARB_CHECK);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto err_ctx_lo;
-		}
-
-		i915_request_add(rq);
-		if (!igt_wait_for_spinner(&spin_lo, rq)) {
-			pr_err("First context failed to start\n");
-			goto err_wedged;
-		}
-
-		rq = spinner_create_request(&spin_hi, ctx_hi, engine,
-					    MI_NOOP);
-		if (IS_ERR(rq)) {
-			igt_spinner_end(&spin_lo);
-			err = PTR_ERR(rq);
-			goto err_ctx_lo;
-		}
-
-		i915_request_add(rq);
-		if (igt_wait_for_spinner(&spin_hi, rq)) {
-			pr_err("Second context overtook first?\n");
-			goto err_wedged;
-		}
-
-		attr.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX);
-		engine->schedule(rq, &attr);
-
-		if (!igt_wait_for_spinner(&spin_hi, rq)) {
-			pr_err("High priority context failed to preempt the low priority context\n");
-			GEM_TRACE_DUMP();
-			goto err_wedged;
-		}
-
-		igt_spinner_end(&spin_hi);
-		igt_spinner_end(&spin_lo);
-
-		if (igt_live_test_end(&t)) {
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-	}
-
-	err = 0;
-err_ctx_lo:
-	kernel_context_close(ctx_lo);
-err_ctx_hi:
-	kernel_context_close(ctx_hi);
-err_spin_lo:
-	igt_spinner_fini(&spin_lo);
-err_spin_hi:
-	igt_spinner_fini(&spin_hi);
-	return err;
-
-err_wedged:
-	igt_spinner_end(&spin_hi);
-	igt_spinner_end(&spin_lo);
-	intel_gt_set_wedged(gt);
-	err = -EIO;
-	goto err_ctx_lo;
-}
-
-struct preempt_client {
-	struct igt_spinner spin;
-	struct i915_gem_context *ctx;
-};
-
-static int preempt_client_init(struct intel_gt *gt, struct preempt_client *c)
-{
-	c->ctx = kernel_context(gt->i915);
-	if (!c->ctx)
-		return -ENOMEM;
-
-	if (igt_spinner_init(&c->spin, gt))
-		goto err_ctx;
-
-	return 0;
-
-err_ctx:
-	kernel_context_close(c->ctx);
-	return -ENOMEM;
-}
-
-static void preempt_client_fini(struct preempt_client *c)
-{
-	igt_spinner_fini(&c->spin);
-	kernel_context_close(c->ctx);
-}
-
-static int live_nopreempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	struct preempt_client a, b;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	/*
-	 * Verify that we can disable preemption for an individual request
-	 * that may be being observed and not want to be interrupted.
-	 */
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (preempt_client_init(gt, &a))
-		return -ENOMEM;
-	if (preempt_client_init(gt, &b))
-		goto err_client_a;
-	b.ctx->sched.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX);
-
-	for_each_engine(engine, gt, id) {
-		struct i915_request *rq_a, *rq_b;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		engine->execlists.preempt_hang.count = 0;
-
-		rq_a = spinner_create_request(&a.spin,
-					      a.ctx, engine,
-					      MI_ARB_CHECK);
-		if (IS_ERR(rq_a)) {
-			err = PTR_ERR(rq_a);
-			goto err_client_b;
-		}
-
-		/* Low priority client, but unpreemptable! */
-		__set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq_a->fence.flags);
-
-		i915_request_add(rq_a);
-		if (!igt_wait_for_spinner(&a.spin, rq_a)) {
-			pr_err("First client failed to start\n");
-			goto err_wedged;
-		}
-
-		rq_b = spinner_create_request(&b.spin,
-					      b.ctx, engine,
-					      MI_ARB_CHECK);
-		if (IS_ERR(rq_b)) {
-			err = PTR_ERR(rq_b);
-			goto err_client_b;
-		}
-
-		i915_request_add(rq_b);
-
-		/* B is much more important than A! (But A is unpreemptable.) */
-		GEM_BUG_ON(rq_prio(rq_b) <= rq_prio(rq_a));
-
-		/* Wait long enough for preemption and timeslicing */
-		if (igt_wait_for_spinner(&b.spin, rq_b)) {
-			pr_err("Second client started too early!\n");
-			goto err_wedged;
-		}
-
-		igt_spinner_end(&a.spin);
-
-		if (!igt_wait_for_spinner(&b.spin, rq_b)) {
-			pr_err("Second client failed to start\n");
-			goto err_wedged;
-		}
-
-		igt_spinner_end(&b.spin);
-
-		if (engine->execlists.preempt_hang.count) {
-			pr_err("Preemption recorded x%d; should have been suppressed!\n",
-			       engine->execlists.preempt_hang.count);
-			err = -EINVAL;
-			goto err_wedged;
-		}
-
-		if (igt_flush_test(gt->i915))
-			goto err_wedged;
-	}
-
-	err = 0;
-err_client_b:
-	preempt_client_fini(&b);
-err_client_a:
-	preempt_client_fini(&a);
-	return err;
-
-err_wedged:
-	igt_spinner_end(&b.spin);
-	igt_spinner_end(&a.spin);
-	intel_gt_set_wedged(gt);
-	err = -EIO;
-	goto err_client_b;
-}
-
-struct live_preempt_cancel {
-	struct intel_engine_cs *engine;
-	struct preempt_client a, b;
-};
-
-static int __cancel_active0(struct live_preempt_cancel *arg)
-{
-	struct i915_request *rq;
-	struct igt_live_test t;
-	int err;
-
-	/* Preempt cancel of ELSP0 */
-	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
-	if (igt_live_test_begin(&t, arg->engine->i915,
-				__func__, arg->engine->name))
-		return -EIO;
-
-	rq = spinner_create_request(&arg->a.spin,
-				    arg->a.ctx, arg->engine,
-				    MI_ARB_CHECK);
-	if (IS_ERR(rq))
-		return PTR_ERR(rq);
-
-	clear_bit(CONTEXT_BANNED, &rq->context->flags);
-	i915_request_get(rq);
-	i915_request_add(rq);
-	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
-		err = -EIO;
-		goto out;
-	}
-
-	intel_context_set_banned(rq->context);
-	err = intel_engine_pulse(arg->engine);
-	if (err)
-		goto out;
-
-	err = wait_for_reset(arg->engine, rq, HZ / 2);
-	if (err) {
-		pr_err("Cancelled inflight0 request did not reset\n");
-		goto out;
-	}
-
-out:
-	i915_request_put(rq);
-	if (igt_live_test_end(&t))
-		err = -EIO;
-	return err;
-}
-
-static int __cancel_active1(struct live_preempt_cancel *arg)
-{
-	struct i915_request *rq[2] = {};
-	struct igt_live_test t;
-	int err;
-
-	/* Preempt cancel of ELSP1 */
-	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
-	if (igt_live_test_begin(&t, arg->engine->i915,
-				__func__, arg->engine->name))
-		return -EIO;
-
-	rq[0] = spinner_create_request(&arg->a.spin,
-				       arg->a.ctx, arg->engine,
-				       MI_NOOP); /* no preemption */
-	if (IS_ERR(rq[0]))
-		return PTR_ERR(rq[0]);
-
-	clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
-	i915_request_get(rq[0]);
-	i915_request_add(rq[0]);
-	if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
-		err = -EIO;
-		goto out;
-	}
-
-	rq[1] = spinner_create_request(&arg->b.spin,
-				       arg->b.ctx, arg->engine,
-				       MI_ARB_CHECK);
-	if (IS_ERR(rq[1])) {
-		err = PTR_ERR(rq[1]);
-		goto out;
-	}
-
-	clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
-	i915_request_get(rq[1]);
-	err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
-	i915_request_add(rq[1]);
-	if (err)
-		goto out;
-
-	intel_context_set_banned(rq[1]->context);
-	err = intel_engine_pulse(arg->engine);
-	if (err)
-		goto out;
-
-	igt_spinner_end(&arg->a.spin);
-	err = wait_for_reset(arg->engine, rq[1], HZ / 2);
-	if (err)
-		goto out;
-
-	if (rq[0]->fence.error != 0) {
-		pr_err("Normal inflight0 request did not complete\n");
-		err = -EINVAL;
-		goto out;
-	}
-
-	if (rq[1]->fence.error != -EIO) {
-		pr_err("Cancelled inflight1 request did not report -EIO\n");
-		err = -EINVAL;
-		goto out;
-	}
-
-out:
-	i915_request_put(rq[1]);
-	i915_request_put(rq[0]);
-	if (igt_live_test_end(&t))
-		err = -EIO;
-	return err;
-}
-
-static int __cancel_queued(struct live_preempt_cancel *arg)
-{
-	struct i915_request *rq[3] = {};
-	struct igt_live_test t;
-	int err;
-
-	/* Full ELSP and one in the wings */
-	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
-	if (igt_live_test_begin(&t, arg->engine->i915,
-				__func__, arg->engine->name))
-		return -EIO;
-
-	rq[0] = spinner_create_request(&arg->a.spin,
-				       arg->a.ctx, arg->engine,
-				       MI_ARB_CHECK);
-	if (IS_ERR(rq[0]))
-		return PTR_ERR(rq[0]);
-
-	clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
-	i915_request_get(rq[0]);
-	i915_request_add(rq[0]);
-	if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
-		err = -EIO;
-		goto out;
-	}
-
-	rq[1] = igt_request_alloc(arg->b.ctx, arg->engine);
-	if (IS_ERR(rq[1])) {
-		err = PTR_ERR(rq[1]);
-		goto out;
-	}
-
-	clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
-	i915_request_get(rq[1]);
-	err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
-	i915_request_add(rq[1]);
-	if (err)
-		goto out;
-
-	rq[2] = spinner_create_request(&arg->b.spin,
-				       arg->a.ctx, arg->engine,
-				       MI_ARB_CHECK);
-	if (IS_ERR(rq[2])) {
-		err = PTR_ERR(rq[2]);
-		goto out;
-	}
-
-	i915_request_get(rq[2]);
-	err = i915_request_await_dma_fence(rq[2], &rq[1]->fence);
-	i915_request_add(rq[2]);
-	if (err)
-		goto out;
-
-	intel_context_set_banned(rq[2]->context);
-	err = intel_engine_pulse(arg->engine);
-	if (err)
-		goto out;
-
-	err = wait_for_reset(arg->engine, rq[2], HZ / 2);
-	if (err)
-		goto out;
-
-	if (rq[0]->fence.error != -EIO) {
-		pr_err("Cancelled inflight0 request did not report -EIO\n");
-		err = -EINVAL;
-		goto out;
-	}
-
-	if (rq[1]->fence.error != 0) {
-		pr_err("Normal inflight1 request did not complete\n");
-		err = -EINVAL;
-		goto out;
-	}
-
-	if (rq[2]->fence.error != -EIO) {
-		pr_err("Cancelled queued request did not report -EIO\n");
-		err = -EINVAL;
-		goto out;
-	}
-
-out:
-	i915_request_put(rq[2]);
-	i915_request_put(rq[1]);
-	i915_request_put(rq[0]);
-	if (igt_live_test_end(&t))
-		err = -EIO;
-	return err;
-}
-
-static int __cancel_hostile(struct live_preempt_cancel *arg)
-{
-	struct i915_request *rq;
-	int err;
-
-	/* Preempt cancel non-preemptible spinner in ELSP0 */
-	if (!IS_ACTIVE(CONFIG_DRM_I915_PREEMPT_TIMEOUT))
-		return 0;
-
-	if (!intel_has_reset_engine(arg->engine->gt))
-		return 0;
-
-	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
-	rq = spinner_create_request(&arg->a.spin,
-				    arg->a.ctx, arg->engine,
-				    MI_NOOP); /* preemption disabled */
-	if (IS_ERR(rq))
-		return PTR_ERR(rq);
-
-	clear_bit(CONTEXT_BANNED, &rq->context->flags);
-	i915_request_get(rq);
-	i915_request_add(rq);
-	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
-		err = -EIO;
-		goto out;
-	}
-
-	intel_context_set_banned(rq->context);
-	err = intel_engine_pulse(arg->engine); /* force reset */
-	if (err)
-		goto out;
-
-	err = wait_for_reset(arg->engine, rq, HZ / 2);
-	if (err) {
-		pr_err("Cancelled inflight0 request did not reset\n");
-		goto out;
-	}
-
-out:
-	i915_request_put(rq);
-	if (igt_flush_test(arg->engine->i915))
-		err = -EIO;
-	return err;
-}
-
-static int live_preempt_cancel(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct live_preempt_cancel data;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	/*
-	 * To cancel an inflight context, we need to first remove it from the
-	 * GPU. That sounds like preemption! Plus a little bit of bookkeeping.
-	 */
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (preempt_client_init(gt, &data.a))
-		return -ENOMEM;
-	if (preempt_client_init(gt, &data.b))
-		goto err_client_a;
-
-	for_each_engine(data.engine, gt, id) {
-		if (!intel_engine_has_preemption(data.engine))
-			continue;
-
-		err = __cancel_active0(&data);
-		if (err)
-			goto err_wedged;
-
-		err = __cancel_active1(&data);
-		if (err)
-			goto err_wedged;
-
-		err = __cancel_queued(&data);
-		if (err)
-			goto err_wedged;
-
-		err = __cancel_hostile(&data);
-		if (err)
-			goto err_wedged;
-	}
-
-	err = 0;
-err_client_b:
-	preempt_client_fini(&data.b);
-err_client_a:
-	preempt_client_fini(&data.a);
-	return err;
-
-err_wedged:
-	GEM_TRACE_DUMP();
-	igt_spinner_end(&data.b.spin);
-	igt_spinner_end(&data.a.spin);
-	intel_gt_set_wedged(gt);
-	goto err_client_b;
-}
-
-static int live_suppress_self_preempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	struct i915_sched_attr attr = {
-		.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX)
-	};
-	struct preempt_client a, b;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	/*
-	 * Verify that if a preemption request does not cause a change in
-	 * the current execution order, the preempt-to-idle injection is
-	 * skipped and that we do not accidentally apply it after the CS
-	 * completion event.
-	 */
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0; /* presume black blox */
-
-	if (intel_vgpu_active(gt->i915))
-		return 0; /* GVT forces single port & request submission */
-
-	if (preempt_client_init(gt, &a))
-		return -ENOMEM;
-	if (preempt_client_init(gt, &b))
-		goto err_client_a;
-
-	for_each_engine(engine, gt, id) {
-		struct i915_request *rq_a, *rq_b;
-		int depth;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (igt_flush_test(gt->i915))
-			goto err_wedged;
-
-		st_engine_heartbeat_disable(engine);
-		engine->execlists.preempt_hang.count = 0;
-
-		rq_a = spinner_create_request(&a.spin,
-					      a.ctx, engine,
-					      MI_NOOP);
-		if (IS_ERR(rq_a)) {
-			err = PTR_ERR(rq_a);
-			st_engine_heartbeat_enable(engine);
-			goto err_client_b;
-		}
-
-		i915_request_add(rq_a);
-		if (!igt_wait_for_spinner(&a.spin, rq_a)) {
-			pr_err("First client failed to start\n");
-			st_engine_heartbeat_enable(engine);
-			goto err_wedged;
-		}
-
-		/* Keep postponing the timer to avoid premature slicing */
-		mod_timer(&engine->execlists.timer, jiffies + HZ);
-		for (depth = 0; depth < 8; depth++) {
-			rq_b = spinner_create_request(&b.spin,
-						      b.ctx, engine,
-						      MI_NOOP);
-			if (IS_ERR(rq_b)) {
-				err = PTR_ERR(rq_b);
-				st_engine_heartbeat_enable(engine);
-				goto err_client_b;
-			}
-			i915_request_add(rq_b);
-
-			GEM_BUG_ON(i915_request_completed(rq_a));
-			engine->schedule(rq_a, &attr);
-			igt_spinner_end(&a.spin);
-
-			if (!igt_wait_for_spinner(&b.spin, rq_b)) {
-				pr_err("Second client failed to start\n");
-				st_engine_heartbeat_enable(engine);
-				goto err_wedged;
-			}
-
-			swap(a, b);
-			rq_a = rq_b;
-		}
-		igt_spinner_end(&a.spin);
-
-		if (engine->execlists.preempt_hang.count) {
-			pr_err("Preemption on %s recorded x%d, depth %d; should have been suppressed!\n",
-			       engine->name,
-			       engine->execlists.preempt_hang.count,
-			       depth);
-			st_engine_heartbeat_enable(engine);
-			err = -EINVAL;
-			goto err_client_b;
-		}
-
-		st_engine_heartbeat_enable(engine);
-		if (igt_flush_test(gt->i915))
-			goto err_wedged;
-	}
-
-	err = 0;
-err_client_b:
-	preempt_client_fini(&b);
-err_client_a:
-	preempt_client_fini(&a);
-	return err;
-
-err_wedged:
-	igt_spinner_end(&b.spin);
-	igt_spinner_end(&a.spin);
-	intel_gt_set_wedged(gt);
-	err = -EIO;
-	goto err_client_b;
-}
-
-static int live_chain_preempt(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	struct preempt_client hi, lo;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	/*
-	 * Build a chain AB...BA between two contexts (A, B) and request
-	 * preemption of the last request. It should then complete before
-	 * the previously submitted spinner in B.
-	 */
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (preempt_client_init(gt, &hi))
-		return -ENOMEM;
-
-	if (preempt_client_init(gt, &lo))
-		goto err_client_hi;
-
-	for_each_engine(engine, gt, id) {
-		struct i915_sched_attr attr = {
-			.priority = I915_USER_PRIORITY(I915_PRIORITY_MAX),
-		};
-		struct igt_live_test t;
-		struct i915_request *rq;
-		int ring_size, count, i;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		rq = spinner_create_request(&lo.spin,
-					    lo.ctx, engine,
-					    MI_ARB_CHECK);
-		if (IS_ERR(rq))
-			goto err_wedged;
-
-		i915_request_get(rq);
-		i915_request_add(rq);
-
-		ring_size = rq->wa_tail - rq->head;
-		if (ring_size < 0)
-			ring_size += rq->ring->size;
-		ring_size = rq->ring->size / ring_size;
-		pr_debug("%s(%s): Using maximum of %d requests\n",
-			 __func__, engine->name, ring_size);
-
-		igt_spinner_end(&lo.spin);
-		if (i915_request_wait(rq, 0, HZ / 2) < 0) {
-			pr_err("Timed out waiting to flush %s\n", engine->name);
-			i915_request_put(rq);
-			goto err_wedged;
-		}
-		i915_request_put(rq);
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			goto err_wedged;
-		}
-
-		for_each_prime_number_from(count, 1, ring_size) {
-			rq = spinner_create_request(&hi.spin,
-						    hi.ctx, engine,
-						    MI_ARB_CHECK);
-			if (IS_ERR(rq))
-				goto err_wedged;
-			i915_request_add(rq);
-			if (!igt_wait_for_spinner(&hi.spin, rq))
-				goto err_wedged;
-
-			rq = spinner_create_request(&lo.spin,
-						    lo.ctx, engine,
-						    MI_ARB_CHECK);
-			if (IS_ERR(rq))
-				goto err_wedged;
-			i915_request_add(rq);
-
-			for (i = 0; i < count; i++) {
-				rq = igt_request_alloc(lo.ctx, engine);
-				if (IS_ERR(rq))
-					goto err_wedged;
-				i915_request_add(rq);
-			}
-
-			rq = igt_request_alloc(hi.ctx, engine);
-			if (IS_ERR(rq))
-				goto err_wedged;
-
-			i915_request_get(rq);
-			i915_request_add(rq);
-			engine->schedule(rq, &attr);
-
-			igt_spinner_end(&hi.spin);
-			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
-				struct drm_printer p =
-					drm_info_printer(gt->i915->drm.dev);
-
-				pr_err("Failed to preempt over chain of %d\n",
-				       count);
-				intel_engine_dump(engine, &p,
-						  "%s\n", engine->name);
-				i915_request_put(rq);
-				goto err_wedged;
-			}
-			igt_spinner_end(&lo.spin);
-			i915_request_put(rq);
-
-			rq = igt_request_alloc(lo.ctx, engine);
-			if (IS_ERR(rq))
-				goto err_wedged;
-
-			i915_request_get(rq);
-			i915_request_add(rq);
-
-			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
-				struct drm_printer p =
-					drm_info_printer(gt->i915->drm.dev);
-
-				pr_err("Failed to flush low priority chain of %d requests\n",
-				       count);
-				intel_engine_dump(engine, &p,
-						  "%s\n", engine->name);
-
-				i915_request_put(rq);
-				goto err_wedged;
-			}
-			i915_request_put(rq);
-		}
-
-		if (igt_live_test_end(&t)) {
-			err = -EIO;
-			goto err_wedged;
-		}
-	}
-
-	err = 0;
-err_client_lo:
-	preempt_client_fini(&lo);
-err_client_hi:
-	preempt_client_fini(&hi);
-	return err;
-
-err_wedged:
-	igt_spinner_end(&hi.spin);
-	igt_spinner_end(&lo.spin);
-	intel_gt_set_wedged(gt);
-	err = -EIO;
-	goto err_client_lo;
-}
-
-static int create_gang(struct intel_engine_cs *engine,
-		       struct i915_request **prev)
-{
-	struct drm_i915_gem_object *obj;
-	struct intel_context *ce;
-	struct i915_request *rq;
-	struct i915_vma *vma;
-	u32 *cs;
-	int err;
-
-	ce = intel_context_create(engine);
-	if (IS_ERR(ce))
-		return PTR_ERR(ce);
-
-	obj = i915_gem_object_create_internal(engine->i915, 4096);
-	if (IS_ERR(obj)) {
-		err = PTR_ERR(obj);
-		goto err_ce;
-	}
-
-	vma = i915_vma_instance(obj, ce->vm, NULL);
-	if (IS_ERR(vma)) {
-		err = PTR_ERR(vma);
-		goto err_obj;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_USER);
-	if (err)
-		goto err_obj;
-
-	cs = i915_gem_object_pin_map(obj, I915_MAP_WC);
-	if (IS_ERR(cs))
-		goto err_obj;
-
-	/* Semaphore target: spin until zero */
-	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
-
-	*cs++ = MI_SEMAPHORE_WAIT |
-		MI_SEMAPHORE_POLL |
-		MI_SEMAPHORE_SAD_EQ_SDD;
-	*cs++ = 0;
-	*cs++ = lower_32_bits(vma->node.start);
-	*cs++ = upper_32_bits(vma->node.start);
-
-	if (*prev) {
-		u64 offset = (*prev)->batch->node.start;
-
-		/* Terminate the spinner in the next lower priority batch. */
-		*cs++ = MI_STORE_DWORD_IMM_GEN4;
-		*cs++ = lower_32_bits(offset);
-		*cs++ = upper_32_bits(offset);
-		*cs++ = 0;
-	}
-
-	*cs++ = MI_BATCH_BUFFER_END;
-	i915_gem_object_flush_map(obj);
-	i915_gem_object_unpin_map(obj);
-
-	rq = intel_context_create_request(ce);
-	if (IS_ERR(rq))
-		goto err_obj;
-
-	rq->batch = i915_vma_get(vma);
-	i915_request_get(rq);
-
-	i915_vma_lock(vma);
-	err = i915_request_await_object(rq, vma->obj, false);
-	if (!err)
-		err = i915_vma_move_to_active(vma, rq, 0);
-	if (!err)
-		err = rq->engine->emit_bb_start(rq,
-						vma->node.start,
-						PAGE_SIZE, 0);
-	i915_vma_unlock(vma);
-	i915_request_add(rq);
-	if (err)
-		goto err_rq;
-
-	i915_gem_object_put(obj);
-	intel_context_put(ce);
-
-	rq->mock.link.next = &(*prev)->mock.link;
-	*prev = rq;
-	return 0;
-
-err_rq:
-	i915_vma_put(rq->batch);
-	i915_request_put(rq);
-err_obj:
-	i915_gem_object_put(obj);
-err_ce:
-	intel_context_put(ce);
-	return err;
-}
-
-static int __live_preempt_ring(struct intel_engine_cs *engine,
-			       struct igt_spinner *spin,
-			       int queue_sz, int ring_sz)
-{
-	struct intel_context *ce[2] = {};
-	struct i915_request *rq;
-	struct igt_live_test t;
-	int err = 0;
-	int n;
-
-	if (igt_live_test_begin(&t, engine->i915, __func__, engine->name))
-		return -EIO;
-
-	for (n = 0; n < ARRAY_SIZE(ce); n++) {
-		struct intel_context *tmp;
-
-		tmp = intel_context_create(engine);
-		if (IS_ERR(tmp)) {
-			err = PTR_ERR(tmp);
-			goto err_ce;
-		}
-
-		tmp->ring = __intel_context_ring_size(ring_sz);
-
-		err = intel_context_pin(tmp);
-		if (err) {
-			intel_context_put(tmp);
-			goto err_ce;
-		}
-
-		memset32(tmp->ring->vaddr,
-			 0xdeadbeef, /* trigger a hang if executed */
-			 tmp->ring->vma->size / sizeof(u32));
-
-		ce[n] = tmp;
-	}
-
-	rq = igt_spinner_create_request(spin, ce[0], MI_ARB_CHECK);
-	if (IS_ERR(rq)) {
-		err = PTR_ERR(rq);
-		goto err_ce;
-	}
-
-	i915_request_get(rq);
-	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
-	i915_request_add(rq);
-
-	if (!igt_wait_for_spinner(spin, rq)) {
-		intel_gt_set_wedged(engine->gt);
-		i915_request_put(rq);
-		err = -ETIME;
-		goto err_ce;
-	}
-
-	/* Fill the ring, until we will cause a wrap */
-	n = 0;
-	while (ce[0]->ring->tail - rq->wa_tail <= queue_sz) {
-		struct i915_request *tmp;
-
-		tmp = intel_context_create_request(ce[0]);
-		if (IS_ERR(tmp)) {
-			err = PTR_ERR(tmp);
-			i915_request_put(rq);
-			goto err_ce;
-		}
-
-		i915_request_add(tmp);
-		intel_engine_flush_submission(engine);
-		n++;
-	}
-	intel_engine_flush_submission(engine);
-	pr_debug("%s: Filled %d with %d nop tails {size:%x, tail:%x, emit:%x, rq.tail:%x}\n",
-		 engine->name, queue_sz, n,
-		 ce[0]->ring->size,
-		 ce[0]->ring->tail,
-		 ce[0]->ring->emit,
-		 rq->tail);
-	i915_request_put(rq);
-
-	/* Create a second request to preempt the first ring */
-	rq = intel_context_create_request(ce[1]);
-	if (IS_ERR(rq)) {
-		err = PTR_ERR(rq);
-		goto err_ce;
-	}
-
-	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
-	i915_request_get(rq);
-	i915_request_add(rq);
-
-	err = wait_for_submit(engine, rq, HZ / 2);
-	i915_request_put(rq);
-	if (err) {
-		pr_err("%s: preemption request was not submited\n",
-		       engine->name);
-		err = -ETIME;
-	}
-
-	pr_debug("%s: ring[0]:{ tail:%x, emit:%x }, ring[1]:{ tail:%x, emit:%x }\n",
-		 engine->name,
-		 ce[0]->ring->tail, ce[0]->ring->emit,
-		 ce[1]->ring->tail, ce[1]->ring->emit);
-
-err_ce:
-	intel_engine_flush_submission(engine);
-	igt_spinner_end(spin);
-	for (n = 0; n < ARRAY_SIZE(ce); n++) {
-		if (IS_ERR_OR_NULL(ce[n]))
-			break;
-
-		intel_context_unpin(ce[n]);
-		intel_context_put(ce[n]);
-	}
-	if (igt_live_test_end(&t))
-		err = -EIO;
-	return err;
-}
-
-static int live_preempt_ring(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	struct igt_spinner spin;
-	enum intel_engine_id id;
-	int err = 0;
-
-	/*
-	 * Check that we rollback large chunks of a ring in order to do a
-	 * preemption event. Similar to live_unlite_ring, but looking at
-	 * ring size rather than the impact of intel_ring_direction().
-	 */
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	for_each_engine(engine, gt, id) {
-		int n;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (!intel_engine_can_store_dword(engine))
-			continue;
-
-		st_engine_heartbeat_disable(engine);
-
-		for (n = 0; n <= 3; n++) {
-			err = __live_preempt_ring(engine, &spin,
-						  n * SZ_4K / 4, SZ_4K);
-			if (err)
-				break;
-		}
-
-		st_engine_heartbeat_enable(engine);
-		if (err)
-			break;
-	}
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_preempt_gang(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	/*
-	 * Build as long a chain of preempters as we can, with each
-	 * request higher priority than the last. Once we are ready, we release
-	 * the last batch which then precolates down the chain, each releasing
-	 * the next oldest in turn. The intent is to simply push as hard as we
-	 * can with the number of preemptions, trying to exceed narrow HW
-	 * limits. At a minimum, we insist that we can sort all the user
-	 * high priority levels into execution order.
-	 */
-
-	for_each_engine(engine, gt, id) {
-		struct i915_request *rq = NULL;
-		struct igt_live_test t;
-		IGT_TIMEOUT(end_time);
-		int prio = 0;
-		int err = 0;
-		u32 *cs;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name))
-			return -EIO;
-
-		do {
-			struct i915_sched_attr attr = {
-				.priority = I915_USER_PRIORITY(prio++),
-			};
-
-			err = create_gang(engine, &rq);
-			if (err)
-				break;
-
-			/* Submit each spinner at increasing priority */
-			engine->schedule(rq, &attr);
-		} while (prio <= I915_PRIORITY_MAX &&
-			 !__igt_timeout(end_time, NULL));
-		pr_debug("%s: Preempt chain of %d requests\n",
-			 engine->name, prio);
-
-		/*
-		 * Such that the last spinner is the highest priority and
-		 * should execute first. When that spinner completes,
-		 * it will terminate the next lowest spinner until there
-		 * are no more spinners and the gang is complete.
-		 */
-		cs = i915_gem_object_pin_map(rq->batch->obj, I915_MAP_WC);
-		if (!IS_ERR(cs)) {
-			*cs = 0;
-			i915_gem_object_unpin_map(rq->batch->obj);
-		} else {
-			err = PTR_ERR(cs);
-			intel_gt_set_wedged(gt);
-		}
-
-		while (rq) { /* wait for each rq from highest to lowest prio */
-			struct i915_request *n = list_next_entry(rq, mock.link);
-
-			if (err == 0 && i915_request_wait(rq, 0, HZ / 5) < 0) {
-				struct drm_printer p =
-					drm_info_printer(engine->i915->drm.dev);
-
-				pr_err("Failed to flush chain of %d requests, at %d\n",
-				       prio, rq_prio(rq) >> I915_USER_PRIORITY_SHIFT);
-				intel_engine_dump(engine, &p,
-						  "%s\n", engine->name);
-
-				err = -ETIME;
-			}
-
-			i915_vma_put(rq->batch);
-			i915_request_put(rq);
-			rq = n;
-		}
-
-		if (igt_live_test_end(&t))
-			err = -EIO;
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-static struct i915_vma *
-create_gpr_user(struct intel_engine_cs *engine,
-		struct i915_vma *result,
-		unsigned int offset)
-{
-	struct drm_i915_gem_object *obj;
-	struct i915_vma *vma;
-	u32 *cs;
-	int err;
-	int i;
-
-	obj = i915_gem_object_create_internal(engine->i915, 4096);
-	if (IS_ERR(obj))
-		return ERR_CAST(obj);
-
-	vma = i915_vma_instance(obj, result->vm, NULL);
-	if (IS_ERR(vma)) {
-		i915_gem_object_put(obj);
-		return vma;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_USER);
-	if (err) {
-		i915_vma_put(vma);
-		return ERR_PTR(err);
-	}
-
-	cs = i915_gem_object_pin_map(obj, I915_MAP_WC);
-	if (IS_ERR(cs)) {
-		i915_vma_put(vma);
-		return ERR_CAST(cs);
-	}
-
-	/* All GPR are clear for new contexts. We use GPR(0) as a constant */
-	*cs++ = MI_LOAD_REGISTER_IMM(1);
-	*cs++ = CS_GPR(engine, 0);
-	*cs++ = 1;
-
-	for (i = 1; i < NUM_GPR; i++) {
-		u64 addr;
-
-		/*
-		 * Perform: GPR[i]++
-		 *
-		 * As we read and write into the context saved GPR[i], if
-		 * we restart this batch buffer from an earlier point, we
-		 * will repeat the increment and store a value > 1.
-		 */
-		*cs++ = MI_MATH(4);
-		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(i));
-		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(0));
-		*cs++ = MI_MATH_ADD;
-		*cs++ = MI_MATH_STORE(MI_MATH_REG(i), MI_MATH_REG_ACCU);
-
-		addr = result->node.start + offset + i * sizeof(*cs);
-		*cs++ = MI_STORE_REGISTER_MEM_GEN8;
-		*cs++ = CS_GPR(engine, 2 * i);
-		*cs++ = lower_32_bits(addr);
-		*cs++ = upper_32_bits(addr);
-
-		*cs++ = MI_SEMAPHORE_WAIT |
-			MI_SEMAPHORE_POLL |
-			MI_SEMAPHORE_SAD_GTE_SDD;
-		*cs++ = i;
-		*cs++ = lower_32_bits(result->node.start);
-		*cs++ = upper_32_bits(result->node.start);
-	}
-
-	*cs++ = MI_BATCH_BUFFER_END;
-	i915_gem_object_flush_map(obj);
-	i915_gem_object_unpin_map(obj);
-
-	return vma;
-}
-
-static struct i915_vma *create_global(struct intel_gt *gt, size_t sz)
-{
-	struct drm_i915_gem_object *obj;
-	struct i915_vma *vma;
-	int err;
-
-	obj = i915_gem_object_create_internal(gt->i915, sz);
-	if (IS_ERR(obj))
-		return ERR_CAST(obj);
-
-	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		i915_gem_object_put(obj);
-		return vma;
-	}
-
-	err = i915_ggtt_pin(vma, NULL, 0, 0);
-	if (err) {
-		i915_vma_put(vma);
-		return ERR_PTR(err);
-	}
-
-	return vma;
-}
-
-static struct i915_request *
-create_gpr_client(struct intel_engine_cs *engine,
-		  struct i915_vma *global,
-		  unsigned int offset)
-{
-	struct i915_vma *batch, *vma;
-	struct intel_context *ce;
-	struct i915_request *rq;
-	int err;
-
-	ce = intel_context_create(engine);
-	if (IS_ERR(ce))
-		return ERR_CAST(ce);
-
-	vma = i915_vma_instance(global->obj, ce->vm, NULL);
-	if (IS_ERR(vma)) {
-		err = PTR_ERR(vma);
-		goto out_ce;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_USER);
-	if (err)
-		goto out_ce;
-
-	batch = create_gpr_user(engine, vma, offset);
-	if (IS_ERR(batch)) {
-		err = PTR_ERR(batch);
-		goto out_vma;
-	}
-
-	rq = intel_context_create_request(ce);
-	if (IS_ERR(rq)) {
-		err = PTR_ERR(rq);
-		goto out_batch;
-	}
-
-	i915_vma_lock(vma);
-	err = i915_request_await_object(rq, vma->obj, false);
-	if (!err)
-		err = i915_vma_move_to_active(vma, rq, 0);
-	i915_vma_unlock(vma);
-
-	i915_vma_lock(batch);
-	if (!err)
-		err = i915_request_await_object(rq, batch->obj, false);
-	if (!err)
-		err = i915_vma_move_to_active(batch, rq, 0);
-	if (!err)
-		err = rq->engine->emit_bb_start(rq,
-						batch->node.start,
-						PAGE_SIZE, 0);
-	i915_vma_unlock(batch);
-	i915_vma_unpin(batch);
-
-	if (!err)
-		i915_request_get(rq);
-	i915_request_add(rq);
-
-out_batch:
-	i915_vma_put(batch);
-out_vma:
-	i915_vma_unpin(vma);
-out_ce:
-	intel_context_put(ce);
-	return err ? ERR_PTR(err) : rq;
-}
-
-static int preempt_user(struct intel_engine_cs *engine,
-			struct i915_vma *global,
-			int id)
-{
-	struct i915_sched_attr attr = {
-		.priority = I915_PRIORITY_MAX
-	};
-	struct i915_request *rq;
-	int err = 0;
-	u32 *cs;
-
-	rq = intel_engine_create_kernel_request(engine);
-	if (IS_ERR(rq))
-		return PTR_ERR(rq);
-
-	cs = intel_ring_begin(rq, 4);
-	if (IS_ERR(cs)) {
-		i915_request_add(rq);
-		return PTR_ERR(cs);
-	}
-
-	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
-	*cs++ = i915_ggtt_offset(global);
-	*cs++ = 0;
-	*cs++ = id;
-
-	intel_ring_advance(rq, cs);
-
-	i915_request_get(rq);
-	i915_request_add(rq);
-
-	engine->schedule(rq, &attr);
-
-	if (i915_request_wait(rq, 0, HZ / 2) < 0)
-		err = -ETIME;
-	i915_request_put(rq);
-
-	return err;
-}
-
-static int live_preempt_user(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *engine;
-	struct i915_vma *global;
-	enum intel_engine_id id;
-	u32 *result;
-	int err = 0;
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	/*
-	 * In our other tests, we look at preemption in carefully
-	 * controlled conditions in the ringbuffer. Since most of the
-	 * time is spent in user batches, most of our preemptions naturally
-	 * occur there. We want to verify that when we preempt inside a batch
-	 * we continue on from the current instruction and do not roll back
-	 * to the start, or another earlier arbitration point.
-	 *
-	 * To verify this, we create a batch which is a mixture of
-	 * MI_MATH (gpr++) MI_SRM (gpr) and preemption points. Then with
-	 * a few preempting contexts thrown into the mix, we look for any
-	 * repeated instructions (which show up as incorrect values).
-	 */
-
-	global = create_global(gt, 4096);
-	if (IS_ERR(global))
-		return PTR_ERR(global);
-
-	result = i915_gem_object_pin_map(global->obj, I915_MAP_WC);
-	if (IS_ERR(result)) {
-		i915_vma_unpin_and_release(&global, 0);
-		return PTR_ERR(result);
-	}
-
-	for_each_engine(engine, gt, id) {
-		struct i915_request *client[3] = {};
-		struct igt_live_test t;
-		int i;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		if (IS_GEN(gt->i915, 8) && engine->class != RENDER_CLASS)
-			continue; /* we need per-context GPR */
-
-		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
-			err = -EIO;
-			break;
-		}
-
-		memset(result, 0, 4096);
-
-		for (i = 0; i < ARRAY_SIZE(client); i++) {
-			struct i915_request *rq;
-
-			rq = create_gpr_client(engine, global,
-					       NUM_GPR * i * sizeof(u32));
-			if (IS_ERR(rq))
-				goto end_test;
-
-			client[i] = rq;
-		}
-
-		/* Continuously preempt the set of 3 running contexts */
-		for (i = 1; i <= NUM_GPR; i++) {
-			err = preempt_user(engine, global, i);
-			if (err)
-				goto end_test;
-		}
-
-		if (READ_ONCE(result[0]) != NUM_GPR) {
-			pr_err("%s: Failed to release semaphore\n",
-			       engine->name);
-			err = -EIO;
-			goto end_test;
-		}
-
-		for (i = 0; i < ARRAY_SIZE(client); i++) {
-			int gpr;
-
-			if (i915_request_wait(client[i], 0, HZ / 2) < 0) {
-				err = -ETIME;
-				goto end_test;
-			}
-
-			for (gpr = 1; gpr < NUM_GPR; gpr++) {
-				if (result[NUM_GPR * i + gpr] != 1) {
-					pr_err("%s: Invalid result, client %d, gpr %d, result: %d\n",
-					       engine->name,
-					       i, gpr, result[NUM_GPR * i + gpr]);
-					err = -EINVAL;
-					goto end_test;
-				}
-			}
-		}
-
-end_test:
-		for (i = 0; i < ARRAY_SIZE(client); i++) {
-			if (!client[i])
-				break;
-
-			i915_request_put(client[i]);
-		}
-
-		/* Flush the semaphores on error */
-		smp_store_mb(result[0], -1);
-		if (igt_live_test_end(&t))
-			err = -EIO;
-		if (err)
-			break;
-	}
-
-	i915_vma_unpin_and_release(&global, I915_VMA_RELEASE_MAP);
-	return err;
-}
-
-static int live_preempt_timeout(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct i915_gem_context *ctx_hi, *ctx_lo;
-	struct igt_spinner spin_lo;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	int err = -ENOMEM;
-
-	/*
-	 * Check that we force preemption to occur by cancelling the previous
-	 * context if it refuses to yield the GPU.
-	 */
-	if (!IS_ACTIVE(CONFIG_DRM_I915_PREEMPT_TIMEOUT))
-		return 0;
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(gt->i915))
-		return 0;
-
-	if (!intel_has_reset_engine(gt))
-		return 0;
-
-	if (igt_spinner_init(&spin_lo, gt))
-		return -ENOMEM;
-
-	ctx_hi = kernel_context(gt->i915);
-	if (!ctx_hi)
-		goto err_spin_lo;
-	ctx_hi->sched.priority =
-		I915_USER_PRIORITY(I915_CONTEXT_MAX_USER_PRIORITY);
-
-	ctx_lo = kernel_context(gt->i915);
-	if (!ctx_lo)
-		goto err_ctx_hi;
-	ctx_lo->sched.priority =
-		I915_USER_PRIORITY(I915_CONTEXT_MIN_USER_PRIORITY);
-
-	for_each_engine(engine, gt, id) {
-		unsigned long saved_timeout;
-		struct i915_request *rq;
-
-		if (!intel_engine_has_preemption(engine))
-			continue;
-
-		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
-					    MI_NOOP); /* preemption disabled */
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto err_ctx_lo;
-		}
-
-		i915_request_add(rq);
-		if (!igt_wait_for_spinner(&spin_lo, rq)) {
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-			goto err_ctx_lo;
-		}
-
-		rq = igt_request_alloc(ctx_hi, engine);
-		if (IS_ERR(rq)) {
-			igt_spinner_end(&spin_lo);
-			err = PTR_ERR(rq);
-			goto err_ctx_lo;
-		}
-
-		/* Flush the previous CS ack before changing timeouts */
-		while (READ_ONCE(engine->execlists.pending[0]))
-			cpu_relax();
-
-		saved_timeout = engine->props.preempt_timeout_ms;
-		engine->props.preempt_timeout_ms = 1; /* in ms, -> 1 jiffie */
-
-		i915_request_get(rq);
-		i915_request_add(rq);
-
-		intel_engine_flush_submission(engine);
-		engine->props.preempt_timeout_ms = saved_timeout;
-
-		if (i915_request_wait(rq, 0, HZ / 10) < 0) {
-			intel_gt_set_wedged(gt);
-			i915_request_put(rq);
-			err = -ETIME;
-			goto err_ctx_lo;
-		}
-
-		igt_spinner_end(&spin_lo);
-		i915_request_put(rq);
-	}
-
-	err = 0;
-err_ctx_lo:
-	kernel_context_close(ctx_lo);
-err_ctx_hi:
-	kernel_context_close(ctx_hi);
-err_spin_lo:
-	igt_spinner_fini(&spin_lo);
-	return err;
-}
-
-static int random_range(struct rnd_state *rnd, int min, int max)
-{
-	return i915_prandom_u32_max_state(max - min, rnd) + min;
-}
-
-static int random_priority(struct rnd_state *rnd)
-{
-	return random_range(rnd, I915_PRIORITY_MIN, I915_PRIORITY_MAX);
-}
-
-struct preempt_smoke {
-	struct intel_gt *gt;
-	struct i915_gem_context **contexts;
-	struct intel_engine_cs *engine;
-	struct drm_i915_gem_object *batch;
-	unsigned int ncontext;
-	struct rnd_state prng;
-	unsigned long count;
-};
-
-static struct i915_gem_context *smoke_context(struct preempt_smoke *smoke)
-{
-	return smoke->contexts[i915_prandom_u32_max_state(smoke->ncontext,
-							  &smoke->prng)];
-}
-
-static int smoke_submit(struct preempt_smoke *smoke,
-			struct i915_gem_context *ctx, int prio,
-			struct drm_i915_gem_object *batch)
-{
-	struct i915_request *rq;
-	struct i915_vma *vma = NULL;
-	int err = 0;
-
-	if (batch) {
-		struct i915_address_space *vm;
-
-		vm = i915_gem_context_get_vm_rcu(ctx);
-		vma = i915_vma_instance(batch, vm, NULL);
-		i915_vm_put(vm);
-		if (IS_ERR(vma))
-			return PTR_ERR(vma);
-
-		err = i915_vma_pin(vma, 0, 0, PIN_USER);
-		if (err)
-			return err;
-	}
-
-	ctx->sched.priority = prio;
-
-	rq = igt_request_alloc(ctx, smoke->engine);
-	if (IS_ERR(rq)) {
-		err = PTR_ERR(rq);
-		goto unpin;
-	}
-
-	if (vma) {
-		i915_vma_lock(vma);
-		err = i915_request_await_object(rq, vma->obj, false);
-		if (!err)
-			err = i915_vma_move_to_active(vma, rq, 0);
-		if (!err)
-			err = rq->engine->emit_bb_start(rq,
-							vma->node.start,
-							PAGE_SIZE, 0);
-		i915_vma_unlock(vma);
-	}
-
-	i915_request_add(rq);
-
-unpin:
-	if (vma)
-		i915_vma_unpin(vma);
-
-	return err;
-}
-
-static int smoke_crescendo_thread(void *arg)
-{
-	struct preempt_smoke *smoke = arg;
-	IGT_TIMEOUT(end_time);
-	unsigned long count;
-
-	count = 0;
-	do {
-		struct i915_gem_context *ctx = smoke_context(smoke);
-		int err;
-
-		err = smoke_submit(smoke,
-				   ctx, count % I915_PRIORITY_MAX,
-				   smoke->batch);
-		if (err)
-			return err;
-
-		count++;
-	} while (count < smoke->ncontext && !__igt_timeout(end_time, NULL));
-
-	smoke->count = count;
-	return 0;
-}
-
-static int smoke_crescendo(struct preempt_smoke *smoke, unsigned int flags)
-#define BATCH BIT(0)
-{
-	struct task_struct *tsk[I915_NUM_ENGINES] = {};
-	struct preempt_smoke arg[I915_NUM_ENGINES];
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	unsigned long count;
-	int err = 0;
-
-	for_each_engine(engine, smoke->gt, id) {
-		arg[id] = *smoke;
-		arg[id].engine = engine;
-		if (!(flags & BATCH))
-			arg[id].batch = NULL;
-		arg[id].count = 0;
-
-		tsk[id] = kthread_run(smoke_crescendo_thread, &arg,
-				      "igt/smoke:%d", id);
-		if (IS_ERR(tsk[id])) {
-			err = PTR_ERR(tsk[id]);
-			break;
-		}
-		get_task_struct(tsk[id]);
-	}
-
-	yield(); /* start all threads before we kthread_stop() */
-
-	count = 0;
-	for_each_engine(engine, smoke->gt, id) {
-		int status;
-
-		if (IS_ERR_OR_NULL(tsk[id]))
-			continue;
-
-		status = kthread_stop(tsk[id]);
-		if (status && !err)
-			err = status;
-
-		count += arg[id].count;
-
-		put_task_struct(tsk[id]);
-	}
-
-	pr_info("Submitted %lu crescendo:%x requests across %d engines and %d contexts\n",
-		count, flags, smoke->gt->info.num_engines, smoke->ncontext);
-	return 0;
-}
-
-static int smoke_random(struct preempt_smoke *smoke, unsigned int flags)
-{
-	enum intel_engine_id id;
-	IGT_TIMEOUT(end_time);
-	unsigned long count;
-
-	count = 0;
-	do {
-		for_each_engine(smoke->engine, smoke->gt, id) {
-			struct i915_gem_context *ctx = smoke_context(smoke);
-			int err;
-
-			err = smoke_submit(smoke,
-					   ctx, random_priority(&smoke->prng),
-					   flags & BATCH ? smoke->batch : NULL);
-			if (err)
-				return err;
-
-			count++;
-		}
-	} while (count < smoke->ncontext && !__igt_timeout(end_time, NULL));
-
-	pr_info("Submitted %lu random:%x requests across %d engines and %d contexts\n",
-		count, flags, smoke->gt->info.num_engines, smoke->ncontext);
-	return 0;
-}
-
-static int live_preempt_smoke(void *arg)
-{
-	struct preempt_smoke smoke = {
-		.gt = arg,
-		.prng = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed),
-		.ncontext = 256,
-	};
-	const unsigned int phase[] = { 0, BATCH };
-	struct igt_live_test t;
-	int err = -ENOMEM;
-	u32 *cs;
-	int n;
-
-	if (!HAS_LOGICAL_RING_PREEMPTION(smoke.gt->i915))
-		return 0;
-
-	smoke.contexts = kmalloc_array(smoke.ncontext,
-				       sizeof(*smoke.contexts),
-				       GFP_KERNEL);
-	if (!smoke.contexts)
-		return -ENOMEM;
-
-	smoke.batch =
-		i915_gem_object_create_internal(smoke.gt->i915, PAGE_SIZE);
-	if (IS_ERR(smoke.batch)) {
-		err = PTR_ERR(smoke.batch);
-		goto err_free;
-	}
-
-	cs = i915_gem_object_pin_map(smoke.batch, I915_MAP_WB);
-	if (IS_ERR(cs)) {
-		err = PTR_ERR(cs);
-		goto err_batch;
-	}
-	for (n = 0; n < PAGE_SIZE / sizeof(*cs) - 1; n++)
-		cs[n] = MI_ARB_CHECK;
-	cs[n] = MI_BATCH_BUFFER_END;
-	i915_gem_object_flush_map(smoke.batch);
-	i915_gem_object_unpin_map(smoke.batch);
-
-	if (igt_live_test_begin(&t, smoke.gt->i915, __func__, "all")) {
-		err = -EIO;
-		goto err_batch;
-	}
-
-	for (n = 0; n < smoke.ncontext; n++) {
-		smoke.contexts[n] = kernel_context(smoke.gt->i915);
-		if (!smoke.contexts[n])
-			goto err_ctx;
-	}
-
-	for (n = 0; n < ARRAY_SIZE(phase); n++) {
-		err = smoke_crescendo(&smoke, phase[n]);
-		if (err)
-			goto err_ctx;
-
-		err = smoke_random(&smoke, phase[n]);
-		if (err)
-			goto err_ctx;
-	}
-
-err_ctx:
-	if (igt_live_test_end(&t))
-		err = -EIO;
-
-	for (n = 0; n < smoke.ncontext; n++) {
-		if (!smoke.contexts[n])
-			break;
-		kernel_context_close(smoke.contexts[n]);
-	}
-
-err_batch:
-	i915_gem_object_put(smoke.batch);
-err_free:
-	kfree(smoke.contexts);
-
-	return err;
-}
-
-static int nop_virtual_engine(struct intel_gt *gt,
-			      struct intel_engine_cs **siblings,
-			      unsigned int nsibling,
-			      unsigned int nctx,
-			      unsigned int flags)
-#define CHAIN BIT(0)
-{
-	IGT_TIMEOUT(end_time);
-	struct i915_request *request[16] = {};
-	struct intel_context *ve[16];
-	unsigned long n, prime, nc;
-	struct igt_live_test t;
-	ktime_t times[2] = {};
-	int err;
-
-	GEM_BUG_ON(!nctx || nctx > ARRAY_SIZE(ve));
-
-	for (n = 0; n < nctx; n++) {
-		ve[n] = intel_execlists_create_virtual(siblings, nsibling);
-		if (IS_ERR(ve[n])) {
-			err = PTR_ERR(ve[n]);
-			nctx = n;
-			goto out;
-		}
-
-		err = intel_context_pin(ve[n]);
-		if (err) {
-			intel_context_put(ve[n]);
-			nctx = n;
-			goto out;
-		}
-	}
-
-	err = igt_live_test_begin(&t, gt->i915, __func__, ve[0]->engine->name);
-	if (err)
-		goto out;
-
-	for_each_prime_number_from(prime, 1, 8192) {
-		times[1] = ktime_get_raw();
-
-		if (flags & CHAIN) {
-			for (nc = 0; nc < nctx; nc++) {
-				for (n = 0; n < prime; n++) {
-					struct i915_request *rq;
-
-					rq = i915_request_create(ve[nc]);
-					if (IS_ERR(rq)) {
-						err = PTR_ERR(rq);
-						goto out;
-					}
-
-					if (request[nc])
-						i915_request_put(request[nc]);
-					request[nc] = i915_request_get(rq);
-					i915_request_add(rq);
-				}
-			}
-		} else {
-			for (n = 0; n < prime; n++) {
-				for (nc = 0; nc < nctx; nc++) {
-					struct i915_request *rq;
-
-					rq = i915_request_create(ve[nc]);
-					if (IS_ERR(rq)) {
-						err = PTR_ERR(rq);
-						goto out;
-					}
-
-					if (request[nc])
-						i915_request_put(request[nc]);
-					request[nc] = i915_request_get(rq);
-					i915_request_add(rq);
-				}
-			}
-		}
-
-		for (nc = 0; nc < nctx; nc++) {
-			if (i915_request_wait(request[nc], 0, HZ / 10) < 0) {
-				pr_err("%s(%s): wait for %llx:%lld timed out\n",
-				       __func__, ve[0]->engine->name,
-				       request[nc]->fence.context,
-				       request[nc]->fence.seqno);
-
-				GEM_TRACE("%s(%s) failed at request %llx:%lld\n",
-					  __func__, ve[0]->engine->name,
-					  request[nc]->fence.context,
-					  request[nc]->fence.seqno);
-				GEM_TRACE_DUMP();
-				intel_gt_set_wedged(gt);
-				break;
-			}
-		}
-
-		times[1] = ktime_sub(ktime_get_raw(), times[1]);
-		if (prime == 1)
-			times[0] = times[1];
-
-		for (nc = 0; nc < nctx; nc++) {
-			i915_request_put(request[nc]);
-			request[nc] = NULL;
-		}
-
-		if (__igt_timeout(end_time, NULL))
-			break;
-	}
-
-	err = igt_live_test_end(&t);
-	if (err)
-		goto out;
-
-	pr_info("Requestx%d latencies on %s: 1 = %lluns, %lu = %lluns\n",
-		nctx, ve[0]->engine->name, ktime_to_ns(times[0]),
-		prime, div64_u64(ktime_to_ns(times[1]), prime));
-
-out:
-	if (igt_flush_test(gt->i915))
-		err = -EIO;
-
-	for (nc = 0; nc < nctx; nc++) {
-		i915_request_put(request[nc]);
-		intel_context_unpin(ve[nc]);
-		intel_context_put(ve[nc]);
-	}
-	return err;
-}
-
-static unsigned int
-__select_siblings(struct intel_gt *gt,
-		  unsigned int class,
-		  struct intel_engine_cs **siblings,
-		  bool (*filter)(const struct intel_engine_cs *))
-{
-	unsigned int n = 0;
-	unsigned int inst;
-
-	for (inst = 0; inst <= MAX_ENGINE_INSTANCE; inst++) {
-		if (!gt->engine_class[class][inst])
-			continue;
-
-		if (filter && !filter(gt->engine_class[class][inst]))
-			continue;
-
-		siblings[n++] = gt->engine_class[class][inst];
-	}
-
-	return n;
-}
-
-static unsigned int
-select_siblings(struct intel_gt *gt,
-		unsigned int class,
-		struct intel_engine_cs **siblings)
-{
-	return __select_siblings(gt, class, siblings, NULL);
-}
-
-static int live_virtual_engine(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	unsigned int class;
-	int err;
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0;
-
-	for_each_engine(engine, gt, id) {
-		err = nop_virtual_engine(gt, &engine, 1, 1, 0);
-		if (err) {
-			pr_err("Failed to wrap engine %s: err=%d\n",
-			       engine->name, err);
-			return err;
-		}
-	}
-
-	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
-		int nsibling, n;
-
-		nsibling = select_siblings(gt, class, siblings);
-		if (nsibling < 2)
-			continue;
-
-		for (n = 1; n <= nsibling + 1; n++) {
-			err = nop_virtual_engine(gt, siblings, nsibling,
-						 n, 0);
-			if (err)
-				return err;
-		}
-
-		err = nop_virtual_engine(gt, siblings, nsibling, n, CHAIN);
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-static int mask_virtual_engine(struct intel_gt *gt,
-			       struct intel_engine_cs **siblings,
-			       unsigned int nsibling)
-{
-	struct i915_request *request[MAX_ENGINE_INSTANCE + 1];
-	struct intel_context *ve;
-	struct igt_live_test t;
-	unsigned int n;
-	int err;
-
-	/*
-	 * Check that by setting the execution mask on a request, we can
-	 * restrict it to our desired engine within the virtual engine.
-	 */
-
-	ve = intel_execlists_create_virtual(siblings, nsibling);
-	if (IS_ERR(ve)) {
-		err = PTR_ERR(ve);
-		goto out_close;
-	}
-
-	err = intel_context_pin(ve);
-	if (err)
-		goto out_put;
-
-	err = igt_live_test_begin(&t, gt->i915, __func__, ve->engine->name);
-	if (err)
-		goto out_unpin;
-
-	for (n = 0; n < nsibling; n++) {
-		request[n] = i915_request_create(ve);
-		if (IS_ERR(request[n])) {
-			err = PTR_ERR(request[n]);
-			nsibling = n;
-			goto out;
-		}
-
-		/* Reverse order as it's more likely to be unnatural */
-		request[n]->execution_mask = siblings[nsibling - n - 1]->mask;
-
-		i915_request_get(request[n]);
-		i915_request_add(request[n]);
-	}
-
-	for (n = 0; n < nsibling; n++) {
-		if (i915_request_wait(request[n], 0, HZ / 10) < 0) {
-			pr_err("%s(%s): wait for %llx:%lld timed out\n",
-			       __func__, ve->engine->name,
-			       request[n]->fence.context,
-			       request[n]->fence.seqno);
-
-			GEM_TRACE("%s(%s) failed at request %llx:%lld\n",
-				  __func__, ve->engine->name,
-				  request[n]->fence.context,
-				  request[n]->fence.seqno);
-			GEM_TRACE_DUMP();
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-			goto out;
-		}
-
-		if (request[n]->engine != siblings[nsibling - n - 1]) {
-			pr_err("Executed on wrong sibling '%s', expected '%s'\n",
-			       request[n]->engine->name,
-			       siblings[nsibling - n - 1]->name);
-			err = -EINVAL;
-			goto out;
-		}
-	}
-
-	err = igt_live_test_end(&t);
-out:
-	if (igt_flush_test(gt->i915))
-		err = -EIO;
-
-	for (n = 0; n < nsibling; n++)
-		i915_request_put(request[n]);
-
-out_unpin:
-	intel_context_unpin(ve);
-out_put:
-	intel_context_put(ve);
-out_close:
-	return err;
-}
-
-static int live_virtual_mask(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
-	unsigned int class;
-	int err;
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0;
-
-	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
-		unsigned int nsibling;
-
-		nsibling = select_siblings(gt, class, siblings);
-		if (nsibling < 2)
-			continue;
-
-		err = mask_virtual_engine(gt, siblings, nsibling);
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-static int slicein_virtual_engine(struct intel_gt *gt,
-				  struct intel_engine_cs **siblings,
-				  unsigned int nsibling)
-{
-	const long timeout = slice_timeout(siblings[0]);
-	struct intel_context *ce;
-	struct i915_request *rq;
-	struct igt_spinner spin;
-	unsigned int n;
-	int err = 0;
-
-	/*
-	 * Virtual requests must take part in timeslicing on the target engines.
-	 */
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	for (n = 0; n < nsibling; n++) {
-		ce = intel_context_create(siblings[n]);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto out;
-		}
-
-		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
-		intel_context_put(ce);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out;
-		}
-
-		i915_request_add(rq);
-	}
-
-	ce = intel_execlists_create_virtual(siblings, nsibling);
-	if (IS_ERR(ce)) {
-		err = PTR_ERR(ce);
-		goto out;
-	}
-
-	rq = intel_context_create_request(ce);
-	intel_context_put(ce);
-	if (IS_ERR(rq)) {
-		err = PTR_ERR(rq);
-		goto out;
-	}
-
-	i915_request_get(rq);
-	i915_request_add(rq);
-	if (i915_request_wait(rq, 0, timeout) < 0) {
-		GEM_TRACE_ERR("%s(%s) failed to slice in virtual request\n",
-			      __func__, rq->engine->name);
-		GEM_TRACE_DUMP();
-		intel_gt_set_wedged(gt);
-		err = -EIO;
-	}
-	i915_request_put(rq);
-
-out:
-	igt_spinner_end(&spin);
-	if (igt_flush_test(gt->i915))
-		err = -EIO;
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int sliceout_virtual_engine(struct intel_gt *gt,
-				   struct intel_engine_cs **siblings,
-				   unsigned int nsibling)
-{
-	const long timeout = slice_timeout(siblings[0]);
-	struct intel_context *ce;
-	struct i915_request *rq;
-	struct igt_spinner spin;
-	unsigned int n;
-	int err = 0;
-
-	/*
-	 * Virtual requests must allow others a fair timeslice.
-	 */
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	/* XXX We do not handle oversubscription and fairness with normal rq */
-	for (n = 0; n < nsibling; n++) {
-		ce = intel_execlists_create_virtual(siblings, nsibling);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto out;
-		}
-
-		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
-		intel_context_put(ce);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out;
-		}
-
-		i915_request_add(rq);
-	}
-
-	for (n = 0; !err && n < nsibling; n++) {
-		ce = intel_context_create(siblings[n]);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto out;
-		}
-
-		rq = intel_context_create_request(ce);
-		intel_context_put(ce);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out;
-		}
-
-		i915_request_get(rq);
-		i915_request_add(rq);
-		if (i915_request_wait(rq, 0, timeout) < 0) {
-			GEM_TRACE_ERR("%s(%s) failed to slice out virtual request\n",
-				      __func__, siblings[n]->name);
-			GEM_TRACE_DUMP();
-			intel_gt_set_wedged(gt);
-			err = -EIO;
-		}
-		i915_request_put(rq);
-	}
-
-out:
-	igt_spinner_end(&spin);
-	if (igt_flush_test(gt->i915))
-		err = -EIO;
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_virtual_slice(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
-	unsigned int class;
-	int err;
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0;
-
-	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
-		unsigned int nsibling;
-
-		nsibling = __select_siblings(gt, class, siblings,
-					     intel_engine_has_timeslices);
-		if (nsibling < 2)
-			continue;
-
-		err = slicein_virtual_engine(gt, siblings, nsibling);
-		if (err)
-			return err;
-
-		err = sliceout_virtual_engine(gt, siblings, nsibling);
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-static int preserved_virtual_engine(struct intel_gt *gt,
-				    struct intel_engine_cs **siblings,
-				    unsigned int nsibling)
-{
-	struct i915_request *last = NULL;
-	struct intel_context *ve;
-	struct i915_vma *scratch;
-	struct igt_live_test t;
-	unsigned int n;
-	int err = 0;
-	u32 *cs;
-
-	scratch = create_scratch(siblings[0]->gt);
-	if (IS_ERR(scratch))
-		return PTR_ERR(scratch);
-
-	err = i915_vma_sync(scratch);
-	if (err)
-		goto out_scratch;
-
-	ve = intel_execlists_create_virtual(siblings, nsibling);
-	if (IS_ERR(ve)) {
-		err = PTR_ERR(ve);
-		goto out_scratch;
-	}
-
-	err = intel_context_pin(ve);
-	if (err)
-		goto out_put;
-
-	err = igt_live_test_begin(&t, gt->i915, __func__, ve->engine->name);
-	if (err)
-		goto out_unpin;
-
-	for (n = 0; n < NUM_GPR_DW; n++) {
-		struct intel_engine_cs *engine = siblings[n % nsibling];
-		struct i915_request *rq;
-
-		rq = i915_request_create(ve);
-		if (IS_ERR(rq)) {
-			err = PTR_ERR(rq);
-			goto out_end;
-		}
-
-		i915_request_put(last);
-		last = i915_request_get(rq);
-
-		cs = intel_ring_begin(rq, 8);
-		if (IS_ERR(cs)) {
-			i915_request_add(rq);
-			err = PTR_ERR(cs);
-			goto out_end;
-		}
-
-		*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
-		*cs++ = CS_GPR(engine, n);
-		*cs++ = i915_ggtt_offset(scratch) + n * sizeof(u32);
-		*cs++ = 0;
-
-		*cs++ = MI_LOAD_REGISTER_IMM(1);
-		*cs++ = CS_GPR(engine, (n + 1) % NUM_GPR_DW);
-		*cs++ = n + 1;
-
-		*cs++ = MI_NOOP;
-		intel_ring_advance(rq, cs);
-
-		/* Restrict this request to run on a particular engine */
-		rq->execution_mask = engine->mask;
-		i915_request_add(rq);
-	}
-
-	if (i915_request_wait(last, 0, HZ / 5) < 0) {
-		err = -ETIME;
-		goto out_end;
-	}
-
-	cs = i915_gem_object_pin_map(scratch->obj, I915_MAP_WB);
-	if (IS_ERR(cs)) {
-		err = PTR_ERR(cs);
-		goto out_end;
-	}
-
-	for (n = 0; n < NUM_GPR_DW; n++) {
-		if (cs[n] != n) {
-			pr_err("Incorrect value[%d] found for GPR[%d]\n",
-			       cs[n], n);
-			err = -EINVAL;
-			break;
-		}
-	}
-
-	i915_gem_object_unpin_map(scratch->obj);
-
-out_end:
-	if (igt_live_test_end(&t))
-		err = -EIO;
-	i915_request_put(last);
-out_unpin:
-	intel_context_unpin(ve);
-out_put:
-	intel_context_put(ve);
-out_scratch:
-	i915_vma_unpin_and_release(&scratch, 0);
-	return err;
-}
-
-static int live_virtual_preserved(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
-	unsigned int class;
-
-	/*
-	 * Check that the context image retains non-privileged (user) registers
-	 * from one engine to the next. For this we check that the CS_GPR
-	 * are preserved.
-	 */
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0;
-
-	/* As we use CS_GPR we cannot run before they existed on all engines. */
-	if (INTEL_GEN(gt->i915) < 9)
-		return 0;
-
-	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
-		int nsibling, err;
-
-		nsibling = select_siblings(gt, class, siblings);
-		if (nsibling < 2)
-			continue;
-
-		err = preserved_virtual_engine(gt, siblings, nsibling);
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-static int bond_virtual_engine(struct intel_gt *gt,
-			       unsigned int class,
-			       struct intel_engine_cs **siblings,
-			       unsigned int nsibling,
-			       unsigned int flags)
-#define BOND_SCHEDULE BIT(0)
-{
-	struct intel_engine_cs *master;
-	struct i915_request *rq[16];
-	enum intel_engine_id id;
-	struct igt_spinner spin;
-	unsigned long n;
-	int err;
-
-	/*
-	 * A set of bonded requests is intended to be run concurrently
-	 * across a number of engines. We use one request per-engine
-	 * and a magic fence to schedule each of the bonded requests
-	 * at the same time. A consequence of our current scheduler is that
-	 * we only move requests to the HW ready queue when the request
-	 * becomes ready, that is when all of its prerequisite fences have
-	 * been signaled. As one of those fences is the master submit fence,
-	 * there is a delay on all secondary fences as the HW may be
-	 * currently busy. Equally, as all the requests are independent,
-	 * they may have other fences that delay individual request
-	 * submission to HW. Ergo, we do not guarantee that all requests are
-	 * immediately submitted to HW at the same time, just that if the
-	 * rules are abided by, they are ready at the same time as the
-	 * first is submitted. Userspace can embed semaphores in its batch
-	 * to ensure parallel execution of its phases as it requires.
-	 * Though naturally it gets requested that perhaps the scheduler should
-	 * take care of parallel execution, even across preemption events on
-	 * different HW. (The proper answer is of course "lalalala".)
-	 *
-	 * With the submit-fence, we have identified three possible phases
-	 * of synchronisation depending on the master fence: queued (not
-	 * ready), executing, and signaled. The first two are quite simple
-	 * and checked below. However, the signaled master fence handling is
-	 * contentious. Currently we do not distinguish between a signaled
-	 * fence and an expired fence, as once signaled it does not convey
-	 * any information about the previous execution. It may even be freed
-	 * and hence checking later it may not exist at all. Ergo we currently
-	 * do not apply the bonding constraint for an already signaled fence,
-	 * as our expectation is that it should not constrain the secondaries
-	 * and is outside of the scope of the bonded request API (i.e. all
-	 * userspace requests are meant to be running in parallel). As
-	 * it imposes no constraint, and is effectively a no-op, we do not
-	 * check below as normal execution flows are checked extensively above.
-	 *
-	 * XXX Is the degenerate handling of signaled submit fences the
-	 * expected behaviour for userpace?
-	 */
-
-	GEM_BUG_ON(nsibling >= ARRAY_SIZE(rq) - 1);
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	err = 0;
-	rq[0] = ERR_PTR(-ENOMEM);
-	for_each_engine(master, gt, id) {
-		struct i915_sw_fence fence = {};
-		struct intel_context *ce;
-
-		if (master->class == class)
-			continue;
-
-		ce = intel_context_create(master);
-		if (IS_ERR(ce)) {
-			err = PTR_ERR(ce);
-			goto out;
-		}
-
-		memset_p((void *)rq, ERR_PTR(-EINVAL), ARRAY_SIZE(rq));
-
-		rq[0] = igt_spinner_create_request(&spin, ce, MI_NOOP);
-		intel_context_put(ce);
-		if (IS_ERR(rq[0])) {
-			err = PTR_ERR(rq[0]);
-			goto out;
-		}
-		i915_request_get(rq[0]);
-
-		if (flags & BOND_SCHEDULE) {
-			onstack_fence_init(&fence);
-			err = i915_sw_fence_await_sw_fence_gfp(&rq[0]->submit,
-							       &fence,
-							       GFP_KERNEL);
-		}
-
-		i915_request_add(rq[0]);
-		if (err < 0)
-			goto out;
-
-		if (!(flags & BOND_SCHEDULE) &&
-		    !igt_wait_for_spinner(&spin, rq[0])) {
-			err = -EIO;
-			goto out;
-		}
-
-		for (n = 0; n < nsibling; n++) {
-			struct intel_context *ve;
-
-			ve = intel_execlists_create_virtual(siblings, nsibling);
-			if (IS_ERR(ve)) {
-				err = PTR_ERR(ve);
-				onstack_fence_fini(&fence);
-				goto out;
-			}
-
-			err = intel_virtual_engine_attach_bond(ve->engine,
-							       master,
-							       siblings[n]);
-			if (err) {
-				intel_context_put(ve);
-				onstack_fence_fini(&fence);
-				goto out;
-			}
-
-			err = intel_context_pin(ve);
-			intel_context_put(ve);
-			if (err) {
-				onstack_fence_fini(&fence);
-				goto out;
-			}
-
-			rq[n + 1] = i915_request_create(ve);
-			intel_context_unpin(ve);
-			if (IS_ERR(rq[n + 1])) {
-				err = PTR_ERR(rq[n + 1]);
-				onstack_fence_fini(&fence);
-				goto out;
-			}
-			i915_request_get(rq[n + 1]);
-
-			err = i915_request_await_execution(rq[n + 1],
-							   &rq[0]->fence,
-							   ve->engine->bond_execute);
-			i915_request_add(rq[n + 1]);
-			if (err < 0) {
-				onstack_fence_fini(&fence);
-				goto out;
-			}
-		}
-		onstack_fence_fini(&fence);
-		intel_engine_flush_submission(master);
-		igt_spinner_end(&spin);
-
-		if (i915_request_wait(rq[0], 0, HZ / 10) < 0) {
-			pr_err("Master request did not execute (on %s)!\n",
-			       rq[0]->engine->name);
-			err = -EIO;
-			goto out;
-		}
-
-		for (n = 0; n < nsibling; n++) {
-			if (i915_request_wait(rq[n + 1], 0,
-					      MAX_SCHEDULE_TIMEOUT) < 0) {
-				err = -EIO;
-				goto out;
-			}
-
-			if (rq[n + 1]->engine != siblings[n]) {
-				pr_err("Bonded request did not execute on target engine: expected %s, used %s; master was %s\n",
-				       siblings[n]->name,
-				       rq[n + 1]->engine->name,
-				       rq[0]->engine->name);
-				err = -EINVAL;
-				goto out;
-			}
-		}
-
-		for (n = 0; !IS_ERR(rq[n]); n++)
-			i915_request_put(rq[n]);
-		rq[0] = ERR_PTR(-ENOMEM);
-	}
-
-out:
-	for (n = 0; !IS_ERR(rq[n]); n++)
-		i915_request_put(rq[n]);
-	if (igt_flush_test(gt->i915))
-		err = -EIO;
-
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_virtual_bond(void *arg)
-{
-	static const struct phase {
-		const char *name;
-		unsigned int flags;
-	} phases[] = {
-		{ "", 0 },
-		{ "schedule", BOND_SCHEDULE },
-		{ },
-	};
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
-	unsigned int class;
-	int err;
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0;
-
-	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
-		const struct phase *p;
-		int nsibling;
-
-		nsibling = select_siblings(gt, class, siblings);
-		if (nsibling < 2)
-			continue;
-
-		for (p = phases; p->name; p++) {
-			err = bond_virtual_engine(gt,
-						  class, siblings, nsibling,
-						  p->flags);
-			if (err) {
-				pr_err("%s(%s): failed class=%d, nsibling=%d, err=%d\n",
-				       __func__, p->name, class, nsibling, err);
-				return err;
-			}
-		}
-	}
-
-	return 0;
-}
-
-static int reset_virtual_engine(struct intel_gt *gt,
-				struct intel_engine_cs **siblings,
-				unsigned int nsibling)
-{
-	struct intel_engine_cs *engine;
-	struct intel_context *ve;
-	struct igt_spinner spin;
-	struct i915_request *rq;
-	unsigned int n;
-	int err = 0;
-
-	/*
-	 * In order to support offline error capture for fast preempt reset,
-	 * we need to decouple the guilty request and ensure that it and its
-	 * descendents are not executed while the capture is in progress.
-	 */
-
-	if (igt_spinner_init(&spin, gt))
-		return -ENOMEM;
-
-	ve = intel_execlists_create_virtual(siblings, nsibling);
-	if (IS_ERR(ve)) {
-		err = PTR_ERR(ve);
-		goto out_spin;
-	}
-
-	for (n = 0; n < nsibling; n++)
-		st_engine_heartbeat_disable(siblings[n]);
-
-	rq = igt_spinner_create_request(&spin, ve, MI_ARB_CHECK);
-	if (IS_ERR(rq)) {
-		err = PTR_ERR(rq);
-		goto out_heartbeat;
-	}
-	i915_request_add(rq);
-
-	if (!igt_wait_for_spinner(&spin, rq)) {
-		intel_gt_set_wedged(gt);
-		err = -ETIME;
-		goto out_heartbeat;
-	}
-
-	engine = rq->engine;
-	GEM_BUG_ON(engine == ve->engine);
-
-	/* Take ownership of the reset and tasklet */
-	if (test_and_set_bit(I915_RESET_ENGINE + engine->id,
-			     &gt->reset.flags)) {
-		intel_gt_set_wedged(gt);
-		err = -EBUSY;
-		goto out_heartbeat;
-	}
-	tasklet_disable(&engine->execlists.tasklet);
-
-	engine->execlists.tasklet.func(engine->execlists.tasklet.data);
-	GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
-
-	/* Fake a preemption event; failed of course */
-	spin_lock_irq(&engine->active.lock);
-	__unwind_incomplete_requests(engine);
-	spin_unlock_irq(&engine->active.lock);
-	GEM_BUG_ON(rq->engine != ve->engine);
-
-	/* Reset the engine while keeping our active request on hold */
-	execlists_hold(engine, rq);
-	GEM_BUG_ON(!i915_request_on_hold(rq));
-
-	intel_engine_reset(engine, NULL);
-	GEM_BUG_ON(rq->fence.error != -EIO);
-
-	/* Release our grasp on the engine, letting CS flow again */
-	tasklet_enable(&engine->execlists.tasklet);
-	clear_and_wake_up_bit(I915_RESET_ENGINE + engine->id, &gt->reset.flags);
-
-	/* Check that we do not resubmit the held request */
-	i915_request_get(rq);
-	if (!i915_request_wait(rq, 0, HZ / 5)) {
-		pr_err("%s: on hold request completed!\n",
-		       engine->name);
-		intel_gt_set_wedged(gt);
-		err = -EIO;
-		goto out_rq;
-	}
-	GEM_BUG_ON(!i915_request_on_hold(rq));
-
-	/* But is resubmitted on release */
-	execlists_unhold(engine, rq);
-	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
-		pr_err("%s: held request did not complete!\n",
-		       engine->name);
-		intel_gt_set_wedged(gt);
-		err = -ETIME;
-	}
-
-out_rq:
-	i915_request_put(rq);
-out_heartbeat:
-	for (n = 0; n < nsibling; n++)
-		st_engine_heartbeat_enable(siblings[n]);
-
-	intel_context_put(ve);
-out_spin:
-	igt_spinner_fini(&spin);
-	return err;
-}
-
-static int live_virtual_reset(void *arg)
-{
-	struct intel_gt *gt = arg;
-	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
-	unsigned int class;
-
-	/*
-	 * Check that we handle a reset event within a virtual engine.
-	 * Only the physical engine is reset, but we have to check the flow
-	 * of the virtual requests around the reset, and make sure it is not
-	 * forgotten.
-	 */
-
-	if (intel_uc_uses_guc_submission(&gt->uc))
-		return 0;
-
-	if (!intel_has_reset_engine(gt))
-		return 0;
-
-	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
-		int nsibling, err;
-
-		nsibling = select_siblings(gt, class, siblings);
-		if (nsibling < 2)
-			continue;
-
-		err = reset_virtual_engine(gt, siblings, nsibling);
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
-int intel_execlists_live_selftests(struct drm_i915_private *i915)
-{
-	static const struct i915_subtest tests[] = {
-		SUBTEST(live_sanitycheck),
-		SUBTEST(live_unlite_switch),
-		SUBTEST(live_unlite_preempt),
-		SUBTEST(live_unlite_ring),
-		SUBTEST(live_pin_rewind),
-		SUBTEST(live_hold_reset),
-		SUBTEST(live_error_interrupt),
-		SUBTEST(live_timeslice_preempt),
-		SUBTEST(live_timeslice_rewind),
-		SUBTEST(live_timeslice_queue),
-		SUBTEST(live_timeslice_nopreempt),
-		SUBTEST(live_busywait_preempt),
-		SUBTEST(live_preempt),
-		SUBTEST(live_late_preempt),
-		SUBTEST(live_nopreempt),
-		SUBTEST(live_preempt_cancel),
-		SUBTEST(live_suppress_self_preempt),
-		SUBTEST(live_chain_preempt),
-		SUBTEST(live_preempt_ring),
-		SUBTEST(live_preempt_gang),
-		SUBTEST(live_preempt_timeout),
-		SUBTEST(live_preempt_user),
-		SUBTEST(live_preempt_smoke),
-		SUBTEST(live_virtual_engine),
-		SUBTEST(live_virtual_mask),
-		SUBTEST(live_virtual_preserved),
-		SUBTEST(live_virtual_slice),
-		SUBTEST(live_virtual_bond),
-		SUBTEST(live_virtual_reset),
-	};
-
-	if (!HAS_EXECLISTS(i915))
-		return 0;
-
-	if (intel_gt_is_wedged(&i915->gt))
-		return 0;
-
-	return intel_gt_live_subtests(tests, &i915->gt);
-}
-
 static int emit_semaphore_signal(struct intel_context *ce, void *slot)
 {
 	const u32 offset =
@@ -4775,9 +139,10 @@ static int live_lrc_layout(void *arg)
 	 * match the layout saved by HW.
 	 */
 
-	lrc = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	lrc = (u32 *)__get_free_page(GFP_KERNEL); /* requires page alignment */
 	if (!lrc)
 		return -ENOMEM;
+	GEM_BUG_ON(offset_in_page(lrc));
 
 	err = 0;
 	for_each_engine(engine, gt, id) {
@@ -4794,15 +159,12 @@ static int live_lrc_layout(void *arg)
 		}
 		hw += LRC_STATE_OFFSET / sizeof(*hw);
 
-		execlists_init_reg_state(memset(lrc, POISON_INUSE, PAGE_SIZE),
-					 engine->kernel_context,
-					 engine,
-					 engine->kernel_context->ring,
-					 true);
+		__lrc_init_regs(memset(lrc, POISON_INUSE, PAGE_SIZE),
+				engine->kernel_context, engine, true);
 
 		dw = 0;
 		do {
-			u32 lri = hw[dw];
+			u32 lri = READ_ONCE(hw[dw]);
 
 			if (lri == 0) {
 				dw++;
@@ -4835,9 +197,11 @@ static int live_lrc_layout(void *arg)
 			dw++;
 
 			while (lri) {
-				if (hw[dw] != lrc[dw]) {
+				u32 offset = READ_ONCE(hw[dw]);
+
+				if (offset != lrc[dw]) {
 					pr_err("%s: Different registers found at dword %d, expected %x, found %x\n",
-					       engine->name, dw, hw[dw], lrc[dw]);
+					       engine->name, dw, offset, lrc[dw]);
 					err = -EINVAL;
 					break;
 				}
@@ -4849,7 +213,7 @@ static int live_lrc_layout(void *arg)
 				dw += 2;
 				lri -= 2;
 			}
-		} while ((lrc[dw] & ~BIT(0)) != MI_BATCH_BUFFER_END);
+		} while (!err && (lrc[dw] & ~BIT(0)) != MI_BATCH_BUFFER_END);
 
 		if (err) {
 			pr_info("%s: HW register image:\n", engine->name);
@@ -4864,7 +228,7 @@ static int live_lrc_layout(void *arg)
 			break;
 	}
 
-	kfree(lrc);
+	free_page((unsigned long)lrc);
 	return err;
 }
 
@@ -5249,6 +613,10 @@ static int __live_lrc_gpr(struct intel_engine_cs *engine,
 		err = emit_semaphore_signal(engine->kernel_context, slot);
 		if (err)
 			goto err_rq;
+
+		err = wait_for_submit(engine, rq, HZ / 2);
+		if (err)
+			goto err_rq;
 	} else {
 		slot[0] = 1;
 		wmb();
@@ -6242,16 +1610,17 @@ static void garbage_reset(struct intel_engine_cs *engine,
 	const unsigned int bit = I915_RESET_ENGINE + engine->id;
 	unsigned long *lock = &engine->gt->reset.flags;
 
-	if (test_and_set_bit(bit, lock))
-		return;
-
-	tasklet_disable(&engine->execlists.tasklet);
+	local_bh_disable();
+	if (!test_and_set_bit(bit, lock)) {
+		tasklet_disable(&engine->execlists.tasklet);
 
-	if (!rq->fence.error)
-		intel_engine_reset(engine, NULL);
+		if (!rq->fence.error)
+			__intel_engine_reset_bh(engine, NULL);
 
-	tasklet_enable(&engine->execlists.tasklet);
-	clear_and_wake_up_bit(bit, lock);
+		tasklet_enable(&engine->execlists.tasklet);
+		clear_and_wake_up_bit(bit, lock);
+	}
+	local_bh_enable();
 }
 
 static struct i915_request *garbage(struct intel_context *ce,
diff --git a/drivers/gpu/drm/i915/gt/selftest_mocs.c b/drivers/gpu/drm/i915/gt/selftest_mocs.c
index b25eba50c88e..cf373c72359e 100644
--- a/drivers/gpu/drm/i915/gt/selftest_mocs.c
+++ b/drivers/gpu/drm/i915/gt/selftest_mocs.c
@@ -5,6 +5,7 @@
  */
 
 #include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
 #include "i915_selftest.h"
 
 #include "gem/selftests/mock_context.h"
@@ -56,33 +57,6 @@ static int request_add_spin(struct i915_request *rq, struct igt_spinner *spin)
 	return err;
 }
 
-static struct i915_vma *create_scratch(struct intel_gt *gt)
-{
-	struct drm_i915_gem_object *obj;
-	struct i915_vma *vma;
-	int err;
-
-	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
-	if (IS_ERR(obj))
-		return ERR_CAST(obj);
-
-	i915_gem_object_set_cache_coherency(obj, I915_CACHING_CACHED);
-
-	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
-	if (IS_ERR(vma)) {
-		i915_gem_object_put(obj);
-		return vma;
-	}
-
-	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
-	if (err) {
-		i915_gem_object_put(obj);
-		return ERR_PTR(err);
-	}
-
-	return vma;
-}
-
 static int live_mocs_init(struct live_mocs *arg, struct intel_gt *gt)
 {
 	struct drm_i915_mocs_table table;
@@ -101,7 +75,7 @@ static int live_mocs_init(struct live_mocs *arg, struct intel_gt *gt)
 	if (flags & (HAS_GLOBAL_MOCS | HAS_ENGINE_MOCS))
 		arg->mocs = table;
 
-	arg->scratch = create_scratch(gt);
+	arg->scratch = __vm_create_scratch_for_read(&gt->ggtt->vm, PAGE_SIZE);
 	if (IS_ERR(arg->scratch))
 		return PTR_ERR(arg->scratch);
 
@@ -125,7 +99,7 @@ static void live_mocs_fini(struct live_mocs *arg)
 
 static int read_regs(struct i915_request *rq,
 		     u32 addr, unsigned int count,
-		     uint32_t *offset)
+		     u32 *offset)
 {
 	unsigned int i;
 	u32 *cs;
@@ -153,7 +127,7 @@ static int read_regs(struct i915_request *rq,
 
 static int read_mocs_table(struct i915_request *rq,
 			   const struct drm_i915_mocs_table *table,
-			   uint32_t *offset)
+			   u32 *offset)
 {
 	u32 addr;
 
@@ -167,7 +141,7 @@ static int read_mocs_table(struct i915_request *rq,
 
 static int read_l3cc_table(struct i915_request *rq,
 			   const struct drm_i915_mocs_table *table,
-			   uint32_t *offset)
+			   u32 *offset)
 {
 	u32 addr = i915_mmio_reg_offset(GEN9_LNCFCMOCS(0));
 
@@ -176,7 +150,7 @@ static int read_l3cc_table(struct i915_request *rq,
 
 static int check_mocs_table(struct intel_engine_cs *engine,
 			    const struct drm_i915_mocs_table *table,
-			    uint32_t **vaddr)
+			    u32 **vaddr)
 {
 	unsigned int i;
 	u32 expect;
@@ -205,7 +179,7 @@ static bool mcr_range(struct drm_i915_private *i915, u32 offset)
 
 static int check_l3cc_table(struct intel_engine_cs *engine,
 			    const struct drm_i915_mocs_table *table,
-			    uint32_t **vaddr)
+			    u32 **vaddr)
 {
 	/* Can we read the MCR range 0xb00 directly? See intel_workarounds! */
 	u32 reg = i915_mmio_reg_offset(GEN9_LNCFCMOCS(0));
@@ -361,29 +335,34 @@ static int active_engine_reset(struct intel_context *ce,
 static int __live_mocs_reset(struct live_mocs *mocs,
 			     struct intel_context *ce)
 {
+	struct intel_gt *gt = ce->engine->gt;
 	int err;
 
-	err = intel_engine_reset(ce->engine, "mocs");
-	if (err)
-		return err;
+	if (intel_has_reset_engine(gt)) {
+		err = intel_engine_reset(ce->engine, "mocs");
+		if (err)
+			return err;
 
-	err = check_mocs_engine(mocs, ce);
-	if (err)
-		return err;
+		err = check_mocs_engine(mocs, ce);
+		if (err)
+			return err;
 
-	err = active_engine_reset(ce, "mocs");
-	if (err)
-		return err;
+		err = active_engine_reset(ce, "mocs");
+		if (err)
+			return err;
 
-	err = check_mocs_engine(mocs, ce);
-	if (err)
-		return err;
+		err = check_mocs_engine(mocs, ce);
+		if (err)
+			return err;
+	}
 
-	intel_gt_reset(ce->engine->gt, ce->engine->mask, "mocs");
+	if (intel_has_gpu_reset(gt)) {
+		intel_gt_reset(gt, ce->engine->mask, "mocs");
 
-	err = check_mocs_engine(mocs, ce);
-	if (err)
-		return err;
+		err = check_mocs_engine(mocs, ce);
+		if (err)
+			return err;
+	}
 
 	return 0;
 }
@@ -398,9 +377,6 @@ static int live_mocs_reset(void *arg)
 
 	/* Check the mocs setup is retained over per-engine and global resets */
 
-	if (!intel_has_reset_engine(gt))
-		return 0;
-
 	err = live_mocs_init(&mocs, gt);
 	if (err)
 		return err;
diff --git a/drivers/gpu/drm/i915/gt/selftest_rc6.c b/drivers/gpu/drm/i915/gt/selftest_rc6.c
index 64ef5ee5decf..61abc0556601 100644
--- a/drivers/gpu/drm/i915/gt/selftest_rc6.c
+++ b/drivers/gpu/drm/i915/gt/selftest_rc6.c
@@ -6,6 +6,7 @@
 
 #include "intel_context.h"
 #include "intel_engine_pm.h"
+#include "intel_gpu_commands.h"
 #include "intel_gt_requests.h"
 #include "intel_ring.h"
 #include "selftest_rc6.h"
diff --git a/drivers/gpu/drm/i915/gt/selftest_reset.c b/drivers/gpu/drm/i915/gt/selftest_reset.c
index ef5aeebbeeb0..8784257ec808 100644
--- a/drivers/gpu/drm/i915/gt/selftest_reset.c
+++ b/drivers/gpu/drm/i915/gt/selftest_reset.c
@@ -9,6 +9,7 @@
 
 #include "i915_memcpy.h"
 #include "i915_selftest.h"
+#include "intel_gpu_commands.h"
 #include "selftests/igt_reset.h"
 #include "selftests/igt_atomic.h"
 #include "selftests/igt_spinner.h"
@@ -95,10 +96,10 @@ __igt_reset_stolen(struct intel_gt *gt,
 		if (!__drm_mm_interval_first(&gt->i915->mm.stolen,
 					     page << PAGE_SHIFT,
 					     ((page + 1) << PAGE_SHIFT) - 1))
-			memset32(s, STACK_MAGIC, PAGE_SIZE / sizeof(u32));
+			memset_io(s, STACK_MAGIC, PAGE_SIZE);
 
-		in = s;
-		if (i915_memcpy_from_wc(tmp, s, PAGE_SIZE))
+		in = (void __force *)s;
+		if (i915_memcpy_from_wc(tmp, in, PAGE_SIZE))
 			in = tmp;
 		crc[page] = crc32_le(0, in, PAGE_SIZE);
 
@@ -133,8 +134,8 @@ __igt_reset_stolen(struct intel_gt *gt,
 				      ggtt->error_capture.start,
 				      PAGE_SIZE);
 
-		in = s;
-		if (i915_memcpy_from_wc(tmp, s, PAGE_SIZE))
+		in = (void __force *)s;
+		if (i915_memcpy_from_wc(tmp, in, PAGE_SIZE))
 			in = tmp;
 		x = crc32_le(0, in, PAGE_SIZE);
 
@@ -320,17 +321,25 @@ static int igt_atomic_engine_reset(void *arg)
 		goto out_unlock;
 
 	for_each_engine(engine, gt, id) {
-		tasklet_disable(&engine->execlists.tasklet);
+		struct tasklet_struct *t = &engine->execlists.tasklet;
+
+		if (t->func)
+			tasklet_disable(t);
 		intel_engine_pm_get(engine);
 
 		for (p = igt_atomic_phases; p->name; p++) {
 			GEM_TRACE("intel_engine_reset(%s) under %s\n",
 				  engine->name, p->name);
+			if (strcmp(p->name, "softirq"))
+				local_bh_disable();
 
 			p->critical_section_begin();
-			err = intel_engine_reset(engine, NULL);
+			err = __intel_engine_reset_bh(engine, NULL);
 			p->critical_section_end();
 
+			if (strcmp(p->name, "softirq"))
+				local_bh_enable();
+
 			if (err) {
 				pr_err("intel_engine_reset(%s) failed under %s\n",
 				       engine->name, p->name);
@@ -339,7 +348,10 @@ static int igt_atomic_engine_reset(void *arg)
 		}
 
 		intel_engine_pm_put(engine);
-		tasklet_enable(&engine->execlists.tasklet);
+		if (t->func) {
+			tasklet_enable(t);
+			tasklet_hi_schedule(t);
+		}
 		if (err)
 			break;
 	}
diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.c b/drivers/gpu/drm/i915/gt/selftest_rps.c
index aa5675ecb5cc..967641fee42a 100644
--- a/drivers/gpu/drm/i915/gt/selftest_rps.c
+++ b/drivers/gpu/drm/i915/gt/selftest_rps.c
@@ -185,7 +185,10 @@ static u8 rps_set_check(struct intel_rps *rps, u8 freq)
 {
 	mutex_lock(&rps->lock);
 	GEM_BUG_ON(!intel_rps_is_active(rps));
-	intel_rps_set(rps, freq);
+	if (wait_for(!intel_rps_set(rps, freq), 50)) {
+		mutex_unlock(&rps->lock);
+		return 0;
+	}
 	GEM_BUG_ON(rps->last_freq != freq);
 	mutex_unlock(&rps->lock);
 
diff --git a/drivers/gpu/drm/i915/gt/selftest_timeline.c b/drivers/gpu/drm/i915/gt/selftest_timeline.c
index 2edf2b15885f..6f3a3687ef0f 100644
--- a/drivers/gpu/drm/i915/gt/selftest_timeline.c
+++ b/drivers/gpu/drm/i915/gt/selftest_timeline.c
@@ -9,6 +9,7 @@
 #include "intel_context.h"
 #include "intel_engine_heartbeat.h"
 #include "intel_engine_pm.h"
+#include "intel_gpu_commands.h"
 #include "intel_gt.h"
 #include "intel_gt_requests.h"
 #include "intel_ring.h"
@@ -1090,12 +1091,6 @@ static int live_hwsp_read(void *arg)
 			}
 			count++;
 
-			if (8 * watcher[1].rq->ring->emit >
-			    3 * watcher[1].rq->ring->size) {
-				i915_request_put(rq);
-				break;
-			}
-
 			/* Flush the timeline before manually wrapping again */
 			if (i915_request_wait(rq,
 					      I915_WAIT_INTERRUPTIBLE,
@@ -1104,9 +1099,14 @@ static int live_hwsp_read(void *arg)
 				i915_request_put(rq);
 				goto out;
 			}
-
 			retire_requests(tl);
 			i915_request_put(rq);
+
+			/* Single requests are limited to half a ring at most */
+			if (8 * watcher[1].rq->ring->emit >
+			    3 * watcher[1].rq->ring->size)
+				break;
+
 		} while (!__igt_timeout(end_time, NULL));
 		WRITE_ONCE(*(u32 *)tl->hwsp_seqno, 0xdeadbeef);
 
diff --git a/drivers/gpu/drm/i915/gt/selftest_workarounds.c b/drivers/gpu/drm/i915/gt/selftest_workarounds.c
index 61a0532d0f3d..2070b91cb607 100644
--- a/drivers/gpu/drm/i915/gt/selftest_workarounds.c
+++ b/drivers/gpu/drm/i915/gt/selftest_workarounds.c
@@ -95,8 +95,9 @@ reference_lists_fini(struct intel_gt *gt, struct wa_lists *lists)
 }
 
 static struct drm_i915_gem_object *
-read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine)
+read_nonprivs(struct intel_context *ce)
 {
+	struct intel_engine_cs *engine = ce->engine;
 	const u32 base = engine->mmio_base;
 	struct drm_i915_gem_object *result;
 	struct i915_request *rq;
@@ -130,7 +131,7 @@ read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine)
 	if (err)
 		goto err_obj;
 
-	rq = igt_request_alloc(ctx, engine);
+	rq = intel_context_create_request(ce);
 	if (IS_ERR(rq)) {
 		err = PTR_ERR(rq);
 		goto err_pin;
@@ -145,7 +146,7 @@ read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine)
 		goto err_req;
 
 	srm = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
-	if (INTEL_GEN(ctx->i915) >= 8)
+	if (INTEL_GEN(engine->i915) >= 8)
 		srm++;
 
 	cs = intel_ring_begin(rq, 4 * RING_MAX_NONPRIV_SLOTS);
@@ -200,16 +201,16 @@ print_results(const struct intel_engine_cs *engine, const u32 *results)
 	}
 }
 
-static int check_whitelist(struct i915_gem_context *ctx,
-			   struct intel_engine_cs *engine)
+static int check_whitelist(struct intel_context *ce)
 {
+	struct intel_engine_cs *engine = ce->engine;
 	struct drm_i915_gem_object *results;
 	struct intel_wedge_me wedge;
 	u32 *vaddr;
 	int err;
 	int i;
 
-	results = read_nonprivs(ctx, engine);
+	results = read_nonprivs(ce);
 	if (IS_ERR(results))
 		return PTR_ERR(results);
 
@@ -293,8 +294,7 @@ static int check_whitelist_across_reset(struct intel_engine_cs *engine,
 					int (*reset)(struct intel_engine_cs *),
 					const char *name)
 {
-	struct drm_i915_private *i915 = engine->i915;
-	struct i915_gem_context *ctx, *tmp;
+	struct intel_context *ce, *tmp;
 	struct igt_spinner spin;
 	intel_wakeref_t wakeref;
 	int err;
@@ -302,15 +302,15 @@ static int check_whitelist_across_reset(struct intel_engine_cs *engine,
 	pr_info("Checking %d whitelisted registers on %s (RING_NONPRIV) [%s]\n",
 		engine->whitelist.count, engine->name, name);
 
-	ctx = kernel_context(i915);
-	if (IS_ERR(ctx))
-		return PTR_ERR(ctx);
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
 
 	err = igt_spinner_init(&spin, engine->gt);
 	if (err)
 		goto out_ctx;
 
-	err = check_whitelist(ctx, engine);
+	err = check_whitelist(ce);
 	if (err) {
 		pr_err("Invalid whitelist *before* %s reset!\n", name);
 		goto out_spin;
@@ -330,22 +330,22 @@ static int check_whitelist_across_reset(struct intel_engine_cs *engine,
 		goto out_spin;
 	}
 
-	err = check_whitelist(ctx, engine);
+	err = check_whitelist(ce);
 	if (err) {
 		pr_err("Whitelist not preserved in context across %s reset!\n",
 		       name);
 		goto out_spin;
 	}
 
-	tmp = kernel_context(i915);
+	tmp = intel_context_create(engine);
 	if (IS_ERR(tmp)) {
 		err = PTR_ERR(tmp);
 		goto out_spin;
 	}
-	kernel_context_close(ctx);
-	ctx = tmp;
+	intel_context_put(ce);
+	ce = tmp;
 
-	err = check_whitelist(ctx, engine);
+	err = check_whitelist(ce);
 	if (err) {
 		pr_err("Invalid whitelist *after* %s reset in fresh context!\n",
 		       name);
@@ -355,7 +355,7 @@ static int check_whitelist_across_reset(struct intel_engine_cs *engine,
 out_spin:
 	igt_spinner_fini(&spin);
 out_ctx:
-	kernel_context_close(ctx);
+	intel_context_put(ce);
 	return err;
 }
 
@@ -486,10 +486,11 @@ static int check_dirty_whitelist(struct intel_context *ce)
 	struct intel_engine_cs *engine = ce->engine;
 	struct i915_vma *scratch;
 	struct i915_vma *batch;
-	int err = 0, i, v;
+	int err = 0, i, v, sz;
 	u32 *cs, *results;
 
-	scratch = create_scratch(ce->vm, 2 * ARRAY_SIZE(values) + 1);
+	sz = (2 * ARRAY_SIZE(values) + 1) * sizeof(u32);
+	scratch = __vm_create_scratch_for_read(ce->vm, sz);
 	if (IS_ERR(scratch))
 		return PTR_ERR(scratch);
 
@@ -786,15 +787,15 @@ out:
 	return err;
 }
 
-static int read_whitelisted_registers(struct i915_gem_context *ctx,
-				      struct intel_engine_cs *engine,
+static int read_whitelisted_registers(struct intel_context *ce,
 				      struct i915_vma *results)
 {
+	struct intel_engine_cs *engine = ce->engine;
 	struct i915_request *rq;
 	int i, err = 0;
 	u32 srm, *cs;
 
-	rq = igt_request_alloc(ctx, engine);
+	rq = intel_context_create_request(ce);
 	if (IS_ERR(rq))
 		return PTR_ERR(rq);
 
@@ -807,7 +808,7 @@ static int read_whitelisted_registers(struct i915_gem_context *ctx,
 		goto err_req;
 
 	srm = MI_STORE_REGISTER_MEM;
-	if (INTEL_GEN(ctx->i915) >= 8)
+	if (INTEL_GEN(engine->i915) >= 8)
 		srm++;
 
 	cs = intel_ring_begin(rq, 4 * engine->whitelist.count);
@@ -834,18 +835,15 @@ err_req:
 	return request_add_sync(rq, err);
 }
 
-static int scrub_whitelisted_registers(struct i915_gem_context *ctx,
-				       struct intel_engine_cs *engine)
+static int scrub_whitelisted_registers(struct intel_context *ce)
 {
-	struct i915_address_space *vm;
+	struct intel_engine_cs *engine = ce->engine;
 	struct i915_request *rq;
 	struct i915_vma *batch;
 	int i, err = 0;
 	u32 *cs;
 
-	vm = i915_gem_context_get_vm_rcu(ctx);
-	batch = create_batch(vm);
-	i915_vm_put(vm);
+	batch = create_batch(ce->vm);
 	if (IS_ERR(batch))
 		return PTR_ERR(batch);
 
@@ -873,7 +871,7 @@ static int scrub_whitelisted_registers(struct i915_gem_context *ctx,
 	i915_gem_object_flush_map(batch->obj);
 	intel_gt_chipset_flush(engine->gt);
 
-	rq = igt_request_alloc(ctx, engine);
+	rq = intel_context_create_request(ce);
 	if (IS_ERR(rq)) {
 		err = PTR_ERR(rq);
 		goto err_unpin;
@@ -1016,7 +1014,6 @@ static int live_isolated_whitelist(void *arg)
 {
 	struct intel_gt *gt = arg;
 	struct {
-		struct i915_gem_context *ctx;
 		struct i915_vma *scratch[2];
 	} client[2] = {};
 	struct intel_engine_cs *engine;
@@ -1032,61 +1029,57 @@ static int live_isolated_whitelist(void *arg)
 		return 0;
 
 	for (i = 0; i < ARRAY_SIZE(client); i++) {
-		struct i915_address_space *vm;
-		struct i915_gem_context *c;
-
-		c = kernel_context(gt->i915);
-		if (IS_ERR(c)) {
-			err = PTR_ERR(c);
-			goto err;
-		}
-
-		vm = i915_gem_context_get_vm_rcu(c);
-
-		client[i].scratch[0] = create_scratch(vm, 1024);
+		client[i].scratch[0] =
+			__vm_create_scratch_for_read(gt->vm, 4096);
 		if (IS_ERR(client[i].scratch[0])) {
 			err = PTR_ERR(client[i].scratch[0]);
-			i915_vm_put(vm);
-			kernel_context_close(c);
 			goto err;
 		}
 
-		client[i].scratch[1] = create_scratch(vm, 1024);
+		client[i].scratch[1] =
+			__vm_create_scratch_for_read(gt->vm, 4096);
 		if (IS_ERR(client[i].scratch[1])) {
 			err = PTR_ERR(client[i].scratch[1]);
 			i915_vma_unpin_and_release(&client[i].scratch[0], 0);
-			i915_vm_put(vm);
-			kernel_context_close(c);
 			goto err;
 		}
-
-		client[i].ctx = c;
-		i915_vm_put(vm);
 	}
 
 	for_each_engine(engine, gt, id) {
+		struct intel_context *ce[2];
+
 		if (!engine->kernel_context->vm)
 			continue;
 
 		if (!whitelist_writable_count(engine))
 			continue;
 
+		ce[0] = intel_context_create(engine);
+		if (IS_ERR(ce[0])) {
+			err = PTR_ERR(ce[0]);
+			break;
+		}
+		ce[1] = intel_context_create(engine);
+		if (IS_ERR(ce[1])) {
+			err = PTR_ERR(ce[1]);
+			intel_context_put(ce[0]);
+			break;
+		}
+
 		/* Read default values */
-		err = read_whitelisted_registers(client[0].ctx, engine,
-						 client[0].scratch[0]);
+		err = read_whitelisted_registers(ce[0], client[0].scratch[0]);
 		if (err)
-			goto err;
+			goto err_ce;
 
 		/* Try to overwrite registers (should only affect ctx0) */
-		err = scrub_whitelisted_registers(client[0].ctx, engine);
+		err = scrub_whitelisted_registers(ce[0]);
 		if (err)
-			goto err;
+			goto err_ce;
 
 		/* Read values from ctx1, we expect these to be defaults */
-		err = read_whitelisted_registers(client[1].ctx, engine,
-						 client[1].scratch[0]);
+		err = read_whitelisted_registers(ce[1], client[1].scratch[0]);
 		if (err)
-			goto err;
+			goto err_ce;
 
 		/* Verify that both reads return the same default values */
 		err = check_whitelisted_registers(engine,
@@ -1094,31 +1087,29 @@ static int live_isolated_whitelist(void *arg)
 						  client[1].scratch[0],
 						  result_eq);
 		if (err)
-			goto err;
+			goto err_ce;
 
 		/* Read back the updated values in ctx0 */
-		err = read_whitelisted_registers(client[0].ctx, engine,
-						 client[0].scratch[1]);
+		err = read_whitelisted_registers(ce[0], client[0].scratch[1]);
 		if (err)
-			goto err;
+			goto err_ce;
 
 		/* User should be granted privilege to overwhite regs */
 		err = check_whitelisted_registers(engine,
 						  client[0].scratch[0],
 						  client[0].scratch[1],
 						  result_neq);
+err_ce:
+		intel_context_put(ce[1]);
+		intel_context_put(ce[0]);
 		if (err)
-			goto err;
+			break;
 	}
 
 err:
 	for (i = 0; i < ARRAY_SIZE(client); i++) {
-		if (!client[i].ctx)
-			break;
-
 		i915_vma_unpin_and_release(&client[i].scratch[1], 0);
 		i915_vma_unpin_and_release(&client[i].scratch[0], 0);
-		kernel_context_close(client[i].ctx);
 	}
 
 	if (igt_flush_test(gt->i915))
@@ -1128,18 +1119,21 @@ err:
 }
 
 static bool
-verify_wa_lists(struct i915_gem_context *ctx, struct wa_lists *lists,
+verify_wa_lists(struct intel_gt *gt, struct wa_lists *lists,
 		const char *str)
 {
-	struct drm_i915_private *i915 = ctx->i915;
-	struct i915_gem_engines_iter it;
-	struct intel_context *ce;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
 	bool ok = true;
 
-	ok &= wa_list_verify(&i915->uncore, &lists->gt_wa_list, str);
+	ok &= wa_list_verify(gt->uncore, &lists->gt_wa_list, str);
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
 
-	for_each_gem_engine(ce, i915_gem_context_engines(ctx), it) {
-		enum intel_engine_id id = ce->engine->id;
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce))
+			return false;
 
 		ok &= engine_wa_list_verify(ce,
 					    &lists->engine[id].wa_list,
@@ -1148,6 +1142,8 @@ verify_wa_lists(struct i915_gem_context *ctx, struct wa_lists *lists,
 		ok &= engine_wa_list_verify(ce,
 					    &lists->engine[id].ctx_wa_list,
 					    str) == 0;
+
+		intel_context_put(ce);
 	}
 
 	return ok;
@@ -1157,7 +1153,6 @@ static int
 live_gpu_reset_workarounds(void *arg)
 {
 	struct intel_gt *gt = arg;
-	struct i915_gem_context *ctx;
 	intel_wakeref_t wakeref;
 	struct wa_lists lists;
 	bool ok;
@@ -1165,12 +1160,6 @@ live_gpu_reset_workarounds(void *arg)
 	if (!intel_has_gpu_reset(gt))
 		return 0;
 
-	ctx = kernel_context(gt->i915);
-	if (IS_ERR(ctx))
-		return PTR_ERR(ctx);
-
-	i915_gem_context_lock_engines(ctx);
-
 	pr_info("Verifying after GPU reset...\n");
 
 	igt_global_reset_lock(gt);
@@ -1178,17 +1167,15 @@ live_gpu_reset_workarounds(void *arg)
 
 	reference_lists_init(gt, &lists);
 
-	ok = verify_wa_lists(ctx, &lists, "before reset");
+	ok = verify_wa_lists(gt, &lists, "before reset");
 	if (!ok)
 		goto out;
 
 	intel_gt_reset(gt, ALL_ENGINES, "live_workarounds");
 
-	ok = verify_wa_lists(ctx, &lists, "after reset");
+	ok = verify_wa_lists(gt, &lists, "after reset");
 
 out:
-	i915_gem_context_unlock_engines(ctx);
-	kernel_context_close(ctx);
 	reference_lists_fini(gt, &lists);
 	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
 	igt_global_reset_unlock(gt);
@@ -1200,8 +1187,8 @@ static int
 live_engine_reset_workarounds(void *arg)
 {
 	struct intel_gt *gt = arg;
-	struct i915_gem_engines_iter it;
-	struct i915_gem_context *ctx;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
 	struct intel_context *ce;
 	struct igt_spinner spin;
 	struct i915_request *rq;
@@ -1212,30 +1199,30 @@ live_engine_reset_workarounds(void *arg)
 	if (!intel_has_reset_engine(gt))
 		return 0;
 
-	ctx = kernel_context(gt->i915);
-	if (IS_ERR(ctx))
-		return PTR_ERR(ctx);
-
 	igt_global_reset_lock(gt);
 	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
 
 	reference_lists_init(gt, &lists);
 
-	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
-		struct intel_engine_cs *engine = ce->engine;
+	for_each_engine(engine, gt, id) {
 		bool ok;
 
 		pr_info("Verifying after %s reset...\n", engine->name);
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			ret = PTR_ERR(ce);
+			break;
+		}
 
-		ok = verify_wa_lists(ctx, &lists, "before reset");
+		ok = verify_wa_lists(gt, &lists, "before reset");
 		if (!ok) {
 			ret = -ESRCH;
 			goto err;
 		}
 
-		intel_engine_reset(engine, "live_workarounds");
+		intel_engine_reset(engine, "live_workarounds:idle");
 
-		ok = verify_wa_lists(ctx, &lists, "after idle reset");
+		ok = verify_wa_lists(gt, &lists, "after idle reset");
 		if (!ok) {
 			ret = -ESRCH;
 			goto err;
@@ -1259,23 +1246,26 @@ live_engine_reset_workarounds(void *arg)
 			goto err;
 		}
 
-		intel_engine_reset(engine, "live_workarounds");
+		intel_engine_reset(engine, "live_workarounds:active");
 
 		igt_spinner_end(&spin);
 		igt_spinner_fini(&spin);
 
-		ok = verify_wa_lists(ctx, &lists, "after busy reset");
+		ok = verify_wa_lists(gt, &lists, "after busy reset");
 		if (!ok) {
 			ret = -ESRCH;
 			goto err;
 		}
-	}
+
 err:
-	i915_gem_context_unlock_engines(ctx);
+		intel_context_put(ce);
+		if (ret)
+			break;
+	}
+
 	reference_lists_fini(gt, &lists);
 	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
 	igt_global_reset_unlock(gt);
-	kernel_context_close(ctx);
 
 	igt_flush_test(gt->i915);
 
diff --git a/drivers/gpu/drm/i915/gt/shmem_utils.c b/drivers/gpu/drm/i915/gt/shmem_utils.c
index 5982b62f913d..a4d8fc9e2374 100644
--- a/drivers/gpu/drm/i915/gt/shmem_utils.c
+++ b/drivers/gpu/drm/i915/gt/shmem_utils.c
@@ -33,7 +33,7 @@ struct file *shmem_create_from_object(struct drm_i915_gem_object *obj)
 	struct file *file;
 	void *ptr;
 
-	if (obj->ops == &i915_gem_shmem_ops) {
+	if (i915_gem_object_is_shmem(obj)) {
 		file = obj->base.filp;
 		atomic_long_inc(&file->f_count);
 		return file;
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.c b/drivers/gpu/drm/i915/gt/uc/intel_guc.c
index 2a343a977987..4545e90e3bf1 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.c
@@ -579,20 +579,8 @@ int intel_guc_reset_engine(struct intel_guc *guc,
  */
 int intel_guc_resume(struct intel_guc *guc)
 {
-	u32 action[] = {
-		INTEL_GUC_ACTION_EXIT_S_STATE,
-		GUC_POWER_D0,
-	};
-
-	/*
-	 * If GuC communication is enabled but submission is not supported,
-	 * we do not need to resume the GuC but we do need to enable the
-	 * GuC communication on resume (above).
-	 */
-	if (!intel_guc_submission_is_used(guc) || !intel_guc_is_ready(guc))
-		return 0;
-
-	return intel_guc_send(guc, action, ARRAY_SIZE(action));
+	/* XXX: to be implemented with submission interface rework */
+	return 0;
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
index e84ab67b317d..bc2ba7d0626c 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc.h
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
@@ -47,13 +47,6 @@ struct intel_guc {
 	struct i915_vma *stage_desc_pool;
 	void *stage_desc_pool_vaddr;
 
-	struct i915_vma *workqueue;
-	void *workqueue_vaddr;
-	spinlock_t wq_lock;
-
-	struct i915_vma *proc_desc;
-	void *proc_desc_vaddr;
-
 	/* Control params for fw initialization */
 	u32 params[GUC_CTL_MAX_DWORDS];
 
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c
index 5212ff844292..17526717368c 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c
@@ -4,6 +4,7 @@
  */
 
 #include "gt/intel_gt.h"
+#include "gt/intel_lrc.h"
 #include "intel_guc_ads.h"
 #include "intel_uc.h"
 #include "i915_drv.h"
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c
index f9d0907ea1a5..2270d6b3b272 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c
@@ -76,7 +76,6 @@ static inline bool guc_ready(struct intel_uncore *uncore, u32 *status)
 
 static int guc_wait_ucode(struct intel_uncore *uncore)
 {
-	struct drm_device *drm = &uncore->i915->drm;
 	u32 status;
 	int ret;
 
@@ -89,11 +88,11 @@ static int guc_wait_ucode(struct intel_uncore *uncore)
 	 * attempt the ucode load again if this happens.)
 	 */
 	ret = wait_for(guc_ready(uncore, &status), 100);
-	DRM_DEBUG_DRIVER("GuC status %#x\n", status);
-
 	if (ret) {
-		drm_err(drm, "GuC load failed: status = 0x%08X\n", status);
-		drm_err(drm, "GuC load failed: status: Reset = %d, "
+		struct drm_device *drm = &uncore->i915->drm;
+
+		drm_dbg(drm, "GuC load failed: status = 0x%08X\n", status);
+		drm_dbg(drm, "GuC load failed: status: Reset = %d, "
 			"BootROM = 0x%02X, UKernel = 0x%02X, "
 			"MIA = 0x%02X, Auth = 0x%02X\n",
 			REG_FIELD_GET(GS_MIA_IN_RESET, status),
@@ -103,12 +102,12 @@ static int guc_wait_ucode(struct intel_uncore *uncore)
 			REG_FIELD_GET(GS_AUTH_STATUS_MASK, status));
 
 		if ((status & GS_BOOTROM_MASK) == GS_BOOTROM_RSA_FAILED) {
-			drm_err(drm, "GuC firmware signature verification failed\n");
+			drm_dbg(drm, "GuC firmware signature verification failed\n");
 			ret = -ENOEXEC;
 		}
 
 		if ((status & GS_UKERNEL_MASK) == GS_UKERNEL_EXCEPTION) {
-			drm_err(drm, "GuC firmware exception. EIP: %#x\n",
+			drm_dbg(drm, "GuC firmware exception. EIP: %#x\n",
 				intel_uncore_read(uncore, SOFT_SCRATCH(13)));
 			ret = -ENXIO;
 		}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
index fdfeb4b9b0f5..23dc0aeaa0ab 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -6,11 +6,14 @@
 #include <linux/circ_buf.h>
 
 #include "gem/i915_gem_context.h"
+#include "gt/gen8_engine_cs.h"
+#include "gt/intel_breadcrumbs.h"
 #include "gt/intel_context.h"
 #include "gt/intel_engine_pm.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_pm.h"
-#include "gt/intel_lrc_reg.h"
+#include "gt/intel_lrc.h"
+#include "gt/intel_mocs.h"
 #include "gt/intel_ring.h"
 
 #include "intel_guc_submission.h"
@@ -54,6 +57,8 @@
  *
  */
 
+#define GUC_REQUEST_SIZE 64 /* bytes */
+
 static inline struct i915_priolist *to_priolist(struct rb_node *rb)
 {
 	return rb_entry(rb, struct i915_priolist, node);
@@ -66,58 +71,6 @@ static struct guc_stage_desc *__get_stage_desc(struct intel_guc *guc, u32 id)
 	return &base[id];
 }
 
-static int guc_workqueue_create(struct intel_guc *guc)
-{
-	return intel_guc_allocate_and_map_vma(guc, GUC_WQ_SIZE, &guc->workqueue,
-					      &guc->workqueue_vaddr);
-}
-
-static void guc_workqueue_destroy(struct intel_guc *guc)
-{
-	i915_vma_unpin_and_release(&guc->workqueue, I915_VMA_RELEASE_MAP);
-}
-
-/*
- * Initialise the process descriptor shared with the GuC firmware.
- */
-static int guc_proc_desc_create(struct intel_guc *guc)
-{
-	const u32 size = PAGE_ALIGN(sizeof(struct guc_process_desc));
-
-	return intel_guc_allocate_and_map_vma(guc, size, &guc->proc_desc,
-					      &guc->proc_desc_vaddr);
-}
-
-static void guc_proc_desc_destroy(struct intel_guc *guc)
-{
-	i915_vma_unpin_and_release(&guc->proc_desc, I915_VMA_RELEASE_MAP);
-}
-
-static void guc_proc_desc_init(struct intel_guc *guc)
-{
-	struct guc_process_desc *desc;
-
-	desc = memset(guc->proc_desc_vaddr, 0, sizeof(*desc));
-
-	/*
-	 * XXX: pDoorbell and WQVBaseAddress are pointers in process address
-	 * space for ring3 clients (set them as in mmap_ioctl) or kernel
-	 * space for kernel clients (map on demand instead? May make debug
-	 * easier to have it mapped).
-	 */
-	desc->wq_base_addr = 0;
-	desc->db_base_addr = 0;
-
-	desc->wq_size_bytes = GUC_WQ_SIZE;
-	desc->wq_status = WQ_STATUS_ACTIVE;
-	desc->priority = GUC_CLIENT_PRIORITY_KMD_NORMAL;
-}
-
-static void guc_proc_desc_fini(struct intel_guc *guc)
-{
-	memset(guc->proc_desc_vaddr, 0, sizeof(struct guc_process_desc));
-}
-
 static int guc_stage_desc_pool_create(struct intel_guc *guc)
 {
 	u32 size = PAGE_ALIGN(sizeof(struct guc_stage_desc) *
@@ -153,8 +106,6 @@ static void guc_stage_desc_init(struct intel_guc *guc)
 	desc->stage_id = 0;
 	desc->priority = GUC_CLIENT_PRIORITY_KMD_NORMAL;
 
-	desc->process_desc = intel_guc_ggtt_offset(guc, guc->proc_desc);
-	desc->wq_addr = intel_guc_ggtt_offset(guc, guc->workqueue);
 	desc->wq_size = GUC_WQ_SIZE;
 }
 
@@ -166,62 +117,9 @@ static void guc_stage_desc_fini(struct intel_guc *guc)
 	memset(desc, 0, sizeof(*desc));
 }
 
-/* Construct a Work Item and append it to the GuC's Work Queue */
-static void guc_wq_item_append(struct intel_guc *guc,
-			       u32 target_engine, u32 context_desc,
-			       u32 ring_tail, u32 fence_id)
-{
-	/* wqi_len is in DWords, and does not include the one-word header */
-	const size_t wqi_size = sizeof(struct guc_wq_item);
-	const u32 wqi_len = wqi_size / sizeof(u32) - 1;
-	struct guc_process_desc *desc = guc->proc_desc_vaddr;
-	struct guc_wq_item *wqi;
-	u32 wq_off;
-
-	lockdep_assert_held(&guc->wq_lock);
-
-	/* For now workqueue item is 4 DWs; workqueue buffer is 2 pages. So we
-	 * should not have the case where structure wqi is across page, neither
-	 * wrapped to the beginning. This simplifies the implementation below.
-	 *
-	 * XXX: if not the case, we need save data to a temp wqi and copy it to
-	 * workqueue buffer dw by dw.
-	 */
-	BUILD_BUG_ON(wqi_size != 16);
-
-	/* We expect the WQ to be active if we're appending items to it */
-	GEM_BUG_ON(desc->wq_status != WQ_STATUS_ACTIVE);
-
-	/* Free space is guaranteed. */
-	wq_off = READ_ONCE(desc->tail);
-	GEM_BUG_ON(CIRC_SPACE(wq_off, READ_ONCE(desc->head),
-			      GUC_WQ_SIZE) < wqi_size);
-	GEM_BUG_ON(wq_off & (wqi_size - 1));
-
-	wqi = guc->workqueue_vaddr + wq_off;
-
-	/* Now fill in the 4-word work queue item */
-	wqi->header = WQ_TYPE_INORDER |
-		      (wqi_len << WQ_LEN_SHIFT) |
-		      (target_engine << WQ_TARGET_SHIFT) |
-		      WQ_NO_WCFLUSH_WAIT;
-	wqi->context_desc = context_desc;
-	wqi->submit_element_info = ring_tail << WQ_RING_TAIL_SHIFT;
-	GEM_BUG_ON(ring_tail > WQ_RING_TAIL_MAX);
-	wqi->fence_id = fence_id;
-
-	/* Make the update visible to GuC */
-	WRITE_ONCE(desc->tail, (wq_off + wqi_size) & (GUC_WQ_SIZE - 1));
-}
-
 static void guc_add_request(struct intel_guc *guc, struct i915_request *rq)
 {
-	struct intel_engine_cs *engine = rq->engine;
-	u32 ctx_desc = rq->context->lrc.ccid;
-	u32 ring_tail = intel_ring_set_tail(rq->ring, rq->tail) / sizeof(u64);
-
-	guc_wq_item_append(guc, engine->guc_id, ctx_desc,
-			   ring_tail, rq->fence.seqno);
+	/* Leaving stub as this function will be used in future patches */
 }
 
 /*
@@ -244,16 +142,12 @@ static void guc_submit(struct intel_engine_cs *engine,
 {
 	struct intel_guc *guc = &engine->gt->uc.guc;
 
-	spin_lock(&guc->wq_lock);
-
 	do {
 		struct i915_request *rq = *out++;
 
 		flush_ggtt_writes(rq->ring->vma);
 		guc_add_request(guc, rq);
 	} while (out != end);
-
-	spin_unlock(&guc->wq_lock);
 }
 
 static inline int rq_prio(const struct i915_request *rq)
@@ -388,17 +282,26 @@ static void guc_reset_prepare(struct intel_engine_cs *engine)
 	__tasklet_disable_sync_once(&execlists->tasklet);
 }
 
-static void
-cancel_port_requests(struct intel_engine_execlists * const execlists)
+static void guc_reset_state(struct intel_context *ce,
+			    struct intel_engine_cs *engine,
+			    u32 head,
+			    bool scrub)
 {
-	struct i915_request * const *port, *rq;
+	GEM_BUG_ON(!intel_context_is_pinned(ce));
 
-	/* Note we are only using the inflight and not the pending queue */
+	/*
+	 * We want a simple context + ring to execute the breadcrumb update.
+	 * We cannot rely on the context being intact across the GPU hang,
+	 * so clear it and rebuild just what we need for the breadcrumb.
+	 * All pending requests for this context will be zapped, and any
+	 * future request will be after userspace has had the opportunity
+	 * to recreate its own state.
+	 */
+	if (scrub)
+		lrc_init_regs(ce, engine, true);
 
-	for (port = execlists->active; (rq = *port); port++)
-		schedule_out(rq);
-	execlists->active =
-		memset(execlists->inflight, 0, sizeof(execlists->inflight));
+	/* Rerun the request; its payload has been neutered (if guilty). */
+	lrc_update_regs(ce, engine, head);
 }
 
 static void guc_reset_rewind(struct intel_engine_cs *engine, bool stalled)
@@ -409,8 +312,6 @@ static void guc_reset_rewind(struct intel_engine_cs *engine, bool stalled)
 
 	spin_lock_irqsave(&engine->active.lock, flags);
 
-	cancel_port_requests(execlists);
-
 	/* Push back any incomplete requests for replay after the reset. */
 	rq = execlists_unwind_incomplete_requests(execlists);
 	if (!rq)
@@ -420,7 +321,7 @@ static void guc_reset_rewind(struct intel_engine_cs *engine, bool stalled)
 		stalled = false;
 
 	__i915_request_reset(rq, stalled);
-	intel_lr_context_reset(engine, rq->context, rq->head, stalled);
+	guc_reset_state(rq->context, engine, rq->head, stalled);
 
 out_unlock:
 	spin_unlock_irqrestore(&engine->active.lock, flags);
@@ -451,9 +352,6 @@ static void guc_reset_cancel(struct intel_engine_cs *engine)
 	 */
 	spin_lock_irqsave(&engine->active.lock, flags);
 
-	/* Cancel the requests on the HW and clear the ELSP tracker. */
-	cancel_port_requests(execlists);
-
 	/* Mark all executing requests as skipped. */
 	list_for_each_entry(rq, &engine->active.requests, sched.link) {
 		i915_request_set_error_once(rq, -EIO);
@@ -497,12 +395,6 @@ static void guc_reset_finish(struct intel_engine_cs *engine)
 }
 
 /*
- * Everything below here is concerned with setup & teardown, and is
- * therefore not part of the somewhat time-critical batch-submission
- * path of guc_submit() above.
- */
-
-/*
  * Set up the memory resources to be shared with the GuC (via the GGTT)
  * at firmware loading time.
  */
@@ -522,30 +414,12 @@ int intel_guc_submission_init(struct intel_guc *guc)
 	 */
 	GEM_BUG_ON(!guc->stage_desc_pool);
 
-	ret = guc_workqueue_create(guc);
-	if (ret)
-		goto err_pool;
-
-	ret = guc_proc_desc_create(guc);
-	if (ret)
-		goto err_workqueue;
-
-	spin_lock_init(&guc->wq_lock);
-
 	return 0;
-
-err_workqueue:
-	guc_workqueue_destroy(guc);
-err_pool:
-	guc_stage_desc_pool_destroy(guc);
-	return ret;
 }
 
 void intel_guc_submission_fini(struct intel_guc *guc)
 {
 	if (guc->stage_desc_pool) {
-		guc_proc_desc_destroy(guc);
-		guc_workqueue_destroy(guc);
 		guc_stage_desc_pool_destroy(guc);
 	}
 }
@@ -576,33 +450,186 @@ static void guc_interrupts_release(struct intel_gt *gt)
 	intel_uncore_rmw(uncore, GEN11_VCS_VECS_INTR_ENABLE, 0, dmask);
 }
 
-static void guc_set_default_submission(struct intel_engine_cs *engine)
+static int guc_context_alloc(struct intel_context *ce)
+{
+	return lrc_alloc(ce, ce->engine);
+}
+
+static int guc_context_pre_pin(struct intel_context *ce,
+			       struct i915_gem_ww_ctx *ww,
+			       void **vaddr)
+{
+	return lrc_pre_pin(ce, ce->engine, ww, vaddr);
+}
+
+static int guc_context_pin(struct intel_context *ce, void *vaddr)
 {
+	return lrc_pin(ce, ce->engine, vaddr);
+}
+
+static const struct intel_context_ops guc_context_ops = {
+	.alloc = guc_context_alloc,
+
+	.pre_pin = guc_context_pre_pin,
+	.pin = guc_context_pin,
+	.unpin = lrc_unpin,
+	.post_unpin = lrc_post_unpin,
+
+	.enter = intel_context_enter_engine,
+	.exit = intel_context_exit_engine,
+
+	.reset = lrc_reset,
+	.destroy = lrc_destroy,
+};
+
+static int guc_request_alloc(struct i915_request *request)
+{
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_pinned(request->context));
+
 	/*
-	 * We inherit a bunch of functions from execlists that we'd like
-	 * to keep using:
-	 *
-	 *    engine->submit_request = execlists_submit_request;
-	 *    engine->cancel_requests = execlists_cancel_requests;
-	 *    engine->schedule = execlists_schedule;
+	 * Flush enough space to reduce the likelihood of waiting after
+	 * we start building the request - in which case we will just
+	 * have to repeat work.
+	 */
+	request->reserved_space += GUC_REQUEST_SIZE;
+
+	/*
+	 * Note that after this point, we have committed to using
+	 * this request as it is being used to both track the
+	 * state of engine initialisation and liveness of the
+	 * golden renderstate above. Think twice before you try
+	 * to cancel/unwind this request now.
+	 */
+
+	/* Unconditionally invalidate GPU caches and TLBs. */
+	ret = request->engine->emit_flush(request, EMIT_INVALIDATE);
+	if (ret)
+		return ret;
+
+	request->reserved_space -= GUC_REQUEST_SIZE;
+	return 0;
+}
+
+static inline void queue_request(struct intel_engine_cs *engine,
+				 struct i915_request *rq,
+				 int prio)
+{
+	GEM_BUG_ON(!list_empty(&rq->sched.link));
+	list_add_tail(&rq->sched.link,
+		      i915_sched_lookup_priolist(engine, prio));
+	set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+}
+
+static void guc_submit_request(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&engine->active.lock, flags);
+
+	queue_request(engine, rq, rq_prio(rq));
+
+	GEM_BUG_ON(RB_EMPTY_ROOT(&engine->execlists.queue.rb_root));
+	GEM_BUG_ON(list_empty(&rq->sched.link));
+
+	tasklet_hi_schedule(&engine->execlists.tasklet);
+
+	spin_unlock_irqrestore(&engine->active.lock, flags);
+}
+
+static void sanitize_hwsp(struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl;
+
+	list_for_each_entry(tl, &engine->status_page.timelines, engine_link)
+		intel_timeline_reset_seqno(tl);
+}
+
+static void guc_sanitize(struct intel_engine_cs *engine)
+{
+	/*
+	 * Poison residual state on resume, in case the suspend didn't!
 	 *
-	 * But we need to override the actual submission backend in order
-	 * to talk to the GuC.
+	 * We have to assume that across suspend/resume (or other loss
+	 * of control) that the contents of our pinned buffers has been
+	 * lost, replaced by garbage. Since this doesn't always happen,
+	 * let's poison such state so that we more quickly spot when
+	 * we falsely assume it has been preserved.
 	 */
-	intel_execlists_set_default_submission(engine);
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
 
-	engine->execlists.tasklet.func = guc_submission_tasklet;
+	/*
+	 * The kernel_context HWSP is stored in the status_page. As above,
+	 * that may be lost on resume/initialisation, and so we need to
+	 * reset the value in the HWSP.
+	 */
+	sanitize_hwsp(engine);
 
-	/* do not use execlists park/unpark */
-	engine->park = engine->unpark = NULL;
+	/* And scrub the dirty cachelines for the HWSP */
+	clflush_cache_range(engine->status_page.addr, PAGE_SIZE);
+}
+
+static void setup_hwsp(struct intel_engine_cs *engine)
+{
+	intel_engine_set_hwsp_writemask(engine, ~0u); /* HWSTAM */
+
+	ENGINE_WRITE_FW(engine,
+			RING_HWS_PGA,
+			i915_ggtt_offset(engine->status_page.vma));
+}
+
+static void start_engine(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE_FW(engine,
+			RING_MODE_GEN7,
+			_MASKED_BIT_ENABLE(GEN11_GFX_DISABLE_LEGACY_MODE));
+
+	ENGINE_WRITE_FW(engine, RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+	ENGINE_POSTING_READ(engine, RING_MI_MODE);
+}
+
+static int guc_resume(struct intel_engine_cs *engine)
+{
+	assert_forcewakes_active(engine->uncore, FORCEWAKE_ALL);
+
+	intel_mocs_init_engine(engine);
+
+	intel_breadcrumbs_reset(engine->breadcrumbs);
+
+	setup_hwsp(engine);
+	start_engine(engine);
+
+	return 0;
+}
+
+static void guc_set_default_submission(struct intel_engine_cs *engine)
+{
+	engine->submit_request = guc_submit_request;
+	engine->schedule = i915_schedule;
+	engine->execlists.tasklet.func = guc_submission_tasklet;
 
 	engine->reset.prepare = guc_reset_prepare;
 	engine->reset.rewind = guc_reset_rewind;
 	engine->reset.cancel = guc_reset_cancel;
 	engine->reset.finish = guc_reset_finish;
 
-	engine->flags &= ~I915_ENGINE_SUPPORTS_STATS;
 	engine->flags |= I915_ENGINE_NEEDS_BREADCRUMB_TASKLET;
+	engine->flags |= I915_ENGINE_HAS_PREEMPTION;
+
+	/*
+	 * TODO: GuC supports timeslicing and semaphores as well, but they're
+	 * handled by the firmware so some minor tweaks are required before
+	 * enabling.
+	 *
+	 * engine->flags |= I915_ENGINE_HAS_TIMESLICES;
+	 * engine->flags |= I915_ENGINE_HAS_SEMAPHORES;
+	 */
+
+	engine->emit_bb_start = gen8_emit_bb_start;
 
 	/*
 	 * For the breadcrumb irq to work we need the interrupts to stay
@@ -613,35 +640,92 @@ static void guc_set_default_submission(struct intel_engine_cs *engine)
 	GEM_BUG_ON(engine->irq_enable || engine->irq_disable);
 }
 
-void intel_guc_submission_enable(struct intel_guc *guc)
+static void guc_release(struct intel_engine_cs *engine)
 {
-	struct intel_gt *gt = guc_to_gt(guc);
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
+	engine->sanitize = NULL; /* no longer in control, nothing to sanitize */
+
+	tasklet_kill(&engine->execlists.tasklet);
+
+	intel_engine_cleanup_common(engine);
+	lrc_fini_wa_ctx(engine);
+}
+
+static void guc_default_vfuncs(struct intel_engine_cs *engine)
+{
+	/* Default vfuncs which can be overridden by each engine. */
+
+	engine->resume = guc_resume;
+
+	engine->cops = &guc_context_ops;
+	engine->request_alloc = guc_request_alloc;
+
+	engine->emit_flush = gen8_emit_flush_xcs;
+	engine->emit_init_breadcrumb = gen8_emit_init_breadcrumb;
+	engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_xcs;
+	if (INTEL_GEN(engine->i915) >= 12) {
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_xcs;
+		engine->emit_flush = gen12_emit_flush_xcs;
+	}
+	engine->set_default_submission = guc_set_default_submission;
+}
+
+static void rcs_submission_override(struct intel_engine_cs *engine)
+{
+	switch (INTEL_GEN(engine->i915)) {
+	case 12:
+		engine->emit_flush = gen12_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_rcs;
+		break;
+	case 11:
+		engine->emit_flush = gen11_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen11_emit_fini_breadcrumb_rcs;
+		break;
+	default:
+		engine->emit_flush = gen8_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_rcs;
+		break;
+	}
+}
+
+static inline void guc_default_irqs(struct intel_engine_cs *engine)
+{
+	engine->irq_keep_mask = GT_RENDER_USER_INTERRUPT;
+}
+
+int intel_guc_submission_setup(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
 
 	/*
-	 * We're using GuC work items for submitting work through GuC. Since
-	 * we're coalescing multiple requests from a single context into a
-	 * single work item prior to assigning it to execlist_port, we can
-	 * never have more work items than the total number of ports (for all
-	 * engines). The GuC firmware is controlling the HEAD of work queue,
-	 * and it is guaranteed that it will remove the work item from the
-	 * queue before our request is completed.
+	 * The setup relies on several assumptions (e.g. irqs always enabled)
+	 * that are only valid on gen11+
 	 */
-	BUILD_BUG_ON(ARRAY_SIZE(engine->execlists.inflight) *
-		     sizeof(struct guc_wq_item) *
-		     I915_NUM_ENGINES > GUC_WQ_SIZE);
+	GEM_BUG_ON(INTEL_GEN(i915) < 11);
+
+	tasklet_init(&engine->execlists.tasklet,
+		     guc_submission_tasklet, (unsigned long)engine);
+
+	guc_default_vfuncs(engine);
+	guc_default_irqs(engine);
 
-	guc_proc_desc_init(guc);
+	if (engine->class == RENDER_CLASS)
+		rcs_submission_override(engine);
+
+	lrc_init_wa_ctx(engine);
+
+	/* Finally, take ownership and responsibility for cleanup! */
+	engine->sanitize = guc_sanitize;
+	engine->release = guc_release;
+
+	return 0;
+}
+
+void intel_guc_submission_enable(struct intel_guc *guc)
+{
 	guc_stage_desc_init(guc);
 
 	/* Take over from manual control of ELSP (execlists) */
-	guc_interrupts_capture(gt);
-
-	for_each_engine(engine, gt, id) {
-		engine->set_default_submission = guc_set_default_submission;
-		engine->set_default_submission(engine);
-	}
+	guc_interrupts_capture(guc_to_gt(guc));
 }
 
 void intel_guc_submission_disable(struct intel_guc *guc)
@@ -655,7 +739,6 @@ void intel_guc_submission_disable(struct intel_guc *guc)
 	guc_interrupts_release(gt);
 
 	guc_stage_desc_fini(guc);
-	guc_proc_desc_fini(guc);
 }
 
 static bool __guc_submission_selected(struct intel_guc *guc)
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h
index 4cf9d3e50263..5f7b9e6347d0 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h
@@ -19,6 +19,7 @@ void intel_guc_submission_disable(struct intel_guc *guc);
 void intel_guc_submission_fini(struct intel_guc *guc);
 int intel_guc_preempt_work_create(struct intel_guc *guc);
 void intel_guc_preempt_work_destroy(struct intel_guc *guc);
+int intel_guc_submission_setup(struct intel_engine_cs *engine);
 bool intel_engine_in_guc_submission_mode(const struct intel_engine_cs *engine);
 
 static inline bool intel_guc_submission_is_supported(struct intel_guc *guc)
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
index 4e6070e95fe9..6abb8f2dc33d 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_uc.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
@@ -15,6 +15,29 @@
 static const struct intel_uc_ops uc_ops_off;
 static const struct intel_uc_ops uc_ops_on;
 
+static void uc_expand_default_options(struct intel_uc *uc)
+{
+	struct drm_i915_private *i915 = uc_to_gt(uc)->i915;
+
+	if (i915->params.enable_guc != -1)
+		return;
+
+	/* Don't enable GuC/HuC on pre-Gen12 */
+	if (INTEL_GEN(i915) < 12) {
+		i915->params.enable_guc = 0;
+		return;
+	}
+
+	/* Don't enable GuC/HuC on older Gen12 platforms */
+	if (IS_TIGERLAKE(i915) || IS_ROCKETLAKE(i915)) {
+		i915->params.enable_guc = 0;
+		return;
+	}
+
+	/* Default: enable HuC authentication only */
+	i915->params.enable_guc = ENABLE_GUC_LOAD_HUC;
+}
+
 /* Reset GuC providing us with fresh state for both GuC and HuC.
  */
 static int __intel_uc_reset_hw(struct intel_uc *uc)
@@ -52,9 +75,6 @@ static void __confirm_options(struct intel_uc *uc)
 		yesno(intel_uc_wants_guc_submission(uc)),
 		yesno(intel_uc_wants_huc(uc)));
 
-	if (i915->params.enable_guc == -1)
-		return;
-
 	if (i915->params.enable_guc == 0) {
 		GEM_BUG_ON(intel_uc_wants_guc(uc));
 		GEM_BUG_ON(intel_uc_wants_guc_submission(uc));
@@ -79,8 +99,7 @@ static void __confirm_options(struct intel_uc *uc)
 			 "Incompatible option enable_guc=%d - %s\n",
 			 i915->params.enable_guc, "GuC submission is N/A");
 
-	if (i915->params.enable_guc & ~(ENABLE_GUC_SUBMISSION |
-					  ENABLE_GUC_LOAD_HUC))
+	if (i915->params.enable_guc & ~ENABLE_GUC_MASK)
 		drm_info(&i915->drm,
 			 "Incompatible option enable_guc=%d - %s\n",
 			 i915->params.enable_guc, "undocumented flag");
@@ -88,6 +107,8 @@ static void __confirm_options(struct intel_uc *uc)
 
 void intel_uc_init_early(struct intel_uc *uc)
 {
+	uc_expand_default_options(uc);
+
 	intel_guc_init_early(&uc->guc);
 	intel_huc_init_early(&uc->huc);
 
@@ -175,19 +196,15 @@ static void guc_get_mmio_msg(struct intel_guc *guc)
 
 static void guc_handle_mmio_msg(struct intel_guc *guc)
 {
-	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
-
 	/* we need communication to be enabled to reply to GuC */
 	GEM_BUG_ON(!guc_communication_enabled(guc));
 
-	if (!guc->mmio_msg)
-		return;
-
-	spin_lock_irq(&i915->irq_lock);
-	intel_guc_to_host_process_recv_msg(guc, &guc->mmio_msg, 1);
-	spin_unlock_irq(&i915->irq_lock);
-
-	guc->mmio_msg = 0;
+	spin_lock_irq(&guc->irq_lock);
+	if (guc->mmio_msg) {
+		intel_guc_to_host_process_recv_msg(guc, &guc->mmio_msg, 1);
+		guc->mmio_msg = 0;
+	}
+	spin_unlock_irq(&guc->irq_lock);
 }
 
 static void guc_reset_interrupts(struct intel_guc *guc)
@@ -207,7 +224,8 @@ static void guc_disable_interrupts(struct intel_guc *guc)
 
 static int guc_enable_communication(struct intel_guc *guc)
 {
-	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = gt->i915;
 	int ret;
 
 	GEM_BUG_ON(guc_communication_enabled(guc));
@@ -227,9 +245,9 @@ static int guc_enable_communication(struct intel_guc *guc)
 	guc_enable_interrupts(guc);
 
 	/* check for CT messages received before we enabled interrupts */
-	spin_lock_irq(&i915->irq_lock);
+	spin_lock_irq(&gt->irq_lock);
 	intel_guc_ct_event_handler(&guc->ct);
-	spin_unlock_irq(&i915->irq_lock);
+	spin_unlock_irq(&gt->irq_lock);
 
 	drm_dbg(&i915->drm, "GuC communication enabled\n");
 
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
index 180c23e2e25e..67b06fde1225 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
@@ -53,6 +53,7 @@ void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw,
 	fw_def(ELKHARTLAKE, 0, guc_def(ehl, 49, 0, 1), huc_def(ehl,  9, 0, 0)) \
 	fw_def(ICELAKE,     0, guc_def(icl, 49, 0, 1), huc_def(icl,  9, 0, 0)) \
 	fw_def(COMETLAKE,   5, guc_def(cml, 49, 0, 1), huc_def(cml,  4, 0, 0)) \
+	fw_def(COMETLAKE,   0, guc_def(kbl, 49, 0, 1), huc_def(kbl,  4, 0, 0)) \
 	fw_def(COFFEELAKE,  0, guc_def(kbl, 49, 0, 1), huc_def(kbl,  4, 0, 0)) \
 	fw_def(GEMINILAKE,  0, guc_def(glk, 49, 0, 1), huc_def(glk,  4, 0, 0)) \
 	fw_def(KABYLAKE,    0, guc_def(kbl, 49, 0, 1), huc_def(kbl,  4, 0, 0)) \
@@ -151,16 +152,11 @@ __uc_fw_auto_select(struct drm_i915_private *i915, struct intel_uc_fw *uc_fw)
 			uc_fw->path = NULL;
 		}
 	}
-
-	/* We don't want to enable GuC/HuC on pre-Gen11 by default */
-	if (i915->params.enable_guc == -1 && p < INTEL_ICELAKE)
-		uc_fw->path = NULL;
 }
 
 static const char *__override_guc_firmware_path(struct drm_i915_private *i915)
 {
-	if (i915->params.enable_guc & (ENABLE_GUC_SUBMISSION |
-				       ENABLE_GUC_LOAD_HUC))
+	if (i915->params.enable_guc & ENABLE_GUC_MASK)
 		return i915->params.guc_firmware_path;
 	return "";
 }
diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c
index 16b582cb97ed..fef1e857cefc 100644
--- a/drivers/gpu/drm/i915/gvt/cmd_parser.c
+++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c
@@ -37,11 +37,19 @@
 #include <linux/slab.h>
 
 #include "i915_drv.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_lrc.h"
 #include "gt/intel_ring.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/shmem_utils.h"
 #include "gvt.h"
 #include "i915_pvinfo.h"
 #include "trace.h"
 
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_context.h"
+
 #define INVALID_OP    (~0U)
 
 #define OP_LEN_MI           9
@@ -454,6 +462,7 @@ enum {
 	RING_BUFFER_INSTRUCTION,
 	BATCH_BUFFER_INSTRUCTION,
 	BATCH_BUFFER_2ND_LEVEL,
+	RING_BUFFER_CTX,
 };
 
 enum {
@@ -495,6 +504,7 @@ struct parser_exec_state {
 	 */
 	int saved_buf_addr_type;
 	bool is_ctx_wa;
+	bool is_init_ctx;
 
 	const struct cmd_info *info;
 
@@ -708,6 +718,11 @@ static inline u32 cmd_val(struct parser_exec_state *s, int index)
 	return *cmd_ptr(s, index);
 }
 
+static inline bool is_init_ctx(struct parser_exec_state *s)
+{
+	return (s->buf_type == RING_BUFFER_CTX && s->is_init_ctx);
+}
+
 static void parser_exec_state_dump(struct parser_exec_state *s)
 {
 	int cnt = 0;
@@ -721,7 +736,8 @@ static void parser_exec_state_dump(struct parser_exec_state *s)
 
 	gvt_dbg_cmd("  %s %s ip_gma(%08lx) ",
 			s->buf_type == RING_BUFFER_INSTRUCTION ?
-			"RING_BUFFER" : "BATCH_BUFFER",
+			"RING_BUFFER" : ((s->buf_type == RING_BUFFER_CTX) ?
+				"CTX_BUFFER" : "BATCH_BUFFER"),
 			s->buf_addr_type == GTT_BUFFER ?
 			"GTT" : "PPGTT", s->ip_gma);
 
@@ -756,7 +772,8 @@ static inline void update_ip_va(struct parser_exec_state *s)
 	if (WARN_ON(s->ring_head == s->ring_tail))
 		return;
 
-	if (s->buf_type == RING_BUFFER_INSTRUCTION) {
+	if (s->buf_type == RING_BUFFER_INSTRUCTION ||
+			s->buf_type == RING_BUFFER_CTX) {
 		unsigned long ring_top = s->ring_start + s->ring_size;
 
 		if (s->ring_head > s->ring_tail) {
@@ -820,68 +837,12 @@ static inline int cmd_length(struct parser_exec_state *s)
 	*addr = val; \
 } while (0)
 
-static bool is_shadowed_mmio(unsigned int offset)
-{
-	bool ret = false;
-
-	if ((offset == 0x2168) || /*BB current head register UDW */
-	    (offset == 0x2140) || /*BB current header register */
-	    (offset == 0x211c) || /*second BB header register UDW */
-	    (offset == 0x2114)) { /*second BB header register UDW */
-		ret = true;
-	}
-	return ret;
-}
-
-static inline bool is_force_nonpriv_mmio(unsigned int offset)
-{
-	return (offset >= 0x24d0 && offset < 0x2500);
-}
-
-static int force_nonpriv_reg_handler(struct parser_exec_state *s,
-		unsigned int offset, unsigned int index, char *cmd)
-{
-	struct intel_gvt *gvt = s->vgpu->gvt;
-	unsigned int data;
-	u32 ring_base;
-	u32 nopid;
-
-	if (!strcmp(cmd, "lri"))
-		data = cmd_val(s, index + 1);
-	else {
-		gvt_err("Unexpected forcenonpriv 0x%x write from cmd %s\n",
-			offset, cmd);
-		return -EINVAL;
-	}
-
-	ring_base = s->engine->mmio_base;
-	nopid = i915_mmio_reg_offset(RING_NOPID(ring_base));
-
-	if (!intel_gvt_in_force_nonpriv_whitelist(gvt, data) &&
-			data != nopid) {
-		gvt_err("Unexpected forcenonpriv 0x%x LRI write, value=0x%x\n",
-			offset, data);
-		patch_value(s, cmd_ptr(s, index), nopid);
-		return 0;
-	}
-	return 0;
-}
-
 static inline bool is_mocs_mmio(unsigned int offset)
 {
 	return ((offset >= 0xc800) && (offset <= 0xcff8)) ||
 		((offset >= 0xb020) && (offset <= 0xb0a0));
 }
 
-static int mocs_cmd_reg_handler(struct parser_exec_state *s,
-				unsigned int offset, unsigned int index)
-{
-	if (!is_mocs_mmio(offset))
-		return -EINVAL;
-	vgpu_vreg(s->vgpu, offset) = cmd_val(s, index + 1);
-	return 0;
-}
-
 static int is_cmd_update_pdps(unsigned int offset,
 			      struct parser_exec_state *s)
 {
@@ -929,6 +890,7 @@ static int cmd_reg_handler(struct parser_exec_state *s,
 	struct intel_vgpu *vgpu = s->vgpu;
 	struct intel_gvt *gvt = vgpu->gvt;
 	u32 ctx_sr_ctl;
+	u32 *vreg, vreg_old;
 
 	if (offset + 4 > gvt->device_info.mmio_size) {
 		gvt_vgpu_err("%s access to (%x) outside of MMIO range\n",
@@ -936,34 +898,101 @@ static int cmd_reg_handler(struct parser_exec_state *s,
 		return -EFAULT;
 	}
 
+	if (is_init_ctx(s)) {
+		struct intel_gvt_mmio_info *mmio_info;
+
+		intel_gvt_mmio_set_cmd_accessible(gvt, offset);
+		mmio_info = intel_gvt_find_mmio_info(gvt, offset);
+		if (mmio_info && mmio_info->write)
+			intel_gvt_mmio_set_cmd_write_patch(gvt, offset);
+		return 0;
+	}
+
 	if (!intel_gvt_mmio_is_cmd_accessible(gvt, offset)) {
 		gvt_vgpu_err("%s access to non-render register (%x)\n",
 				cmd, offset);
 		return -EBADRQC;
 	}
 
-	if (is_shadowed_mmio(offset)) {
-		gvt_vgpu_err("found access of shadowed MMIO %x\n", offset);
-		return 0;
+	if (!strncmp(cmd, "srm", 3) ||
+			!strncmp(cmd, "lrm", 3)) {
+		if (offset != i915_mmio_reg_offset(GEN8_L3SQCREG4) &&
+				offset != 0x21f0) {
+			gvt_vgpu_err("%s access to register (%x)\n",
+					cmd, offset);
+			return -EPERM;
+		} else
+			return 0;
 	}
 
-	if (is_mocs_mmio(offset) &&
-	    mocs_cmd_reg_handler(s, offset, index))
-		return -EINVAL;
+	if (!strncmp(cmd, "lrr-src", 7) ||
+			!strncmp(cmd, "lrr-dst", 7)) {
+		gvt_vgpu_err("not allowed cmd %s\n", cmd);
+		return -EPERM;
+	}
+
+	if (!strncmp(cmd, "pipe_ctrl", 9)) {
+		/* TODO: add LRI POST logic here */
+		return 0;
+	}
 
-	if (is_force_nonpriv_mmio(offset) &&
-		force_nonpriv_reg_handler(s, offset, index, cmd))
+	if (strncmp(cmd, "lri", 3))
 		return -EPERM;
 
+	/* below are all lri handlers */
+	vreg = &vgpu_vreg(s->vgpu, offset);
+	if (!intel_gvt_mmio_is_cmd_accessible(gvt, offset)) {
+		gvt_vgpu_err("%s access to non-render register (%x)\n",
+				cmd, offset);
+		return -EBADRQC;
+	}
+
+	if (is_cmd_update_pdps(offset, s) &&
+	    cmd_pdp_mmio_update_handler(s, offset, index))
+		return -EINVAL;
+
 	if (offset == i915_mmio_reg_offset(DERRMR) ||
 		offset == i915_mmio_reg_offset(FORCEWAKE_MT)) {
 		/* Writing to HW VGT_PVINFO_PAGE offset will be discarded */
 		patch_value(s, cmd_ptr(s, index), VGT_PVINFO_PAGE);
 	}
 
-	if (is_cmd_update_pdps(offset, s) &&
-	    cmd_pdp_mmio_update_handler(s, offset, index))
-		return -EINVAL;
+	if (is_mocs_mmio(offset))
+		*vreg = cmd_val(s, index + 1);
+
+	vreg_old = *vreg;
+
+	if (intel_gvt_mmio_is_cmd_write_patch(gvt, offset)) {
+		u32 cmdval_new, cmdval;
+		struct intel_gvt_mmio_info *mmio_info;
+
+		cmdval = cmd_val(s, index + 1);
+
+		mmio_info = intel_gvt_find_mmio_info(gvt, offset);
+		if (!mmio_info) {
+			cmdval_new = cmdval;
+		} else {
+			u64 ro_mask = mmio_info->ro_mask;
+			int ret;
+
+			if (likely(!ro_mask))
+				ret = mmio_info->write(s->vgpu, offset,
+						&cmdval, 4);
+			else {
+				gvt_vgpu_err("try to write RO reg %x\n",
+						offset);
+				ret = -EBADRQC;
+			}
+			if (ret)
+				return ret;
+			cmdval_new = *vreg;
+		}
+		if (cmdval_new != cmdval)
+			patch_value(s, cmd_ptr(s, index+1), cmdval_new);
+	}
+
+	/* only patch cmd. restore vreg value if changed in mmio write handler*/
+	*vreg = vreg_old;
 
 	/* TODO
 	 * In order to let workload with inhibit context to generate
@@ -1215,6 +1244,8 @@ static int cmd_handler_mi_batch_buffer_end(struct parser_exec_state *s)
 		s->buf_type = BATCH_BUFFER_INSTRUCTION;
 		ret = ip_gma_set(s, s->ret_ip_gma_bb);
 		s->buf_addr_type = s->saved_buf_addr_type;
+	} else if (s->buf_type == RING_BUFFER_CTX) {
+		ret = ip_gma_set(s, s->ring_tail);
 	} else {
 		s->buf_type = RING_BUFFER_INSTRUCTION;
 		s->buf_addr_type = GTT_BUFFER;
@@ -2763,7 +2794,8 @@ static int command_scan(struct parser_exec_state *s,
 	gma_bottom = rb_start +  rb_len;
 
 	while (s->ip_gma != gma_tail) {
-		if (s->buf_type == RING_BUFFER_INSTRUCTION) {
+		if (s->buf_type == RING_BUFFER_INSTRUCTION ||
+				s->buf_type == RING_BUFFER_CTX) {
 			if (!(s->ip_gma >= rb_start) ||
 				!(s->ip_gma < gma_bottom)) {
 				gvt_vgpu_err("ip_gma %lx out of ring scope."
@@ -3056,6 +3088,118 @@ int intel_gvt_scan_and_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 	return 0;
 }
 
+/* generate dummy contexts by sending empty requests to HW, and let
+ * the HW to fill Engine Contexts. This dummy contexts are used for
+ * initialization purpose (update reg whitelist), so referred to as
+ * init context here
+ */
+void intel_gvt_update_reg_whitelist(struct intel_vgpu *vgpu)
+{
+	const unsigned long start = LRC_STATE_PN * PAGE_SIZE;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	if (gvt->is_reg_whitelist_updated)
+		return;
+
+	/* scan init ctx to update cmd accessible list */
+	for_each_engine(engine, gvt->gt, id) {
+		struct parser_exec_state s;
+		void *vaddr;
+		int ret;
+
+		if (!engine->default_state)
+			continue;
+
+		vaddr = shmem_pin_map(engine->default_state);
+		if (IS_ERR(vaddr)) {
+			gvt_err("failed to map %s->default state, err:%zd\n",
+				engine->name, PTR_ERR(vaddr));
+			return;
+		}
+
+		s.buf_type = RING_BUFFER_CTX;
+		s.buf_addr_type = GTT_BUFFER;
+		s.vgpu = vgpu;
+		s.engine = engine;
+		s.ring_start = 0;
+		s.ring_size = engine->context_size - start;
+		s.ring_head = 0;
+		s.ring_tail = s.ring_size;
+		s.rb_va = vaddr + start;
+		s.workload = NULL;
+		s.is_ctx_wa = false;
+		s.is_init_ctx = true;
+
+		/* skipping the first RING_CTX_SIZE(0x50) dwords */
+		ret = ip_gma_set(&s, RING_CTX_SIZE);
+		if (ret == 0) {
+			ret = command_scan(&s, 0, s.ring_size, 0, s.ring_size);
+			if (ret)
+				gvt_err("Scan init ctx error\n");
+		}
+
+		shmem_unpin_map(engine->default_state, vaddr);
+		if (ret)
+			return;
+	}
+
+	gvt->is_reg_whitelist_updated = true;
+}
+
+int intel_gvt_scan_engine_context(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	unsigned long gma_head, gma_tail, gma_start, ctx_size;
+	struct parser_exec_state s;
+	int ring_id = workload->engine->id;
+	struct intel_context *ce = vgpu->submission.shadow[ring_id];
+	int ret;
+
+	GEM_BUG_ON(atomic_read(&ce->pin_count) < 0);
+
+	ctx_size = workload->engine->context_size - PAGE_SIZE;
+
+	/* Only ring contxt is loaded to HW for inhibit context, no need to
+	 * scan engine context
+	 */
+	if (is_inhibit_context(ce))
+		return 0;
+
+	gma_start = i915_ggtt_offset(ce->state) + LRC_STATE_PN*PAGE_SIZE;
+	gma_head = 0;
+	gma_tail = ctx_size;
+
+	s.buf_type = RING_BUFFER_CTX;
+	s.buf_addr_type = GTT_BUFFER;
+	s.vgpu = workload->vgpu;
+	s.engine = workload->engine;
+	s.ring_start = gma_start;
+	s.ring_size = ctx_size;
+	s.ring_head = gma_start + gma_head;
+	s.ring_tail = gma_start + gma_tail;
+	s.rb_va = ce->lrc_reg_state;
+	s.workload = workload;
+	s.is_ctx_wa = false;
+	s.is_init_ctx = false;
+
+	/* don't scan the first RING_CTX_SIZE(0x50) dwords, as it's ring
+	 * context
+	 */
+	ret = ip_gma_set(&s, gma_start + gma_head + RING_CTX_SIZE);
+	if (ret)
+		goto out;
+
+	ret = command_scan(&s, gma_head, gma_tail,
+		gma_start, ctx_size);
+out:
+	if (ret)
+		gvt_vgpu_err("scan shadow ctx error\n");
+
+	return ret;
+}
+
 static int init_cmd_table(struct intel_gvt *gvt)
 {
 	unsigned int gen_type = intel_gvt_get_device_type(gvt);
diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.h b/drivers/gpu/drm/i915/gvt/cmd_parser.h
index ab25d151932a..416d345e2816 100644
--- a/drivers/gpu/drm/i915/gvt/cmd_parser.h
+++ b/drivers/gpu/drm/i915/gvt/cmd_parser.h
@@ -40,6 +40,7 @@
 
 struct intel_gvt;
 struct intel_shadow_wa_ctx;
+struct intel_vgpu;
 struct intel_vgpu_workload;
 
 void intel_gvt_clean_cmd_parser(struct intel_gvt *gvt);
@@ -50,4 +51,8 @@ int intel_gvt_scan_and_shadow_ringbuffer(struct intel_vgpu_workload *workload);
 
 int intel_gvt_scan_and_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx);
 
+void intel_gvt_update_reg_whitelist(struct intel_vgpu *vgpu);
+
+int intel_gvt_scan_engine_context(struct intel_vgpu_workload *workload);
+
 #endif
diff --git a/drivers/gpu/drm/i915/gvt/display.c b/drivers/gpu/drm/i915/gvt/display.c
index a15f87539657..62a5b0dd2003 100644
--- a/drivers/gpu/drm/i915/gvt/display.c
+++ b/drivers/gpu/drm/i915/gvt/display.c
@@ -217,6 +217,15 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu)
 				  DDI_BUF_CTL_ENABLE);
 			vgpu_vreg_t(vgpu, DDI_BUF_CTL(port)) |= DDI_BUF_IS_IDLE;
 		}
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+			~(PORTA_HOTPLUG_ENABLE | PORTA_HOTPLUG_STATUS_MASK);
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+			~(PORTB_HOTPLUG_ENABLE | PORTB_HOTPLUG_STATUS_MASK);
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+			~(PORTC_HOTPLUG_ENABLE | PORTC_HOTPLUG_STATUS_MASK);
+		/* No hpd_invert set in vgpu vbt, need to clear invert mask */
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &= ~BXT_DDI_HPD_INVERT_MASK;
+		vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &= ~BXT_DE_PORT_HOTPLUG_MASK;
 
 		vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) &= ~(BIT(0) | BIT(1));
 		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) &=
@@ -273,6 +282,8 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu)
 			vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_EDP)) |=
 				(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
 				 TRANS_DDI_FUNC_ENABLE);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTA_HOTPLUG_ENABLE;
 			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
 				GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
 		}
@@ -301,6 +312,8 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu)
 				(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
 				 (PORT_B << TRANS_DDI_PORT_SHIFT) |
 				 TRANS_DDI_FUNC_ENABLE);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTB_HOTPLUG_ENABLE;
 			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
 				GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
 		}
@@ -329,6 +342,8 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu)
 				(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
 				 (PORT_B << TRANS_DDI_PORT_SHIFT) |
 				 TRANS_DDI_FUNC_ENABLE);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTC_HOTPLUG_ENABLE;
 			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
 				GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
 		}
@@ -661,44 +676,62 @@ void intel_vgpu_emulate_hotplug(struct intel_vgpu *vgpu, bool connected)
 				PORTD_HOTPLUG_STATUS_MASK;
 		intel_vgpu_trigger_virtual_event(vgpu, DP_D_HOTPLUG);
 	} else if (IS_BROXTON(i915)) {
-		if (connected) {
-			if (intel_vgpu_has_monitor_on_port(vgpu, PORT_A)) {
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_A)) {
+			if (connected) {
 				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
 					GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+			} else {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
+					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
 			}
-			if (intel_vgpu_has_monitor_on_port(vgpu, PORT_B)) {
-				vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
-					SFUSE_STRAP_DDIB_DETECTED;
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_IIR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+				~PORTA_HOTPLUG_STATUS_MASK;
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTA_HOTPLUG_LONG_DETECT;
+			intel_vgpu_trigger_virtual_event(vgpu, DP_A_HOTPLUG);
+		}
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_B)) {
+			if (connected) {
 				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
 					GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
-			}
-			if (intel_vgpu_has_monitor_on_port(vgpu, PORT_C)) {
 				vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
-					SFUSE_STRAP_DDIC_DETECTED;
-				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
-					GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
-			}
-		} else {
-			if (intel_vgpu_has_monitor_on_port(vgpu, PORT_A)) {
+					SFUSE_STRAP_DDIB_DETECTED;
+			} else {
 				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
-					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
-			}
-			if (intel_vgpu_has_monitor_on_port(vgpu, PORT_B)) {
+					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
 				vgpu_vreg_t(vgpu, SFUSE_STRAP) &=
 					~SFUSE_STRAP_DDIB_DETECTED;
-				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
-					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
 			}
-			if (intel_vgpu_has_monitor_on_port(vgpu, PORT_C)) {
-				vgpu_vreg_t(vgpu, SFUSE_STRAP) &=
-					~SFUSE_STRAP_DDIC_DETECTED;
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_IIR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+				~PORTB_HOTPLUG_STATUS_MASK;
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTB_HOTPLUG_LONG_DETECT;
+			intel_vgpu_trigger_virtual_event(vgpu, DP_B_HOTPLUG);
+		}
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_C)) {
+			if (connected) {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+					GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+				vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
+					SFUSE_STRAP_DDIC_DETECTED;
+			} else {
 				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
 					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+				vgpu_vreg_t(vgpu, SFUSE_STRAP) &=
+					~SFUSE_STRAP_DDIC_DETECTED;
 			}
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_IIR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+				~PORTC_HOTPLUG_STATUS_MASK;
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTC_HOTPLUG_LONG_DETECT;
+			intel_vgpu_trigger_virtual_event(vgpu, DP_C_HOTPLUG);
 		}
-		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
-			PORTB_HOTPLUG_STATUS_MASK;
-		intel_vgpu_trigger_virtual_event(vgpu, DP_B_HOTPLUG);
 	}
 }
 
diff --git a/drivers/gpu/drm/i915/gvt/execlist.c b/drivers/gpu/drm/i915/gvt/execlist.c
index 158873f269b1..c8dcda6d4f0d 100644
--- a/drivers/gpu/drm/i915/gvt/execlist.c
+++ b/drivers/gpu/drm/i915/gvt/execlist.c
@@ -522,12 +522,11 @@ static void init_vgpu_execlist(struct intel_vgpu *vgpu,
 static void clean_execlist(struct intel_vgpu *vgpu,
 			   intel_engine_mask_t engine_mask)
 {
-	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
-	struct intel_engine_cs *engine;
 	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_engine_cs *engine;
 	intel_engine_mask_t tmp;
 
-	for_each_engine_masked(engine, &dev_priv->gt, engine_mask, tmp) {
+	for_each_engine_masked(engine, vgpu->gvt->gt, engine_mask, tmp) {
 		kfree(s->ring_scan_buffer[engine->id]);
 		s->ring_scan_buffer[engine->id] = NULL;
 		s->ring_scan_buffer_size[engine->id] = 0;
@@ -537,11 +536,10 @@ static void clean_execlist(struct intel_vgpu *vgpu,
 static void reset_execlist(struct intel_vgpu *vgpu,
 			   intel_engine_mask_t engine_mask)
 {
-	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
 	struct intel_engine_cs *engine;
 	intel_engine_mask_t tmp;
 
-	for_each_engine_masked(engine, &dev_priv->gt, engine_mask, tmp)
+	for_each_engine_masked(engine, vgpu->gvt->gt, engine_mask, tmp)
 		init_vgpu_execlist(vgpu, engine);
 }
 
diff --git a/drivers/gpu/drm/i915/gvt/execlist.h b/drivers/gpu/drm/i915/gvt/execlist.h
index d62cd14605a3..84ad74b37d66 100644
--- a/drivers/gpu/drm/i915/gvt/execlist.h
+++ b/drivers/gpu/drm/i915/gvt/execlist.h
@@ -182,7 +182,4 @@ int intel_vgpu_init_execlist(struct intel_vgpu *vgpu);
 int intel_vgpu_submit_execlist(struct intel_vgpu *vgpu,
 			       const struct intel_engine_cs *engine);
 
-void intel_vgpu_reset_execlist(struct intel_vgpu *vgpu,
-			       intel_engine_mask_t engine_mask);
-
 #endif /*_GVT_EXECLIST_H_*/
diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.h b/drivers/gpu/drm/i915/gvt/fb_decoder.h
index 67b6ede9e707..0daa3931aef7 100644
--- a/drivers/gpu/drm/i915/gvt/fb_decoder.h
+++ b/drivers/gpu/drm/i915/gvt/fb_decoder.h
@@ -38,6 +38,10 @@
 
 #include <linux/types.h>
 
+#include "display/intel_display.h"
+
+struct intel_vgpu;
+
 #define _PLANE_CTL_FORMAT_SHIFT		24
 #define _PLANE_CTL_TILED_SHIFT		10
 #define _PIPE_V_SRCSZ_SHIFT		0
@@ -98,8 +102,6 @@ enum DDI_PORT {
 	DDI_PORT_E	= 4
 };
 
-struct intel_gvt;
-
 /* color space conversion and gamma correction are not included */
 struct intel_vgpu_primary_plane_format {
 	u8	enabled;	/* plane is enabled */
diff --git a/drivers/gpu/drm/i915/gvt/gtt.h b/drivers/gpu/drm/i915/gvt/gtt.h
index b0e173f2d990..3bf45672ef98 100644
--- a/drivers/gpu/drm/i915/gvt/gtt.h
+++ b/drivers/gpu/drm/i915/gvt/gtt.h
@@ -34,10 +34,19 @@
 #ifndef _GVT_GTT_H_
 #define _GVT_GTT_H_
 
-#define I915_GTT_PAGE_SHIFT         12
+#include <linux/kernel.h>
+#include <linux/kref.h>
+#include <linux/mutex.h>
+#include <linux/radix-tree.h>
+
+#include "gt/intel_gtt.h"
 
+struct intel_gvt;
+struct intel_vgpu;
 struct intel_vgpu_mm;
 
+#define I915_GTT_PAGE_SHIFT         12
+
 #define INTEL_GVT_INVALID_ADDR (~0UL)
 
 struct intel_gvt_gtt_entry {
diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h
index cf3578e3f4dd..03c993d68f10 100644
--- a/drivers/gpu/drm/i915/gvt/gvt.h
+++ b/drivers/gpu/drm/i915/gvt/gvt.h
@@ -33,6 +33,10 @@
 #ifndef _GVT_H_
 #define _GVT_H_
 
+#include <uapi/linux/pci_regs.h>
+
+#include "i915_drv.h"
+
 #include "debug.h"
 #include "hypercall.h"
 #include "mmio.h"
@@ -244,7 +248,7 @@ struct gvt_mmio_block {
 #define INTEL_GVT_MMIO_HASH_BITS 11
 
 struct intel_gvt_mmio {
-	u8 *mmio_attribute;
+	u16 *mmio_attribute;
 /* Register contains RO bits */
 #define F_RO		(1 << 0)
 /* Register contains graphics address */
@@ -263,6 +267,8 @@ struct intel_gvt_mmio {
  * logical context image
  */
 #define F_SR_IN_CTX	(1 << 7)
+/* Value of command write of this reg needs to be patched */
+#define F_CMD_WRITE_PATCH	(1 << 8)
 
 	struct gvt_mmio_block *mmio_block;
 	unsigned int num_mmio_block;
@@ -329,6 +335,7 @@ struct intel_gvt {
 		u32 *mocs_mmio_offset_list;
 		u32 mocs_mmio_offset_list_cnt;
 	} engine_mmio_list;
+	bool is_reg_whitelist_updated;
 
 	struct dentry *debugfs_root;
 };
@@ -412,6 +419,9 @@ int intel_gvt_load_firmware(struct intel_gvt *gvt);
 #define vgpu_fence_base(vgpu) (vgpu->fence.base)
 #define vgpu_fence_sz(vgpu) (vgpu->fence.size)
 
+/* ring context size i.e. the first 0x50 dwords*/
+#define RING_CTX_SIZE 320
+
 struct intel_vgpu_creation_params {
 	__u64 handle;
 	__u64 low_gm_sz;  /* in MB */
@@ -683,6 +693,35 @@ static inline void intel_gvt_mmio_set_sr_in_ctx(
 }
 
 void intel_gvt_debugfs_add_vgpu(struct intel_vgpu *vgpu);
+/**
+ * intel_gvt_mmio_set_cmd_write_patch -
+ *				mark an MMIO if its cmd write needs to be
+ *				patched
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ */
+static inline void intel_gvt_mmio_set_cmd_write_patch(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	gvt->mmio.mmio_attribute[offset >> 2] |= F_CMD_WRITE_PATCH;
+}
+
+/**
+ * intel_gvt_mmio_is_cmd_write_patch - check if an mmio's cmd access needs to
+ * be patched
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * True if GPU commmand write to an MMIO should be patched
+ */
+static inline bool intel_gvt_mmio_is_cmd_write_patch(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	return gvt->mmio.mmio_attribute[offset >> 2] & F_CMD_WRITE_PATCH;
+}
+
 void intel_gvt_debugfs_remove_vgpu(struct intel_vgpu *vgpu);
 void intel_gvt_debugfs_init(struct intel_gvt *gvt);
 void intel_gvt_debugfs_clean(struct intel_gvt *gvt);
diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c
index 3a4f5590640f..6eeaeecb7f85 100644
--- a/drivers/gpu/drm/i915/gvt/handlers.c
+++ b/drivers/gpu/drm/i915/gvt/handlers.c
@@ -83,7 +83,7 @@ static void write_vreg(struct intel_vgpu *vgpu, unsigned int offset,
 	memcpy(&vgpu_vreg(vgpu, offset), p_data, bytes);
 }
 
-static struct intel_gvt_mmio_info *find_mmio_info(struct intel_gvt *gvt,
+struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt,
 						  unsigned int offset)
 {
 	struct intel_gvt_mmio_info *e;
@@ -96,7 +96,7 @@ static struct intel_gvt_mmio_info *find_mmio_info(struct intel_gvt *gvt,
 }
 
 static int new_mmio_info(struct intel_gvt *gvt,
-		u32 offset, u8 flags, u32 size,
+		u32 offset, u16 flags, u32 size,
 		u32 addr_mask, u32 ro_mask, u32 device,
 		gvt_mmio_func read, gvt_mmio_func write)
 {
@@ -118,7 +118,7 @@ static int new_mmio_info(struct intel_gvt *gvt,
 			return -ENOMEM;
 
 		info->offset = i;
-		p = find_mmio_info(gvt, info->offset);
+		p = intel_gvt_find_mmio_info(gvt, info->offset);
 		if (p) {
 			WARN(1, "dup mmio definition offset %x\n",
 				info->offset);
@@ -1651,7 +1651,7 @@ static int edp_psr_imr_iir_write(struct intel_vgpu *vgpu,
 	return 0;
 }
 
-/**
+/*
  * FixMe:
  * If guest fills non-priv batch buffer on ApolloLake/Broxton as Mesa i965 did:
  * 717e7539124d (i965: Use a WC map and memcpy for the batch instead of pwrite.)
@@ -1965,7 +1965,8 @@ static int init_generic_mmio_info(struct intel_gvt *gvt)
 
 	/* RING MODE */
 #define RING_REG(base) _MMIO((base) + 0x29c)
-	MMIO_RING_DFH(RING_REG, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL,
+	MMIO_RING_DFH(RING_REG, D_ALL,
+		F_MODE_MASK | F_CMD_ACCESS | F_CMD_WRITE_PATCH, NULL,
 		ring_mode_mmio_write);
 #undef RING_REG
 
@@ -2885,8 +2886,8 @@ static int init_bdw_mmio_info(struct intel_gvt *gvt)
 	MMIO_DFH(_MMIO(0xb10c), D_BDW, F_CMD_ACCESS, NULL, NULL);
 	MMIO_D(_MMIO(0xb110), D_BDW);
 
-	MMIO_F(_MMIO(0x24d0), 48, F_CMD_ACCESS, 0, 0, D_BDW_PLUS,
-		NULL, force_nonpriv_write);
+	MMIO_F(_MMIO(0x24d0), 48, F_CMD_ACCESS | F_CMD_WRITE_PATCH, 0, 0,
+		D_BDW_PLUS, NULL, force_nonpriv_write);
 
 	MMIO_D(_MMIO(0x44484), D_BDW_PLUS);
 	MMIO_D(_MMIO(0x4448c), D_BDW_PLUS);
@@ -3626,7 +3627,7 @@ int intel_vgpu_mmio_reg_rw(struct intel_vgpu *vgpu, unsigned int offset,
 	/*
 	 * Normal tracked MMIOs.
 	 */
-	mmio_info = find_mmio_info(gvt, offset);
+	mmio_info = intel_gvt_find_mmio_info(gvt, offset);
 	if (!mmio_info) {
 		gvt_dbg_mmio("untracked MMIO %08x len %d\n", offset, bytes);
 		goto default_rw;
@@ -3686,8 +3687,7 @@ void intel_gvt_restore_fence(struct intel_gvt *gvt)
 	}
 }
 
-static inline int mmio_pm_restore_handler(struct intel_gvt *gvt,
-					  u32 offset, void *data)
+static int mmio_pm_restore_handler(struct intel_gvt *gvt, u32 offset, void *data)
 {
 	struct intel_vgpu *vgpu = data;
 	struct drm_i915_private *dev_priv = gvt->gt->i915;
diff --git a/drivers/gpu/drm/i915/gvt/interrupt.h b/drivers/gpu/drm/i915/gvt/interrupt.h
index fcd663811d37..287cd142629e 100644
--- a/drivers/gpu/drm/i915/gvt/interrupt.h
+++ b/drivers/gpu/drm/i915/gvt/interrupt.h
@@ -32,7 +32,10 @@
 #ifndef _GVT_INTERRUPT_H_
 #define _GVT_INTERRUPT_H_
 
-#include <linux/types.h>
+#include <linux/hrtimer.h>
+#include <linux/kernel.h>
+
+#include "i915_reg.h"
 
 enum intel_gvt_event_type {
 	RCS_MI_USER_INTERRUPT = 0,
diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c
index 60f1a386dd06..b4348256ae95 100644
--- a/drivers/gpu/drm/i915/gvt/kvmgt.c
+++ b/drivers/gpu/drm/i915/gvt/kvmgt.c
@@ -1703,7 +1703,7 @@ static int kvmgt_page_track_add(unsigned long handle, u64 gfn)
 		return -EINVAL;
 	}
 
-	spin_lock(&kvm->mmu_lock);
+	write_lock(&kvm->mmu_lock);
 
 	if (kvmgt_gfn_is_write_protected(info, gfn))
 		goto out;
@@ -1712,7 +1712,7 @@ static int kvmgt_page_track_add(unsigned long handle, u64 gfn)
 	kvmgt_protect_table_add(info, gfn);
 
 out:
-	spin_unlock(&kvm->mmu_lock);
+	write_unlock(&kvm->mmu_lock);
 	srcu_read_unlock(&kvm->srcu, idx);
 	return 0;
 }
@@ -1737,7 +1737,7 @@ static int kvmgt_page_track_remove(unsigned long handle, u64 gfn)
 		return -EINVAL;
 	}
 
-	spin_lock(&kvm->mmu_lock);
+	write_lock(&kvm->mmu_lock);
 
 	if (!kvmgt_gfn_is_write_protected(info, gfn))
 		goto out;
@@ -1746,7 +1746,7 @@ static int kvmgt_page_track_remove(unsigned long handle, u64 gfn)
 	kvmgt_protect_table_del(info, gfn);
 
 out:
-	spin_unlock(&kvm->mmu_lock);
+	write_unlock(&kvm->mmu_lock);
 	srcu_read_unlock(&kvm->srcu, idx);
 	return 0;
 }
@@ -1772,7 +1772,7 @@ static void kvmgt_page_track_flush_slot(struct kvm *kvm,
 	struct kvmgt_guest_info *info = container_of(node,
 					struct kvmgt_guest_info, track_node);
 
-	spin_lock(&kvm->mmu_lock);
+	write_lock(&kvm->mmu_lock);
 	for (i = 0; i < slot->npages; i++) {
 		gfn = slot->base_gfn + i;
 		if (kvmgt_gfn_is_write_protected(info, gfn)) {
@@ -1781,7 +1781,7 @@ static void kvmgt_page_track_flush_slot(struct kvm *kvm,
 			kvmgt_protect_table_del(info, gfn);
 		}
 	}
-	spin_unlock(&kvm->mmu_lock);
+	write_unlock(&kvm->mmu_lock);
 }
 
 static bool __kvmgt_vgpu_exist(struct intel_vgpu *vgpu, struct kvm *kvm)
diff --git a/drivers/gpu/drm/i915/gvt/mmio.h b/drivers/gpu/drm/i915/gvt/mmio.h
index 9e862dc73579..7c26af39fbfc 100644
--- a/drivers/gpu/drm/i915/gvt/mmio.h
+++ b/drivers/gpu/drm/i915/gvt/mmio.h
@@ -80,6 +80,9 @@ int intel_gvt_for_each_tracked_mmio(struct intel_gvt *gvt,
 	int (*handler)(struct intel_gvt *gvt, u32 offset, void *data),
 	void *data);
 
+struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt,
+						  unsigned int offset);
+
 int intel_vgpu_init_mmio(struct intel_vgpu *vgpu);
 void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu, bool dmlr);
 void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu);
diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c
index afe574d6b3b5..c9589e26af93 100644
--- a/drivers/gpu/drm/i915/gvt/mmio_context.c
+++ b/drivers/gpu/drm/i915/gvt/mmio_context.c
@@ -35,6 +35,7 @@
 
 #include "i915_drv.h"
 #include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_ring.h"
 #include "gvt.h"
 #include "trace.h"
diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.h b/drivers/gpu/drm/i915/gvt/mmio_context.h
index 3b25e7fe32f6..b6b69777af49 100644
--- a/drivers/gpu/drm/i915/gvt/mmio_context.h
+++ b/drivers/gpu/drm/i915/gvt/mmio_context.h
@@ -36,6 +36,18 @@
 #ifndef __GVT_RENDER_H__
 #define __GVT_RENDER_H__
 
+#include <linux/types.h>
+
+#include "gt/intel_engine_types.h"
+#include "gt/intel_lrc_reg.h"
+#include "i915_reg.h"
+
+struct i915_request;
+struct intel_context;
+struct intel_engine_cs;
+struct intel_gvt;
+struct intel_vgpu;
+
 struct engine_mmio {
 	enum intel_engine_id id;
 	i915_reg_t reg;
diff --git a/drivers/gpu/drm/i915/gvt/mpt.h b/drivers/gpu/drm/i915/gvt/mpt.h
index 6f92cde71971..550a456e936f 100644
--- a/drivers/gpu/drm/i915/gvt/mpt.h
+++ b/drivers/gpu/drm/i915/gvt/mpt.h
@@ -33,6 +33,8 @@
 #ifndef _GVT_MPT_H_
 #define _GVT_MPT_H_
 
+#include "gvt.h"
+
 /**
  * DOC: Hypervisor Service APIs for GVT-g Core Logic
  *
diff --git a/drivers/gpu/drm/i915/gvt/reg.h b/drivers/gpu/drm/i915/gvt/reg.h
index b58860dee970..244cc7320b54 100644
--- a/drivers/gpu/drm/i915/gvt/reg.h
+++ b/drivers/gpu/drm/i915/gvt/reg.h
@@ -133,4 +133,6 @@
 #define RING_GFX_MODE(base)	_MMIO((base) + 0x29c)
 #define VF_GUARDBAND		_MMIO(0x83a4)
 
+
+#define BCS_TILE_REGISTER_VAL_OFFSET (0x43*4)
 #endif
diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
index aed2ef6466a2..fc735692f21f 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.c
+++ b/drivers/gpu/drm/i915/gvt/scheduler.c
@@ -37,6 +37,8 @@
 
 #include "gem/i915_gem_pm.h"
 #include "gt/intel_context.h"
+#include "gt/intel_execlists_submission.h"
+#include "gt/intel_lrc.h"
 #include "gt/intel_ring.h"
 
 #include "i915_drv.h"
@@ -135,6 +137,7 @@ static int populate_shadow_context(struct intel_vgpu_workload *workload)
 	int i;
 	bool skip = false;
 	int ring_id = workload->engine->id;
+	int ret;
 
 	GEM_BUG_ON(!intel_context_is_pinned(ctx));
 
@@ -161,16 +164,24 @@ static int populate_shadow_context(struct intel_vgpu_workload *workload)
 		COPY_REG(bb_per_ctx_ptr);
 		COPY_REG(rcs_indirect_ctx);
 		COPY_REG(rcs_indirect_ctx_offset);
-	}
+	} else if (workload->engine->id == BCS0)
+		intel_gvt_hypervisor_read_gpa(vgpu,
+				workload->ring_context_gpa +
+				BCS_TILE_REGISTER_VAL_OFFSET,
+				(void *)shadow_ring_context +
+				BCS_TILE_REGISTER_VAL_OFFSET, 4);
 #undef COPY_REG
 #undef COPY_REG_MASKED
 
+	/* don't copy Ring Context (the first 0x50 dwords),
+	 * only copy the Engine Context part from guest
+	 */
 	intel_gvt_hypervisor_read_gpa(vgpu,
 			workload->ring_context_gpa +
-			sizeof(*shadow_ring_context),
+			RING_CTX_SIZE,
 			(void *)shadow_ring_context +
-			sizeof(*shadow_ring_context),
-			I915_GTT_PAGE_SIZE - sizeof(*shadow_ring_context));
+			RING_CTX_SIZE,
+			I915_GTT_PAGE_SIZE - RING_CTX_SIZE);
 
 	sr_oa_regs(workload, (u32 *)shadow_ring_context, false);
 
@@ -237,6 +248,11 @@ read:
 		gpa_size = I915_GTT_PAGE_SIZE;
 		dst = context_base + (i << I915_GTT_PAGE_SHIFT);
 	}
+	ret = intel_gvt_scan_engine_context(workload);
+	if (ret) {
+		gvt_vgpu_err("invalid cmd found in guest context pages\n");
+		return ret;
+	}
 	s->last_ctx[ring_id].valid = true;
 	return 0;
 }
@@ -396,7 +412,9 @@ static void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 	if (!wa_ctx->indirect_ctx.obj)
 		return;
 
+	i915_gem_object_lock(wa_ctx->indirect_ctx.obj, NULL);
 	i915_gem_object_unpin_map(wa_ctx->indirect_ctx.obj);
+	i915_gem_object_unlock(wa_ctx->indirect_ctx.obj);
 	i915_gem_object_put(wa_ctx->indirect_ctx.obj);
 
 	wa_ctx->indirect_ctx.obj = NULL;
@@ -504,6 +522,7 @@ static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 	struct intel_gvt *gvt = workload->vgpu->gvt;
 	const int gmadr_bytes = gvt->device_info.gmadr_bytes_in_cmd;
 	struct intel_vgpu_shadow_bb *bb;
+	struct i915_gem_ww_ctx ww;
 	int ret;
 
 	list_for_each_entry(bb, &workload->shadow_bb, list) {
@@ -528,10 +547,19 @@ static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 		 * directly
 		 */
 		if (!bb->ppgtt) {
-			bb->vma = i915_gem_object_ggtt_pin(bb->obj,
-							   NULL, 0, 0, 0);
+			i915_gem_ww_ctx_init(&ww, false);
+retry:
+			i915_gem_object_lock(bb->obj, &ww);
+
+			bb->vma = i915_gem_object_ggtt_pin_ww(bb->obj, &ww,
+							      NULL, 0, 0, 0);
 			if (IS_ERR(bb->vma)) {
 				ret = PTR_ERR(bb->vma);
+				if (ret == -EDEADLK) {
+					ret = i915_gem_ww_ctx_backoff(&ww);
+					if (!ret)
+						goto retry;
+				}
 				goto err;
 			}
 
@@ -545,13 +573,15 @@ static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 						      0);
 			if (ret)
 				goto err;
-		}
 
-		/* No one is going to touch shadow bb from now on. */
-		i915_gem_object_flush_map(bb->obj);
+			/* No one is going to touch shadow bb from now on. */
+			i915_gem_object_flush_map(bb->obj);
+			i915_gem_object_unlock(bb->obj);
+		}
 	}
 	return 0;
 err:
+	i915_gem_ww_ctx_fini(&ww);
 	release_shadow_batch_buffer(workload);
 	return ret;
 }
@@ -578,14 +608,29 @@ static int prepare_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 	unsigned char *per_ctx_va =
 		(unsigned char *)wa_ctx->indirect_ctx.shadow_va +
 		wa_ctx->indirect_ctx.size;
+	struct i915_gem_ww_ctx ww;
+	int ret;
 
 	if (wa_ctx->indirect_ctx.size == 0)
 		return 0;
 
-	vma = i915_gem_object_ggtt_pin(wa_ctx->indirect_ctx.obj, NULL,
-				       0, CACHELINE_BYTES, 0);
-	if (IS_ERR(vma))
-		return PTR_ERR(vma);
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	i915_gem_object_lock(wa_ctx->indirect_ctx.obj, &ww);
+
+	vma = i915_gem_object_ggtt_pin_ww(wa_ctx->indirect_ctx.obj, &ww, NULL,
+					  0, CACHELINE_BYTES, 0);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		if (ret == -EDEADLK) {
+			ret = i915_gem_ww_ctx_backoff(&ww);
+			if (!ret)
+				goto retry;
+		}
+		return ret;
+	}
+
+	i915_gem_object_unlock(wa_ctx->indirect_ctx.obj);
 
 	/* FIXME: we are not tracking our pinned VMA leaving it
 	 * up to the core to fix up the stray pin_count upon
@@ -619,12 +664,14 @@ static void release_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 
 	list_for_each_entry_safe(bb, pos, &workload->shadow_bb, list) {
 		if (bb->obj) {
+			i915_gem_object_lock(bb->obj, NULL);
 			if (bb->va && !IS_ERR(bb->va))
 				i915_gem_object_unpin_map(bb->obj);
 
 			if (bb->vma && !IS_ERR(bb->vma))
 				i915_vma_unpin(bb->vma);
 
+			i915_gem_object_unlock(bb->obj);
 			i915_gem_object_put(bb->obj);
 		}
 		list_del(&bb->list);
@@ -999,13 +1046,12 @@ void intel_vgpu_clean_workloads(struct intel_vgpu *vgpu,
 				intel_engine_mask_t engine_mask)
 {
 	struct intel_vgpu_submission *s = &vgpu->submission;
-	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
 	struct intel_engine_cs *engine;
 	struct intel_vgpu_workload *pos, *n;
 	intel_engine_mask_t tmp;
 
 	/* free the unsubmited workloads in the queues. */
-	for_each_engine_masked(engine, &dev_priv->gt, engine_mask, tmp) {
+	for_each_engine_masked(engine, vgpu->gvt->gt, engine_mask, tmp) {
 		list_for_each_entry_safe(pos, n,
 			&s->workload_q_head[engine->id], list) {
 			list_del_init(&pos->list);
diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h
index 64e7a0b791c3..7c86984a842f 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.h
+++ b/drivers/gpu/drm/i915/gvt/scheduler.h
@@ -36,6 +36,11 @@
 #ifndef _GVT_SCHEDULER_H_
 #define _GVT_SCHEDULER_H_
 
+#include "gt/intel_engine_types.h"
+
+#include "execlist.h"
+#include "interrupt.h"
+
 struct intel_gvt_workload_scheduler {
 	struct intel_vgpu *current_vgpu;
 	struct intel_vgpu *next_vgpu;
diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c
index e49944fde333..6a16d0ca7cda 100644
--- a/drivers/gpu/drm/i915/gvt/vgpu.c
+++ b/drivers/gpu/drm/i915/gvt/vgpu.c
@@ -437,10 +437,9 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt,
 	if (ret)
 		goto out_clean_sched_policy;
 
-	if (IS_BROADWELL(dev_priv))
+	if (IS_BROADWELL(dev_priv) || IS_BROXTON(dev_priv))
 		ret = intel_gvt_hypervisor_set_edid(vgpu, PORT_B);
-	/* FixMe: Re-enable APL/BXT once vfio_edid enabled */
-	else if (!IS_BROXTON(dev_priv))
+	else
 		ret = intel_gvt_hypervisor_set_edid(vgpu, PORT_D);
 	if (ret)
 		goto out_clean_sched_policy;
@@ -500,9 +499,11 @@ struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt,
 
 	mutex_lock(&gvt->lock);
 	vgpu = __intel_gvt_create_vgpu(gvt, &param);
-	if (!IS_ERR(vgpu))
+	if (!IS_ERR(vgpu)) {
 		/* calculate left instance change for types */
 		intel_gvt_update_vgpu_types(gvt);
+		intel_gvt_update_reg_whitelist(vgpu);
+	}
 	mutex_unlock(&gvt->lock);
 
 	return vgpu;
diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
index 10a865f3dc09..3bc616cc1ad2 100644
--- a/drivers/gpu/drm/i915/i915_active.c
+++ b/drivers/gpu/drm/i915/i915_active.c
@@ -159,8 +159,7 @@ __active_retire(struct i915_active *ref)
 		GEM_BUG_ON(ref->tree.rb_node != &ref->cache->node);
 
 		/* Make the cached node available for reuse with any timeline */
-		if (IS_ENABLED(CONFIG_64BIT))
-			ref->cache->timeline = 0; /* needs cmpxchg(u64) */
+		ref->cache->timeline = 0; /* needs cmpxchg(u64) */
 	}
 
 	spin_unlock_irqrestore(&ref->tree_lock, flags);
@@ -256,7 +255,6 @@ static struct active_node *__active_lookup(struct i915_active *ref, u64 idx)
 		if (cached == idx)
 			return it;
 
-#ifdef CONFIG_64BIT /* for cmpxchg(u64) */
 		/*
 		 * An unclaimed cache [.timeline=0] can only be claimed once.
 		 *
@@ -267,9 +265,8 @@ static struct active_node *__active_lookup(struct i915_active *ref, u64 idx)
 		 * only the winner of that race will cmpxchg return the old
 		 * value of 0).
 		 */
-		if (!cached && !cmpxchg(&it->timeline, 0, idx))
+		if (!cached && !cmpxchg64(&it->timeline, 0, idx))
 			return it;
-#endif
 	}
 
 	BUILD_BUG_ON(offsetof(typeof(*it), node));
@@ -631,24 +628,26 @@ static int flush_lazy_signals(struct i915_active *ref)
 
 int __i915_active_wait(struct i915_active *ref, int state)
 {
-	int err;
-
 	might_sleep();
 
-	if (!i915_active_acquire_if_busy(ref))
-		return 0;
-
 	/* Any fence added after the wait begins will not be auto-signaled */
-	err = flush_lazy_signals(ref);
-	i915_active_release(ref);
-	if (err)
-		return err;
+	if (i915_active_acquire_if_busy(ref)) {
+		int err;
 
-	if (!i915_active_is_idle(ref) &&
-	    ___wait_var_event(ref, i915_active_is_idle(ref),
-			      state, 0, 0, schedule()))
-		return -EINTR;
+		err = flush_lazy_signals(ref);
+		i915_active_release(ref);
+		if (err)
+			return err;
 
+		if (___wait_var_event(ref, i915_active_is_idle(ref),
+				      state, 0, 0, schedule()))
+			return -EINTR;
+	}
+
+	/*
+	 * After the wait is complete, the caller may free the active.
+	 * We have to flush any concurrent retirement before returning.
+	 */
 	flush_work(&ref->work);
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
index 93265951fdbb..5f86f5b2caf6 100644
--- a/drivers/gpu/drm/i915/i915_cmd_parser.c
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -26,6 +26,7 @@
  */
 
 #include "gt/intel_engine.h"
+#include "gt/intel_gpu_commands.h"
 
 #include "i915_drv.h"
 #include "i915_memcpy.h"
@@ -939,7 +940,7 @@ static void fini_hash_table(struct intel_engine_cs *engine)
  * struct intel_engine_cs based on whether the platform requires software
  * command parsing.
  */
-void intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
+int intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
 {
 	const struct drm_i915_cmd_table *cmd_tables;
 	int cmd_table_count;
@@ -947,7 +948,7 @@ void intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
 
 	if (!IS_GEN(engine->i915, 7) && !(IS_GEN(engine->i915, 9) &&
 					  engine->class == COPY_ENGINE_CLASS))
-		return;
+		return 0;
 
 	switch (engine->class) {
 	case RENDER_CLASS:
@@ -1012,19 +1013,19 @@ void intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
 		break;
 	default:
 		MISSING_CASE(engine->class);
-		return;
+		goto out;
 	}
 
 	if (!validate_cmds_sorted(engine, cmd_tables, cmd_table_count)) {
 		drm_err(&engine->i915->drm,
 			"%s: command descriptions are not sorted\n",
 			engine->name);
-		return;
+		goto out;
 	}
 	if (!validate_regs_sorted(engine)) {
 		drm_err(&engine->i915->drm,
 			"%s: registers are not sorted\n", engine->name);
-		return;
+		goto out;
 	}
 
 	ret = init_hash_table(engine, cmd_tables, cmd_table_count);
@@ -1032,10 +1033,17 @@ void intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
 		drm_err(&engine->i915->drm,
 			"%s: initialised failed!\n", engine->name);
 		fini_hash_table(engine);
-		return;
+		goto out;
 	}
 
 	engine->flags |= I915_ENGINE_USING_CMD_PARSER;
+
+out:
+	if (intel_engine_requires_cmd_parser(engine) &&
+	    !intel_engine_using_cmd_parser(engine))
+		return -EINVAL;
+
+	return 0;
 }
 
 /**
@@ -1142,7 +1150,7 @@ static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
 	void *dst, *src;
 	int ret;
 
-	dst = i915_gem_object_pin_map(dst_obj, I915_MAP_FORCE_WB);
+	dst = i915_gem_object_pin_map(dst_obj, I915_MAP_WB);
 	if (IS_ERR(dst))
 		return dst;
 
@@ -1166,7 +1174,7 @@ static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
 		}
 	}
 	if (IS_ERR(src)) {
-		unsigned long x, n;
+		unsigned long x, n, remain;
 		void *ptr;
 
 		/*
@@ -1177,14 +1185,15 @@ static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
 		 * We don't care about copying too much here as we only
 		 * validate up to the end of the batch.
 		 */
+		remain = length;
 		if (!(dst_obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
-			length = round_up(length,
+			remain = round_up(remain,
 					  boot_cpu_data.x86_clflush_size);
 
 		ptr = dst;
 		x = offset_in_page(offset);
-		for (n = offset >> PAGE_SHIFT; length; n++) {
-			int len = min(length, PAGE_SIZE - x);
+		for (n = offset >> PAGE_SHIFT; remain; n++) {
+			int len = min(remain, PAGE_SIZE - x);
 
 			src = kmap_atomic(i915_gem_object_get_page(src_obj, n));
 			if (needs_clflush)
@@ -1193,13 +1202,15 @@ static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
 			kunmap_atomic(src);
 
 			ptr += len;
-			length -= len;
+			remain -= len;
 			x = 0;
 		}
 	}
 
 	i915_gem_object_unpin_pages(src_obj);
 
+	memset32(dst + length, 0, (dst_obj->base.size - length) / sizeof(u32));
+
 	/* dst_obj is returned with vmap pinned */
 	return dst;
 }
@@ -1392,11 +1403,6 @@ static unsigned long *alloc_whitelist(u32 batch_length)
 
 #define LENGTH_BIAS 2
 
-static bool shadow_needs_clflush(struct drm_i915_gem_object *obj)
-{
-	return !(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE);
-}
-
 /**
  * intel_engine_cmd_parser() - parse a batch buffer for privilege violations
  * @engine: the engine on which the batch is to execute
@@ -1538,16 +1544,9 @@ int intel_engine_cmd_parser(struct intel_engine_cs *engine,
 				ret = 0; /* allow execution */
 			}
 		}
-
-		if (shadow_needs_clflush(shadow->obj))
-			drm_clflush_virt_range(batch_end, 8);
 	}
 
-	if (shadow_needs_clflush(shadow->obj)) {
-		void *ptr = page_mask_bits(shadow->obj->mm.mapping);
-
-		drm_clflush_virt_range(ptr, (void *)(cmd + 1) - ptr);
-	}
+	i915_gem_object_flush_map(shadow->obj);
 
 	if (!IS_ERR_OR_NULL(jump_whitelist))
 		kfree(jump_whitelist);
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index a61e9b095896..88336ff4bf09 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -45,6 +45,7 @@
 #include "i915_debugfs.h"
 #include "i915_debugfs_params.h"
 #include "i915_irq.h"
+#include "i915_scheduler.h"
 #include "i915_trace.h"
 #include "intel_pm.h"
 #include "intel_sideband.h"
@@ -209,7 +210,7 @@ i915_debugfs_describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 	spin_unlock(&obj->vma.lock);
 
 	seq_printf(m, " (pinned x %d)", pin_count);
-	if (obj->stolen)
+	if (i915_gem_object_is_stolen(obj))
 		seq_printf(m, " (stolen: %08llx)", obj->stolen->start);
 	if (i915_gem_object_is_framebuffer(obj))
 		seq_printf(m, " (fb)");
@@ -219,145 +220,6 @@ i915_debugfs_describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 		seq_printf(m, " (%s)", engine->name);
 }
 
-struct file_stats {
-	struct i915_address_space *vm;
-	unsigned long count;
-	u64 total;
-	u64 active, inactive;
-	u64 closed;
-};
-
-static int per_file_stats(int id, void *ptr, void *data)
-{
-	struct drm_i915_gem_object *obj = ptr;
-	struct file_stats *stats = data;
-	struct i915_vma *vma;
-
-	if (IS_ERR_OR_NULL(obj) || !kref_get_unless_zero(&obj->base.refcount))
-		return 0;
-
-	stats->count++;
-	stats->total += obj->base.size;
-
-	spin_lock(&obj->vma.lock);
-	if (!stats->vm) {
-		for_each_ggtt_vma(vma, obj) {
-			if (!drm_mm_node_allocated(&vma->node))
-				continue;
-
-			if (i915_vma_is_active(vma))
-				stats->active += vma->node.size;
-			else
-				stats->inactive += vma->node.size;
-
-			if (i915_vma_is_closed(vma))
-				stats->closed += vma->node.size;
-		}
-	} else {
-		struct rb_node *p = obj->vma.tree.rb_node;
-
-		while (p) {
-			long cmp;
-
-			vma = rb_entry(p, typeof(*vma), obj_node);
-			cmp = i915_vma_compare(vma, stats->vm, NULL);
-			if (cmp == 0) {
-				if (drm_mm_node_allocated(&vma->node)) {
-					if (i915_vma_is_active(vma))
-						stats->active += vma->node.size;
-					else
-						stats->inactive += vma->node.size;
-
-					if (i915_vma_is_closed(vma))
-						stats->closed += vma->node.size;
-				}
-				break;
-			}
-			if (cmp < 0)
-				p = p->rb_right;
-			else
-				p = p->rb_left;
-		}
-	}
-	spin_unlock(&obj->vma.lock);
-
-	i915_gem_object_put(obj);
-	return 0;
-}
-
-#define print_file_stats(m, name, stats) do { \
-	if (stats.count) \
-		seq_printf(m, "%s: %lu objects, %llu bytes (%llu active, %llu inactive, %llu closed)\n", \
-			   name, \
-			   stats.count, \
-			   stats.total, \
-			   stats.active, \
-			   stats.inactive, \
-			   stats.closed); \
-} while (0)
-
-static void print_context_stats(struct seq_file *m,
-				struct drm_i915_private *i915)
-{
-	struct file_stats kstats = {};
-	struct i915_gem_context *ctx, *cn;
-
-	spin_lock(&i915->gem.contexts.lock);
-	list_for_each_entry_safe(ctx, cn, &i915->gem.contexts.list, link) {
-		struct i915_gem_engines_iter it;
-		struct intel_context *ce;
-
-		if (!kref_get_unless_zero(&ctx->ref))
-			continue;
-
-		spin_unlock(&i915->gem.contexts.lock);
-
-		for_each_gem_engine(ce,
-				    i915_gem_context_lock_engines(ctx), it) {
-			if (intel_context_pin_if_active(ce)) {
-				rcu_read_lock();
-				if (ce->state)
-					per_file_stats(0,
-						       ce->state->obj, &kstats);
-				per_file_stats(0, ce->ring->vma->obj, &kstats);
-				rcu_read_unlock();
-				intel_context_unpin(ce);
-			}
-		}
-		i915_gem_context_unlock_engines(ctx);
-
-		mutex_lock(&ctx->mutex);
-		if (!IS_ERR_OR_NULL(ctx->file_priv)) {
-			struct file_stats stats = {
-				.vm = rcu_access_pointer(ctx->vm),
-			};
-			struct drm_file *file = ctx->file_priv->file;
-			struct task_struct *task;
-			char name[80];
-
-			rcu_read_lock();
-			idr_for_each(&file->object_idr, per_file_stats, &stats);
-			rcu_read_unlock();
-
-			rcu_read_lock();
-			task = pid_task(ctx->pid ?: file->pid, PIDTYPE_PID);
-			snprintf(name, sizeof(name), "%s",
-				 task ? task->comm : "<unknown>");
-			rcu_read_unlock();
-
-			print_file_stats(m, name, stats);
-		}
-		mutex_unlock(&ctx->mutex);
-
-		spin_lock(&i915->gem.contexts.lock);
-		list_safe_reset_next(ctx, cn, link);
-		i915_gem_context_put(ctx);
-	}
-	spin_unlock(&i915->gem.contexts.lock);
-
-	print_file_stats(m, "[k]contexts", kstats);
-}
-
 static int i915_gem_object_info(struct seq_file *m, void *data)
 {
 	struct drm_i915_private *i915 = node_to_i915(m->private);
@@ -371,9 +233,6 @@ static int i915_gem_object_info(struct seq_file *m, void *data)
 	for_each_memory_region(mr, i915, id)
 		seq_printf(m, "%s: total:%pa, available:%pa bytes\n",
 			   mr->name, &mr->total, &mr->avail);
-	seq_putc(m, '\n');
-
-	print_context_stats(m, i915);
 
 	return 0;
 }
@@ -634,27 +493,27 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		seq_printf(m, "RPDECLIMIT: 0x%08x\n", rpdeclimit);
 		seq_printf(m, "RPNSWREQ: %dMHz\n", reqf);
 		seq_printf(m, "CAGF: %dMHz\n", cagf);
-		seq_printf(m, "RP CUR UP EI: %d (%dns)\n",
+		seq_printf(m, "RP CUR UP EI: %d (%lldns)\n",
 			   rpupei,
 			   intel_gt_pm_interval_to_ns(&dev_priv->gt, rpupei));
-		seq_printf(m, "RP CUR UP: %d (%dun)\n",
+		seq_printf(m, "RP CUR UP: %d (%lldun)\n",
 			   rpcurup,
 			   intel_gt_pm_interval_to_ns(&dev_priv->gt, rpcurup));
-		seq_printf(m, "RP PREV UP: %d (%dns)\n",
+		seq_printf(m, "RP PREV UP: %d (%lldns)\n",
 			   rpprevup,
 			   intel_gt_pm_interval_to_ns(&dev_priv->gt, rpprevup));
 		seq_printf(m, "Up threshold: %d%%\n",
 			   rps->power.up_threshold);
 
-		seq_printf(m, "RP CUR DOWN EI: %d (%dns)\n",
+		seq_printf(m, "RP CUR DOWN EI: %d (%lldns)\n",
 			   rpdownei,
 			   intel_gt_pm_interval_to_ns(&dev_priv->gt,
 						      rpdownei));
-		seq_printf(m, "RP CUR DOWN: %d (%dns)\n",
+		seq_printf(m, "RP CUR DOWN: %d (%lldns)\n",
 			   rpcurdown,
 			   intel_gt_pm_interval_to_ns(&dev_priv->gt,
 						      rpcurdown));
-		seq_printf(m, "RP PREV DOWN: %d (%dns)\n",
+		seq_printf(m, "RP PREV DOWN: %d (%lldns)\n",
 			   rpprevdown,
 			   intel_gt_pm_interval_to_ns(&dev_priv->gt,
 						      rpprevdown));
@@ -810,7 +669,7 @@ static int i915_rps_boost_info(struct seq_file *m, void *data)
 		   intel_gpu_freq(rps, rps->efficient_freq),
 		   intel_gpu_freq(rps, rps->boost_freq));
 
-	seq_printf(m, "Wait boosts: %d\n", atomic_read(&rps->boosts));
+	seq_printf(m, "Wait boosts: %d\n", READ_ONCE(rps->boosts));
 
 	return 0;
 }
@@ -850,24 +709,28 @@ static int i915_runtime_pm_status(struct seq_file *m, void *unused)
 
 static int i915_engine_info(struct seq_file *m, void *unused)
 {
-	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_i915_private *i915 = node_to_i915(m->private);
 	struct intel_engine_cs *engine;
 	intel_wakeref_t wakeref;
 	struct drm_printer p;
 
-	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
 
-	seq_printf(m, "GT awake? %s [%d]\n",
-		   yesno(dev_priv->gt.awake),
-		   atomic_read(&dev_priv->gt.wakeref.count));
-	seq_printf(m, "CS timestamp frequency: %u Hz\n",
-		   RUNTIME_INFO(dev_priv)->cs_timestamp_frequency_hz);
+	seq_printf(m, "GT awake? %s [%d], %llums\n",
+		   yesno(i915->gt.awake),
+		   atomic_read(&i915->gt.wakeref.count),
+		   ktime_to_ms(intel_gt_get_awake_time(&i915->gt)));
+	seq_printf(m, "CS timestamp frequency: %u Hz, %d ns\n",
+		   i915->gt.clock_frequency,
+		   i915->gt.clock_period_ns);
 
 	p = drm_seq_file_printer(m);
-	for_each_uabi_engine(engine, dev_priv)
+	for_each_uabi_engine(engine, i915)
 		intel_engine_dump(engine, &p, "%s\n", engine->name);
 
-	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+	intel_gt_show_timelines(&i915->gt, &p, i915_request_show_with_schedule);
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
 
 	return 0;
 }
@@ -945,7 +808,7 @@ i915_perf_noa_delay_set(void *data, u64 val)
 	 * This would lead to infinite waits as we're doing timestamp
 	 * difference on the CS with only 32bits.
 	 */
-	if (i915_cs_timestamp_ns_to_ticks(i915, val) > U32_MAX)
+	if (intel_gt_ns_to_clock_interval(&i915->gt, val) > U32_MAX)
 		return -EINVAL;
 
 	atomic64_set(&i915->perf.noa_programming_delay, val);
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 249f765993f7..8e9cb44e66e5 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -58,12 +58,14 @@
 #include "display/intel_hotplug.h"
 #include "display/intel_overlay.h"
 #include "display/intel_pipe_crc.h"
+#include "display/intel_pps.h"
 #include "display/intel_sprite.h"
 #include "display/intel_vga.h"
 
 #include "gem/i915_gem_context.h"
 #include "gem/i915_gem_ioctls.h"
 #include "gem/i915_gem_mman.h"
+#include "gem/i915_gem_pm.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_pm.h"
 #include "gt/intel_rc6.h"
@@ -410,6 +412,7 @@ static int i915_driver_mmio_probe(struct drm_i915_private *dev_priv)
 
 	/* Try to make sure MCHBAR is enabled before poking at it */
 	intel_setup_mchbar(dev_priv);
+	intel_device_info_runtime_init(dev_priv);
 
 	ret = intel_gt_init_mmio(&dev_priv->gt);
 	if (ret)
@@ -516,8 +519,6 @@ static int i915_driver_hw_probe(struct drm_i915_private *dev_priv)
 	if (i915_inject_probe_failure(dev_priv))
 		return -ENODEV;
 
-	intel_device_info_runtime_init(dev_priv);
-
 	if (HAS_PPGTT(dev_priv)) {
 		if (intel_vgpu_active(dev_priv) &&
 		    !intel_vgpu_has_full_ppgtt(dev_priv)) {
@@ -609,14 +610,15 @@ static int i915_driver_hw_probe(struct drm_i915_private *dev_priv)
 		goto err_msi;
 
 	intel_opregion_setup(dev_priv);
+
+	intel_pcode_init(dev_priv);
+
 	/*
-	 * Fill the dram structure to get the system raw bandwidth and
-	 * dram info. This will be used for memory latency calculation.
+	 * Fill the dram structure to get the system dram info. This will be
+	 * used for memory latency calculation.
 	 */
 	intel_dram_detect(dev_priv);
 
-	intel_pcode_init(dev_priv);
-
 	intel_bw_init_hw(dev_priv);
 
 	return 0;
@@ -1046,6 +1048,8 @@ static void intel_shutdown_encoders(struct drm_i915_private *dev_priv)
 
 void i915_driver_shutdown(struct drm_i915_private *i915)
 {
+	disable_rpm_wakeref_asserts(&i915->runtime_pm);
+
 	i915_gem_suspend(i915);
 
 	drm_kms_helper_poll_disable(&i915->drm);
@@ -1059,6 +1063,8 @@ void i915_driver_shutdown(struct drm_i915_private *i915)
 
 	intel_suspend_encoders(i915);
 	intel_shutdown_encoders(i915);
+
+	enable_rpm_wakeref_asserts(&i915->runtime_pm);
 }
 
 static bool suspend_to_idle(struct drm_i915_private *dev_priv)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 2688f3e3e349..cb62ddba2035 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -79,9 +79,9 @@
 #include "gem/i915_gem_shrinker.h"
 #include "gem/i915_gem_stolen.h"
 
-#include "gt/intel_lrc.h"
 #include "gt/intel_engine.h"
 #include "gt/intel_gt_types.h"
+#include "gt/intel_region_lmem.h"
 #include "gt/intel_workarounds.h"
 #include "gt/uc/intel_uc.h"
 
@@ -103,7 +103,6 @@
 #include "i915_vma.h"
 #include "i915_irq.h"
 
-#include "intel_region_lmem.h"
 
 /* General customization:
  */
@@ -1123,23 +1122,11 @@ struct drm_i915_private {
 		 * crtc_state->wm.need_postvbl_update.
 		 */
 		struct mutex wm_mutex;
-
-		/*
-		 * Set during HW readout of watermarks/DDB.  Some platforms
-		 * need to know when we're still using BIOS-provided values
-		 * (which we don't fully trust).
-		 *
-		 * FIXME get rid of this.
-		 */
-		bool distrust_bios_wm;
 	} wm;
 
 	struct dram_info {
-		bool valid;
-		bool is_16gb_dimm;
+		bool wm_lv_0_adjust_needed;
 		u8 num_channels;
-		u8 ranks;
-		u32 bandwidth_kbps;
 		bool symmetric_memory;
 		enum intel_dram_type {
 			INTEL_DRAM_UNKNOWN,
@@ -1148,6 +1135,7 @@ struct drm_i915_private {
 			INTEL_DRAM_LPDDR3,
 			INTEL_DRAM_LPDDR4
 		} type;
+		u8 num_qgv_points;
 	} dram_info;
 
 	struct intel_bw_info {
@@ -1170,9 +1158,6 @@ struct drm_i915_private {
 		struct i915_gem_contexts {
 			spinlock_t lock; /* locks list */
 			struct list_head list;
-
-			struct llist_head free_list;
-			struct work_struct free_work;
 		} contexts;
 
 		/*
@@ -1186,6 +1171,8 @@ struct drm_i915_private {
 		struct file *mmap_singleton;
 	} gem;
 
+	u8 framestart_delay;
+
 	u8 pch_ssc_use;
 
 	/* For i915gm/i945gm vblank irq workaround */
@@ -1347,7 +1334,7 @@ intel_subplatform(const struct intel_runtime_info *info, enum intel_platform p)
 {
 	const unsigned int pi = __platform_mask_index(info, p);
 
-	return info->platform_mask[pi] & INTEL_SUBPLATFORM_BITS;
+	return info->platform_mask[pi] & ((1 << INTEL_SUBPLATFORM_BITS) - 1);
 }
 
 static __always_inline bool
@@ -1570,16 +1557,30 @@ enum {
 	TGL_REVID_D0,
 };
 
-extern const struct i915_rev_steppings tgl_uy_revids[];
-extern const struct i915_rev_steppings tgl_revids[];
+#define TGL_UY_REVIDS_SIZE	4
+#define TGL_REVIDS_SIZE		2
+
+extern const struct i915_rev_steppings tgl_uy_revids[TGL_UY_REVIDS_SIZE];
+extern const struct i915_rev_steppings tgl_revids[TGL_REVIDS_SIZE];
 
 static inline const struct i915_rev_steppings *
 tgl_revids_get(struct drm_i915_private *dev_priv)
 {
-	if (IS_TGL_U(dev_priv) || IS_TGL_Y(dev_priv))
-		return &tgl_uy_revids[INTEL_REVID(dev_priv)];
-	else
-		return &tgl_revids[INTEL_REVID(dev_priv)];
+	u8 revid = INTEL_REVID(dev_priv);
+	u8 size;
+	const struct i915_rev_steppings *tgl_revid_tbl;
+
+	if (IS_TGL_U(dev_priv) || IS_TGL_Y(dev_priv)) {
+		tgl_revid_tbl = tgl_uy_revids;
+		size = ARRAY_SIZE(tgl_uy_revids);
+	} else {
+		tgl_revid_tbl = tgl_revids;
+		size = ARRAY_SIZE(tgl_revids);
+	}
+
+	revid = min_t(u8, revid, size - 1);
+
+	return &tgl_revid_tbl[revid];
 }
 
 #define IS_TGL_DISP_REVID(p, since, until) \
@@ -1589,14 +1590,14 @@ tgl_revids_get(struct drm_i915_private *dev_priv)
 
 #define IS_TGL_UY_GT_REVID(p, since, until) \
 	((IS_TGL_U(p) || IS_TGL_Y(p)) && \
-	 tgl_uy_revids[INTEL_REVID(p)].gt_stepping >= (since) && \
-	 tgl_uy_revids[INTEL_REVID(p)].gt_stepping <= (until))
+	 tgl_revids_get(p)->gt_stepping >= (since) && \
+	 tgl_revids_get(p)->gt_stepping <= (until))
 
 #define IS_TGL_GT_REVID(p, since, until) \
 	(IS_TIGERLAKE(p) && \
 	 !(IS_TGL_U(p) || IS_TGL_Y(p)) && \
-	 tgl_revids[INTEL_REVID(p)].gt_stepping >= (since) && \
-	 tgl_revids[INTEL_REVID(p)].gt_stepping <= (until))
+	 tgl_revids_get(p)->gt_stepping >= (since) && \
+	 tgl_revids_get(p)->gt_stepping <= (until))
 
 #define RKL_REVID_A0		0x0
 #define RKL_REVID_B0		0x1
@@ -1647,8 +1648,6 @@ tgl_revids_get(struct drm_i915_private *dev_priv)
 		(INTEL_INFO(dev_priv)->has_logical_ring_contexts)
 #define HAS_LOGICAL_RING_ELSQ(dev_priv) \
 		(INTEL_INFO(dev_priv)->has_logical_ring_elsq)
-#define HAS_LOGICAL_RING_PREEMPTION(dev_priv) \
-		(INTEL_INFO(dev_priv)->has_logical_ring_preemption)
 
 #define HAS_MASTER_UNIT_IRQ(dev_priv) (INTEL_INFO(dev_priv)->has_master_unit_irq)
 
@@ -1750,10 +1749,17 @@ tgl_revids_get(struct drm_i915_private *dev_priv)
 
 #define HAS_DISPLAY(dev_priv) (INTEL_INFO(dev_priv)->pipe_mask != 0)
 
+#define HAS_VRR(i915)	(INTEL_GEN(i915) >= 12)
+
 /* Only valid when HAS_DISPLAY() is true */
 #define INTEL_DISPLAY_ENABLED(dev_priv) \
 	(drm_WARN_ON(&(dev_priv)->drm, !HAS_DISPLAY(dev_priv)), !(dev_priv)->params.disable_display)
 
+static inline bool run_as_guest(void)
+{
+	return !hypervisor_is_type(X86_HYPER_NATIVE);
+}
+
 static inline bool intel_vtd_active(void)
 {
 #ifdef CONFIG_INTEL_IOMMU
@@ -1762,7 +1768,7 @@ static inline bool intel_vtd_active(void)
 #endif
 
 	/* Running as a guest, we assume the host is enforcing VT'd */
-	return !hypervisor_is_type(X86_HYPER_NATIVE);
+	return run_as_guest();
 }
 
 static inline bool intel_scanout_needs_vtd_wa(struct drm_i915_private *dev_priv)
@@ -1794,8 +1800,6 @@ int i915_gem_init_userptr(struct drm_i915_private *dev_priv);
 void i915_gem_cleanup_userptr(struct drm_i915_private *dev_priv);
 void i915_gem_init_early(struct drm_i915_private *dev_priv);
 void i915_gem_cleanup_early(struct drm_i915_private *dev_priv);
-int i915_gem_freeze(struct drm_i915_private *dev_priv);
-int i915_gem_freeze_late(struct drm_i915_private *dev_priv);
 
 struct intel_memory_region *i915_gem_shmem_setup(struct drm_i915_private *i915);
 
@@ -1948,7 +1952,7 @@ const char *i915_cache_level_str(struct drm_i915_private *i915, int type);
 
 /* i915_cmd_parser.c */
 int i915_cmd_parser_get_version(struct drm_i915_private *dev_priv);
-void intel_engine_init_cmd_parser(struct intel_engine_cs *engine);
+int intel_engine_init_cmd_parser(struct intel_engine_cs *engine);
 void intel_engine_cleanup_cmd_parser(struct intel_engine_cs *engine);
 int intel_engine_cmd_parser(struct intel_engine_cs *engine,
 			    struct i915_vma *batch,
@@ -1990,16 +1994,4 @@ i915_coherent_map_type(struct drm_i915_private *i915)
 	return HAS_LLC(i915) ? I915_MAP_WB : I915_MAP_WC;
 }
 
-static inline u64 i915_cs_timestamp_ns_to_ticks(struct drm_i915_private *i915, u64 val)
-{
-	return DIV_ROUND_UP_ULL(val * RUNTIME_INFO(i915)->cs_timestamp_frequency_hz,
-				1000000000);
-}
-
-static inline u64 i915_cs_timestamp_ticks_to_ns(struct drm_i915_private *i915, u64 val)
-{
-	return div_u64(val * 1000000000,
-		       RUNTIME_INFO(i915)->cs_timestamp_frequency_hz);
-}
-
 #endif
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 58276694c848..aa4490934469 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -180,108 +180,6 @@ try_again:
 }
 
 static int
-i915_gem_create(struct drm_file *file,
-		struct intel_memory_region *mr,
-		u64 *size_p,
-		u32 *handle_p)
-{
-	struct drm_i915_gem_object *obj;
-	u32 handle;
-	u64 size;
-	int ret;
-
-	GEM_BUG_ON(!is_power_of_2(mr->min_page_size));
-	size = round_up(*size_p, mr->min_page_size);
-	if (size == 0)
-		return -EINVAL;
-
-	/* For most of the ABI (e.g. mmap) we think in system pages */
-	GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE));
-
-	/* Allocate the new object */
-	obj = i915_gem_object_create_region(mr, size, 0);
-	if (IS_ERR(obj))
-		return PTR_ERR(obj);
-
-	ret = drm_gem_handle_create(file, &obj->base, &handle);
-	/* drop reference from allocate - handle holds it now */
-	i915_gem_object_put(obj);
-	if (ret)
-		return ret;
-
-	*handle_p = handle;
-	*size_p = size;
-	return 0;
-}
-
-int
-i915_gem_dumb_create(struct drm_file *file,
-		     struct drm_device *dev,
-		     struct drm_mode_create_dumb *args)
-{
-	enum intel_memory_type mem_type;
-	int cpp = DIV_ROUND_UP(args->bpp, 8);
-	u32 format;
-
-	switch (cpp) {
-	case 1:
-		format = DRM_FORMAT_C8;
-		break;
-	case 2:
-		format = DRM_FORMAT_RGB565;
-		break;
-	case 4:
-		format = DRM_FORMAT_XRGB8888;
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	/* have to work out size/pitch and return them */
-	args->pitch = ALIGN(args->width * cpp, 64);
-
-	/* align stride to page size so that we can remap */
-	if (args->pitch > intel_plane_fb_max_stride(to_i915(dev), format,
-						    DRM_FORMAT_MOD_LINEAR))
-		args->pitch = ALIGN(args->pitch, 4096);
-
-	if (args->pitch < args->width)
-		return -EINVAL;
-
-	args->size = mul_u32_u32(args->pitch, args->height);
-
-	mem_type = INTEL_MEMORY_SYSTEM;
-	if (HAS_LMEM(to_i915(dev)))
-		mem_type = INTEL_MEMORY_LOCAL;
-
-	return i915_gem_create(file,
-			       intel_memory_region_by_type(to_i915(dev),
-							   mem_type),
-			       &args->size, &args->handle);
-}
-
-/**
- * Creates a new mm object and returns a handle to it.
- * @dev: drm device pointer
- * @data: ioctl data blob
- * @file: drm file pointer
- */
-int
-i915_gem_create_ioctl(struct drm_device *dev, void *data,
-		      struct drm_file *file)
-{
-	struct drm_i915_private *i915 = to_i915(dev);
-	struct drm_i915_gem_create *args = data;
-
-	i915_gem_flush_free_objects(i915);
-
-	return i915_gem_create(file,
-			       intel_memory_region_by_type(i915,
-							   INTEL_MEMORY_SYSTEM),
-			       &args->size, &args->handle);
-}
-
-static int
 shmem_pread(struct page *page, int offset, int len, char __user *user_data,
 	    bool needs_clflush)
 {
@@ -1059,14 +957,14 @@ i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
 	    i915_gem_object_is_tiled(obj) &&
 	    i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
 		if (obj->mm.madv == I915_MADV_WILLNEED) {
-			GEM_BUG_ON(!obj->mm.quirked);
-			__i915_gem_object_unpin_pages(obj);
-			obj->mm.quirked = false;
+			GEM_BUG_ON(!i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_clear_tiling_quirk(obj);
+			i915_gem_object_make_shrinkable(obj);
 		}
 		if (args->madv == I915_MADV_WILLNEED) {
-			GEM_BUG_ON(obj->mm.quirked);
-			__i915_gem_object_pin_pages(obj);
-			obj->mm.quirked = true;
+			GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_make_unshrinkable(obj);
+			i915_gem_object_set_tiling_quirk(obj);
 		}
 	}
 
@@ -1207,8 +1105,6 @@ void i915_gem_driver_remove(struct drm_i915_private *dev_priv)
 
 void i915_gem_driver_release(struct drm_i915_private *dev_priv)
 {
-	i915_gem_driver_release__contexts(dev_priv);
-
 	intel_gt_driver_release(&dev_priv->gt);
 
 	intel_wa_list_free(&dev_priv->gt_wa_list);
@@ -1249,53 +1145,6 @@ void i915_gem_cleanup_early(struct drm_i915_private *dev_priv)
 	drm_WARN_ON(&dev_priv->drm, dev_priv->mm.shrink_count);
 }
 
-int i915_gem_freeze(struct drm_i915_private *dev_priv)
-{
-	/* Discard all purgeable objects, let userspace recover those as
-	 * required after resuming.
-	 */
-	i915_gem_shrink_all(dev_priv);
-
-	return 0;
-}
-
-int i915_gem_freeze_late(struct drm_i915_private *i915)
-{
-	struct drm_i915_gem_object *obj;
-	intel_wakeref_t wakeref;
-
-	/*
-	 * Called just before we write the hibernation image.
-	 *
-	 * We need to update the domain tracking to reflect that the CPU
-	 * will be accessing all the pages to create and restore from the
-	 * hibernation, and so upon restoration those pages will be in the
-	 * CPU domain.
-	 *
-	 * To make sure the hibernation image contains the latest state,
-	 * we update that state just before writing out the image.
-	 *
-	 * To try and reduce the hibernation image, we manually shrink
-	 * the objects as well, see i915_gem_freeze()
-	 */
-
-	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
-
-	i915_gem_shrink(i915, -1UL, NULL, ~0);
-	i915_gem_drain_freed_objects(i915);
-
-	list_for_each_entry(obj, &i915->mm.shrink_list, mm.link) {
-		i915_gem_object_lock(obj, NULL);
-		drm_WARN_ON(&i915->drm,
-			    i915_gem_object_set_to_cpu_domain(obj, true));
-		i915_gem_object_unlock(obj);
-	}
-
-	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
-
-	return 0;
-}
-
 int i915_gem_open(struct drm_i915_private *i915, struct drm_file *file)
 {
 	struct drm_i915_file_private *file_priv;
diff --git a/drivers/gpu/drm/i915/i915_gem.h b/drivers/gpu/drm/i915/i915_gem.h
index a4cad3f154ca..e622aee6e4be 100644
--- a/drivers/gpu/drm/i915/i915_gem.h
+++ b/drivers/gpu/drm/i915/i915_gem.h
@@ -38,11 +38,18 @@ struct drm_i915_private;
 
 #define GEM_SHOW_DEBUG() drm_debug_enabled(DRM_UT_DRIVER)
 
+#ifdef CONFIG_DRM_I915_DEBUG_GEM_ONCE
+#define __GEM_BUG(cond) BUG()
+#else
+#define __GEM_BUG(cond) \
+	WARN(1, "%s:%d GEM_BUG_ON(%s)\n", __func__, __LINE__, __stringify(cond))
+#endif
+
 #define GEM_BUG_ON(condition) do { if (unlikely((condition))) {	\
 		GEM_TRACE_ERR("%s:%d GEM_BUG_ON(%s)\n", \
 			      __func__, __LINE__, __stringify(condition)); \
 		GEM_TRACE_DUMP(); \
-		BUG(); \
+		__GEM_BUG(condition); \
 		} \
 	} while(0)
 #define GEM_WARN_ON(expr) WARN_ON(expr)
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index e1a66c8245b8..4d2d59a9942b 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -61,6 +61,17 @@ mark_free(struct drm_mm_scan *scan,
 	return drm_mm_scan_add_block(scan, &vma->node);
 }
 
+static bool defer_evict(struct i915_vma *vma)
+{
+	if (i915_vma_is_active(vma))
+		return true;
+
+	if (i915_vma_is_scanout(vma))
+		return true;
+
+	return false;
+}
+
 /**
  * i915_gem_evict_something - Evict vmas to make room for binding a new one
  * @vm: address space to evict from
@@ -150,7 +161,7 @@ search_again:
 		 * To notice when we complete one full cycle, we record the
 		 * first active element seen, before moving it to the tail.
 		 */
-		if (active != ERR_PTR(-EAGAIN) && i915_vma_is_active(vma)) {
+		if (active != ERR_PTR(-EAGAIN) && defer_evict(vma)) {
 			if (!active)
 				active = vma;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index c5ee1567f3d1..3ee2f682eff6 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -55,22 +55,17 @@ int i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj,
 void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj,
 			       struct sg_table *pages)
 {
-	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
-	struct device *kdev = &dev_priv->drm.pdev->dev;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-
-	if (unlikely(ggtt->do_idle_maps)) {
-		/* XXX This does not prevent more requests being submitted! */
-		if (intel_gt_retire_requests_timeout(ggtt->vm.gt,
-						     -MAX_SCHEDULE_TIMEOUT)) {
-			drm_err(&dev_priv->drm,
-				"Failed to wait for idle; VT'd may hang.\n");
-			/* Wait a bit, in hopes it avoids the hang */
-			udelay(10);
-		}
-	}
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = &i915->ggtt;
+
+	/* XXX This does not prevent more requests being submitted! */
+	if (unlikely(ggtt->do_idle_maps))
+		/* Wait a bit, in the hope it avoids the hang */
+		usleep_range(100, 250);
 
-	dma_unmap_sg(kdev, pages->sgl, pages->nents, PCI_DMA_BIDIRECTIONAL);
+	dma_unmap_sg(&i915->drm.pdev->dev,
+		     pages->sgl, pages->nents,
+		     PCI_DMA_BIDIRECTIONAL);
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/i915_getparam.c b/drivers/gpu/drm/i915/i915_getparam.c
index f96032c60a12..75c3bfc2486e 100644
--- a/drivers/gpu/drm/i915/i915_getparam.c
+++ b/drivers/gpu/drm/i915/i915_getparam.c
@@ -154,7 +154,7 @@ int i915_getparam_ioctl(struct drm_device *dev, void *data,
 			return -ENODEV;
 		break;
 	case I915_PARAM_CS_TIMESTAMP_FREQUENCY:
-		value = RUNTIME_INFO(i915)->cs_timestamp_frequency_hz;
+		value = i915->gt.clock_frequency;
 		break;
 	case I915_PARAM_MMAP_GTT_COHERENT:
 		value = INTEL_INFO(i915)->has_coherent_ggtt;
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index d8cac4c5881f..f962693404b7 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -485,7 +485,7 @@ static void error_print_context(struct drm_i915_error_state_buf *m,
 				const char *header,
 				const struct i915_gem_context_coredump *ctx)
 {
-	const u32 period = RUNTIME_INFO(m->i915)->cs_timestamp_period_ns;
+	const u32 period = m->i915->gt.clock_period_ns;
 
 	err_printf(m, "%s%s[%d] prio %d, guilty %d active %d, runtime total %lluns, avg %lluns\n",
 		   header, ctx->comm, ctx->pid, ctx->sched_attr.priority,
@@ -1051,7 +1051,9 @@ i915_vma_coredump_create(const struct intel_gt *gt,
 		for_each_sgt_daddr(dma, iter, vma->pages) {
 			void __iomem *s;
 
-			s = io_mapping_map_wc(&mem->iomap, dma, PAGE_SIZE);
+			s = io_mapping_map_wc(&mem->iomap,
+					      dma - mem->region.start,
+					      PAGE_SIZE);
 			ret = compress_page(compress,
 					    (void __force *)s, dst,
 					    true);
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index dd1971040bbc..1a701367a718 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -718,25 +718,15 @@ u32 g4x_get_vblank_counter(struct drm_crtc *crtc)
 	return intel_uncore_read(&dev_priv->uncore, PIPE_FRMCOUNT_G4X(pipe));
 }
 
-/*
- * On certain encoders on certain platforms, pipe
- * scanline register will not work to get the scanline,
- * since the timings are driven from the PORT or issues
- * with scanline register updates.
- * This function will use Framestamp and current
- * timestamp registers to calculate the scanline.
- */
-static u32 __intel_get_crtc_scanline_from_timestamp(struct intel_crtc *crtc)
+static u32 intel_crtc_scanlines_since_frame_timestamp(struct intel_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct drm_vblank_crtc *vblank =
 		&crtc->base.dev->vblank[drm_crtc_index(&crtc->base)];
 	const struct drm_display_mode *mode = &vblank->hwmode;
-	u32 vblank_start = mode->crtc_vblank_start;
-	u32 vtotal = mode->crtc_vtotal;
 	u32 htotal = mode->crtc_htotal;
 	u32 clock = mode->crtc_clock;
-	u32 scanline, scan_prev_time, scan_curr_time, scan_post_time;
+	u32 scan_prev_time, scan_curr_time, scan_post_time;
 
 	/*
 	 * To avoid the race condition where we might cross into the
@@ -763,8 +753,28 @@ static u32 __intel_get_crtc_scanline_from_timestamp(struct intel_crtc *crtc)
 						  PIPE_FRMTMSTMP(crtc->pipe));
 	} while (scan_post_time != scan_prev_time);
 
-	scanline = div_u64(mul_u32_u32(scan_curr_time - scan_prev_time,
-					clock), 1000 * htotal);
+	return div_u64(mul_u32_u32(scan_curr_time - scan_prev_time,
+				   clock), 1000 * htotal);
+}
+
+/*
+ * On certain encoders on certain platforms, pipe
+ * scanline register will not work to get the scanline,
+ * since the timings are driven from the PORT or issues
+ * with scanline register updates.
+ * This function will use Framestamp and current
+ * timestamp registers to calculate the scanline.
+ */
+static u32 __intel_get_crtc_scanline_from_timestamp(struct intel_crtc *crtc)
+{
+	struct drm_vblank_crtc *vblank =
+		&crtc->base.dev->vblank[drm_crtc_index(&crtc->base)];
+	const struct drm_display_mode *mode = &vblank->hwmode;
+	u32 vblank_start = mode->crtc_vblank_start;
+	u32 vtotal = mode->crtc_vtotal;
+	u32 scanline;
+
+	scanline = intel_crtc_scanlines_since_frame_timestamp(crtc);
 	scanline = min(scanline, vtotal - 1);
 	scanline = (scanline + vblank_start) % vtotal;
 
@@ -883,7 +893,20 @@ static bool i915_get_crtc_scanoutpos(struct drm_crtc *_crtc,
 	if (stime)
 		*stime = ktime_get();
 
-	if (use_scanline_counter) {
+	if (crtc->mode_flags & I915_MODE_FLAG_VRR) {
+		int scanlines = intel_crtc_scanlines_since_frame_timestamp(crtc);
+
+		position = __intel_get_crtc_scanline(crtc);
+
+		/*
+		 * Already exiting vblank? If so, shift our position
+		 * so it looks like we're already apporaching the full
+		 * vblank end. This should make the generated timestamp
+		 * more or less match when the active portion will start.
+		 */
+		if (position >= vbl_start && scanlines < position)
+			position = min(crtc->vmax_vblank_start + scanlines, vtotal - 1);
+	} else if (use_scanline_counter) {
 		/* No obvious pixelcount register. Only query vertical
 		 * scanout position from Display scan line register.
 		 */
@@ -1517,6 +1540,9 @@ static void valleyview_pipestat_irq_handler(struct drm_i915_private *dev_priv,
 		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
 			intel_handle_vblank(dev_priv, pipe);
 
+		if (pipe_stats[pipe] & PLANE_FLIP_DONE_INT_STATUS_VLV)
+			flip_done_handler(dev_priv, pipe);
+
 		if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
 			i9xx_pipe_crc_irq_handler(dev_priv, pipe);
 
@@ -2029,6 +2055,9 @@ static void ilk_display_irq_handler(struct drm_i915_private *dev_priv,
 		if (de_iir & DE_PIPE_VBLANK(pipe))
 			intel_handle_vblank(dev_priv, pipe);
 
+		if (de_iir & DE_PLANE_FLIP_DONE(pipe))
+			flip_done_handler(dev_priv, pipe);
+
 		if (de_iir & DE_PIPE_FIFO_UNDERRUN(pipe))
 			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
 
@@ -2079,8 +2108,11 @@ static void ivb_display_irq_handler(struct drm_i915_private *dev_priv,
 		intel_opregion_asle_intr(dev_priv);
 
 	for_each_pipe(dev_priv, pipe) {
-		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)))
+		if (de_iir & DE_PIPE_VBLANK_IVB(pipe))
 			intel_handle_vblank(dev_priv, pipe);
+
+		if (de_iir & DE_PLANE_FLIP_DONE_IVB(pipe))
+			flip_done_handler(dev_priv, pipe);
 	}
 
 	/* check event from PCH */
@@ -2357,6 +2389,14 @@ static void gen11_dsi_te_interrupt_handler(struct drm_i915_private *dev_priv,
 	intel_uncore_write(&dev_priv->uncore, DSI_INTR_IDENT_REG(port), tmp);
 }
 
+static u32 gen8_de_pipe_flip_done_mask(struct drm_i915_private *i915)
+{
+	if (INTEL_GEN(i915) >= 9)
+		return GEN9_PIPE_PLANE1_FLIP_DONE;
+	else
+		return GEN8_PIPE_PRIMARY_FLIP_DONE;
+}
+
 static irqreturn_t
 gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
 {
@@ -2459,7 +2499,7 @@ gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
 		if (iir & GEN8_PIPE_VBLANK)
 			intel_handle_vblank(dev_priv, pipe);
 
-		if (iir & GEN9_PIPE_PLANE1_FLIP_DONE)
+		if (iir & gen8_de_pipe_flip_done_mask(dev_priv))
 			flip_done_handler(dev_priv, pipe);
 
 		if (iir & GEN8_PIPE_CDCLK_CRC_DONE)
@@ -2822,19 +2862,6 @@ int bdw_enable_vblank(struct drm_crtc *crtc)
 	return 0;
 }
 
-void skl_enable_flip_done(struct intel_crtc *crtc)
-{
-	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
-	enum pipe pipe = crtc->pipe;
-	unsigned long irqflags;
-
-	spin_lock_irqsave(&i915->irq_lock, irqflags);
-
-	bdw_enable_pipe_irq(i915, pipe, GEN9_PIPE_PLANE1_FLIP_DONE);
-
-	spin_unlock_irqrestore(&i915->irq_lock, irqflags);
-}
-
 /* Called from drm generic code, passed 'crtc' which
  * we use as a pipe index
  */
@@ -2899,19 +2926,6 @@ void bdw_disable_vblank(struct drm_crtc *crtc)
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 }
 
-void skl_disable_flip_done(struct intel_crtc *crtc)
-{
-	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
-	enum pipe pipe = crtc->pipe;
-	unsigned long irqflags;
-
-	spin_lock_irqsave(&i915->irq_lock, irqflags);
-
-	bdw_disable_pipe_irq(i915, pipe, GEN9_PIPE_PLANE1_FLIP_DONE);
-
-	spin_unlock_irqrestore(&i915->irq_lock, irqflags);
-}
-
 static void ibx_irq_reset(struct drm_i915_private *dev_priv)
 {
 	struct intel_uncore *uncore = &dev_priv->uncore;
@@ -3104,13 +3118,10 @@ void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv,
 				     u8 pipe_mask)
 {
 	struct intel_uncore *uncore = &dev_priv->uncore;
-
-	u32 extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN;
+	u32 extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN |
+		gen8_de_pipe_flip_done_mask(dev_priv);
 	enum pipe pipe;
 
-	if (INTEL_GEN(dev_priv) >= 9)
-		extra_ier |= GEN9_PIPE_PLANE1_FLIP_DONE;
-
 	spin_lock_irq(&dev_priv->irq_lock);
 
 	if (!intel_irqs_enabled(dev_priv)) {
@@ -3585,6 +3596,9 @@ static void ilk_irq_postinstall(struct drm_i915_private *dev_priv)
 				DE_PCH_EVENT_IVB | DE_AUX_CHANNEL_A_IVB);
 		extra_mask = (DE_PIPEC_VBLANK_IVB | DE_PIPEB_VBLANK_IVB |
 			      DE_PIPEA_VBLANK_IVB | DE_ERR_INT_IVB |
+			      DE_PLANE_FLIP_DONE_IVB(PLANE_C) |
+			      DE_PLANE_FLIP_DONE_IVB(PLANE_B) |
+			      DE_PLANE_FLIP_DONE_IVB(PLANE_A) |
 			      DE_DP_A_HOTPLUG_IVB);
 	} else {
 		display_mask = (DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT |
@@ -3592,6 +3606,8 @@ static void ilk_irq_postinstall(struct drm_i915_private *dev_priv)
 				DE_PIPEA_CRC_DONE | DE_POISON);
 		extra_mask = (DE_PIPEA_VBLANK | DE_PIPEB_VBLANK |
 			      DE_PIPEB_FIFO_UNDERRUN | DE_PIPEA_FIFO_UNDERRUN |
+			      DE_PLANE_FLIP_DONE(PLANE_A) |
+			      DE_PLANE_FLIP_DONE(PLANE_B) |
 			      DE_DP_A_HOTPLUG);
 	}
 
@@ -3682,11 +3698,9 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 			de_port_masked |= DSI0_TE | DSI1_TE;
 	}
 
-	de_pipe_enables = de_pipe_masked | GEN8_PIPE_VBLANK |
-					   GEN8_PIPE_FIFO_UNDERRUN;
-
-	if (INTEL_GEN(dev_priv) >= 9)
-		de_pipe_enables |= GEN9_PIPE_PLANE1_FLIP_DONE;
+	de_pipe_enables = de_pipe_masked |
+		GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN |
+		gen8_de_pipe_flip_done_mask(dev_priv);
 
 	de_port_enables = de_port_masked;
 	if (IS_GEN9_LP(dev_priv))
diff --git a/drivers/gpu/drm/i915/i915_irq.h b/drivers/gpu/drm/i915/i915_irq.h
index 2efe609519ca..25f25cd95818 100644
--- a/drivers/gpu/drm/i915/i915_irq.h
+++ b/drivers/gpu/drm/i915/i915_irq.h
@@ -118,9 +118,6 @@ void i965_disable_vblank(struct drm_crtc *crtc);
 void ilk_disable_vblank(struct drm_crtc *crtc);
 void bdw_disable_vblank(struct drm_crtc *crtc);
 
-void skl_enable_flip_done(struct intel_crtc *crtc);
-void skl_disable_flip_done(struct intel_crtc *crtc);
-
 void gen2_irq_reset(struct intel_uncore *uncore);
 void gen3_irq_reset(struct intel_uncore *uncore, i915_reg_t imr,
 		    i915_reg_t iir, i915_reg_t ier);
diff --git a/drivers/gpu/drm/i915/i915_mitigations.c b/drivers/gpu/drm/i915/i915_mitigations.c
new file mode 100644
index 000000000000..84f12598d145
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_mitigations.c
@@ -0,0 +1,146 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/kernel.h>
+#include <linux/moduleparam.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+
+#include "i915_drv.h"
+#include "i915_mitigations.h"
+
+static unsigned long mitigations __read_mostly = ~0UL;
+
+enum {
+	CLEAR_RESIDUALS = 0,
+};
+
+static const char * const names[] = {
+	[CLEAR_RESIDUALS] = "residuals",
+};
+
+bool i915_mitigate_clear_residuals(void)
+{
+	return READ_ONCE(mitigations) & BIT(CLEAR_RESIDUALS);
+}
+
+static int mitigations_set(const char *val, const struct kernel_param *kp)
+{
+	unsigned long new = ~0UL;
+	char *str, *sep, *tok;
+	bool first = true;
+	int err = 0;
+
+	BUILD_BUG_ON(ARRAY_SIZE(names) >= BITS_PER_TYPE(mitigations));
+
+	str = kstrdup(val, GFP_KERNEL);
+	if (!str)
+		return -ENOMEM;
+
+	for (sep = str; (tok = strsep(&sep, ","));) {
+		bool enable = true;
+		int i;
+
+		/* Be tolerant of leading/trailing whitespace */
+		tok = strim(tok);
+
+		if (first) {
+			first = false;
+
+			if (!strcmp(tok, "auto"))
+				continue;
+
+			new = 0;
+			if (!strcmp(tok, "off"))
+				continue;
+		}
+
+		if (*tok == '!') {
+			enable = !enable;
+			tok++;
+		}
+
+		if (!strncmp(tok, "no", 2)) {
+			enable = !enable;
+			tok += 2;
+		}
+
+		if (*tok == '\0')
+			continue;
+
+		for (i = 0; i < ARRAY_SIZE(names); i++) {
+			if (!strcmp(tok, names[i])) {
+				if (enable)
+					new |= BIT(i);
+				else
+					new &= ~BIT(i);
+				break;
+			}
+		}
+		if (i == ARRAY_SIZE(names)) {
+			pr_err("Bad \"%s.mitigations=%s\", '%s' is unknown\n",
+			       DRIVER_NAME, val, tok);
+			err = -EINVAL;
+			break;
+		}
+	}
+	kfree(str);
+	if (err)
+		return err;
+
+	WRITE_ONCE(mitigations, new);
+	return 0;
+}
+
+static int mitigations_get(char *buffer, const struct kernel_param *kp)
+{
+	unsigned long local = READ_ONCE(mitigations);
+	int count, i;
+	bool enable;
+
+	if (!local)
+		return scnprintf(buffer, PAGE_SIZE, "%s\n", "off");
+
+	if (local & BIT(BITS_PER_LONG - 1)) {
+		count = scnprintf(buffer, PAGE_SIZE, "%s,", "auto");
+		enable = false;
+	} else {
+		enable = true;
+		count = 0;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(names); i++) {
+		if ((local & BIT(i)) != enable)
+			continue;
+
+		count += scnprintf(buffer + count, PAGE_SIZE - count,
+				   "%s%s,", enable ? "" : "!", names[i]);
+	}
+
+	buffer[count - 1] = '\n';
+	return count;
+}
+
+static const struct kernel_param_ops ops = {
+	.set = mitigations_set,
+	.get = mitigations_get,
+};
+
+module_param_cb_unsafe(mitigations, &ops, NULL, 0600);
+MODULE_PARM_DESC(mitigations,
+"Selectively enable security mitigations for all Intel® GPUs in the system.\n"
+"\n"
+"  auto -- enables all mitigations required for the platform [default]\n"
+"  off  -- disables all mitigations\n"
+"\n"
+"Individual mitigations can be enabled by passing a comma-separated string,\n"
+"e.g. mitigations=residuals to enable only clearing residuals or\n"
+"mitigations=auto,noresiduals to disable only the clear residual mitigation.\n"
+"Either '!' or 'no' may be used to switch from enabling the mitigation to\n"
+"disabling it.\n"
+"\n"
+"Active mitigations for Ivybridge, Baytrail, Haswell:\n"
+"  residuals -- clear all thread-local registers between contexts"
+);
diff --git a/drivers/gpu/drm/i915/i915_mitigations.h b/drivers/gpu/drm/i915/i915_mitigations.h
new file mode 100644
index 000000000000..1359d8135287
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_mitigations.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_MITIGATIONS_H__
+#define __I915_MITIGATIONS_H__
+
+#include <linux/types.h>
+
+bool i915_mitigate_clear_residuals(void);
+
+#endif /* __I915_MITIGATIONS_H__ */
diff --git a/drivers/gpu/drm/i915/i915_mm.c b/drivers/gpu/drm/i915/i915_mm.c
index 43039dc8c607..666808cb3a32 100644
--- a/drivers/gpu/drm/i915/i915_mm.c
+++ b/drivers/gpu/drm/i915/i915_mm.c
@@ -62,7 +62,7 @@ static int remap_sg(pte_t *pte, unsigned long addr, void *data)
 {
 	struct remap_pfn *r = data;
 
-	if (GEM_WARN_ON(!r->sgt.pfn))
+	if (GEM_WARN_ON(!r->sgt.sgp))
 		return -EINVAL;
 
 	/* Special PTE are not associated with any struct page */
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index 7f139ea4a90b..6939634e56ed 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -185,7 +185,7 @@ i915_param_named_unsafe(inject_probe_failure, uint, 0400,
 
 i915_param_named(enable_dpcd_backlight, int, 0400,
 	"Enable support for DPCD backlight control"
-	"(-1=use per-VBT LFP backlight type setting [default], 0=disabled, 1=enabled)");
+	"(-1=use per-VBT LFP backlight type setting [default], 0=disabled, 1=enable, 2=force VESA interface, 3=force Intel interface)");
 
 #if IS_ENABLED(CONFIG_DRM_I915_GVT)
 i915_param_named(enable_gvt, bool, 0400,
diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h
index 330c03e2b4f7..f031966af5b7 100644
--- a/drivers/gpu/drm/i915/i915_params.h
+++ b/drivers/gpu/drm/i915/i915_params.h
@@ -32,6 +32,7 @@ struct drm_printer;
 
 #define ENABLE_GUC_SUBMISSION		BIT(0)
 #define ENABLE_GUC_LOAD_HUC		BIT(1)
+#define ENABLE_GUC_MASK			GENMASK(1, 0)
 
 /*
  * Invoke param, a function-like macro, for each i915 param, with arguments:
diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c
index 11fe790b1969..020b5f561f07 100644
--- a/drivers/gpu/drm/i915/i915_pci.c
+++ b/drivers/gpu/drm/i915/i915_pci.c
@@ -455,6 +455,7 @@ static const struct intel_device_info snb_m_gt2_info = {
 	.has_llc = 1, \
 	.has_rc6 = 1, \
 	.has_rc6p = 1, \
+	.has_reset_engine = true, \
 	.has_rps = true, \
 	.dma_mask_size = 40, \
 	.ppgtt_type = INTEL_PPGTT_ALIASING, \
@@ -513,6 +514,7 @@ static const struct intel_device_info vlv_info = {
 	.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B),
 	.has_runtime_pm = 1,
 	.has_rc6 = 1,
+	.has_reset_engine = true,
 	.has_rps = true,
 	.display.has_gmch = 1,
 	.display.has_hotplug = 1,
@@ -571,8 +573,7 @@ static const struct intel_device_info hsw_gt3_info = {
 	.dma_mask_size = 39, \
 	.ppgtt_type = INTEL_PPGTT_FULL, \
 	.ppgtt_size = 48, \
-	.has_64bit_reloc = 1, \
-	.has_reset_engine = 1
+	.has_64bit_reloc = 1
 
 #define BDW_PLATFORM \
 	GEN8_FEATURES, \
@@ -639,7 +640,6 @@ static const struct intel_device_info chv_info = {
 	GEN8_FEATURES, \
 	GEN(9), \
 	GEN9_DEFAULT_PAGE_SIZES, \
-	.has_logical_ring_preemption = 1, \
 	.display.has_csr = 1, \
 	.has_gt_uc = 1, \
 	.display.has_hdcp = 1, \
@@ -700,7 +700,6 @@ static const struct intel_device_info skl_gt4_info = {
 	.has_rps = true, \
 	.display.has_dp_mst = 1, \
 	.has_logical_ring_contexts = 1, \
-	.has_logical_ring_preemption = 1, \
 	.has_gt_uc = 1, \
 	.dma_mask_size = 39, \
 	.ppgtt_type = INTEL_PPGTT_FULL, \
diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
index 649c26518d26..112ba5f2ce90 100644
--- a/drivers/gpu/drm/i915/i915_perf.c
+++ b/drivers/gpu/drm/i915/i915_perf.c
@@ -198,8 +198,11 @@
 #include "gem/i915_gem_context.h"
 #include "gt/intel_engine_pm.h"
 #include "gt/intel_engine_user.h"
+#include "gt/intel_execlists_submission.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
-#include "gt/intel_lrc_reg.h"
+#include "gt/intel_gt_clock_utils.h"
+#include "gt/intel_lrc.h"
 #include "gt/intel_ring.h"
 
 #include "i915_drv.h"
@@ -1635,7 +1638,8 @@ static int alloc_noa_wait(struct i915_perf_stream *stream)
 	struct drm_i915_gem_object *bo;
 	struct i915_vma *vma;
 	const u64 delay_ticks = 0xffffffffffffffff -
-		i915_cs_timestamp_ns_to_ticks(i915, atomic64_read(&stream->perf->noa_programming_delay));
+		intel_gt_ns_to_clock_interval(stream->perf->i915->ggtt.vm.gt,
+					      atomic64_read(&stream->perf->noa_programming_delay));
 	const u32 base = stream->engine->mmio_base;
 #define CS_GPR(x) GEN8_RING_CS_GPR(base, x)
 	u32 *batch, *ts0, *cs, *jump;
@@ -3516,7 +3520,8 @@ err:
 
 static u64 oa_exponent_to_ns(struct i915_perf *perf, int exponent)
 {
-	return i915_cs_timestamp_ticks_to_ns(perf->i915, 2ULL << exponent);
+	return intel_gt_clock_interval_to_ns(perf->i915->ggtt.vm.gt,
+					     2ULL << exponent);
 }
 
 /**
@@ -4370,11 +4375,11 @@ void i915_perf_init(struct drm_i915_private *i915)
 	if (perf->ops.enable_metric_set) {
 		mutex_init(&perf->lock);
 
-		oa_sample_rate_hard_limit =
-			RUNTIME_INFO(i915)->cs_timestamp_frequency_hz / 2;
+		/* Choose a representative limit */
+		oa_sample_rate_hard_limit = i915->gt.clock_frequency / 2;
 
 		mutex_init(&perf->metrics_lock);
-		idr_init(&perf->metrics_idr);
+		idr_init_base(&perf->metrics_idr, 1);
 
 		/* We set up some ratelimit state to potentially throttle any
 		 * _NOTES about spurious, invalid OA reports which we don't
diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c
index d76685ce0399..2b88c0baa1bf 100644
--- a/drivers/gpu/drm/i915/i915_pmu.c
+++ b/drivers/gpu/drm/i915/i915_pmu.c
@@ -26,8 +26,6 @@
 	 BIT(I915_SAMPLE_WAIT) | \
 	 BIT(I915_SAMPLE_SEMA))
 
-#define ENGINE_SAMPLE_BITS (1 << I915_PMU_SAMPLE_BITS)
-
 static cpumask_t i915_pmu_cpumask;
 static unsigned int i915_pmu_target_cpu = -1;
 
@@ -56,17 +54,42 @@ static bool is_engine_config(u64 config)
 	return config < __I915_PMU_OTHER(0);
 }
 
-static unsigned int config_enabled_bit(u64 config)
+static unsigned int other_bit(const u64 config)
+{
+	unsigned int val;
+
+	switch (config) {
+	case I915_PMU_ACTUAL_FREQUENCY:
+		val =  __I915_PMU_ACTUAL_FREQUENCY_ENABLED;
+		break;
+	case I915_PMU_REQUESTED_FREQUENCY:
+		val = __I915_PMU_REQUESTED_FREQUENCY_ENABLED;
+		break;
+	case I915_PMU_RC6_RESIDENCY:
+		val = __I915_PMU_RC6_RESIDENCY_ENABLED;
+		break;
+	default:
+		/*
+		 * Events that do not require sampling, or tracking state
+		 * transitions between enabled and disabled can be ignored.
+		 */
+		return -1;
+	}
+
+	return I915_ENGINE_SAMPLE_COUNT + val;
+}
+
+static unsigned int config_bit(const u64 config)
 {
 	if (is_engine_config(config))
 		return engine_config_sample(config);
 	else
-		return ENGINE_SAMPLE_BITS + (config - __I915_PMU_OTHER(0));
+		return other_bit(config);
 }
 
-static u64 config_enabled_mask(u64 config)
+static u64 config_mask(u64 config)
 {
-	return BIT_ULL(config_enabled_bit(config));
+	return BIT_ULL(config_bit(config));
 }
 
 static bool is_engine_event(struct perf_event *event)
@@ -74,15 +97,15 @@ static bool is_engine_event(struct perf_event *event)
 	return is_engine_config(event->attr.config);
 }
 
-static unsigned int event_enabled_bit(struct perf_event *event)
+static unsigned int event_bit(struct perf_event *event)
 {
-	return config_enabled_bit(event->attr.config);
+	return config_bit(event->attr.config);
 }
 
 static bool pmu_needs_timer(struct i915_pmu *pmu, bool gpu_active)
 {
 	struct drm_i915_private *i915 = container_of(pmu, typeof(*i915), pmu);
-	u64 enable;
+	u32 enable;
 
 	/*
 	 * Only some counters need the sampling timer.
@@ -95,8 +118,8 @@ static bool pmu_needs_timer(struct i915_pmu *pmu, bool gpu_active)
 	 * Mask out all the ones which do not need the timer, or in
 	 * other words keep all the ones that could need the timer.
 	 */
-	enable &= config_enabled_mask(I915_PMU_ACTUAL_FREQUENCY) |
-		  config_enabled_mask(I915_PMU_REQUESTED_FREQUENCY) |
+	enable &= config_mask(I915_PMU_ACTUAL_FREQUENCY) |
+		  config_mask(I915_PMU_REQUESTED_FREQUENCY) |
 		  ENGINE_SAMPLE_MASK;
 
 	/*
@@ -137,11 +160,9 @@ static u64 __get_rc6(struct intel_gt *gt)
 	return val;
 }
 
-#if IS_ENABLED(CONFIG_PM)
-
-static inline s64 ktime_since(const ktime_t kt)
+static inline s64 ktime_since_raw(const ktime_t kt)
 {
-	return ktime_to_ns(ktime_sub(ktime_get(), kt));
+	return ktime_to_ns(ktime_sub(ktime_get_raw(), kt));
 }
 
 static u64 get_rc6(struct intel_gt *gt)
@@ -170,7 +191,7 @@ static u64 get_rc6(struct intel_gt *gt)
 		 * on top of the last known real value, as the approximated RC6
 		 * counter value.
 		 */
-		val = ktime_since(pmu->sleep_last);
+		val = ktime_since_raw(pmu->sleep_last);
 		val += pmu->sample[__I915_SAMPLE_RC6].cur;
 	}
 
@@ -184,26 +205,26 @@ static u64 get_rc6(struct intel_gt *gt)
 	return val;
 }
 
-static void park_rc6(struct drm_i915_private *i915)
+static void init_rc6(struct i915_pmu *pmu)
 {
-	struct i915_pmu *pmu = &i915->pmu;
+	struct drm_i915_private *i915 = container_of(pmu, typeof(*i915), pmu);
+	intel_wakeref_t wakeref;
 
-	if (pmu->enable & config_enabled_mask(I915_PMU_RC6_RESIDENCY))
+	with_intel_runtime_pm(i915->gt.uncore->rpm, wakeref) {
 		pmu->sample[__I915_SAMPLE_RC6].cur = __get_rc6(&i915->gt);
-
-	pmu->sleep_last = ktime_get();
+		pmu->sample[__I915_SAMPLE_RC6_LAST_REPORTED].cur =
+					pmu->sample[__I915_SAMPLE_RC6].cur;
+		pmu->sleep_last = ktime_get_raw();
+	}
 }
 
-#else
-
-static u64 get_rc6(struct intel_gt *gt)
+static void park_rc6(struct drm_i915_private *i915)
 {
-	return __get_rc6(gt);
-}
-
-static void park_rc6(struct drm_i915_private *i915) {}
+	struct i915_pmu *pmu = &i915->pmu;
 
-#endif
+	pmu->sample[__I915_SAMPLE_RC6].cur = __get_rc6(&i915->gt);
+	pmu->sleep_last = ktime_get_raw();
+}
 
 static void __i915_pmu_maybe_start_timer(struct i915_pmu *pmu)
 {
@@ -343,8 +364,8 @@ add_sample_mult(struct i915_pmu_sample *sample, u32 val, u32 mul)
 static bool frequency_sampling_enabled(struct i915_pmu *pmu)
 {
 	return pmu->enable &
-	       (config_enabled_mask(I915_PMU_ACTUAL_FREQUENCY) |
-		config_enabled_mask(I915_PMU_REQUESTED_FREQUENCY));
+	       (config_mask(I915_PMU_ACTUAL_FREQUENCY) |
+		config_mask(I915_PMU_REQUESTED_FREQUENCY));
 }
 
 static void
@@ -362,7 +383,7 @@ frequency_sample(struct intel_gt *gt, unsigned int period_ns)
 	if (!intel_gt_pm_get_if_awake(gt))
 		return;
 
-	if (pmu->enable & config_enabled_mask(I915_PMU_ACTUAL_FREQUENCY)) {
+	if (pmu->enable & config_mask(I915_PMU_ACTUAL_FREQUENCY)) {
 		u32 val;
 
 		/*
@@ -384,7 +405,7 @@ frequency_sample(struct intel_gt *gt, unsigned int period_ns)
 				intel_gpu_freq(rps, val), period_ns / 1000);
 	}
 
-	if (pmu->enable & config_enabled_mask(I915_PMU_REQUESTED_FREQUENCY)) {
+	if (pmu->enable & config_mask(I915_PMU_REQUESTED_FREQUENCY)) {
 		add_sample_mult(&pmu->sample[__I915_SAMPLE_FREQ_REQ],
 				intel_gpu_freq(rps, rps->cur_freq),
 				period_ns / 1000);
@@ -471,6 +492,8 @@ config_status(struct drm_i915_private *i915, u64 config)
 		if (!HAS_RC6(i915))
 			return -ENODEV;
 		break;
+	case I915_PMU_SOFTWARE_GT_AWAKE_TIME:
+		break;
 	default:
 		return -ENOENT;
 	}
@@ -578,6 +601,9 @@ static u64 __i915_pmu_event_read(struct perf_event *event)
 		case I915_PMU_RC6_RESIDENCY:
 			val = get_rc6(&i915->gt);
 			break;
+		case I915_PMU_SOFTWARE_GT_AWAKE_TIME:
+			val = ktime_to_ns(intel_gt_get_awake_time(&i915->gt));
+			break;
 		}
 	}
 
@@ -610,12 +636,14 @@ static void i915_pmu_enable(struct perf_event *event)
 {
 	struct drm_i915_private *i915 =
 		container_of(event->pmu, typeof(*i915), pmu.base);
-	unsigned int bit = event_enabled_bit(event);
 	struct i915_pmu *pmu = &i915->pmu;
-	intel_wakeref_t wakeref;
 	unsigned long flags;
+	unsigned int bit;
+
+	bit = event_bit(event);
+	if (bit == -1)
+		goto update;
 
-	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
 	spin_lock_irqsave(&pmu->lock, flags);
 
 	/*
@@ -626,13 +654,6 @@ static void i915_pmu_enable(struct perf_event *event)
 	GEM_BUG_ON(bit >= ARRAY_SIZE(pmu->enable_count));
 	GEM_BUG_ON(pmu->enable_count[bit] == ~0);
 
-	if (pmu->enable_count[bit] == 0 &&
-	    config_enabled_mask(I915_PMU_RC6_RESIDENCY) & BIT_ULL(bit)) {
-		pmu->sample[__I915_SAMPLE_RC6_LAST_REPORTED].cur = 0;
-		pmu->sample[__I915_SAMPLE_RC6].cur = __get_rc6(&i915->gt);
-		pmu->sleep_last = ktime_get();
-	}
-
 	pmu->enable |= BIT_ULL(bit);
 	pmu->enable_count[bit]++;
 
@@ -667,24 +688,26 @@ static void i915_pmu_enable(struct perf_event *event)
 
 	spin_unlock_irqrestore(&pmu->lock, flags);
 
+update:
 	/*
 	 * Store the current counter value so we can report the correct delta
 	 * for all listeners. Even when the event was already enabled and has
 	 * an existing non-zero value.
 	 */
 	local64_set(&event->hw.prev_count, __i915_pmu_event_read(event));
-
-	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
 }
 
 static void i915_pmu_disable(struct perf_event *event)
 {
 	struct drm_i915_private *i915 =
 		container_of(event->pmu, typeof(*i915), pmu.base);
-	unsigned int bit = event_enabled_bit(event);
+	unsigned int bit = event_bit(event);
 	struct i915_pmu *pmu = &i915->pmu;
 	unsigned long flags;
 
+	if (bit == -1)
+		return;
+
 	spin_lock_irqsave(&pmu->lock, flags);
 
 	if (is_engine_event(event)) {
@@ -881,6 +904,7 @@ create_event_attributes(struct i915_pmu *pmu)
 		__event(I915_PMU_REQUESTED_FREQUENCY, "requested-frequency", "M"),
 		__event(I915_PMU_INTERRUPTS, "interrupts", NULL),
 		__event(I915_PMU_RC6_RESIDENCY, "rc6-residency", "ns"),
+		__event(I915_PMU_SOFTWARE_GT_AWAKE_TIME, "software-gt-awake-time", "ns"),
 	};
 	static const struct {
 		enum drm_i915_pmu_engine_sample sample;
@@ -1130,6 +1154,7 @@ void i915_pmu_register(struct drm_i915_private *i915)
 	hrtimer_init(&pmu->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
 	pmu->timer.function = i915_sample;
 	pmu->cpuhp.cpu = -1;
+	init_rc6(pmu);
 
 	if (!is_igp(i915)) {
 		pmu->name = kasprintf(GFP_KERNEL,
diff --git a/drivers/gpu/drm/i915/i915_pmu.h b/drivers/gpu/drm/i915/i915_pmu.h
index 8405d6da5b9a..60f9595f902c 100644
--- a/drivers/gpu/drm/i915/i915_pmu.h
+++ b/drivers/gpu/drm/i915/i915_pmu.h
@@ -14,6 +14,21 @@
 
 struct drm_i915_private;
 
+/**
+ * Non-engine events that we need to track enabled-disabled transition and
+ * current state.
+ */
+enum i915_pmu_tracked_events {
+	__I915_PMU_ACTUAL_FREQUENCY_ENABLED = 0,
+	__I915_PMU_REQUESTED_FREQUENCY_ENABLED,
+	__I915_PMU_RC6_RESIDENCY_ENABLED,
+	__I915_PMU_TRACKED_EVENT_COUNT, /* count marker */
+};
+
+/**
+ * Slots used from the sampling timer (non-engine events) with some extras for
+ * convenience.
+ */
 enum {
 	__I915_SAMPLE_FREQ_ACT = 0,
 	__I915_SAMPLE_FREQ_REQ,
@@ -28,8 +43,7 @@ enum {
  * It is also used to know to needed number of event reference counters.
  */
 #define I915_PMU_MASK_BITS \
-	((1 << I915_PMU_SAMPLE_BITS) + \
-	 (I915_PMU_LAST + 1 - __I915_PMU_OTHER(0)))
+	(I915_ENGINE_SAMPLE_COUNT + __I915_PMU_TRACKED_EVENT_COUNT)
 
 #define I915_ENGINE_SAMPLE_COUNT (I915_SAMPLE_SEMA + 1)
 
@@ -66,18 +80,17 @@ struct i915_pmu {
 	 */
 	struct hrtimer timer;
 	/**
-	 * @enable: Bitmask of all currently enabled events.
+	 * @enable: Bitmask of specific enabled events.
+	 *
+	 * For some events we need to track their state and do some internal
+	 * house keeping.
 	 *
-	 * Bits are derived from uAPI event numbers in a way that low 16 bits
-	 * correspond to engine event _sample_ _type_ (I915_SAMPLE_QUEUED is
-	 * bit 0), and higher bits correspond to other events (for instance
-	 * I915_PMU_ACTUAL_FREQUENCY is bit 16 etc).
+	 * Each engine event sampler type and event listed in enum
+	 * i915_pmu_tracked_events gets a bit in this field.
 	 *
-	 * In other words, low 16 bits are not per engine but per engine
-	 * sampler type, while the upper bits are directly mapped to other
-	 * event types.
+	 * Low bits are engine samplers and other events continue from there.
 	 */
-	u64 enable;
+	u32 enable;
 
 	/**
 	 * @timer_last:
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 0023c023f472..7146cd0f3256 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2928,8 +2928,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 
 #define HDPORT_STATE			_MMIO(0x45050)
 #define   HDPORT_DPLL_USED_MASK		REG_GENMASK(14, 12)
-#define   HDPORT_PHY_USED_DP(phy)	REG_BIT(2 * (phy) + 2)
-#define   HDPORT_PHY_USED_HDMI(phy)	REG_BIT(2 * (phy) + 1)
+#define   HDPORT_DDI_USED(phy)		REG_BIT(2 * (phy) + 1)
 #define   HDPORT_ENABLED		REG_BIT(0)
 
 /* Make render/texture TLB fetches lower priorty than associated data
@@ -4347,13 +4346,13 @@ enum {
 #define _TRANS_VRR_CTL_B		0x61420
 #define _TRANS_VRR_CTL_C		0x62420
 #define _TRANS_VRR_CTL_D		0x63420
-#define TRANS_VRR_CTL(trans)		_MMIO_TRANS2(trans, _TRANS_VRR_CTL_A)
-#define   VRR_CTL_VRR_ENABLE		REG_BIT(31)
-#define   VRR_CTL_IGN_MAX_SHIFT		REG_BIT(30)
-#define   VRR_CTL_FLIP_LINE_EN		REG_BIT(29)
-#define   VRR_CTL_LINE_COUNT_MASK	REG_GENMASK(10, 3)
-#define   VRR_CTL_LINE_COUNT(x)		REG_FIELD_PREP(VRR_CTL_LINE_COUNT_MASK, (x))
-#define   VRR_CTL_SW_FULLLINE_COUNT	REG_BIT(0)
+#define TRANS_VRR_CTL(trans)			_MMIO_TRANS2(trans, _TRANS_VRR_CTL_A)
+#define   VRR_CTL_VRR_ENABLE			REG_BIT(31)
+#define   VRR_CTL_IGN_MAX_SHIFT			REG_BIT(30)
+#define   VRR_CTL_FLIP_LINE_EN			REG_BIT(29)
+#define   VRR_CTL_PIPELINE_FULL_MASK		REG_GENMASK(10, 3)
+#define   VRR_CTL_PIPELINE_FULL(x)		REG_FIELD_PREP(VRR_CTL_PIPELINE_FULL_MASK, (x))
+#define   VRR_CTL_PIPELINE_FULL_OVERRIDE	REG_BIT(0)
 
 #define _TRANS_VRR_VMAX_A		0x60424
 #define _TRANS_VRR_VMAX_B		0x61424
@@ -6579,6 +6578,7 @@ enum {
 #define TGL_CURSOR_D_OFFSET 0x73080
 
 /* Display A control */
+#define _DSPAADDR_VLV				0x7017C /* vlv/chv */
 #define _DSPACNTR				0x70180
 #define   DISPLAY_PLANE_ENABLE			(1 << 31)
 #define   DISPLAY_PLANE_DISABLE			0
@@ -6615,6 +6615,7 @@ enum {
 #define   DISPPLANE_ROTATE_180			(1 << 15)
 #define   DISPPLANE_TRICKLE_FEED_DISABLE	(1 << 14) /* Ironlake */
 #define   DISPPLANE_TILED			(1 << 10)
+#define   DISPPLANE_ASYNC_FLIP			(1 << 9) /* g4x+ */
 #define   DISPPLANE_MIRROR			(1 << 8) /* CHV pipe B */
 #define _DSPAADDR				0x70184
 #define _DSPASTRIDE				0x70188
@@ -6626,6 +6627,7 @@ enum {
 #define _DSPASURFLIVE				0x701AC
 #define _DSPAGAMC				0x701E0
 
+#define DSPADDR_VLV(plane)	_MMIO_PIPE2(plane, _DSPAADDR_VLV)
 #define DSPCNTR(plane)		_MMIO_PIPE2(plane, _DSPACNTR)
 #define DSPADDR(plane)		_MMIO_PIPE2(plane, _DSPAADDR)
 #define DSPSTRIDE(plane)	_MMIO_PIPE2(plane, _DSPASTRIDE)
@@ -7071,6 +7073,8 @@ enum {
 #define _PLANE_KEYMAX_1_A			0x701a0
 #define _PLANE_KEYMAX_2_A			0x702a0
 #define  PLANE_KEYMAX_ALPHA(a)			((a) << 24)
+#define _PLANE_CC_VAL_1_A			0x701b4
+#define _PLANE_CC_VAL_2_A			0x702b4
 #define _PLANE_AUX_DIST_1_A			0x701c0
 #define _PLANE_AUX_DIST_2_A			0x702c0
 #define _PLANE_AUX_OFFSET_1_A			0x701c4
@@ -7112,6 +7116,13 @@ enum {
 #define _PLANE_NV12_BUF_CFG_1_A		0x70278
 #define _PLANE_NV12_BUF_CFG_2_A		0x70378
 
+#define _PLANE_CC_VAL_1_B			0x711b4
+#define _PLANE_CC_VAL_2_B			0x712b4
+#define _PLANE_CC_VAL_1(pipe)	_PIPE(pipe, _PLANE_CC_VAL_1_A, _PLANE_CC_VAL_1_B)
+#define _PLANE_CC_VAL_2(pipe)	_PIPE(pipe, _PLANE_CC_VAL_2_A, _PLANE_CC_VAL_2_B)
+#define PLANE_CC_VAL(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_CC_VAL_1(pipe), _PLANE_CC_VAL_2(pipe))
+
 /* Input CSC Register Definitions */
 #define _PLANE_INPUT_CSC_RY_GY_1_A	0x701E0
 #define _PLANE_INPUT_CSC_RY_GY_2_A	0x702E0
@@ -8214,6 +8225,7 @@ enum {
 #define  GEN11_LQSC_CLEAN_EVICT_DISABLE		(1 << 6)
 #define  GEN8_LQSC_RO_PERF_DIS			(1 << 27)
 #define  GEN8_LQSC_FLUSH_COHERENT_LINES		(1 << 21)
+#define  GEN8_LQSQ_NONIA_COHERENT_ATOMICS_ENABLE REG_BIT(22)
 
 /* GEN8 chicken */
 #define HDC_CHICKEN0				_MMIO(0x7300)
@@ -9873,6 +9885,7 @@ enum skl_power_gate {
 					  _PORTD_HDCP2_BASE, \
 					  _PORTE_HDCP2_BASE, \
 					  _PORTF_HDCP2_BASE) + (x))
+
 #define PORT_HDCP2_AUTH(port)		_PORT_HDCP2_BASE(port, 0x98)
 #define _TRANSA_HDCP2_AUTH		0x66498
 #define _TRANSB_HDCP2_AUTH		0x66598
@@ -9912,6 +9925,44 @@ enum skl_power_gate {
 					 TRANS_HDCP2_STATUS(trans) : \
 					 PORT_HDCP2_STATUS(port))
 
+#define _PIPEA_HDCP2_STREAM_STATUS	0x668C0
+#define _PIPEB_HDCP2_STREAM_STATUS	0x665C0
+#define _PIPEC_HDCP2_STREAM_STATUS	0x666C0
+#define _PIPED_HDCP2_STREAM_STATUS	0x667C0
+#define PIPE_HDCP2_STREAM_STATUS(pipe)		_MMIO(_PICK((pipe), \
+						      _PIPEA_HDCP2_STREAM_STATUS, \
+						      _PIPEB_HDCP2_STREAM_STATUS, \
+						      _PIPEC_HDCP2_STREAM_STATUS, \
+						      _PIPED_HDCP2_STREAM_STATUS))
+
+#define _TRANSA_HDCP2_STREAM_STATUS		0x664C0
+#define _TRANSB_HDCP2_STREAM_STATUS		0x665C0
+#define TRANS_HDCP2_STREAM_STATUS(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP2_STREAM_STATUS, \
+						    _TRANSB_HDCP2_STREAM_STATUS)
+#define   STREAM_ENCRYPTION_STATUS	BIT(31)
+#define   STREAM_TYPE_STATUS		BIT(30)
+#define HDCP2_STREAM_STATUS(dev_priv, trans, port) \
+					(INTEL_GEN(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_STREAM_STATUS(trans) : \
+					 PIPE_HDCP2_STREAM_STATUS(pipe))
+
+#define _PORTA_HDCP2_AUTH_STREAM		0x66F00
+#define _PORTB_HDCP2_AUTH_STREAM		0x66F04
+#define PORT_HDCP2_AUTH_STREAM(port)	_MMIO_PORT(port, \
+						   _PORTA_HDCP2_AUTH_STREAM, \
+						   _PORTB_HDCP2_AUTH_STREAM)
+#define _TRANSA_HDCP2_AUTH_STREAM		0x66F00
+#define _TRANSB_HDCP2_AUTH_STREAM		0x66F04
+#define TRANS_HDCP2_AUTH_STREAM(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP2_AUTH_STREAM, \
+						    _TRANSB_HDCP2_AUTH_STREAM)
+#define   AUTH_STREAM_TYPE		BIT(31)
+#define HDCP2_AUTH_STREAM(dev_priv, trans, port) \
+					(INTEL_GEN(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_AUTH_STREAM(trans) : \
+					 PORT_HDCP2_AUTH_STREAM(port))
+
 /* Per-pipe DDI Function Control */
 #define _TRANS_DDI_FUNC_CTL_A		0x60400
 #define _TRANS_DDI_FUNC_CTL_B		0x61400
@@ -9961,6 +10012,7 @@ enum skl_power_gate {
 #define  TRANS_DDI_DP_VC_PAYLOAD_ALLOC	(1 << 8)
 #define  TRANS_DDI_HDMI_SCRAMBLER_CTS_ENABLE (1 << 7)
 #define  TRANS_DDI_HDMI_SCRAMBLER_RESET_FREQ (1 << 6)
+#define  TRANS_DDI_HDCP_SELECT		REG_BIT(5)
 #define  TRANS_DDI_BFI_ENABLE		(1 << 4)
 #define  TRANS_DDI_HIGH_TMDS_CHAR_RATE	(1 << 4)
 #define  TRANS_DDI_HDMI_SCRAMBLING	(1 << 0)
@@ -12056,6 +12108,12 @@ enum skl_power_gate {
 #define __GEN11_VCS2_MOCS0	0x10000
 #define GEN11_MFX2_MOCS(i)	_MMIO(__GEN11_VCS2_MOCS0 + (i) * 4)
 
+#define GEN9_SCRATCH_LNCF1		_MMIO(0xb008)
+#define   GEN9_LNCF_NONIA_COHERENT_ATOMICS_ENABLE REG_BIT(0)
+
+#define GEN9_SCRATCH1			_MMIO(0xb11c)
+#define   EVICTION_PERF_FIX_ENABLE	REG_BIT(8)
+
 #define GEN10_SCRATCH_LNCF2		_MMIO(0xb0a0)
 #define   PMFLUSHDONE_LNICRSDROP	(1 << 20)
 #define   PMFLUSH_GAPL3UNBLOCK		(1 << 21)
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 5385b081a376..22e39d938f17 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -33,6 +33,7 @@
 #include "gem/i915_gem_context.h"
 #include "gt/intel_breadcrumbs.h"
 #include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_ring.h"
 #include "gt/intel_rps.h"
 
@@ -275,7 +276,7 @@ static void remove_from_engine(struct i915_request *rq)
 
 bool i915_request_retire(struct i915_request *rq)
 {
-	if (!i915_request_completed(rq))
+	if (!__i915_request_is_complete(rq))
 		return false;
 
 	RQ_TRACE(rq, "\n");
@@ -306,10 +307,8 @@ bool i915_request_retire(struct i915_request *rq)
 		spin_unlock_irq(&rq->lock);
 	}
 
-	if (i915_request_has_waitboost(rq)) {
-		GEM_BUG_ON(!atomic_read(&rq->engine->gt->rps.num_waiters));
+	if (test_and_set_bit(I915_FENCE_FLAG_BOOST, &rq->fence.flags))
 		atomic_dec(&rq->engine->gt->rps.num_waiters);
-	}
 
 	/*
 	 * We only loosely track inflight requests across preemption,
@@ -321,7 +320,8 @@ bool i915_request_retire(struct i915_request *rq)
 	 * after removing the breadcrumb and signaling it, so that we do not
 	 * inadvertently attach the breadcrumb to a completed request.
 	 */
-	remove_from_engine(rq);
+	if (!list_empty(&rq->sched.link))
+		remove_from_engine(rq);
 	GEM_BUG_ON(!llist_empty(&rq->execute_cb));
 
 	__list_del_entry(&rq->link); /* poison neither prev/next (RCU walks) */
@@ -342,8 +342,7 @@ void i915_request_retire_upto(struct i915_request *rq)
 	struct i915_request *tmp;
 
 	RQ_TRACE(rq, "\n");
-
-	GEM_BUG_ON(!i915_request_completed(rq));
+	GEM_BUG_ON(!__i915_request_is_complete(rq));
 
 	do {
 		tmp = list_first_entry(&tl->requests, typeof(*tmp), link);
@@ -488,6 +487,8 @@ void __i915_request_skip(struct i915_request *rq)
 	if (rq->infix == rq->postfix)
 		return;
 
+	RQ_TRACE(rq, "error: %d\n", rq->fence.error);
+
 	/*
 	 * As this request likely depends on state from the lost
 	 * context, clear out all the user operations leaving the
@@ -513,6 +514,17 @@ void i915_request_set_error_once(struct i915_request *rq, int error)
 	} while (!try_cmpxchg(&rq->fence.error, &old, error));
 }
 
+void i915_request_mark_eio(struct i915_request *rq)
+{
+	if (__i915_request_is_complete(rq))
+		return;
+
+	GEM_BUG_ON(i915_request_signaled(rq));
+
+	i915_request_set_error_once(rq, -EIO);
+	i915_request_mark_complete(rq);
+}
+
 bool __i915_request_submit(struct i915_request *request)
 {
 	struct intel_engine_cs *engine = request->engine;
@@ -539,12 +551,10 @@ bool __i915_request_submit(struct i915_request *request)
 	 * dropped upon retiring. (Otherwise if resubmit a *retired*
 	 * request, this would be a horrible use-after-free.)
 	 */
-	if (i915_request_completed(request))
-		goto xfer;
-
-	if (unlikely(intel_context_is_closed(request->context) &&
-		     !intel_engine_has_heartbeat(engine)))
-		intel_context_set_banned(request->context);
+	if (__i915_request_is_complete(request)) {
+		list_del_init(&request->sched.link);
+		goto active;
+	}
 
 	if (unlikely(intel_context_is_banned(request->context)))
 		i915_request_set_error_once(request, -EIO);
@@ -579,11 +589,11 @@ bool __i915_request_submit(struct i915_request *request)
 	engine->serial++;
 	result = true;
 
-xfer:
-	if (!test_and_set_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags)) {
-		list_move_tail(&request->sched.link, &engine->active.requests);
-		clear_bit(I915_FENCE_FLAG_PQUEUE, &request->fence.flags);
-	}
+	GEM_BUG_ON(test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags));
+	list_move_tail(&request->sched.link, &engine->active.requests);
+active:
+	clear_bit(I915_FENCE_FLAG_PQUEUE, &request->fence.flags);
+	set_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags);
 
 	/*
 	 * XXX Rollback bonded-execution on __i915_request_unsubmit()?
@@ -643,7 +653,7 @@ void __i915_request_unsubmit(struct i915_request *request)
 		i915_request_cancel_breadcrumb(request);
 
 	/* We've already spun, don't charge on resubmitting. */
-	if (request->sched.semaphores && i915_request_started(request))
+	if (request->sched.semaphores && __i915_request_has_started(request))
 		request->sched.semaphores = 0;
 
 	/*
@@ -855,7 +865,7 @@ __i915_request_create(struct intel_context *ce, gfp_t gfp)
 	RCU_INIT_POINTER(rq->timeline, tl);
 	RCU_INIT_POINTER(rq->hwsp_cacheline, tl->hwsp_cacheline);
 	rq->hwsp_seqno = tl->hwsp_seqno;
-	GEM_BUG_ON(i915_request_completed(rq));
+	GEM_BUG_ON(__i915_request_is_complete(rq));
 
 	rq->rcustate = get_state_synchronize_rcu(); /* acts as smp_mb() */
 
@@ -961,15 +971,22 @@ i915_request_await_start(struct i915_request *rq, struct i915_request *signal)
 	if (i915_request_started(signal))
 		return 0;
 
+	/*
+	 * The caller holds a reference on @signal, but we do not serialise
+	 * against it being retired and removed from the lists.
+	 *
+	 * We do not hold a reference to the request before @signal, and
+	 * so must be very careful to ensure that it is not _recycled_ as
+	 * we follow the link backwards.
+	 */
 	fence = NULL;
 	rcu_read_lock();
-	spin_lock_irq(&signal->lock);
 	do {
 		struct list_head *pos = READ_ONCE(signal->link.prev);
 		struct i915_request *prev;
 
 		/* Confirm signal has not been retired, the link is valid */
-		if (unlikely(i915_request_started(signal)))
+		if (unlikely(__i915_request_has_started(signal)))
 			break;
 
 		/* Is signal the earliest request on its timeline? */
@@ -994,7 +1011,6 @@ i915_request_await_start(struct i915_request *rq, struct i915_request *signal)
 
 		fence = &prev->fence;
 	} while (0);
-	spin_unlock_irq(&signal->lock);
 	rcu_read_unlock();
 	if (!fence)
 		return 0;
@@ -1511,7 +1527,7 @@ __i915_request_add_to_timeline(struct i915_request *rq)
 	 */
 	prev = to_request(__i915_active_fence_set(&timeline->last_request,
 						  &rq->fence));
-	if (prev && !i915_request_completed(prev)) {
+	if (prev && !__i915_request_is_complete(prev)) {
 		/*
 		 * The requests are supposed to be kept in order. However,
 		 * we need to be wary in case the timeline->last_request
@@ -1582,6 +1598,12 @@ struct i915_request *__i915_request_commit(struct i915_request *rq)
 	return __i915_request_add_to_timeline(rq);
 }
 
+void __i915_request_queue_bh(struct i915_request *rq)
+{
+	i915_sw_fence_commit(&rq->semaphore);
+	i915_sw_fence_commit(&rq->submit);
+}
+
 void __i915_request_queue(struct i915_request *rq,
 			  const struct i915_sched_attr *attr)
 {
@@ -1598,8 +1620,10 @@ void __i915_request_queue(struct i915_request *rq,
 	 */
 	if (attr && rq->engine->schedule)
 		rq->engine->schedule(rq, attr);
-	i915_sw_fence_commit(&rq->semaphore);
-	i915_sw_fence_commit(&rq->submit);
+
+	local_bh_disable();
+	__i915_request_queue_bh(rq);
+	local_bh_enable(); /* kick tasklets */
 }
 
 void i915_request_add(struct i915_request *rq)
@@ -1823,7 +1847,7 @@ long i915_request_wait(struct i915_request *rq,
 	 * for unhappy HW.
 	 */
 	if (i915_request_is_ready(rq))
-		intel_engine_flush_submission(rq->engine);
+		__intel_engine_flush_submission(rq->engine, false);
 
 	for (;;) {
 		set_current_state(state);
@@ -1855,6 +1879,106 @@ out:
 	return timeout;
 }
 
+static int print_sched_attr(const struct i915_sched_attr *attr,
+			    char *buf, int x, int len)
+{
+	if (attr->priority == I915_PRIORITY_INVALID)
+		return x;
+
+	x += snprintf(buf + x, len - x,
+		      " prio=%d", attr->priority);
+
+	return x;
+}
+
+static char queue_status(const struct i915_request *rq)
+{
+	if (i915_request_is_active(rq))
+		return 'E';
+
+	if (i915_request_is_ready(rq))
+		return intel_engine_is_virtual(rq->engine) ? 'V' : 'R';
+
+	return 'U';
+}
+
+static const char *run_status(const struct i915_request *rq)
+{
+	if (__i915_request_is_complete(rq))
+		return "!";
+
+	if (__i915_request_has_started(rq))
+		return "*";
+
+	if (!i915_sw_fence_signaled(&rq->semaphore))
+		return "&";
+
+	return "";
+}
+
+static const char *fence_status(const struct i915_request *rq)
+{
+	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags))
+		return "+";
+
+	if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &rq->fence.flags))
+		return "-";
+
+	return "";
+}
+
+void i915_request_show(struct drm_printer *m,
+		       const struct i915_request *rq,
+		       const char *prefix,
+		       int indent)
+{
+	const char *name = rq->fence.ops->get_timeline_name((struct dma_fence *)&rq->fence);
+	char buf[80] = "";
+	int x = 0;
+
+	/*
+	 * The prefix is used to show the queue status, for which we use
+	 * the following flags:
+	 *
+	 *  U [Unready]
+	 *    - initial status upon being submitted by the user
+	 *
+	 *    - the request is not ready for execution as it is waiting
+	 *      for external fences
+	 *
+	 *  R [Ready]
+	 *    - all fences the request was waiting on have been signaled,
+	 *      and the request is now ready for execution and will be
+	 *      in a backend queue
+	 *
+	 *    - a ready request may still need to wait on semaphores
+	 *      [internal fences]
+	 *
+	 *  V [Ready/virtual]
+	 *    - same as ready, but queued over multiple backends
+	 *
+	 *  E [Executing]
+	 *    - the request has been transferred from the backend queue and
+	 *      submitted for execution on HW
+	 *
+	 *    - a completed request may still be regarded as executing, its
+	 *      status may not be updated until it is retired and removed
+	 *      from the lists
+	 */
+
+	x = print_sched_attr(&rq->sched.attr, buf, x, sizeof(buf));
+
+	drm_printf(m, "%s%.*s%c %llx:%lld%s%s %s @ %dms: %s\n",
+		   prefix, indent, "                ",
+		   queue_status(rq),
+		   rq->fence.context, rq->fence.seqno,
+		   run_status(rq),
+		   fence_status(rq),
+		   buf,
+		   jiffies_to_msecs(jiffies - rq->emitted_jiffies),
+		   name);
+}
+
 #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
 #include "selftests/mock_request.c"
 #include "selftests/i915_request.c"
diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h
index 620b6fab2c5c..1bfe214a47e9 100644
--- a/drivers/gpu/drm/i915/i915_request.h
+++ b/drivers/gpu/drm/i915/i915_request.h
@@ -43,6 +43,7 @@
 
 struct drm_file;
 struct drm_i915_gem_object;
+struct drm_printer;
 struct i915_request;
 
 struct i915_capture_list {
@@ -308,12 +309,14 @@ __i915_request_create(struct intel_context *ce, gfp_t gfp);
 struct i915_request * __must_check
 i915_request_create(struct intel_context *ce);
 
-void i915_request_set_error_once(struct i915_request *rq, int error);
 void __i915_request_skip(struct i915_request *rq);
+void i915_request_set_error_once(struct i915_request *rq, int error);
+void i915_request_mark_eio(struct i915_request *rq);
 
 struct i915_request *__i915_request_commit(struct i915_request *request);
 void __i915_request_queue(struct i915_request *rq,
 			  const struct i915_sched_attr *attr);
+void __i915_request_queue_bh(struct i915_request *rq);
 
 bool i915_request_retire(struct i915_request *rq);
 void i915_request_retire_upto(struct i915_request *rq);
@@ -371,6 +374,11 @@ long i915_request_wait(struct i915_request *rq,
 #define I915_WAIT_PRIORITY	BIT(1) /* small priority bump for the request */
 #define I915_WAIT_ALL		BIT(2) /* used by i915_gem_object_wait() */
 
+void i915_request_show(struct drm_printer *m,
+		       const struct i915_request *rq,
+		       const char *prefix,
+		       int indent);
+
 static inline bool i915_request_signaled(const struct i915_request *rq)
 {
 	/* The request may live longer than its HWSP, so check flags first! */
@@ -434,7 +442,7 @@ static inline u32 hwsp_seqno(const struct i915_request *rq)
 
 static inline bool __i915_request_has_started(const struct i915_request *rq)
 {
-	return i915_seqno_passed(hwsp_seqno(rq), rq->fence.seqno - 1);
+	return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno - 1);
 }
 
 /**
@@ -465,11 +473,19 @@ static inline bool __i915_request_has_started(const struct i915_request *rq)
  */
 static inline bool i915_request_started(const struct i915_request *rq)
 {
+	bool result;
+
 	if (i915_request_signaled(rq))
 		return true;
 
-	/* Remember: started but may have since been preempted! */
-	return __i915_request_has_started(rq);
+	result = true;
+	rcu_read_lock(); /* the HWSP may be freed at runtime */
+	if (likely(!i915_request_signaled(rq)))
+		/* Remember: started but may have since been preempted! */
+		result = __i915_request_has_started(rq);
+	rcu_read_unlock();
+
+	return result;
 }
 
 /**
@@ -482,10 +498,16 @@ static inline bool i915_request_started(const struct i915_request *rq)
  */
 static inline bool i915_request_is_running(const struct i915_request *rq)
 {
+	bool result;
+
 	if (!i915_request_is_active(rq))
 		return false;
 
-	return __i915_request_has_started(rq);
+	rcu_read_lock();
+	result = __i915_request_has_started(rq) && i915_request_is_active(rq);
+	rcu_read_unlock();
+
+	return result;
 }
 
 /**
@@ -509,12 +531,25 @@ static inline bool i915_request_is_ready(const struct i915_request *rq)
 	return !list_empty(&rq->sched.link);
 }
 
+static inline bool __i915_request_is_complete(const struct i915_request *rq)
+{
+	return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno);
+}
+
 static inline bool i915_request_completed(const struct i915_request *rq)
 {
+	bool result;
+
 	if (i915_request_signaled(rq))
 		return true;
 
-	return i915_seqno_passed(hwsp_seqno(rq), rq->fence.seqno);
+	result = true;
+	rcu_read_lock(); /* the HWSP may be freed at runtime */
+	if (likely(!i915_request_signaled(rq)))
+		result = __i915_request_is_complete(rq);
+	rcu_read_unlock();
+
+	return result;
 }
 
 static inline void i915_request_mark_complete(struct i915_request *rq)
diff --git a/drivers/gpu/drm/i915/i915_scheduler.c b/drivers/gpu/drm/i915/i915_scheduler.c
index cbb880b10c65..7144239f08df 100644
--- a/drivers/gpu/drm/i915/i915_scheduler.c
+++ b/drivers/gpu/drm/i915/i915_scheduler.c
@@ -458,14 +458,10 @@ int i915_sched_node_add_dependency(struct i915_sched_node *node,
 	if (!dep)
 		return -ENOMEM;
 
-	local_bh_disable();
-
 	if (!__i915_sched_node_add_dependency(node, signal, dep,
 					      flags | I915_DEPENDENCY_ALLOC))
 		i915_dependency_free(dep);
 
-	local_bh_enable(); /* kick submission tasklet */
-
 	return 0;
 }
 
@@ -504,6 +500,34 @@ void i915_sched_node_fini(struct i915_sched_node *node)
 	spin_unlock_irq(&schedule_lock);
 }
 
+void i915_request_show_with_schedule(struct drm_printer *m,
+				     const struct i915_request *rq,
+				     const char *prefix,
+				     int indent)
+{
+	struct i915_dependency *dep;
+
+	i915_request_show(m, rq, prefix, indent);
+	if (i915_request_completed(rq))
+		return;
+
+	rcu_read_lock();
+	for_each_signaler(dep, rq) {
+		const struct i915_request *signaler =
+			node_to_request(dep->signaler);
+
+		/* Dependencies along the same timeline are expected. */
+		if (signaler->timeline == rq->timeline)
+			continue;
+
+		if (__i915_request_is_complete(signaler))
+			continue;
+
+		i915_request_show(m, signaler, prefix, indent + 2);
+	}
+	rcu_read_unlock();
+}
+
 static void i915_global_scheduler_shrink(void)
 {
 	kmem_cache_shrink(global.slab_dependencies);
diff --git a/drivers/gpu/drm/i915/i915_scheduler.h b/drivers/gpu/drm/i915/i915_scheduler.h
index 6f0bf00fc569..4501e5ac2637 100644
--- a/drivers/gpu/drm/i915/i915_scheduler.h
+++ b/drivers/gpu/drm/i915/i915_scheduler.h
@@ -13,6 +13,8 @@
 
 #include "i915_scheduler_types.h"
 
+struct drm_printer;
+
 #define priolist_for_each_request(it, plist, idx) \
 	for (idx = 0; idx < ARRAY_SIZE((plist)->requests); idx++) \
 		list_for_each_entry(it, &(plist)->requests[idx], sched.link)
@@ -54,4 +56,9 @@ static inline void i915_priolist_free(struct i915_priolist *p)
 		__i915_priolist_free(p);
 }
 
+void i915_request_show_with_schedule(struct drm_printer *m,
+				     const struct i915_request *rq,
+				     const char *prefix,
+				     int indent);
+
 #endif /* _I915_SCHEDULER_H_ */
diff --git a/drivers/gpu/drm/i915/i915_scheduler_types.h b/drivers/gpu/drm/i915/i915_scheduler_types.h
index f72e6c397b08..343ed44d5ed4 100644
--- a/drivers/gpu/drm/i915/i915_scheduler_types.h
+++ b/drivers/gpu/drm/i915/i915_scheduler_types.h
@@ -81,4 +81,14 @@ struct i915_dependency {
 #define I915_DEPENDENCY_WEAK		BIT(2)
 };
 
+#define for_each_waiter(p__, rq__) \
+	list_for_each_entry_lockless(p__, \
+				     &(rq__)->sched.waiters_list, \
+				     wait_link)
+
+#define for_each_signaler(p__, rq__) \
+	list_for_each_entry_rcu(p__, \
+				&(rq__)->sched.signalers_list, \
+				signal_link)
+
 #endif /* _I915_SCHEDULER_TYPES_H_ */
diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c
index 038d4c6884c5..2744558f3050 100644
--- a/drivers/gpu/drm/i915/i915_sw_fence.c
+++ b/drivers/gpu/drm/i915/i915_sw_fence.c
@@ -18,10 +18,15 @@
 #define I915_SW_FENCE_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
 #endif
 
-#define I915_SW_FENCE_FLAG_ALLOC BIT(3) /* after WQ_FLAG_* for safety */
-
 static DEFINE_SPINLOCK(i915_sw_fence_lock);
 
+#define WQ_FLAG_BITS \
+	BITS_PER_TYPE(typeof_member(struct wait_queue_entry, flags))
+
+/* after WQ_FLAG_* for safety */
+#define I915_SW_FENCE_FLAG_FENCE BIT(WQ_FLAG_BITS - 1)
+#define I915_SW_FENCE_FLAG_ALLOC BIT(WQ_FLAG_BITS - 2)
+
 enum {
 	DEBUG_FENCE_IDLE = 0,
 	DEBUG_FENCE_NOTIFY,
@@ -154,10 +159,10 @@ static void __i915_sw_fence_wake_up_all(struct i915_sw_fence *fence,
 	spin_lock_irqsave_nested(&x->lock, flags, 1 + !!continuation);
 	if (continuation) {
 		list_for_each_entry_safe(pos, next, &x->head, entry) {
-			if (pos->func == autoremove_wake_function)
-				pos->func(pos, TASK_NORMAL, 0, continuation);
-			else
+			if (pos->flags & I915_SW_FENCE_FLAG_FENCE)
 				list_move_tail(&pos->entry, continuation);
+			else
+				pos->func(pos, TASK_NORMAL, 0, continuation);
 		}
 	} else {
 		LIST_HEAD(extra);
@@ -166,9 +171,9 @@ static void __i915_sw_fence_wake_up_all(struct i915_sw_fence *fence,
 			list_for_each_entry_safe(pos, next, &x->head, entry) {
 				int wake_flags;
 
-				wake_flags = fence->error;
-				if (pos->func == autoremove_wake_function)
-					wake_flags = 0;
+				wake_flags = 0;
+				if (pos->flags & I915_SW_FENCE_FLAG_FENCE)
+					wake_flags = fence->error;
 
 				pos->func(pos, TASK_NORMAL, wake_flags, &extra);
 			}
@@ -332,8 +337,8 @@ static int __i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
 					  struct i915_sw_fence *signaler,
 					  wait_queue_entry_t *wq, gfp_t gfp)
 {
+	unsigned int pending;
 	unsigned long flags;
-	int pending;
 
 	debug_fence_assert(fence);
 	might_sleep_if(gfpflags_allow_blocking(gfp));
@@ -349,7 +354,7 @@ static int __i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
 	if (unlikely(i915_sw_fence_check_if_after(fence, signaler)))
 		return -EINVAL;
 
-	pending = 0;
+	pending = I915_SW_FENCE_FLAG_FENCE;
 	if (!wq) {
 		wq = kmalloc(sizeof(*wq), gfp);
 		if (!wq) {
diff --git a/drivers/gpu/drm/i915/i915_utils.c b/drivers/gpu/drm/i915/i915_utils.c
index 4c305d838016..f9e780dee9de 100644
--- a/drivers/gpu/drm/i915/i915_utils.c
+++ b/drivers/gpu/drm/i915/i915_utils.c
@@ -87,7 +87,7 @@ bool i915_error_injected(void)
 
 void cancel_timer(struct timer_list *t)
 {
-	if (!READ_ONCE(t->expires))
+	if (!timer_active(t))
 		return;
 
 	del_timer(t);
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
index 54773371e6bd..abd4dcd9f79c 100644
--- a/drivers/gpu/drm/i915/i915_utils.h
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -438,9 +438,14 @@ static inline void __add_taint_for_CI(unsigned int taint)
 void cancel_timer(struct timer_list *t);
 void set_timer_ms(struct timer_list *t, unsigned long timeout);
 
+static inline bool timer_active(const struct timer_list *t)
+{
+	return READ_ONCE(t->expires);
+}
+
 static inline bool timer_expired(const struct timer_list *t)
 {
-	return READ_ONCE(t->expires) && !timer_pending(t);
+	return timer_active(t) && !timer_pending(t);
 }
 
 /*
diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h
index 5b3a3c653454..a64adc8c883b 100644
--- a/drivers/gpu/drm/i915/i915_vma.h
+++ b/drivers/gpu/drm/i915/i915_vma.h
@@ -363,6 +363,21 @@ i915_vma_unpin_fence(struct i915_vma *vma)
 
 void i915_vma_parked(struct intel_gt *gt);
 
+static inline bool i915_vma_is_scanout(const struct i915_vma *vma)
+{
+	return test_bit(I915_VMA_SCANOUT_BIT, __i915_vma_flags(vma));
+}
+
+static inline void i915_vma_mark_scanout(struct i915_vma *vma)
+{
+	set_bit(I915_VMA_SCANOUT_BIT, __i915_vma_flags(vma));
+}
+
+static inline void i915_vma_clear_scanout(struct i915_vma *vma)
+{
+	clear_bit(I915_VMA_SCANOUT_BIT, __i915_vma_flags(vma));
+}
+
 #define for_each_until(cond) if (cond) break; else
 
 /**
diff --git a/drivers/gpu/drm/i915/i915_vma_types.h b/drivers/gpu/drm/i915/i915_vma_types.h
index 9e9082dc8f4b..f5cb848b7a7e 100644
--- a/drivers/gpu/drm/i915/i915_vma_types.h
+++ b/drivers/gpu/drm/i915/i915_vma_types.h
@@ -249,6 +249,9 @@ struct i915_vma {
 #define I915_VMA_USERFAULT	((int)BIT(I915_VMA_USERFAULT_BIT))
 #define I915_VMA_GGTT_WRITE	((int)BIT(I915_VMA_GGTT_WRITE_BIT))
 
+#define I915_VMA_SCANOUT_BIT	18
+#define I915_VMA_SCANOUT	((int)BIT(I915_VMA_SCANOUT_BIT))
+
 	struct i915_active active;
 
 #define I915_VMA_PAGES_BIAS 24
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
index ef767f04c37c..f2d5ae59081e 100644
--- a/drivers/gpu/drm/i915/intel_device_info.c
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -117,150 +117,6 @@ void intel_device_info_print_runtime(const struct intel_runtime_info *info,
 				     struct drm_printer *p)
 {
 	drm_printf(p, "rawclk rate: %u kHz\n", info->rawclk_freq);
-	drm_printf(p, "CS timestamp frequency: %u Hz\n",
-		   info->cs_timestamp_frequency_hz);
-}
-
-static u32 read_reference_ts_freq(struct drm_i915_private *dev_priv)
-{
-	u32 ts_override = intel_uncore_read(&dev_priv->uncore,
-					    GEN9_TIMESTAMP_OVERRIDE);
-	u32 base_freq, frac_freq;
-
-	base_freq = ((ts_override & GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_MASK) >>
-		     GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_SHIFT) + 1;
-	base_freq *= 1000000;
-
-	frac_freq = ((ts_override &
-		      GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_MASK) >>
-		     GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_SHIFT);
-	frac_freq = 1000000 / (frac_freq + 1);
-
-	return base_freq + frac_freq;
-}
-
-static u32 gen10_get_crystal_clock_freq(struct drm_i915_private *dev_priv,
-					u32 rpm_config_reg)
-{
-	u32 f19_2_mhz = 19200000;
-	u32 f24_mhz = 24000000;
-	u32 crystal_clock = (rpm_config_reg &
-			     GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >>
-			    GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT;
-
-	switch (crystal_clock) {
-	case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ:
-		return f19_2_mhz;
-	case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ:
-		return f24_mhz;
-	default:
-		MISSING_CASE(crystal_clock);
-		return 0;
-	}
-}
-
-static u32 gen11_get_crystal_clock_freq(struct drm_i915_private *dev_priv,
-					u32 rpm_config_reg)
-{
-	u32 f19_2_mhz = 19200000;
-	u32 f24_mhz = 24000000;
-	u32 f25_mhz = 25000000;
-	u32 f38_4_mhz = 38400000;
-	u32 crystal_clock = (rpm_config_reg &
-			     GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >>
-			    GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT;
-
-	switch (crystal_clock) {
-	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ:
-		return f24_mhz;
-	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ:
-		return f19_2_mhz;
-	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ:
-		return f38_4_mhz;
-	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ:
-		return f25_mhz;
-	default:
-		MISSING_CASE(crystal_clock);
-		return 0;
-	}
-}
-
-static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv)
-{
-	struct intel_uncore *uncore = &dev_priv->uncore;
-	u32 f12_5_mhz = 12500000;
-	u32 f19_2_mhz = 19200000;
-	u32 f24_mhz = 24000000;
-
-	if (INTEL_GEN(dev_priv) <= 4) {
-		/* PRMs say:
-		 *
-		 *     "The value in this register increments once every 16
-		 *      hclks." (through the “Clocking Configuration�
-		 *      (“CLKCFG�) MCHBAR register)
-		 */
-		return RUNTIME_INFO(dev_priv)->rawclk_freq * 1000 / 16;
-	} else if (INTEL_GEN(dev_priv) <= 8) {
-		/* PRMs say:
-		 *
-		 *     "The PCU TSC counts 10ns increments; this timestamp
-		 *      reflects bits 38:3 of the TSC (i.e. 80ns granularity,
-		 *      rolling over every 1.5 hours).
-		 */
-		return f12_5_mhz;
-	} else if (INTEL_GEN(dev_priv) <= 9) {
-		u32 ctc_reg = intel_uncore_read(uncore, CTC_MODE);
-		u32 freq = 0;
-
-		if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) {
-			freq = read_reference_ts_freq(dev_priv);
-		} else {
-			freq = IS_GEN9_LP(dev_priv) ? f19_2_mhz : f24_mhz;
-
-			/* Now figure out how the command stream's timestamp
-			 * register increments from this frequency (it might
-			 * increment only every few clock cycle).
-			 */
-			freq >>= 3 - ((ctc_reg & CTC_SHIFT_PARAMETER_MASK) >>
-				      CTC_SHIFT_PARAMETER_SHIFT);
-		}
-
-		return freq;
-	} else if (INTEL_GEN(dev_priv) <= 12) {
-		u32 ctc_reg = intel_uncore_read(uncore, CTC_MODE);
-		u32 freq = 0;
-
-		/* First figure out the reference frequency. There are 2 ways
-		 * we can compute the frequency, either through the
-		 * TIMESTAMP_OVERRIDE register or through RPM_CONFIG. CTC_MODE
-		 * tells us which one we should use.
-		 */
-		if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) {
-			freq = read_reference_ts_freq(dev_priv);
-		} else {
-			u32 rpm_config_reg = intel_uncore_read(uncore, RPM_CONFIG0);
-
-			if (INTEL_GEN(dev_priv) <= 10)
-				freq = gen10_get_crystal_clock_freq(dev_priv,
-								rpm_config_reg);
-			else
-				freq = gen11_get_crystal_clock_freq(dev_priv,
-								rpm_config_reg);
-
-			/* Now figure out how the command stream's timestamp
-			 * register increments from this frequency (it might
-			 * increment only every few clock cycle).
-			 */
-			freq >>= 3 - ((rpm_config_reg &
-				       GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK) >>
-				      GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT);
-		}
-
-		return freq;
-	}
-
-	MISSING_CASE("Unknown gen, unable to read command streamer timestamp frequency\n");
-	return 0;
 }
 
 #undef INTEL_VGA_DEVICE
@@ -505,19 +361,6 @@ void intel_device_info_runtime_init(struct drm_i915_private *dev_priv)
 	runtime->rawclk_freq = intel_read_rawclk(dev_priv);
 	drm_dbg(&dev_priv->drm, "rawclk rate: %d kHz\n", runtime->rawclk_freq);
 
-	/* Initialize command stream timestamp frequency */
-	runtime->cs_timestamp_frequency_hz =
-		read_timestamp_frequency(dev_priv);
-	if (runtime->cs_timestamp_frequency_hz) {
-		runtime->cs_timestamp_period_ns =
-			i915_cs_timestamp_ticks_to_ns(dev_priv, 1);
-		drm_dbg(&dev_priv->drm,
-			"CS timestamp wraparound in %lldms\n",
-			div_u64(mul_u32_u32(runtime->cs_timestamp_period_ns,
-					    S32_MAX),
-				USEC_PER_SEC));
-	}
-
 	if (!HAS_DISPLAY(dev_priv)) {
 		dev_priv->drm.driver_features &= ~(DRIVER_MODESET |
 						   DRIVER_ATOMIC);
diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h
index d92fa041c700..cf2d528c6e9b 100644
--- a/drivers/gpu/drm/i915/intel_device_info.h
+++ b/drivers/gpu/drm/i915/intel_device_info.h
@@ -123,7 +123,6 @@ enum intel_ppgtt_type {
 	func(has_llc); \
 	func(has_logical_ring_contexts); \
 	func(has_logical_ring_elsq); \
-	func(has_logical_ring_preemption); \
 	func(has_master_unit_irq); \
 	func(has_pooled_eu); \
 	func(has_rc6); \
@@ -224,9 +223,6 @@ struct intel_runtime_info {
 	u8 num_scalers[I915_MAX_PIPES];
 
 	u32 rawclk_freq;
-
-	u32 cs_timestamp_frequency_hz;
-	u32 cs_timestamp_period_ns;
 };
 
 struct intel_driver_caps {
diff --git a/drivers/gpu/drm/i915/intel_dram.c b/drivers/gpu/drm/i915/intel_dram.c
index 4754296a250e..73d256fc6830 100644
--- a/drivers/gpu/drm/i915/intel_dram.c
+++ b/drivers/gpu/drm/i915/intel_dram.c
@@ -5,6 +5,7 @@
 
 #include "i915_drv.h"
 #include "intel_dram.h"
+#include "intel_sideband.h"
 
 struct dram_dimm_info {
 	u16 size;
@@ -201,22 +202,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
 		return -EINVAL;
 	}
 
-	/*
-	 * If any of the channel is single rank channel, worst case output
-	 * will be same as if single rank memory, so consider single rank
-	 * memory.
-	 */
-	if (ch0.ranks == 1 || ch1.ranks == 1)
-		dram_info->ranks = 1;
-	else
-		dram_info->ranks = max(ch0.ranks, ch1.ranks);
-
-	if (dram_info->ranks == 0) {
+	if (ch0.ranks == 0 && ch1.ranks == 0) {
 		drm_info(&i915->drm, "couldn't get memory rank information\n");
 		return -EINVAL;
 	}
 
-	dram_info->is_16gb_dimm = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
+	dram_info->wm_lv_0_adjust_needed = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
 
 	dram_info->symmetric_memory = intel_is_dram_symmetric(&ch0, &ch1);
 
@@ -269,16 +260,12 @@ skl_get_dram_info(struct drm_i915_private *i915)
 	mem_freq_khz = DIV_ROUND_UP((val & SKL_REQ_DATA_MASK) *
 				    SKL_MEMORY_FREQ_MULTIPLIER_HZ, 1000);
 
-	dram_info->bandwidth_kbps = dram_info->num_channels *
-		mem_freq_khz * 8;
-
-	if (dram_info->bandwidth_kbps == 0) {
+	if (dram_info->num_channels * mem_freq_khz == 0) {
 		drm_info(&i915->drm,
 			 "Couldn't get system memory bandwidth\n");
 		return -EINVAL;
 	}
 
-	dram_info->valid = true;
 	return 0;
 }
 
@@ -365,7 +352,7 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
 	struct dram_info *dram_info = &i915->dram_info;
 	u32 dram_channels;
 	u32 mem_freq_khz, val;
-	u8 num_active_channels;
+	u8 num_active_channels, valid_ranks = 0;
 	int i;
 
 	val = intel_uncore_read(&i915->uncore, BXT_P_CR_MC_BIOS_REQ_0_0_0);
@@ -375,10 +362,7 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
 	dram_channels = val & BXT_DRAM_CHANNEL_ACTIVE_MASK;
 	num_active_channels = hweight32(dram_channels);
 
-	/* Each active bit represents 4-byte channel */
-	dram_info->bandwidth_kbps = (mem_freq_khz * num_active_channels * 4);
-
-	if (dram_info->bandwidth_kbps == 0) {
+	if (mem_freq_khz * num_active_channels == 0) {
 		drm_info(&i915->drm,
 			 "Couldn't get system memory bandwidth\n");
 		return -EINVAL;
@@ -410,57 +394,125 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
 			    dimm.size, dimm.width, dimm.ranks,
 			    intel_dram_type_str(type));
 
-		/*
-		 * If any of the channel is single rank channel,
-		 * worst case output will be same as if single rank
-		 * memory, so consider single rank memory.
-		 */
-		if (dram_info->ranks == 0)
-			dram_info->ranks = dimm.ranks;
-		else if (dimm.ranks == 1)
-			dram_info->ranks = 1;
+		if (valid_ranks == 0)
+			valid_ranks = dimm.ranks;
 
 		if (type != INTEL_DRAM_UNKNOWN)
 			dram_info->type = type;
 	}
 
-	if (dram_info->type == INTEL_DRAM_UNKNOWN || dram_info->ranks == 0) {
+	if (dram_info->type == INTEL_DRAM_UNKNOWN || valid_ranks == 0) {
 		drm_info(&i915->drm, "couldn't get memory information\n");
 		return -EINVAL;
 	}
 
-	dram_info->valid = true;
+	return 0;
+}
+
+static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
+{
+	struct dram_info *dram_info = &dev_priv->dram_info;
+	u32 val = 0;
+	int ret;
+
+	ret = sandybridge_pcode_read(dev_priv,
+				     ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
+				     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO,
+				     &val, NULL);
+	if (ret)
+		return ret;
+
+	if (IS_GEN(dev_priv, 12)) {
+		switch (val & 0xf) {
+		case 0:
+			dram_info->type = INTEL_DRAM_DDR4;
+			break;
+		case 3:
+			dram_info->type = INTEL_DRAM_LPDDR4;
+			break;
+		case 4:
+			dram_info->type = INTEL_DRAM_DDR3;
+			break;
+		case 5:
+			dram_info->type = INTEL_DRAM_LPDDR3;
+			break;
+		default:
+			MISSING_CASE(val & 0xf);
+			return -1;
+		}
+	} else {
+		switch (val & 0xf) {
+		case 0:
+			dram_info->type = INTEL_DRAM_DDR4;
+			break;
+		case 1:
+			dram_info->type = INTEL_DRAM_DDR3;
+			break;
+		case 2:
+			dram_info->type = INTEL_DRAM_LPDDR3;
+			break;
+		case 3:
+			dram_info->type = INTEL_DRAM_LPDDR4;
+			break;
+		default:
+			MISSING_CASE(val & 0xf);
+			return -1;
+		}
+	}
+
+	dram_info->num_channels = (val & 0xf0) >> 4;
+	dram_info->num_qgv_points = (val & 0xf00) >> 8;
 
 	return 0;
 }
 
+static int gen11_get_dram_info(struct drm_i915_private *i915)
+{
+	int ret = skl_get_dram_info(i915);
+
+	if (ret)
+		return ret;
+
+	return icl_pcode_read_mem_global_info(i915);
+}
+
+static int gen12_get_dram_info(struct drm_i915_private *i915)
+{
+	/* Always needed for GEN12+ */
+	i915->dram_info.wm_lv_0_adjust_needed = true;
+
+	return icl_pcode_read_mem_global_info(i915);
+}
+
 void intel_dram_detect(struct drm_i915_private *i915)
 {
 	struct dram_info *dram_info = &i915->dram_info;
 	int ret;
 
 	/*
-	 * Assume 16Gb DIMMs are present until proven otherwise.
-	 * This is only used for the level 0 watermark latency
-	 * w/a which does not apply to bxt/glk.
+	 * Assume level 0 watermark latency adjustment is needed until proven
+	 * otherwise, this w/a is not needed by bxt/glk.
 	 */
-	dram_info->is_16gb_dimm = !IS_GEN9_LP(i915);
+	dram_info->wm_lv_0_adjust_needed = !IS_GEN9_LP(i915);
 
 	if (INTEL_GEN(i915) < 9 || !HAS_DISPLAY(i915))
 		return;
 
-	if (IS_GEN9_LP(i915))
+	if (INTEL_GEN(i915) >= 12)
+		ret = gen12_get_dram_info(i915);
+	else if (INTEL_GEN(i915) >= 11)
+		ret = gen11_get_dram_info(i915);
+	else if (IS_GEN9_LP(i915))
 		ret = bxt_get_dram_info(i915);
 	else
 		ret = skl_get_dram_info(i915);
 	if (ret)
 		return;
 
-	drm_dbg_kms(&i915->drm, "DRAM bandwidth: %u kBps, channels: %u\n",
-		    dram_info->bandwidth_kbps, dram_info->num_channels);
+	drm_dbg_kms(&i915->drm, "DRAM channels: %u\n", dram_info->num_channels);
 
-	drm_dbg_kms(&i915->drm, "DRAM ranks: %u, 16Gb DIMMs: %s\n",
-		    dram_info->ranks, yesno(dram_info->is_16gb_dimm));
+	drm_dbg_kms(&i915->drm, "Watermark level 0 adjustment needed: %s\n",
+		    yesno(dram_info->wm_lv_0_adjust_needed));
 }
 
 static u32 gen9_edram_size_mb(struct drm_i915_private *i915, u32 cap)
diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c
index b326993a1026..1bfcdd89b241 100644
--- a/drivers/gpu/drm/i915/intel_memory_region.c
+++ b/drivers/gpu/drm/i915/intel_memory_region.c
@@ -10,7 +10,7 @@
 #define REGION_MAP(type, inst) \
 	BIT((type) + INTEL_MEMORY_TYPE_SHIFT) | BIT(inst)
 
-const u32 intel_region_map[] = {
+static const u32 intel_region_map[] = {
 	[INTEL_REGION_SMEM] = REGION_MAP(INTEL_MEMORY_SYSTEM, 0),
 	[INTEL_REGION_LMEM] = REGION_MAP(INTEL_MEMORY_LOCAL, 0),
 	[INTEL_REGION_STOLEN] = REGION_MAP(INTEL_MEMORY_STOLEN, 0),
diff --git a/drivers/gpu/drm/i915/intel_memory_region.h b/drivers/gpu/drm/i915/intel_memory_region.h
index 232490d89a83..6ffc0673f005 100644
--- a/drivers/gpu/drm/i915/intel_memory_region.h
+++ b/drivers/gpu/drm/i915/intel_memory_region.h
@@ -51,21 +51,16 @@ enum intel_region_id {
 	for (id = 0; id < ARRAY_SIZE((i915)->mm.regions); id++) \
 		for_each_if((mr) = (i915)->mm.regions[id])
 
-/**
- * Memory regions encoded as type | instance
- */
-extern const u32 intel_region_map[];
-
 struct intel_memory_region_ops {
 	unsigned int flags;
 
 	int (*init)(struct intel_memory_region *mem);
 	void (*release)(struct intel_memory_region *mem);
 
-	struct drm_i915_gem_object *
-	(*create_object)(struct intel_memory_region *mem,
-			 resource_size_t size,
-			 unsigned int flags);
+	int (*init_object)(struct intel_memory_region *mem,
+			   struct drm_i915_gem_object *obj,
+			   resource_size_t size,
+			   unsigned int flags);
 };
 
 struct intel_memory_region {
diff --git a/drivers/gpu/drm/i915/intel_pch.c b/drivers/gpu/drm/i915/intel_pch.c
index f31c0dabd0cc..ecaf314d60b6 100644
--- a/drivers/gpu/drm/i915/intel_pch.c
+++ b/drivers/gpu/drm/i915/intel_pch.c
@@ -143,8 +143,9 @@ static bool intel_is_virt_pch(unsigned short id,
 		 sdevice == PCI_SUBDEVICE_ID_QEMU));
 }
 
-static unsigned short
-intel_virt_detect_pch(const struct drm_i915_private *dev_priv)
+static void
+intel_virt_detect_pch(const struct drm_i915_private *dev_priv,
+		      unsigned short *pch_id, enum intel_pch *pch_type)
 {
 	unsigned short id = 0;
 
@@ -181,12 +182,21 @@ intel_virt_detect_pch(const struct drm_i915_private *dev_priv)
 	else
 		drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n");
 
-	return id;
+	*pch_type = intel_pch_type(dev_priv, id);
+
+	/* Sanity check virtual PCH id */
+	if (drm_WARN_ON(&dev_priv->drm,
+			id && *pch_type == PCH_NONE))
+		id = 0;
+
+	*pch_id = id;
 }
 
 void intel_detect_pch(struct drm_i915_private *dev_priv)
 {
 	struct pci_dev *pch = NULL;
+	unsigned short id;
+	enum intel_pch pch_type;
 
 	/* DG1 has south engine display on the same PCI device */
 	if (IS_DG1(dev_priv)) {
@@ -206,9 +216,6 @@ void intel_detect_pch(struct drm_i915_private *dev_priv)
 	 * of only checking the first one.
 	 */
 	while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) {
-		unsigned short id;
-		enum intel_pch pch_type;
-
 		if (pch->vendor != PCI_VENDOR_ID_INTEL)
 			continue;
 
@@ -221,14 +228,7 @@ void intel_detect_pch(struct drm_i915_private *dev_priv)
 			break;
 		} else if (intel_is_virt_pch(id, pch->subsystem_vendor,
 					     pch->subsystem_device)) {
-			id = intel_virt_detect_pch(dev_priv);
-			pch_type = intel_pch_type(dev_priv, id);
-
-			/* Sanity check virtual PCH id */
-			if (drm_WARN_ON(&dev_priv->drm,
-					id && pch_type == PCH_NONE))
-				id = 0;
-
+			intel_virt_detect_pch(dev_priv, &id, &pch_type);
 			dev_priv->pch_type = pch_type;
 			dev_priv->pch_id = id;
 			break;
@@ -244,10 +244,15 @@ void intel_detect_pch(struct drm_i915_private *dev_priv)
 			    "Display disabled, reverting to NOP PCH\n");
 		dev_priv->pch_type = PCH_NOP;
 		dev_priv->pch_id = 0;
+	} else if (!pch) {
+		if (run_as_guest() && HAS_DISPLAY(dev_priv)) {
+			intel_virt_detect_pch(dev_priv, &id, &pch_type);
+			dev_priv->pch_type = pch_type;
+			dev_priv->pch_id = id;
+		} else {
+			drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
+		}
 	}
 
-	if (!pch)
-		drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
-
 	pci_dev_put(pch);
 }
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index bbc73df7f753..0c3e63f27c29 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -2930,7 +2930,7 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv,
 		 * any underrun. If not able to get Dimm info assume 16GB dimm
 		 * to avoid any underrun.
 		 */
-		if (dev_priv->dram_info.is_16gb_dimm)
+		if (dev_priv->dram_info.wm_lv_0_adjust_needed)
 			wm[0] += 1;
 
 	} else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
@@ -4017,43 +4017,48 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	return 0;
 }
 
-/*
- * Calculate initial DBuf slice offset, based on slice size
- * and mask(i.e if slice size is 1024 and second slice is enabled
- * offset would be 1024)
- */
-static unsigned int
-icl_get_first_dbuf_slice_offset(u32 dbuf_slice_mask,
-				u32 slice_size,
-				u32 ddb_size)
+static int intel_dbuf_size(struct drm_i915_private *dev_priv)
 {
-	unsigned int offset = 0;
+	int ddb_size = INTEL_INFO(dev_priv)->ddb_size;
 
-	if (!dbuf_slice_mask)
-		return 0;
+	drm_WARN_ON(&dev_priv->drm, ddb_size == 0);
 
-	offset = (ffs(dbuf_slice_mask) - 1) * slice_size;
+	if (INTEL_GEN(dev_priv) < 11)
+		return ddb_size - 4; /* 4 blocks for bypass path allocation */
 
-	WARN_ON(offset >= ddb_size);
-	return offset;
+	return ddb_size;
 }
 
-u16 intel_get_ddb_size(struct drm_i915_private *dev_priv)
+static int intel_dbuf_slice_size(struct drm_i915_private *dev_priv)
 {
-	u16 ddb_size = INTEL_INFO(dev_priv)->ddb_size;
-	drm_WARN_ON(&dev_priv->drm, ddb_size == 0);
+	return intel_dbuf_size(dev_priv) /
+		INTEL_INFO(dev_priv)->num_supported_dbuf_slices;
+}
 
-	if (INTEL_GEN(dev_priv) < 11)
-		return ddb_size - 4; /* 4 blocks for bypass path allocation */
+static void
+skl_ddb_entry_for_slices(struct drm_i915_private *dev_priv, u8 slice_mask,
+			 struct skl_ddb_entry *ddb)
+{
+	int slice_size = intel_dbuf_slice_size(dev_priv);
 
-	return ddb_size;
+	if (!slice_mask) {
+		ddb->start = 0;
+		ddb->end = 0;
+		return;
+	}
+
+	ddb->start = (ffs(slice_mask) - 1) * slice_size;
+	ddb->end = fls(slice_mask) * slice_size;
+
+	WARN_ON(ddb->start >= ddb->end);
+	WARN_ON(ddb->end > intel_dbuf_size(dev_priv));
 }
 
 u32 skl_ddb_dbuf_slice_mask(struct drm_i915_private *dev_priv,
 			    const struct skl_ddb_entry *entry)
 {
 	u32 slice_mask = 0;
-	u16 ddb_size = intel_get_ddb_size(dev_priv);
+	u16 ddb_size = intel_dbuf_size(dev_priv);
 	u16 num_supported_slices = INTEL_INFO(dev_priv)->num_supported_dbuf_slices;
 	u16 slice_size = ddb_size / num_supported_slices;
 	u16 start_slice;
@@ -4077,116 +4082,40 @@ u32 skl_ddb_dbuf_slice_mask(struct drm_i915_private *dev_priv,
 	return slice_mask;
 }
 
-static u8 skl_compute_dbuf_slices(const struct intel_crtc_state *crtc_state,
-				  u8 active_pipes);
-
-static int
-skl_ddb_get_pipe_allocation_limits(struct drm_i915_private *dev_priv,
-				   const struct intel_crtc_state *crtc_state,
-				   const u64 total_data_rate,
-				   struct skl_ddb_entry *alloc, /* out */
-				   int *num_active /* out */)
-{
-	struct drm_atomic_state *state = crtc_state->uapi.state;
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_crtc *for_crtc = crtc_state->uapi.crtc;
-	const struct intel_crtc *crtc;
-	u32 pipe_width = 0, total_width_in_range = 0, width_before_pipe_in_range = 0;
-	enum pipe for_pipe = to_intel_crtc(for_crtc)->pipe;
-	struct intel_dbuf_state *new_dbuf_state =
-		intel_atomic_get_new_dbuf_state(intel_state);
-	const struct intel_dbuf_state *old_dbuf_state =
-		intel_atomic_get_old_dbuf_state(intel_state);
-	u8 active_pipes = new_dbuf_state->active_pipes;
-	u16 ddb_size;
-	u32 ddb_range_size;
-	u32 i;
-	u32 dbuf_slice_mask;
-	u32 offset;
-	u32 slice_size;
-	u32 total_slice_mask;
-	u32 start, end;
-	int ret;
-
-	*num_active = hweight8(active_pipes);
-
-	if (!crtc_state->hw.active) {
-		alloc->start = 0;
-		alloc->end = 0;
-		return 0;
-	}
-
-	ddb_size = intel_get_ddb_size(dev_priv);
-
-	slice_size = ddb_size / INTEL_INFO(dev_priv)->num_supported_dbuf_slices;
+static unsigned int intel_crtc_ddb_weight(const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode;
+	int hdisplay, vdisplay;
 
-	/*
-	 * If the state doesn't change the active CRTC's or there is no
-	 * modeset request, then there's no need to recalculate;
-	 * the existing pipe allocation limits should remain unchanged.
-	 * Note that we're safe from racing commits since any racing commit
-	 * that changes the active CRTC list or do modeset would need to
-	 * grab _all_ crtc locks, including the one we currently hold.
-	 */
-	if (old_dbuf_state->active_pipes == new_dbuf_state->active_pipes &&
-	    !dev_priv->wm.distrust_bios_wm) {
-		/*
-		 * alloc may be cleared by clear_intel_crtc_state,
-		 * copy from old state to be sure
-		 *
-		 * FIXME get rid of this mess
-		 */
-		*alloc = to_intel_crtc_state(for_crtc->state)->wm.skl.ddb;
+	if (!crtc_state->hw.active)
 		return 0;
-	}
-
-	/*
-	 * Get allowed DBuf slices for correspondent pipe and platform.
-	 */
-	dbuf_slice_mask = skl_compute_dbuf_slices(crtc_state, active_pipes);
-
-	/*
-	 * Figure out at which DBuf slice we start, i.e if we start at Dbuf S2
-	 * and slice size is 1024, the offset would be 1024
-	 */
-	offset = icl_get_first_dbuf_slice_offset(dbuf_slice_mask,
-						 slice_size, ddb_size);
-
-	/*
-	 * Figure out total size of allowed DBuf slices, which is basically
-	 * a number of allowed slices for that pipe multiplied by slice size.
-	 * Inside of this
-	 * range ddb entries are still allocated in proportion to display width.
-	 */
-	ddb_range_size = hweight8(dbuf_slice_mask) * slice_size;
 
 	/*
 	 * Watermark/ddb requirement highly depends upon width of the
 	 * framebuffer, So instead of allocating DDB equally among pipes
 	 * distribute DDB based on resolution/width of the display.
 	 */
-	total_slice_mask = dbuf_slice_mask;
-	for_each_new_intel_crtc_in_state(intel_state, crtc, crtc_state, i) {
-		const struct drm_display_mode *pipe_mode =
-			&crtc_state->hw.pipe_mode;
-		enum pipe pipe = crtc->pipe;
-		int hdisplay, vdisplay;
-		u32 pipe_dbuf_slice_mask;
+	drm_mode_get_hv_timing(pipe_mode, &hdisplay, &vdisplay);
 
-		if (!crtc_state->hw.active)
-			continue;
+	return hdisplay;
+}
 
-		pipe_dbuf_slice_mask = skl_compute_dbuf_slices(crtc_state,
-							       active_pipes);
+static void intel_crtc_dbuf_weights(const struct intel_dbuf_state *dbuf_state,
+				    enum pipe for_pipe,
+				    unsigned int *weight_start,
+				    unsigned int *weight_end,
+				    unsigned int *weight_total)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(dbuf_state->base.state->base.dev);
+	enum pipe pipe;
 
-		/*
-		 * According to BSpec pipe can share one dbuf slice with another
-		 * pipes or pipe can use multiple dbufs, in both cases we
-		 * account for other pipes only if they have exactly same mask.
-		 * However we need to account how many slices we should enable
-		 * in total.
-		 */
-		total_slice_mask |= pipe_dbuf_slice_mask;
+	*weight_start = 0;
+	*weight_end = 0;
+	*weight_total = 0;
+
+	for_each_pipe(dev_priv, pipe) {
+		int weight = dbuf_state->weight[pipe];
 
 		/*
 		 * Do not account pipes using other slice sets
@@ -4195,42 +4124,78 @@ skl_ddb_get_pipe_allocation_limits(struct drm_i915_private *dev_priv,
 		 * i.e no partial intersection), so it is enough to check for
 		 * equality for now.
 		 */
-		if (dbuf_slice_mask != pipe_dbuf_slice_mask)
+		if (dbuf_state->slices[pipe] != dbuf_state->slices[for_pipe])
 			continue;
 
-		drm_mode_get_hv_timing(pipe_mode, &hdisplay, &vdisplay);
+		*weight_total += weight;
+		if (pipe < for_pipe) {
+			*weight_start += weight;
+			*weight_end += weight;
+		} else if (pipe == for_pipe) {
+			*weight_end += weight;
+		}
+	}
+}
 
-		total_width_in_range += hdisplay;
+static int
+skl_crtc_allocate_ddb(struct intel_atomic_state *state, struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	unsigned int weight_total, weight_start, weight_end;
+	const struct intel_dbuf_state *old_dbuf_state =
+		intel_atomic_get_old_dbuf_state(state);
+	struct intel_dbuf_state *new_dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+	struct intel_crtc_state *crtc_state;
+	struct skl_ddb_entry ddb_slices;
+	enum pipe pipe = crtc->pipe;
+	u32 ddb_range_size;
+	u32 dbuf_slice_mask;
+	u32 start, end;
+	int ret;
 
-		if (pipe < for_pipe)
-			width_before_pipe_in_range += hdisplay;
-		else if (pipe == for_pipe)
-			pipe_width = hdisplay;
+	if (new_dbuf_state->weight[pipe] == 0) {
+		new_dbuf_state->ddb[pipe].start = 0;
+		new_dbuf_state->ddb[pipe].end = 0;
+		goto out;
 	}
 
-	/*
-	 * FIXME: For now we always enable slice S1 as per
-	 * the Bspec display initialization sequence.
-	 */
-	new_dbuf_state->enabled_slices = total_slice_mask | BIT(DBUF_S1);
+	dbuf_slice_mask = new_dbuf_state->slices[pipe];
 
-	if (old_dbuf_state->enabled_slices != new_dbuf_state->enabled_slices) {
-		ret = intel_atomic_serialize_global_state(&new_dbuf_state->base);
-		if (ret)
-			return ret;
-	}
+	skl_ddb_entry_for_slices(dev_priv, dbuf_slice_mask, &ddb_slices);
+	ddb_range_size = skl_ddb_entry_size(&ddb_slices);
+
+	intel_crtc_dbuf_weights(new_dbuf_state, pipe,
+				&weight_start, &weight_end, &weight_total);
+
+	start = ddb_range_size * weight_start / weight_total;
+	end = ddb_range_size * weight_end / weight_total;
+
+	new_dbuf_state->ddb[pipe].start = ddb_slices.start + start;
+	new_dbuf_state->ddb[pipe].end = ddb_slices.start + end;
+
+out:
+	if (skl_ddb_entry_equal(&old_dbuf_state->ddb[pipe],
+				&new_dbuf_state->ddb[pipe]))
+		return 0;
+
+	ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
+	if (ret)
+		return ret;
 
-	start = ddb_range_size * width_before_pipe_in_range / total_width_in_range;
-	end = ddb_range_size *
-		(width_before_pipe_in_range + pipe_width) / total_width_in_range;
+	crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+	if (IS_ERR(crtc_state))
+		return PTR_ERR(crtc_state);
 
-	alloc->start = offset + start;
-	alloc->end = offset + end;
+	crtc_state->wm.skl.ddb = new_dbuf_state->ddb[pipe];
 
 	drm_dbg_kms(&dev_priv->drm,
-		    "[CRTC:%d:%s] dbuf slices 0x%x, ddb (%d - %d), active pipes 0x%x\n",
-		    for_crtc->base.id, for_crtc->name,
-		    dbuf_slice_mask, alloc->start, alloc->end, active_pipes);
+		    "[CRTC:%d:%s] dbuf slices 0x%x -> 0x%x, ddb (%d - %d) -> (%d - %d), active pipes 0x%x -> 0x%x\n",
+		    crtc->base.base.id, crtc->base.name,
+		    old_dbuf_state->slices[pipe], new_dbuf_state->slices[pipe],
+		    old_dbuf_state->ddb[pipe].start, old_dbuf_state->ddb[pipe].end,
+		    new_dbuf_state->ddb[pipe].start, new_dbuf_state->ddb[pipe].end,
+		    old_dbuf_state->active_pipes, new_dbuf_state->active_pipes);
 
 	return 0;
 }
@@ -4632,10 +4597,8 @@ static u8 tgl_compute_dbuf_slices(enum pipe pipe, u8 active_pipes)
 	return compute_dbuf_slices(pipe, active_pipes, tgl_allowed_dbufs);
 }
 
-static u8 skl_compute_dbuf_slices(const struct intel_crtc_state *crtc_state,
-				  u8 active_pipes)
+static u8 skl_compute_dbuf_slices(struct intel_crtc *crtc, u8 active_pipes)
 {
-	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
 	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	enum pipe pipe = crtc->pipe;
 
@@ -4798,55 +4761,30 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 }
 
 static int
-skl_allocate_pipe_ddb(struct intel_atomic_state *state,
-		      struct intel_crtc *crtc)
+skl_allocate_plane_ddb(struct intel_atomic_state *state,
+		       struct intel_crtc *crtc)
 {
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_crtc_state *crtc_state =
 		intel_atomic_get_new_crtc_state(state, crtc);
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	struct skl_ddb_entry *alloc = &crtc_state->wm.skl.ddb;
+	const struct intel_dbuf_state *dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+	const struct skl_ddb_entry *alloc = &dbuf_state->ddb[crtc->pipe];
+	int num_active = hweight8(dbuf_state->active_pipes);
 	u16 alloc_size, start = 0;
 	u16 total[I915_MAX_PLANES] = {};
 	u16 uv_total[I915_MAX_PLANES] = {};
 	u64 total_data_rate;
 	enum plane_id plane_id;
-	int num_active;
 	u32 blocks;
 	int level;
-	int ret;
 
 	/* Clear the partitioning for disabled planes. */
 	memset(crtc_state->wm.skl.plane_ddb_y, 0, sizeof(crtc_state->wm.skl.plane_ddb_y));
 	memset(crtc_state->wm.skl.plane_ddb_uv, 0, sizeof(crtc_state->wm.skl.plane_ddb_uv));
 
-	if (!crtc_state->hw.active) {
-		struct intel_atomic_state *state =
-			to_intel_atomic_state(crtc_state->uapi.state);
-		struct intel_dbuf_state *new_dbuf_state =
-			intel_atomic_get_new_dbuf_state(state);
-		const struct intel_dbuf_state *old_dbuf_state =
-			intel_atomic_get_old_dbuf_state(state);
-
-		/*
-		 * FIXME hack to make sure we compute this sensibly when
-		 * turning off all the pipes. Otherwise we leave it at
-		 * whatever we had previously, and then runtime PM will
-		 * mess it up by turning off all but S1. Remove this
-		 * once the dbuf state computation flow becomes sane.
-		 */
-		if (new_dbuf_state->active_pipes == 0) {
-			new_dbuf_state->enabled_slices = BIT(DBUF_S1);
-
-			if (old_dbuf_state->enabled_slices != new_dbuf_state->enabled_slices) {
-				ret = intel_atomic_serialize_global_state(&new_dbuf_state->base);
-				if (ret)
-					return ret;
-			}
-		}
-
-		alloc->start = alloc->end = 0;
+	if (!crtc_state->hw.active)
 		return 0;
-	}
 
 	if (INTEL_GEN(dev_priv) >= 11)
 		total_data_rate =
@@ -4855,12 +4793,6 @@ skl_allocate_pipe_ddb(struct intel_atomic_state *state,
 		total_data_rate =
 			skl_get_total_relative_data_rate(state, crtc);
 
-	ret = skl_ddb_get_pipe_allocation_limits(dev_priv, crtc_state,
-						 total_data_rate,
-						 alloc, &num_active);
-	if (ret)
-		return ret;
-
 	alloc_size = skl_ddb_entry_size(alloc);
 	if (alloc_size == 0)
 		return 0;
@@ -5731,6 +5663,18 @@ static bool skl_ddb_entries_overlap(const struct skl_ddb_entry *a,
 	return a->start < b->end && b->start < a->end;
 }
 
+static void skl_ddb_entry_union(struct skl_ddb_entry *a,
+				const struct skl_ddb_entry *b)
+{
+	if (a->end && b->end) {
+		a->start = min(a->start, b->start);
+		a->end = max(a->end, b->end);
+	} else if (b->end) {
+		a->start = b->start;
+		a->end = b->end;
+	}
+}
+
 bool skl_ddb_allocation_overlaps(const struct skl_ddb_entry *ddb,
 				 const struct skl_ddb_entry *entries,
 				 int num_entries, int ignore_idx)
@@ -5775,39 +5719,114 @@ skl_ddb_add_affected_planes(const struct intel_crtc_state *old_crtc_state,
 	return 0;
 }
 
+static u8 intel_dbuf_enabled_slices(const struct intel_dbuf_state *dbuf_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(dbuf_state->base.state->base.dev);
+	u8 enabled_slices;
+	enum pipe pipe;
+
+	/*
+	 * FIXME: For now we always enable slice S1 as per
+	 * the Bspec display initialization sequence.
+	 */
+	enabled_slices = BIT(DBUF_S1);
+
+	for_each_pipe(dev_priv, pipe)
+		enabled_slices |= dbuf_state->slices[pipe];
+
+	return enabled_slices;
+}
+
 static int
 skl_compute_ddb(struct intel_atomic_state *state)
 {
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	const struct intel_dbuf_state *old_dbuf_state;
-	const struct intel_dbuf_state *new_dbuf_state;
+	struct intel_dbuf_state *new_dbuf_state = NULL;
 	const struct intel_crtc_state *old_crtc_state;
 	struct intel_crtc_state *new_crtc_state;
 	struct intel_crtc *crtc;
 	int ret, i;
 
-	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
-					    new_crtc_state, i) {
-		ret = skl_allocate_pipe_ddb(state, crtc);
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		new_dbuf_state = intel_atomic_get_dbuf_state(state);
+		if (IS_ERR(new_dbuf_state))
+			return PTR_ERR(new_dbuf_state);
+
+		old_dbuf_state = intel_atomic_get_old_dbuf_state(state);
+		break;
+	}
+
+	if (!new_dbuf_state)
+		return 0;
+
+	new_dbuf_state->active_pipes =
+		intel_calc_active_pipes(state, old_dbuf_state->active_pipes);
+
+	if (old_dbuf_state->active_pipes != new_dbuf_state->active_pipes) {
+		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
 		if (ret)
 			return ret;
+	}
 
-		ret = skl_ddb_add_affected_planes(old_crtc_state,
-						  new_crtc_state);
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		enum pipe pipe = crtc->pipe;
+
+		new_dbuf_state->slices[pipe] =
+			skl_compute_dbuf_slices(crtc, new_dbuf_state->active_pipes);
+
+		if (old_dbuf_state->slices[pipe] == new_dbuf_state->slices[pipe])
+			continue;
+
+		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
 		if (ret)
 			return ret;
 	}
 
-	old_dbuf_state = intel_atomic_get_old_dbuf_state(state);
-	new_dbuf_state = intel_atomic_get_new_dbuf_state(state);
+	new_dbuf_state->enabled_slices = intel_dbuf_enabled_slices(new_dbuf_state);
+
+	if (old_dbuf_state->enabled_slices != new_dbuf_state->enabled_slices) {
+		ret = intel_atomic_serialize_global_state(&new_dbuf_state->base);
+		if (ret)
+			return ret;
 
-	if (new_dbuf_state &&
-	    new_dbuf_state->enabled_slices != old_dbuf_state->enabled_slices)
 		drm_dbg_kms(&dev_priv->drm,
 			    "Enabled dbuf slices 0x%x -> 0x%x (out of %d dbuf slices)\n",
 			    old_dbuf_state->enabled_slices,
 			    new_dbuf_state->enabled_slices,
 			    INTEL_INFO(dev_priv)->num_supported_dbuf_slices);
+	}
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		enum pipe pipe = crtc->pipe;
+
+		new_dbuf_state->weight[pipe] = intel_crtc_ddb_weight(new_crtc_state);
+
+		if (old_dbuf_state->weight[pipe] == new_dbuf_state->weight[pipe])
+			continue;
+
+		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
+		if (ret)
+			return ret;
+	}
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		ret = skl_crtc_allocate_ddb(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		ret = skl_allocate_plane_ddb(state, crtc);
+		if (ret)
+			return ret;
+
+		ret = skl_ddb_add_affected_planes(old_crtc_state,
+						  new_crtc_state);
+		if (ret)
+			return ret;
+	}
 
 	return 0;
 }
@@ -5944,83 +5963,6 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 	}
 }
 
-static int intel_add_affected_pipes(struct intel_atomic_state *state,
-				    u8 pipe_mask)
-{
-	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
-	struct intel_crtc *crtc;
-
-	for_each_intel_crtc(&dev_priv->drm, crtc) {
-		struct intel_crtc_state *crtc_state;
-
-		if ((pipe_mask & BIT(crtc->pipe)) == 0)
-			continue;
-
-		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
-		if (IS_ERR(crtc_state))
-			return PTR_ERR(crtc_state);
-	}
-
-	return 0;
-}
-
-static int
-skl_ddb_add_affected_pipes(struct intel_atomic_state *state)
-{
-	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
-	struct intel_crtc_state *crtc_state;
-	struct intel_crtc *crtc;
-	int i, ret;
-
-	if (dev_priv->wm.distrust_bios_wm) {
-		/*
-		 * skl_ddb_get_pipe_allocation_limits() currently requires
-		 * all active pipes to be included in the state so that
-		 * it can redistribute the dbuf among them, and it really
-		 * wants to recompute things when distrust_bios_wm is set
-		 * so we add all the pipes to the state.
-		 */
-		ret = intel_add_affected_pipes(state, ~0);
-		if (ret)
-			return ret;
-	}
-
-	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
-		struct intel_dbuf_state *new_dbuf_state;
-		const struct intel_dbuf_state *old_dbuf_state;
-
-		new_dbuf_state = intel_atomic_get_dbuf_state(state);
-		if (IS_ERR(new_dbuf_state))
-			return PTR_ERR(new_dbuf_state);
-
-		old_dbuf_state = intel_atomic_get_old_dbuf_state(state);
-
-		new_dbuf_state->active_pipes =
-			intel_calc_active_pipes(state, old_dbuf_state->active_pipes);
-
-		if (old_dbuf_state->active_pipes == new_dbuf_state->active_pipes)
-			break;
-
-		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
-		if (ret)
-			return ret;
-
-		/*
-		 * skl_ddb_get_pipe_allocation_limits() currently requires
-		 * all active pipes to be included in the state so that
-		 * it can redistribute the dbuf among them.
-		 */
-		ret = intel_add_affected_pipes(state,
-					       new_dbuf_state->active_pipes);
-		if (ret)
-			return ret;
-
-		break;
-	}
-
-	return 0;
-}
-
 /*
  * To make sure the cursor watermark registers are always consistent
  * with our computed state the following scenario needs special
@@ -6088,15 +6030,6 @@ skl_compute_wm(struct intel_atomic_state *state)
 	struct intel_crtc_state *new_crtc_state;
 	int ret, i;
 
-	ret = skl_ddb_add_affected_pipes(state);
-	if (ret)
-		return ret;
-
-	/*
-	 * Calculate WM's for all pipes that are part of this transaction.
-	 * Note that skl_ddb_add_affected_pipes may have added more CRTC's that
-	 * weren't otherwise being modified if pipe allocations had to change.
-	 */
 	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
 		ret = skl_build_pipe_wm(state, crtc);
 		if (ret)
@@ -6255,20 +6188,49 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 
 void skl_wm_get_hw_state(struct drm_i915_private *dev_priv)
 {
+	struct intel_dbuf_state *dbuf_state =
+		to_intel_dbuf_state(dev_priv->dbuf.obj.state);
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *crtc_state;
 
 	for_each_intel_crtc(&dev_priv->drm, crtc) {
-		crtc_state = to_intel_crtc_state(crtc->base.state);
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		enum pipe pipe = crtc->pipe;
+		enum plane_id plane_id;
 
 		skl_pipe_wm_get_hw_state(crtc, &crtc_state->wm.skl.optimal);
 		crtc_state->wm.skl.raw = crtc_state->wm.skl.optimal;
-	}
 
-	if (dev_priv->active_pipes) {
-		/* Fully recompute DDB on first atomic commit */
-		dev_priv->wm.distrust_bios_wm = true;
+		memset(&dbuf_state->ddb[pipe], 0, sizeof(dbuf_state->ddb[pipe]));
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			struct skl_ddb_entry *ddb_y =
+				&crtc_state->wm.skl.plane_ddb_y[plane_id];
+			struct skl_ddb_entry *ddb_uv =
+				&crtc_state->wm.skl.plane_ddb_uv[plane_id];
+
+			skl_ddb_get_hw_plane_state(dev_priv, crtc->pipe,
+						   plane_id, ddb_y, ddb_uv);
+
+			skl_ddb_entry_union(&dbuf_state->ddb[pipe], ddb_y);
+			skl_ddb_entry_union(&dbuf_state->ddb[pipe], ddb_uv);
+		}
+
+		dbuf_state->slices[pipe] =
+			skl_compute_dbuf_slices(crtc, dbuf_state->active_pipes);
+
+		dbuf_state->weight[pipe] = intel_crtc_ddb_weight(crtc_state);
+
+		crtc_state->wm.skl.ddb = dbuf_state->ddb[pipe];
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "[CRTC:%d:%s] dbuf slices 0x%x, ddb (%d - %d), active pipes 0x%x\n",
+			    crtc->base.base.id, crtc->base.name,
+			    dbuf_state->slices[pipe], dbuf_state->ddb[pipe].start,
+			    dbuf_state->ddb[pipe].end, dbuf_state->active_pipes);
 	}
+
+	dbuf_state->enabled_slices = dev_priv->dbuf.enabled_slices;
 }
 
 static void ilk_pipe_wm_get_hw_state(struct intel_crtc *crtc)
@@ -7103,24 +7065,26 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
 			 0, CNL_DELAY_PMRSP);
 }
 
-static void tgl_init_clock_gating(struct drm_i915_private *dev_priv)
+static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
 {
-	/* Wa_1409120013:tgl */
+	/* Wa_1409120013:tgl,rkl,adl_s,dg1 */
 	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
-		   ILK_DPFC_CHICKEN_COMP_DUMMY_PIXEL);
+			   ILK_DPFC_CHICKEN_COMP_DUMMY_PIXEL);
 
 	/* Wa_1409825376:tgl (pre-prod)*/
 	if (IS_TGL_DISP_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_B1))
 		intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_3) |
 			   TGL_VRH_GATING_DIS);
 
-	/* Wa_14011059788:tgl */
+	/* Wa_14011059788:tgl,rkl,adl_s,dg1 */
 	intel_uncore_rmw(&dev_priv->uncore, GEN10_DFR_RATIO_EN_AND_CHICKEN,
 			 0, DFR_DISABLE);
 }
 
 static void dg1_init_clock_gating(struct drm_i915_private *dev_priv)
 {
+	gen12lp_init_clock_gating(dev_priv);
+
 	/* Wa_1409836686:dg1[a0] */
 	if (IS_DG1_REVID(dev_priv, DG1_REVID_A0, DG1_REVID_A0))
 		intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_3) |
@@ -7583,7 +7547,7 @@ void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv)
 	if (IS_DG1(dev_priv))
 		dev_priv->display.init_clock_gating = dg1_init_clock_gating;
 	else if (IS_GEN(dev_priv, 12))
-		dev_priv->display.init_clock_gating = tgl_init_clock_gating;
+		dev_priv->display.init_clock_gating = gen12lp_init_clock_gating;
 	else if (IS_GEN(dev_priv, 11))
 		dev_priv->display.init_clock_gating = icl_init_clock_gating;
 	else if (IS_CANNONLAKE(dev_priv))
diff --git a/drivers/gpu/drm/i915/intel_pm.h b/drivers/gpu/drm/i915/intel_pm.h
index eab83e251dd5..97550cf0b6df 100644
--- a/drivers/gpu/drm/i915/intel_pm.h
+++ b/drivers/gpu/drm/i915/intel_pm.h
@@ -9,8 +9,10 @@
 #include <linux/types.h>
 
 #include "display/intel_bw.h"
+#include "display/intel_display.h"
 #include "display/intel_global_state.h"
 
+#include "i915_drv.h"
 #include "i915_reg.h"
 
 struct drm_device;
@@ -40,7 +42,6 @@ void skl_pipe_ddb_get_hw_state(struct intel_crtc *crtc,
 			       struct skl_ddb_entry *ddb_y,
 			       struct skl_ddb_entry *ddb_uv);
 void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv);
-u16 intel_get_ddb_size(struct drm_i915_private *dev_priv);
 u32 skl_ddb_dbuf_slice_mask(struct drm_i915_private *dev_priv,
 			    const struct skl_ddb_entry *entry);
 void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
@@ -69,6 +70,10 @@ bool intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable);
 struct intel_dbuf_state {
 	struct intel_global_state base;
 
+	struct skl_ddb_entry ddb[I915_MAX_PIPES];
+	unsigned int weight[I915_MAX_PIPES];
+	u8 slices[I915_MAX_PIPES];
+
 	u8 enabled_slices;
 	u8 active_pipes;
 };
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem.c b/drivers/gpu/drm/i915/selftests/i915_gem.c
index 412e21604a05..dc394fb7ccfa 100644
--- a/drivers/gpu/drm/i915/selftests/i915_gem.c
+++ b/drivers/gpu/drm/i915/selftests/i915_gem.c
@@ -8,6 +8,7 @@
 
 #include "gem/selftests/igt_gem_utils.h"
 #include "gem/selftests/mock_context.h"
+#include "gem/i915_gem_pm.h"
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_pm.h"
 
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
index f88473d396f4..f99bb0113726 100644
--- a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
@@ -38,8 +38,8 @@ static void quirk_add(struct drm_i915_gem_object *obj,
 		      struct list_head *objects)
 {
 	/* quirk is only for live tiled objects, use it to declare ownership */
-	GEM_BUG_ON(obj->mm.quirked);
-	obj->mm.quirked = true;
+	GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+	i915_gem_object_set_tiling_quirk(obj);
 	list_add(&obj->st_link, objects);
 }
 
@@ -85,7 +85,7 @@ static void unpin_ggtt(struct i915_ggtt *ggtt)
 	struct i915_vma *vma;
 
 	list_for_each_entry(vma, &ggtt->vm.bound_list, vm_link)
-		if (vma->obj->mm.quirked)
+		if (i915_gem_object_has_tiling_quirk(vma->obj))
 			i915_vma_unpin(vma);
 }
 
@@ -94,8 +94,8 @@ static void cleanup_objects(struct i915_ggtt *ggtt, struct list_head *list)
 	struct drm_i915_gem_object *obj, *on;
 
 	list_for_each_entry_safe(obj, on, list, st_link) {
-		GEM_BUG_ON(!obj->mm.quirked);
-		obj->mm.quirked = false;
+		GEM_BUG_ON(!i915_gem_object_has_tiling_quirk(obj));
+		i915_gem_object_set_tiling_quirk(obj);
 		i915_gem_object_put(obj);
 	}
 
@@ -442,28 +442,22 @@ static int igt_evict_contexts(void *arg)
 	/* Overfill the GGTT with context objects and so try to evict one. */
 	for_each_engine(engine, gt, id) {
 		struct i915_sw_fence fence;
-		struct file *file;
-
-		file = mock_file(i915);
-		if (IS_ERR(file)) {
-			err = PTR_ERR(file);
-			break;
-		}
 
 		count = 0;
 		onstack_fence_init(&fence);
 		do {
+			struct intel_context *ce;
 			struct i915_request *rq;
-			struct i915_gem_context *ctx;
 
-			ctx = live_context(i915, file);
-			if (IS_ERR(ctx))
+			ce = intel_context_create(engine);
+			if (IS_ERR(ce))
 				break;
 
 			/* We will need some GGTT space for the rq's context */
 			igt_evict_ctl.fail_if_busy = true;
-			rq = igt_request_alloc(ctx, engine);
+			rq = intel_context_create_request(ce);
 			igt_evict_ctl.fail_if_busy = false;
+			intel_context_put(ce);
 
 			if (IS_ERR(rq)) {
 				/* When full, fail_if_busy will trigger EBUSY */
@@ -490,8 +484,6 @@ static int igt_evict_contexts(void *arg)
 		onstack_fence_fini(&fence);
 		pr_info("Submitted %lu contexts/requests on %s\n",
 			count, engine->name);
-
-		fput(file);
 		if (err)
 			break;
 	}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
index c53a222e3dec..c1adea8765a9 100644
--- a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
@@ -28,6 +28,7 @@
 #include "gem/i915_gem_context.h"
 #include "gem/selftests/mock_context.h"
 #include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
 
 #include "i915_random.h"
 #include "i915_selftest.h"
@@ -1880,7 +1881,7 @@ static int igt_cs_tlb(void *arg)
 	vma = i915_vma_instance(out, vm, NULL);
 	if (IS_ERR(vma)) {
 		err = PTR_ERR(vma);
-		goto out_put_batch;
+		goto out_put_out;
 	}
 
 	err = i915_vma_pin(vma, 0, 0,
diff --git a/drivers/gpu/drm/i915/selftests/i915_perf.c b/drivers/gpu/drm/i915/selftests/i915_perf.c
index debbac660519..e9d86dab8677 100644
--- a/drivers/gpu/drm/i915/selftests/i915_perf.c
+++ b/drivers/gpu/drm/i915/selftests/i915_perf.c
@@ -262,7 +262,7 @@ static int live_noa_delay(void *arg)
 
 	delay = intel_read_status_page(stream->engine, 0x102);
 	delay -= intel_read_status_page(stream->engine, 0x100);
-	delay = i915_cs_timestamp_ticks_to_ns(i915, delay);
+	delay = intel_gt_clock_interval_to_ns(stream->engine->gt, delay);
 	pr_info("GPU delay: %uns, expected %lluns\n",
 		delay, expected);
 
diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c
index e424a6d1a68c..d2a678a2497e 100644
--- a/drivers/gpu/drm/i915/selftests/i915_request.c
+++ b/drivers/gpu/drm/i915/selftests/i915_request.c
@@ -33,6 +33,7 @@
 #include "gt/intel_engine_pm.h"
 #include "gt/intel_engine_user.h"
 #include "gt/intel_gt.h"
+#include "gt/intel_gt_clock_utils.h"
 #include "gt/intel_gt_requests.h"
 #include "gt/selftest_engine_heartbeat.h"
 
@@ -1560,7 +1561,7 @@ static u32 trifilter(u32 *a)
 
 static u64 cycles_to_ns(struct intel_engine_cs *engine, u32 cycles)
 {
-	u64 ns = i915_cs_timestamp_ticks_to_ns(engine->i915, cycles);
+	u64 ns = intel_gt_clock_interval_to_ns(engine->gt, cycles);
 
 	return DIV_ROUND_CLOSEST(ns, 1 << TF_BIAS);
 }
@@ -1932,9 +1933,7 @@ static int measure_inter_request(struct intel_context *ce)
 		intel_ring_advance(rq, cs);
 		i915_request_add(rq);
 	}
-	local_bh_disable();
 	i915_sw_fence_commit(submit);
-	local_bh_enable();
 	intel_engine_flush_submission(ce->engine);
 	heap_fence_put(submit);
 
@@ -2220,11 +2219,9 @@ static int measure_completion(struct intel_context *ce)
 		intel_ring_advance(rq, cs);
 
 		dma_fence_add_callback(&rq->fence, &cb.base, signal_cb);
-
-		local_bh_disable();
 		i915_request_add(rq);
-		local_bh_enable();
 
+		intel_engine_flush_submission(ce->engine);
 		if (wait_for(READ_ONCE(sema[i]) == -1, 50)) {
 			err = -EIO;
 			goto err;
diff --git a/drivers/gpu/drm/i915/selftests/igt_spinner.c b/drivers/gpu/drm/i915/selftests/igt_spinner.c
index ec0ecb4e4ca6..83f6e5f31fb3 100644
--- a/drivers/gpu/drm/i915/selftests/igt_spinner.c
+++ b/drivers/gpu/drm/i915/selftests/igt_spinner.c
@@ -3,6 +3,7 @@
  *
  * Copyright © 2018 Intel Corporation
  */
+#include "gt/intel_gpu_commands.h"
 #include "gt/intel_gt.h"
 
 #include "gem/selftests/igt_gem_utils.h"
@@ -219,6 +220,9 @@ void igt_spinner_fini(struct igt_spinner *spin)
 
 bool igt_wait_for_spinner(struct igt_spinner *spin, struct i915_request *rq)
 {
+	if (i915_request_is_ready(rq))
+		intel_engine_flush_submission(rq->engine);
+
 	return !(wait_for_us(i915_seqno_passed(hws_seqno(spin, rq),
 					       rq->fence.seqno),
 			     100) &&
diff --git a/drivers/gpu/drm/i915/selftests/intel_memory_region.c b/drivers/gpu/drm/i915/selftests/intel_memory_region.c
index 0aeba8e3af28..ce7adfa3bca0 100644
--- a/drivers/gpu/drm/i915/selftests/intel_memory_region.c
+++ b/drivers/gpu/drm/i915/selftests/intel_memory_region.c
@@ -129,6 +129,21 @@ static void igt_object_release(struct drm_i915_gem_object *obj)
 	i915_gem_object_put(obj);
 }
 
+static bool is_contiguous(struct drm_i915_gem_object *obj)
+{
+	struct scatterlist *sg;
+	dma_addr_t addr = -1;
+
+	for (sg = obj->mm.pages->sgl; sg; sg = sg_next(sg)) {
+		if (addr != -1 && sg_dma_address(sg) != addr)
+			return false;
+
+		addr = sg_dma_address(sg) + sg_dma_len(sg);
+	}
+
+	return true;
+}
+
 static int igt_mock_contiguous(void *arg)
 {
 	struct intel_memory_region *mem = arg;
@@ -150,8 +165,8 @@ static int igt_mock_contiguous(void *arg)
 	if (IS_ERR(obj))
 		return PTR_ERR(obj);
 
-	if (obj->mm.pages->nents != 1) {
-		pr_err("%s min object spans multiple sg entries\n", __func__);
+	if (!is_contiguous(obj)) {
+		pr_err("%s min object spans disjoint sg entries\n", __func__);
 		err = -EINVAL;
 		goto err_close_objects;
 	}
@@ -163,8 +178,8 @@ static int igt_mock_contiguous(void *arg)
 	if (IS_ERR(obj))
 		return PTR_ERR(obj);
 
-	if (obj->mm.pages->nents != 1) {
-		pr_err("%s max object spans multiple sg entries\n", __func__);
+	if (!is_contiguous(obj)) {
+		pr_err("%s max object spans disjoint sg entries\n", __func__);
 		err = -EINVAL;
 		goto err_close_objects;
 	}
@@ -189,8 +204,8 @@ static int igt_mock_contiguous(void *arg)
 		goto err_close_objects;
 	}
 
-	if (obj->mm.pages->nents != 1) {
-		pr_err("%s object spans multiple sg entries\n", __func__);
+	if (!is_contiguous(obj)) {
+		pr_err("%s object spans disjoint sg entries\n", __func__);
 		err = -EINVAL;
 		goto err_close_objects;
 	}
@@ -337,6 +352,68 @@ out_put:
 	return err;
 }
 
+#ifndef SZ_8G
+#define SZ_8G BIT_ULL(33)
+#endif
+
+static int igt_mock_max_segment(void *arg)
+{
+	const unsigned int max_segment = i915_sg_segment_size();
+	struct intel_memory_region *mem = arg;
+	struct drm_i915_private *i915 = mem->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_buddy_block *block;
+	struct scatterlist *sg;
+	LIST_HEAD(objects);
+	u64 size;
+	int err = 0;
+
+	/*
+	 * While we may create very large contiguous blocks, we may need
+	 * to break those down for consumption elsewhere. In particular,
+	 * dma-mapping with scatterlist elements have an implicit limit of
+	 * UINT_MAX on each element.
+	 */
+
+	size = SZ_8G;
+	mem = mock_region_create(i915, 0, size, PAGE_SIZE, 0);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+
+	obj = igt_object_create(mem, &objects, size, 0);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_put;
+	}
+
+	size = 0;
+	list_for_each_entry(block, &obj->mm.blocks, link) {
+		if (i915_buddy_block_size(&mem->mm, block) > size)
+			size = i915_buddy_block_size(&mem->mm, block);
+	}
+	if (size < max_segment) {
+		pr_err("%s: Failed to create a huge contiguous block [> %u], largest block %lld\n",
+		       __func__, max_segment, size);
+		err = -EINVAL;
+		goto out_close;
+	}
+
+	for (sg = obj->mm.pages->sgl; sg; sg = sg_next(sg)) {
+		if (sg->length > max_segment) {
+			pr_err("%s: Created an oversized scatterlist entry, %u > %u\n",
+			       __func__, sg->length, max_segment);
+			err = -EINVAL;
+			goto out_close;
+		}
+	}
+
+out_close:
+	close_objects(mem, &objects);
+out_put:
+	intel_memory_region_put(mem);
+	return err;
+}
+
 static int igt_gpu_write_dw(struct intel_context *ce,
 			    struct i915_vma *vma,
 			    u32 dword,
@@ -775,14 +852,22 @@ static int _perf_memcpy(struct intel_memory_region *src_mr,
 		}
 
 		sort(t, ARRAY_SIZE(t), sizeof(*t), wrap_ktime_compare, NULL);
+		if (t[0] <= 0) {
+			/* ignore the impossible to protect our sanity */
+			pr_debug("Skipping %s src(%s, %s) -> dst(%s, %s) %14s %4lluKiB copy, unstable measurement [%lld, %lld]\n",
+				 __func__,
+				 src_mr->name, repr_type(src_type),
+				 dst_mr->name, repr_type(dst_type),
+				 tests[i].name, size >> 10,
+				 t[0], t[4]);
+			continue;
+		}
+
 		pr_info("%s src(%s, %s) -> dst(%s, %s) %14s %4llu KiB copy: %5lld MiB/s\n",
 			__func__,
-			src_mr->name,
-			repr_type(src_type),
-			dst_mr->name,
-			repr_type(dst_type),
-			tests[i].name,
-			size >> 10,
+			src_mr->name, repr_type(src_type),
+			dst_mr->name, repr_type(dst_type),
+			tests[i].name, size >> 10,
 			div64_u64(mul_u32_u32(4 * size,
 					      1000 * 1000 * 1000),
 				  t[1] + 2 * t[2] + t[3]) >> 20);
@@ -848,6 +933,7 @@ int intel_memory_region_mock_selftests(void)
 		SUBTEST(igt_mock_fill),
 		SUBTEST(igt_mock_contiguous),
 		SUBTEST(igt_mock_splintered_region),
+		SUBTEST(igt_mock_max_segment),
 	};
 	struct intel_memory_region *mem;
 	struct drm_i915_private *i915;
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
index e946bd2087d8..0188f877cab2 100644
--- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
@@ -64,8 +64,6 @@ static void mock_device_release(struct drm_device *dev)
 	mock_device_flush(i915);
 	intel_gt_driver_remove(&i915->gt);
 
-	i915_gem_driver_release__contexts(i915);
-
 	i915_gem_drain_workqueue(i915);
 	i915_gem_drain_freed_objects(i915);
 
diff --git a/drivers/gpu/drm/i915/selftests/mock_region.c b/drivers/gpu/drm/i915/selftests/mock_region.c
index 979d96f27c43..3c6021415274 100644
--- a/drivers/gpu/drm/i915/selftests/mock_region.c
+++ b/drivers/gpu/drm/i915/selftests/mock_region.c
@@ -15,21 +15,16 @@ static const struct drm_i915_gem_object_ops mock_region_obj_ops = {
 	.release = i915_gem_object_release_memory_region,
 };
 
-static struct drm_i915_gem_object *
-mock_object_create(struct intel_memory_region *mem,
-		   resource_size_t size,
-		   unsigned int flags)
+static int mock_object_init(struct intel_memory_region *mem,
+			    struct drm_i915_gem_object *obj,
+			    resource_size_t size,
+			    unsigned int flags)
 {
 	static struct lock_class_key lock_class;
 	struct drm_i915_private *i915 = mem->i915;
-	struct drm_i915_gem_object *obj;
 
 	if (size > mem->mm.size)
-		return ERR_PTR(-E2BIG);
-
-	obj = i915_gem_object_alloc();
-	if (!obj)
-		return ERR_PTR(-ENOMEM);
+		return -E2BIG;
 
 	drm_gem_private_object_init(&i915->drm, &obj->base, size);
 	i915_gem_object_init(obj, &mock_region_obj_ops, &lock_class);
@@ -40,13 +35,13 @@ mock_object_create(struct intel_memory_region *mem,
 
 	i915_gem_object_init_memory_region(obj, mem, flags);
 
-	return obj;
+	return 0;
 }
 
 static const struct intel_memory_region_ops mock_region_ops = {
 	.init = intel_memory_region_init_buddy,
 	.release = intel_memory_region_release_buddy,
-	.create_object = mock_object_create,
+	.init_object = mock_object_init,
 };
 
 struct intel_memory_region *
diff --git a/drivers/gpu/drm/imx/Kconfig b/drivers/gpu/drm/imx/Kconfig
index 73fe2bc5633c..b5fa0e45a839 100644
--- a/drivers/gpu/drm/imx/Kconfig
+++ b/drivers/gpu/drm/imx/Kconfig
@@ -37,7 +37,7 @@ config DRM_IMX_LDB
 config DRM_IMX_HDMI
 	tristate "Freescale i.MX DRM HDMI"
 	select DRM_DW_HDMI
-	depends on DRM_IMX
+	depends on DRM_IMX && OF
 	help
 	  Choose this if you want to use HDMI on i.MX6.
 
diff --git a/drivers/gpu/drm/mediatek/Makefile b/drivers/gpu/drm/mediatek/Makefile
index a892edec5563..dc54a7a69005 100644
--- a/drivers/gpu/drm/mediatek/Makefile
+++ b/drivers/gpu/drm/mediatek/Makefile
@@ -1,10 +1,11 @@
 # SPDX-License-Identifier: GPL-2.0
 
-mediatek-drm-y := mtk_disp_color.o \
+mediatek-drm-y := mtk_disp_ccorr.o \
+		  mtk_disp_color.o \
+		  mtk_disp_gamma.o \
 		  mtk_disp_ovl.o \
 		  mtk_disp_rdma.o \
 		  mtk_drm_crtc.o \
-		  mtk_drm_ddp.o \
 		  mtk_drm_ddp_comp.o \
 		  mtk_drm_drv.o \
 		  mtk_drm_gem.o \
diff --git a/drivers/gpu/drm/mediatek/mtk_disp_ccorr.c b/drivers/gpu/drm/mediatek/mtk_disp_ccorr.c
new file mode 100644
index 000000000000..141cb36b9c07
--- /dev/null
+++ b/drivers/gpu/drm/mediatek/mtk_disp_ccorr.c
@@ -0,0 +1,223 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/of_irq.h>
+#include <linux/platform_device.h>
+#include <linux/soc/mediatek/mtk-cmdq.h>
+
+#include "mtk_disp_drv.h"
+#include "mtk_drm_crtc.h"
+#include "mtk_drm_ddp_comp.h"
+
+#define DISP_CCORR_EN				0x0000
+#define CCORR_EN					BIT(0)
+#define DISP_CCORR_CFG				0x0020
+#define CCORR_RELAY_MODE				BIT(0)
+#define CCORR_ENGINE_EN					BIT(1)
+#define CCORR_GAMMA_OFF					BIT(2)
+#define CCORR_WGAMUT_SRC_CLIP				BIT(3)
+#define DISP_CCORR_SIZE				0x0030
+#define DISP_CCORR_COEF_0			0x0080
+#define DISP_CCORR_COEF_1			0x0084
+#define DISP_CCORR_COEF_2			0x0088
+#define DISP_CCORR_COEF_3			0x008C
+#define DISP_CCORR_COEF_4			0x0090
+
+struct mtk_disp_ccorr_data {
+	u32 matrix_bits;
+};
+
+/**
+ * struct mtk_disp_ccorr - DISP_CCORR driver structure
+ * @ddp_comp - structure containing type enum and hardware resources
+ * @crtc - associated crtc to report irq events to
+ */
+struct mtk_disp_ccorr {
+	struct clk *clk;
+	void __iomem *regs;
+	struct cmdq_client_reg cmdq_reg;
+	const struct mtk_disp_ccorr_data	*data;
+};
+
+int mtk_ccorr_clk_enable(struct device *dev)
+{
+	struct mtk_disp_ccorr *ccorr = dev_get_drvdata(dev);
+
+	return clk_prepare_enable(ccorr->clk);
+}
+
+void mtk_ccorr_clk_disable(struct device *dev)
+{
+	struct mtk_disp_ccorr *ccorr = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(ccorr->clk);
+}
+
+void mtk_ccorr_config(struct device *dev, unsigned int w,
+			     unsigned int h, unsigned int vrefresh,
+			     unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+{
+	struct mtk_disp_ccorr *ccorr = dev_get_drvdata(dev);
+
+	mtk_ddp_write(cmdq_pkt, w << 16 | h, &ccorr->cmdq_reg, ccorr->regs,
+		      DISP_CCORR_SIZE);
+	mtk_ddp_write(cmdq_pkt, CCORR_ENGINE_EN, &ccorr->cmdq_reg, ccorr->regs,
+		      DISP_CCORR_CFG);
+}
+
+void mtk_ccorr_start(struct device *dev)
+{
+	struct mtk_disp_ccorr *ccorr = dev_get_drvdata(dev);
+
+	writel(CCORR_EN, ccorr->regs + DISP_CCORR_EN);
+}
+
+void mtk_ccorr_stop(struct device *dev)
+{
+	struct mtk_disp_ccorr *ccorr = dev_get_drvdata(dev);
+
+	writel_relaxed(0x0, ccorr->regs + DISP_CCORR_EN);
+}
+
+/* Converts a DRM S31.32 value to the HW S1.n format. */
+static u16 mtk_ctm_s31_32_to_s1_n(u64 in, u32 n)
+{
+	u16 r;
+
+	/* Sign bit. */
+	r = in & BIT_ULL(63) ? BIT(n + 1) : 0;
+
+	if ((in & GENMASK_ULL(62, 33)) > 0) {
+		/* identity value 0x100000000 -> 0x400(mt8183), */
+		/* identity value 0x100000000 -> 0x800(mt8192), */
+		/* if bigger this, set it to max 0x7ff. */
+		r |= GENMASK(n, 0);
+	} else {
+		/* take the n+1 most important bits. */
+		r |= (in >> (32 - n)) & GENMASK(n, 0);
+	}
+
+	return r;
+}
+
+void mtk_ccorr_ctm_set(struct device *dev, struct drm_crtc_state *state)
+{
+	struct mtk_disp_ccorr *ccorr = dev_get_drvdata(dev);
+	struct drm_property_blob *blob = state->ctm;
+	struct drm_color_ctm *ctm;
+	const u64 *input;
+	uint16_t coeffs[9] = { 0 };
+	int i;
+	struct cmdq_pkt *cmdq_pkt = NULL;
+	u32 matrix_bits = ccorr->data->matrix_bits;
+
+	if (!blob)
+		return;
+
+	ctm = (struct drm_color_ctm *)blob->data;
+	input = ctm->matrix;
+
+	for (i = 0; i < ARRAY_SIZE(coeffs); i++)
+		coeffs[i] = mtk_ctm_s31_32_to_s1_n(input[i], matrix_bits);
+
+	mtk_ddp_write(cmdq_pkt, coeffs[0] << 16 | coeffs[1],
+		      &ccorr->cmdq_reg, ccorr->regs, DISP_CCORR_COEF_0);
+	mtk_ddp_write(cmdq_pkt, coeffs[2] << 16 | coeffs[3],
+		      &ccorr->cmdq_reg, ccorr->regs, DISP_CCORR_COEF_1);
+	mtk_ddp_write(cmdq_pkt, coeffs[4] << 16 | coeffs[5],
+		      &ccorr->cmdq_reg, ccorr->regs, DISP_CCORR_COEF_2);
+	mtk_ddp_write(cmdq_pkt, coeffs[6] << 16 | coeffs[7],
+		      &ccorr->cmdq_reg, ccorr->regs, DISP_CCORR_COEF_3);
+	mtk_ddp_write(cmdq_pkt, coeffs[8] << 16,
+		      &ccorr->cmdq_reg, ccorr->regs, DISP_CCORR_COEF_4);
+}
+
+static int mtk_disp_ccorr_bind(struct device *dev, struct device *master,
+			       void *data)
+{
+	return 0;
+}
+
+static void mtk_disp_ccorr_unbind(struct device *dev, struct device *master,
+				  void *data)
+{
+}
+
+static const struct component_ops mtk_disp_ccorr_component_ops = {
+	.bind	= mtk_disp_ccorr_bind,
+	.unbind	= mtk_disp_ccorr_unbind,
+};
+
+static int mtk_disp_ccorr_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct mtk_disp_ccorr *priv;
+	struct resource *res;
+	int ret;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to get ccorr clk\n");
+		return PTR_ERR(priv->clk);
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	priv->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(priv->regs)) {
+		dev_err(dev, "failed to ioremap ccorr\n");
+		return PTR_ERR(priv->regs);
+	}
+
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	ret = cmdq_dev_get_client_reg(dev, &priv->cmdq_reg, 0);
+	if (ret)
+		dev_dbg(dev, "get mediatek,gce-client-reg fail!\n");
+#endif
+
+	priv->data = of_device_get_match_data(dev);
+	platform_set_drvdata(pdev, priv);
+
+	ret = component_add(dev, &mtk_disp_ccorr_component_ops);
+	if (ret)
+		dev_err(dev, "Failed to add component: %d\n", ret);
+
+	return ret;
+}
+
+static int mtk_disp_ccorr_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &mtk_disp_ccorr_component_ops);
+
+	return 0;
+}
+
+static const struct mtk_disp_ccorr_data mt8183_ccorr_driver_data = {
+	.matrix_bits = 10,
+};
+
+static const struct of_device_id mtk_disp_ccorr_driver_dt_match[] = {
+	{ .compatible = "mediatek,mt8183-disp-ccorr",
+	  .data = &mt8183_ccorr_driver_data},
+	{},
+};
+MODULE_DEVICE_TABLE(of, mtk_disp_ccorr_driver_dt_match);
+
+struct platform_driver mtk_disp_ccorr_driver = {
+	.probe		= mtk_disp_ccorr_probe,
+	.remove		= mtk_disp_ccorr_remove,
+	.driver		= {
+		.name	= "mediatek-disp-ccorr",
+		.owner	= THIS_MODULE,
+		.of_match_table = mtk_disp_ccorr_driver_dt_match,
+	},
+};
diff --git a/drivers/gpu/drm/mediatek/mtk_disp_color.c b/drivers/gpu/drm/mediatek/mtk_disp_color.c
index 6048cbc9f0ec..63f411ab393b 100644
--- a/drivers/gpu/drm/mediatek/mtk_disp_color.c
+++ b/drivers/gpu/drm/mediatek/mtk_disp_color.c
@@ -11,6 +11,7 @@
 #include <linux/platform_device.h>
 #include <linux/soc/mediatek/mtk-cmdq.h>
 
+#include "mtk_disp_drv.h"
 #include "mtk_drm_crtc.h"
 #include "mtk_drm_ddp_comp.h"
 
@@ -36,64 +37,55 @@ struct mtk_disp_color_data {
  * @data: platform colour driver data
  */
 struct mtk_disp_color {
-	struct mtk_ddp_comp			ddp_comp;
 	struct drm_crtc				*crtc;
+	struct clk				*clk;
+	void __iomem				*regs;
+	struct cmdq_client_reg			cmdq_reg;
 	const struct mtk_disp_color_data	*data;
 };
 
-static inline struct mtk_disp_color *comp_to_color(struct mtk_ddp_comp *comp)
+int mtk_color_clk_enable(struct device *dev)
 {
-	return container_of(comp, struct mtk_disp_color, ddp_comp);
+	struct mtk_disp_color *color = dev_get_drvdata(dev);
+
+	return clk_prepare_enable(color->clk);
 }
 
-static void mtk_color_config(struct mtk_ddp_comp *comp, unsigned int w,
-			     unsigned int h, unsigned int vrefresh,
-			     unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+void mtk_color_clk_disable(struct device *dev)
 {
-	struct mtk_disp_color *color = comp_to_color(comp);
+	struct mtk_disp_color *color = dev_get_drvdata(dev);
 
-	mtk_ddp_write(cmdq_pkt, w, comp, DISP_COLOR_WIDTH(color));
-	mtk_ddp_write(cmdq_pkt, h, comp, DISP_COLOR_HEIGHT(color));
+	clk_disable_unprepare(color->clk);
 }
 
-static void mtk_color_start(struct mtk_ddp_comp *comp)
+void mtk_color_config(struct device *dev, unsigned int w,
+		      unsigned int h, unsigned int vrefresh,
+		      unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
 {
-	struct mtk_disp_color *color = comp_to_color(comp);
+	struct mtk_disp_color *color = dev_get_drvdata(dev);
 
-	writel(COLOR_BYPASS_ALL | COLOR_SEQ_SEL,
-	       comp->regs + DISP_COLOR_CFG_MAIN);
-	writel(0x1, comp->regs + DISP_COLOR_START(color));
+	mtk_ddp_write(cmdq_pkt, w, &color->cmdq_reg, color->regs, DISP_COLOR_WIDTH(color));
+	mtk_ddp_write(cmdq_pkt, h, &color->cmdq_reg, color->regs, DISP_COLOR_HEIGHT(color));
 }
 
-static const struct mtk_ddp_comp_funcs mtk_disp_color_funcs = {
-	.config = mtk_color_config,
-	.start = mtk_color_start,
-};
+void mtk_color_start(struct device *dev)
+{
+	struct mtk_disp_color *color = dev_get_drvdata(dev);
+
+	writel(COLOR_BYPASS_ALL | COLOR_SEQ_SEL,
+	       color->regs + DISP_COLOR_CFG_MAIN);
+	writel(0x1, color->regs + DISP_COLOR_START(color));
+}
 
 static int mtk_disp_color_bind(struct device *dev, struct device *master,
 			       void *data)
 {
-	struct mtk_disp_color *priv = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
-	int ret;
-
-	ret = mtk_ddp_comp_register(drm_dev, &priv->ddp_comp);
-	if (ret < 0) {
-		dev_err(dev, "Failed to register component %pOF: %d\n",
-			dev->of_node, ret);
-		return ret;
-	}
-
 	return 0;
 }
 
 static void mtk_disp_color_unbind(struct device *dev, struct device *master,
 				  void *data)
 {
-	struct mtk_disp_color *priv = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
-
-	mtk_ddp_comp_unregister(drm_dev, &priv->ddp_comp);
 }
 
 static const struct component_ops mtk_disp_color_component_ops = {
@@ -105,31 +97,32 @@ static int mtk_disp_color_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct mtk_disp_color *priv;
-	int comp_id;
+	struct resource *res;
 	int ret;
 
 	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv)
 		return -ENOMEM;
 
-	comp_id = mtk_ddp_comp_get_id(dev->of_node, MTK_DISP_COLOR);
-	if (comp_id < 0) {
-		dev_err(dev, "Failed to identify by alias: %d\n", comp_id);
-		return comp_id;
+	priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to get color clk\n");
+		return PTR_ERR(priv->clk);
 	}
 
-	ret = mtk_ddp_comp_init(dev, dev->of_node, &priv->ddp_comp, comp_id,
-				&mtk_disp_color_funcs);
-	if (ret) {
-		if (ret != -EPROBE_DEFER)
-			dev_err(dev, "Failed to initialize component: %d\n",
-				ret);
-
-		return ret;
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	priv->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(priv->regs)) {
+		dev_err(dev, "failed to ioremap color\n");
+		return PTR_ERR(priv->regs);
 	}
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	ret = cmdq_dev_get_client_reg(dev, &priv->cmdq_reg, 0);
+	if (ret)
+		dev_dbg(dev, "get mediatek,gce-client-reg fail!\n");
+#endif
 
 	priv->data = of_device_get_match_data(dev);
-
 	platform_set_drvdata(pdev, priv);
 
 	ret = component_add(dev, &mtk_disp_color_component_ops);
@@ -141,8 +134,6 @@ static int mtk_disp_color_probe(struct platform_device *pdev)
 
 static int mtk_disp_color_remove(struct platform_device *pdev)
 {
-	component_del(&pdev->dev, &mtk_disp_color_component_ops);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/mediatek/mtk_disp_drv.h b/drivers/gpu/drm/mediatek/mtk_disp_drv.h
new file mode 100644
index 000000000000..cafd9df2d63b
--- /dev/null
+++ b/drivers/gpu/drm/mediatek/mtk_disp_drv.h
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2020 MediaTek Inc.
+ */
+
+#ifndef _MTK_DISP_DRV_H_
+#define _MTK_DISP_DRV_H_
+
+#include <linux/soc/mediatek/mtk-cmdq.h>
+#include "mtk_drm_plane.h"
+
+void mtk_ccorr_ctm_set(struct device *dev, struct drm_crtc_state *state);
+int mtk_ccorr_clk_enable(struct device *dev);
+void mtk_ccorr_clk_disable(struct device *dev);
+void mtk_ccorr_config(struct device *dev, unsigned int w,
+		      unsigned int h, unsigned int vrefresh,
+		      unsigned int bpc, struct cmdq_pkt *cmdq_pkt);
+void mtk_ccorr_start(struct device *dev);
+void mtk_ccorr_stop(struct device *dev);
+
+void mtk_color_bypass_shadow(struct device *dev);
+int mtk_color_clk_enable(struct device *dev);
+void mtk_color_clk_disable(struct device *dev);
+void mtk_color_config(struct device *dev, unsigned int w,
+		      unsigned int h, unsigned int vrefresh,
+		      unsigned int bpc, struct cmdq_pkt *cmdq_pkt);
+void mtk_color_start(struct device *dev);
+
+void mtk_dither_set_common(void __iomem *regs, struct cmdq_client_reg *cmdq_reg,
+			   unsigned int bpc, unsigned int cfg,
+			   unsigned int dither_en, struct cmdq_pkt *cmdq_pkt);
+
+void mtk_dpi_start(struct device *dev);
+void mtk_dpi_stop(struct device *dev);
+
+void mtk_dsi_ddp_start(struct device *dev);
+void mtk_dsi_ddp_stop(struct device *dev);
+
+int mtk_gamma_clk_enable(struct device *dev);
+void mtk_gamma_clk_disable(struct device *dev);
+void mtk_gamma_config(struct device *dev, unsigned int w,
+		      unsigned int h, unsigned int vrefresh,
+		      unsigned int bpc, struct cmdq_pkt *cmdq_pkt);
+void mtk_gamma_set(struct device *dev, struct drm_crtc_state *state);
+void mtk_gamma_set_common(void __iomem *regs, struct drm_crtc_state *state);
+void mtk_gamma_start(struct device *dev);
+void mtk_gamma_stop(struct device *dev);
+
+void mtk_ovl_bgclr_in_on(struct device *dev);
+void mtk_ovl_bgclr_in_off(struct device *dev);
+void mtk_ovl_bypass_shadow(struct device *dev);
+int mtk_ovl_clk_enable(struct device *dev);
+void mtk_ovl_clk_disable(struct device *dev);
+void mtk_ovl_config(struct device *dev, unsigned int w,
+		    unsigned int h, unsigned int vrefresh,
+		    unsigned int bpc, struct cmdq_pkt *cmdq_pkt);
+int mtk_ovl_layer_check(struct device *dev, unsigned int idx,
+			struct mtk_plane_state *mtk_state);
+void mtk_ovl_layer_config(struct device *dev, unsigned int idx,
+			  struct mtk_plane_state *state,
+			  struct cmdq_pkt *cmdq_pkt);
+unsigned int mtk_ovl_layer_nr(struct device *dev);
+void mtk_ovl_layer_on(struct device *dev, unsigned int idx,
+		      struct cmdq_pkt *cmdq_pkt);
+void mtk_ovl_layer_off(struct device *dev, unsigned int idx,
+		       struct cmdq_pkt *cmdq_pkt);
+void mtk_ovl_start(struct device *dev);
+void mtk_ovl_stop(struct device *dev);
+unsigned int mtk_ovl_supported_rotations(struct device *dev);
+void mtk_ovl_enable_vblank(struct device *dev,
+			   void (*vblank_cb)(void *),
+			   void *vblank_cb_data);
+void mtk_ovl_disable_vblank(struct device *dev);
+
+void mtk_rdma_bypass_shadow(struct device *dev);
+int mtk_rdma_clk_enable(struct device *dev);
+void mtk_rdma_clk_disable(struct device *dev);
+void mtk_rdma_config(struct device *dev, unsigned int width,
+		     unsigned int height, unsigned int vrefresh,
+		     unsigned int bpc, struct cmdq_pkt *cmdq_pkt);
+unsigned int mtk_rdma_layer_nr(struct device *dev);
+void mtk_rdma_layer_config(struct device *dev, unsigned int idx,
+			   struct mtk_plane_state *state,
+			   struct cmdq_pkt *cmdq_pkt);
+void mtk_rdma_start(struct device *dev);
+void mtk_rdma_stop(struct device *dev);
+void mtk_rdma_enable_vblank(struct device *dev,
+			    void (*vblank_cb)(void *),
+			    void *vblank_cb_data);
+void mtk_rdma_disable_vblank(struct device *dev);
+
+#endif
diff --git a/drivers/gpu/drm/mediatek/mtk_disp_gamma.c b/drivers/gpu/drm/mediatek/mtk_disp_gamma.c
new file mode 100644
index 000000000000..3ebf91e0ab41
--- /dev/null
+++ b/drivers/gpu/drm/mediatek/mtk_disp_gamma.c
@@ -0,0 +1,197 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/of_irq.h>
+#include <linux/platform_device.h>
+#include <linux/soc/mediatek/mtk-cmdq.h>
+
+#include "mtk_disp_drv.h"
+#include "mtk_drm_crtc.h"
+#include "mtk_drm_ddp_comp.h"
+
+#define DISP_GAMMA_EN				0x0000
+#define GAMMA_EN					BIT(0)
+#define DISP_GAMMA_CFG				0x0020
+#define GAMMA_LUT_EN					BIT(1)
+#define GAMMA_DITHERING					BIT(2)
+#define DISP_GAMMA_SIZE				0x0030
+#define DISP_GAMMA_LUT				0x0700
+
+#define LUT_10BIT_MASK				0x03ff
+
+struct mtk_disp_gamma_data {
+	bool has_dither;
+};
+
+/**
+ * struct mtk_disp_gamma - DISP_GAMMA driver structure
+ * @ddp_comp - structure containing type enum and hardware resources
+ * @crtc - associated crtc to report irq events to
+ */
+struct mtk_disp_gamma {
+	struct clk *clk;
+	void __iomem *regs;
+	struct cmdq_client_reg cmdq_reg;
+	const struct mtk_disp_gamma_data *data;
+};
+
+int mtk_gamma_clk_enable(struct device *dev)
+{
+	struct mtk_disp_gamma *gamma = dev_get_drvdata(dev);
+
+	return clk_prepare_enable(gamma->clk);
+}
+
+void mtk_gamma_clk_disable(struct device *dev)
+{
+	struct mtk_disp_gamma *gamma = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(gamma->clk);
+}
+
+void mtk_gamma_set_common(void __iomem *regs, struct drm_crtc_state *state)
+{
+	unsigned int i, reg;
+	struct drm_color_lut *lut;
+	void __iomem *lut_base;
+	u32 word;
+
+	if (state->gamma_lut) {
+		reg = readl(regs + DISP_GAMMA_CFG);
+		reg = reg | GAMMA_LUT_EN;
+		writel(reg, regs + DISP_GAMMA_CFG);
+		lut_base = regs + DISP_GAMMA_LUT;
+		lut = (struct drm_color_lut *)state->gamma_lut->data;
+		for (i = 0; i < MTK_LUT_SIZE; i++) {
+			word = (((lut[i].red >> 6) & LUT_10BIT_MASK) << 20) +
+				(((lut[i].green >> 6) & LUT_10BIT_MASK) << 10) +
+				((lut[i].blue >> 6) & LUT_10BIT_MASK);
+			writel(word, (lut_base + i * 4));
+		}
+	}
+}
+
+void mtk_gamma_set(struct device *dev, struct drm_crtc_state *state)
+{
+	struct mtk_disp_gamma *gamma = dev_get_drvdata(dev);
+
+	mtk_gamma_set_common(gamma->regs, state);
+}
+
+void mtk_gamma_config(struct device *dev, unsigned int w,
+		      unsigned int h, unsigned int vrefresh,
+		      unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+{
+	struct mtk_disp_gamma *gamma = dev_get_drvdata(dev);
+
+	mtk_ddp_write(cmdq_pkt, h << 16 | w, &gamma->cmdq_reg, gamma->regs,
+		      DISP_GAMMA_SIZE);
+	if (gamma->data && gamma->data->has_dither)
+		mtk_dither_set_common(gamma->regs, &gamma->cmdq_reg, bpc,
+				      DISP_GAMMA_CFG, GAMMA_DITHERING, cmdq_pkt);
+}
+
+void mtk_gamma_start(struct device *dev)
+{
+	struct mtk_disp_gamma *gamma = dev_get_drvdata(dev);
+
+	writel(GAMMA_EN, gamma->regs + DISP_GAMMA_EN);
+}
+
+void mtk_gamma_stop(struct device *dev)
+{
+	struct mtk_disp_gamma *gamma = dev_get_drvdata(dev);
+
+	writel_relaxed(0x0, gamma->regs + DISP_GAMMA_EN);
+}
+
+static int mtk_disp_gamma_bind(struct device *dev, struct device *master,
+			       void *data)
+{
+	return 0;
+}
+
+static void mtk_disp_gamma_unbind(struct device *dev, struct device *master,
+				  void *data)
+{
+}
+
+static const struct component_ops mtk_disp_gamma_component_ops = {
+	.bind	= mtk_disp_gamma_bind,
+	.unbind = mtk_disp_gamma_unbind,
+};
+
+static int mtk_disp_gamma_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct mtk_disp_gamma *priv;
+	struct resource *res;
+	int ret;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to get gamma clk\n");
+		return PTR_ERR(priv->clk);
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	priv->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(priv->regs)) {
+		dev_err(dev, "failed to ioremap gamma\n");
+		return PTR_ERR(priv->regs);
+	}
+
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	ret = cmdq_dev_get_client_reg(dev, &priv->cmdq_reg, 0);
+	if (ret)
+		dev_dbg(dev, "get mediatek,gce-client-reg fail!\n");
+#endif
+
+	priv->data = of_device_get_match_data(dev);
+	platform_set_drvdata(pdev, priv);
+
+	ret = component_add(dev, &mtk_disp_gamma_component_ops);
+	if (ret)
+		dev_err(dev, "Failed to add component: %d\n", ret);
+
+	return ret;
+}
+
+static int mtk_disp_gamma_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &mtk_disp_gamma_component_ops);
+
+	return 0;
+}
+
+static const struct mtk_disp_gamma_data mt8173_gamma_driver_data = {
+	.has_dither = true,
+};
+
+static const struct of_device_id mtk_disp_gamma_driver_dt_match[] = {
+	{ .compatible = "mediatek,mt8173-disp-gamma",
+	  .data = &mt8173_gamma_driver_data},
+	{ .compatible = "mediatek,mt8183-disp-gamma"},
+	{},
+};
+MODULE_DEVICE_TABLE(of, mtk_disp_gamma_driver_dt_match);
+
+struct platform_driver mtk_disp_gamma_driver = {
+	.probe		= mtk_disp_gamma_probe,
+	.remove		= mtk_disp_gamma_remove,
+	.driver		= {
+		.name	= "mediatek-disp-gamma",
+		.owner	= THIS_MODULE,
+		.of_match_table = mtk_disp_gamma_driver_dt_match,
+	},
+};
diff --git a/drivers/gpu/drm/mediatek/mtk_disp_ovl.c b/drivers/gpu/drm/mediatek/mtk_disp_ovl.c
index 74ef6fc0528b..961f87f8d4d1 100644
--- a/drivers/gpu/drm/mediatek/mtk_disp_ovl.c
+++ b/drivers/gpu/drm/mediatek/mtk_disp_ovl.c
@@ -13,6 +13,7 @@
 #include <linux/platform_device.h>
 #include <linux/soc/mediatek/mtk-cmdq.h>
 
+#include "mtk_disp_drv.h"
 #include "mtk_drm_crtc.h"
 #include "mtk_drm_ddp_comp.h"
 
@@ -23,6 +24,7 @@
 #define DISP_REG_OVL_RST			0x0014
 #define DISP_REG_OVL_ROI_SIZE			0x0020
 #define DISP_REG_OVL_DATAPATH_CON		0x0024
+#define OVL_LAYER_SMI_ID_EN				BIT(0)
 #define OVL_BGCLR_SEL_IN				BIT(2)
 #define DISP_REG_OVL_ROI_BGCLR			0x0028
 #define DISP_REG_OVL_SRC_CON			0x002c
@@ -61,6 +63,7 @@ struct mtk_disp_ovl_data {
 	unsigned int gmc_bits;
 	unsigned int layer_nr;
 	bool fmt_rgb565_is_0;
+	bool smi_id_en;
 };
 
 /**
@@ -70,88 +73,124 @@ struct mtk_disp_ovl_data {
  * @data: platform data
  */
 struct mtk_disp_ovl {
-	struct mtk_ddp_comp		ddp_comp;
 	struct drm_crtc			*crtc;
+	struct clk			*clk;
+	void __iomem			*regs;
+	struct cmdq_client_reg		cmdq_reg;
 	const struct mtk_disp_ovl_data	*data;
+	void				(*vblank_cb)(void *data);
+	void				*vblank_cb_data;
 };
 
-static inline struct mtk_disp_ovl *comp_to_ovl(struct mtk_ddp_comp *comp)
-{
-	return container_of(comp, struct mtk_disp_ovl, ddp_comp);
-}
-
 static irqreturn_t mtk_disp_ovl_irq_handler(int irq, void *dev_id)
 {
 	struct mtk_disp_ovl *priv = dev_id;
-	struct mtk_ddp_comp *ovl = &priv->ddp_comp;
 
 	/* Clear frame completion interrupt */
-	writel(0x0, ovl->regs + DISP_REG_OVL_INTSTA);
+	writel(0x0, priv->regs + DISP_REG_OVL_INTSTA);
 
-	if (!priv->crtc)
+	if (!priv->vblank_cb)
 		return IRQ_NONE;
 
-	mtk_crtc_ddp_irq(priv->crtc, ovl);
+	priv->vblank_cb(priv->vblank_cb_data);
 
 	return IRQ_HANDLED;
 }
 
-static void mtk_ovl_enable_vblank(struct mtk_ddp_comp *comp,
-				  struct drm_crtc *crtc)
+void mtk_ovl_enable_vblank(struct device *dev,
+			   void (*vblank_cb)(void *),
+			   void *vblank_cb_data)
+{
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
+
+	ovl->vblank_cb = vblank_cb;
+	ovl->vblank_cb_data = vblank_cb_data;
+	writel(0x0, ovl->regs + DISP_REG_OVL_INTSTA);
+	writel_relaxed(OVL_FME_CPL_INT, ovl->regs + DISP_REG_OVL_INTEN);
+}
+
+void mtk_ovl_disable_vblank(struct device *dev)
+{
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
+
+	ovl->vblank_cb = NULL;
+	ovl->vblank_cb_data = NULL;
+	writel_relaxed(0x0, ovl->regs + DISP_REG_OVL_INTEN);
+}
+
+int mtk_ovl_clk_enable(struct device *dev)
 {
-	struct mtk_disp_ovl *ovl = comp_to_ovl(comp);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 
-	ovl->crtc = crtc;
-	writel(0x0, comp->regs + DISP_REG_OVL_INTSTA);
-	writel_relaxed(OVL_FME_CPL_INT, comp->regs + DISP_REG_OVL_INTEN);
+	return clk_prepare_enable(ovl->clk);
 }
 
-static void mtk_ovl_disable_vblank(struct mtk_ddp_comp *comp)
+void mtk_ovl_clk_disable(struct device *dev)
 {
-	struct mtk_disp_ovl *ovl = comp_to_ovl(comp);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 
-	ovl->crtc = NULL;
-	writel_relaxed(0x0, comp->regs + DISP_REG_OVL_INTEN);
+	clk_disable_unprepare(ovl->clk);
 }
 
-static void mtk_ovl_start(struct mtk_ddp_comp *comp)
+void mtk_ovl_start(struct device *dev)
 {
-	writel_relaxed(0x1, comp->regs + DISP_REG_OVL_EN);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
+
+	if (ovl->data->smi_id_en) {
+		unsigned int reg;
+
+		reg = readl(ovl->regs + DISP_REG_OVL_DATAPATH_CON);
+		reg = reg | OVL_LAYER_SMI_ID_EN;
+		writel_relaxed(reg, ovl->regs + DISP_REG_OVL_DATAPATH_CON);
+	}
+	writel_relaxed(0x1, ovl->regs + DISP_REG_OVL_EN);
 }
 
-static void mtk_ovl_stop(struct mtk_ddp_comp *comp)
+void mtk_ovl_stop(struct device *dev)
 {
-	writel_relaxed(0x0, comp->regs + DISP_REG_OVL_EN);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
+
+	writel_relaxed(0x0, ovl->regs + DISP_REG_OVL_EN);
+	if (ovl->data->smi_id_en) {
+		unsigned int reg;
+
+		reg = readl(ovl->regs + DISP_REG_OVL_DATAPATH_CON);
+		reg = reg & ~OVL_LAYER_SMI_ID_EN;
+		writel_relaxed(reg, ovl->regs + DISP_REG_OVL_DATAPATH_CON);
+	}
+
 }
 
-static void mtk_ovl_config(struct mtk_ddp_comp *comp, unsigned int w,
-			   unsigned int h, unsigned int vrefresh,
-			   unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+void mtk_ovl_config(struct device *dev, unsigned int w,
+		    unsigned int h, unsigned int vrefresh,
+		    unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
 {
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
+
 	if (w != 0 && h != 0)
-		mtk_ddp_write_relaxed(cmdq_pkt, h << 16 | w, comp,
+		mtk_ddp_write_relaxed(cmdq_pkt, h << 16 | w, &ovl->cmdq_reg, ovl->regs,
 				      DISP_REG_OVL_ROI_SIZE);
-	mtk_ddp_write_relaxed(cmdq_pkt, 0x0, comp, DISP_REG_OVL_ROI_BGCLR);
+	mtk_ddp_write_relaxed(cmdq_pkt, 0x0, &ovl->cmdq_reg, ovl->regs, DISP_REG_OVL_ROI_BGCLR);
 
-	mtk_ddp_write(cmdq_pkt, 0x1, comp, DISP_REG_OVL_RST);
-	mtk_ddp_write(cmdq_pkt, 0x0, comp, DISP_REG_OVL_RST);
+	mtk_ddp_write(cmdq_pkt, 0x1, &ovl->cmdq_reg, ovl->regs, DISP_REG_OVL_RST);
+	mtk_ddp_write(cmdq_pkt, 0x0, &ovl->cmdq_reg, ovl->regs, DISP_REG_OVL_RST);
 }
 
-static unsigned int mtk_ovl_layer_nr(struct mtk_ddp_comp *comp)
+unsigned int mtk_ovl_layer_nr(struct device *dev)
 {
-	struct mtk_disp_ovl *ovl = comp_to_ovl(comp);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 
 	return ovl->data->layer_nr;
 }
 
-static unsigned int mtk_ovl_supported_rotations(struct mtk_ddp_comp *comp)
+unsigned int mtk_ovl_supported_rotations(struct device *dev)
 {
 	return DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180 |
 	       DRM_MODE_REFLECT_X | DRM_MODE_REFLECT_Y;
 }
 
-static int mtk_ovl_layer_check(struct mtk_ddp_comp *comp, unsigned int idx,
-			       struct mtk_plane_state *mtk_state)
+int mtk_ovl_layer_check(struct device *dev, unsigned int idx,
+			struct mtk_plane_state *mtk_state)
 {
 	struct drm_plane_state *state = &mtk_state->base;
 	unsigned int rotation = 0;
@@ -178,15 +217,15 @@ static int mtk_ovl_layer_check(struct mtk_ddp_comp *comp, unsigned int idx,
 	return 0;
 }
 
-static void mtk_ovl_layer_on(struct mtk_ddp_comp *comp, unsigned int idx,
-			     struct cmdq_pkt *cmdq_pkt)
+void mtk_ovl_layer_on(struct device *dev, unsigned int idx,
+		      struct cmdq_pkt *cmdq_pkt)
 {
 	unsigned int gmc_thrshd_l;
 	unsigned int gmc_thrshd_h;
 	unsigned int gmc_value;
-	struct mtk_disp_ovl *ovl = comp_to_ovl(comp);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 
-	mtk_ddp_write(cmdq_pkt, 0x1, comp,
+	mtk_ddp_write(cmdq_pkt, 0x1, &ovl->cmdq_reg, ovl->regs,
 		      DISP_REG_OVL_RDMA_CTRL(idx));
 	gmc_thrshd_l = GMC_THRESHOLD_LOW >>
 		      (GMC_THRESHOLD_BITS - ovl->data->gmc_bits);
@@ -198,17 +237,19 @@ static void mtk_ovl_layer_on(struct mtk_ddp_comp *comp, unsigned int idx,
 		gmc_value = gmc_thrshd_l | gmc_thrshd_l << 8 |
 			    gmc_thrshd_h << 16 | gmc_thrshd_h << 24;
 	mtk_ddp_write(cmdq_pkt, gmc_value,
-		      comp, DISP_REG_OVL_RDMA_GMC(idx));
-	mtk_ddp_write_mask(cmdq_pkt, BIT(idx), comp,
+		      &ovl->cmdq_reg, ovl->regs, DISP_REG_OVL_RDMA_GMC(idx));
+	mtk_ddp_write_mask(cmdq_pkt, BIT(idx), &ovl->cmdq_reg, ovl->regs,
 			   DISP_REG_OVL_SRC_CON, BIT(idx));
 }
 
-static void mtk_ovl_layer_off(struct mtk_ddp_comp *comp, unsigned int idx,
-			      struct cmdq_pkt *cmdq_pkt)
+void mtk_ovl_layer_off(struct device *dev, unsigned int idx,
+		       struct cmdq_pkt *cmdq_pkt)
 {
-	mtk_ddp_write_mask(cmdq_pkt, 0, comp,
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
+
+	mtk_ddp_write_mask(cmdq_pkt, 0, &ovl->cmdq_reg, ovl->regs,
 			   DISP_REG_OVL_SRC_CON, BIT(idx));
-	mtk_ddp_write(cmdq_pkt, 0, comp,
+	mtk_ddp_write(cmdq_pkt, 0, &ovl->cmdq_reg, ovl->regs,
 		      DISP_REG_OVL_RDMA_CTRL(idx));
 }
 
@@ -248,11 +289,11 @@ static unsigned int ovl_fmt_convert(struct mtk_disp_ovl *ovl, unsigned int fmt)
 	}
 }
 
-static void mtk_ovl_layer_config(struct mtk_ddp_comp *comp, unsigned int idx,
-				 struct mtk_plane_state *state,
-				 struct cmdq_pkt *cmdq_pkt)
+void mtk_ovl_layer_config(struct device *dev, unsigned int idx,
+			  struct mtk_plane_state *state,
+			  struct cmdq_pkt *cmdq_pkt)
 {
-	struct mtk_disp_ovl *ovl = comp_to_ovl(comp);
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 	struct mtk_plane_pending_state *pending = &state->pending;
 	unsigned int addr = pending->addr;
 	unsigned int pitch = pending->pitch & 0xffff;
@@ -262,12 +303,12 @@ static void mtk_ovl_layer_config(struct mtk_ddp_comp *comp, unsigned int idx,
 	unsigned int con;
 
 	if (!pending->enable) {
-		mtk_ovl_layer_off(comp, idx, cmdq_pkt);
+		mtk_ovl_layer_off(dev, idx, cmdq_pkt);
 		return;
 	}
 
 	con = ovl_fmt_convert(ovl, fmt);
-	if (state->base.fb->format->has_alpha)
+	if (state->base.fb && state->base.fb->format->has_alpha)
 		con |= OVL_CON_AEN | OVL_CON_ALPHA;
 
 	if (pending->rotation & DRM_MODE_REFLECT_Y) {
@@ -280,76 +321,49 @@ static void mtk_ovl_layer_config(struct mtk_ddp_comp *comp, unsigned int idx,
 		addr += pending->pitch - 1;
 	}
 
-	mtk_ddp_write_relaxed(cmdq_pkt, con, comp,
+	mtk_ddp_write_relaxed(cmdq_pkt, con, &ovl->cmdq_reg, ovl->regs,
 			      DISP_REG_OVL_CON(idx));
-	mtk_ddp_write_relaxed(cmdq_pkt, pitch, comp,
+	mtk_ddp_write_relaxed(cmdq_pkt, pitch, &ovl->cmdq_reg, ovl->regs,
 			      DISP_REG_OVL_PITCH(idx));
-	mtk_ddp_write_relaxed(cmdq_pkt, src_size, comp,
+	mtk_ddp_write_relaxed(cmdq_pkt, src_size, &ovl->cmdq_reg, ovl->regs,
 			      DISP_REG_OVL_SRC_SIZE(idx));
-	mtk_ddp_write_relaxed(cmdq_pkt, offset, comp,
+	mtk_ddp_write_relaxed(cmdq_pkt, offset, &ovl->cmdq_reg, ovl->regs,
 			      DISP_REG_OVL_OFFSET(idx));
-	mtk_ddp_write_relaxed(cmdq_pkt, addr, comp,
+	mtk_ddp_write_relaxed(cmdq_pkt, addr, &ovl->cmdq_reg, ovl->regs,
 			      DISP_REG_OVL_ADDR(ovl, idx));
 
-	mtk_ovl_layer_on(comp, idx, cmdq_pkt);
+	mtk_ovl_layer_on(dev, idx, cmdq_pkt);
 }
 
-static void mtk_ovl_bgclr_in_on(struct mtk_ddp_comp *comp)
+void mtk_ovl_bgclr_in_on(struct device *dev)
 {
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 	unsigned int reg;
 
-	reg = readl(comp->regs + DISP_REG_OVL_DATAPATH_CON);
+	reg = readl(ovl->regs + DISP_REG_OVL_DATAPATH_CON);
 	reg = reg | OVL_BGCLR_SEL_IN;
-	writel(reg, comp->regs + DISP_REG_OVL_DATAPATH_CON);
+	writel(reg, ovl->regs + DISP_REG_OVL_DATAPATH_CON);
 }
 
-static void mtk_ovl_bgclr_in_off(struct mtk_ddp_comp *comp)
+void mtk_ovl_bgclr_in_off(struct device *dev)
 {
+	struct mtk_disp_ovl *ovl = dev_get_drvdata(dev);
 	unsigned int reg;
 
-	reg = readl(comp->regs + DISP_REG_OVL_DATAPATH_CON);
+	reg = readl(ovl->regs + DISP_REG_OVL_DATAPATH_CON);
 	reg = reg & ~OVL_BGCLR_SEL_IN;
-	writel(reg, comp->regs + DISP_REG_OVL_DATAPATH_CON);
+	writel(reg, ovl->regs + DISP_REG_OVL_DATAPATH_CON);
 }
 
-static const struct mtk_ddp_comp_funcs mtk_disp_ovl_funcs = {
-	.config = mtk_ovl_config,
-	.start = mtk_ovl_start,
-	.stop = mtk_ovl_stop,
-	.enable_vblank = mtk_ovl_enable_vblank,
-	.disable_vblank = mtk_ovl_disable_vblank,
-	.supported_rotations = mtk_ovl_supported_rotations,
-	.layer_nr = mtk_ovl_layer_nr,
-	.layer_check = mtk_ovl_layer_check,
-	.layer_config = mtk_ovl_layer_config,
-	.bgclr_in_on = mtk_ovl_bgclr_in_on,
-	.bgclr_in_off = mtk_ovl_bgclr_in_off,
-};
-
 static int mtk_disp_ovl_bind(struct device *dev, struct device *master,
 			     void *data)
 {
-	struct mtk_disp_ovl *priv = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
-	int ret;
-
-	ret = mtk_ddp_comp_register(drm_dev, &priv->ddp_comp);
-	if (ret < 0) {
-		dev_err(dev, "Failed to register component %pOF: %d\n",
-			dev->of_node, ret);
-		return ret;
-	}
-
 	return 0;
 }
 
 static void mtk_disp_ovl_unbind(struct device *dev, struct device *master,
 				void *data)
 {
-	struct mtk_disp_ovl *priv = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
-
-	mtk_ddp_comp_unregister(drm_dev, &priv->ddp_comp);
 }
 
 static const struct component_ops mtk_disp_ovl_component_ops = {
@@ -361,7 +375,7 @@ static int mtk_disp_ovl_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct mtk_disp_ovl *priv;
-	int comp_id;
+	struct resource *res;
 	int irq;
 	int ret;
 
@@ -373,27 +387,25 @@ static int mtk_disp_ovl_probe(struct platform_device *pdev)
 	if (irq < 0)
 		return irq;
 
-	priv->data = of_device_get_match_data(dev);
-
-	comp_id = mtk_ddp_comp_get_id(dev->of_node,
-				      priv->data->layer_nr == 4 ?
-				      MTK_DISP_OVL :
-				      MTK_DISP_OVL_2L);
-	if (comp_id < 0) {
-		dev_err(dev, "Failed to identify by alias: %d\n", comp_id);
-		return comp_id;
+	priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to get ovl clk\n");
+		return PTR_ERR(priv->clk);
 	}
 
-	ret = mtk_ddp_comp_init(dev, dev->of_node, &priv->ddp_comp, comp_id,
-				&mtk_disp_ovl_funcs);
-	if (ret) {
-		if (ret != -EPROBE_DEFER)
-			dev_err(dev, "Failed to initialize component: %d\n",
-				ret);
-
-		return ret;
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	priv->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(priv->regs)) {
+		dev_err(dev, "failed to ioremap ovl\n");
+		return PTR_ERR(priv->regs);
 	}
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	ret = cmdq_dev_get_client_reg(dev, &priv->cmdq_reg, 0);
+	if (ret)
+		dev_dbg(dev, "get mediatek,gce-client-reg fail!\n");
+#endif
 
+	priv->data = of_device_get_match_data(dev);
 	platform_set_drvdata(pdev, priv);
 
 	ret = devm_request_irq(dev, irq, mtk_disp_ovl_irq_handler,
@@ -412,8 +424,6 @@ static int mtk_disp_ovl_probe(struct platform_device *pdev)
 
 static int mtk_disp_ovl_remove(struct platform_device *pdev)
 {
-	component_del(&pdev->dev, &mtk_disp_ovl_component_ops);
-
 	return 0;
 }
 
@@ -431,11 +441,29 @@ static const struct mtk_disp_ovl_data mt8173_ovl_driver_data = {
 	.fmt_rgb565_is_0 = true,
 };
 
+static const struct mtk_disp_ovl_data mt8183_ovl_driver_data = {
+	.addr = DISP_REG_OVL_ADDR_MT8173,
+	.gmc_bits = 10,
+	.layer_nr = 4,
+	.fmt_rgb565_is_0 = true,
+};
+
+static const struct mtk_disp_ovl_data mt8183_ovl_2l_driver_data = {
+	.addr = DISP_REG_OVL_ADDR_MT8173,
+	.gmc_bits = 10,
+	.layer_nr = 2,
+	.fmt_rgb565_is_0 = true,
+};
+
 static const struct of_device_id mtk_disp_ovl_driver_dt_match[] = {
 	{ .compatible = "mediatek,mt2701-disp-ovl",
 	  .data = &mt2701_ovl_driver_data},
 	{ .compatible = "mediatek,mt8173-disp-ovl",
 	  .data = &mt8173_ovl_driver_data},
+	{ .compatible = "mediatek,mt8183-disp-ovl",
+	  .data = &mt8183_ovl_driver_data},
+	{ .compatible = "mediatek,mt8183-disp-ovl-2l",
+	  .data = &mt8183_ovl_2l_driver_data},
 	{},
 };
 MODULE_DEVICE_TABLE(of, mtk_disp_ovl_driver_dt_match);
diff --git a/drivers/gpu/drm/mediatek/mtk_disp_rdma.c b/drivers/gpu/drm/mediatek/mtk_disp_rdma.c
index d46b8ae1d080..728aaadfea8c 100644
--- a/drivers/gpu/drm/mediatek/mtk_disp_rdma.c
+++ b/drivers/gpu/drm/mediatek/mtk_disp_rdma.c
@@ -11,6 +11,7 @@
 #include <linux/platform_device.h>
 #include <linux/soc/mediatek/mtk-cmdq.h>
 
+#include "mtk_disp_drv.h"
 #include "mtk_drm_crtc.h"
 #include "mtk_drm_ddp_comp.h"
 
@@ -61,83 +62,105 @@ struct mtk_disp_rdma_data {
  * @data: local driver data
  */
 struct mtk_disp_rdma {
-	struct mtk_ddp_comp		ddp_comp;
-	struct drm_crtc			*crtc;
+	struct clk			*clk;
+	void __iomem			*regs;
+	struct cmdq_client_reg		cmdq_reg;
 	const struct mtk_disp_rdma_data	*data;
+	void				(*vblank_cb)(void *data);
+	void				*vblank_cb_data;
+	u32				fifo_size;
 };
 
-static inline struct mtk_disp_rdma *comp_to_rdma(struct mtk_ddp_comp *comp)
-{
-	return container_of(comp, struct mtk_disp_rdma, ddp_comp);
-}
-
 static irqreturn_t mtk_disp_rdma_irq_handler(int irq, void *dev_id)
 {
 	struct mtk_disp_rdma *priv = dev_id;
-	struct mtk_ddp_comp *rdma = &priv->ddp_comp;
 
 	/* Clear frame completion interrupt */
-	writel(0x0, rdma->regs + DISP_REG_RDMA_INT_STATUS);
+	writel(0x0, priv->regs + DISP_REG_RDMA_INT_STATUS);
 
-	if (!priv->crtc)
+	if (!priv->vblank_cb)
 		return IRQ_NONE;
 
-	mtk_crtc_ddp_irq(priv->crtc, rdma);
+	priv->vblank_cb(priv->vblank_cb_data);
 
 	return IRQ_HANDLED;
 }
 
-static void rdma_update_bits(struct mtk_ddp_comp *comp, unsigned int reg,
+static void rdma_update_bits(struct device *dev, unsigned int reg,
 			     unsigned int mask, unsigned int val)
 {
-	unsigned int tmp = readl(comp->regs + reg);
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
+	unsigned int tmp = readl(rdma->regs + reg);
 
 	tmp = (tmp & ~mask) | (val & mask);
-	writel(tmp, comp->regs + reg);
+	writel(tmp, rdma->regs + reg);
 }
 
-static void mtk_rdma_enable_vblank(struct mtk_ddp_comp *comp,
-				   struct drm_crtc *crtc)
+void mtk_rdma_enable_vblank(struct device *dev,
+			    void (*vblank_cb)(void *),
+			    void *vblank_cb_data)
 {
-	struct mtk_disp_rdma *rdma = comp_to_rdma(comp);
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
 
-	rdma->crtc = crtc;
-	rdma_update_bits(comp, DISP_REG_RDMA_INT_ENABLE, RDMA_FRAME_END_INT,
+	rdma->vblank_cb = vblank_cb;
+	rdma->vblank_cb_data = vblank_cb_data;
+	rdma_update_bits(dev, DISP_REG_RDMA_INT_ENABLE, RDMA_FRAME_END_INT,
 			 RDMA_FRAME_END_INT);
 }
 
-static void mtk_rdma_disable_vblank(struct mtk_ddp_comp *comp)
+void mtk_rdma_disable_vblank(struct device *dev)
+{
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
+
+	rdma->vblank_cb = NULL;
+	rdma->vblank_cb_data = NULL;
+	rdma_update_bits(dev, DISP_REG_RDMA_INT_ENABLE, RDMA_FRAME_END_INT, 0);
+}
+
+int mtk_rdma_clk_enable(struct device *dev)
 {
-	struct mtk_disp_rdma *rdma = comp_to_rdma(comp);
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
 
-	rdma->crtc = NULL;
-	rdma_update_bits(comp, DISP_REG_RDMA_INT_ENABLE, RDMA_FRAME_END_INT, 0);
+	return clk_prepare_enable(rdma->clk);
 }
 
-static void mtk_rdma_start(struct mtk_ddp_comp *comp)
+void mtk_rdma_clk_disable(struct device *dev)
 {
-	rdma_update_bits(comp, DISP_REG_RDMA_GLOBAL_CON, RDMA_ENGINE_EN,
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(rdma->clk);
+}
+
+void mtk_rdma_start(struct device *dev)
+{
+	rdma_update_bits(dev, DISP_REG_RDMA_GLOBAL_CON, RDMA_ENGINE_EN,
 			 RDMA_ENGINE_EN);
 }
 
-static void mtk_rdma_stop(struct mtk_ddp_comp *comp)
+void mtk_rdma_stop(struct device *dev)
 {
-	rdma_update_bits(comp, DISP_REG_RDMA_GLOBAL_CON, RDMA_ENGINE_EN, 0);
+	rdma_update_bits(dev, DISP_REG_RDMA_GLOBAL_CON, RDMA_ENGINE_EN, 0);
 }
 
-static void mtk_rdma_config(struct mtk_ddp_comp *comp, unsigned int width,
-			    unsigned int height, unsigned int vrefresh,
-			    unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+void mtk_rdma_config(struct device *dev, unsigned int width,
+		     unsigned int height, unsigned int vrefresh,
+		     unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
 {
 	unsigned int threshold;
 	unsigned int reg;
-	struct mtk_disp_rdma *rdma = comp_to_rdma(comp);
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
+	u32 rdma_fifo_size;
 
-	mtk_ddp_write_mask(cmdq_pkt, width, comp,
+	mtk_ddp_write_mask(cmdq_pkt, width, &rdma->cmdq_reg, rdma->regs,
 			   DISP_REG_RDMA_SIZE_CON_0, 0xfff);
-	mtk_ddp_write_mask(cmdq_pkt, height, comp,
+	mtk_ddp_write_mask(cmdq_pkt, height, &rdma->cmdq_reg, rdma->regs,
 			   DISP_REG_RDMA_SIZE_CON_1, 0xfffff);
 
+	if (rdma->fifo_size)
+		rdma_fifo_size = rdma->fifo_size;
+	else
+		rdma_fifo_size = RDMA_FIFO_SIZE(rdma);
+
 	/*
 	 * Enable FIFO underflow since DSI and DPI can't be blocked.
 	 * Keep the FIFO pseudo size reset default of 8 KiB. Set the
@@ -146,9 +169,9 @@ static void mtk_rdma_config(struct mtk_ddp_comp *comp, unsigned int width,
 	 */
 	threshold = width * height * vrefresh * 4 * 7 / 1000000;
 	reg = RDMA_FIFO_UNDERFLOW_EN |
-	      RDMA_FIFO_PSEUDO_SIZE(RDMA_FIFO_SIZE(rdma)) |
+	      RDMA_FIFO_PSEUDO_SIZE(rdma_fifo_size) |
 	      RDMA_OUTPUT_VALID_FIFO_THRESHOLD(threshold);
-	mtk_ddp_write(cmdq_pkt, reg, comp, DISP_REG_RDMA_FIFO_CON);
+	mtk_ddp_write(cmdq_pkt, reg, &rdma->cmdq_reg, rdma->regs, DISP_REG_RDMA_FIFO_CON);
 }
 
 static unsigned int rdma_fmt_convert(struct mtk_disp_rdma *rdma,
@@ -188,16 +211,16 @@ static unsigned int rdma_fmt_convert(struct mtk_disp_rdma *rdma,
 	}
 }
 
-static unsigned int mtk_rdma_layer_nr(struct mtk_ddp_comp *comp)
+unsigned int mtk_rdma_layer_nr(struct device *dev)
 {
 	return 1;
 }
 
-static void mtk_rdma_layer_config(struct mtk_ddp_comp *comp, unsigned int idx,
-				  struct mtk_plane_state *state,
-				  struct cmdq_pkt *cmdq_pkt)
+void mtk_rdma_layer_config(struct device *dev, unsigned int idx,
+			   struct mtk_plane_state *state,
+			   struct cmdq_pkt *cmdq_pkt)
 {
-	struct mtk_disp_rdma *rdma = comp_to_rdma(comp);
+	struct mtk_disp_rdma *rdma = dev_get_drvdata(dev);
 	struct mtk_plane_pending_state *pending = &state->pending;
 	unsigned int addr = pending->addr;
 	unsigned int pitch = pending->pitch & 0xffff;
@@ -205,53 +228,34 @@ static void mtk_rdma_layer_config(struct mtk_ddp_comp *comp, unsigned int idx,
 	unsigned int con;
 
 	con = rdma_fmt_convert(rdma, fmt);
-	mtk_ddp_write_relaxed(cmdq_pkt, con, comp, DISP_RDMA_MEM_CON);
+	mtk_ddp_write_relaxed(cmdq_pkt, con, &rdma->cmdq_reg, rdma->regs, DISP_RDMA_MEM_CON);
 
 	if (fmt == DRM_FORMAT_UYVY || fmt == DRM_FORMAT_YUYV) {
-		mtk_ddp_write_mask(cmdq_pkt, RDMA_MATRIX_ENABLE, comp,
+		mtk_ddp_write_mask(cmdq_pkt, RDMA_MATRIX_ENABLE, &rdma->cmdq_reg, rdma->regs,
 				   DISP_REG_RDMA_SIZE_CON_0,
 				   RDMA_MATRIX_ENABLE);
 		mtk_ddp_write_mask(cmdq_pkt, RDMA_MATRIX_INT_MTX_BT601_to_RGB,
-				   comp, DISP_REG_RDMA_SIZE_CON_0,
+				   &rdma->cmdq_reg, rdma->regs, DISP_REG_RDMA_SIZE_CON_0,
 				   RDMA_MATRIX_INT_MTX_SEL);
 	} else {
-		mtk_ddp_write_mask(cmdq_pkt, 0, comp,
+		mtk_ddp_write_mask(cmdq_pkt, 0, &rdma->cmdq_reg, rdma->regs,
 				   DISP_REG_RDMA_SIZE_CON_0,
 				   RDMA_MATRIX_ENABLE);
 	}
-	mtk_ddp_write_relaxed(cmdq_pkt, addr, comp, DISP_RDMA_MEM_START_ADDR);
-	mtk_ddp_write_relaxed(cmdq_pkt, pitch, comp, DISP_RDMA_MEM_SRC_PITCH);
-	mtk_ddp_write(cmdq_pkt, RDMA_MEM_GMC, comp,
+	mtk_ddp_write_relaxed(cmdq_pkt, addr, &rdma->cmdq_reg, rdma->regs,
+			      DISP_RDMA_MEM_START_ADDR);
+	mtk_ddp_write_relaxed(cmdq_pkt, pitch, &rdma->cmdq_reg, rdma->regs,
+			      DISP_RDMA_MEM_SRC_PITCH);
+	mtk_ddp_write(cmdq_pkt, RDMA_MEM_GMC, &rdma->cmdq_reg, rdma->regs,
 		      DISP_RDMA_MEM_GMC_SETTING_0);
-	mtk_ddp_write_mask(cmdq_pkt, RDMA_MODE_MEMORY, comp,
+	mtk_ddp_write_mask(cmdq_pkt, RDMA_MODE_MEMORY, &rdma->cmdq_reg, rdma->regs,
 			   DISP_REG_RDMA_GLOBAL_CON, RDMA_MODE_MEMORY);
 
 }
 
-static const struct mtk_ddp_comp_funcs mtk_disp_rdma_funcs = {
-	.config = mtk_rdma_config,
-	.start = mtk_rdma_start,
-	.stop = mtk_rdma_stop,
-	.enable_vblank = mtk_rdma_enable_vblank,
-	.disable_vblank = mtk_rdma_disable_vblank,
-	.layer_nr = mtk_rdma_layer_nr,
-	.layer_config = mtk_rdma_layer_config,
-};
-
 static int mtk_disp_rdma_bind(struct device *dev, struct device *master,
 			      void *data)
 {
-	struct mtk_disp_rdma *priv = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
-	int ret;
-
-	ret = mtk_ddp_comp_register(drm_dev, &priv->ddp_comp);
-	if (ret < 0) {
-		dev_err(dev, "Failed to register component %pOF: %d\n",
-			dev->of_node, ret);
-		return ret;
-	}
-
 	return 0;
 
 }
@@ -259,10 +263,6 @@ static int mtk_disp_rdma_bind(struct device *dev, struct device *master,
 static void mtk_disp_rdma_unbind(struct device *dev, struct device *master,
 				 void *data)
 {
-	struct mtk_disp_rdma *priv = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
-
-	mtk_ddp_comp_unregister(drm_dev, &priv->ddp_comp);
 }
 
 static const struct component_ops mtk_disp_rdma_component_ops = {
@@ -274,7 +274,7 @@ static int mtk_disp_rdma_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct mtk_disp_rdma *priv;
-	int comp_id;
+	struct resource *res;
 	int irq;
 	int ret;
 
@@ -286,25 +286,37 @@ static int mtk_disp_rdma_probe(struct platform_device *pdev)
 	if (irq < 0)
 		return irq;
 
-	comp_id = mtk_ddp_comp_get_id(dev->of_node, MTK_DISP_RDMA);
-	if (comp_id < 0) {
-		dev_err(dev, "Failed to identify by alias: %d\n", comp_id);
-		return comp_id;
+	priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to get rdma clk\n");
+		return PTR_ERR(priv->clk);
 	}
 
-	ret = mtk_ddp_comp_init(dev, dev->of_node, &priv->ddp_comp, comp_id,
-				&mtk_disp_rdma_funcs);
-	if (ret) {
-		if (ret != -EPROBE_DEFER)
-			dev_err(dev, "Failed to initialize component: %d\n",
-				ret);
-
-		return ret;
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	priv->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(priv->regs)) {
+		dev_err(dev, "failed to ioremap rdma\n");
+		return PTR_ERR(priv->regs);
+	}
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	ret = cmdq_dev_get_client_reg(dev, &priv->cmdq_reg, 0);
+	if (ret)
+		dev_dbg(dev, "get mediatek,gce-client-reg fail!\n");
+#endif
+
+	if (of_find_property(dev->of_node, "mediatek,rdma-fifo-size", &ret)) {
+		ret = of_property_read_u32(dev->of_node,
+					   "mediatek,rdma-fifo-size",
+					   &priv->fifo_size);
+		if (ret) {
+			dev_err(dev, "Failed to get rdma fifo size\n");
+			return ret;
+		}
 	}
 
 	/* Disable and clear pending interrupts */
-	writel(0x0, priv->ddp_comp.regs + DISP_REG_RDMA_INT_ENABLE);
-	writel(0x0, priv->ddp_comp.regs + DISP_REG_RDMA_INT_STATUS);
+	writel(0x0, priv->regs + DISP_REG_RDMA_INT_ENABLE);
+	writel(0x0, priv->regs + DISP_REG_RDMA_INT_STATUS);
 
 	ret = devm_request_irq(dev, irq, mtk_disp_rdma_irq_handler,
 			       IRQF_TRIGGER_NONE, dev_name(dev), priv);
@@ -339,11 +351,17 @@ static const struct mtk_disp_rdma_data mt8173_rdma_driver_data = {
 	.fifo_size = SZ_8K,
 };
 
+static const struct mtk_disp_rdma_data mt8183_rdma_driver_data = {
+	.fifo_size = 5 * SZ_1K,
+};
+
 static const struct of_device_id mtk_disp_rdma_driver_dt_match[] = {
 	{ .compatible = "mediatek,mt2701-disp-rdma",
 	  .data = &mt2701_rdma_driver_data},
 	{ .compatible = "mediatek,mt8173-disp-rdma",
 	  .data = &mt8173_rdma_driver_data},
+	{ .compatible = "mediatek,mt8183-disp-rdma",
+	  .data = &mt8183_rdma_driver_data},
 	{},
 };
 MODULE_DEVICE_TABLE(of, mtk_disp_rdma_driver_dt_match);
diff --git a/drivers/gpu/drm/mediatek/mtk_dpi.c b/drivers/gpu/drm/mediatek/mtk_dpi.c
index 52f11a63a330..b05f900d9322 100644
--- a/drivers/gpu/drm/mediatek/mtk_dpi.c
+++ b/drivers/gpu/drm/mediatek/mtk_dpi.c
@@ -20,10 +20,12 @@
 
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_bridge.h>
+#include <drm/drm_bridge_connector.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_of.h>
 #include <drm/drm_simple_kms_helper.h>
 
+#include "mtk_disp_drv.h"
 #include "mtk_dpi_regs.h"
 #include "mtk_drm_ddp_comp.h"
 
@@ -62,10 +64,10 @@ enum mtk_dpi_out_color_format {
 };
 
 struct mtk_dpi {
-	struct mtk_ddp_comp ddp_comp;
 	struct drm_encoder encoder;
 	struct drm_bridge bridge;
 	struct drm_bridge *next_bridge;
+	struct drm_connector *connector;
 	void __iomem *regs;
 	struct device *dev;
 	struct clk *engine_clk;
@@ -562,53 +564,50 @@ static const struct drm_bridge_funcs mtk_dpi_bridge_funcs = {
 	.enable = mtk_dpi_bridge_enable,
 };
 
-static void mtk_dpi_start(struct mtk_ddp_comp *comp)
+void mtk_dpi_start(struct device *dev)
 {
-	struct mtk_dpi *dpi = container_of(comp, struct mtk_dpi, ddp_comp);
+	struct mtk_dpi *dpi = dev_get_drvdata(dev);
 
 	mtk_dpi_power_on(dpi);
 }
 
-static void mtk_dpi_stop(struct mtk_ddp_comp *comp)
+void mtk_dpi_stop(struct device *dev)
 {
-	struct mtk_dpi *dpi = container_of(comp, struct mtk_dpi, ddp_comp);
+	struct mtk_dpi *dpi = dev_get_drvdata(dev);
 
 	mtk_dpi_power_off(dpi);
 }
 
-static const struct mtk_ddp_comp_funcs mtk_dpi_funcs = {
-	.start = mtk_dpi_start,
-	.stop = mtk_dpi_stop,
-};
-
 static int mtk_dpi_bind(struct device *dev, struct device *master, void *data)
 {
 	struct mtk_dpi *dpi = dev_get_drvdata(dev);
 	struct drm_device *drm_dev = data;
 	int ret;
 
-	ret = mtk_ddp_comp_register(drm_dev, &dpi->ddp_comp);
-	if (ret < 0) {
-		dev_err(dev, "Failed to register component %pOF: %d\n",
-			dev->of_node, ret);
-		return ret;
-	}
-
 	ret = drm_simple_encoder_init(drm_dev, &dpi->encoder,
 				      DRM_MODE_ENCODER_TMDS);
 	if (ret) {
 		dev_err(dev, "Failed to initialize decoder: %d\n", ret);
-		goto err_unregister;
+		return ret;
 	}
 
-	dpi->encoder.possible_crtcs = mtk_drm_find_possible_crtc_by_comp(drm_dev, dpi->ddp_comp);
+	dpi->encoder.possible_crtcs = mtk_drm_find_possible_crtc_by_comp(drm_dev, dpi->dev);
 
-	ret = drm_bridge_attach(&dpi->encoder, &dpi->bridge, NULL, 0);
+	ret = drm_bridge_attach(&dpi->encoder, &dpi->bridge, NULL,
+				DRM_BRIDGE_ATTACH_NO_CONNECTOR);
 	if (ret) {
 		dev_err(dev, "Failed to attach bridge: %d\n", ret);
 		goto err_cleanup;
 	}
 
+	dpi->connector = drm_bridge_connector_init(drm_dev, &dpi->encoder);
+	if (IS_ERR(dpi->connector)) {
+		dev_err(dev, "Unable to create bridge connector\n");
+		ret = PTR_ERR(dpi->connector);
+		goto err_cleanup;
+	}
+	drm_connector_attach_encoder(dpi->connector, &dpi->encoder);
+
 	dpi->bit_num = MTK_DPI_OUT_BIT_NUM_8BITS;
 	dpi->channel_swap = MTK_DPI_OUT_CHANNEL_SWAP_RGB;
 	dpi->yc_map = MTK_DPI_OUT_YC_MAP_RGB;
@@ -618,8 +617,6 @@ static int mtk_dpi_bind(struct device *dev, struct device *master, void *data)
 
 err_cleanup:
 	drm_encoder_cleanup(&dpi->encoder);
-err_unregister:
-	mtk_ddp_comp_unregister(drm_dev, &dpi->ddp_comp);
 	return ret;
 }
 
@@ -627,10 +624,8 @@ static void mtk_dpi_unbind(struct device *dev, struct device *master,
 			   void *data)
 {
 	struct mtk_dpi *dpi = dev_get_drvdata(dev);
-	struct drm_device *drm_dev = data;
 
 	drm_encoder_cleanup(&dpi->encoder);
-	mtk_ddp_comp_unregister(drm_dev, &dpi->ddp_comp);
 }
 
 static const struct component_ops mtk_dpi_component_ops = {
@@ -691,7 +686,6 @@ static int mtk_dpi_probe(struct platform_device *pdev)
 	struct device *dev = &pdev->dev;
 	struct mtk_dpi *dpi;
 	struct resource *mem;
-	int comp_id;
 	int ret;
 
 	dpi = devm_kzalloc(dev, sizeof(*dpi), GFP_KERNEL);
@@ -769,19 +763,6 @@ static int mtk_dpi_probe(struct platform_device *pdev)
 
 	dev_info(dev, "Found bridge node: %pOF\n", dpi->next_bridge->of_node);
 
-	comp_id = mtk_ddp_comp_get_id(dev->of_node, MTK_DPI);
-	if (comp_id < 0) {
-		dev_err(dev, "Failed to identify by alias: %d\n", comp_id);
-		return comp_id;
-	}
-
-	ret = mtk_ddp_comp_init(dev, dev->of_node, &dpi->ddp_comp, comp_id,
-				&mtk_dpi_funcs);
-	if (ret) {
-		dev_err(dev, "Failed to initialize component: %d\n", ret);
-		return ret;
-	}
-
 	platform_set_drvdata(pdev, dpi);
 
 	dpi->bridge.funcs = &mtk_dpi_bridge_funcs;
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_crtc.c b/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
index 70ded25fb92c..54ab3a324752 100644
--- a/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
+++ b/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
@@ -7,6 +7,7 @@
 #include <linux/pm_runtime.h>
 #include <linux/soc/mediatek/mtk-cmdq.h>
 #include <linux/soc/mediatek/mtk-mmsys.h>
+#include <linux/soc/mediatek/mtk-mutex.h>
 
 #include <asm/barrier.h>
 #include <soc/mediatek/smi.h>
@@ -19,7 +20,6 @@
 
 #include "mtk_drm_drv.h"
 #include "mtk_drm_crtc.h"
-#include "mtk_drm_ddp.h"
 #include "mtk_drm_ddp_comp.h"
 #include "mtk_drm_gem.h"
 #include "mtk_drm_plane.h"
@@ -55,7 +55,7 @@ struct mtk_drm_crtc {
 #endif
 
 	struct device			*mmsys_dev;
-	struct mtk_disp_mutex		*mutex;
+	struct mtk_mutex		*mutex;
 	unsigned int			ddp_comp_nr;
 	struct mtk_ddp_comp		**ddp_comp;
 
@@ -107,7 +107,7 @@ static void mtk_drm_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
 
-	mtk_disp_mutex_put(mtk_crtc->mutex);
+	mtk_mutex_put(mtk_crtc->mutex);
 
 	drm_crtc_cleanup(crtc);
 }
@@ -169,31 +169,13 @@ static void mtk_drm_crtc_mode_set_nofb(struct drm_crtc *crtc)
 	state->pending_config = true;
 }
 
-static int mtk_drm_crtc_enable_vblank(struct drm_crtc *crtc)
-{
-	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
-	struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[0];
-
-	mtk_ddp_comp_enable_vblank(comp, &mtk_crtc->base);
-
-	return 0;
-}
-
-static void mtk_drm_crtc_disable_vblank(struct drm_crtc *crtc)
-{
-	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
-	struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[0];
-
-	mtk_ddp_comp_disable_vblank(comp);
-}
-
 static int mtk_crtc_ddp_clk_enable(struct mtk_drm_crtc *mtk_crtc)
 {
 	int ret;
 	int i;
 
 	for (i = 0; i < mtk_crtc->ddp_comp_nr; i++) {
-		ret = clk_prepare_enable(mtk_crtc->ddp_comp[i]->clk);
+		ret = mtk_ddp_comp_clk_enable(mtk_crtc->ddp_comp[i]);
 		if (ret) {
 			DRM_ERROR("Failed to enable clock %d: %d\n", i, ret);
 			goto err;
@@ -203,7 +185,7 @@ static int mtk_crtc_ddp_clk_enable(struct mtk_drm_crtc *mtk_crtc)
 	return 0;
 err:
 	while (--i >= 0)
-		clk_disable_unprepare(mtk_crtc->ddp_comp[i]->clk);
+		mtk_ddp_comp_clk_disable(mtk_crtc->ddp_comp[i]);
 	return ret;
 }
 
@@ -212,7 +194,7 @@ static void mtk_crtc_ddp_clk_disable(struct mtk_drm_crtc *mtk_crtc)
 	int i;
 
 	for (i = 0; i < mtk_crtc->ddp_comp_nr; i++)
-		clk_disable_unprepare(mtk_crtc->ddp_comp[i]->clk);
+		mtk_ddp_comp_clk_disable(mtk_crtc->ddp_comp[i]);
 }
 
 static
@@ -283,7 +265,7 @@ static int mtk_crtc_ddp_hw_init(struct mtk_drm_crtc *mtk_crtc)
 		return ret;
 	}
 
-	ret = mtk_disp_mutex_prepare(mtk_crtc->mutex);
+	ret = mtk_mutex_prepare(mtk_crtc->mutex);
 	if (ret < 0) {
 		DRM_ERROR("Failed to enable mutex clock: %d\n", ret);
 		goto err_pm_runtime_put;
@@ -299,11 +281,11 @@ static int mtk_crtc_ddp_hw_init(struct mtk_drm_crtc *mtk_crtc)
 		mtk_mmsys_ddp_connect(mtk_crtc->mmsys_dev,
 				      mtk_crtc->ddp_comp[i]->id,
 				      mtk_crtc->ddp_comp[i + 1]->id);
-		mtk_disp_mutex_add_comp(mtk_crtc->mutex,
+		mtk_mutex_add_comp(mtk_crtc->mutex,
 					mtk_crtc->ddp_comp[i]->id);
 	}
-	mtk_disp_mutex_add_comp(mtk_crtc->mutex, mtk_crtc->ddp_comp[i]->id);
-	mtk_disp_mutex_enable(mtk_crtc->mutex);
+	mtk_mutex_add_comp(mtk_crtc->mutex, mtk_crtc->ddp_comp[i]->id);
+	mtk_mutex_enable(mtk_crtc->mutex);
 
 	for (i = 0; i < mtk_crtc->ddp_comp_nr; i++) {
 		struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[i];
@@ -332,7 +314,7 @@ static int mtk_crtc_ddp_hw_init(struct mtk_drm_crtc *mtk_crtc)
 	return 0;
 
 err_mutex_unprepare:
-	mtk_disp_mutex_unprepare(mtk_crtc->mutex);
+	mtk_mutex_unprepare(mtk_crtc->mutex);
 err_pm_runtime_put:
 	pm_runtime_put(crtc->dev->dev);
 	return ret;
@@ -351,19 +333,19 @@ static void mtk_crtc_ddp_hw_fini(struct mtk_drm_crtc *mtk_crtc)
 	}
 
 	for (i = 0; i < mtk_crtc->ddp_comp_nr; i++)
-		mtk_disp_mutex_remove_comp(mtk_crtc->mutex,
+		mtk_mutex_remove_comp(mtk_crtc->mutex,
 					   mtk_crtc->ddp_comp[i]->id);
-	mtk_disp_mutex_disable(mtk_crtc->mutex);
+	mtk_mutex_disable(mtk_crtc->mutex);
 	for (i = 0; i < mtk_crtc->ddp_comp_nr - 1; i++) {
 		mtk_mmsys_ddp_disconnect(mtk_crtc->mmsys_dev,
 					 mtk_crtc->ddp_comp[i]->id,
 					 mtk_crtc->ddp_comp[i + 1]->id);
-		mtk_disp_mutex_remove_comp(mtk_crtc->mutex,
+		mtk_mutex_remove_comp(mtk_crtc->mutex,
 					   mtk_crtc->ddp_comp[i]->id);
 	}
-	mtk_disp_mutex_remove_comp(mtk_crtc->mutex, mtk_crtc->ddp_comp[i]->id);
+	mtk_mutex_remove_comp(mtk_crtc->mutex, mtk_crtc->ddp_comp[i]->id);
 	mtk_crtc_ddp_clk_disable(mtk_crtc);
-	mtk_disp_mutex_unprepare(mtk_crtc->mutex);
+	mtk_mutex_unprepare(mtk_crtc->mutex);
 
 	pm_runtime_put(drm->dev);
 
@@ -475,9 +457,9 @@ static void mtk_drm_crtc_hw_config(struct mtk_drm_crtc *mtk_crtc)
 		mtk_crtc->pending_async_planes = true;
 
 	if (priv->data->shadow_register) {
-		mtk_disp_mutex_acquire(mtk_crtc->mutex);
+		mtk_mutex_acquire(mtk_crtc->mutex);
 		mtk_crtc_ddp_config(crtc, NULL);
-		mtk_disp_mutex_release(mtk_crtc->mutex);
+		mtk_mutex_release(mtk_crtc->mutex);
 	}
 #if IS_REACHABLE(CONFIG_MTK_CMDQ)
 	if (mtk_crtc->cmdq_client) {
@@ -493,6 +475,40 @@ static void mtk_drm_crtc_hw_config(struct mtk_drm_crtc *mtk_crtc)
 	mutex_unlock(&mtk_crtc->hw_lock);
 }
 
+static void mtk_crtc_ddp_irq(void *data)
+{
+	struct drm_crtc *crtc = data;
+	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
+	struct mtk_drm_private *priv = crtc->dev->dev_private;
+
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	if (!priv->data->shadow_register && !mtk_crtc->cmdq_client)
+#else
+	if (!priv->data->shadow_register)
+#endif
+		mtk_crtc_ddp_config(crtc, NULL);
+
+	mtk_drm_finish_page_flip(mtk_crtc);
+}
+
+static int mtk_drm_crtc_enable_vblank(struct drm_crtc *crtc)
+{
+	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
+	struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[0];
+
+	mtk_ddp_comp_enable_vblank(comp, mtk_crtc_ddp_irq, &mtk_crtc->base);
+
+	return 0;
+}
+
+static void mtk_drm_crtc_disable_vblank(struct drm_crtc *crtc)
+{
+	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
+	struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[0];
+
+	mtk_ddp_comp_disable_vblank(comp);
+}
+
 int mtk_drm_crtc_plane_check(struct drm_crtc *crtc, struct drm_plane *plane,
 			     struct mtk_plane_state *state)
 {
@@ -661,21 +677,6 @@ err_cleanup_crtc:
 	return ret;
 }
 
-void mtk_crtc_ddp_irq(struct drm_crtc *crtc, struct mtk_ddp_comp *comp)
-{
-	struct mtk_drm_crtc *mtk_crtc = to_mtk_crtc(crtc);
-	struct mtk_drm_private *priv = crtc->dev->dev_private;
-
-#if IS_REACHABLE(CONFIG_MTK_CMDQ)
-	if (!priv->data->shadow_register && !mtk_crtc->cmdq_client)
-#else
-	if (!priv->data->shadow_register)
-#endif
-		mtk_crtc_ddp_config(crtc, NULL);
-
-	mtk_drm_finish_page_flip(mtk_crtc);
-}
-
 static int mtk_drm_crtc_num_comp_planes(struct mtk_drm_crtc *mtk_crtc,
 					int comp_idx)
 {
@@ -771,7 +772,7 @@ int mtk_drm_crtc_create(struct drm_device *drm_dev,
 	if (!mtk_crtc->ddp_comp)
 		return -ENOMEM;
 
-	mtk_crtc->mutex = mtk_disp_mutex_get(priv->mutex_dev, pipe);
+	mtk_crtc->mutex = mtk_mutex_get(priv->mutex_dev);
 	if (IS_ERR(mtk_crtc->mutex)) {
 		ret = PTR_ERR(mtk_crtc->mutex);
 		dev_err(dev, "Failed to get mutex: %d\n", ret);
@@ -784,7 +785,7 @@ int mtk_drm_crtc_create(struct drm_device *drm_dev,
 		struct device_node *node;
 
 		node = priv->comp_node[comp_id];
-		comp = priv->ddp_comp[comp_id];
+		comp = &priv->ddp_comp[comp_id];
 		if (!comp) {
 			dev_err(dev, "Component %pOF not initialized\n", node);
 			ret = -ENODEV;
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_crtc.h b/drivers/gpu/drm/mediatek/mtk_drm_crtc.h
index de0479cd1e72..cb9a36c48d4f 100644
--- a/drivers/gpu/drm/mediatek/mtk_drm_crtc.h
+++ b/drivers/gpu/drm/mediatek/mtk_drm_crtc.h
@@ -15,7 +15,6 @@
 #define MTK_MIN_BPC	3
 
 void mtk_drm_crtc_commit(struct drm_crtc *crtc);
-void mtk_crtc_ddp_irq(struct drm_crtc *crtc, struct mtk_ddp_comp *comp);
 int mtk_drm_crtc_create(struct drm_device *drm_dev,
 			const enum mtk_ddp_comp_id *path,
 			unsigned int path_len);
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_ddp.c b/drivers/gpu/drm/mediatek/mtk_drm_ddp.c
deleted file mode 100644
index 1f99db6b1a42..000000000000
--- a/drivers/gpu/drm/mediatek/mtk_drm_ddp.c
+++ /dev/null
@@ -1,464 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-only
-/*
- * Copyright (c) 2015 MediaTek Inc.
- */
-
-#include <linux/clk.h>
-#include <linux/iopoll.h>
-#include <linux/module.h>
-#include <linux/of_device.h>
-#include <linux/platform_device.h>
-#include <linux/regmap.h>
-
-#include "mtk_drm_ddp.h"
-#include "mtk_drm_ddp_comp.h"
-
-#define MT2701_DISP_MUTEX0_MOD0			0x2c
-#define MT2701_DISP_MUTEX0_SOF0			0x30
-
-#define DISP_REG_MUTEX_EN(n)			(0x20 + 0x20 * (n))
-#define DISP_REG_MUTEX(n)			(0x24 + 0x20 * (n))
-#define DISP_REG_MUTEX_RST(n)			(0x28 + 0x20 * (n))
-#define DISP_REG_MUTEX_MOD(mutex_mod_reg, n)	(mutex_mod_reg + 0x20 * (n))
-#define DISP_REG_MUTEX_SOF(mutex_sof_reg, n)	(mutex_sof_reg + 0x20 * (n))
-#define DISP_REG_MUTEX_MOD2(n)			(0x34 + 0x20 * (n))
-
-#define INT_MUTEX				BIT(1)
-
-#define MT8167_MUTEX_MOD_DISP_PWM		1
-#define MT8167_MUTEX_MOD_DISP_OVL0		6
-#define MT8167_MUTEX_MOD_DISP_OVL1		7
-#define MT8167_MUTEX_MOD_DISP_RDMA0		8
-#define MT8167_MUTEX_MOD_DISP_RDMA1		9
-#define MT8167_MUTEX_MOD_DISP_WDMA0		10
-#define MT8167_MUTEX_MOD_DISP_CCORR		11
-#define MT8167_MUTEX_MOD_DISP_COLOR		12
-#define MT8167_MUTEX_MOD_DISP_AAL		13
-#define MT8167_MUTEX_MOD_DISP_GAMMA		14
-#define MT8167_MUTEX_MOD_DISP_DITHER		15
-#define MT8167_MUTEX_MOD_DISP_UFOE		16
-
-#define MT8173_MUTEX_MOD_DISP_OVL0		11
-#define MT8173_MUTEX_MOD_DISP_OVL1		12
-#define MT8173_MUTEX_MOD_DISP_RDMA0		13
-#define MT8173_MUTEX_MOD_DISP_RDMA1		14
-#define MT8173_MUTEX_MOD_DISP_RDMA2		15
-#define MT8173_MUTEX_MOD_DISP_WDMA0		16
-#define MT8173_MUTEX_MOD_DISP_WDMA1		17
-#define MT8173_MUTEX_MOD_DISP_COLOR0		18
-#define MT8173_MUTEX_MOD_DISP_COLOR1		19
-#define MT8173_MUTEX_MOD_DISP_AAL		20
-#define MT8173_MUTEX_MOD_DISP_GAMMA		21
-#define MT8173_MUTEX_MOD_DISP_UFOE		22
-#define MT8173_MUTEX_MOD_DISP_PWM0		23
-#define MT8173_MUTEX_MOD_DISP_PWM1		24
-#define MT8173_MUTEX_MOD_DISP_OD		25
-
-#define MT2712_MUTEX_MOD_DISP_PWM2		10
-#define MT2712_MUTEX_MOD_DISP_OVL0		11
-#define MT2712_MUTEX_MOD_DISP_OVL1		12
-#define MT2712_MUTEX_MOD_DISP_RDMA0		13
-#define MT2712_MUTEX_MOD_DISP_RDMA1		14
-#define MT2712_MUTEX_MOD_DISP_RDMA2		15
-#define MT2712_MUTEX_MOD_DISP_WDMA0		16
-#define MT2712_MUTEX_MOD_DISP_WDMA1		17
-#define MT2712_MUTEX_MOD_DISP_COLOR0		18
-#define MT2712_MUTEX_MOD_DISP_COLOR1		19
-#define MT2712_MUTEX_MOD_DISP_AAL0		20
-#define MT2712_MUTEX_MOD_DISP_UFOE		22
-#define MT2712_MUTEX_MOD_DISP_PWM0		23
-#define MT2712_MUTEX_MOD_DISP_PWM1		24
-#define MT2712_MUTEX_MOD_DISP_OD0		25
-#define MT2712_MUTEX_MOD2_DISP_AAL1		33
-#define MT2712_MUTEX_MOD2_DISP_OD1		34
-
-#define MT2701_MUTEX_MOD_DISP_OVL		3
-#define MT2701_MUTEX_MOD_DISP_WDMA		6
-#define MT2701_MUTEX_MOD_DISP_COLOR		7
-#define MT2701_MUTEX_MOD_DISP_BLS		9
-#define MT2701_MUTEX_MOD_DISP_RDMA0		10
-#define MT2701_MUTEX_MOD_DISP_RDMA1		12
-
-#define MUTEX_SOF_SINGLE_MODE		0
-#define MUTEX_SOF_DSI0			1
-#define MUTEX_SOF_DSI1			2
-#define MUTEX_SOF_DPI0			3
-#define MUTEX_SOF_DPI1			4
-#define MUTEX_SOF_DSI2			5
-#define MUTEX_SOF_DSI3			6
-#define MT8167_MUTEX_SOF_DPI0		2
-#define MT8167_MUTEX_SOF_DPI1		3
-
-
-struct mtk_disp_mutex {
-	int id;
-	bool claimed;
-};
-
-enum mtk_ddp_mutex_sof_id {
-	DDP_MUTEX_SOF_SINGLE_MODE,
-	DDP_MUTEX_SOF_DSI0,
-	DDP_MUTEX_SOF_DSI1,
-	DDP_MUTEX_SOF_DPI0,
-	DDP_MUTEX_SOF_DPI1,
-	DDP_MUTEX_SOF_DSI2,
-	DDP_MUTEX_SOF_DSI3,
-};
-
-struct mtk_ddp_data {
-	const unsigned int *mutex_mod;
-	const unsigned int *mutex_sof;
-	const unsigned int mutex_mod_reg;
-	const unsigned int mutex_sof_reg;
-	const bool no_clk;
-};
-
-struct mtk_ddp {
-	struct device			*dev;
-	struct clk			*clk;
-	void __iomem			*regs;
-	struct mtk_disp_mutex		mutex[10];
-	const struct mtk_ddp_data	*data;
-};
-
-static const unsigned int mt2701_mutex_mod[DDP_COMPONENT_ID_MAX] = {
-	[DDP_COMPONENT_BLS] = MT2701_MUTEX_MOD_DISP_BLS,
-	[DDP_COMPONENT_COLOR0] = MT2701_MUTEX_MOD_DISP_COLOR,
-	[DDP_COMPONENT_OVL0] = MT2701_MUTEX_MOD_DISP_OVL,
-	[DDP_COMPONENT_RDMA0] = MT2701_MUTEX_MOD_DISP_RDMA0,
-	[DDP_COMPONENT_RDMA1] = MT2701_MUTEX_MOD_DISP_RDMA1,
-	[DDP_COMPONENT_WDMA0] = MT2701_MUTEX_MOD_DISP_WDMA,
-};
-
-static const unsigned int mt2712_mutex_mod[DDP_COMPONENT_ID_MAX] = {
-	[DDP_COMPONENT_AAL0] = MT2712_MUTEX_MOD_DISP_AAL0,
-	[DDP_COMPONENT_AAL1] = MT2712_MUTEX_MOD2_DISP_AAL1,
-	[DDP_COMPONENT_COLOR0] = MT2712_MUTEX_MOD_DISP_COLOR0,
-	[DDP_COMPONENT_COLOR1] = MT2712_MUTEX_MOD_DISP_COLOR1,
-	[DDP_COMPONENT_OD0] = MT2712_MUTEX_MOD_DISP_OD0,
-	[DDP_COMPONENT_OD1] = MT2712_MUTEX_MOD2_DISP_OD1,
-	[DDP_COMPONENT_OVL0] = MT2712_MUTEX_MOD_DISP_OVL0,
-	[DDP_COMPONENT_OVL1] = MT2712_MUTEX_MOD_DISP_OVL1,
-	[DDP_COMPONENT_PWM0] = MT2712_MUTEX_MOD_DISP_PWM0,
-	[DDP_COMPONENT_PWM1] = MT2712_MUTEX_MOD_DISP_PWM1,
-	[DDP_COMPONENT_PWM2] = MT2712_MUTEX_MOD_DISP_PWM2,
-	[DDP_COMPONENT_RDMA0] = MT2712_MUTEX_MOD_DISP_RDMA0,
-	[DDP_COMPONENT_RDMA1] = MT2712_MUTEX_MOD_DISP_RDMA1,
-	[DDP_COMPONENT_RDMA2] = MT2712_MUTEX_MOD_DISP_RDMA2,
-	[DDP_COMPONENT_UFOE] = MT2712_MUTEX_MOD_DISP_UFOE,
-	[DDP_COMPONENT_WDMA0] = MT2712_MUTEX_MOD_DISP_WDMA0,
-	[DDP_COMPONENT_WDMA1] = MT2712_MUTEX_MOD_DISP_WDMA1,
-};
-
-static const unsigned int mt8167_mutex_mod[DDP_COMPONENT_ID_MAX] = {
-	[DDP_COMPONENT_AAL0] = MT8167_MUTEX_MOD_DISP_AAL,
-	[DDP_COMPONENT_CCORR] = MT8167_MUTEX_MOD_DISP_CCORR,
-	[DDP_COMPONENT_COLOR0] = MT8167_MUTEX_MOD_DISP_COLOR,
-	[DDP_COMPONENT_DITHER] = MT8167_MUTEX_MOD_DISP_DITHER,
-	[DDP_COMPONENT_GAMMA] = MT8167_MUTEX_MOD_DISP_GAMMA,
-	[DDP_COMPONENT_OVL0] = MT8167_MUTEX_MOD_DISP_OVL0,
-	[DDP_COMPONENT_OVL1] = MT8167_MUTEX_MOD_DISP_OVL1,
-	[DDP_COMPONENT_PWM0] = MT8167_MUTEX_MOD_DISP_PWM,
-	[DDP_COMPONENT_RDMA0] = MT8167_MUTEX_MOD_DISP_RDMA0,
-	[DDP_COMPONENT_RDMA1] = MT8167_MUTEX_MOD_DISP_RDMA1,
-	[DDP_COMPONENT_UFOE] = MT8167_MUTEX_MOD_DISP_UFOE,
-	[DDP_COMPONENT_WDMA0] = MT8167_MUTEX_MOD_DISP_WDMA0,
-};
-
-static const unsigned int mt8173_mutex_mod[DDP_COMPONENT_ID_MAX] = {
-	[DDP_COMPONENT_AAL0] = MT8173_MUTEX_MOD_DISP_AAL,
-	[DDP_COMPONENT_COLOR0] = MT8173_MUTEX_MOD_DISP_COLOR0,
-	[DDP_COMPONENT_COLOR1] = MT8173_MUTEX_MOD_DISP_COLOR1,
-	[DDP_COMPONENT_GAMMA] = MT8173_MUTEX_MOD_DISP_GAMMA,
-	[DDP_COMPONENT_OD0] = MT8173_MUTEX_MOD_DISP_OD,
-	[DDP_COMPONENT_OVL0] = MT8173_MUTEX_MOD_DISP_OVL0,
-	[DDP_COMPONENT_OVL1] = MT8173_MUTEX_MOD_DISP_OVL1,
-	[DDP_COMPONENT_PWM0] = MT8173_MUTEX_MOD_DISP_PWM0,
-	[DDP_COMPONENT_PWM1] = MT8173_MUTEX_MOD_DISP_PWM1,
-	[DDP_COMPONENT_RDMA0] = MT8173_MUTEX_MOD_DISP_RDMA0,
-	[DDP_COMPONENT_RDMA1] = MT8173_MUTEX_MOD_DISP_RDMA1,
-	[DDP_COMPONENT_RDMA2] = MT8173_MUTEX_MOD_DISP_RDMA2,
-	[DDP_COMPONENT_UFOE] = MT8173_MUTEX_MOD_DISP_UFOE,
-	[DDP_COMPONENT_WDMA0] = MT8173_MUTEX_MOD_DISP_WDMA0,
-	[DDP_COMPONENT_WDMA1] = MT8173_MUTEX_MOD_DISP_WDMA1,
-};
-
-static const unsigned int mt2712_mutex_sof[DDP_MUTEX_SOF_DSI3 + 1] = {
-	[DDP_MUTEX_SOF_SINGLE_MODE] = MUTEX_SOF_SINGLE_MODE,
-	[DDP_MUTEX_SOF_DSI0] = MUTEX_SOF_DSI0,
-	[DDP_MUTEX_SOF_DSI1] = MUTEX_SOF_DSI1,
-	[DDP_MUTEX_SOF_DPI0] = MUTEX_SOF_DPI0,
-	[DDP_MUTEX_SOF_DPI1] = MUTEX_SOF_DPI1,
-	[DDP_MUTEX_SOF_DSI2] = MUTEX_SOF_DSI2,
-	[DDP_MUTEX_SOF_DSI3] = MUTEX_SOF_DSI3,
-};
-
-static const unsigned int mt8167_mutex_sof[DDP_MUTEX_SOF_DSI3 + 1] = {
-	[DDP_MUTEX_SOF_SINGLE_MODE] = MUTEX_SOF_SINGLE_MODE,
-	[DDP_MUTEX_SOF_DSI0] = MUTEX_SOF_DSI0,
-	[DDP_MUTEX_SOF_DPI0] = MT8167_MUTEX_SOF_DPI0,
-	[DDP_MUTEX_SOF_DPI1] = MT8167_MUTEX_SOF_DPI1,
-};
-
-static const struct mtk_ddp_data mt2701_ddp_driver_data = {
-	.mutex_mod = mt2701_mutex_mod,
-	.mutex_sof = mt2712_mutex_sof,
-	.mutex_mod_reg = MT2701_DISP_MUTEX0_MOD0,
-	.mutex_sof_reg = MT2701_DISP_MUTEX0_SOF0,
-};
-
-static const struct mtk_ddp_data mt2712_ddp_driver_data = {
-	.mutex_mod = mt2712_mutex_mod,
-	.mutex_sof = mt2712_mutex_sof,
-	.mutex_mod_reg = MT2701_DISP_MUTEX0_MOD0,
-	.mutex_sof_reg = MT2701_DISP_MUTEX0_SOF0,
-};
-
-static const struct mtk_ddp_data mt8167_ddp_driver_data = {
-	.mutex_mod = mt8167_mutex_mod,
-	.mutex_sof = mt8167_mutex_sof,
-	.mutex_mod_reg = MT2701_DISP_MUTEX0_MOD0,
-	.mutex_sof_reg = MT2701_DISP_MUTEX0_SOF0,
-	.no_clk = true,
-};
-
-static const struct mtk_ddp_data mt8173_ddp_driver_data = {
-	.mutex_mod = mt8173_mutex_mod,
-	.mutex_sof = mt2712_mutex_sof,
-	.mutex_mod_reg = MT2701_DISP_MUTEX0_MOD0,
-	.mutex_sof_reg = MT2701_DISP_MUTEX0_SOF0,
-};
-
-struct mtk_disp_mutex *mtk_disp_mutex_get(struct device *dev, unsigned int id)
-{
-	struct mtk_ddp *ddp = dev_get_drvdata(dev);
-
-	if (id >= 10)
-		return ERR_PTR(-EINVAL);
-	if (ddp->mutex[id].claimed)
-		return ERR_PTR(-EBUSY);
-
-	ddp->mutex[id].claimed = true;
-
-	return &ddp->mutex[id];
-}
-
-void mtk_disp_mutex_put(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-
-	WARN_ON(&ddp->mutex[mutex->id] != mutex);
-
-	mutex->claimed = false;
-}
-
-int mtk_disp_mutex_prepare(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-	return clk_prepare_enable(ddp->clk);
-}
-
-void mtk_disp_mutex_unprepare(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-	clk_disable_unprepare(ddp->clk);
-}
-
-void mtk_disp_mutex_add_comp(struct mtk_disp_mutex *mutex,
-			     enum mtk_ddp_comp_id id)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-	unsigned int reg;
-	unsigned int sof_id;
-	unsigned int offset;
-
-	WARN_ON(&ddp->mutex[mutex->id] != mutex);
-
-	switch (id) {
-	case DDP_COMPONENT_DSI0:
-		sof_id = DDP_MUTEX_SOF_DSI0;
-		break;
-	case DDP_COMPONENT_DSI1:
-		sof_id = DDP_MUTEX_SOF_DSI0;
-		break;
-	case DDP_COMPONENT_DSI2:
-		sof_id = DDP_MUTEX_SOF_DSI2;
-		break;
-	case DDP_COMPONENT_DSI3:
-		sof_id = DDP_MUTEX_SOF_DSI3;
-		break;
-	case DDP_COMPONENT_DPI0:
-		sof_id = DDP_MUTEX_SOF_DPI0;
-		break;
-	case DDP_COMPONENT_DPI1:
-		sof_id = DDP_MUTEX_SOF_DPI1;
-		break;
-	default:
-		if (ddp->data->mutex_mod[id] < 32) {
-			offset = DISP_REG_MUTEX_MOD(ddp->data->mutex_mod_reg,
-						    mutex->id);
-			reg = readl_relaxed(ddp->regs + offset);
-			reg |= 1 << ddp->data->mutex_mod[id];
-			writel_relaxed(reg, ddp->regs + offset);
-		} else {
-			offset = DISP_REG_MUTEX_MOD2(mutex->id);
-			reg = readl_relaxed(ddp->regs + offset);
-			reg |= 1 << (ddp->data->mutex_mod[id] - 32);
-			writel_relaxed(reg, ddp->regs + offset);
-		}
-		return;
-	}
-
-	writel_relaxed(ddp->data->mutex_sof[sof_id],
-		       ddp->regs +
-		       DISP_REG_MUTEX_SOF(ddp->data->mutex_sof_reg, mutex->id));
-}
-
-void mtk_disp_mutex_remove_comp(struct mtk_disp_mutex *mutex,
-				enum mtk_ddp_comp_id id)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-	unsigned int reg;
-	unsigned int offset;
-
-	WARN_ON(&ddp->mutex[mutex->id] != mutex);
-
-	switch (id) {
-	case DDP_COMPONENT_DSI0:
-	case DDP_COMPONENT_DSI1:
-	case DDP_COMPONENT_DSI2:
-	case DDP_COMPONENT_DSI3:
-	case DDP_COMPONENT_DPI0:
-	case DDP_COMPONENT_DPI1:
-		writel_relaxed(MUTEX_SOF_SINGLE_MODE,
-			       ddp->regs +
-			       DISP_REG_MUTEX_SOF(ddp->data->mutex_sof_reg,
-						  mutex->id));
-		break;
-	default:
-		if (ddp->data->mutex_mod[id] < 32) {
-			offset = DISP_REG_MUTEX_MOD(ddp->data->mutex_mod_reg,
-						    mutex->id);
-			reg = readl_relaxed(ddp->regs + offset);
-			reg &= ~(1 << ddp->data->mutex_mod[id]);
-			writel_relaxed(reg, ddp->regs + offset);
-		} else {
-			offset = DISP_REG_MUTEX_MOD2(mutex->id);
-			reg = readl_relaxed(ddp->regs + offset);
-			reg &= ~(1 << (ddp->data->mutex_mod[id] - 32));
-			writel_relaxed(reg, ddp->regs + offset);
-		}
-		break;
-	}
-}
-
-void mtk_disp_mutex_enable(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-
-	WARN_ON(&ddp->mutex[mutex->id] != mutex);
-
-	writel(1, ddp->regs + DISP_REG_MUTEX_EN(mutex->id));
-}
-
-void mtk_disp_mutex_disable(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-
-	WARN_ON(&ddp->mutex[mutex->id] != mutex);
-
-	writel(0, ddp->regs + DISP_REG_MUTEX_EN(mutex->id));
-}
-
-void mtk_disp_mutex_acquire(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-	u32 tmp;
-
-	writel(1, ddp->regs + DISP_REG_MUTEX_EN(mutex->id));
-	writel(1, ddp->regs + DISP_REG_MUTEX(mutex->id));
-	if (readl_poll_timeout_atomic(ddp->regs + DISP_REG_MUTEX(mutex->id),
-				      tmp, tmp & INT_MUTEX, 1, 10000))
-		pr_err("could not acquire mutex %d\n", mutex->id);
-}
-
-void mtk_disp_mutex_release(struct mtk_disp_mutex *mutex)
-{
-	struct mtk_ddp *ddp = container_of(mutex, struct mtk_ddp,
-					   mutex[mutex->id]);
-
-	writel(0, ddp->regs + DISP_REG_MUTEX(mutex->id));
-}
-
-static int mtk_ddp_probe(struct platform_device *pdev)
-{
-	struct device *dev = &pdev->dev;
-	struct mtk_ddp *ddp;
-	struct resource *regs;
-	int i;
-
-	ddp = devm_kzalloc(dev, sizeof(*ddp), GFP_KERNEL);
-	if (!ddp)
-		return -ENOMEM;
-
-	for (i = 0; i < 10; i++)
-		ddp->mutex[i].id = i;
-
-	ddp->data = of_device_get_match_data(dev);
-
-	if (!ddp->data->no_clk) {
-		ddp->clk = devm_clk_get(dev, NULL);
-		if (IS_ERR(ddp->clk)) {
-			if (PTR_ERR(ddp->clk) != -EPROBE_DEFER)
-				dev_err(dev, "Failed to get clock\n");
-			return PTR_ERR(ddp->clk);
-		}
-	}
-
-	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	ddp->regs = devm_ioremap_resource(dev, regs);
-	if (IS_ERR(ddp->regs)) {
-		dev_err(dev, "Failed to map mutex registers\n");
-		return PTR_ERR(ddp->regs);
-	}
-
-	platform_set_drvdata(pdev, ddp);
-
-	return 0;
-}
-
-static int mtk_ddp_remove(struct platform_device *pdev)
-{
-	return 0;
-}
-
-static const struct of_device_id ddp_driver_dt_match[] = {
-	{ .compatible = "mediatek,mt2701-disp-mutex",
-	  .data = &mt2701_ddp_driver_data},
-	{ .compatible = "mediatek,mt2712-disp-mutex",
-	  .data = &mt2712_ddp_driver_data},
-	{ .compatible = "mediatek,mt8167-disp-mutex",
-	  .data = &mt8167_ddp_driver_data},
-	{ .compatible = "mediatek,mt8173-disp-mutex",
-	  .data = &mt8173_ddp_driver_data},
-	{},
-};
-MODULE_DEVICE_TABLE(of, ddp_driver_dt_match);
-
-struct platform_driver mtk_ddp_driver = {
-	.probe		= mtk_ddp_probe,
-	.remove		= mtk_ddp_remove,
-	.driver		= {
-		.name	= "mediatek-ddp",
-		.owner	= THIS_MODULE,
-		.of_match_table = ddp_driver_dt_match,
-	},
-};
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_ddp.h b/drivers/gpu/drm/mediatek/mtk_drm_ddp.h
deleted file mode 100644
index 6b691a57be4a..000000000000
--- a/drivers/gpu/drm/mediatek/mtk_drm_ddp.h
+++ /dev/null
@@ -1,28 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-only */
-/*
- * Copyright (c) 2015 MediaTek Inc.
- */
-
-#ifndef MTK_DRM_DDP_H
-#define MTK_DRM_DDP_H
-
-#include "mtk_drm_ddp_comp.h"
-
-struct regmap;
-struct device;
-struct mtk_disp_mutex;
-
-struct mtk_disp_mutex *mtk_disp_mutex_get(struct device *dev, unsigned int id);
-int mtk_disp_mutex_prepare(struct mtk_disp_mutex *mutex);
-void mtk_disp_mutex_add_comp(struct mtk_disp_mutex *mutex,
-			     enum mtk_ddp_comp_id id);
-void mtk_disp_mutex_enable(struct mtk_disp_mutex *mutex);
-void mtk_disp_mutex_disable(struct mtk_disp_mutex *mutex);
-void mtk_disp_mutex_remove_comp(struct mtk_disp_mutex *mutex,
-				enum mtk_ddp_comp_id id);
-void mtk_disp_mutex_unprepare(struct mtk_disp_mutex *mutex);
-void mtk_disp_mutex_put(struct mtk_disp_mutex *mutex);
-void mtk_disp_mutex_acquire(struct mtk_disp_mutex *mutex);
-void mtk_disp_mutex_release(struct mtk_disp_mutex *mutex);
-
-#endif /* MTK_DRM_DDP_H */
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c b/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c
index 3064eac1a750..75bc00e17fc4 100644
--- a/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c
+++ b/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c
@@ -9,12 +9,12 @@
 #include <linux/clk.h>
 #include <linux/of.h>
 #include <linux/of_address.h>
-#include <linux/of_irq.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
 #include <linux/soc/mediatek/mtk-cmdq.h>
 #include <drm/drm_print.h>
 
+#include "mtk_disp_drv.h"
 #include "mtk_drm_drv.h"
 #include "mtk_drm_plane.h"
 #include "mtk_drm_ddp_comp.h"
@@ -35,31 +35,13 @@
 #define DISP_AAL_EN				0x0000
 #define DISP_AAL_SIZE				0x0030
 
-#define DISP_CCORR_EN				0x0000
-#define CCORR_EN				BIT(0)
-#define DISP_CCORR_CFG				0x0020
-#define CCORR_RELAY_MODE			BIT(0)
-#define CCORR_ENGINE_EN				BIT(1)
-#define CCORR_GAMMA_OFF				BIT(2)
-#define CCORR_WGAMUT_SRC_CLIP			BIT(3)
-#define DISP_CCORR_SIZE				0x0030
-#define DISP_CCORR_COEF_0			0x0080
-#define DISP_CCORR_COEF_1			0x0084
-#define DISP_CCORR_COEF_2			0x0088
-#define DISP_CCORR_COEF_3			0x008C
-#define DISP_CCORR_COEF_4			0x0090
-
 #define DISP_DITHER_EN				0x0000
 #define DITHER_EN				BIT(0)
 #define DISP_DITHER_CFG				0x0020
 #define DITHER_RELAY_MODE			BIT(0)
+#define DITHER_ENGINE_EN			BIT(1)
 #define DISP_DITHER_SIZE			0x0030
 
-#define DISP_GAMMA_EN				0x0000
-#define DISP_GAMMA_CFG				0x0020
-#define DISP_GAMMA_SIZE				0x0030
-#define DISP_GAMMA_LUT				0x0700
-
 #define LUT_10BIT_MASK				0x03ff
 
 #define OD_RELAYMODE				BIT(0)
@@ -68,9 +50,6 @@
 
 #define AAL_EN					BIT(0)
 
-#define GAMMA_EN				BIT(0)
-#define GAMMA_LUT_EN				BIT(1)
-
 #define DISP_DITHERING				BIT(2)
 #define DITHER_LSB_ERR_SHIFT_R(x)		(((x) & 0x7) << 28)
 #define DITHER_OVFLW_BIT_R(x)			(((x) & 0x7) << 24)
@@ -86,262 +65,233 @@
 #define DITHER_ADD_LSHIFT_G(x)			(((x) & 0x7) << 4)
 #define DITHER_ADD_RSHIFT_G(x)			(((x) & 0x7) << 0)
 
+struct mtk_ddp_comp_dev {
+	struct clk *clk;
+	void __iomem *regs;
+	struct cmdq_client_reg cmdq_reg;
+};
+
 void mtk_ddp_write(struct cmdq_pkt *cmdq_pkt, unsigned int value,
-		   struct mtk_ddp_comp *comp, unsigned int offset)
+		   struct cmdq_client_reg *cmdq_reg, void __iomem *regs,
+		   unsigned int offset)
 {
 #if IS_REACHABLE(CONFIG_MTK_CMDQ)
 	if (cmdq_pkt)
-		cmdq_pkt_write(cmdq_pkt, comp->subsys,
-			       comp->regs_pa + offset, value);
+		cmdq_pkt_write(cmdq_pkt, cmdq_reg->subsys,
+			       cmdq_reg->offset + offset, value);
 	else
 #endif
-		writel(value, comp->regs + offset);
+		writel(value, regs + offset);
 }
 
 void mtk_ddp_write_relaxed(struct cmdq_pkt *cmdq_pkt, unsigned int value,
-			   struct mtk_ddp_comp *comp,
+			   struct cmdq_client_reg *cmdq_reg, void __iomem *regs,
 			   unsigned int offset)
 {
 #if IS_REACHABLE(CONFIG_MTK_CMDQ)
 	if (cmdq_pkt)
-		cmdq_pkt_write(cmdq_pkt, comp->subsys,
-			       comp->regs_pa + offset, value);
+		cmdq_pkt_write(cmdq_pkt, cmdq_reg->subsys,
+			       cmdq_reg->offset + offset, value);
 	else
 #endif
-		writel_relaxed(value, comp->regs + offset);
+		writel_relaxed(value, regs + offset);
 }
 
-void mtk_ddp_write_mask(struct cmdq_pkt *cmdq_pkt,
-			unsigned int value,
-			struct mtk_ddp_comp *comp,
-			unsigned int offset,
-			unsigned int mask)
+void mtk_ddp_write_mask(struct cmdq_pkt *cmdq_pkt, unsigned int value,
+			struct cmdq_client_reg *cmdq_reg, void __iomem *regs,
+			unsigned int offset, unsigned int mask)
 {
 #if IS_REACHABLE(CONFIG_MTK_CMDQ)
 	if (cmdq_pkt) {
-		cmdq_pkt_write_mask(cmdq_pkt, comp->subsys,
-				    comp->regs_pa + offset, value, mask);
+		cmdq_pkt_write_mask(cmdq_pkt, cmdq_reg->subsys,
+				    cmdq_reg->offset + offset, value, mask);
 	} else {
 #endif
-		u32 tmp = readl(comp->regs + offset);
+		u32 tmp = readl(regs + offset);
 
 		tmp = (tmp & ~mask) | (value & mask);
-		writel(tmp, comp->regs + offset);
+		writel(tmp, regs + offset);
 #if IS_REACHABLE(CONFIG_MTK_CMDQ)
 	}
 #endif
 }
 
-void mtk_dither_set(struct mtk_ddp_comp *comp, unsigned int bpc,
-		    unsigned int CFG, struct cmdq_pkt *cmdq_pkt)
+static int mtk_ddp_clk_enable(struct device *dev)
+{
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
+
+	return clk_prepare_enable(priv->clk);
+}
+
+static void mtk_ddp_clk_disable(struct device *dev)
+{
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(priv->clk);
+}
+
+void mtk_dither_set_common(void __iomem *regs, struct cmdq_client_reg *cmdq_reg,
+			   unsigned int bpc, unsigned int cfg,
+			   unsigned int dither_en, struct cmdq_pkt *cmdq_pkt)
 {
 	/* If bpc equal to 0, the dithering function didn't be enabled */
 	if (bpc == 0)
 		return;
 
 	if (bpc >= MTK_MIN_BPC) {
-		mtk_ddp_write(cmdq_pkt, 0, comp, DISP_DITHER_5);
-		mtk_ddp_write(cmdq_pkt, 0, comp, DISP_DITHER_7);
+		mtk_ddp_write(cmdq_pkt, 0, cmdq_reg, regs, DISP_DITHER_5);
+		mtk_ddp_write(cmdq_pkt, 0, cmdq_reg, regs, DISP_DITHER_7);
 		mtk_ddp_write(cmdq_pkt,
 			      DITHER_LSB_ERR_SHIFT_R(MTK_MAX_BPC - bpc) |
 			      DITHER_ADD_LSHIFT_R(MTK_MAX_BPC - bpc) |
 			      DITHER_NEW_BIT_MODE,
-			      comp, DISP_DITHER_15);
+			      cmdq_reg, regs, DISP_DITHER_15);
 		mtk_ddp_write(cmdq_pkt,
 			      DITHER_LSB_ERR_SHIFT_B(MTK_MAX_BPC - bpc) |
 			      DITHER_ADD_LSHIFT_B(MTK_MAX_BPC - bpc) |
 			      DITHER_LSB_ERR_SHIFT_G(MTK_MAX_BPC - bpc) |
 			      DITHER_ADD_LSHIFT_G(MTK_MAX_BPC - bpc),
-			      comp, DISP_DITHER_16);
-		mtk_ddp_write(cmdq_pkt, DISP_DITHERING, comp, CFG);
+			      cmdq_reg, regs, DISP_DITHER_16);
+		mtk_ddp_write(cmdq_pkt, dither_en, cmdq_reg, regs, cfg);
 	}
 }
 
-static void mtk_od_config(struct mtk_ddp_comp *comp, unsigned int w,
-			  unsigned int h, unsigned int vrefresh,
-			  unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+static void mtk_dither_set(struct device *dev, unsigned int bpc,
+		    unsigned int cfg, struct cmdq_pkt *cmdq_pkt)
 {
-	mtk_ddp_write(cmdq_pkt, w << 16 | h, comp, DISP_OD_SIZE);
-	mtk_ddp_write(cmdq_pkt, OD_RELAYMODE, comp, DISP_OD_CFG);
-	mtk_dither_set(comp, bpc, DISP_OD_CFG, cmdq_pkt);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_od_start(struct mtk_ddp_comp *comp)
-{
-	writel(1, comp->regs + DISP_OD_EN);
+	mtk_dither_set_common(priv->regs, &priv->cmdq_reg, bpc, cfg,
+			      DISP_DITHERING, cmdq_pkt);
 }
 
-static void mtk_ufoe_start(struct mtk_ddp_comp *comp)
+static void mtk_od_config(struct device *dev, unsigned int w,
+			  unsigned int h, unsigned int vrefresh,
+			  unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
 {
-	writel(UFO_BYPASS, comp->regs + DISP_REG_UFO_START);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_aal_config(struct mtk_ddp_comp *comp, unsigned int w,
-			   unsigned int h, unsigned int vrefresh,
-			   unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
-{
-	mtk_ddp_write(cmdq_pkt, h << 16 | w, comp, DISP_AAL_SIZE);
+	mtk_ddp_write(cmdq_pkt, w << 16 | h, &priv->cmdq_reg, priv->regs, DISP_OD_SIZE);
+	mtk_ddp_write(cmdq_pkt, OD_RELAYMODE, &priv->cmdq_reg, priv->regs, DISP_OD_CFG);
+	mtk_dither_set(dev, bpc, DISP_OD_CFG, cmdq_pkt);
 }
 
-static void mtk_aal_start(struct mtk_ddp_comp *comp)
+static void mtk_od_start(struct device *dev)
 {
-	writel(AAL_EN, comp->regs + DISP_AAL_EN);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_aal_stop(struct mtk_ddp_comp *comp)
-{
-	writel_relaxed(0x0, comp->regs + DISP_AAL_EN);
+	writel(1, priv->regs + DISP_OD_EN);
 }
 
-static void mtk_ccorr_config(struct mtk_ddp_comp *comp, unsigned int w,
-			     unsigned int h, unsigned int vrefresh,
-			     unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+static void mtk_ufoe_start(struct device *dev)
 {
-	mtk_ddp_write(cmdq_pkt, h << 16 | w, comp, DISP_CCORR_SIZE);
-	mtk_ddp_write(cmdq_pkt, CCORR_ENGINE_EN, comp, DISP_CCORR_CFG);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_ccorr_start(struct mtk_ddp_comp *comp)
-{
-	writel(CCORR_EN, comp->regs + DISP_CCORR_EN);
+	writel(UFO_BYPASS, priv->regs + DISP_REG_UFO_START);
 }
 
-static void mtk_ccorr_stop(struct mtk_ddp_comp *comp)
+static void mtk_aal_config(struct device *dev, unsigned int w,
+			   unsigned int h, unsigned int vrefresh,
+			   unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
 {
-	writel_relaxed(0x0, comp->regs + DISP_CCORR_EN);
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
+
+	mtk_ddp_write(cmdq_pkt, w << 16 | h, &priv->cmdq_reg, priv->regs, DISP_AAL_SIZE);
 }
 
-/* Converts a DRM S31.32 value to the HW S1.10 format. */
-static u16 mtk_ctm_s31_32_to_s1_10(u64 in)
+static void mtk_aal_gamma_set(struct device *dev, struct drm_crtc_state *state)
 {
-	u16 r;
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-	/* Sign bit. */
-	r = in & BIT_ULL(63) ? BIT(11) : 0;
+	mtk_gamma_set_common(priv->regs, state);
+}
 
-	if ((in & GENMASK_ULL(62, 33)) > 0) {
-		/* identity value 0x100000000 -> 0x400, */
-		/* if bigger this, set it to max 0x7ff. */
-		r |= GENMASK(10, 0);
-	} else {
-		/* take the 11 most important bits. */
-		r |= (in >> 22) & GENMASK(10, 0);
-	}
+static void mtk_aal_start(struct device *dev)
+{
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-	return r;
+	writel(AAL_EN, priv->regs + DISP_AAL_EN);
 }
 
-static void mtk_ccorr_ctm_set(struct mtk_ddp_comp *comp,
-			      struct drm_crtc_state *state)
+static void mtk_aal_stop(struct device *dev)
 {
-	struct drm_property_blob *blob = state->ctm;
-	struct drm_color_ctm *ctm;
-	const u64 *input;
-	uint16_t coeffs[9] = { 0 };
-	int i;
-	struct cmdq_pkt *cmdq_pkt = NULL;
-
-	if (!blob)
-		return;
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-	ctm = (struct drm_color_ctm *)blob->data;
-	input = ctm->matrix;
-
-	for (i = 0; i < ARRAY_SIZE(coeffs); i++)
-		coeffs[i] = mtk_ctm_s31_32_to_s1_10(input[i]);
-
-	mtk_ddp_write(cmdq_pkt, coeffs[0] << 16 | coeffs[1],
-		      comp, DISP_CCORR_COEF_0);
-	mtk_ddp_write(cmdq_pkt, coeffs[2] << 16 | coeffs[3],
-		      comp, DISP_CCORR_COEF_1);
-	mtk_ddp_write(cmdq_pkt, coeffs[4] << 16 | coeffs[5],
-		      comp, DISP_CCORR_COEF_2);
-	mtk_ddp_write(cmdq_pkt, coeffs[6] << 16 | coeffs[7],
-		      comp, DISP_CCORR_COEF_3);
-	mtk_ddp_write(cmdq_pkt, coeffs[8] << 16,
-		      comp, DISP_CCORR_COEF_4);
+	writel_relaxed(0x0, priv->regs + DISP_AAL_EN);
 }
 
-static void mtk_dither_config(struct mtk_ddp_comp *comp, unsigned int w,
+static void mtk_dither_config(struct device *dev, unsigned int w,
 			      unsigned int h, unsigned int vrefresh,
 			      unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
 {
-	mtk_ddp_write(cmdq_pkt, h << 16 | w, comp, DISP_DITHER_SIZE);
-	mtk_ddp_write(cmdq_pkt, DITHER_RELAY_MODE, comp, DISP_DITHER_CFG);
-}
-
-static void mtk_dither_start(struct mtk_ddp_comp *comp)
-{
-	writel(DITHER_EN, comp->regs + DISP_DITHER_EN);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_dither_stop(struct mtk_ddp_comp *comp)
-{
-	writel_relaxed(0x0, comp->regs + DISP_DITHER_EN);
+	mtk_ddp_write(cmdq_pkt, h << 16 | w, &priv->cmdq_reg, priv->regs, DISP_DITHER_SIZE);
+	mtk_ddp_write(cmdq_pkt, DITHER_RELAY_MODE, &priv->cmdq_reg, priv->regs, DISP_DITHER_CFG);
+	mtk_dither_set_common(priv->regs, &priv->cmdq_reg, bpc, DISP_DITHER_CFG,
+			      DITHER_ENGINE_EN, cmdq_pkt);
 }
 
-static void mtk_gamma_config(struct mtk_ddp_comp *comp, unsigned int w,
-			     unsigned int h, unsigned int vrefresh,
-			     unsigned int bpc, struct cmdq_pkt *cmdq_pkt)
+static void mtk_dither_start(struct device *dev)
 {
-	mtk_ddp_write(cmdq_pkt, h << 16 | w, comp, DISP_GAMMA_SIZE);
-	mtk_dither_set(comp, bpc, DISP_GAMMA_CFG, cmdq_pkt);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_gamma_start(struct mtk_ddp_comp *comp)
-{
-	writel(GAMMA_EN, comp->regs  + DISP_GAMMA_EN);
+	writel(DITHER_EN, priv->regs + DISP_DITHER_EN);
 }
 
-static void mtk_gamma_stop(struct mtk_ddp_comp *comp)
+static void mtk_dither_stop(struct device *dev)
 {
-	writel_relaxed(0x0, comp->regs  + DISP_GAMMA_EN);
-}
+	struct mtk_ddp_comp_dev *priv = dev_get_drvdata(dev);
 
-static void mtk_gamma_set(struct mtk_ddp_comp *comp,
-			  struct drm_crtc_state *state)
-{
-	unsigned int i, reg;
-	struct drm_color_lut *lut;
-	void __iomem *lut_base;
-	u32 word;
-
-	if (state->gamma_lut) {
-		reg = readl(comp->regs + DISP_GAMMA_CFG);
-		reg = reg | GAMMA_LUT_EN;
-		writel(reg, comp->regs + DISP_GAMMA_CFG);
-		lut_base = comp->regs + DISP_GAMMA_LUT;
-		lut = (struct drm_color_lut *)state->gamma_lut->data;
-		for (i = 0; i < MTK_LUT_SIZE; i++) {
-			word = (((lut[i].red >> 6) & LUT_10BIT_MASK) << 20) +
-				(((lut[i].green >> 6) & LUT_10BIT_MASK) << 10) +
-				((lut[i].blue >> 6) & LUT_10BIT_MASK);
-			writel(word, (lut_base + i * 4));
-		}
-	}
+	writel_relaxed(0x0, priv->regs + DISP_DITHER_EN);
 }
 
 static const struct mtk_ddp_comp_funcs ddp_aal = {
-	.gamma_set = mtk_gamma_set,
+	.clk_enable = mtk_ddp_clk_enable,
+	.clk_disable = mtk_ddp_clk_disable,
+	.gamma_set = mtk_aal_gamma_set,
 	.config = mtk_aal_config,
 	.start = mtk_aal_start,
 	.stop = mtk_aal_stop,
 };
 
 static const struct mtk_ddp_comp_funcs ddp_ccorr = {
+	.clk_enable = mtk_ccorr_clk_enable,
+	.clk_disable = mtk_ccorr_clk_disable,
 	.config = mtk_ccorr_config,
 	.start = mtk_ccorr_start,
 	.stop = mtk_ccorr_stop,
 	.ctm_set = mtk_ccorr_ctm_set,
 };
 
+static const struct mtk_ddp_comp_funcs ddp_color = {
+	.clk_enable = mtk_color_clk_enable,
+	.clk_disable = mtk_color_clk_disable,
+	.config = mtk_color_config,
+	.start = mtk_color_start,
+};
+
 static const struct mtk_ddp_comp_funcs ddp_dither = {
+	.clk_enable = mtk_ddp_clk_enable,
+	.clk_disable = mtk_ddp_clk_disable,
 	.config = mtk_dither_config,
 	.start = mtk_dither_start,
 	.stop = mtk_dither_stop,
 };
 
+static const struct mtk_ddp_comp_funcs ddp_dpi = {
+	.start = mtk_dpi_start,
+	.stop = mtk_dpi_stop,
+};
+
+static const struct mtk_ddp_comp_funcs ddp_dsi = {
+	.start = mtk_dsi_ddp_start,
+	.stop = mtk_dsi_ddp_stop,
+};
+
 static const struct mtk_ddp_comp_funcs ddp_gamma = {
+	.clk_enable = mtk_gamma_clk_enable,
+	.clk_disable = mtk_gamma_clk_disable,
 	.gamma_set = mtk_gamma_set,
 	.config = mtk_gamma_config,
 	.start = mtk_gamma_start,
@@ -349,11 +299,43 @@ static const struct mtk_ddp_comp_funcs ddp_gamma = {
 };
 
 static const struct mtk_ddp_comp_funcs ddp_od = {
+	.clk_enable = mtk_ddp_clk_enable,
+	.clk_disable = mtk_ddp_clk_disable,
 	.config = mtk_od_config,
 	.start = mtk_od_start,
 };
 
+static const struct mtk_ddp_comp_funcs ddp_ovl = {
+	.clk_enable = mtk_ovl_clk_enable,
+	.clk_disable = mtk_ovl_clk_disable,
+	.config = mtk_ovl_config,
+	.start = mtk_ovl_start,
+	.stop = mtk_ovl_stop,
+	.enable_vblank = mtk_ovl_enable_vblank,
+	.disable_vblank = mtk_ovl_disable_vblank,
+	.supported_rotations = mtk_ovl_supported_rotations,
+	.layer_nr = mtk_ovl_layer_nr,
+	.layer_check = mtk_ovl_layer_check,
+	.layer_config = mtk_ovl_layer_config,
+	.bgclr_in_on = mtk_ovl_bgclr_in_on,
+	.bgclr_in_off = mtk_ovl_bgclr_in_off,
+};
+
+static const struct mtk_ddp_comp_funcs ddp_rdma = {
+	.clk_enable = mtk_rdma_clk_enable,
+	.clk_disable = mtk_rdma_clk_disable,
+	.config = mtk_rdma_config,
+	.start = mtk_rdma_start,
+	.stop = mtk_rdma_stop,
+	.enable_vblank = mtk_rdma_enable_vblank,
+	.disable_vblank = mtk_rdma_disable_vblank,
+	.layer_nr = mtk_rdma_layer_nr,
+	.layer_config = mtk_rdma_layer_config,
+};
+
 static const struct mtk_ddp_comp_funcs ddp_ufoe = {
+	.clk_enable = mtk_ddp_clk_enable,
+	.clk_disable = mtk_ddp_clk_disable,
 	.start = mtk_ufoe_start,
 };
 
@@ -387,36 +369,37 @@ static const struct mtk_ddp_comp_match mtk_ddp_matches[DDP_COMPONENT_ID_MAX] = {
 	[DDP_COMPONENT_AAL1]	= { MTK_DISP_AAL,	1, &ddp_aal },
 	[DDP_COMPONENT_BLS]	= { MTK_DISP_BLS,	0, NULL },
 	[DDP_COMPONENT_CCORR]	= { MTK_DISP_CCORR,	0, &ddp_ccorr },
-	[DDP_COMPONENT_COLOR0]	= { MTK_DISP_COLOR,	0, NULL },
-	[DDP_COMPONENT_COLOR1]	= { MTK_DISP_COLOR,	1, NULL },
+	[DDP_COMPONENT_COLOR0]	= { MTK_DISP_COLOR,	0, &ddp_color },
+	[DDP_COMPONENT_COLOR1]	= { MTK_DISP_COLOR,	1, &ddp_color },
 	[DDP_COMPONENT_DITHER]	= { MTK_DISP_DITHER,	0, &ddp_dither },
-	[DDP_COMPONENT_DPI0]	= { MTK_DPI,		0, NULL },
-	[DDP_COMPONENT_DPI1]	= { MTK_DPI,		1, NULL },
-	[DDP_COMPONENT_DSI0]	= { MTK_DSI,		0, NULL },
-	[DDP_COMPONENT_DSI1]	= { MTK_DSI,		1, NULL },
-	[DDP_COMPONENT_DSI2]	= { MTK_DSI,		2, NULL },
-	[DDP_COMPONENT_DSI3]	= { MTK_DSI,		3, NULL },
+	[DDP_COMPONENT_DPI0]	= { MTK_DPI,		0, &ddp_dpi },
+	[DDP_COMPONENT_DPI1]	= { MTK_DPI,		1, &ddp_dpi },
+	[DDP_COMPONENT_DSI0]	= { MTK_DSI,		0, &ddp_dsi },
+	[DDP_COMPONENT_DSI1]	= { MTK_DSI,		1, &ddp_dsi },
+	[DDP_COMPONENT_DSI2]	= { MTK_DSI,		2, &ddp_dsi },
+	[DDP_COMPONENT_DSI3]	= { MTK_DSI,		3, &ddp_dsi },
 	[DDP_COMPONENT_GAMMA]	= { MTK_DISP_GAMMA,	0, &ddp_gamma },
 	[DDP_COMPONENT_OD0]	= { MTK_DISP_OD,	0, &ddp_od },
 	[DDP_COMPONENT_OD1]	= { MTK_DISP_OD,	1, &ddp_od },
-	[DDP_COMPONENT_OVL0]	= { MTK_DISP_OVL,	0, NULL },
-	[DDP_COMPONENT_OVL1]	= { MTK_DISP_OVL,	1, NULL },
-	[DDP_COMPONENT_OVL_2L0]	= { MTK_DISP_OVL_2L,	0, NULL },
-	[DDP_COMPONENT_OVL_2L1]	= { MTK_DISP_OVL_2L,	1, NULL },
+	[DDP_COMPONENT_OVL0]	= { MTK_DISP_OVL,	0, &ddp_ovl },
+	[DDP_COMPONENT_OVL1]	= { MTK_DISP_OVL,	1, &ddp_ovl },
+	[DDP_COMPONENT_OVL_2L0]	= { MTK_DISP_OVL_2L,	0, &ddp_ovl },
+	[DDP_COMPONENT_OVL_2L1]	= { MTK_DISP_OVL_2L,	1, &ddp_ovl },
 	[DDP_COMPONENT_PWM0]	= { MTK_DISP_PWM,	0, NULL },
 	[DDP_COMPONENT_PWM1]	= { MTK_DISP_PWM,	1, NULL },
 	[DDP_COMPONENT_PWM2]	= { MTK_DISP_PWM,	2, NULL },
-	[DDP_COMPONENT_RDMA0]	= { MTK_DISP_RDMA,	0, NULL },
-	[DDP_COMPONENT_RDMA1]	= { MTK_DISP_RDMA,	1, NULL },
-	[DDP_COMPONENT_RDMA2]	= { MTK_DISP_RDMA,	2, NULL },
+	[DDP_COMPONENT_RDMA0]	= { MTK_DISP_RDMA,	0, &ddp_rdma },
+	[DDP_COMPONENT_RDMA1]	= { MTK_DISP_RDMA,	1, &ddp_rdma },
+	[DDP_COMPONENT_RDMA2]	= { MTK_DISP_RDMA,	2, &ddp_rdma },
 	[DDP_COMPONENT_UFOE]	= { MTK_DISP_UFOE,	0, &ddp_ufoe },
 	[DDP_COMPONENT_WDMA0]	= { MTK_DISP_WDMA,	0, NULL },
 	[DDP_COMPONENT_WDMA1]	= { MTK_DISP_WDMA,	1, NULL },
 };
 
-static bool mtk_drm_find_comp_in_ddp(struct mtk_ddp_comp ddp_comp,
+static bool mtk_drm_find_comp_in_ddp(struct device *dev,
 				     const enum mtk_ddp_comp_id *path,
-				     unsigned int path_len)
+				     unsigned int path_len,
+				     struct mtk_ddp_comp *ddp_comp)
 {
 	unsigned int i;
 
@@ -424,7 +407,7 @@ static bool mtk_drm_find_comp_in_ddp(struct mtk_ddp_comp ddp_comp,
 		return false;
 
 	for (i = 0U; i < path_len; i++)
-		if (ddp_comp.id == path[i])
+		if (dev == ddp_comp[path[i]].dev)
 			return true;
 
 	return false;
@@ -446,18 +429,19 @@ int mtk_ddp_comp_get_id(struct device_node *node,
 }
 
 unsigned int mtk_drm_find_possible_crtc_by_comp(struct drm_device *drm,
-						struct mtk_ddp_comp ddp_comp)
+						struct device *dev)
 {
 	struct mtk_drm_private *private = drm->dev_private;
 	unsigned int ret = 0;
 
-	if (mtk_drm_find_comp_in_ddp(ddp_comp, private->data->main_path, private->data->main_len))
+	if (mtk_drm_find_comp_in_ddp(dev, private->data->main_path, private->data->main_len,
+				     private->ddp_comp))
 		ret = BIT(0);
-	else if (mtk_drm_find_comp_in_ddp(ddp_comp, private->data->ext_path,
-					  private->data->ext_len))
+	else if (mtk_drm_find_comp_in_ddp(dev, private->data->ext_path,
+					  private->data->ext_len, private->ddp_comp))
 		ret = BIT(1);
-	else if (mtk_drm_find_comp_in_ddp(ddp_comp, private->data->third_path,
-					  private->data->third_len))
+	else if (mtk_drm_find_comp_in_ddp(dev, private->data->third_path,
+					  private->data->third_len, private->ddp_comp))
 		ret = BIT(2);
 	else
 		DRM_INFO("Failed to find comp in ddp table\n");
@@ -465,59 +449,15 @@ unsigned int mtk_drm_find_possible_crtc_by_comp(struct drm_device *drm,
 	return ret;
 }
 
-int mtk_ddp_comp_init(struct device *dev, struct device_node *node,
-		      struct mtk_ddp_comp *comp, enum mtk_ddp_comp_id comp_id,
-		      const struct mtk_ddp_comp_funcs *funcs)
+static int mtk_ddp_get_larb_dev(struct device_node *node, struct mtk_ddp_comp *comp,
+				struct device *dev)
 {
-	enum mtk_ddp_comp_type type;
 	struct device_node *larb_node;
 	struct platform_device *larb_pdev;
-#if IS_REACHABLE(CONFIG_MTK_CMDQ)
-	struct resource res;
-	struct cmdq_client_reg cmdq_reg;
-	int ret;
-#endif
-
-	if (comp_id < 0 || comp_id >= DDP_COMPONENT_ID_MAX)
-		return -EINVAL;
-
-	type = mtk_ddp_matches[comp_id].type;
-
-	comp->id = comp_id;
-	comp->funcs = funcs ?: mtk_ddp_matches[comp_id].funcs;
-
-	if (comp_id == DDP_COMPONENT_BLS ||
-	    comp_id == DDP_COMPONENT_DPI0 ||
-	    comp_id == DDP_COMPONENT_DPI1 ||
-	    comp_id == DDP_COMPONENT_DSI0 ||
-	    comp_id == DDP_COMPONENT_DSI1 ||
-	    comp_id == DDP_COMPONENT_DSI2 ||
-	    comp_id == DDP_COMPONENT_DSI3 ||
-	    comp_id == DDP_COMPONENT_PWM0) {
-		comp->regs = NULL;
-		comp->clk = NULL;
-		comp->irq = 0;
-		return 0;
-	}
-
-	comp->regs = of_iomap(node, 0);
-	comp->irq = of_irq_get(node, 0);
-	comp->clk = of_clk_get(node, 0);
-	if (IS_ERR(comp->clk))
-		return PTR_ERR(comp->clk);
-
-	/* Only DMA capable components need the LARB property */
-	comp->larb_dev = NULL;
-	if (type != MTK_DISP_OVL &&
-	    type != MTK_DISP_OVL_2L &&
-	    type != MTK_DISP_RDMA &&
-	    type != MTK_DISP_WDMA)
-		return 0;
 
 	larb_node = of_parse_phandle(node, "mediatek,larb", 0);
 	if (!larb_node) {
-		dev_err(dev,
-			"Missing mediadek,larb phandle in %pOF node\n", node);
+		dev_err(dev, "Missing mediadek,larb phandle in %pOF node\n", node);
 		return -EINVAL;
 	}
 
@@ -528,40 +468,71 @@ int mtk_ddp_comp_init(struct device *dev, struct device_node *node,
 		return -EPROBE_DEFER;
 	}
 	of_node_put(larb_node);
-
 	comp->larb_dev = &larb_pdev->dev;
 
-#if IS_REACHABLE(CONFIG_MTK_CMDQ)
-	if (of_address_to_resource(node, 0, &res) != 0) {
-		dev_err(dev, "Missing reg in %s node\n", node->full_name);
-		put_device(&larb_pdev->dev);
-		return -EINVAL;
-	}
-	comp->regs_pa = res.start;
-
-	ret = cmdq_dev_get_client_reg(dev, &cmdq_reg, 0);
-	if (ret)
-		dev_dbg(dev, "get mediatek,gce-client-reg fail!\n");
-	else
-		comp->subsys = cmdq_reg.subsys;
-#endif
 	return 0;
 }
 
-int mtk_ddp_comp_register(struct drm_device *drm, struct mtk_ddp_comp *comp)
+int mtk_ddp_comp_init(struct device_node *node, struct mtk_ddp_comp *comp,
+		      enum mtk_ddp_comp_id comp_id)
 {
-	struct mtk_drm_private *private = drm->dev_private;
+	struct platform_device *comp_pdev;
+	enum mtk_ddp_comp_type type;
+	struct mtk_ddp_comp_dev *priv;
+	int ret;
 
-	if (private->ddp_comp[comp->id])
-		return -EBUSY;
+	if (comp_id < 0 || comp_id >= DDP_COMPONENT_ID_MAX)
+		return -EINVAL;
 
-	private->ddp_comp[comp->id] = comp;
-	return 0;
-}
+	type = mtk_ddp_matches[comp_id].type;
 
-void mtk_ddp_comp_unregister(struct drm_device *drm, struct mtk_ddp_comp *comp)
-{
-	struct mtk_drm_private *private = drm->dev_private;
+	comp->id = comp_id;
+	comp->funcs = mtk_ddp_matches[comp_id].funcs;
+	comp_pdev = of_find_device_by_node(node);
+	if (!comp_pdev) {
+		DRM_INFO("Waiting for device %s\n", node->full_name);
+		return -EPROBE_DEFER;
+	}
+	comp->dev = &comp_pdev->dev;
 
-	private->ddp_comp[comp->id] = NULL;
+	/* Only DMA capable components need the LARB property */
+	if (type == MTK_DISP_OVL ||
+	    type == MTK_DISP_OVL_2L ||
+	    type == MTK_DISP_RDMA ||
+	    type == MTK_DISP_WDMA) {
+		ret = mtk_ddp_get_larb_dev(node, comp, comp->dev);
+		if (ret)
+			return ret;
+	}
+
+	if (type == MTK_DISP_BLS ||
+	    type == MTK_DISP_CCORR ||
+	    type == MTK_DISP_COLOR ||
+	    type == MTK_DISP_GAMMA ||
+	    type == MTK_DPI ||
+	    type == MTK_DSI ||
+	    type == MTK_DISP_OVL ||
+	    type == MTK_DISP_OVL_2L ||
+	    type == MTK_DISP_PWM ||
+	    type == MTK_DISP_RDMA)
+		return 0;
+
+	priv = devm_kzalloc(comp->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->regs = of_iomap(node, 0);
+	priv->clk = of_clk_get(node, 0);
+	if (IS_ERR(priv->clk))
+		return PTR_ERR(priv->clk);
+
+#if IS_REACHABLE(CONFIG_MTK_CMDQ)
+	ret = cmdq_dev_get_client_reg(comp->dev, &priv->cmdq_reg, 0);
+	if (ret)
+		dev_dbg(comp->dev, "get mediatek,gce-client-reg fail!\n");
+#endif
+
+	platform_set_drvdata(comp_pdev, priv);
+
+	return 0;
 }
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.h b/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.h
index 5aa52b7afeec..bb914d976cf5 100644
--- a/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.h
+++ b/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.h
@@ -7,6 +7,7 @@
 #define MTK_DRM_DDP_COMP_H
 
 #include <linux/io.h>
+#include <linux/soc/mediatek/mtk-cmdq.h>
 #include <linux/soc/mediatek/mtk-mmsys.h>
 
 struct device;
@@ -39,79 +40,95 @@ enum mtk_ddp_comp_type {
 struct mtk_ddp_comp;
 struct cmdq_pkt;
 struct mtk_ddp_comp_funcs {
-	void (*config)(struct mtk_ddp_comp *comp, unsigned int w,
+	int (*clk_enable)(struct device *dev);
+	void (*clk_disable)(struct device *dev);
+	void (*config)(struct device *dev, unsigned int w,
 		       unsigned int h, unsigned int vrefresh,
 		       unsigned int bpc, struct cmdq_pkt *cmdq_pkt);
-	void (*start)(struct mtk_ddp_comp *comp);
-	void (*stop)(struct mtk_ddp_comp *comp);
-	void (*enable_vblank)(struct mtk_ddp_comp *comp, struct drm_crtc *crtc);
-	void (*disable_vblank)(struct mtk_ddp_comp *comp);
-	unsigned int (*supported_rotations)(struct mtk_ddp_comp *comp);
-	unsigned int (*layer_nr)(struct mtk_ddp_comp *comp);
-	int (*layer_check)(struct mtk_ddp_comp *comp,
+	void (*start)(struct device *dev);
+	void (*stop)(struct device *dev);
+	void (*enable_vblank)(struct device *dev,
+			      void (*vblank_cb)(void *),
+			      void *vblank_cb_data);
+	void (*disable_vblank)(struct device *dev);
+	unsigned int (*supported_rotations)(struct device *dev);
+	unsigned int (*layer_nr)(struct device *dev);
+	int (*layer_check)(struct device *dev,
 			   unsigned int idx,
 			   struct mtk_plane_state *state);
-	void (*layer_config)(struct mtk_ddp_comp *comp, unsigned int idx,
+	void (*layer_config)(struct device *dev, unsigned int idx,
 			     struct mtk_plane_state *state,
 			     struct cmdq_pkt *cmdq_pkt);
-	void (*gamma_set)(struct mtk_ddp_comp *comp,
+	void (*gamma_set)(struct device *dev,
 			  struct drm_crtc_state *state);
-	void (*bgclr_in_on)(struct mtk_ddp_comp *comp);
-	void (*bgclr_in_off)(struct mtk_ddp_comp *comp);
-	void (*ctm_set)(struct mtk_ddp_comp *comp,
+	void (*bgclr_in_on)(struct device *dev);
+	void (*bgclr_in_off)(struct device *dev);
+	void (*ctm_set)(struct device *dev,
 			struct drm_crtc_state *state);
 };
 
 struct mtk_ddp_comp {
-	struct clk *clk;
-	void __iomem *regs;
+	struct device *dev;
 	int irq;
 	struct device *larb_dev;
 	enum mtk_ddp_comp_id id;
 	const struct mtk_ddp_comp_funcs *funcs;
-	resource_size_t regs_pa;
-	u8 subsys;
 };
 
+static inline int mtk_ddp_comp_clk_enable(struct mtk_ddp_comp *comp)
+{
+	if (comp->funcs && comp->funcs->clk_enable)
+		return comp->funcs->clk_enable(comp->dev);
+
+	return 0;
+}
+
+static inline void mtk_ddp_comp_clk_disable(struct mtk_ddp_comp *comp)
+{
+	if (comp->funcs && comp->funcs->clk_disable)
+		comp->funcs->clk_disable(comp->dev);
+}
+
 static inline void mtk_ddp_comp_config(struct mtk_ddp_comp *comp,
 				       unsigned int w, unsigned int h,
 				       unsigned int vrefresh, unsigned int bpc,
 				       struct cmdq_pkt *cmdq_pkt)
 {
 	if (comp->funcs && comp->funcs->config)
-		comp->funcs->config(comp, w, h, vrefresh, bpc, cmdq_pkt);
+		comp->funcs->config(comp->dev, w, h, vrefresh, bpc, cmdq_pkt);
 }
 
 static inline void mtk_ddp_comp_start(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->start)
-		comp->funcs->start(comp);
+		comp->funcs->start(comp->dev);
 }
 
 static inline void mtk_ddp_comp_stop(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->stop)
-		comp->funcs->stop(comp);
+		comp->funcs->stop(comp->dev);
 }
 
 static inline void mtk_ddp_comp_enable_vblank(struct mtk_ddp_comp *comp,
-					      struct drm_crtc *crtc)
+					      void (*vblank_cb)(void *),
+					      void *vblank_cb_data)
 {
 	if (comp->funcs && comp->funcs->enable_vblank)
-		comp->funcs->enable_vblank(comp, crtc);
+		comp->funcs->enable_vblank(comp->dev, vblank_cb, vblank_cb_data);
 }
 
 static inline void mtk_ddp_comp_disable_vblank(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->disable_vblank)
-		comp->funcs->disable_vblank(comp);
+		comp->funcs->disable_vblank(comp->dev);
 }
 
 static inline
 unsigned int mtk_ddp_comp_supported_rotations(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->supported_rotations)
-		return comp->funcs->supported_rotations(comp);
+		return comp->funcs->supported_rotations(comp->dev);
 
 	return 0;
 }
@@ -119,7 +136,7 @@ unsigned int mtk_ddp_comp_supported_rotations(struct mtk_ddp_comp *comp)
 static inline unsigned int mtk_ddp_comp_layer_nr(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->layer_nr)
-		return comp->funcs->layer_nr(comp);
+		return comp->funcs->layer_nr(comp->dev);
 
 	return 0;
 }
@@ -129,7 +146,7 @@ static inline int mtk_ddp_comp_layer_check(struct mtk_ddp_comp *comp,
 					   struct mtk_plane_state *state)
 {
 	if (comp->funcs && comp->funcs->layer_check)
-		return comp->funcs->layer_check(comp, idx, state);
+		return comp->funcs->layer_check(comp->dev, idx, state);
 	return 0;
 }
 
@@ -139,52 +156,49 @@ static inline void mtk_ddp_comp_layer_config(struct mtk_ddp_comp *comp,
 					     struct cmdq_pkt *cmdq_pkt)
 {
 	if (comp->funcs && comp->funcs->layer_config)
-		comp->funcs->layer_config(comp, idx, state, cmdq_pkt);
+		comp->funcs->layer_config(comp->dev, idx, state, cmdq_pkt);
 }
 
 static inline void mtk_ddp_gamma_set(struct mtk_ddp_comp *comp,
 				     struct drm_crtc_state *state)
 {
 	if (comp->funcs && comp->funcs->gamma_set)
-		comp->funcs->gamma_set(comp, state);
+		comp->funcs->gamma_set(comp->dev, state);
 }
 
 static inline void mtk_ddp_comp_bgclr_in_on(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->bgclr_in_on)
-		comp->funcs->bgclr_in_on(comp);
+		comp->funcs->bgclr_in_on(comp->dev);
 }
 
 static inline void mtk_ddp_comp_bgclr_in_off(struct mtk_ddp_comp *comp)
 {
 	if (comp->funcs && comp->funcs->bgclr_in_off)
-		comp->funcs->bgclr_in_off(comp);
+		comp->funcs->bgclr_in_off(comp->dev);
 }
 
 static inline void mtk_ddp_ctm_set(struct mtk_ddp_comp *comp,
 				   struct drm_crtc_state *state)
 {
 	if (comp->funcs && comp->funcs->ctm_set)
-		comp->funcs->ctm_set(comp, state);
+		comp->funcs->ctm_set(comp->dev, state);
 }
 
 int mtk_ddp_comp_get_id(struct device_node *node,
 			enum mtk_ddp_comp_type comp_type);
 unsigned int mtk_drm_find_possible_crtc_by_comp(struct drm_device *drm,
-						struct mtk_ddp_comp ddp_comp);
-int mtk_ddp_comp_init(struct device *dev, struct device_node *comp_node,
-		      struct mtk_ddp_comp *comp, enum mtk_ddp_comp_id comp_id,
-		      const struct mtk_ddp_comp_funcs *funcs);
-int mtk_ddp_comp_register(struct drm_device *drm, struct mtk_ddp_comp *comp);
-void mtk_ddp_comp_unregister(struct drm_device *drm, struct mtk_ddp_comp *comp);
-void mtk_dither_set(struct mtk_ddp_comp *comp, unsigned int bpc,
-		    unsigned int CFG, struct cmdq_pkt *cmdq_pkt);
+						struct device *dev);
+int mtk_ddp_comp_init(struct device_node *comp_node, struct mtk_ddp_comp *comp,
+		      enum mtk_ddp_comp_id comp_id);
 enum mtk_ddp_comp_type mtk_ddp_comp_get_type(enum mtk_ddp_comp_id comp_id);
 void mtk_ddp_write(struct cmdq_pkt *cmdq_pkt, unsigned int value,
-		   struct mtk_ddp_comp *comp, unsigned int offset);
+		   struct cmdq_client_reg *cmdq_reg, void __iomem *regs,
+		   unsigned int offset);
 void mtk_ddp_write_relaxed(struct cmdq_pkt *cmdq_pkt, unsigned int value,
-			   struct mtk_ddp_comp *comp, unsigned int offset);
+			   struct cmdq_client_reg *cmdq_reg, void __iomem *regs,
+			   unsigned int offset);
 void mtk_ddp_write_mask(struct cmdq_pkt *cmdq_pkt, unsigned int value,
-			struct mtk_ddp_comp *comp, unsigned int offset,
-			unsigned int mask);
+			struct cmdq_client_reg *cmdq_reg, void __iomem *regs,
+			unsigned int offset, unsigned int mask);
 #endif /* MTK_DRM_DDP_COMP_H */
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c
index 2f717df28a77..b013d56d2777 100644
--- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c
+++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c
@@ -10,7 +10,6 @@
 #include <linux/of_address.h>
 #include <linux/of_platform.h>
 #include <linux/pm_runtime.h>
-#include <linux/soc/mediatek/mtk-mmsys.h>
 #include <linux/dma-mapping.h>
 
 #include <drm/drm_atomic.h>
@@ -26,7 +25,6 @@
 #include <drm/drm_vblank.h>
 
 #include "mtk_drm_crtc.h"
-#include "mtk_drm_ddp.h"
 #include "mtk_drm_ddp_comp.h"
 #include "mtk_drm_drv.h"
 #include "mtk_drm_gem.h"
@@ -131,6 +129,24 @@ static const enum mtk_ddp_comp_id mt8173_mtk_ddp_ext[] = {
 	DDP_COMPONENT_DPI0,
 };
 
+static const enum mtk_ddp_comp_id mt8183_mtk_ddp_main[] = {
+	DDP_COMPONENT_OVL0,
+	DDP_COMPONENT_OVL_2L0,
+	DDP_COMPONENT_RDMA0,
+	DDP_COMPONENT_COLOR0,
+	DDP_COMPONENT_CCORR,
+	DDP_COMPONENT_AAL0,
+	DDP_COMPONENT_GAMMA,
+	DDP_COMPONENT_DITHER,
+	DDP_COMPONENT_DSI0,
+};
+
+static const enum mtk_ddp_comp_id mt8183_mtk_ddp_ext[] = {
+	DDP_COMPONENT_OVL_2L1,
+	DDP_COMPONENT_RDMA1,
+	DDP_COMPONENT_DPI0,
+};
+
 static const struct mtk_mmsys_driver_data mt2701_mmsys_driver_data = {
 	.main_path = mt2701_mtk_ddp_main,
 	.main_len = ARRAY_SIZE(mt2701_mtk_ddp_main),
@@ -163,6 +179,13 @@ static const struct mtk_mmsys_driver_data mt8173_mmsys_driver_data = {
 	.ext_len = ARRAY_SIZE(mt8173_mtk_ddp_ext),
 };
 
+static const struct mtk_mmsys_driver_data mt8183_mmsys_driver_data = {
+	.main_path = mt8183_mtk_ddp_main,
+	.main_len = ARRAY_SIZE(mt8183_mtk_ddp_main),
+	.ext_path = mt8183_mtk_ddp_ext,
+	.ext_len = ARRAY_SIZE(mt8183_mtk_ddp_ext),
+};
+
 static int mtk_drm_kms_init(struct drm_device *drm)
 {
 	struct mtk_drm_private *private = drm->dev_private;
@@ -377,12 +400,20 @@ static const struct of_device_id mtk_ddp_comp_dt_ids[] = {
 	  .data = (void *)MTK_DISP_OVL },
 	{ .compatible = "mediatek,mt8173-disp-ovl",
 	  .data = (void *)MTK_DISP_OVL },
+	{ .compatible = "mediatek,mt8183-disp-ovl",
+	  .data = (void *)MTK_DISP_OVL },
+	{ .compatible = "mediatek,mt8183-disp-ovl-2l",
+	  .data = (void *)MTK_DISP_OVL_2L },
 	{ .compatible = "mediatek,mt2701-disp-rdma",
 	  .data = (void *)MTK_DISP_RDMA },
 	{ .compatible = "mediatek,mt8173-disp-rdma",
 	  .data = (void *)MTK_DISP_RDMA },
+	{ .compatible = "mediatek,mt8183-disp-rdma",
+	  .data = (void *)MTK_DISP_RDMA },
 	{ .compatible = "mediatek,mt8173-disp-wdma",
 	  .data = (void *)MTK_DISP_WDMA },
+	{ .compatible = "mediatek,mt8183-disp-ccorr",
+	  .data = (void *)MTK_DISP_CCORR },
 	{ .compatible = "mediatek,mt2701-disp-color",
 	  .data = (void *)MTK_DISP_COLOR },
 	{ .compatible = "mediatek,mt8173-disp-color",
@@ -391,22 +422,32 @@ static const struct of_device_id mtk_ddp_comp_dt_ids[] = {
 	  .data = (void *)MTK_DISP_AAL},
 	{ .compatible = "mediatek,mt8173-disp-gamma",
 	  .data = (void *)MTK_DISP_GAMMA, },
+	{ .compatible = "mediatek,mt8183-disp-gamma",
+	  .data = (void *)MTK_DISP_GAMMA, },
+	{ .compatible = "mediatek,mt8183-disp-dither",
+	  .data = (void *)MTK_DISP_DITHER },
 	{ .compatible = "mediatek,mt8173-disp-ufoe",
 	  .data = (void *)MTK_DISP_UFOE },
 	{ .compatible = "mediatek,mt2701-dsi",
 	  .data = (void *)MTK_DSI },
 	{ .compatible = "mediatek,mt8173-dsi",
 	  .data = (void *)MTK_DSI },
+	{ .compatible = "mediatek,mt8183-dsi",
+	  .data = (void *)MTK_DSI },
 	{ .compatible = "mediatek,mt2701-dpi",
 	  .data = (void *)MTK_DPI },
 	{ .compatible = "mediatek,mt8173-dpi",
 	  .data = (void *)MTK_DPI },
+	{ .compatible = "mediatek,mt8183-dpi",
+	  .data = (void *)MTK_DPI },
 	{ .compatible = "mediatek,mt2701-disp-mutex",
 	  .data = (void *)MTK_DISP_MUTEX },
 	{ .compatible = "mediatek,mt2712-disp-mutex",
 	  .data = (void *)MTK_DISP_MUTEX },
 	{ .compatible = "mediatek,mt8173-disp-mutex",
 	  .data = (void *)MTK_DISP_MUTEX },
+	{ .compatible = "mediatek,mt8183-disp-mutex",
+	  .data = (void *)MTK_DISP_MUTEX },
 	{ .compatible = "mediatek,mt2701-disp-pwm",
 	  .data = (void *)MTK_DISP_BLS },
 	{ .compatible = "mediatek,mt8173-disp-pwm",
@@ -425,6 +466,8 @@ static const struct of_device_id mtk_drm_of_ids[] = {
 	  .data = &mt2712_mmsys_driver_data},
 	{ .compatible = "mediatek,mt8173-mmsys",
 	  .data = &mt8173_mmsys_driver_data},
+	{ .compatible = "mediatek,mt8183-mmsys",
+	  .data = &mt8183_mmsys_driver_data},
 	{ }
 };
 
@@ -488,11 +531,13 @@ static int mtk_drm_probe(struct platform_device *pdev)
 		private->comp_node[comp_id] = of_node_get(node);
 
 		/*
-		 * Currently only the COLOR, OVL, RDMA, DSI, and DPI blocks have
-		 * separate component platform drivers and initialize their own
+		 * Currently only the CCORR, COLOR, GAMMA, OVL, RDMA, DSI, and DPI
+		 * blocks have separate component platform drivers and initialize their own
 		 * DDP component structure. The others are initialized here.
 		 */
-		if (comp_type == MTK_DISP_COLOR ||
+		if (comp_type == MTK_DISP_CCORR ||
+		    comp_type == MTK_DISP_COLOR ||
+		    comp_type == MTK_DISP_GAMMA ||
 		    comp_type == MTK_DISP_OVL ||
 		    comp_type == MTK_DISP_OVL_2L ||
 		    comp_type == MTK_DISP_RDMA ||
@@ -502,24 +547,12 @@ static int mtk_drm_probe(struct platform_device *pdev)
 				 node);
 			drm_of_component_match_add(dev, &match, compare_of,
 						   node);
-		} else {
-			struct mtk_ddp_comp *comp;
-
-			comp = devm_kzalloc(dev, sizeof(*comp), GFP_KERNEL);
-			if (!comp) {
-				ret = -ENOMEM;
-				of_node_put(node);
-				goto err_node;
-			}
-
-			ret = mtk_ddp_comp_init(dev->parent, node, comp,
-						comp_id, NULL);
-			if (ret) {
-				of_node_put(node);
-				goto err_node;
-			}
-
-			private->ddp_comp[comp_id] = comp;
+		}
+
+		ret = mtk_ddp_comp_init(node, &private->ddp_comp[comp_id], comp_id);
+		if (ret) {
+			of_node_put(node);
+			goto err_node;
 		}
 	}
 
@@ -545,10 +578,8 @@ err_node:
 	of_node_put(private->mutex_node);
 	for (i = 0; i < DDP_COMPONENT_ID_MAX; i++) {
 		of_node_put(private->comp_node[i]);
-		if (private->ddp_comp[i]) {
-			put_device(private->ddp_comp[i]->larb_dev);
-			private->ddp_comp[i] = NULL;
-		}
+		if (private->ddp_comp[i].larb_dev)
+			put_device(private->ddp_comp[i].larb_dev);
 	}
 	return ret;
 }
@@ -604,8 +635,9 @@ static struct platform_driver mtk_drm_platform_driver = {
 };
 
 static struct platform_driver * const mtk_drm_drivers[] = {
-	&mtk_ddp_driver,
+	&mtk_disp_ccorr_driver,
 	&mtk_disp_color_driver,
+	&mtk_disp_gamma_driver,
 	&mtk_disp_ovl_driver,
 	&mtk_disp_rdma_driver,
 	&mtk_dpi_driver,
diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.h b/drivers/gpu/drm/mediatek/mtk_drm_drv.h
index 5d771cf0bf25..637f5669e895 100644
--- a/drivers/gpu/drm/mediatek/mtk_drm_drv.h
+++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.h
@@ -41,13 +41,14 @@ struct mtk_drm_private {
 	struct device *mutex_dev;
 	struct device *mmsys_dev;
 	struct device_node *comp_node[DDP_COMPONENT_ID_MAX];
-	struct mtk_ddp_comp *ddp_comp[DDP_COMPONENT_ID_MAX];
+	struct mtk_ddp_comp ddp_comp[DDP_COMPONENT_ID_MAX];
 	const struct mtk_mmsys_driver_data *data;
 	struct drm_atomic_state *suspend_state;
 };
 
-extern struct platform_driver mtk_ddp_driver;
+extern struct platform_driver mtk_disp_ccorr_driver;
 extern struct platform_driver mtk_disp_color_driver;
+extern struct platform_driver mtk_disp_gamma_driver;
 extern struct platform_driver mtk_disp_ovl_driver;
 extern struct platform_driver mtk_disp_rdma_driver;
 extern struct platform_driver mtk_dpi_driver;
diff --git a/drivers/gpu/drm/mediatek/mtk_dsi.c b/drivers/gpu/drm/mediatek/mtk_dsi.c
index 65fd99c528af..a1ff152ef468 100644
--- a/drivers/gpu/drm/mediatek/mtk_dsi.c
+++ b/drivers/gpu/drm/mediatek/mtk_dsi.c
@@ -25,6 +25,7 @@
 #include <drm/drm_probe_helper.h>
 #include <drm/drm_simple_kms_helper.h>
 
+#include "mtk_disp_drv.h"
 #include "mtk_drm_ddp_comp.h"
 
 #define DSI_START		0x00
@@ -178,7 +179,6 @@ struct mtk_dsi_driver_data {
 };
 
 struct mtk_dsi {
-	struct mtk_ddp_comp ddp_comp;
 	struct device *dev;
 	struct mipi_dsi_host host;
 	struct drm_encoder encoder;
@@ -767,25 +767,20 @@ static const struct drm_bridge_funcs mtk_dsi_bridge_funcs = {
 	.mode_set = mtk_dsi_bridge_mode_set,
 };
 
-static void mtk_dsi_ddp_start(struct mtk_ddp_comp *comp)
+void mtk_dsi_ddp_start(struct device *dev)
 {
-	struct mtk_dsi *dsi = container_of(comp, struct mtk_dsi, ddp_comp);
+	struct mtk_dsi *dsi = dev_get_drvdata(dev);
 
 	mtk_dsi_poweron(dsi);
 }
 
-static void mtk_dsi_ddp_stop(struct mtk_ddp_comp *comp)
+void mtk_dsi_ddp_stop(struct device *dev)
 {
-	struct mtk_dsi *dsi = container_of(comp, struct mtk_dsi, ddp_comp);
+	struct mtk_dsi *dsi = dev_get_drvdata(dev);
 
 	mtk_dsi_poweroff(dsi);
 }
 
-static const struct mtk_ddp_comp_funcs mtk_dsi_funcs = {
-	.start = mtk_dsi_ddp_start,
-	.stop = mtk_dsi_ddp_stop,
-};
-
 static int mtk_dsi_host_attach(struct mipi_dsi_host *host,
 			       struct mipi_dsi_device *device)
 {
@@ -952,7 +947,7 @@ static int mtk_dsi_encoder_init(struct drm_device *drm, struct mtk_dsi *dsi)
 		return ret;
 	}
 
-	dsi->encoder.possible_crtcs = mtk_drm_find_possible_crtc_by_comp(drm, dsi->ddp_comp);
+	dsi->encoder.possible_crtcs = mtk_drm_find_possible_crtc_by_comp(drm, dsi->host.dev);
 
 	ret = drm_bridge_attach(&dsi->encoder, &dsi->bridge, NULL,
 				DRM_BRIDGE_ATTACH_NO_CONNECTOR);
@@ -980,32 +975,17 @@ static int mtk_dsi_bind(struct device *dev, struct device *master, void *data)
 	struct drm_device *drm = data;
 	struct mtk_dsi *dsi = dev_get_drvdata(dev);
 
-	ret = mtk_ddp_comp_register(drm, &dsi->ddp_comp);
-	if (ret < 0) {
-		dev_err(dev, "Failed to register component %pOF: %d\n",
-			dev->of_node, ret);
-		return ret;
-	}
-
 	ret = mtk_dsi_encoder_init(drm, dsi);
-	if (ret)
-		goto err_unregister;
 
-	return 0;
-
-err_unregister:
-	mtk_ddp_comp_unregister(drm, &dsi->ddp_comp);
 	return ret;
 }
 
 static void mtk_dsi_unbind(struct device *dev, struct device *master,
 			   void *data)
 {
-	struct drm_device *drm = data;
 	struct mtk_dsi *dsi = dev_get_drvdata(dev);
 
 	drm_encoder_cleanup(&dsi->encoder);
-	mtk_ddp_comp_unregister(drm, &dsi->ddp_comp);
 }
 
 static const struct component_ops mtk_dsi_component_ops = {
@@ -1020,7 +1000,6 @@ static int mtk_dsi_probe(struct platform_device *pdev)
 	struct drm_panel *panel;
 	struct resource *regs;
 	int irq_num;
-	int comp_id;
 	int ret;
 
 	dsi = devm_kzalloc(dev, sizeof(*dsi), GFP_KERNEL);
@@ -1090,20 +1069,6 @@ static int mtk_dsi_probe(struct platform_device *pdev)
 		goto err_unregister_host;
 	}
 
-	comp_id = mtk_ddp_comp_get_id(dev->of_node, MTK_DSI);
-	if (comp_id < 0) {
-		dev_err(dev, "Failed to identify by alias: %d\n", comp_id);
-		ret = comp_id;
-		goto err_unregister_host;
-	}
-
-	ret = mtk_ddp_comp_init(dev, dev->of_node, &dsi->ddp_comp, comp_id,
-				&mtk_dsi_funcs);
-	if (ret) {
-		dev_err(dev, "Failed to initialize component: %d\n", ret);
-		goto err_unregister_host;
-	}
-
 	irq_num = platform_get_irq(pdev, 0);
 	if (irq_num < 0) {
 		dev_err(&pdev->dev, "failed to get dsi irq_num: %d\n", irq_num);
@@ -1111,9 +1076,8 @@ static int mtk_dsi_probe(struct platform_device *pdev)
 		goto err_unregister_host;
 	}
 
-	irq_set_status_flags(irq_num, IRQ_TYPE_LEVEL_LOW);
 	ret = devm_request_irq(&pdev->dev, irq_num, mtk_dsi_irq,
-			       IRQF_TRIGGER_LOW, dev_name(&pdev->dev), dsi);
+			       IRQF_TRIGGER_NONE, dev_name(&pdev->dev), dsi);
 	if (ret) {
 		dev_err(&pdev->dev, "failed to request mediatek dsi irq\n");
 		goto err_unregister_host;
diff --git a/drivers/gpu/drm/meson/meson_drv.c b/drivers/gpu/drm/meson/meson_drv.c
index 42c5d3246cfc..453d8b4c5763 100644
--- a/drivers/gpu/drm/meson/meson_drv.c
+++ b/drivers/gpu/drm/meson/meson_drv.c
@@ -482,6 +482,16 @@ static int meson_probe_remote(struct platform_device *pdev,
 	return count;
 }
 
+static void meson_drv_shutdown(struct platform_device *pdev)
+{
+	struct meson_drm *priv = dev_get_drvdata(&pdev->dev);
+	struct drm_device *drm = priv->drm;
+
+	DRM_DEBUG_DRIVER("\n");
+	drm_kms_helper_poll_fini(drm);
+	drm_atomic_helper_shutdown(drm);
+}
+
 static int meson_drv_probe(struct platform_device *pdev)
 {
 	struct component_match *match = NULL;
@@ -553,6 +563,7 @@ static const struct dev_pm_ops meson_drv_pm_ops = {
 
 static struct platform_driver meson_drm_platform_driver = {
 	.probe      = meson_drv_probe,
+	.shutdown   = meson_drv_shutdown,
 	.driver     = {
 		.name	= "meson-drm",
 		.of_match_table = dt_match,
diff --git a/drivers/gpu/drm/msm/adreno/a2xx_gpu.c b/drivers/gpu/drm/msm/adreno/a2xx_gpu.c
index 7e82c41a85f1..bdc989183c64 100644
--- a/drivers/gpu/drm/msm/adreno/a2xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a2xx_gpu.c
@@ -534,8 +534,10 @@ struct msm_gpu *a2xx_gpu_init(struct drm_device *dev)
 
 	if (!gpu->aspace) {
 		dev_err(dev->dev, "No memory protection without MMU\n");
-		ret = -ENXIO;
-		goto fail;
+		if (!allow_vram_carveout) {
+			ret = -ENXIO;
+			goto fail;
+		}
 	}
 
 	return gpu;
diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
index 93da6683a866..4534633fe7cd 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
@@ -564,8 +564,10 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 		 * implement a cmdstream validator.
 		 */
 		DRM_DEV_ERROR(dev->dev, "No memory protection without IOMMU\n");
-		ret = -ENXIO;
-		goto fail;
+		if (!allow_vram_carveout) {
+			ret = -ENXIO;
+			goto fail;
+		}
 	}
 
 	icc_path = devm_of_icc_get(&pdev->dev, "gfx-mem");
diff --git a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
index c0be3a0f36b2..82bebb40234d 100644
--- a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
@@ -692,8 +692,10 @@ struct msm_gpu *a4xx_gpu_init(struct drm_device *dev)
 		 * implement a cmdstream validator.
 		 */
 		DRM_DEV_ERROR(dev->dev, "No memory protection without IOMMU\n");
-		ret = -ENXIO;
-		goto fail;
+		if (!allow_vram_carveout) {
+			ret = -ENXIO;
+			goto fail;
+		}
 	}
 
 	icc_path = devm_of_icc_get(&pdev->dev, "gfx-mem");
diff --git a/drivers/gpu/drm/msm/adreno/a5xx.xml.h b/drivers/gpu/drm/msm/adreno/a5xx.xml.h
index 346cc6ff3a36..7b9fcfe95c04 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a5xx.xml.h
@@ -2367,6 +2367,8 @@ static inline uint32_t A5XX_VSC_RESOLVE_CNTL_Y(uint32_t val)
 
 #define REG_A5XX_UCHE_ADDR_MODE_CNTL				0x00000e80
 
+#define REG_A5XX_UCHE_MODE_CNTL					0x00000e81
+
 #define REG_A5XX_UCHE_SVM_CNTL					0x00000e82
 
 #define REG_A5XX_UCHE_WRITE_THRU_BASE_LO			0x00000e87
diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
index a5af223eaf50..7e553d3efeb2 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
@@ -222,7 +222,7 @@ static void a5xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit)
 	a5xx_preempt_trigger(gpu);
 }
 
-static const struct {
+static const struct adreno_five_hwcg_regs {
 	u32 offset;
 	u32 value;
 } a5xx_hwcg[] = {
@@ -318,16 +318,124 @@ static const struct {
 	{REG_A5XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM, 0x00004000},
 	{REG_A5XX_RBBM_CLOCK_DELAY_GPC, 0x00000200},
 	{REG_A5XX_RBBM_CLOCK_DELAY_VFD, 0x00002222}
+}, a50x_hwcg[] = {
+	{REG_A5XX_RBBM_CLOCK_CNTL_SP0, 0x02222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_SP0, 0x02222220},
+	{REG_A5XX_RBBM_CLOCK_HYST_SP0, 0x0000F3CF},
+	{REG_A5XX_RBBM_CLOCK_DELAY_SP0, 0x00000080},
+	{REG_A5XX_RBBM_CLOCK_CNTL_TP0, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_TP0, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL3_TP0, 0x00002222},
+	{REG_A5XX_RBBM_CLOCK_HYST_TP0, 0x77777777},
+	{REG_A5XX_RBBM_CLOCK_HYST2_TP0, 0x77777777},
+	{REG_A5XX_RBBM_CLOCK_HYST3_TP0, 0x00007777},
+	{REG_A5XX_RBBM_CLOCK_DELAY_TP0, 0x11111111},
+	{REG_A5XX_RBBM_CLOCK_DELAY2_TP0, 0x11111111},
+	{REG_A5XX_RBBM_CLOCK_DELAY3_TP0, 0x00001111},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_UCHE, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL3_UCHE, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL4_UCHE, 0x00222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL_UCHE, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_HYST_UCHE, 0x00FFFFF4},
+	{REG_A5XX_RBBM_CLOCK_DELAY_UCHE, 0x00000002},
+	{REG_A5XX_RBBM_CLOCK_CNTL_RB0, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_RB0, 0x00222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL_CCU0, 0x00022220},
+	{REG_A5XX_RBBM_CLOCK_CNTL_RAC, 0x05522222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_RAC, 0x00505555},
+	{REG_A5XX_RBBM_CLOCK_HYST_RB_CCU0, 0x04040404},
+	{REG_A5XX_RBBM_CLOCK_HYST_RAC, 0x07444044},
+	{REG_A5XX_RBBM_CLOCK_DELAY_RB_CCU_L1_0, 0x00000002},
+	{REG_A5XX_RBBM_CLOCK_DELAY_RAC, 0x00010011},
+	{REG_A5XX_RBBM_CLOCK_CNTL_TSE_RAS_RBBM, 0x04222222},
+	{REG_A5XX_RBBM_CLOCK_MODE_GPC, 0x02222222},
+	{REG_A5XX_RBBM_CLOCK_MODE_VFD, 0x00002222},
+	{REG_A5XX_RBBM_CLOCK_HYST_TSE_RAS_RBBM, 0x00000000},
+	{REG_A5XX_RBBM_CLOCK_HYST_GPC, 0x04104004},
+	{REG_A5XX_RBBM_CLOCK_HYST_VFD, 0x00000000},
+	{REG_A5XX_RBBM_CLOCK_DELAY_HLSQ, 0x00000000},
+	{REG_A5XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM, 0x00004000},
+	{REG_A5XX_RBBM_CLOCK_DELAY_GPC, 0x00000200},
+	{REG_A5XX_RBBM_CLOCK_DELAY_VFD, 0x00002222},
+}, a512_hwcg[] = {
+	{REG_A5XX_RBBM_CLOCK_CNTL_SP0, 0x02222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL_SP1, 0x02222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_SP0, 0x02222220},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_SP1, 0x02222220},
+	{REG_A5XX_RBBM_CLOCK_HYST_SP0, 0x0000F3CF},
+	{REG_A5XX_RBBM_CLOCK_HYST_SP1, 0x0000F3CF},
+	{REG_A5XX_RBBM_CLOCK_DELAY_SP0, 0x00000080},
+	{REG_A5XX_RBBM_CLOCK_DELAY_SP1, 0x00000080},
+	{REG_A5XX_RBBM_CLOCK_CNTL_TP0, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL_TP1, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_TP0, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_TP1, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL3_TP0, 0x00002222},
+	{REG_A5XX_RBBM_CLOCK_CNTL3_TP1, 0x00002222},
+	{REG_A5XX_RBBM_CLOCK_HYST_TP0, 0x77777777},
+	{REG_A5XX_RBBM_CLOCK_HYST_TP1, 0x77777777},
+	{REG_A5XX_RBBM_CLOCK_HYST2_TP0, 0x77777777},
+	{REG_A5XX_RBBM_CLOCK_HYST2_TP1, 0x77777777},
+	{REG_A5XX_RBBM_CLOCK_HYST3_TP0, 0x00007777},
+	{REG_A5XX_RBBM_CLOCK_HYST3_TP1, 0x00007777},
+	{REG_A5XX_RBBM_CLOCK_DELAY_TP0, 0x11111111},
+	{REG_A5XX_RBBM_CLOCK_DELAY_TP1, 0x11111111},
+	{REG_A5XX_RBBM_CLOCK_DELAY2_TP0, 0x11111111},
+	{REG_A5XX_RBBM_CLOCK_DELAY2_TP1, 0x11111111},
+	{REG_A5XX_RBBM_CLOCK_DELAY3_TP0, 0x00001111},
+	{REG_A5XX_RBBM_CLOCK_DELAY3_TP1, 0x00001111},
+	{REG_A5XX_RBBM_CLOCK_CNTL_UCHE, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_UCHE, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL3_UCHE, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL4_UCHE, 0x00222222},
+	{REG_A5XX_RBBM_CLOCK_HYST_UCHE, 0x00444444},
+	{REG_A5XX_RBBM_CLOCK_DELAY_UCHE, 0x00000002},
+	{REG_A5XX_RBBM_CLOCK_CNTL_RB0, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL_RB1, 0x22222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_RB0, 0x00222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_RB1, 0x00222222},
+	{REG_A5XX_RBBM_CLOCK_CNTL_CCU0, 0x00022220},
+	{REG_A5XX_RBBM_CLOCK_CNTL_CCU1, 0x00022220},
+	{REG_A5XX_RBBM_CLOCK_CNTL_RAC, 0x05522222},
+	{REG_A5XX_RBBM_CLOCK_CNTL2_RAC, 0x00505555},
+	{REG_A5XX_RBBM_CLOCK_HYST_RB_CCU0, 0x04040404},
+	{REG_A5XX_RBBM_CLOCK_HYST_RB_CCU1, 0x04040404},
+	{REG_A5XX_RBBM_CLOCK_HYST_RAC, 0x07444044},
+	{REG_A5XX_RBBM_CLOCK_DELAY_RB_CCU_L1_0, 0x00000002},
+	{REG_A5XX_RBBM_CLOCK_DELAY_RB_CCU_L1_1, 0x00000002},
+	{REG_A5XX_RBBM_CLOCK_DELAY_RAC, 0x00010011},
+	{REG_A5XX_RBBM_CLOCK_CNTL_TSE_RAS_RBBM, 0x04222222},
+	{REG_A5XX_RBBM_CLOCK_MODE_GPC, 0x02222222},
+	{REG_A5XX_RBBM_CLOCK_MODE_VFD, 0x00002222},
+	{REG_A5XX_RBBM_CLOCK_HYST_TSE_RAS_RBBM, 0x00000000},
+	{REG_A5XX_RBBM_CLOCK_HYST_GPC, 0x04104004},
+	{REG_A5XX_RBBM_CLOCK_HYST_VFD, 0x00000000},
+	{REG_A5XX_RBBM_CLOCK_DELAY_HLSQ, 0x00000000},
+	{REG_A5XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM, 0x00004000},
+	{REG_A5XX_RBBM_CLOCK_DELAY_GPC, 0x00000200},
+	{REG_A5XX_RBBM_CLOCK_DELAY_VFD, 0x00002222},
 };
 
 void a5xx_set_hwcg(struct msm_gpu *gpu, bool state)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
-	unsigned int i;
+	const struct adreno_five_hwcg_regs *regs;
+	unsigned int i, sz;
+
+	if (adreno_is_a508(adreno_gpu)) {
+		regs = a50x_hwcg;
+		sz = ARRAY_SIZE(a50x_hwcg);
+	} else if (adreno_is_a509(adreno_gpu) || adreno_is_a512(adreno_gpu)) {
+		regs = a512_hwcg;
+		sz = ARRAY_SIZE(a512_hwcg);
+	} else {
+		regs = a5xx_hwcg;
+		sz = ARRAY_SIZE(a5xx_hwcg);
+	}
 
-	for (i = 0; i < ARRAY_SIZE(a5xx_hwcg); i++)
-		gpu_write(gpu, a5xx_hwcg[i].offset,
-			state ? a5xx_hwcg[i].value : 0);
+	for (i = 0; i < sz; i++)
+		gpu_write(gpu, regs[i].offset,
+			  state ? regs[i].value : 0);
 
 	if (adreno_is_a540(adreno_gpu)) {
 		gpu_write(gpu, REG_A5XX_RBBM_CLOCK_DELAY_GPMU, state ? 0x00000770 : 0);
@@ -538,11 +646,13 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
 	struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu);
+	u32 regbit;
 	int ret;
 
 	gpu_write(gpu, REG_A5XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x00000003);
 
-	if (adreno_is_a540(adreno_gpu))
+	if (adreno_is_a509(adreno_gpu) || adreno_is_a512(adreno_gpu) ||
+	    adreno_is_a540(adreno_gpu))
 		gpu_write(gpu, REG_A5XX_VBIF_GATE_OFF_WRREQ_EN, 0x00000009);
 
 	/* Make all blocks contribute to the GPU BUSY perf counter */
@@ -604,29 +714,48 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 		0x00100000 + adreno_gpu->gmem - 1);
 	gpu_write(gpu, REG_A5XX_UCHE_GMEM_RANGE_MAX_HI, 0x00000000);
 
-	if (adreno_is_a510(adreno_gpu)) {
+	if (adreno_is_a508(adreno_gpu) || adreno_is_a510(adreno_gpu)) {
 		gpu_write(gpu, REG_A5XX_CP_MEQ_THRESHOLDS, 0x20);
-		gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x20);
+		if (adreno_is_a508(adreno_gpu))
+			gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x400);
+		else
+			gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x20);
 		gpu_write(gpu, REG_A5XX_CP_ROQ_THRESHOLDS_2, 0x40000030);
 		gpu_write(gpu, REG_A5XX_CP_ROQ_THRESHOLDS_1, 0x20100D0A);
-		gpu_write(gpu, REG_A5XX_PC_DBG_ECO_CNTL,
-			  (0x200 << 11 | 0x200 << 22));
 	} else {
 		gpu_write(gpu, REG_A5XX_CP_MEQ_THRESHOLDS, 0x40);
 		if (adreno_is_a530(adreno_gpu))
 			gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x40);
-		if (adreno_is_a540(adreno_gpu))
+		else
 			gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x400);
 		gpu_write(gpu, REG_A5XX_CP_ROQ_THRESHOLDS_2, 0x80000060);
 		gpu_write(gpu, REG_A5XX_CP_ROQ_THRESHOLDS_1, 0x40201B16);
+	}
+
+	if (adreno_is_a508(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_PC_DBG_ECO_CNTL,
+			  (0x100 << 11 | 0x100 << 22));
+	else if (adreno_is_a509(adreno_gpu) || adreno_is_a510(adreno_gpu) ||
+		 adreno_is_a512(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_PC_DBG_ECO_CNTL,
+			  (0x200 << 11 | 0x200 << 22));
+	else
 		gpu_write(gpu, REG_A5XX_PC_DBG_ECO_CNTL,
 			  (0x400 << 11 | 0x300 << 22));
-	}
 
 	if (adreno_gpu->info->quirks & ADRENO_QUIRK_TWO_PASS_USE_WFI)
 		gpu_rmw(gpu, REG_A5XX_PC_DBG_ECO_CNTL, 0, (1 << 8));
 
-	gpu_write(gpu, REG_A5XX_PC_DBG_ECO_CNTL, 0xc0200100);
+	/*
+	 * Disable the RB sampler datapath DP2 clock gating optimization
+	 * for 1-SP GPUs, as it is enabled by default.
+	 */
+	if (adreno_is_a508(adreno_gpu) || adreno_is_a509(adreno_gpu) ||
+	    adreno_is_a512(adreno_gpu))
+		gpu_rmw(gpu, REG_A5XX_RB_DBG_ECO_CNTL, 0, (1 << 9));
+
+	/* Disable UCHE global filter as SP can invalidate/flush independently */
+	gpu_write(gpu, REG_A5XX_UCHE_MODE_CNTL, BIT(29));
 
 	/* Enable USE_RETENTION_FLOPS */
 	gpu_write(gpu, REG_A5XX_CP_CHICKEN_DBG, 0x02000000);
@@ -653,10 +782,20 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 	gpu_write(gpu, REG_A5XX_RBBM_AHB_CNTL2, 0x0000003F);
 
 	/* Set the highest bank bit */
-	gpu_write(gpu, REG_A5XX_TPL1_MODE_CNTL, 2 << 7);
-	gpu_write(gpu, REG_A5XX_RB_MODE_CNTL, 2 << 1);
 	if (adreno_is_a540(adreno_gpu))
-		gpu_write(gpu, REG_A5XX_UCHE_DBG_ECO_CNTL_2, 2);
+		regbit = 2;
+	else
+		regbit = 1;
+
+	gpu_write(gpu, REG_A5XX_TPL1_MODE_CNTL, regbit << 7);
+	gpu_write(gpu, REG_A5XX_RB_MODE_CNTL, regbit << 1);
+
+	if (adreno_is_a509(adreno_gpu) || adreno_is_a512(adreno_gpu) ||
+	    adreno_is_a540(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_UCHE_DBG_ECO_CNTL_2, regbit);
+
+	/* Disable All flat shading optimization (ALLFLATOPTDIS) */
+	gpu_rmw(gpu, REG_A5XX_VPC_DBG_ECO_CNTL, 0, (1 << 10));
 
 	/* Protect registers from the CP */
 	gpu_write(gpu, REG_A5XX_CP_PROTECT_CNTL, 0x00000007);
@@ -688,12 +827,14 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 
 	/* VPC */
 	gpu_write(gpu, REG_A5XX_CP_PROTECT(14), ADRENO_PROTECT_RW(0xE68, 8));
-	gpu_write(gpu, REG_A5XX_CP_PROTECT(15), ADRENO_PROTECT_RW(0xE70, 4));
+	gpu_write(gpu, REG_A5XX_CP_PROTECT(15), ADRENO_PROTECT_RW(0xE70, 16));
 
 	/* UCHE */
 	gpu_write(gpu, REG_A5XX_CP_PROTECT(16), ADRENO_PROTECT_RW(0xE80, 16));
 
-	if (adreno_is_a530(adreno_gpu) || adreno_is_a510(adreno_gpu))
+	if (adreno_is_a508(adreno_gpu) || adreno_is_a509(adreno_gpu) ||
+	    adreno_is_a510(adreno_gpu) || adreno_is_a512(adreno_gpu) ||
+	    adreno_is_a530(adreno_gpu))
 		gpu_write(gpu, REG_A5XX_CP_PROTECT(17),
 			ADRENO_PROTECT_RW(0x10000, 0x8000));
 
@@ -735,7 +876,8 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 	if (ret)
 		return ret;
 
-	if (!adreno_is_a510(adreno_gpu))
+	if (!(adreno_is_a508(adreno_gpu) || adreno_is_a509(adreno_gpu) ||
+	      adreno_is_a510(adreno_gpu) || adreno_is_a512(adreno_gpu)))
 		a5xx_gpmu_ucode_init(gpu);
 
 	ret = a5xx_ucode_init(gpu);
@@ -1168,7 +1310,8 @@ static int a5xx_pm_resume(struct msm_gpu *gpu)
 	if (ret)
 		return ret;
 
-	if (adreno_is_a510(adreno_gpu)) {
+	/* Adreno 508, 509, 510, 512 needs manual RBBM sus/res control */
+	if (!(adreno_is_a530(adreno_gpu) || adreno_is_a540(adreno_gpu))) {
 		/* Halt the sp_input_clk at HM level */
 		gpu_write(gpu, REG_A5XX_RBBM_CLOCK_CNTL, 0x00000055);
 		a5xx_set_hwcg(gpu, true);
@@ -1210,8 +1353,8 @@ static int a5xx_pm_suspend(struct msm_gpu *gpu)
 	u32 mask = 0xf;
 	int i, ret;
 
-	/* A510 has 3 XIN ports in VBIF */
-	if (adreno_is_a510(adreno_gpu))
+	/* A508, A510 have 3 XIN ports in VBIF */
+	if (adreno_is_a508(adreno_gpu) || adreno_is_a510(adreno_gpu))
 		mask = 0x7;
 
 	/* Clear the VBIF pipe before shutting down */
@@ -1223,10 +1366,12 @@ static int a5xx_pm_suspend(struct msm_gpu *gpu)
 
 	/*
 	 * Reset the VBIF before power collapse to avoid issue with FIFO
-	 * entries
+	 * entries on Adreno A510 and A530 (the others will tend to lock up)
 	 */
-	gpu_write(gpu, REG_A5XX_RBBM_BLOCK_SW_RESET_CMD, 0x003C0000);
-	gpu_write(gpu, REG_A5XX_RBBM_BLOCK_SW_RESET_CMD, 0x00000000);
+	if (adreno_is_a510(adreno_gpu) || adreno_is_a530(adreno_gpu)) {
+		gpu_write(gpu, REG_A5XX_RBBM_BLOCK_SW_RESET_CMD, 0x003C0000);
+		gpu_write(gpu, REG_A5XX_RBBM_BLOCK_SW_RESET_CMD, 0x00000000);
+	}
 
 	ret = msm_gpu_pm_suspend(gpu);
 	if (ret)
diff --git a/drivers/gpu/drm/msm/adreno/a5xx_power.c b/drivers/gpu/drm/msm/adreno/a5xx_power.c
index f176a6f3eff6..5ccc9da455a1 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_power.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_power.c
@@ -298,7 +298,7 @@ int a5xx_power_init(struct msm_gpu *gpu)
 	int ret;
 
 	/* Not all A5xx chips have a GPMU */
-	if (adreno_is_a510(adreno_gpu))
+	if (!(adreno_is_a530(adreno_gpu) || adreno_is_a540(adreno_gpu)))
 		return 0;
 
 	/* Set up the limits management */
@@ -330,7 +330,7 @@ void a5xx_gpmu_ucode_init(struct msm_gpu *gpu)
 	unsigned int *data, *ptr, *cmds;
 	unsigned int cmds_size;
 
-	if (adreno_is_a510(adreno_gpu))
+	if (!(adreno_is_a530(adreno_gpu) || adreno_is_a540(adreno_gpu)))
 		return;
 
 	if (a5xx_gpu->gpmu_bo)
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
index e6703ae98760..71c917f909af 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
@@ -134,7 +134,7 @@ void a6xx_gmu_set_freq(struct msm_gpu *gpu, struct dev_pm_opp *opp)
 
 	if (!gmu->legacy) {
 		a6xx_hfi_set_freq(gmu, perf_index);
-		dev_pm_opp_set_bw(&gpu->pdev->dev, opp);
+		dev_pm_opp_set_opp(&gpu->pdev->dev, opp);
 		pm_runtime_put(gmu->dev);
 		return;
 	}
@@ -158,7 +158,7 @@ void a6xx_gmu_set_freq(struct msm_gpu *gpu, struct dev_pm_opp *opp)
 	if (ret)
 		dev_err(gmu->dev, "GMU set GPU frequency error: %d\n", ret);
 
-	dev_pm_opp_set_bw(&gpu->pdev->dev, opp);
+	dev_pm_opp_set_opp(&gpu->pdev->dev, opp);
 	pm_runtime_put(gmu->dev);
 }
 
@@ -245,37 +245,66 @@ static int a6xx_gmu_hfi_start(struct a6xx_gmu *gmu)
 	return ret;
 }
 
+struct a6xx_gmu_oob_bits {
+	int set, ack, set_new, ack_new;
+	const char *name;
+};
+
+/* These are the interrupt / ack bits for each OOB request that are set
+ * in a6xx_gmu_set_oob and a6xx_clear_oob
+ */
+static const struct a6xx_gmu_oob_bits a6xx_gmu_oob_bits[] = {
+	[GMU_OOB_GPU_SET] = {
+		.name = "GPU_SET",
+		.set = 16,
+		.ack = 24,
+		.set_new = 30,
+		.ack_new = 31,
+	},
+
+	[GMU_OOB_PERFCOUNTER_SET] = {
+		.name = "PERFCOUNTER",
+		.set = 17,
+		.ack = 25,
+		.set_new = 28,
+		.ack_new = 30,
+	},
+
+	[GMU_OOB_BOOT_SLUMBER] = {
+		.name = "BOOT_SLUMBER",
+		.set = 22,
+		.ack = 30,
+	},
+
+	[GMU_OOB_DCVS_SET] = {
+		.name = "GPU_DCVS",
+		.set = 23,
+		.ack = 31,
+	},
+};
+
 /* Trigger a OOB (out of band) request to the GMU */
 int a6xx_gmu_set_oob(struct a6xx_gmu *gmu, enum a6xx_gmu_oob_state state)
 {
 	int ret;
 	u32 val;
 	int request, ack;
-	const char *name;
 
-	switch (state) {
-	case GMU_OOB_GPU_SET:
-		if (gmu->legacy) {
-			request = GMU_OOB_GPU_SET_REQUEST;
-			ack = GMU_OOB_GPU_SET_ACK;
-		} else {
-			request = GMU_OOB_GPU_SET_REQUEST_NEW;
-			ack = GMU_OOB_GPU_SET_ACK_NEW;
-		}
-		name = "GPU_SET";
-		break;
-	case GMU_OOB_BOOT_SLUMBER:
-		request = GMU_OOB_BOOT_SLUMBER_REQUEST;
-		ack = GMU_OOB_BOOT_SLUMBER_ACK;
-		name = "BOOT_SLUMBER";
-		break;
-	case GMU_OOB_DCVS_SET:
-		request = GMU_OOB_DCVS_REQUEST;
-		ack = GMU_OOB_DCVS_ACK;
-		name = "GPU_DCVS";
-		break;
-	default:
+	if (state >= ARRAY_SIZE(a6xx_gmu_oob_bits))
 		return -EINVAL;
+
+	if (gmu->legacy) {
+		request = a6xx_gmu_oob_bits[state].set;
+		ack = a6xx_gmu_oob_bits[state].ack;
+	} else {
+		request = a6xx_gmu_oob_bits[state].set_new;
+		ack = a6xx_gmu_oob_bits[state].ack_new;
+		if (!request || !ack) {
+			DRM_DEV_ERROR(gmu->dev,
+				      "Invalid non-legacy GMU request %s\n",
+				      a6xx_gmu_oob_bits[state].name);
+			return -EINVAL;
+		}
 	}
 
 	/* Trigger the equested OOB operation */
@@ -288,7 +317,7 @@ int a6xx_gmu_set_oob(struct a6xx_gmu *gmu, enum a6xx_gmu_oob_state state)
 	if (ret)
 		DRM_DEV_ERROR(gmu->dev,
 			"Timeout waiting for GMU OOB set %s: 0x%x\n",
-				name,
+				a6xx_gmu_oob_bits[state].name,
 				gmu_read(gmu, REG_A6XX_GMU_GMU2HOST_INTR_INFO));
 
 	/* Clear the acknowledge interrupt */
@@ -300,27 +329,17 @@ int a6xx_gmu_set_oob(struct a6xx_gmu *gmu, enum a6xx_gmu_oob_state state)
 /* Clear a pending OOB state in the GMU */
 void a6xx_gmu_clear_oob(struct a6xx_gmu *gmu, enum a6xx_gmu_oob_state state)
 {
-	if (!gmu->legacy) {
-		WARN_ON(state != GMU_OOB_GPU_SET);
-		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
-			1 << GMU_OOB_GPU_SET_CLEAR_NEW);
+	int bit;
+
+	if (state >= ARRAY_SIZE(a6xx_gmu_oob_bits))
 		return;
-	}
 
-	switch (state) {
-	case GMU_OOB_GPU_SET:
-		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
-			1 << GMU_OOB_GPU_SET_CLEAR);
-		break;
-	case GMU_OOB_BOOT_SLUMBER:
-		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
-			1 << GMU_OOB_BOOT_SLUMBER_CLEAR);
-		break;
-	case GMU_OOB_DCVS_SET:
-		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
-			1 << GMU_OOB_DCVS_CLEAR);
-		break;
-	}
+	if (gmu->legacy)
+		bit = a6xx_gmu_oob_bits[state].ack;
+	else
+		bit = a6xx_gmu_oob_bits[state].ack_new;
+
+	gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET, bit);
 }
 
 /* Enable CPU control of SPTP power power collapse */
@@ -866,7 +885,7 @@ static void a6xx_gmu_set_initial_bw(struct msm_gpu *gpu, struct a6xx_gmu *gmu)
 	if (IS_ERR_OR_NULL(gpu_opp))
 		return;
 
-	dev_pm_opp_set_bw(&gpu->pdev->dev, gpu_opp);
+	dev_pm_opp_set_opp(&gpu->pdev->dev, gpu_opp);
 	dev_pm_opp_put(gpu_opp);
 }
 
@@ -1072,7 +1091,7 @@ int a6xx_gmu_stop(struct a6xx_gpu *a6xx_gpu)
 		a6xx_gmu_shutdown(gmu);
 
 	/* Remove the bus vote */
-	dev_pm_opp_set_bw(&gpu->pdev->dev, NULL);
+	dev_pm_opp_set_opp(&gpu->pdev->dev, NULL);
 
 	/*
 	 * Make sure the GX domain is off before turning off the GMU (CX)
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h
index c6d2bced8e5d..71dfa60070cc 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h
@@ -153,44 +153,27 @@ static inline void gmu_write_rscc(struct a6xx_gmu *gmu, u32 offset, u32 value)
  */
 
 enum a6xx_gmu_oob_state {
+	/*
+	 * Let the GMU know that a boot or slumber operation has started. The value in
+	 * REG_A6XX_GMU_BOOT_SLUMBER_OPTION lets the GMU know which operation we are
+	 * doing
+	 */
 	GMU_OOB_BOOT_SLUMBER = 0,
+	/*
+	 * Let the GMU know to not turn off any GPU registers while the CPU is in a
+	 * critical section
+	 */
 	GMU_OOB_GPU_SET,
+	/*
+	 * Set a new power level for the GPU when the CPU is doing frequency scaling
+	 */
 	GMU_OOB_DCVS_SET,
+	/*
+	 * Used to keep the GPU on for CPU-side reads of performance counters.
+	 */
+	GMU_OOB_PERFCOUNTER_SET,
 };
 
-/* These are the interrupt / ack bits for each OOB request that are set
- * in a6xx_gmu_set_oob and a6xx_clear_oob
- */
-
-/*
- * Let the GMU know that a boot or slumber operation has started. The value in
- * REG_A6XX_GMU_BOOT_SLUMBER_OPTION lets the GMU know which operation we are
- * doing
- */
-#define GMU_OOB_BOOT_SLUMBER_REQUEST	22
-#define GMU_OOB_BOOT_SLUMBER_ACK	30
-#define GMU_OOB_BOOT_SLUMBER_CLEAR	30
-
-/*
- * Set a new power level for the GPU when the CPU is doing frequency scaling
- */
-#define GMU_OOB_DCVS_REQUEST	23
-#define GMU_OOB_DCVS_ACK	31
-#define GMU_OOB_DCVS_CLEAR	31
-
-/*
- * Let the GMU know to not turn off any GPU registers while the CPU is in a
- * critical section
- */
-#define GMU_OOB_GPU_SET_REQUEST	16
-#define GMU_OOB_GPU_SET_ACK	24
-#define GMU_OOB_GPU_SET_CLEAR	24
-
-#define GMU_OOB_GPU_SET_REQUEST_NEW	30
-#define GMU_OOB_GPU_SET_ACK_NEW		31
-#define GMU_OOB_GPU_SET_CLEAR_NEW	31
-
-
 void a6xx_hfi_init(struct a6xx_gmu *gmu);
 int a6xx_hfi_start(struct a6xx_gmu *gmu, int boot_state);
 void a6xx_hfi_stop(struct a6xx_gmu *gmu);
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
index 130661898546..ba8e9d3cf0fe 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
@@ -10,6 +10,7 @@
 
 #include <linux/bitfield.h>
 #include <linux/devfreq.h>
+#include <linux/nvmem-consumer.h>
 #include <linux/soc/qcom/llcc-qcom.h>
 
 #define GPU_PAS_ID 13
@@ -1117,7 +1118,7 @@ static void a6xx_llc_slices_init(struct platform_device *pdev,
 	a6xx_gpu->llc_slice = llcc_slice_getd(LLCC_GPU);
 	a6xx_gpu->htw_llc_slice = llcc_slice_getd(LLCC_GPUHTW);
 
-	if (IS_ERR(a6xx_gpu->llc_slice) && IS_ERR(a6xx_gpu->htw_llc_slice))
+	if (IS_ERR_OR_NULL(a6xx_gpu->llc_slice) && IS_ERR_OR_NULL(a6xx_gpu->htw_llc_slice))
 		a6xx_gpu->llc_mmio = ERR_PTR(-EINVAL);
 }
 
@@ -1169,14 +1170,18 @@ static int a6xx_get_timestamp(struct msm_gpu *gpu, uint64_t *value)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
 	struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
+	static DEFINE_MUTEX(perfcounter_oob);
+
+	mutex_lock(&perfcounter_oob);
 
 	/* Force the GPU power on so we can read this register */
-	a6xx_gmu_set_oob(&a6xx_gpu->gmu, GMU_OOB_GPU_SET);
+	a6xx_gmu_set_oob(&a6xx_gpu->gmu, GMU_OOB_PERFCOUNTER_SET);
 
 	*value = gpu_read64(gpu, REG_A6XX_RBBM_PERFCTR_CP_0_LO,
 		REG_A6XX_RBBM_PERFCTR_CP_0_HI);
 
-	a6xx_gmu_clear_oob(&a6xx_gpu->gmu, GMU_OOB_GPU_SET);
+	a6xx_gmu_clear_oob(&a6xx_gpu->gmu, GMU_OOB_PERFCOUNTER_SET);
+	mutex_unlock(&perfcounter_oob);
 	return 0;
 }
 
@@ -1208,6 +1213,10 @@ static void a6xx_destroy(struct msm_gpu *gpu)
 	a6xx_gmu_remove(a6xx_gpu);
 
 	adreno_gpu_cleanup(adreno_gpu);
+
+	if (a6xx_gpu->opp_table)
+		dev_pm_opp_put_supported_hw(a6xx_gpu->opp_table);
+
 	kfree(a6xx_gpu);
 }
 
@@ -1240,6 +1249,50 @@ static unsigned long a6xx_gpu_busy(struct msm_gpu *gpu)
 }
 
 static struct msm_gem_address_space *
+a6xx_create_address_space(struct msm_gpu *gpu, struct platform_device *pdev)
+{
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
+	struct iommu_domain *iommu;
+	struct msm_mmu *mmu;
+	struct msm_gem_address_space *aspace;
+	u64 start, size;
+
+	iommu = iommu_domain_alloc(&platform_bus_type);
+	if (!iommu)
+		return NULL;
+
+	/*
+	 * This allows GPU to set the bus attributes required to use system
+	 * cache on behalf of the iommu page table walker.
+	 */
+	if (!IS_ERR_OR_NULL(a6xx_gpu->htw_llc_slice))
+		adreno_set_llc_attributes(iommu);
+
+	mmu = msm_iommu_new(&pdev->dev, iommu);
+	if (IS_ERR(mmu)) {
+		iommu_domain_free(iommu);
+		return ERR_CAST(mmu);
+	}
+
+	/*
+	 * Use the aperture start or SZ_16M, whichever is greater. This will
+	 * ensure that we align with the allocated pagetable range while still
+	 * allowing room in the lower 32 bits for GMEM and whatnot
+	 */
+	start = max_t(u64, SZ_16M, iommu->geometry.aperture_start);
+	size = iommu->geometry.aperture_end - start + 1;
+
+	aspace = msm_gem_address_space_create(mmu, "gpu",
+		start & GENMASK_ULL(48, 0), size);
+
+	if (IS_ERR(aspace) && !IS_ERR(mmu))
+		mmu->funcs->destroy(mmu);
+
+	return aspace;
+}
+
+static struct msm_gem_address_space *
 a6xx_create_private_address_space(struct msm_gpu *gpu)
 {
 	struct msm_mmu *mmu;
@@ -1264,6 +1317,78 @@ static uint32_t a6xx_get_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
 	return ring->memptrs->rptr = gpu_read(gpu, REG_A6XX_CP_RB_RPTR);
 }
 
+static u32 a618_get_speed_bin(u32 fuse)
+{
+	if (fuse == 0)
+		return 0;
+	else if (fuse == 169)
+		return 1;
+	else if (fuse == 174)
+		return 2;
+
+	return UINT_MAX;
+}
+
+static u32 fuse_to_supp_hw(struct device *dev, u32 revn, u32 fuse)
+{
+	u32 val = UINT_MAX;
+
+	if (revn == 618)
+		val = a618_get_speed_bin(fuse);
+
+	if (val == UINT_MAX) {
+		DRM_DEV_ERROR(dev,
+			"missing support for speed-bin: %u. Some OPPs may not be supported by hardware",
+			fuse);
+		return UINT_MAX;
+	}
+
+	return (1 << val);
+}
+
+static int a6xx_set_supported_hw(struct device *dev, struct a6xx_gpu *a6xx_gpu,
+		u32 revn)
+{
+	struct opp_table *opp_table;
+	struct nvmem_cell *cell;
+	u32 supp_hw = UINT_MAX;
+	void *buf;
+
+	cell = nvmem_cell_get(dev, "speed_bin");
+	/*
+	 * -ENOENT means that the platform doesn't support speedbin which is
+	 * fine
+	 */
+	if (PTR_ERR(cell) == -ENOENT)
+		return 0;
+	else if (IS_ERR(cell)) {
+		DRM_DEV_ERROR(dev,
+				"failed to read speed-bin. Some OPPs may not be supported by hardware");
+		goto done;
+	}
+
+	buf = nvmem_cell_read(cell, NULL);
+	if (IS_ERR(buf)) {
+		nvmem_cell_put(cell);
+		DRM_DEV_ERROR(dev,
+				"failed to read speed-bin. Some OPPs may not be supported by hardware");
+		goto done;
+	}
+
+	supp_hw = fuse_to_supp_hw(dev, revn, *((u32 *) buf));
+
+	kfree(buf);
+	nvmem_cell_put(cell);
+
+done:
+	opp_table = dev_pm_opp_set_supported_hw(dev, &supp_hw, 1);
+	if (IS_ERR(opp_table))
+		return PTR_ERR(opp_table);
+
+	a6xx_gpu->opp_table = opp_table;
+	return 0;
+}
+
 static const struct adreno_gpu_funcs funcs = {
 	.base = {
 		.get_param = adreno_get_param,
@@ -1285,7 +1410,7 @@ static const struct adreno_gpu_funcs funcs = {
 		.gpu_state_get = a6xx_gpu_state_get,
 		.gpu_state_put = a6xx_gpu_state_put,
 #endif
-		.create_address_space = adreno_iommu_create_address_space,
+		.create_address_space = a6xx_create_address_space,
 		.create_private_address_space = a6xx_create_private_address_space,
 		.get_rptr = a6xx_get_rptr,
 	},
@@ -1325,6 +1450,12 @@ struct msm_gpu *a6xx_gpu_init(struct drm_device *dev)
 
 	a6xx_llc_slices_init(pdev, a6xx_gpu);
 
+	ret = a6xx_set_supported_hw(&pdev->dev, a6xx_gpu, info->revn);
+	if (ret) {
+		a6xx_destroy(&(a6xx_gpu->base.base));
+		return ERR_PTR(ret);
+	}
+
 	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, 1);
 	if (ret) {
 		a6xx_destroy(&(a6xx_gpu->base.base));
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h
index e793d329e77b..ce0610c5256f 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h
@@ -33,6 +33,8 @@ struct a6xx_gpu {
 	void *llc_slice;
 	void *htw_llc_slice;
 	bool have_mmu500;
+
+	struct opp_table *opp_table;
 };
 
 #define to_a6xx_gpu(x) container_of(x, struct a6xx_gpu, base)
diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c
index 87c8b033ad1a..600d445fabe8 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_device.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_device.c
@@ -18,6 +18,10 @@ bool snapshot_debugbus = false;
 MODULE_PARM_DESC(snapshot_debugbus, "Include debugbus sections in GPU devcoredump (if not fused off)");
 module_param_named(snapshot_debugbus, snapshot_debugbus, bool, 0600);
 
+bool allow_vram_carveout = false;
+MODULE_PARM_DESC(allow_vram_carveout, "Allow using VRAM Carveout, in place of IOMMU");
+module_param_named(allow_vram_carveout, allow_vram_carveout, bool, 0600);
+
 static const struct adreno_info gpulist[] = {
 	{
 		.rev   = ADRENO_REV(2, 0, 0, 0),
@@ -130,6 +134,41 @@ static const struct adreno_info gpulist[] = {
 		.inactive_period = DRM_MSM_INACTIVE_PERIOD,
 		.init  = a4xx_gpu_init,
 	}, {
+		.rev   = ADRENO_REV(5, 0, 8, ANY_ID),
+		.revn = 508,
+		.name = "A508",
+		.fw = {
+			[ADRENO_FW_PM4] = "a530_pm4.fw",
+			[ADRENO_FW_PFP] = "a530_pfp.fw",
+		},
+		.gmem = (SZ_128K + SZ_8K),
+		/*
+		 * Increase inactive period to 250 to avoid bouncing
+		 * the GDSC which appears to make it grumpy
+		 */
+		.inactive_period = 250,
+		.quirks = ADRENO_QUIRK_LMLOADKILL_DISABLE,
+		.init = a5xx_gpu_init,
+		.zapfw = "a508_zap.mdt",
+	}, {
+		.rev   = ADRENO_REV(5, 0, 9, ANY_ID),
+		.revn = 509,
+		.name = "A509",
+		.fw = {
+			[ADRENO_FW_PM4] = "a530_pm4.fw",
+			[ADRENO_FW_PFP] = "a530_pfp.fw",
+		},
+		.gmem = (SZ_256K + SZ_16K),
+		/*
+		 * Increase inactive period to 250 to avoid bouncing
+		 * the GDSC which appears to make it grumpy
+		 */
+		.inactive_period = 250,
+		.quirks = ADRENO_QUIRK_LMLOADKILL_DISABLE,
+		.init = a5xx_gpu_init,
+		/* Adreno 509 uses the same ZAP as 512 */
+		.zapfw = "a512_zap.mdt",
+	}, {
 		.rev   = ADRENO_REV(5, 1, 0, ANY_ID),
 		.revn = 510,
 		.name = "A510",
@@ -145,6 +184,23 @@ static const struct adreno_info gpulist[] = {
 		.inactive_period = 250,
 		.init = a5xx_gpu_init,
 	}, {
+		.rev   = ADRENO_REV(5, 1, 2, ANY_ID),
+		.revn = 512,
+		.name = "A512",
+		.fw = {
+			[ADRENO_FW_PM4] = "a530_pm4.fw",
+			[ADRENO_FW_PFP] = "a530_pfp.fw",
+		},
+		.gmem = (SZ_256K + SZ_16K),
+		/*
+		 * Increase inactive period to 250 to avoid bouncing
+		 * the GDSC which appears to make it grumpy
+		 */
+		.inactive_period = 250,
+		.quirks = ADRENO_QUIRK_LMLOADKILL_DISABLE,
+		.init = a5xx_gpu_init,
+		.zapfw = "a512_zap.mdt",
+	}, {
 		.rev = ADRENO_REV(5, 3, 0, 2),
 		.revn = 530,
 		.name = "A530",
@@ -164,7 +220,7 @@ static const struct adreno_info gpulist[] = {
 		.init = a5xx_gpu_init,
 		.zapfw = "a530_zap.mdt",
 	}, {
-		.rev = ADRENO_REV(5, 4, 0, 2),
+		.rev = ADRENO_REV(5, 4, 0, ANY_ID),
 		.revn = 540,
 		.name = "A540",
 		.fw = {
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
index 6cf9975e951e..0f184c3dd9d9 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
@@ -186,13 +186,18 @@ int adreno_zap_shader_load(struct msm_gpu *gpu, u32 pasid)
 	return zap_shader_load_mdt(gpu, adreno_gpu->info->zapfw, pasid);
 }
 
+void adreno_set_llc_attributes(struct iommu_domain *iommu)
+{
+	struct io_pgtable_domain_attr pgtbl_cfg;
+
+	pgtbl_cfg.quirks = IO_PGTABLE_QUIRK_ARM_OUTER_WBWA;
+	iommu_domain_set_attr(iommu, DOMAIN_ATTR_IO_PGTABLE_CFG, &pgtbl_cfg);
+}
+
 struct msm_gem_address_space *
 adreno_iommu_create_address_space(struct msm_gpu *gpu,
 		struct platform_device *pdev)
 {
-	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
-	struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
-	struct io_pgtable_domain_attr pgtbl_cfg;
 	struct iommu_domain *iommu;
 	struct msm_mmu *mmu;
 	struct msm_gem_address_space *aspace;
@@ -202,15 +207,6 @@ adreno_iommu_create_address_space(struct msm_gpu *gpu,
 	if (!iommu)
 		return NULL;
 
-	/*
-	 * This allows GPU to set the bus attributes required to use system
-	 * cache on behalf of the iommu page table walker.
-	 */
-	if (!IS_ERR(a6xx_gpu->htw_llc_slice)) {
-		pgtbl_cfg.quirks = IO_PGTABLE_QUIRK_ARM_OUTER_WBWA;
-		iommu_domain_set_attr(iommu, DOMAIN_ATTR_IO_PGTABLE_CFG, &pgtbl_cfg);
-	}
-
 	mmu = msm_iommu_new(&pdev->dev, iommu);
 	if (IS_ERR(mmu)) {
 		iommu_domain_free(iommu);
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
index c3775f79525a..ccac275aa7a2 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
@@ -18,6 +18,7 @@
 #include "adreno_pm4.xml.h"
 
 extern bool snapshot_debugbus;
+extern bool allow_vram_carveout;
 
 enum {
 	ADRENO_FW_PM4 = 0,
@@ -196,11 +197,26 @@ static inline int adreno_is_a430(struct adreno_gpu *gpu)
        return gpu->revn == 430;
 }
 
+static inline int adreno_is_a508(struct adreno_gpu *gpu)
+{
+	return gpu->revn == 508;
+}
+
+static inline int adreno_is_a509(struct adreno_gpu *gpu)
+{
+	return gpu->revn == 509;
+}
+
 static inline int adreno_is_a510(struct adreno_gpu *gpu)
 {
 	return gpu->revn == 510;
 }
 
+static inline int adreno_is_a512(struct adreno_gpu *gpu)
+{
+	return gpu->revn == 512;
+}
+
 static inline int adreno_is_a530(struct adreno_gpu *gpu)
 {
 	return gpu->revn == 530;
@@ -272,6 +288,8 @@ struct msm_gem_address_space *
 adreno_iommu_create_address_space(struct msm_gpu *gpu,
 		struct platform_device *pdev);
 
+void adreno_set_llc_attributes(struct iommu_domain *iommu);
+
 /*
  * For a5xx and a6xx targets load the zap shader that is used to pull the GPU
  * out of secure mode
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
index 5a056c1191df..b2be39b9144e 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
@@ -4,8 +4,10 @@
  */
 
 #define pr_fmt(fmt)	"[drm:%s:%d] " fmt, __func__, __LINE__
+#include <linux/delay.h>
 #include "dpu_encoder_phys.h"
 #include "dpu_hw_interrupts.h"
+#include "dpu_hw_pingpong.h"
 #include "dpu_core_irq.h"
 #include "dpu_formats.h"
 #include "dpu_trace.h"
@@ -35,6 +37,8 @@
 
 #define DPU_ENC_WR_PTR_START_TIMEOUT_US 20000
 
+#define DPU_ENC_MAX_POLL_TIMEOUT_US	2000
+
 static bool dpu_encoder_phys_cmd_is_master(struct dpu_encoder_phys *phys_enc)
 {
 	return (phys_enc->split_role != ENC_ROLE_SLAVE) ? true : false;
@@ -368,15 +372,12 @@ static void dpu_encoder_phys_cmd_tearcheck_config(
 	tc_cfg.vsync_count = vsync_hz /
 				(mode->vtotal * drm_mode_vrefresh(mode));
 
-	/* enable external TE after kickoff to avoid premature autorefresh */
-	tc_cfg.hw_vsync_mode = 0;
-
 	/*
-	 * By setting sync_cfg_height to near max register value, we essentially
-	 * disable dpu hw generated TE signal, since hw TE will arrive first.
-	 * Only caveat is if due to error, we hit wrap-around.
+	 * Set the sync_cfg_height to twice vtotal so that if we lose a
+	 * TE event coming from the display TE pin we won't stall immediately
 	 */
-	tc_cfg.sync_cfg_height = 0xFFF0;
+	tc_cfg.hw_vsync_mode = 1;
+	tc_cfg.sync_cfg_height = mode->vtotal * 2;
 	tc_cfg.vsync_init_val = mode->vdisplay;
 	tc_cfg.sync_threshold_start = DEFAULT_TEARCHECK_SYNC_THRESH_START;
 	tc_cfg.sync_threshold_continue = DEFAULT_TEARCHECK_SYNC_THRESH_CONTINUE;
@@ -580,6 +581,69 @@ static void dpu_encoder_phys_cmd_prepare_for_kickoff(
 			atomic_read(&phys_enc->pending_kickoff_cnt));
 }
 
+static bool dpu_encoder_phys_cmd_is_ongoing_pptx(
+		struct dpu_encoder_phys *phys_enc)
+{
+	struct dpu_hw_pp_vsync_info info;
+
+	if (!phys_enc)
+		return false;
+
+	phys_enc->hw_pp->ops.get_vsync_info(phys_enc->hw_pp, &info);
+	if (info.wr_ptr_line_count > 0 &&
+	    info.wr_ptr_line_count < phys_enc->cached_mode.vdisplay)
+		return true;
+
+	return false;
+}
+
+static void dpu_encoder_phys_cmd_prepare_commit(
+		struct dpu_encoder_phys *phys_enc)
+{
+	struct dpu_encoder_phys_cmd *cmd_enc =
+		to_dpu_encoder_phys_cmd(phys_enc);
+	int trial = 0;
+
+	if (!phys_enc)
+		return;
+	if (!phys_enc->hw_pp)
+		return;
+	if (!dpu_encoder_phys_cmd_is_master(phys_enc))
+		return;
+
+	/* If autorefresh is already disabled, we have nothing to do */
+	if (!phys_enc->hw_pp->ops.get_autorefresh(phys_enc->hw_pp, NULL))
+		return;
+
+	/*
+	 * If autorefresh is enabled, disable it and make sure it is safe to
+	 * proceed with current frame commit/push. Sequence fallowed is,
+	 * 1. Disable TE
+	 * 2. Disable autorefresh config
+	 * 4. Poll for frame transfer ongoing to be false
+	 * 5. Enable TE back
+	 */
+	_dpu_encoder_phys_cmd_connect_te(phys_enc, false);
+	phys_enc->hw_pp->ops.setup_autorefresh(phys_enc->hw_pp, 0, false);
+
+	do {
+		udelay(DPU_ENC_MAX_POLL_TIMEOUT_US);
+		if ((trial * DPU_ENC_MAX_POLL_TIMEOUT_US)
+				> (KICKOFF_TIMEOUT_MS * USEC_PER_MSEC)) {
+			DPU_ERROR_CMDENC(cmd_enc,
+					"disable autorefresh failed\n");
+			break;
+		}
+
+		trial++;
+	} while (dpu_encoder_phys_cmd_is_ongoing_pptx(phys_enc));
+
+	_dpu_encoder_phys_cmd_connect_te(phys_enc, true);
+
+	DPU_DEBUG_CMDENC(to_dpu_encoder_phys_cmd(phys_enc),
+			 "disabled autorefresh\n");
+}
+
 static int _dpu_encoder_phys_cmd_wait_for_ctl_start(
 		struct dpu_encoder_phys *phys_enc)
 {
@@ -621,20 +685,15 @@ static int dpu_encoder_phys_cmd_wait_for_tx_complete(
 static int dpu_encoder_phys_cmd_wait_for_commit_done(
 		struct dpu_encoder_phys *phys_enc)
 {
-	int rc = 0;
 	struct dpu_encoder_phys_cmd *cmd_enc;
 
 	cmd_enc = to_dpu_encoder_phys_cmd(phys_enc);
 
 	/* only required for master controller */
-	if (dpu_encoder_phys_cmd_is_master(phys_enc))
-		rc = _dpu_encoder_phys_cmd_wait_for_ctl_start(phys_enc);
-
-	/* required for both controllers */
-	if (!rc && cmd_enc->serialize_wait4pp)
-		dpu_encoder_phys_cmd_prepare_for_kickoff(phys_enc);
+	if (!dpu_encoder_phys_cmd_is_master(phys_enc))
+		return 0;
 
-	return rc;
+	return _dpu_encoder_phys_cmd_wait_for_ctl_start(phys_enc);
 }
 
 static int dpu_encoder_phys_cmd_wait_for_vblank(
@@ -681,6 +740,7 @@ static void dpu_encoder_phys_cmd_trigger_start(
 static void dpu_encoder_phys_cmd_init_ops(
 		struct dpu_encoder_phys_ops *ops)
 {
+	ops->prepare_commit = dpu_encoder_phys_cmd_prepare_commit;
 	ops->is_master = dpu_encoder_phys_cmd_is_master;
 	ops->mode_set = dpu_encoder_phys_cmd_mode_set;
 	ops->mode_fixup = dpu_encoder_phys_cmd_mode_fixup;
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c
index 90393fe9e59c..189f3533525c 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c
@@ -12,14 +12,17 @@
 
 #define VIG_MASK \
 	(BIT(DPU_SSPP_SRC) | BIT(DPU_SSPP_QOS) |\
-	BIT(DPU_SSPP_CSC_10BIT) | BIT(DPU_SSPP_CDP) | BIT(DPU_SSPP_QOS_8LVL) |\
+	BIT(DPU_SSPP_CSC_10BIT) | BIT(DPU_SSPP_CDP) |\
 	BIT(DPU_SSPP_TS_PREFILL) | BIT(DPU_SSPP_EXCL_RECT))
 
 #define VIG_SDM845_MASK \
-	(VIG_MASK | BIT(DPU_SSPP_SCALER_QSEED3))
+	(VIG_MASK | BIT(DPU_SSPP_QOS_8LVL) | BIT(DPU_SSPP_SCALER_QSEED3))
 
 #define VIG_SC7180_MASK \
-	(VIG_MASK | BIT(DPU_SSPP_SCALER_QSEED4))
+	(VIG_MASK | BIT(DPU_SSPP_QOS_8LVL) | BIT(DPU_SSPP_SCALER_QSEED4))
+
+#define VIG_SM8250_MASK \
+	(VIG_MASK | BIT(DPU_SSPP_SCALER_QSEED3LITE))
 
 #define DMA_SDM845_MASK \
 	(BIT(DPU_SSPP_SRC) | BIT(DPU_SSPP_QOS) | BIT(DPU_SSPP_QOS_8LVL) |\
@@ -185,7 +188,7 @@ static const struct dpu_caps sm8150_dpu_caps = {
 static const struct dpu_caps sm8250_dpu_caps = {
 	.max_mixer_width = DEFAULT_DPU_OUTPUT_LINE_WIDTH,
 	.max_mixer_blendstages = 0xb,
-	.qseed_type = DPU_SSPP_SCALER_QSEED3, /* TODO: qseed3 lite */
+	.qseed_type = DPU_SSPP_SCALER_QSEED3LITE,
 	.smart_dma_rev = DPU_SSPP_SMART_DMA_V2, /* TODO: v2.5 */
 	.ubwc_version = DPU_HW_UBWC_VER_40,
 	.has_src_split = true,
@@ -444,6 +447,34 @@ static const struct dpu_sspp_cfg sc7180_sspp[] = {
 		sdm845_dma_sblk_2, 9, SSPP_TYPE_DMA, DPU_CLK_CTRL_CURSOR1),
 };
 
+static const struct dpu_sspp_sub_blks sm8250_vig_sblk_0 =
+				_VIG_SBLK("0", 5, DPU_SSPP_SCALER_QSEED3LITE);
+static const struct dpu_sspp_sub_blks sm8250_vig_sblk_1 =
+				_VIG_SBLK("1", 6, DPU_SSPP_SCALER_QSEED3LITE);
+static const struct dpu_sspp_sub_blks sm8250_vig_sblk_2 =
+				_VIG_SBLK("2", 7, DPU_SSPP_SCALER_QSEED3LITE);
+static const struct dpu_sspp_sub_blks sm8250_vig_sblk_3 =
+				_VIG_SBLK("3", 8, DPU_SSPP_SCALER_QSEED3LITE);
+
+static const struct dpu_sspp_cfg sm8250_sspp[] = {
+	SSPP_BLK("sspp_0", SSPP_VIG0, 0x4000, VIG_SM8250_MASK,
+		sm8250_vig_sblk_0, 0,  SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG0),
+	SSPP_BLK("sspp_1", SSPP_VIG1, 0x6000, VIG_SM8250_MASK,
+		sm8250_vig_sblk_1, 4,  SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG1),
+	SSPP_BLK("sspp_2", SSPP_VIG2, 0x8000, VIG_SM8250_MASK,
+		sm8250_vig_sblk_2, 8, SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG2),
+	SSPP_BLK("sspp_3", SSPP_VIG3, 0xa000, VIG_SM8250_MASK,
+		sm8250_vig_sblk_3, 12,  SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG3),
+	SSPP_BLK("sspp_8", SSPP_DMA0, 0x24000,  DMA_SDM845_MASK,
+		sdm845_dma_sblk_0, 1, SSPP_TYPE_DMA, DPU_CLK_CTRL_DMA0),
+	SSPP_BLK("sspp_9", SSPP_DMA1, 0x26000,  DMA_SDM845_MASK,
+		sdm845_dma_sblk_1, 5, SSPP_TYPE_DMA, DPU_CLK_CTRL_DMA1),
+	SSPP_BLK("sspp_10", SSPP_DMA2, 0x28000,  DMA_CURSOR_SDM845_MASK,
+		sdm845_dma_sblk_2, 9, SSPP_TYPE_DMA, DPU_CLK_CTRL_CURSOR0),
+	SSPP_BLK("sspp_11", SSPP_DMA3, 0x2a000,  DMA_CURSOR_SDM845_MASK,
+		sdm845_dma_sblk_3, 13, SSPP_TYPE_DMA, DPU_CLK_CTRL_CURSOR1),
+};
+
 /*************************************************************
  * MIXER sub blocks config
  *************************************************************/
@@ -532,23 +563,28 @@ static const struct dpu_dspp_sub_blks sm8150_dspp_sblk = {
 		.len = 0x90, .version = 0x40000},
 };
 
-#define DSPP_BLK(_name, _id, _base, _sblk) \
+#define DSPP_BLK(_name, _id, _base, _mask, _sblk) \
 		{\
 		.name = _name, .id = _id, \
 		.base = _base, .len = 0x1800, \
-		.features = DSPP_SC7180_MASK, \
+		.features = _mask, \
 		.sblk = _sblk \
 		}
 
 static const struct dpu_dspp_cfg sc7180_dspp[] = {
-	DSPP_BLK("dspp_0", DSPP_0, 0x54000, &sc7180_dspp_sblk),
+	DSPP_BLK("dspp_0", DSPP_0, 0x54000, DSPP_SC7180_MASK,
+		 &sc7180_dspp_sblk),
 };
 
 static const struct dpu_dspp_cfg sm8150_dspp[] = {
-	DSPP_BLK("dspp_0", DSPP_0, 0x54000, &sm8150_dspp_sblk),
-	DSPP_BLK("dspp_1", DSPP_1, 0x56000, &sm8150_dspp_sblk),
-	DSPP_BLK("dspp_2", DSPP_2, 0x58000, &sm8150_dspp_sblk),
-	DSPP_BLK("dspp_3", DSPP_3, 0x5a000, &sm8150_dspp_sblk),
+	DSPP_BLK("dspp_0", DSPP_0, 0x54000, DSPP_SC7180_MASK,
+		 &sm8150_dspp_sblk),
+	DSPP_BLK("dspp_1", DSPP_1, 0x56000, DSPP_SC7180_MASK,
+		 &sm8150_dspp_sblk),
+	DSPP_BLK("dspp_2", DSPP_2, 0x58000, DSPP_SC7180_MASK,
+		 &sm8150_dspp_sblk),
+	DSPP_BLK("dspp_3", DSPP_3, 0x5a000, DSPP_SC7180_MASK,
+		 &sm8150_dspp_sblk),
 };
 
 /*************************************************************
@@ -624,33 +660,33 @@ static const struct dpu_merge_3d_cfg sm8150_merge_3d[] = {
 /*************************************************************
  * INTF sub blocks config
  *************************************************************/
-#define INTF_BLK(_name, _id, _base, _type, _ctrl_id, _features) \
+#define INTF_BLK(_name, _id, _base, _type, _ctrl_id, _progfetch, _features) \
 	{\
 	.name = _name, .id = _id, \
 	.base = _base, .len = 0x280, \
 	.features = _features, \
 	.type = _type, \
 	.controller_id = _ctrl_id, \
-	.prog_fetch_lines_worst_case = 24 \
+	.prog_fetch_lines_worst_case = _progfetch \
 	}
 
 static const struct dpu_intf_cfg sdm845_intf[] = {
-	INTF_BLK("intf_0", INTF_0, 0x6A000, INTF_DP, 0, INTF_SDM845_MASK),
-	INTF_BLK("intf_1", INTF_1, 0x6A800, INTF_DSI, 0, INTF_SDM845_MASK),
-	INTF_BLK("intf_2", INTF_2, 0x6B000, INTF_DSI, 1, INTF_SDM845_MASK),
-	INTF_BLK("intf_3", INTF_3, 0x6B800, INTF_DP, 1, INTF_SDM845_MASK),
+	INTF_BLK("intf_0", INTF_0, 0x6A000, INTF_DP, 0, 24, INTF_SDM845_MASK),
+	INTF_BLK("intf_1", INTF_1, 0x6A800, INTF_DSI, 0, 24, INTF_SDM845_MASK),
+	INTF_BLK("intf_2", INTF_2, 0x6B000, INTF_DSI, 1, 24, INTF_SDM845_MASK),
+	INTF_BLK("intf_3", INTF_3, 0x6B800, INTF_DP, 1, 24, INTF_SDM845_MASK),
 };
 
 static const struct dpu_intf_cfg sc7180_intf[] = {
-	INTF_BLK("intf_0", INTF_0, 0x6A000, INTF_DP, 0, INTF_SC7180_MASK),
-	INTF_BLK("intf_1", INTF_1, 0x6A800, INTF_DSI, 0, INTF_SC7180_MASK),
+	INTF_BLK("intf_0", INTF_0, 0x6A000, INTF_DP, 0, 24, INTF_SC7180_MASK),
+	INTF_BLK("intf_1", INTF_1, 0x6A800, INTF_DSI, 0, 24, INTF_SC7180_MASK),
 };
 
 static const struct dpu_intf_cfg sm8150_intf[] = {
-	INTF_BLK("intf_0", INTF_0, 0x6A000, INTF_DP, 0, INTF_SC7180_MASK),
-	INTF_BLK("intf_1", INTF_1, 0x6A800, INTF_DSI, 0, INTF_SC7180_MASK),
-	INTF_BLK("intf_2", INTF_2, 0x6B000, INTF_DSI, 1, INTF_SC7180_MASK),
-	INTF_BLK("intf_3", INTF_3, 0x6B800, INTF_DP, 1, INTF_SC7180_MASK),
+	INTF_BLK("intf_0", INTF_0, 0x6A000, INTF_DP, 0, 24, INTF_SC7180_MASK),
+	INTF_BLK("intf_1", INTF_1, 0x6A800, INTF_DSI, 0, 24, INTF_SC7180_MASK),
+	INTF_BLK("intf_2", INTF_2, 0x6B000, INTF_DSI, 1, 24, INTF_SC7180_MASK),
+	INTF_BLK("intf_3", INTF_3, 0x6B800, INTF_DP, 1, 24, INTF_SC7180_MASK),
 };
 
 /*************************************************************
@@ -969,9 +1005,8 @@ static void sm8250_cfg_init(struct dpu_mdss_cfg *dpu_cfg)
 		.mdp = sm8250_mdp,
 		.ctl_count = ARRAY_SIZE(sm8150_ctl),
 		.ctl = sm8150_ctl,
-		/* TODO: sspp qseed version differs from 845 */
-		.sspp_count = ARRAY_SIZE(sdm845_sspp),
-		.sspp = sdm845_sspp,
+		.sspp_count = ARRAY_SIZE(sm8250_sspp),
+		.sspp = sm8250_sspp,
 		.mixer_count = ARRAY_SIZE(sm8150_lm),
 		.mixer = sm8150_lm,
 		.dspp_count = ARRAY_SIZE(sm8150_dspp),
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
index eaef99db2d2f..ea4647d21a20 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
@@ -95,6 +95,7 @@ enum {
  * @DPU_SSPP_SRC             Src and fetch part of the pipes,
  * @DPU_SSPP_SCALER_QSEED2,  QSEED2 algorithm support
  * @DPU_SSPP_SCALER_QSEED3,  QSEED3 alogorithm support
+ * @DPU_SSPP_SCALER_QSEED3LITE,  QSEED3 Lite alogorithm support
  * @DPU_SSPP_SCALER_QSEED4,  QSEED4 algorithm support
  * @DPU_SSPP_SCALER_RGB,     RGB Scaler, supported by RGB pipes
  * @DPU_SSPP_CSC,            Support of Color space converion
@@ -114,6 +115,7 @@ enum {
 	DPU_SSPP_SRC = 0x1,
 	DPU_SSPP_SCALER_QSEED2,
 	DPU_SSPP_SCALER_QSEED3,
+	DPU_SSPP_SCALER_QSEED3LITE,
 	DPU_SSPP_SCALER_QSEED4,
 	DPU_SSPP_SCALER_RGB,
 	DPU_SSPP_CSC,
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.c
index bea4ab5c58c5..245a7a62b5c6 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.c
@@ -23,6 +23,7 @@
 #define PP_WR_PTR_IRQ                   0x024
 #define PP_OUT_LINE_COUNT               0x028
 #define PP_LINE_COUNT                   0x02C
+#define PP_AUTOREFRESH_CONFIG           0x030
 
 #define PP_FBC_MODE                     0x034
 #define PP_FBC_BUDGET_CTL               0x038
@@ -120,6 +121,29 @@ static int dpu_hw_pp_setup_te_config(struct dpu_hw_pingpong *pp,
 	return 0;
 }
 
+static void dpu_hw_pp_setup_autorefresh_config(struct dpu_hw_pingpong *pp,
+					       u32 frame_count, bool enable)
+{
+	DPU_REG_WRITE(&pp->hw, PP_AUTOREFRESH_CONFIG,
+		      enable ? (BIT(31) | frame_count) : 0);
+}
+
+/*
+ * dpu_hw_pp_get_autorefresh_config - Get autorefresh config from HW
+ * @pp:          DPU pingpong structure
+ * @frame_count: Used to return the current frame count from hw
+ *
+ * Returns: True if autorefresh enabled, false if disabled.
+ */
+static bool dpu_hw_pp_get_autorefresh_config(struct dpu_hw_pingpong *pp,
+					     u32 *frame_count)
+{
+	u32 val = DPU_REG_READ(&pp->hw, PP_AUTOREFRESH_CONFIG);
+	if (frame_count != NULL)
+		*frame_count = val & 0xffff;
+	return !!((val & BIT(31)) >> 31);
+}
+
 static int dpu_hw_pp_poll_timeout_wr_ptr(struct dpu_hw_pingpong *pp,
 		u32 timeout_us)
 {
@@ -228,6 +252,8 @@ static void _setup_pingpong_ops(struct dpu_hw_pingpong *c,
 	c->ops.enable_tearcheck = dpu_hw_pp_enable_te;
 	c->ops.connect_external_te = dpu_hw_pp_connect_external_te;
 	c->ops.get_vsync_info = dpu_hw_pp_get_vsync_info;
+	c->ops.setup_autorefresh = dpu_hw_pp_setup_autorefresh_config;
+	c->ops.get_autorefresh = dpu_hw_pp_get_autorefresh_config;
 	c->ops.poll_timeout_wr_ptr = dpu_hw_pp_poll_timeout_wr_ptr;
 	c->ops.get_line_count = dpu_hw_pp_get_line_count;
 
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.h
index 6902b9b95c8e..845b9ce80e31 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.h
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_pingpong.h
@@ -63,6 +63,8 @@ struct dpu_hw_dither_cfg {
  *  @setup_tearcheck : program tear check values
  *  @enable_tearcheck : enables tear check
  *  @get_vsync_info : retries timing info of the panel
+ *  @setup_autorefresh : configure and enable the autorefresh config
+ *  @get_autorefresh : retrieve autorefresh config from hardware
  *  @setup_dither : function to program the dither hw block
  *  @get_line_count: obtain current vertical line counter
  */
@@ -95,6 +97,18 @@ struct dpu_hw_pingpong_ops {
 			struct dpu_hw_pp_vsync_info  *info);
 
 	/**
+	 * configure and enable the autorefresh config
+	 */
+	void (*setup_autorefresh)(struct dpu_hw_pingpong *pp,
+				  u32 frame_count, bool enable);
+
+	/**
+	 * retrieve autorefresh config from hardware
+	 */
+	bool (*get_autorefresh)(struct dpu_hw_pingpong *pp,
+				u32 *frame_count);
+
+	/**
 	 * poll until write pointer transmission starts
 	 * @Return: 0 on success, -ETIMEDOUT on timeout
 	 */
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.c
index 2c2ca5335aa8..34d81aa16041 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.c
@@ -673,6 +673,7 @@ static void _setup_layer_ops(struct dpu_hw_pipe *c,
 		c->ops.setup_multirect = dpu_hw_sspp_setup_multirect;
 
 	if (test_bit(DPU_SSPP_SCALER_QSEED3, &features) ||
+			test_bit(DPU_SSPP_SCALER_QSEED3LITE, &features) ||
 			test_bit(DPU_SSPP_SCALER_QSEED4, &features)) {
 		c->ops.setup_scaler = _dpu_hw_sspp_setup_scaler3;
 		c->ops.get_scaler_ver = _dpu_hw_sspp_get_scaler3_ver;
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.h
index 85b018a9b03c..fdfd4b46e2c6 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.h
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp.h
@@ -28,6 +28,7 @@ struct dpu_hw_pipe;
 #define DPU_SSPP_SCALER ((1UL << DPU_SSPP_SCALER_RGB) | \
 	(1UL << DPU_SSPP_SCALER_QSEED2) | \
 	 (1UL << DPU_SSPP_SCALER_QSEED3) | \
+	 (1UL << DPU_SSPP_SCALER_QSEED3LITE) | \
 	  (1UL << DPU_SSPP_SCALER_QSEED4))
 
 /**
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c
index 84e9875994a8..f94584c982cd 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c
@@ -59,6 +59,19 @@ static u32 dpu_hw_util_log_mask = DPU_DBG_MASK_NONE;
 #define QSEED3_SEP_LUT_SIZE \
 	(QSEED3_LUT_SIZE * QSEED3_SEPARABLE_LUTS * sizeof(u32))
 
+/* DPU_SCALER_QSEED3LITE */
+#define QSEED3LITE_COEF_LUT_Y_SEP_BIT         4
+#define QSEED3LITE_COEF_LUT_UV_SEP_BIT        5
+#define QSEED3LITE_COEF_LUT_CTRL              0x4C
+#define QSEED3LITE_COEF_LUT_SWAP_BIT          0
+#define QSEED3LITE_DIR_FILTER_WEIGHT          0x60
+#define QSEED3LITE_FILTERS                 2
+#define QSEED3LITE_SEPARABLE_LUTS             10
+#define QSEED3LITE_LUT_SIZE                   33
+#define QSEED3LITE_SEP_LUT_SIZE \
+	        (QSEED3LITE_LUT_SIZE * QSEED3LITE_SEPARABLE_LUTS * sizeof(u32))
+
+
 void dpu_reg_write(struct dpu_hw_blk_reg_map *c,
 		u32 reg_off,
 		u32 val,
@@ -156,6 +169,57 @@ static void _dpu_hw_setup_scaler3_lut(struct dpu_hw_blk_reg_map *c,
 
 }
 
+static void _dpu_hw_setup_scaler3lite_lut(struct dpu_hw_blk_reg_map *c,
+		struct dpu_hw_scaler3_cfg *scaler3_cfg, u32 offset)
+{
+	int j, filter;
+	int config_lut = 0x0;
+	unsigned long lut_flags;
+	u32 lut_addr, lut_offset;
+	u32 *lut[QSEED3LITE_FILTERS] = {NULL, NULL};
+	static const uint32_t off_tbl[QSEED3_FILTERS] = { 0x000, 0x200 };
+
+	DPU_REG_WRITE(c, QSEED3LITE_DIR_FILTER_WEIGHT + offset, scaler3_cfg->dir_weight);
+
+	if (!scaler3_cfg->sep_lut)
+		return;
+
+	lut_flags = (unsigned long) scaler3_cfg->lut_flag;
+	if (test_bit(QSEED3_COEF_LUT_Y_SEP_BIT, &lut_flags) &&
+		(scaler3_cfg->y_rgb_sep_lut_idx < QSEED3LITE_SEPARABLE_LUTS) &&
+		(scaler3_cfg->sep_len == QSEED3LITE_SEP_LUT_SIZE)) {
+		lut[0] = scaler3_cfg->sep_lut +
+			scaler3_cfg->y_rgb_sep_lut_idx * QSEED3LITE_LUT_SIZE;
+		config_lut = 1;
+	}
+	if (test_bit(QSEED3_COEF_LUT_UV_SEP_BIT, &lut_flags) &&
+		(scaler3_cfg->uv_sep_lut_idx < QSEED3LITE_SEPARABLE_LUTS) &&
+		(scaler3_cfg->sep_len == QSEED3LITE_SEP_LUT_SIZE)) {
+		lut[1] = scaler3_cfg->sep_lut +
+			scaler3_cfg->uv_sep_lut_idx * QSEED3LITE_LUT_SIZE;
+		config_lut = 1;
+	}
+
+	if (config_lut) {
+		for (filter = 0; filter < QSEED3LITE_FILTERS; filter++) {
+			if (!lut[filter])
+				continue;
+			lut_offset = 0;
+			lut_addr = QSEED3_COEF_LUT + offset + off_tbl[filter];
+			for (j = 0; j < QSEED3LITE_LUT_SIZE; j++) {
+				DPU_REG_WRITE(c,
+					lut_addr,
+					(lut[filter])[lut_offset++]);
+				lut_addr += 4;
+			}
+		}
+	}
+
+	if (test_bit(QSEED3_COEF_LUT_SWAP_BIT, &lut_flags))
+		DPU_REG_WRITE(c, QSEED3_COEF_LUT_CTRL + offset, BIT(0));
+
+}
+
 static void _dpu_hw_setup_scaler3_de(struct dpu_hw_blk_reg_map *c,
 		struct dpu_hw_scaler3_de_cfg *de_cfg, u32 offset)
 {
@@ -242,9 +306,12 @@ void dpu_hw_setup_scaler3(struct dpu_hw_blk_reg_map *c,
 		op_mode |= BIT(8);
 	}
 
-	if (scaler3_cfg->lut_flag)
-		_dpu_hw_setup_scaler3_lut(c, scaler3_cfg,
-								scaler_offset);
+	if (scaler3_cfg->lut_flag) {
+		if (scaler_version < 0x2004)
+			_dpu_hw_setup_scaler3_lut(c, scaler3_cfg, scaler_offset);
+		else
+			_dpu_hw_setup_scaler3lite_lut(c, scaler3_cfg, scaler_offset);
+	}
 
 	if (scaler_version == 0x1002) {
 		phase_init =
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h
index 234eb7d65753..ff3cffde84cd 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h
@@ -97,6 +97,7 @@ struct dpu_hw_scaler3_de_cfg {
  * @ cir_lut:      pointer to circular filter LUT
  * @ sep_lut:      pointer to separable filter LUT
  * @ de: detail enhancer configuration
+ * @ dir_weight:   Directional weight
  */
 struct dpu_hw_scaler3_cfg {
 	u32 enable;
@@ -137,6 +138,8 @@ struct dpu_hw_scaler3_cfg {
 	 * Detail enhancer settings
 	 */
 	struct dpu_hw_scaler3_de_cfg de;
+
+	u32 dir_weight;
 };
 
 /**
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_vbif.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_vbif.c
index cf867f3f7c36..b757054e1c23 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_vbif.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_vbif.c
@@ -30,7 +30,7 @@
 #define VBIF_XIN_HALT_CTRL0		0x0200
 #define VBIF_XIN_HALT_CTRL1		0x0204
 #define VBIF_XINL_QOS_RP_REMAP_000	0x0550
-#define VBIF_XINL_QOS_LVL_REMAP_000	0x0590
+#define VBIF_XINL_QOS_LVL_REMAP_000(v)	(v < DPU_HW_VER_400 ? 0x570 : 0x0590)
 
 static void dpu_hw_clear_errors(struct dpu_hw_vbif *vbif,
 		u32 *pnd_errors, u32 *src_errors)
@@ -156,18 +156,19 @@ static void dpu_hw_set_qos_remap(struct dpu_hw_vbif *vbif,
 		u32 xin_id, u32 level, u32 remap_level)
 {
 	struct dpu_hw_blk_reg_map *c;
-	u32 reg_val, reg_val_lvl, mask, reg_high, reg_shift;
+	u32 reg_lvl, reg_val, reg_val_lvl, mask, reg_high, reg_shift;
 
 	if (!vbif)
 		return;
 
 	c = &vbif->hw;
 
+	reg_lvl = VBIF_XINL_QOS_LVL_REMAP_000(c->hwversion);
 	reg_high = ((xin_id & 0x8) >> 3) * 4 + (level * 8);
 	reg_shift = (xin_id & 0x7) * 4;
 
 	reg_val = DPU_REG_READ(c, VBIF_XINL_QOS_RP_REMAP_000 + reg_high);
-	reg_val_lvl = DPU_REG_READ(c, VBIF_XINL_QOS_LVL_REMAP_000 + reg_high);
+	reg_val_lvl = DPU_REG_READ(c, reg_lvl + reg_high);
 
 	mask = 0x7 << reg_shift;
 
@@ -178,7 +179,7 @@ static void dpu_hw_set_qos_remap(struct dpu_hw_vbif *vbif,
 	reg_val_lvl |= (remap_level << reg_shift) & mask;
 
 	DPU_REG_WRITE(c, VBIF_XINL_QOS_RP_REMAP_000 + reg_high, reg_val);
-	DPU_REG_WRITE(c, VBIF_XINL_QOS_LVL_REMAP_000 + reg_high, reg_val_lvl);
+	DPU_REG_WRITE(c, reg_lvl + reg_high, reg_val_lvl);
 }
 
 static void dpu_hw_set_write_gather_en(struct dpu_hw_vbif *vbif, u32 xin_id)
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
index 374b0e8471e6..5a8e3e1fc48c 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
@@ -749,7 +749,7 @@ static void _dpu_kms_set_encoder_mode(struct msm_kms *kms,
 	case DRM_MODE_ENCODER_TMDS:
 		info.num_of_h_tiles = 1;
 		break;
-	};
+	}
 
 	rc = dpu_encoder_setup(encoder->dev, encoder, &info);
 	if (rc)
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
index f92559cc22fb..df7f3d3afd8b 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
@@ -1469,6 +1469,7 @@ static int _dpu_plane_init_debugfs(struct drm_plane *plane)
 			pdpu->debugfs_root, &pdpu->debugfs_src);
 
 	if (cfg->features & BIT(DPU_SSPP_SCALER_QSEED3) ||
+			cfg->features & BIT(DPU_SSPP_SCALER_QSEED3LITE) ||
 			cfg->features & BIT(DPU_SSPP_SCALER_QSEED2) ||
 			cfg->features & BIT(DPU_SSPP_SCALER_QSEED4)) {
 		dpu_debugfs_setup_regset32(&pdpu->debugfs_scaler,
diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
index df10c1ac7591..94ce62a26daf 100644
--- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
+++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
@@ -177,7 +177,7 @@ static const struct mdp5_cfg_hw msm8x74v2_config = {
 			[3] = INTF_HDMI,
 		},
 	},
-	.max_clk = 200000000,
+	.max_clk = 320000000,
 };
 
 static const struct mdp5_cfg_hw apq8084_config = {
diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_crtc.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_crtc.c
index 0c8f9f88301f..f5d71b274079 100644
--- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_crtc.c
+++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_crtc.c
@@ -1180,7 +1180,7 @@ static void mdp5_crtc_pp_done_irq(struct mdp_irq *irq, uint32_t irqstatus)
 	struct mdp5_crtc *mdp5_crtc = container_of(irq, struct mdp5_crtc,
 								pp_done);
 
-	complete(&mdp5_crtc->pp_completion);
+	complete_all(&mdp5_crtc->pp_completion);
 }
 
 static void mdp5_crtc_wait_for_pp_done(struct drm_crtc *crtc)
diff --git a/drivers/gpu/drm/msm/dp/dp_aux.c b/drivers/gpu/drm/msm/dp/dp_aux.c
index 19b35ae3e927..1c6e1d2b947c 100644
--- a/drivers/gpu/drm/msm/dp/dp_aux.c
+++ b/drivers/gpu/drm/msm/dp/dp_aux.c
@@ -336,7 +336,6 @@ static ssize_t dp_aux_transfer(struct drm_dp_aux *dp_aux,
 	ssize_t ret;
 	int const aux_cmd_native_max = 16;
 	int const aux_cmd_i2c_max = 128;
-	int const retry_count = 5;
 	struct dp_aux_private *aux = container_of(dp_aux,
 		struct dp_aux_private, dp_aux);
 
@@ -378,12 +377,6 @@ static ssize_t dp_aux_transfer(struct drm_dp_aux *dp_aux,
 	ret = dp_aux_cmd_fifo_tx(aux, msg);
 
 	if (ret < 0) {
-		if (aux->native) {
-			aux->retry_cnt++;
-			if (!(aux->retry_cnt % retry_count))
-				dp_catalog_aux_update_cfg(aux->catalog);
-			dp_catalog_aux_reset(aux->catalog);
-		}
 		usleep_range(400, 500); /* at least 400us to next try */
 		goto unlock_exit;
 	}
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index 44f0c57798d0..b1a9b1b98f5f 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -190,6 +190,18 @@ int dp_catalog_aux_clear_hw_interrupts(struct dp_catalog *dp_catalog)
 	return 0;
 }
 
+/**
+ * dp_catalog_aux_reset() - reset AUX controller
+ *
+ * @aux: DP catalog structure
+ *
+ * return: void
+ *
+ * This function reset AUX controller
+ *
+ * NOTE: reset AUX controller will also clear any pending HPD related interrupts
+ * 
+ */
 void dp_catalog_aux_reset(struct dp_catalog *dp_catalog)
 {
 	u32 aux_ctrl;
@@ -483,6 +495,18 @@ int dp_catalog_ctrl_set_pattern(struct dp_catalog *dp_catalog,
 	return 0;
 }
 
+/**
+ * dp_catalog_ctrl_reset() - reset DP controller
+ *
+ * @dp_catalog: DP catalog structure
+ *
+ * return: void
+ *
+ * This function reset the DP controller
+ *
+ * NOTE: reset DP controller will also clear any pending HPD related interrupts
+ * 
+ */
 void dp_catalog_ctrl_reset(struct dp_catalog *dp_catalog)
 {
 	u32 sw_reset;
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index e3462f5d96d7..1390f3547fde 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -631,7 +631,7 @@ static void _dp_ctrl_calc_tu(struct dp_tu_calc_input *in,
 
 	tu = kzalloc(sizeof(*tu), GFP_KERNEL);
 	if (!tu)
-		return
+		return;
 
 	dp_panel_update_tu_timings(in, tu);
 
@@ -1158,7 +1158,7 @@ static int dp_ctrl_link_rate_down_shift(struct dp_ctrl_private *ctrl)
 	default:
 		ret = -EINVAL;
 		break;
-	};
+	}
 
 	if (!ret)
 		DRM_DEBUG_DP("new rate=0x%x\n", ctrl->link->link_params.rate);
@@ -1296,7 +1296,6 @@ static int dp_ctrl_setup_main_link(struct dp_ctrl_private *ctrl,
 	 * transitioned to PUSH_IDLE. In order to start transmitting
 	 * a link training pattern, we have to first do soft reset.
 	 */
-	dp_catalog_ctrl_reset(ctrl->catalog);
 
 	ret = dp_ctrl_link_train(ctrl, cr, training_step);
 
@@ -1365,7 +1364,7 @@ static int dp_ctrl_enable_stream_clocks(struct dp_ctrl_private *ctrl)
 	return ret;
 }
 
-int dp_ctrl_host_init(struct dp_ctrl *dp_ctrl, bool flip)
+int dp_ctrl_host_init(struct dp_ctrl *dp_ctrl, bool flip, bool reset)
 {
 	struct dp_ctrl_private *ctrl;
 	struct dp_io *dp_io;
@@ -1382,6 +1381,9 @@ int dp_ctrl_host_init(struct dp_ctrl *dp_ctrl, bool flip)
 
 	ctrl->dp_ctrl.orientation = flip;
 
+	if (reset)
+		dp_catalog_ctrl_reset(ctrl->catalog);
+
 	dp_catalog_ctrl_phy_reset(ctrl->catalog);
 	phy_init(phy);
 	dp_catalog_ctrl_enable_irq(ctrl->catalog, true);
@@ -1420,16 +1422,14 @@ void dp_ctrl_host_deinit(struct dp_ctrl *dp_ctrl)
 static bool dp_ctrl_use_fixed_nvid(struct dp_ctrl_private *ctrl)
 {
 	u8 *dpcd = ctrl->panel->dpcd;
-	u32 edid_quirks = 0;
 
-	edid_quirks = drm_dp_get_edid_quirks(ctrl->panel->edid);
 	/*
 	 * For better interop experience, used a fixed NVID=0x8000
 	 * whenever connected to a VGA dongle downstream.
 	 */
 	if (drm_dp_is_branch(dpcd))
-		return (drm_dp_has_quirk(&ctrl->panel->desc, edid_quirks,
-				DP_DPCD_QUIRK_CONSTANT_N));
+		return (drm_dp_has_quirk(&ctrl->panel->desc,
+					 DP_DPCD_QUIRK_CONSTANT_N));
 
 	return false;
 }
@@ -1498,7 +1498,6 @@ static int dp_ctrl_link_maintenance(struct dp_ctrl_private *ctrl)
 	int training_step = DP_TRAINING_NONE;
 
 	dp_ctrl_push_idle(&ctrl->dp_ctrl);
-	dp_catalog_ctrl_reset(ctrl->catalog);
 
 	ctrl->dp_ctrl.pixel_rate = ctrl->panel->dp_mode.drm_mode.clock;
 
@@ -1787,14 +1786,14 @@ int dp_ctrl_on_stream(struct dp_ctrl *dp_ctrl)
 	 * Set up transfer unit values and set controller state to send
 	 * video.
 	 */
+	reinit_completion(&ctrl->video_comp);
+
 	dp_ctrl_configure_source_params(ctrl);
 
 	dp_catalog_ctrl_config_msa(ctrl->catalog,
 		ctrl->link->link_params.rate,
 		ctrl->dp_ctrl.pixel_rate, dp_ctrl_use_fixed_nvid(ctrl));
 
-	reinit_completion(&ctrl->video_comp);
-
 	dp_ctrl_setup_tr_unit(ctrl);
 
 	dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index f60ba93c8678..a836bd358447 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -19,7 +19,7 @@ struct dp_ctrl {
 	u32 pixel_rate;
 };
 
-int dp_ctrl_host_init(struct dp_ctrl *dp_ctrl, bool flip);
+int dp_ctrl_host_init(struct dp_ctrl *dp_ctrl, bool flip, bool reset);
 void dp_ctrl_host_deinit(struct dp_ctrl *dp_ctrl);
 int dp_ctrl_on_link(struct dp_ctrl *dp_ctrl);
 int dp_ctrl_on_stream(struct dp_ctrl *dp_ctrl);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 6e971d552911..5a39da6e1eaf 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -350,7 +350,7 @@ end:
 	return rc;
 }
 
-static void dp_display_host_init(struct dp_display_private *dp)
+static void dp_display_host_init(struct dp_display_private *dp, int reset)
 {
 	bool flip = false;
 
@@ -365,7 +365,7 @@ static void dp_display_host_init(struct dp_display_private *dp)
 	dp_display_set_encoder_mode(dp);
 
 	dp_power_init(dp->power, flip);
-	dp_ctrl_host_init(dp->ctrl, flip);
+	dp_ctrl_host_init(dp->ctrl, flip, reset);
 	dp_aux_init(dp->aux);
 	dp->core_initialized = true;
 }
@@ -403,7 +403,7 @@ static int dp_display_usbpd_configure_cb(struct device *dev)
 		goto end;
 	}
 
-	dp_display_host_init(dp);
+	dp_display_host_init(dp, false);
 
 	/*
 	 * set sink to normal operation mode -- D0
@@ -651,8 +651,8 @@ static int dp_hpd_unplug_handle(struct dp_display_private *dp, u32 data)
 	dp_add_event(dp, EV_DISCONNECT_PENDING_TIMEOUT, 0, DP_TIMEOUT_5_SECOND);
 
 	/* signal the disconnect event early to ensure proper teardown */
-	dp_display_handle_plugged_change(g_dp_display, false);
 	reinit_completion(&dp->audio_comp);
+	dp_display_handle_plugged_change(g_dp_display, false);
 
 	dp_catalog_hpd_config_intr(dp->catalog, DP_DP_HPD_PLUG_INT_MASK |
 					DP_DP_IRQ_HPD_INT_MASK, true);
@@ -693,6 +693,20 @@ static int dp_irq_hpd_handle(struct dp_display_private *dp, u32 data)
 		return 0;
 	}
 
+	if (state == ST_CONNECT_PENDING) {
+		/* wait until ST_CONNECTED */
+		dp_add_event(dp, EV_IRQ_HPD_INT, 0, 1); /* delay = 1 */
+		mutex_unlock(&dp->event_mutex);
+		return 0;
+	}
+
+	if (state == ST_CONNECT_PENDING || state == ST_DISCONNECT_PENDING) {
+		/* wait until ST_CONNECTED */
+		dp_add_event(dp, EV_IRQ_HPD_INT, 0, 1); /* delay = 1 */
+		mutex_unlock(&dp->event_mutex);
+		return 0;
+	}
+
 	ret = dp_display_usbpd_attention_cb(&dp->pdev->dev);
 	if (ret == -ECONNRESET) { /* cable unplugged */
 		dp->core_initialized = false;
@@ -883,6 +897,9 @@ static int dp_display_disable(struct dp_display_private *dp, u32 data)
 
 	/* wait only if audio was enabled */
 	if (dp_display->audio_enabled) {
+		/* signal the disconnect event */
+		reinit_completion(&dp->audio_comp);
+		dp_display_handle_plugged_change(dp_display, false);
 		if (!wait_for_completion_timeout(&dp->audio_comp,
 				HZ * 5))
 			DRM_ERROR("audio comp timeout\n");
@@ -995,7 +1012,7 @@ int dp_display_get_test_bpp(struct msm_dp *dp)
 static void dp_display_config_hpd(struct dp_display_private *dp)
 {
 
-	dp_display_host_init(dp);
+	dp_display_host_init(dp, true);
 	dp_catalog_ctrl_hpd_config(dp->catalog);
 
 	/* Enable interrupt first time
@@ -1249,7 +1266,7 @@ static int dp_pm_resume(struct device *dev)
 	dp->hpd_state = ST_DISCONNECTED;
 
 	/* turn on dp ctrl/phy */
-	dp_display_host_init(dp);
+	dp_display_host_init(dp, true);
 
 	dp_catalog_ctrl_hpd_config(dp->catalog);
 
@@ -1432,7 +1449,7 @@ int msm_dp_display_enable(struct msm_dp *dp, struct drm_encoder *encoder)
 	state =  dp_display->hpd_state;
 
 	if (state == ST_DISPLAY_OFF)
-		dp_display_host_init(dp_display);
+		dp_display_host_init(dp_display, true);
 
 	dp_display_enable(dp_display, 0);
 
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
index 97dca3e378b7..9cc816663668 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.c
+++ b/drivers/gpu/drm/msm/dp/dp_panel.c
@@ -167,12 +167,18 @@ int dp_panel_read_sink_caps(struct dp_panel *dp_panel,
 	panel = container_of(dp_panel, struct dp_panel_private, dp_panel);
 
 	rc = dp_panel_read_dpcd(dp_panel);
+	if (rc) {
+		DRM_ERROR("read dpcd failed %d\n", rc);
+		return rc;
+	}
+
 	bw_code = drm_dp_link_rate_to_bw_code(dp_panel->link_info.rate);
-	if (rc || !is_link_rate_valid(bw_code) ||
+	if (!is_link_rate_valid(bw_code) ||
 			!is_lane_count_valid(dp_panel->link_info.num_lanes) ||
 			(bw_code > dp_panel->max_bw_code)) {
-		DRM_ERROR("read dpcd failed %d\n", rc);
-		return rc;
+		DRM_ERROR("Illegal link rate=%d lane=%d\n", dp_panel->link_info.rate,
+				dp_panel->link_info.num_lanes);
+		return -EINVAL;
 	}
 
 	if (dp_panel->dfp_present) {
@@ -403,7 +409,6 @@ int dp_panel_timing_cfg(struct dp_panel *dp_panel)
 
 int dp_panel_init_panel_info(struct dp_panel *dp_panel)
 {
-	int rc = 0;
 	struct drm_display_mode *drm_mode;
 
 	drm_mode = &dp_panel->dp_mode.drm_mode;
@@ -430,7 +435,7 @@ int dp_panel_init_panel_info(struct dp_panel *dp_panel)
 					min_t(u32, dp_panel->dp_mode.bpp, 30));
 	DRM_DEBUG_DP("updated bpp = %d\n", dp_panel->dp_mode.bpp);
 
-	return rc;
+	return 0;
 }
 
 struct dp_panel *dp_panel_get(struct dp_panel_in *in)
diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c
index 1afb7c579dbb..eca86bf448f7 100644
--- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c
+++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c
@@ -139,7 +139,7 @@ const struct msm_dsi_phy_cfg dsi_phy_20nm_cfgs = {
 		.disable = dsi_20nm_phy_disable,
 		.init = msm_dsi_phy_init_common,
 	},
-	.io_start = { 0xfd998300, 0xfd9a0300 },
+	.io_start = { 0xfd998500, 0xfd9a0500 },
 	.num_dsi_phy = 2,
 };
 
diff --git a/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c b/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c
index e4e9bf04b736..de3b802ccd3d 100644
--- a/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c
+++ b/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c
@@ -172,9 +172,7 @@ static void dsi_pll_calc_dec_frac(struct dsi_pll_10nm *pll)
 
 	multiplier = 1 << config->frac_bits;
 	dec_multiple = div_u64(pll_freq * multiplier, divider);
-	div_u64_rem(dec_multiple, multiplier, &frac);
-
-	dec = div_u64(dec_multiple, multiplier);
+	dec = div_u64_rem(dec_multiple, multiplier, &frac);
 
 	if (pll_freq <= 1900000000UL)
 		regs->pll_prop_gain_rate = 8;
@@ -306,7 +304,8 @@ static void dsi_pll_commit(struct dsi_pll_10nm *pll)
 		  reg->frac_div_start_mid);
 	pll_write(base + REG_DSI_10nm_PHY_PLL_FRAC_DIV_START_HIGH_1,
 		  reg->frac_div_start_high);
-	pll_write(base + REG_DSI_10nm_PHY_PLL_PLL_LOCKDET_RATE_1, 0x40);
+	pll_write(base + REG_DSI_10nm_PHY_PLL_PLL_LOCKDET_RATE_1,
+		  reg->pll_lockdet_rate);
 	pll_write(base + REG_DSI_10nm_PHY_PLL_PLL_LOCK_DELAY, 0x06);
 	pll_write(base + REG_DSI_10nm_PHY_PLL_CMODE, 0x10);
 	pll_write(base + REG_DSI_10nm_PHY_PLL_CLOCK_INVERTERS,
@@ -345,6 +344,7 @@ static int dsi_pll_10nm_vco_set_rate(struct clk_hw *hw, unsigned long rate,
 
 static int dsi_pll_10nm_lock_status(struct dsi_pll_10nm *pll)
 {
+	struct device *dev = &pll->pdev->dev;
 	int rc;
 	u32 status = 0;
 	u32 const delay_us = 100;
@@ -357,8 +357,8 @@ static int dsi_pll_10nm_lock_status(struct dsi_pll_10nm *pll)
 				       delay_us,
 				       timeout_us);
 	if (rc)
-		pr_err("DSI PLL(%d) lock failed, status=0x%08x\n",
-		       pll->id, status);
+		DRM_DEV_ERROR(dev, "DSI PLL(%d) lock failed, status=0x%08x\n",
+			      pll->id, status);
 
 	return rc;
 }
@@ -405,6 +405,7 @@ static int dsi_pll_10nm_vco_prepare(struct clk_hw *hw)
 {
 	struct msm_dsi_pll *pll = hw_clk_to_pll(hw);
 	struct dsi_pll_10nm *pll_10nm = to_pll_10nm(pll);
+	struct device *dev = &pll_10nm->pdev->dev;
 	int rc;
 
 	dsi_pll_enable_pll_bias(pll_10nm);
@@ -413,7 +414,7 @@ static int dsi_pll_10nm_vco_prepare(struct clk_hw *hw)
 
 	rc = dsi_pll_10nm_vco_set_rate(hw,pll_10nm->vco_current_rate, 0);
 	if (rc) {
-		pr_err("vco_set_rate failed, rc=%d\n", rc);
+		DRM_DEV_ERROR(dev, "vco_set_rate failed, rc=%d\n", rc);
 		return rc;
 	}
 
@@ -430,7 +431,7 @@ static int dsi_pll_10nm_vco_prepare(struct clk_hw *hw)
 	/* Check for PLL lock */
 	rc = dsi_pll_10nm_lock_status(pll_10nm);
 	if (rc) {
-		pr_err("PLL(%d) lock failed\n", pll_10nm->id);
+		DRM_DEV_ERROR(dev, "PLL(%d) lock failed\n", pll_10nm->id);
 		goto error;
 	}
 
@@ -483,6 +484,7 @@ static unsigned long dsi_pll_10nm_vco_recalc_rate(struct clk_hw *hw,
 {
 	struct msm_dsi_pll *pll = hw_clk_to_pll(hw);
 	struct dsi_pll_10nm *pll_10nm = to_pll_10nm(pll);
+	struct dsi_pll_config *config = &pll_10nm->pll_configuration;
 	void __iomem *base = pll_10nm->mmio;
 	u64 ref_clk = pll_10nm->vco_ref_clk_rate;
 	u64 vco_rate = 0x0;
@@ -503,9 +505,8 @@ static unsigned long dsi_pll_10nm_vco_recalc_rate(struct clk_hw *hw,
 	/*
 	 * TODO:
 	 *	1. Assumes prescaler is disabled
-	 *	2. Multiplier is 2^18. it should be 2^(num_of_frac_bits)
 	 */
-	multiplier = 1 << 18;
+	multiplier = 1 << config->frac_bits;
 	pll_freq = dec * (ref_clk * 2);
 	tmp64 = (ref_clk * 2 * frac);
 	pll_freq += div_u64(tmp64, multiplier);
diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c
index 535a0263ceeb..94525ac76d4e 100644
--- a/drivers/gpu/drm/msm/msm_drv.c
+++ b/drivers/gpu/drm/msm/msm_drv.c
@@ -457,14 +457,14 @@ static int msm_drm_init(struct device *dev, const struct drm_driver *drv)
 
 	drm_mode_config_init(ddev);
 
-	/* Bind all our sub-components: */
-	ret = component_bind_all(dev, ddev);
+	ret = msm_init_vram(ddev);
 	if (ret)
 		goto err_destroy_mdss;
 
-	ret = msm_init_vram(ddev);
+	/* Bind all our sub-components: */
+	ret = component_bind_all(dev, ddev);
 	if (ret)
-		goto err_msm_uninit;
+		goto err_destroy_mdss;
 
 	dma_set_max_seg_size(dev, UINT_MAX);
 
@@ -788,9 +788,10 @@ static int msm_ioctl_gem_info_iova(struct drm_device *dev,
 		struct drm_file *file, struct drm_gem_object *obj,
 		uint64_t *iova)
 {
+	struct msm_drm_private *priv = dev->dev_private;
 	struct msm_file_private *ctx = file->driver_priv;
 
-	if (!ctx->aspace)
+	if (!priv->gpu)
 		return -EINVAL;
 
 	/*
diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index 45e23125dfdb..f091c1e164fa 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -96,6 +96,8 @@ static struct page **get_pages(struct drm_gem_object *obj)
 {
 	struct msm_gem_object *msm_obj = to_msm_bo(obj);
 
+	WARN_ON(!msm_gem_is_locked(obj));
+
 	if (!msm_obj->pages) {
 		struct drm_device *dev = obj->dev;
 		struct page **p;
@@ -985,8 +987,9 @@ void msm_gem_free_object(struct drm_gem_object *obj)
 		/* Don't drop the pages for imported dmabuf, as they are not
 		 * ours, just free the array we allocated:
 		 */
-		if (msm_obj->pages)
-			kvfree(msm_obj->pages);
+		kvfree(msm_obj->pages);
+
+		put_iova_vmas(obj);
 
 		/* dma_buf_detach() grabs resv lock, so we need to unlock
 		 * prior to drm_prime_gem_destroy
@@ -997,11 +1000,10 @@ void msm_gem_free_object(struct drm_gem_object *obj)
 	} else {
 		msm_gem_vunmap(obj);
 		put_pages(obj);
+		put_iova_vmas(obj);
 		msm_gem_unlock(obj);
 	}
 
-	put_iova_vmas(obj);
-
 	drm_gem_object_release(obj);
 
 	kfree(msm_obj);
@@ -1115,6 +1117,8 @@ static struct drm_gem_object *_msm_gem_new(struct drm_device *dev,
 		struct msm_gem_vma *vma;
 		struct page **pages;
 
+		drm_gem_private_object_init(dev, obj, size);
+
 		msm_gem_lock(obj);
 
 		vma = add_vma(obj, NULL);
@@ -1126,9 +1130,9 @@ static struct drm_gem_object *_msm_gem_new(struct drm_device *dev,
 
 		to_msm_bo(obj)->vram_node = &vma->node;
 
-		drm_gem_private_object_init(dev, obj, size);
-
+		msm_gem_lock(obj);
 		pages = get_pages(obj);
+		msm_gem_unlock(obj);
 		if (IS_ERR(pages)) {
 			ret = PTR_ERR(pages);
 			goto fail;
diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c
index d04c349d8112..5480852bdeda 100644
--- a/drivers/gpu/drm/msm/msm_gem_submit.c
+++ b/drivers/gpu/drm/msm/msm_gem_submit.c
@@ -198,6 +198,8 @@ static int submit_lookup_cmds(struct msm_gem_submit *submit,
 		submit->cmd[i].idx  = submit_cmd.submit_idx;
 		submit->cmd[i].nr_relocs = submit_cmd.nr_relocs;
 
+		userptr = u64_to_user_ptr(submit_cmd.relocs);
+
 		sz = array_size(submit_cmd.nr_relocs,
 				sizeof(struct drm_msm_gem_submit_reloc));
 		/* check for overflow: */
diff --git a/drivers/gpu/drm/msm/msm_kms.h b/drivers/gpu/drm/msm/msm_kms.h
index d8151a89e163..4735251a394d 100644
--- a/drivers/gpu/drm/msm/msm_kms.h
+++ b/drivers/gpu/drm/msm/msm_kms.h
@@ -157,6 +157,7 @@ struct msm_kms {
 	 * from the crtc's pending_timer close to end of the frame:
 	 */
 	struct mutex commit_lock[MAX_CRTCS];
+	struct lock_class_key commit_lock_keys[MAX_CRTCS];
 	unsigned pending_crtc_mask;
 	struct msm_pending_timer pending_timers[MAX_CRTCS];
 };
@@ -166,8 +167,11 @@ static inline int msm_kms_init(struct msm_kms *kms,
 {
 	unsigned i, ret;
 
-	for (i = 0; i < ARRAY_SIZE(kms->commit_lock); i++)
-		mutex_init(&kms->commit_lock[i]);
+	for (i = 0; i < ARRAY_SIZE(kms->commit_lock); i++) {
+		lockdep_register_key(&kms->commit_lock_keys[i]);
+		__mutex_init(&kms->commit_lock[i], "&kms->commit_lock[i]",
+			     &kms->commit_lock_keys[i]);
+	}
 
 	kms->funcs = funcs;
 
diff --git a/drivers/gpu/drm/nouveau/dispnv50/Kbuild b/drivers/gpu/drm/nouveau/dispnv50/Kbuild
index 6fdddb266fb1..4488e1c061b3 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/Kbuild
+++ b/drivers/gpu/drm/nouveau/dispnv50/Kbuild
@@ -37,6 +37,7 @@ nouveau-y += dispnv50/wimmc37b.o
 nouveau-y += dispnv50/wndw.o
 nouveau-y += dispnv50/wndwc37e.o
 nouveau-y += dispnv50/wndwc57e.o
+nouveau-y += dispnv50/wndwc67e.o
 
 nouveau-y += dispnv50/base.o
 nouveau-y += dispnv50/base507c.o
diff --git a/drivers/gpu/drm/nouveau/dispnv50/base507c.c b/drivers/gpu/drm/nouveau/dispnv50/base507c.c
index 302d4e6fc52f..788db043a342 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/base507c.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/base507c.c
@@ -88,7 +88,11 @@ base507c_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
 			  NVVAL(NV507C, SET_CONVERSION, OFS, 0x64));
 	} else {
 		PUSH_MTHD(push, NV507C, SET_PROCESSING,
-			  NVDEF(NV507C, SET_PROCESSING, USE_GAIN_OFS, DISABLE));
+			  NVDEF(NV507C, SET_PROCESSING, USE_GAIN_OFS, DISABLE),
+
+					SET_CONVERSION,
+			  NVVAL(NV507C, SET_CONVERSION, GAIN, 0) |
+			  NVVAL(NV507C, SET_CONVERSION, OFS, 0));
 	}
 
 	PUSH_MTHD(push, NV507C, SURFACE_SET_OFFSET(0, 0), asyw->image.offset[0] >> 8);
diff --git a/drivers/gpu/drm/nouveau/dispnv50/base827c.c b/drivers/gpu/drm/nouveau/dispnv50/base827c.c
index 18d34096f125..093d4ba6910e 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/base827c.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/base827c.c
@@ -49,7 +49,11 @@ base827c_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
 			  NVVAL(NV827C, SET_CONVERSION, OFS, 0x64));
 	} else {
 		PUSH_MTHD(push, NV827C, SET_PROCESSING,
-			  NVDEF(NV827C, SET_PROCESSING, USE_GAIN_OFS, DISABLE));
+			  NVDEF(NV827C, SET_PROCESSING, USE_GAIN_OFS, DISABLE),
+
+					SET_CONVERSION,
+			  NVVAL(NV827C, SET_CONVERSION, GAIN, 0) |
+			  NVVAL(NV827C, SET_CONVERSION, OFS, 0));
 	}
 
 	PUSH_MTHD(push, NV827C, SURFACE_SET_OFFSET(0, 0), asyw->image.offset[0] >> 8,
diff --git a/drivers/gpu/drm/nouveau/dispnv50/core.c b/drivers/gpu/drm/nouveau/dispnv50/core.c
index 27ea3f34706d..abefc2343443 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/core.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/core.c
@@ -42,6 +42,7 @@ nv50_core_new(struct nouveau_drm *drm, struct nv50_core **pcore)
 		int version;
 		int (*new)(struct nouveau_drm *, s32, struct nv50_core **);
 	} cores[] = {
+		{ GA102_DISP_CORE_CHANNEL_DMA, 0, corec57d_new },
 		{ TU102_DISP_CORE_CHANNEL_DMA, 0, corec57d_new },
 		{ GV100_DISP_CORE_CHANNEL_DMA, 0, corec37d_new },
 		{ GP102_DISP_CORE_CHANNEL_DMA, 0, core917d_new },
diff --git a/drivers/gpu/drm/nouveau/dispnv50/core507d.c b/drivers/gpu/drm/nouveau/dispnv50/core507d.c
index e6f16a7750f0..1a1d806e0b01 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/core507d.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/core507d.c
@@ -36,7 +36,7 @@ core507d_update(struct nv50_core *core, u32 *interlock, bool ntfy)
 	struct nvif_push *push = core->chan.push;
 	int ret;
 
-	if ((ret = PUSH_WAIT(push, 5)))
+	if ((ret = PUSH_WAIT(push, (ntfy ? 2 : 0) + 3)))
 		return ret;
 
 	if (ntfy) {
diff --git a/drivers/gpu/drm/nouveau/dispnv50/corec37d.c b/drivers/gpu/drm/nouveau/dispnv50/corec37d.c
index 9035d3ab062c..42f877f2ced2 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/corec37d.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/corec37d.c
@@ -54,7 +54,7 @@ corec37d_update(struct nv50_core *core, u32 *interlock, bool ntfy)
 	struct nvif_push *push = core->chan.push;
 	int ret;
 
-	if ((ret = PUSH_WAIT(push, 9)))
+	if ((ret = PUSH_WAIT(push, (ntfy ? 2 * 2 : 0) + 5)))
 		return ret;
 
 	if (ntfy) {
diff --git a/drivers/gpu/drm/nouveau/dispnv50/curs.c b/drivers/gpu/drm/nouveau/dispnv50/curs.c
index 121c24a18f11..31d8b2e4791d 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/curs.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/curs.c
@@ -31,6 +31,7 @@ nv50_curs_new(struct nouveau_drm *drm, int head, struct nv50_wndw **pwndw)
 		int version;
 		int (*new)(struct nouveau_drm *, int, s32, struct nv50_wndw **);
 	} curses[] = {
+		{ GA102_DISP_CURSOR, 0, cursc37a_new },
 		{ TU102_DISP_CURSOR, 0, cursc37a_new },
 		{ GV100_DISP_CURSOR, 0, cursc37a_new },
 		{ GK104_DISP_CURSOR, 0, curs907a_new },
diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c
index 33fff388dd83..196612addfd6 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c
@@ -220,9 +220,13 @@ nv50_dmac_wait(struct nvif_push *push, u32 size)
 	return 0;
 }
 
+MODULE_PARM_DESC(kms_vram_pushbuf, "Place EVO/NVD push buffers in VRAM (default: auto)");
+static int nv50_dmac_vram_pushbuf = -1;
+module_param_named(kms_vram_pushbuf, nv50_dmac_vram_pushbuf, int, 0400);
+
 int
 nv50_dmac_create(struct nvif_device *device, struct nvif_object *disp,
-		 const s32 *oclass, u8 head, void *data, u32 size, u64 syncbuf,
+		 const s32 *oclass, u8 head, void *data, u32 size, s64 syncbuf,
 		 struct nv50_dmac *dmac)
 {
 	struct nouveau_cli *cli = (void *)device->object.client;
@@ -241,7 +245,8 @@ nv50_dmac_create(struct nvif_device *device, struct nvif_object *disp,
 	 *
 	 * This appears to match NVIDIA's behaviour on Pascal.
 	 */
-	if (device->info.family == NV_DEVICE_INFO_V0_PASCAL)
+	if ((nv50_dmac_vram_pushbuf > 0) ||
+	    (nv50_dmac_vram_pushbuf < 0 && device->info.family == NV_DEVICE_INFO_V0_PASCAL))
 		type |= NVIF_MEM_VRAM;
 
 	ret = nvif_mem_ctor_map(&cli->mmu, "kmsChanPush", type, 0x1000,
@@ -271,7 +276,7 @@ nv50_dmac_create(struct nvif_device *device, struct nvif_object *disp,
 	if (ret)
 		return ret;
 
-	if (!syncbuf)
+	if (syncbuf < 0)
 		return 0;
 
 	ret = nvif_object_ctor(&dmac->base.user, "kmsSyncCtxDma", NV50_DISP_HANDLE_SYNCBUF,
@@ -305,6 +310,14 @@ nv50_dmac_create(struct nvif_device *device, struct nvif_object *disp,
  * Output path helpers
  *****************************************************************************/
 static void
+nv50_outp_dump_caps(struct nouveau_drm *drm,
+		    struct nouveau_encoder *outp)
+{
+	NV_DEBUG(drm, "%s caps: dp_interlace=%d\n",
+		 outp->base.base.name, outp->caps.dp_interlace);
+}
+
+static void
 nv50_outp_release(struct nouveau_encoder *nv_encoder)
 {
 	struct nv50_disp *disp = nv50_disp(nv_encoder->base.base.dev);
@@ -419,8 +432,7 @@ nv50_outp_atomic_check(struct drm_encoder *encoder,
 }
 
 struct nouveau_connector *
-nv50_outp_get_new_connector(struct nouveau_encoder *outp,
-			    struct drm_atomic_state *state)
+nv50_outp_get_new_connector(struct drm_atomic_state *state, struct nouveau_encoder *outp)
 {
 	struct drm_connector *connector;
 	struct drm_connector_state *connector_state;
@@ -436,8 +448,7 @@ nv50_outp_get_new_connector(struct nouveau_encoder *outp,
 }
 
 struct nouveau_connector *
-nv50_outp_get_old_connector(struct nouveau_encoder *outp,
-			    struct drm_atomic_state *state)
+nv50_outp_get_old_connector(struct drm_atomic_state *state, struct nouveau_encoder *outp)
 {
 	struct drm_connector *connector;
 	struct drm_connector_state *connector_state;
@@ -452,27 +463,44 @@ nv50_outp_get_old_connector(struct nouveau_encoder *outp,
 	return NULL;
 }
 
+static struct nouveau_crtc *
+nv50_outp_get_new_crtc(const struct drm_atomic_state *state, const struct nouveau_encoder *outp)
+{
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *crtc_state;
+	const u32 mask = drm_encoder_mask(&outp->base.base);
+	int i;
+
+	for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
+		if (crtc_state->encoder_mask & mask)
+			return nouveau_crtc(crtc);
+	}
+
+	return NULL;
+}
+
 /******************************************************************************
  * DAC
  *****************************************************************************/
 static void
-nv50_dac_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_dac_atomic_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_core *core = nv50_disp(encoder->dev)->core;
 	const u32 ctrl = NVDEF(NV507D, DAC_SET_CONTROL, OWNER, NONE);
-	if (nv_encoder->crtc)
-		core->func->dac->ctrl(core, nv_encoder->or, ctrl, NULL);
+
+	core->func->dac->ctrl(core, nv_encoder->or, ctrl, NULL);
 	nv_encoder->crtc = NULL;
 	nv50_outp_release(nv_encoder);
 }
 
 static void
-nv50_dac_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_dac_atomic_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nv50_head_atom *asyh = nv50_head_atom(nv_crtc->base.state);
+	struct nouveau_crtc *nv_crtc = nv50_outp_get_new_crtc(state, nv_encoder);
+	struct nv50_head_atom *asyh =
+		nv50_head_atom(drm_atomic_get_new_crtc_state(state, &nv_crtc->base));
 	struct nv50_core *core = nv50_disp(encoder->dev)->core;
 	u32 ctrl = 0;
 
@@ -493,7 +521,7 @@ nv50_dac_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 	core->func->dac->ctrl(core, nv_encoder->or, ctrl, asyh);
 	asyh->or.depth = 0;
 
-	nv_encoder->crtc = encoder->crtc;
+	nv_encoder->crtc = &nv_crtc->base;
 }
 
 static enum drm_connector_status
@@ -526,8 +554,8 @@ nv50_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector)
 static const struct drm_encoder_helper_funcs
 nv50_dac_help = {
 	.atomic_check = nv50_outp_atomic_check,
-	.atomic_enable = nv50_dac_enable,
-	.atomic_disable = nv50_dac_disable,
+	.atomic_enable = nv50_dac_atomic_enable,
+	.atomic_disable = nv50_dac_atomic_disable,
 	.detect = nv50_dac_detect
 };
 
@@ -593,34 +621,27 @@ nv50_audio_component_get_eld(struct device *kdev, int port, int dev_id,
 	struct nouveau_drm *drm = nouveau_drm(drm_dev);
 	struct drm_encoder *encoder;
 	struct nouveau_encoder *nv_encoder;
-	struct drm_connector *connector;
 	struct nouveau_crtc *nv_crtc;
-	struct drm_connector_list_iter conn_iter;
 	int ret = 0;
 
 	*enabled = false;
 
+	mutex_lock(&drm->audio.lock);
+
 	drm_for_each_encoder(encoder, drm->dev) {
 		struct nouveau_connector *nv_connector = NULL;
 
+		if (encoder->encoder_type == DRM_MODE_ENCODER_DPMST)
+			continue; /* TODO */
+
 		nv_encoder = nouveau_encoder(encoder);
+		nv_connector = nouveau_connector(nv_encoder->audio.connector);
+		nv_crtc = nouveau_crtc(nv_encoder->crtc);
 
-		drm_connector_list_iter_begin(drm_dev, &conn_iter);
-		drm_for_each_connector_iter(connector, &conn_iter) {
-			if (connector->state->best_encoder == encoder) {
-				nv_connector = nouveau_connector(connector);
-				break;
-			}
-		}
-		drm_connector_list_iter_end(&conn_iter);
-		if (!nv_connector)
+		if (!nv_crtc || nv_encoder->or != port || nv_crtc->index != dev_id)
 			continue;
 
-		nv_crtc = nouveau_crtc(encoder->crtc);
-		if (!nv_crtc || nv_encoder->or != port ||
-		    nv_crtc->index != dev_id)
-			continue;
-		*enabled = nv_encoder->audio;
+		*enabled = nv_encoder->audio.enabled;
 		if (*enabled) {
 			ret = drm_eld_size(nv_connector->base.eld);
 			memcpy(buf, nv_connector->base.eld,
@@ -629,6 +650,8 @@ nv50_audio_component_get_eld(struct device *kdev, int port, int dev_id,
 		break;
 	}
 
+	mutex_unlock(&drm->audio.lock);
+
 	return ret;
 }
 
@@ -678,17 +701,22 @@ static const struct component_ops nv50_audio_component_bind_ops = {
 static void
 nv50_audio_component_init(struct nouveau_drm *drm)
 {
-	if (!component_add(drm->dev->dev, &nv50_audio_component_bind_ops))
-		drm->audio.component_registered = true;
+	if (component_add(drm->dev->dev, &nv50_audio_component_bind_ops))
+		return;
+
+	drm->audio.component_registered = true;
+	mutex_init(&drm->audio.lock);
 }
 
 static void
 nv50_audio_component_fini(struct nouveau_drm *drm)
 {
-	if (drm->audio.component_registered) {
-		component_del(drm->dev->dev, &nv50_audio_component_bind_ops);
-		drm->audio.component_registered = false;
-	}
+	if (!drm->audio.component_registered)
+		return;
+
+	component_del(drm->dev->dev, &nv50_audio_component_bind_ops);
+	drm->audio.component_registered = false;
+	mutex_destroy(&drm->audio.lock);
 }
 
 /******************************************************************************
@@ -711,24 +739,25 @@ nv50_audio_disable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 				(0x0100 << nv_crtc->index),
 	};
 
-	if (!nv_encoder->audio)
-		return;
-
-	nv_encoder->audio = false;
-	nvif_mthd(&disp->disp->object, 0, &args, sizeof(args));
+	mutex_lock(&drm->audio.lock);
+	if (nv_encoder->audio.enabled) {
+		nv_encoder->audio.enabled = false;
+		nv_encoder->audio.connector = NULL;
+		nvif_mthd(&disp->disp->object, 0, &args, sizeof(args));
+	}
+	mutex_unlock(&drm->audio.lock);
 
 	nv50_audio_component_eld_notify(drm->audio.component, nv_encoder->or,
 					nv_crtc->index);
 }
 
 static void
-nv50_audio_enable(struct drm_encoder *encoder, struct drm_atomic_state *state,
+nv50_audio_enable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc,
+		  struct nouveau_connector *nv_connector, struct drm_atomic_state *state,
 		  struct drm_display_mode *mode)
 {
 	struct nouveau_drm *drm = nouveau_drm(encoder->dev);
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector;
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
 	struct __packed {
 		struct {
@@ -744,15 +773,19 @@ nv50_audio_enable(struct drm_encoder *encoder, struct drm_atomic_state *state,
 				     (0x0100 << nv_crtc->index),
 	};
 
-	nv_connector = nv50_outp_get_new_connector(nv_encoder, state);
 	if (!drm_detect_monitor_audio(nv_connector->edid))
 		return;
 
+	mutex_lock(&drm->audio.lock);
+
 	memcpy(args.data, nv_connector->base.eld, sizeof(args.data));
 
 	nvif_mthd(&disp->disp->object, 0, &args,
 		  sizeof(args.base) + drm_eld_size(args.data));
-	nv_encoder->audio = true;
+	nv_encoder->audio.enabled = true;
+	nv_encoder->audio.connector = &nv_connector->base;
+
+	mutex_unlock(&drm->audio.lock);
 
 	nv50_audio_component_eld_notify(drm->audio.component, nv_encoder->or,
 					nv_crtc->index);
@@ -781,12 +814,12 @@ nv50_hdmi_disable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 }
 
 static void
-nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_atomic_state *state,
+nv50_hdmi_enable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc,
+		 struct nouveau_connector *nv_connector, struct drm_atomic_state *state,
 		 struct drm_display_mode *mode)
 {
 	struct nouveau_drm *drm = nouveau_drm(encoder->dev);
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
 	struct {
 		struct nv50_disp_mthd_v1 base;
@@ -801,7 +834,6 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_atomic_state *state,
 		.pwr.state = 1,
 		.pwr.rekey = 56, /* binary driver, and tegra, constant */
 	};
-	struct nouveau_connector *nv_connector;
 	struct drm_hdmi_info *hdmi;
 	u32 max_ac_packet;
 	union hdmi_infoframe avi_frame;
@@ -811,7 +843,6 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_atomic_state *state,
 	int ret;
 	int size;
 
-	nv_connector = nv50_outp_get_new_connector(nv_encoder, state);
 	if (!drm_detect_hdmi_monitor(nv_connector->edid))
 		return;
 
@@ -857,7 +888,7 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_atomic_state *state,
 		+ args.pwr.vendor_infoframe_length;
 	nvif_mthd(&disp->disp->object, 0, &args, size);
 
-	nv50_audio_enable(encoder, state, mode);
+	nv50_audio_enable(encoder, nv_crtc, nv_connector, state, mode);
 
 	/* If SCDC is supported by the downstream monitor, update
 	 * divider / scrambling settings to what we programmed above.
@@ -898,6 +929,7 @@ struct nv50_mstc {
 struct nv50_msto {
 	struct drm_encoder encoder;
 
+	/* head is statically assigned on msto creation */
 	struct nv50_head *head;
 	struct nv50_mstc *mstc;
 	bool disabled;
@@ -1056,11 +1088,12 @@ nv50_dp_bpc_to_depth(unsigned int bpc)
 }
 
 static void
-nv50_msto_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_msto_atomic_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
-	struct nv50_head *head = nv50_head(encoder->crtc);
-	struct nv50_head_atom *armh = nv50_head_atom(head->base.base.state);
 	struct nv50_msto *msto = nv50_msto(encoder);
+	struct nv50_head *head = msto->head;
+	struct nv50_head_atom *asyh =
+		nv50_head_atom(drm_atomic_get_new_crtc_state(state, &head->base.base));
 	struct nv50_mstc *mstc = NULL;
 	struct nv50_mstm *mstm = NULL;
 	struct drm_connector *connector;
@@ -1081,8 +1114,7 @@ nv50_msto_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 	if (WARN_ON(!mstc))
 		return;
 
-	r = drm_dp_mst_allocate_vcpi(&mstm->mgr, mstc->port, armh->dp.pbn,
-				     armh->dp.tu);
+	r = drm_dp_mst_allocate_vcpi(&mstm->mgr, mstc->port, asyh->dp.pbn, asyh->dp.tu);
 	if (!r)
 		DRM_DEBUG_KMS("Failed to allocate VCPI\n");
 
@@ -1094,15 +1126,15 @@ nv50_msto_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 	else
 		proto = NV917D_SOR_SET_CONTROL_PROTOCOL_DP_B;
 
-	mstm->outp->update(mstm->outp, head->base.index, armh, proto,
-			   nv50_dp_bpc_to_depth(armh->or.bpc));
+	mstm->outp->update(mstm->outp, head->base.index, asyh, proto,
+			   nv50_dp_bpc_to_depth(asyh->or.bpc));
 
 	msto->mstc = mstc;
 	mstm->modified = true;
 }
 
 static void
-nv50_msto_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_msto_atomic_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nv50_msto *msto = nv50_msto(encoder);
 	struct nv50_mstc *mstc = msto->mstc;
@@ -1119,8 +1151,8 @@ nv50_msto_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 
 static const struct drm_encoder_helper_funcs
 nv50_msto_help = {
-	.atomic_disable = nv50_msto_disable,
-	.atomic_enable = nv50_msto_enable,
+	.atomic_disable = nv50_msto_atomic_disable,
+	.atomic_enable = nv50_msto_atomic_enable,
 	.atomic_check = nv50_msto_atomic_check,
 };
 
@@ -1616,43 +1648,38 @@ nv50_sor_update(struct nouveau_encoder *nv_encoder, u8 head,
 }
 
 static void
-nv50_sor_disable(struct drm_encoder *encoder,
-		 struct drm_atomic_state *state)
+nv50_sor_atomic_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(nv_encoder->crtc);
-	struct nouveau_connector *nv_connector =
-		nv50_outp_get_old_connector(nv_encoder, state);
-
-	nv_encoder->crtc = NULL;
-
-	if (nv_crtc) {
-		struct drm_dp_aux *aux = &nv_connector->aux;
-		u8 pwr;
+	struct nouveau_connector *nv_connector = nv50_outp_get_old_connector(state, nv_encoder);
+	struct drm_dp_aux *aux = &nv_connector->aux;
+	u8 pwr;
 
-		if (nv_encoder->dcb->type == DCB_OUTPUT_DP) {
-			int ret = drm_dp_dpcd_readb(aux, DP_SET_POWER, &pwr);
+	if (nv_encoder->dcb->type == DCB_OUTPUT_DP) {
+		int ret = drm_dp_dpcd_readb(aux, DP_SET_POWER, &pwr);
 
-			if (ret == 0) {
-				pwr &= ~DP_SET_POWER_MASK;
-				pwr |=  DP_SET_POWER_D3;
-				drm_dp_dpcd_writeb(aux, DP_SET_POWER, pwr);
-			}
+		if (ret == 0) {
+			pwr &= ~DP_SET_POWER_MASK;
+			pwr |=  DP_SET_POWER_D3;
+			drm_dp_dpcd_writeb(aux, DP_SET_POWER, pwr);
 		}
-
-		nv_encoder->update(nv_encoder, nv_crtc->index, NULL, 0, 0);
-		nv50_audio_disable(encoder, nv_crtc);
-		nv50_hdmi_disable(&nv_encoder->base.base, nv_crtc);
-		nv50_outp_release(nv_encoder);
 	}
+
+	nv_encoder->update(nv_encoder, nv_crtc->index, NULL, 0, 0);
+	nv50_audio_disable(encoder, nv_crtc);
+	nv50_hdmi_disable(&nv_encoder->base.base, nv_crtc);
+	nv50_outp_release(nv_encoder);
+	nv_encoder->crtc = NULL;
 }
 
 static void
-nv50_sor_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_sor_atomic_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nv50_head_atom *asyh = nv50_head_atom(nv_crtc->base.state);
+	struct nouveau_crtc *nv_crtc = nv50_outp_get_new_crtc(state, nv_encoder);
+	struct nv50_head_atom *asyh =
+		nv50_head_atom(drm_atomic_get_new_crtc_state(state, &nv_crtc->base));
 	struct drm_display_mode *mode = &asyh->state.adjusted_mode;
 	struct {
 		struct nv50_disp_mthd_v1 base;
@@ -1672,8 +1699,8 @@ nv50_sor_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 	u8 proto = NV507D_SOR_SET_CONTROL_PROTOCOL_CUSTOM;
 	u8 depth = NV837D_SOR_SET_CONTROL_PIXEL_DEPTH_DEFAULT;
 
-	nv_connector = nv50_outp_get_new_connector(nv_encoder, state);
-	nv_encoder->crtc = encoder->crtc;
+	nv_connector = nv50_outp_get_new_connector(state, nv_encoder);
+	nv_encoder->crtc = &nv_crtc->base;
 
 	if ((disp->disp->object.oclass == GT214_DISP ||
 	     disp->disp->object.oclass >= GF110_DISP) &&
@@ -1699,7 +1726,7 @@ nv50_sor_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 			proto = NV507D_SOR_SET_CONTROL_PROTOCOL_SINGLE_TMDS_B;
 		}
 
-		nv50_hdmi_enable(&nv_encoder->base.base, state, mode);
+		nv50_hdmi_enable(&nv_encoder->base.base, nv_crtc, nv_connector, state, mode);
 		break;
 	case DCB_OUTPUT_LVDS:
 		proto = NV507D_SOR_SET_CONTROL_PROTOCOL_LVDS_CUSTOM;
@@ -1740,7 +1767,7 @@ nv50_sor_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 		else
 			proto = NV887D_SOR_SET_CONTROL_PROTOCOL_DP_B;
 
-		nv50_audio_enable(encoder, state, mode);
+		nv50_audio_enable(encoder, nv_crtc, nv_connector, state, mode);
 		break;
 	default:
 		BUG();
@@ -1753,8 +1780,8 @@ nv50_sor_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 static const struct drm_encoder_helper_funcs
 nv50_sor_help = {
 	.atomic_check = nv50_outp_atomic_check,
-	.atomic_enable = nv50_sor_enable,
-	.atomic_disable = nv50_sor_disable,
+	.atomic_enable = nv50_sor_atomic_enable,
+	.atomic_disable = nv50_sor_atomic_disable,
 };
 
 static void
@@ -1821,6 +1848,7 @@ nv50_sor_create(struct drm_connector *connector, struct dcb_output *dcbe)
 	drm_connector_attach_encoder(connector, encoder);
 
 	disp->core->func->sor->get_caps(disp, nv_encoder, ffs(dcbe->or) - 1);
+	nv50_outp_dump_caps(drm, nv_encoder);
 
 	if (dcbe->type == DCB_OUTPUT_DP) {
 		struct nvkm_i2c_aux *aux =
@@ -1875,23 +1903,24 @@ nv50_pior_atomic_check(struct drm_encoder *encoder,
 }
 
 static void
-nv50_pior_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_pior_atomic_disable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_core *core = nv50_disp(encoder->dev)->core;
 	const u32 ctrl = NVDEF(NV507D, PIOR_SET_CONTROL, OWNER, NONE);
-	if (nv_encoder->crtc)
-		core->func->pior->ctrl(core, nv_encoder->or, ctrl, NULL);
+
+	core->func->pior->ctrl(core, nv_encoder->or, ctrl, NULL);
 	nv_encoder->crtc = NULL;
 	nv50_outp_release(nv_encoder);
 }
 
 static void
-nv50_pior_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
+nv50_pior_atomic_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nv50_head_atom *asyh = nv50_head_atom(nv_crtc->base.state);
+	struct nouveau_crtc *nv_crtc = nv50_outp_get_new_crtc(state, nv_encoder);
+	struct nv50_head_atom *asyh =
+		nv50_head_atom(drm_atomic_get_new_crtc_state(state, &nv_crtc->base));
 	struct nv50_core *core = nv50_disp(encoder->dev)->core;
 	u32 ctrl = 0;
 
@@ -1929,8 +1958,8 @@ nv50_pior_enable(struct drm_encoder *encoder, struct drm_atomic_state *state)
 static const struct drm_encoder_helper_funcs
 nv50_pior_help = {
 	.atomic_check = nv50_pior_atomic_check,
-	.atomic_enable = nv50_pior_enable,
-	.atomic_disable = nv50_pior_disable,
+	.atomic_enable = nv50_pior_atomic_enable,
+	.atomic_disable = nv50_pior_atomic_disable,
 };
 
 static void
@@ -1991,6 +2020,7 @@ nv50_pior_create(struct drm_connector *connector, struct dcb_output *dcbe)
 	drm_connector_attach_encoder(connector, encoder);
 
 	disp->core->func->pior->get_caps(disp, nv_encoder, ffs(dcbe->or) - 1);
+	nv50_outp_dump_caps(drm, nv_encoder);
 
 	return 0;
 }
@@ -2663,6 +2693,14 @@ nv50_display_create(struct drm_device *dev)
 	else
 		nouveau_display(dev)->format_modifiers = disp50xx_modifiers;
 
+	if (disp->disp->object.oclass >= GK104_DISP) {
+		dev->mode_config.cursor_width = 256;
+		dev->mode_config.cursor_height = 256;
+	} else {
+		dev->mode_config.cursor_width = 64;
+		dev->mode_config.cursor_height = 64;
+	}
+
 	/* create crtc objects to represent the hw heads */
 	if (disp->disp->object.oclass >= GV100_DISP)
 		crtcs = nvif_rd32(&device->object, 0x610060) & 0xff;
diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.h b/drivers/gpu/drm/nouveau/dispnv50/disp.h
index 92bddc083617..38dec11e7dda 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/disp.h
+++ b/drivers/gpu/drm/nouveau/dispnv50/disp.h
@@ -95,7 +95,7 @@ struct nv50_outp_atom {
 
 int nv50_dmac_create(struct nvif_device *device, struct nvif_object *disp,
 		     const s32 *oclass, u8 head, void *data, u32 size,
-		     u64 syncbuf, struct nv50_dmac *dmac);
+		     s64 syncbuf, struct nv50_dmac *dmac);
 void nv50_dmac_destroy(struct nv50_dmac *);
 
 /*
diff --git a/drivers/gpu/drm/nouveau/dispnv50/head907d.c b/drivers/gpu/drm/nouveau/dispnv50/head907d.c
index 8f860e9c5224..85648d790743 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/head907d.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/head907d.c
@@ -322,7 +322,7 @@ head907d_mode(struct nv50_head *head, struct nv50_head_atom *asyh)
 	const int i = head->base.index;
 	int ret;
 
-	if ((ret = PUSH_WAIT(push, 14)))
+	if ((ret = PUSH_WAIT(push, 13)))
 		return ret;
 
 	PUSH_MTHD(push, NV907D, HEAD_SET_OVERSCAN_COLOR(i),
@@ -353,14 +353,7 @@ head907d_mode(struct nv50_head *head, struct nv50_head_atom *asyh)
 	PUSH_MTHD(push, NV907D, HEAD_SET_DEFAULT_BASE_COLOR(i),
 		  NVVAL(NV907D, HEAD_SET_DEFAULT_BASE_COLOR, RED, 0) |
 		  NVVAL(NV907D, HEAD_SET_DEFAULT_BASE_COLOR, GREEN, 0) |
-		  NVVAL(NV907D, HEAD_SET_DEFAULT_BASE_COLOR, BLUE, 0),
-
-				HEAD_SET_CRC_CONTROL(i),
-		  NVDEF(NV907D, HEAD_SET_CRC_CONTROL, CONTROLLING_CHANNEL, CORE) |
-		  NVDEF(NV907D, HEAD_SET_CRC_CONTROL, EXPECT_BUFFER_COLLAPSE, FALSE) |
-		  NVDEF(NV907D, HEAD_SET_CRC_CONTROL, TIMESTAMP_MODE, FALSE) |
-		  NVDEF(NV907D, HEAD_SET_CRC_CONTROL, PRIMARY_OUTPUT, NONE) |
-		  NVDEF(NV907D, HEAD_SET_CRC_CONTROL, SECONDARY_OUTPUT, NONE));
+		  NVVAL(NV907D, HEAD_SET_DEFAULT_BASE_COLOR, BLUE, 0));
 
 	PUSH_MTHD(push, NV907D, HEAD_SET_PIXEL_CLOCK_FREQUENCY(i),
 		  NVVAL(NV907D, HEAD_SET_PIXEL_CLOCK_FREQUENCY, HERTZ, m->clock * 1000) |
diff --git a/drivers/gpu/drm/nouveau/dispnv50/head917d.c b/drivers/gpu/drm/nouveau/dispnv50/head917d.c
index a5d827403660..ea9f8667305e 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/head917d.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/head917d.c
@@ -22,6 +22,7 @@
 #include "head.h"
 #include "core.h"
 
+#include "nvif/push.h"
 #include <nvif/push507c.h>
 
 #include <nvhw/class/cl917d.h>
@@ -73,6 +74,31 @@ head917d_base(struct nv50_head *head, struct nv50_head_atom *asyh)
 	return 0;
 }
 
+static int
+head917d_curs_set(struct nv50_head *head, struct nv50_head_atom *asyh)
+{
+	struct nvif_push *push = nv50_disp(head->base.base.dev)->core->chan.push;
+	const int i = head->base.index;
+	int ret;
+
+	ret = PUSH_WAIT(push, 5);
+	if (ret)
+		return ret;
+
+	PUSH_MTHD(push, NV917D, HEAD_SET_CONTROL_CURSOR(i),
+		  NVDEF(NV917D, HEAD_SET_CONTROL_CURSOR, ENABLE, ENABLE) |
+		  NVVAL(NV917D, HEAD_SET_CONTROL_CURSOR, FORMAT, asyh->curs.format) |
+		  NVVAL(NV917D, HEAD_SET_CONTROL_CURSOR, SIZE, asyh->curs.layout) |
+		  NVVAL(NV917D, HEAD_SET_CONTROL_CURSOR, HOT_SPOT_X, 0) |
+		  NVVAL(NV917D, HEAD_SET_CONTROL_CURSOR, HOT_SPOT_Y, 0) |
+		  NVDEF(NV917D, HEAD_SET_CONTROL_CURSOR, COMPOSITION, ALPHA_BLEND),
+
+				HEAD_SET_OFFSET_CURSOR(i), asyh->curs.offset >> 8);
+
+	PUSH_MTHD(push, NV917D, HEAD_SET_CONTEXT_DMA_CURSOR(i), asyh->curs.handle);
+	return 0;
+}
+
 int
 head917d_curs_layout(struct nv50_head *head, struct nv50_wndw_atom *asyw,
 		     struct nv50_head_atom *asyh)
@@ -101,7 +127,7 @@ head917d = {
 	.core_clr = head907d_core_clr,
 	.curs_layout = head917d_curs_layout,
 	.curs_format = head507d_curs_format,
-	.curs_set = head907d_curs_set,
+	.curs_set = head917d_curs_set,
 	.curs_clr = head907d_curs_clr,
 	.base = head917d_base,
 	.ovly = head907d_ovly,
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wimm.c b/drivers/gpu/drm/nouveau/dispnv50/wimm.c
index a1ac153d5e98..566fbddfc8d7 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wimm.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/wimm.c
@@ -31,6 +31,7 @@ nv50_wimm_init(struct nouveau_drm *drm, struct nv50_wndw *wndw)
 		int version;
 		int (*init)(struct nouveau_drm *, s32, struct nv50_wndw *);
 	} wimms[] = {
+		{ GA102_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init },
 		{ TU102_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init },
 		{ GV100_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init },
 		{}
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wimmc37b.c b/drivers/gpu/drm/nouveau/dispnv50/wimmc37b.c
index 685b70871324..b390029c69ec 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wimmc37b.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/wimmc37b.c
@@ -76,7 +76,7 @@ wimmc37b_init_(const struct nv50_wimm_func *func, struct nouveau_drm *drm,
 	int ret;
 
 	ret = nv50_dmac_create(&drm->client.device, &disp->disp->object,
-			       &oclass, 0, &args, sizeof(args), 0,
+			       &oclass, 0, &args, sizeof(args), -1,
 			       &wndw->wimm);
 	if (ret) {
 		NV_ERROR(drm, "wimm%04x allocation failed: %d\n", oclass, ret);
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.c b/drivers/gpu/drm/nouveau/dispnv50/wndw.c
index 1382ccdfe7e3..0cb1f9d848d3 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wndw.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.c
@@ -706,6 +706,11 @@ nv50_wndw_init(struct nv50_wndw *wndw)
 	nvif_notify_get(&wndw->notify);
 }
 
+static const u64 nv50_cursor_format_modifiers[] = {
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_INVALID,
+};
+
 int
 nv50_wndw_new_(const struct nv50_wndw_func *func, struct drm_device *dev,
 	       enum drm_plane_type type, const char *name, int index,
@@ -717,6 +722,7 @@ nv50_wndw_new_(const struct nv50_wndw_func *func, struct drm_device *dev,
 	struct nvif_mmu *mmu = &drm->client.mmu;
 	struct nv50_disp *disp = nv50_disp(dev);
 	struct nv50_wndw *wndw;
+	const u64 *format_modifiers;
 	int nformat;
 	int ret;
 
@@ -732,10 +738,13 @@ nv50_wndw_new_(const struct nv50_wndw_func *func, struct drm_device *dev,
 
 	for (nformat = 0; format[nformat]; nformat++);
 
-	ret = drm_universal_plane_init(dev, &wndw->plane, heads, &nv50_wndw,
-				       format, nformat,
-				       nouveau_display(dev)->format_modifiers,
-				       type, "%s-%d", name, index);
+	if (type == DRM_PLANE_TYPE_CURSOR)
+		format_modifiers = nv50_cursor_format_modifiers;
+	else
+		format_modifiers = nouveau_display(dev)->format_modifiers;
+
+	ret = drm_universal_plane_init(dev, &wndw->plane, heads, &nv50_wndw, format, nformat,
+				       format_modifiers, type, "%s-%d", name, index);
 	if (ret) {
 		kfree(*pwndw);
 		*pwndw = NULL;
@@ -788,6 +797,7 @@ nv50_wndw_new(struct nouveau_drm *drm, enum drm_plane_type type, int index,
 		int (*new)(struct nouveau_drm *, enum drm_plane_type,
 			   int, s32, struct nv50_wndw **);
 	} wndws[] = {
+		{ GA102_DISP_WINDOW_CHANNEL_DMA, 0, wndwc67e_new },
 		{ TU102_DISP_WINDOW_CHANNEL_DMA, 0, wndwc57e_new },
 		{ GV100_DISP_WINDOW_CHANNEL_DMA, 0, wndwc37e_new },
 		{}
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.h b/drivers/gpu/drm/nouveau/dispnv50/wndw.h
index 3278e2880034..f4e0c5080034 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wndw.h
+++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.h
@@ -129,6 +129,14 @@ int wndwc37e_update(struct nv50_wndw *, u32 *);
 
 int wndwc57e_new(struct nouveau_drm *, enum drm_plane_type, int, s32,
 		 struct nv50_wndw **);
+bool wndwc57e_ilut(struct nv50_wndw *, struct nv50_wndw_atom *, int);
+int wndwc57e_ilut_set(struct nv50_wndw *, struct nv50_wndw_atom *);
+int wndwc57e_ilut_clr(struct nv50_wndw *);
+int wndwc57e_csc_set(struct nv50_wndw *, struct nv50_wndw_atom *);
+int wndwc57e_csc_clr(struct nv50_wndw *);
+
+int wndwc67e_new(struct nouveau_drm *, enum drm_plane_type, int, s32,
+		 struct nv50_wndw **);
 
 int nv50_wndw_new(struct nouveau_drm *, enum drm_plane_type, int index,
 		  struct nv50_wndw **);
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c b/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c
index 429be0bb0222..abdd3bb658b3 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c
@@ -80,7 +80,7 @@ wndwc57e_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
 	return 0;
 }
 
-static int
+int
 wndwc57e_csc_clr(struct nv50_wndw *wndw)
 {
 	struct nvif_push *push = wndw->wndw.push;
@@ -98,7 +98,7 @@ wndwc57e_csc_clr(struct nv50_wndw *wndw)
 	return 0;
 }
 
-static int
+int
 wndwc57e_csc_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
 {
 	struct nvif_push *push = wndw->wndw.push;
@@ -111,7 +111,7 @@ wndwc57e_csc_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
 	return 0;
 }
 
-static int
+int
 wndwc57e_ilut_clr(struct nv50_wndw *wndw)
 {
 	struct nvif_push *push = wndw->wndw.push;
@@ -124,7 +124,7 @@ wndwc57e_ilut_clr(struct nv50_wndw *wndw)
 	return 0;
 }
 
-static int
+int
 wndwc57e_ilut_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
 {
 	struct nvif_push *push = wndw->wndw.push;
@@ -179,7 +179,7 @@ wndwc57e_ilut_load(struct drm_color_lut *in, int size, void __iomem *mem)
 	writew(readw(mem - 4), mem + 4);
 }
 
-static bool
+bool
 wndwc57e_ilut(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw, int size)
 {
 	if (size = size ? size : 1024, size != 256 && size != 1024)
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndwc67e.c b/drivers/gpu/drm/nouveau/dispnv50/wndwc67e.c
new file mode 100644
index 000000000000..7a370fa1df20
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/dispnv50/wndwc67e.c
@@ -0,0 +1,106 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "wndw.h"
+#include "atom.h"
+
+#include <nvif/pushc37b.h>
+
+#include <nvhw/class/clc57e.h>
+
+static int
+wndwc67e_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw)
+{
+	struct nvif_push *push = wndw->wndw.push;
+	int ret;
+
+	if ((ret = PUSH_WAIT(push, 17)))
+		return ret;
+
+	PUSH_MTHD(push, NVC57E, SET_PRESENT_CONTROL,
+		  NVVAL(NVC57E, SET_PRESENT_CONTROL, MIN_PRESENT_INTERVAL, asyw->image.interval) |
+		  NVVAL(NVC57E, SET_PRESENT_CONTROL, BEGIN_MODE, asyw->image.mode) |
+		  NVDEF(NVC57E, SET_PRESENT_CONTROL, TIMESTAMP_MODE, DISABLE));
+
+	PUSH_MTHD(push, NVC57E, SET_SIZE,
+		  NVVAL(NVC57E, SET_SIZE, WIDTH, asyw->image.w) |
+		  NVVAL(NVC57E, SET_SIZE, HEIGHT, asyw->image.h),
+
+				SET_STORAGE,
+		  NVVAL(NVC57E, SET_STORAGE, BLOCK_HEIGHT, asyw->image.blockh),
+
+				SET_PARAMS,
+		  NVVAL(NVC57E, SET_PARAMS, FORMAT, asyw->image.format) |
+		  NVDEF(NVC57E, SET_PARAMS, CLAMP_BEFORE_BLEND, DISABLE) |
+		  NVDEF(NVC57E, SET_PARAMS, SWAP_UV, DISABLE) |
+		  NVDEF(NVC57E, SET_PARAMS, FMT_ROUNDING_MODE, ROUND_TO_NEAREST),
+
+				SET_PLANAR_STORAGE(0),
+		  NVVAL(NVC57E, SET_PLANAR_STORAGE, PITCH, asyw->image.blocks[0]) |
+		  NVVAL(NVC57E, SET_PLANAR_STORAGE, PITCH, asyw->image.pitch[0] >> 6));
+
+	PUSH_MTHD(push, NVC57E, SET_CONTEXT_DMA_ISO(0), asyw->image.handle, 1);
+	PUSH_MTHD(push, NVC57E, SET_OFFSET(0), asyw->image.offset[0] >> 8);
+
+	PUSH_MTHD(push, NVC57E, SET_POINT_IN(0),
+		  NVVAL(NVC57E, SET_POINT_IN, X, asyw->state.src_x >> 16) |
+		  NVVAL(NVC57E, SET_POINT_IN, Y, asyw->state.src_y >> 16));
+
+	PUSH_MTHD(push, NVC57E, SET_SIZE_IN,
+		  NVVAL(NVC57E, SET_SIZE_IN, WIDTH, asyw->state.src_w >> 16) |
+		  NVVAL(NVC57E, SET_SIZE_IN, HEIGHT, asyw->state.src_h >> 16));
+
+	PUSH_MTHD(push, NVC57E, SET_SIZE_OUT,
+		  NVVAL(NVC57E, SET_SIZE_OUT, WIDTH, asyw->state.crtc_w) |
+		  NVVAL(NVC57E, SET_SIZE_OUT, HEIGHT, asyw->state.crtc_h));
+	return 0;
+}
+
+static const struct nv50_wndw_func
+wndwc67e = {
+	.acquire = wndwc37e_acquire,
+	.release = wndwc37e_release,
+	.sema_set = wndwc37e_sema_set,
+	.sema_clr = wndwc37e_sema_clr,
+	.ntfy_set = wndwc37e_ntfy_set,
+	.ntfy_clr = wndwc37e_ntfy_clr,
+	.ntfy_reset = corec37d_ntfy_init,
+	.ntfy_wait_begun = base507c_ntfy_wait_begun,
+	.ilut = wndwc57e_ilut,
+	.ilut_identity = true,
+	.ilut_size = 1024,
+	.xlut_set = wndwc57e_ilut_set,
+	.xlut_clr = wndwc57e_ilut_clr,
+	.csc = base907c_csc,
+	.csc_set = wndwc57e_csc_set,
+	.csc_clr = wndwc57e_csc_clr,
+	.image_set = wndwc67e_image_set,
+	.image_clr = wndwc37e_image_clr,
+	.blend_set = wndwc37e_blend_set,
+	.update = wndwc37e_update,
+};
+
+int
+wndwc67e_new(struct nouveau_drm *drm, enum drm_plane_type type, int index,
+	     s32 oclass, struct nv50_wndw **pwndw)
+{
+	return wndwc37e_new_(&wndwc67e, drm, type, index, oclass, BIT(index >> 1), pwndw);
+}
diff --git a/drivers/gpu/drm/nouveau/include/nvhw/class/cl917d.h b/drivers/gpu/drm/nouveau/include/nvhw/class/cl917d.h
index 2a2612d6e1e0..fb223723a38a 100644
--- a/drivers/gpu/drm/nouveau/include/nvhw/class/cl917d.h
+++ b/drivers/gpu/drm/nouveau/include/nvhw/class/cl917d.h
@@ -66,6 +66,10 @@
 #define NV917D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_ALPHA_BLEND                  (0x00000000)
 #define NV917D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_PREMULT_ALPHA_BLEND          (0x00000001)
 #define NV917D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_XOR                          (0x00000002)
+#define NV917D_HEAD_SET_OFFSET_CURSOR(a)                                        (0x00000484 + (a)*0x00000300)
+#define NV917D_HEAD_SET_OFFSET_CURSOR_ORIGIN                                    31:0
+#define NV917D_HEAD_SET_CONTEXT_DMA_CURSOR(a)                                   (0x0000048C + (a)*0x00000300)
+#define NV917D_HEAD_SET_CONTEXT_DMA_CURSOR_HANDLE                               31:0
 #define NV917D_HEAD_SET_DITHER_CONTROL(a)                                       (0x000004A0 + (a)*0x00000300)
 #define NV917D_HEAD_SET_DITHER_CONTROL_ENABLE                                   0:0
 #define NV917D_HEAD_SET_DITHER_CONTROL_ENABLE_DISABLE                           (0x00000000)
diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h
index cd9a2e687bb6..0b86c44878e0 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h
@@ -33,6 +33,7 @@ struct nv_device_info_v0 {
 #define NV_DEVICE_INFO_V0_PASCAL                                           0x0a
 #define NV_DEVICE_INFO_V0_VOLTA                                            0x0b
 #define NV_DEVICE_INFO_V0_TURING                                           0x0c
+#define NV_DEVICE_INFO_V0_AMPERE                                           0x0d
 	__u8  family;
 	__u8  pad06[2];
 	__u64 ram_size;
@@ -59,37 +60,33 @@ struct nv_device_time_v0 {
 
 #define NV_DEVICE_INFO_UNIT                               (0xffffffffULL << 32)
 #define NV_DEVICE_INFO(n)                          ((n) | (0x00000000ULL << 32))
-#define NV_DEVICE_FIFO(n)                          ((n) | (0x00000001ULL << 32))
+#define NV_DEVICE_HOST(n)                          ((n) | (0x00000001ULL << 32))
 
-/* This will be returned for unsupported queries. */
+/* This will be returned in the mthd field for unsupported queries. */
 #define NV_DEVICE_INFO_INVALID                                           ~0ULL
 
-/* These return a mask of available engines of particular type. */
-#define NV_DEVICE_INFO_ENGINE_SW                     NV_DEVICE_INFO(0x00000000)
-#define NV_DEVICE_INFO_ENGINE_GR                     NV_DEVICE_INFO(0x00000001)
-#define NV_DEVICE_INFO_ENGINE_MPEG                   NV_DEVICE_INFO(0x00000002)
-#define NV_DEVICE_INFO_ENGINE_ME                     NV_DEVICE_INFO(0x00000003)
-#define NV_DEVICE_INFO_ENGINE_CIPHER                 NV_DEVICE_INFO(0x00000004)
-#define NV_DEVICE_INFO_ENGINE_BSP                    NV_DEVICE_INFO(0x00000005)
-#define NV_DEVICE_INFO_ENGINE_VP                     NV_DEVICE_INFO(0x00000006)
-#define NV_DEVICE_INFO_ENGINE_CE                     NV_DEVICE_INFO(0x00000007)
-#define NV_DEVICE_INFO_ENGINE_SEC                    NV_DEVICE_INFO(0x00000008)
-#define NV_DEVICE_INFO_ENGINE_MSVLD                  NV_DEVICE_INFO(0x00000009)
-#define NV_DEVICE_INFO_ENGINE_MSPDEC                 NV_DEVICE_INFO(0x0000000a)
-#define NV_DEVICE_INFO_ENGINE_MSPPP                  NV_DEVICE_INFO(0x0000000b)
-#define NV_DEVICE_INFO_ENGINE_MSENC                  NV_DEVICE_INFO(0x0000000c)
-#define NV_DEVICE_INFO_ENGINE_VIC                    NV_DEVICE_INFO(0x0000000d)
-#define NV_DEVICE_INFO_ENGINE_SEC2                   NV_DEVICE_INFO(0x0000000e)
-#define NV_DEVICE_INFO_ENGINE_NVDEC                  NV_DEVICE_INFO(0x0000000f)
-#define NV_DEVICE_INFO_ENGINE_NVENC                  NV_DEVICE_INFO(0x00000010)
-
+/* Returns the number of available runlists. */
+#define NV_DEVICE_HOST_RUNLISTS                       NV_DEVICE_HOST(0x00000000)
 /* Returns the number of available channels. */
-#define NV_DEVICE_FIFO_CHANNELS                      NV_DEVICE_FIFO(0x00000000)
-
-/* Returns a mask of available runlists. */
-#define NV_DEVICE_FIFO_RUNLISTS                      NV_DEVICE_FIFO(0x00000001)
+#define NV_DEVICE_HOST_CHANNELS                       NV_DEVICE_HOST(0x00000001)
 
-/* These return a mask of engines available on a particular runlist. */
-#define NV_DEVICE_FIFO_RUNLIST_ENGINES(n)     ((n) + NV_DEVICE_FIFO(0x00000010))
-#define NV_DEVICE_FIFO_RUNLIST_ENGINES__SIZE                                64
+/* Returns a mask of available engine types on runlist(data). */
+#define NV_DEVICE_HOST_RUNLIST_ENGINES                NV_DEVICE_HOST(0x00000100)
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_SW                            0x00000001
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_GR                            0x00000002
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_MPEG                          0x00000004
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_ME                            0x00000008
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_CIPHER                        0x00000010
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_BSP                           0x00000020
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_VP                            0x00000040
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_CE                            0x00000080
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_SEC                           0x00000100
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_MSVLD                         0x00000200
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_MSPDEC                        0x00000400
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_MSPPP                         0x00000800
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_MSENC                         0x00001000
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_VIC                           0x00002000
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_SEC2                          0x00004000
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_NVDEC                         0x00008000
+#define NV_DEVICE_HOST_RUNLIST_ENGINES_NVENC                         0x00010000
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvif/class.h b/drivers/gpu/drm/nouveau/include/nvif/class.h
index 2c79beb41126..ba2c28ea43d2 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/class.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/class.h
@@ -88,6 +88,7 @@
 #define GP102_DISP                                    /* cl5070.h */ 0x00009870
 #define GV100_DISP                                    /* cl5070.h */ 0x0000c370
 #define TU102_DISP                                    /* cl5070.h */ 0x0000c570
+#define GA102_DISP                                    /* cl5070.h */ 0x0000c670
 
 #define GV100_DISP_CAPS                                              0x0000c373
 
@@ -103,6 +104,7 @@
 #define GK104_DISP_CURSOR                             /* cl507a.h */ 0x0000917a
 #define GV100_DISP_CURSOR                             /* cl507a.h */ 0x0000c37a
 #define TU102_DISP_CURSOR                             /* cl507a.h */ 0x0000c57a
+#define GA102_DISP_CURSOR                             /* cl507a.h */ 0x0000c67a
 
 #define NV50_DISP_OVERLAY                             /* cl507b.h */ 0x0000507b
 #define G82_DISP_OVERLAY                              /* cl507b.h */ 0x0000827b
@@ -112,6 +114,7 @@
 
 #define GV100_DISP_WINDOW_IMM_CHANNEL_DMA             /* clc37b.h */ 0x0000c37b
 #define TU102_DISP_WINDOW_IMM_CHANNEL_DMA             /* clc37b.h */ 0x0000c57b
+#define GA102_DISP_WINDOW_IMM_CHANNEL_DMA             /* clc37b.h */ 0x0000c67b
 
 #define NV50_DISP_BASE_CHANNEL_DMA                    /* cl507c.h */ 0x0000507c
 #define G82_DISP_BASE_CHANNEL_DMA                     /* cl507c.h */ 0x0000827c
@@ -135,6 +138,7 @@
 #define GP102_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000987d
 #define GV100_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000c37d
 #define TU102_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000c57d
+#define GA102_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000c67d
 
 #define NV50_DISP_OVERLAY_CHANNEL_DMA                 /* cl507e.h */ 0x0000507e
 #define G82_DISP_OVERLAY_CHANNEL_DMA                  /* cl507e.h */ 0x0000827e
@@ -145,6 +149,7 @@
 
 #define GV100_DISP_WINDOW_CHANNEL_DMA                 /* clc37e.h */ 0x0000c37e
 #define TU102_DISP_WINDOW_CHANNEL_DMA                 /* clc37e.h */ 0x0000c57e
+#define GA102_DISP_WINDOW_CHANNEL_DMA                 /* clc37e.h */ 0x0000c67e
 
 #define NV50_TESLA                                                   0x00005097
 #define G82_TESLA                                                    0x00008297
diff --git a/drivers/gpu/drm/nouveau/include/nvif/fifo.h b/drivers/gpu/drm/nouveau/include/nvif/fifo.h
index e9468c9f9abf..d351ac890ca1 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/fifo.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/fifo.h
@@ -2,15 +2,15 @@
 #define __NVIF_FIFO_H__
 #include <nvif/device.h>
 
-/* Returns mask of runlists that support a NV_DEVICE_INFO_ENGINE_* type. */
+/* Returns mask of runlists that support a NV_DEVICE_INFO_RUNLIST_ENGINES_* type. */
 u64 nvif_fifo_runlist(struct nvif_device *, u64 engine);
 
 /* CE-supporting runlists (excluding GRCE, if others exist). */
 static inline u64
 nvif_fifo_runlist_ce(struct nvif_device *device)
 {
-	u64 runmgr = nvif_fifo_runlist(device, NV_DEVICE_INFO_ENGINE_GR);
-	u64 runmce = nvif_fifo_runlist(device, NV_DEVICE_INFO_ENGINE_CE);
+	u64 runmgr = nvif_fifo_runlist(device, NV_DEVICE_HOST_RUNLIST_ENGINES_GR);
+	u64 runmce = nvif_fifo_runlist(device, NV_DEVICE_HOST_RUNLIST_ENGINES_CE);
 	if (runmce && !(runmce &= ~runmgr))
 		runmce = runmgr;
 	return runmce;
diff --git a/drivers/gpu/drm/nouveau/include/nvif/push.h b/drivers/gpu/drm/nouveau/include/nvif/push.h
index 168d7694ede5..6d3a8a3d2087 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/push.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/push.h
@@ -123,131 +123,131 @@ PUSH_KICK(struct nvif_push *push)
 } while(0)
 #endif
 
-#define PUSH_1(X,f,ds,n,c,o,p,s,mA,dA) do {                            \
-	PUSH_##o##_HDR((p), s, mA, (c)+(n));                           \
-	PUSH_##f(X, (p), X##mA, 1, o, (dA), ds, "");                   \
+#define PUSH_1(X,f,ds,n,o,p,s,mA,dA) do {                             \
+	PUSH_##o##_HDR((p), s, mA, (ds)+(n));                         \
+	PUSH_##f(X, (p), X##mA, 1, o, (dA), ds, "");                  \
 } while(0)
-#define PUSH_2(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (1?PUSH_##o##_INC), "mthd1");       \
-	PUSH_1(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_2(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (1?PUSH_##o##_INC), "mthd1");      \
+	PUSH_1(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_3(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd2");       \
-	PUSH_2(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_3(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd2");      \
+	PUSH_2(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_4(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd3");       \
-	PUSH_3(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_4(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd3");      \
+	PUSH_3(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_5(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd4");       \
-	PUSH_4(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_5(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd4");      \
+	PUSH_4(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_6(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd5");       \
-	PUSH_5(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_6(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd5");      \
+	PUSH_5(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_7(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd6");       \
-	PUSH_6(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_7(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd6");      \
+	PUSH_6(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_8(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd7");       \
-	PUSH_7(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_8(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd7");      \
+	PUSH_7(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_9(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                 \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd8");       \
-	PUSH_8(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_9(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                  \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd8");      \
+	PUSH_8(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
-#define PUSH_10(X,f,ds,n,c,o,p,s,mB,dB,mA,dA,a...) do {                \
-	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd9");       \
-	PUSH_9(X, DATA_, 1, ds, (c)+(n), o, (p), s, X##mA, (dA), ##a); \
-	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                   \
+#define PUSH_10(X,f,ds,n,o,p,s,mB,dB,mA,dA,a...) do {                 \
+	PUSH_ASSERT((mB) - (mA) == (0?PUSH_##o##_INC), "mthd9");      \
+	PUSH_9(X, DATA_, 1, (ds) + (n), o, (p), s, X##mA, (dA), ##a); \
+	PUSH_##f(X, (p), X##mB, 0, o, (dB), ds, "");                  \
 } while(0)
 
-#define PUSH_1D(X,o,p,s,mA,dA)                            \
-	PUSH_1(X, DATA_, 1, 1, 0, o, (p), s, X##mA, (dA))
-#define PUSH_2D(X,o,p,s,mA,dA,mB,dB)                      \
-	PUSH_2(X, DATA_, 1, 1, 0, o, (p), s, X##mB, (dB), \
-					     X##mA, (dA))
-#define PUSH_3D(X,o,p,s,mA,dA,mB,dB,mC,dC)                \
-	PUSH_3(X, DATA_, 1, 1, 0, o, (p), s, X##mC, (dC), \
-					     X##mB, (dB), \
-					     X##mA, (dA))
-#define PUSH_4D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD)          \
-	PUSH_4(X, DATA_, 1, 1, 0, o, (p), s, X##mD, (dD), \
-					     X##mC, (dC), \
-					     X##mB, (dB), \
-					     X##mA, (dA))
-#define PUSH_5D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE)    \
-	PUSH_5(X, DATA_, 1, 1, 0, o, (p), s, X##mE, (dE), \
-					     X##mD, (dD), \
-					     X##mC, (dC), \
-					     X##mB, (dB), \
-					     X##mA, (dA))
+#define PUSH_1D(X,o,p,s,mA,dA)                         \
+	PUSH_1(X, DATA_, 1, 0, o, (p), s, X##mA, (dA))
+#define PUSH_2D(X,o,p,s,mA,dA,mB,dB)                   \
+	PUSH_2(X, DATA_, 1, 0, o, (p), s, X##mB, (dB), \
+					  X##mA, (dA))
+#define PUSH_3D(X,o,p,s,mA,dA,mB,dB,mC,dC)             \
+	PUSH_3(X, DATA_, 1, 0, o, (p), s, X##mC, (dC), \
+					  X##mB, (dB), \
+					  X##mA, (dA))
+#define PUSH_4D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD)       \
+	PUSH_4(X, DATA_, 1, 0, o, (p), s, X##mD, (dD), \
+					  X##mC, (dC), \
+					  X##mB, (dB), \
+					  X##mA, (dA))
+#define PUSH_5D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE) \
+	PUSH_5(X, DATA_, 1, 0, o, (p), s, X##mE, (dE), \
+					  X##mD, (dD), \
+					  X##mC, (dC), \
+					  X##mB, (dB), \
+					  X##mA, (dA))
 #define PUSH_6D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE,mF,dF) \
-	PUSH_6(X, DATA_, 1, 1, 0, o, (p), s, X##mF, (dF),    \
-					     X##mE, (dE),    \
-					     X##mD, (dD),    \
-					     X##mC, (dC),    \
-					     X##mB, (dB),    \
-					     X##mA, (dA))
+	PUSH_6(X, DATA_, 1, 0, o, (p), s, X##mF, (dF),       \
+					  X##mE, (dE),       \
+					  X##mD, (dD),       \
+					  X##mC, (dC),       \
+					  X##mB, (dB),       \
+					  X##mA, (dA))
 #define PUSH_7D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE,mF,dF,mG,dG) \
-	PUSH_7(X, DATA_, 1, 1, 0, o, (p), s, X##mG, (dG),          \
-					     X##mF, (dF),          \
-					     X##mE, (dE),          \
-					     X##mD, (dD),          \
-					     X##mC, (dC),          \
-					     X##mB, (dB),          \
-					     X##mA, (dA))
+	PUSH_7(X, DATA_, 1, 0, o, (p), s, X##mG, (dG),             \
+					  X##mF, (dF),             \
+					  X##mE, (dE),             \
+					  X##mD, (dD),             \
+					  X##mC, (dC),             \
+					  X##mB, (dB),             \
+					  X##mA, (dA))
 #define PUSH_8D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE,mF,dF,mG,dG,mH,dH) \
-	PUSH_8(X, DATA_, 1, 1, 0, o, (p), s, X##mH, (dH),                \
-					     X##mG, (dG),                \
-					     X##mF, (dF),                \
-					     X##mE, (dE),                \
-					     X##mD, (dD),                \
-					     X##mC, (dC),                \
-					     X##mB, (dB),                \
-					     X##mA, (dA))
+	PUSH_8(X, DATA_, 1, 0, o, (p), s, X##mH, (dH),                   \
+					  X##mG, (dG),                   \
+					  X##mF, (dF),                   \
+					  X##mE, (dE),                   \
+					  X##mD, (dD),                   \
+					  X##mC, (dC),                   \
+					  X##mB, (dB),                   \
+					  X##mA, (dA))
 #define PUSH_9D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE,mF,dF,mG,dG,mH,dH,mI,dI) \
-	PUSH_9(X, DATA_, 1, 1, 0, o, (p), s, X##mI, (dI),                      \
-					     X##mH, (dH),                      \
-					     X##mG, (dG),                      \
-					     X##mF, (dF),                      \
-					     X##mE, (dE),                      \
-					     X##mD, (dD),                      \
-					     X##mC, (dC),                      \
-					     X##mB, (dB),                      \
-					     X##mA, (dA))
+	PUSH_9(X, DATA_, 1, 0, o, (p), s, X##mI, (dI),                         \
+					  X##mH, (dH),                         \
+					  X##mG, (dG),                         \
+					  X##mF, (dF),                         \
+					  X##mE, (dE),                         \
+					  X##mD, (dD),                         \
+					  X##mC, (dC),                         \
+					  X##mB, (dB),                         \
+					  X##mA, (dA))
 #define PUSH_10D(X,o,p,s,mA,dA,mB,dB,mC,dC,mD,dD,mE,dE,mF,dF,mG,dG,mH,dH,mI,dI,mJ,dJ) \
-	PUSH_10(X, DATA_, 1, 1, 0, o, (p), s, X##mJ, (dJ),                            \
-					      X##mI, (dI),                            \
-					      X##mH, (dH),                            \
-					      X##mG, (dG),                            \
-					      X##mF, (dF),                            \
-					      X##mE, (dE),                            \
-					      X##mD, (dD),                            \
-					      X##mC, (dC),                            \
-					      X##mB, (dB),                            \
-					      X##mA, (dA))
+	PUSH_10(X, DATA_, 1, 0, o, (p), s, X##mJ, (dJ),                               \
+					   X##mI, (dI),                               \
+					   X##mH, (dH),                               \
+					   X##mG, (dG),                               \
+					   X##mF, (dF),                               \
+					   X##mE, (dE),                               \
+					   X##mD, (dD),                               \
+					   X##mC, (dC),                               \
+					   X##mB, (dB),                               \
+					   X##mA, (dA))
 
-#define PUSH_1P(X,o,p,s,mA,dp,ds)                           \
-	PUSH_1(X, DATAp, ds, ds, 0, o, (p), s, X##mA, (dp))
-#define PUSH_2P(X,o,p,s,mA,dA,mB,dp,ds)                     \
-	PUSH_2(X, DATAp, ds, ds, 0, o, (p), s, X##mB, (dp), \
-					       X##mA, (dA))
-#define PUSH_3P(X,o,p,s,mA,dA,mB,dB,mC,dp,ds)               \
-	PUSH_3(X, DATAp, ds, ds, 0, o, (p), s, X##mC, (dp), \
-					       X##mB, (dB), \
-					       X##mA, (dA))
+#define PUSH_1P(X,o,p,s,mA,dp,ds)                       \
+	PUSH_1(X, DATAp, ds, 0, o, (p), s, X##mA, (dp))
+#define PUSH_2P(X,o,p,s,mA,dA,mB,dp,ds)                 \
+	PUSH_2(X, DATAp, ds, 0, o, (p), s, X##mB, (dp), \
+					   X##mA, (dA))
+#define PUSH_3P(X,o,p,s,mA,dA,mB,dB,mC,dp,ds)           \
+	PUSH_3(X, DATAp, ds, 0, o, (p), s, X##mC, (dp), \
+					   X##mB, (dB), \
+					   X##mA, (dA))
 
 #define PUSH_(A,B,C,D,E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,IMPL,...) IMPL
 #define PUSH(A...) PUSH_(A, PUSH_10P, PUSH_10D,          \
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h
index 5c007ce62fc3..a18b6cfda07e 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h
@@ -3,79 +3,7 @@
 #define __NVKM_DEVICE_H__
 #include <core/oclass.h>
 #include <core/event.h>
-
-enum nvkm_devidx {
-	NVKM_SUBDEV_PCI,
-	NVKM_SUBDEV_VBIOS,
-	NVKM_SUBDEV_DEVINIT,
-	NVKM_SUBDEV_TOP,
-	NVKM_SUBDEV_IBUS,
-	NVKM_SUBDEV_GPIO,
-	NVKM_SUBDEV_I2C,
-	NVKM_SUBDEV_FUSE,
-	NVKM_SUBDEV_MXM,
-	NVKM_SUBDEV_MC,
-	NVKM_SUBDEV_BUS,
-	NVKM_SUBDEV_TIMER,
-	NVKM_SUBDEV_INSTMEM,
-	NVKM_SUBDEV_FB,
-	NVKM_SUBDEV_LTC,
-	NVKM_SUBDEV_MMU,
-	NVKM_SUBDEV_BAR,
-	NVKM_SUBDEV_FAULT,
-	NVKM_SUBDEV_ACR,
-	NVKM_SUBDEV_PMU,
-	NVKM_SUBDEV_VOLT,
-	NVKM_SUBDEV_ICCSENSE,
-	NVKM_SUBDEV_THERM,
-	NVKM_SUBDEV_CLK,
-	NVKM_SUBDEV_GSP,
-
-	NVKM_ENGINE_BSP,
-
-	NVKM_ENGINE_CE0,
-	NVKM_ENGINE_CE1,
-	NVKM_ENGINE_CE2,
-	NVKM_ENGINE_CE3,
-	NVKM_ENGINE_CE4,
-	NVKM_ENGINE_CE5,
-	NVKM_ENGINE_CE6,
-	NVKM_ENGINE_CE7,
-	NVKM_ENGINE_CE8,
-	NVKM_ENGINE_CE_LAST = NVKM_ENGINE_CE8,
-
-	NVKM_ENGINE_CIPHER,
-	NVKM_ENGINE_DISP,
-	NVKM_ENGINE_DMAOBJ,
-	NVKM_ENGINE_FIFO,
-	NVKM_ENGINE_GR,
-	NVKM_ENGINE_IFB,
-	NVKM_ENGINE_ME,
-	NVKM_ENGINE_MPEG,
-	NVKM_ENGINE_MSENC,
-	NVKM_ENGINE_MSPDEC,
-	NVKM_ENGINE_MSPPP,
-	NVKM_ENGINE_MSVLD,
-
-	NVKM_ENGINE_NVENC0,
-	NVKM_ENGINE_NVENC1,
-	NVKM_ENGINE_NVENC2,
-	NVKM_ENGINE_NVENC_LAST = NVKM_ENGINE_NVENC2,
-
-	NVKM_ENGINE_NVDEC0,
-	NVKM_ENGINE_NVDEC1,
-	NVKM_ENGINE_NVDEC2,
-	NVKM_ENGINE_NVDEC_LAST = NVKM_ENGINE_NVDEC2,
-
-	NVKM_ENGINE_PM,
-	NVKM_ENGINE_SEC,
-	NVKM_ENGINE_SEC2,
-	NVKM_ENGINE_SW,
-	NVKM_ENGINE_VIC,
-	NVKM_ENGINE_VP,
-
-	NVKM_SUBDEV_NR
-};
+enum nvkm_subdev_type;
 
 enum nvkm_device_type {
 	NVKM_DEVICE_PCI,
@@ -102,7 +30,6 @@ struct nvkm_device {
 
 	struct nvkm_event event;
 
-	u64 disable_mask;
 	u32 debug;
 
 	const struct nvkm_device_chip *chip;
@@ -120,6 +47,7 @@ struct nvkm_device {
 		GP100    = 0x130,
 		GV100    = 0x140,
 		TU100    = 0x160,
+		GA100    = 0x170,
 	} card_type;
 	u32 chipset;
 	u8  chiprev;
@@ -129,58 +57,16 @@ struct nvkm_device {
 		struct notifier_block nb;
 	} acpi;
 
-	struct nvkm_acr *acr;
-	struct nvkm_bar *bar;
-	struct nvkm_bios *bios;
-	struct nvkm_bus *bus;
-	struct nvkm_clk *clk;
-	struct nvkm_devinit *devinit;
-	struct nvkm_fault *fault;
-	struct nvkm_fb *fb;
-	struct nvkm_fuse *fuse;
-	struct nvkm_gpio *gpio;
-	struct nvkm_gsp *gsp;
-	struct nvkm_i2c *i2c;
-	struct nvkm_subdev *ibus;
-	struct nvkm_iccsense *iccsense;
-	struct nvkm_instmem *imem;
-	struct nvkm_ltc *ltc;
-	struct nvkm_mc *mc;
-	struct nvkm_mmu *mmu;
-	struct nvkm_subdev *mxm;
-	struct nvkm_pci *pci;
-	struct nvkm_pmu *pmu;
-	struct nvkm_therm *therm;
-	struct nvkm_timer *timer;
-	struct nvkm_top *top;
-	struct nvkm_volt *volt;
-
-	struct nvkm_engine *bsp;
-	struct nvkm_engine *ce[9];
-	struct nvkm_engine *cipher;
-	struct nvkm_disp *disp;
-	struct nvkm_dma *dma;
-	struct nvkm_fifo *fifo;
-	struct nvkm_gr *gr;
-	struct nvkm_engine *ifb;
-	struct nvkm_engine *me;
-	struct nvkm_engine *mpeg;
-	struct nvkm_engine *msenc;
-	struct nvkm_engine *mspdec;
-	struct nvkm_engine *msppp;
-	struct nvkm_engine *msvld;
-	struct nvkm_nvenc *nvenc[3];
-	struct nvkm_nvdec *nvdec[3];
-	struct nvkm_pm *pm;
-	struct nvkm_engine *sec;
-	struct nvkm_sec2 *sec2;
-	struct nvkm_sw *sw;
-	struct nvkm_engine *vic;
-	struct nvkm_engine *vp;
+#define NVKM_LAYOUT_ONCE(type,data,ptr) data *ptr;
+#define NVKM_LAYOUT_INST(type,data,ptr,cnt) data *ptr[cnt];
+#include <core/layout.h>
+#undef NVKM_LAYOUT_INST
+#undef NVKM_LAYOUT_ONCE
+	struct list_head subdev;
 };
 
-struct nvkm_subdev *nvkm_device_subdev(struct nvkm_device *, int index);
-struct nvkm_engine *nvkm_device_engine(struct nvkm_device *, int index);
+struct nvkm_subdev *nvkm_device_subdev(struct nvkm_device *, int type, int inst);
+struct nvkm_engine *nvkm_device_engine(struct nvkm_device *, int type, int inst);
 
 struct nvkm_device_func {
 	struct nvkm_device_pci *(*pci)(struct nvkm_device *);
@@ -201,55 +87,15 @@ struct nvkm_device_quirk {
 
 struct nvkm_device_chip {
 	const char *name;
-
-	int (*acr     )(struct nvkm_device *, int idx, struct nvkm_acr **);
-	int (*bar     )(struct nvkm_device *, int idx, struct nvkm_bar **);
-	int (*bios    )(struct nvkm_device *, int idx, struct nvkm_bios **);
-	int (*bus     )(struct nvkm_device *, int idx, struct nvkm_bus **);
-	int (*clk     )(struct nvkm_device *, int idx, struct nvkm_clk **);
-	int (*devinit )(struct nvkm_device *, int idx, struct nvkm_devinit **);
-	int (*fault   )(struct nvkm_device *, int idx, struct nvkm_fault **);
-	int (*fb      )(struct nvkm_device *, int idx, struct nvkm_fb **);
-	int (*fuse    )(struct nvkm_device *, int idx, struct nvkm_fuse **);
-	int (*gpio    )(struct nvkm_device *, int idx, struct nvkm_gpio **);
-	int (*gsp     )(struct nvkm_device *, int idx, struct nvkm_gsp **);
-	int (*i2c     )(struct nvkm_device *, int idx, struct nvkm_i2c **);
-	int (*ibus    )(struct nvkm_device *, int idx, struct nvkm_subdev **);
-	int (*iccsense)(struct nvkm_device *, int idx, struct nvkm_iccsense **);
-	int (*imem    )(struct nvkm_device *, int idx, struct nvkm_instmem **);
-	int (*ltc     )(struct nvkm_device *, int idx, struct nvkm_ltc **);
-	int (*mc      )(struct nvkm_device *, int idx, struct nvkm_mc **);
-	int (*mmu     )(struct nvkm_device *, int idx, struct nvkm_mmu **);
-	int (*mxm     )(struct nvkm_device *, int idx, struct nvkm_subdev **);
-	int (*pci     )(struct nvkm_device *, int idx, struct nvkm_pci **);
-	int (*pmu     )(struct nvkm_device *, int idx, struct nvkm_pmu **);
-	int (*therm   )(struct nvkm_device *, int idx, struct nvkm_therm **);
-	int (*timer   )(struct nvkm_device *, int idx, struct nvkm_timer **);
-	int (*top     )(struct nvkm_device *, int idx, struct nvkm_top **);
-	int (*volt    )(struct nvkm_device *, int idx, struct nvkm_volt **);
-
-	int (*bsp     )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*ce[9]   )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*cipher  )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*disp    )(struct nvkm_device *, int idx, struct nvkm_disp **);
-	int (*dma     )(struct nvkm_device *, int idx, struct nvkm_dma **);
-	int (*fifo    )(struct nvkm_device *, int idx, struct nvkm_fifo **);
-	int (*gr      )(struct nvkm_device *, int idx, struct nvkm_gr **);
-	int (*ifb     )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*me      )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*mpeg    )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*msenc   )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*mspdec  )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*msppp   )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*msvld   )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*nvenc[3])(struct nvkm_device *, int idx, struct nvkm_nvenc **);
-	int (*nvdec[3])(struct nvkm_device *, int idx, struct nvkm_nvdec **);
-	int (*pm      )(struct nvkm_device *, int idx, struct nvkm_pm **);
-	int (*sec     )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*sec2    )(struct nvkm_device *, int idx, struct nvkm_sec2 **);
-	int (*sw      )(struct nvkm_device *, int idx, struct nvkm_sw **);
-	int (*vic     )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*vp      )(struct nvkm_device *, int idx, struct nvkm_engine **);
+#define NVKM_LAYOUT_ONCE(type,data,ptr,...)                                                  \
+	struct {                                                                             \
+		u32 inst;                                                                    \
+		int (*ctor)(struct nvkm_device *, enum nvkm_subdev_type, int inst, data **); \
+	} ptr;
+#define NVKM_LAYOUT_INST(A...) NVKM_LAYOUT_ONCE(A)
+#include <core/layout.h>
+#undef NVKM_LAYOUT_INST
+#undef NVKM_LAYOUT_ONCE
 };
 
 struct nvkm_device *nvkm_device_find(u64 name);
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/engine.h b/drivers/gpu/drm/nouveau/include/nvkm/core/engine.h
index c6b401a6ea23..e58923b67d74 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/engine.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/engine.h
@@ -6,12 +6,18 @@
 struct nvkm_fifo_chan;
 struct nvkm_fb_tile;
 
+extern const struct nvkm_subdev_func nvkm_engine;
+
 struct nvkm_engine {
 	const struct nvkm_engine_func *func;
 	struct nvkm_subdev subdev;
 	spinlock_t lock;
 
-	int usecount;
+	struct {
+		refcount_t refcount;
+		struct mutex mutex;
+		bool enabled;
+	} use;
 };
 
 struct nvkm_engine_func {
@@ -42,9 +48,10 @@ struct nvkm_engine_func {
 };
 
 int nvkm_engine_ctor(const struct nvkm_engine_func *, struct nvkm_device *,
-		     int index, bool enable, struct nvkm_engine *);
+		     enum nvkm_subdev_type, int inst, bool enable, struct nvkm_engine *);
 int nvkm_engine_new_(const struct nvkm_engine_func *, struct nvkm_device *,
-		     int index, bool enable, struct nvkm_engine **);
+		     enum nvkm_subdev_type, int, bool enable, struct nvkm_engine **);
+
 struct nvkm_engine *nvkm_engine_ref(struct nvkm_engine *);
 void nvkm_engine_unref(struct nvkm_engine **);
 void nvkm_engine_tile(struct nvkm_engine *, int region);
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/enum.h b/drivers/gpu/drm/nouveau/include/nvkm/core/enum.h
index ce98efd4b209..070462be35d9 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/enum.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/enum.h
@@ -8,6 +8,7 @@ struct nvkm_enum {
 	const char *name;
 	const void *data;
 	u32 data2;
+	int inst;
 };
 
 const struct nvkm_enum *nvkm_enum_find(const struct nvkm_enum *, u32 value);
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/falcon.h b/drivers/gpu/drm/nouveau/include/nvkm/core/falcon.h
index 3981cb106aae..fd9a3f9a518e 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/falcon.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/falcon.h
@@ -21,11 +21,11 @@ void nvkm_falcon_v1_disable(struct nvkm_falcon *);
 void gp102_sec2_flcn_bind_context(struct nvkm_falcon *, struct nvkm_memory *);
 int gp102_sec2_flcn_enable(struct nvkm_falcon *);
 
-#define FLCN_PRINTK(t,f,fmt,a...) do {                                         \
-	if (nvkm_subdev_name[(f)->owner->index] != (f)->name)                  \
-		nvkm_##t((f)->owner, "%s: "fmt"\n", (f)->name, ##a);           \
-	else                                                                   \
-		nvkm_##t((f)->owner, fmt"\n", ##a);                            \
+#define FLCN_PRINTK(t,f,fmt,a...) do {                               \
+	if ((f)->owner->name != (f)->name)                           \
+		nvkm_##t((f)->owner, "%s: "fmt"\n", (f)->name, ##a); \
+	else                                                         \
+		nvkm_##t((f)->owner, fmt"\n", ##a);                  \
 } while(0)
 #define FLCN_DBG(f,fmt,a...) FLCN_PRINTK(debug, (f), fmt, ##a)
 #define FLCN_ERR(f,fmt,a...) FLCN_PRINTK(error, (f), fmt, ##a)
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h b/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h
new file mode 100644
index 000000000000..7afe1579b20f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h
@@ -0,0 +1,53 @@
+/* SPDX-License-Identifier: MIT */
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_PCI     , struct nvkm_pci     ,      pci)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_VBIOS   , struct nvkm_bios    ,     bios)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_DEVINIT , struct nvkm_devinit ,  devinit)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_TOP     , struct nvkm_top     ,      top)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_PRIVRING, struct nvkm_subdev  , privring)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_GPIO    , struct nvkm_gpio    ,     gpio)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_I2C     , struct nvkm_i2c     ,      i2c)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_FUSE    , struct nvkm_fuse    ,     fuse)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_MXM     , struct nvkm_subdev  ,      mxm)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_MC      , struct nvkm_mc      ,       mc)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_BUS     , struct nvkm_bus     ,      bus)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_TIMER   , struct nvkm_timer   ,    timer)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_INSTMEM , struct nvkm_instmem ,     imem)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_FB      , struct nvkm_fb      ,       fb)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_LTC     , struct nvkm_ltc     ,      ltc)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_MMU     , struct nvkm_mmu     ,      mmu)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_BAR     , struct nvkm_bar     ,      bar)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_FAULT   , struct nvkm_fault   ,    fault)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_ACR     , struct nvkm_acr     ,      acr)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_PMU     , struct nvkm_pmu     ,      pmu)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_VOLT    , struct nvkm_volt    ,     volt)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_ICCSENSE, struct nvkm_iccsense, iccsense)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_THERM   , struct nvkm_therm   ,    therm)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_CLK     , struct nvkm_clk     ,      clk)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_GSP     , struct nvkm_gsp     ,      gsp)
+NVKM_LAYOUT_INST(NVKM_SUBDEV_IOCTRL  , struct nvkm_subdev  ,   ioctrl, 3)
+NVKM_LAYOUT_ONCE(NVKM_SUBDEV_FLA     , struct nvkm_subdev  ,      fla)
+
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_BSP     , struct nvkm_engine  ,      bsp)
+NVKM_LAYOUT_INST(NVKM_ENGINE_CE      , struct nvkm_engine  ,       ce, 10)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_CIPHER  , struct nvkm_engine  ,   cipher)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_DISP    , struct nvkm_disp    ,     disp)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_DMAOBJ  , struct nvkm_dma     ,      dma)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_FIFO    , struct nvkm_fifo    ,     fifo)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_GR      , struct nvkm_gr      ,       gr)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_IFB     , struct nvkm_engine  ,      ifb)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_ME      , struct nvkm_engine  ,       me)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_MPEG    , struct nvkm_engine  ,     mpeg)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSENC   , struct nvkm_engine  ,    msenc)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSPDEC  , struct nvkm_engine  ,   mspdec)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSPPP   , struct nvkm_engine  ,    msppp)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSVLD   , struct nvkm_engine  ,    msvld)
+NVKM_LAYOUT_INST(NVKM_ENGINE_NVDEC   , struct nvkm_nvdec   ,    nvdec, 5)
+NVKM_LAYOUT_INST(NVKM_ENGINE_NVENC   , struct nvkm_nvenc   ,    nvenc, 3)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_NVJPG   , struct nvkm_engine  ,    nvjpg)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_OFA     , struct nvkm_engine  ,      ofa)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_PM      , struct nvkm_pm      ,       pm)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_SEC     , struct nvkm_engine  ,      sec)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_SEC2    , struct nvkm_sec2    ,     sec2)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_SW      , struct nvkm_sw      ,       sw)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_VIC     , struct nvkm_engine  ,      vic)
+NVKM_LAYOUT_ONCE(NVKM_ENGINE_VP      , struct nvkm_engine  ,       vp)
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/subdev.h b/drivers/gpu/drm/nouveau/include/nvkm/core/subdev.h
index 76288c682e9e..1665738948fb 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/subdev.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/subdev.h
@@ -3,13 +3,25 @@
 #define __NVKM_SUBDEV_H__
 #include <core/device.h>
 
+enum nvkm_subdev_type {
+#define NVKM_LAYOUT_ONCE(t,s,p,...) t,
+#define NVKM_LAYOUT_INST NVKM_LAYOUT_ONCE
+#include <core/layout.h>
+#undef NVKM_LAYOUT_INST
+#undef NVKM_LAYOUT_ONCE
+	NVKM_SUBDEV_NR
+};
+
 struct nvkm_subdev {
 	const struct nvkm_subdev_func *func;
 	struct nvkm_device *device;
-	enum nvkm_devidx index;
-	struct mutex mutex;
+	enum nvkm_subdev_type type;
+	int inst;
+	char name[16];
 	u32 debug;
+	struct list_head head;
 
+	void **pself;
 	bool oneinit;
 };
 
@@ -23,11 +35,12 @@ struct nvkm_subdev_func {
 	void (*intr)(struct nvkm_subdev *);
 };
 
-extern const char *nvkm_subdev_name[NVKM_SUBDEV_NR];
-int nvkm_subdev_new_(const struct nvkm_subdev_func *, struct nvkm_device *,
-		     int index, struct nvkm_subdev **);
+extern const char *nvkm_subdev_type[NVKM_SUBDEV_NR];
+int nvkm_subdev_new_(const struct nvkm_subdev_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		     int inst, struct nvkm_subdev **);
 void nvkm_subdev_ctor(const struct nvkm_subdev_func *, struct nvkm_device *,
-		      int index, struct nvkm_subdev *);
+		      enum nvkm_subdev_type, int inst, struct nvkm_subdev *);
+void nvkm_subdev_disable(struct nvkm_device *, enum nvkm_subdev_type, int inst);
 void nvkm_subdev_del(struct nvkm_subdev **);
 int  nvkm_subdev_preinit(struct nvkm_subdev *);
 int  nvkm_subdev_init(struct nvkm_subdev *);
@@ -38,10 +51,8 @@ void nvkm_subdev_intr(struct nvkm_subdev *);
 /* subdev logging */
 #define nvkm_printk_(s,l,p,f,a...) do {                                        \
 	const struct nvkm_subdev *_subdev = (s);                               \
-	if (CONFIG_NOUVEAU_DEBUG >= (l) && _subdev->debug >= (l)) {            \
-		dev_##p(_subdev->device->dev, "%s: "f,                         \
-			nvkm_subdev_name[_subdev->index], ##a);                \
-	}                                                                      \
+	if (CONFIG_NOUVEAU_DEBUG >= (l) && _subdev->debug >= (l))              \
+		dev_##p(_subdev->device->dev, "%s: "f, _subdev->name, ##a);    \
 } while(0)
 #define nvkm_printk(s,l,p,f,a...) nvkm_printk_((s), NV_DBG_##l, p, f, ##a)
 #define nvkm_fatal(s,f,a...) nvkm_printk((s), FATAL,   crit, f, ##a)
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/bsp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/bsp.h
index f938f024db81..d5530faf025e 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/bsp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/bsp.h
@@ -2,5 +2,5 @@
 #ifndef __NVKM_BSP_H__
 #define __NVKM_BSP_H__
 #include <engine/xtensa.h>
-int g84_bsp_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g84_bsp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h
index 86f420f4630b..cfd2da8e66fe 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h
@@ -3,13 +3,13 @@
 #define __NVKM_CE_H__
 #include <engine/falcon.h>
 
-int gt215_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gf100_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gk104_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gm107_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gm200_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gp100_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gp102_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gv100_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
-int tu102_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
+int gt215_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gf100_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gk104_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gm107_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gm200_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gp100_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gp102_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gv100_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int tu102_ce_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/cipher.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/cipher.h
index 66c5c5e27520..9da9176bbbbf 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/cipher.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/cipher.h
@@ -2,5 +2,5 @@
 #ifndef __NVKM_CIPHER_H__
 #define __NVKM_CIPHER_H__
 #include <core/engine.h>
-int g84_cipher_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g84_cipher_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h
index 5a96c942d912..d08d3337ba0d 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h
@@ -17,24 +17,28 @@ struct nvkm_disp {
 	struct nvkm_event hpd;
 	struct nvkm_event vblank;
 
-	struct nvkm_oproxy *client;
+	struct {
+		spinlock_t lock;
+		struct nvkm_oproxy *object;
+	} client;
 };
 
-int nv04_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int nv50_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int g84_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gt200_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int g94_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int mcp77_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gt215_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int mcp89_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gf119_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gk104_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gk110_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gm107_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gm200_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gp100_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gp102_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int gv100_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
-int tu102_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
+int nv04_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int nv50_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int g84_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gt200_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int g94_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int mcp77_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gt215_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int mcp89_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gf119_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gk104_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gk110_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gm107_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gm200_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gp100_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gp102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int gv100_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int tu102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
+int ga102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/dma.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/dma.h
index 2e12cdb6bb93..a003da39fd13 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/dma.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/dma.h
@@ -23,9 +23,9 @@ struct nvkm_dma {
 
 struct nvkm_dmaobj *nvkm_dmaobj_search(struct nvkm_client *, u64 object);
 
-int nv04_dma_new(struct nvkm_device *, int, struct nvkm_dma **);
-int nv50_dma_new(struct nvkm_device *, int, struct nvkm_dma **);
-int gf100_dma_new(struct nvkm_device *, int, struct nvkm_dma **);
-int gf119_dma_new(struct nvkm_device *, int, struct nvkm_dma **);
-int gv100_dma_new(struct nvkm_device *, int, struct nvkm_dma **);
+int nv04_dma_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_dma **);
+int nv50_dma_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_dma **);
+int gf100_dma_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_dma **);
+int gf119_dma_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_dma **);
+int gv100_dma_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_dma **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/falcon.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/falcon.h
index 27c1f868552c..306125d17ece 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/falcon.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/falcon.h
@@ -64,7 +64,7 @@ int nvkm_falcon_get(struct nvkm_falcon *, const struct nvkm_subdev *);
 void nvkm_falcon_put(struct nvkm_falcon *, const struct nvkm_subdev *);
 
 int nvkm_falcon_new_(const struct nvkm_falcon_func *, struct nvkm_device *,
-		     int index, bool enable, u32 addr, struct nvkm_engine **);
+		     enum nvkm_subdev_type, int inst, bool enable, u32 addr, struct nvkm_engine **);
 
 struct nvkm_falcon_func {
 	struct {
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h
index b335f3a1e66d..54fab7cc36c1 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h
@@ -7,6 +7,7 @@
 struct nvkm_fault_data;
 
 #define NVKM_FIFO_CHID_NR 4096
+#define NVKM_FIFO_ENGN_NR 16
 
 struct nvkm_fifo_engn {
 	struct nvkm_object *object;
@@ -17,7 +18,7 @@ struct nvkm_fifo_engn {
 struct nvkm_fifo_chan {
 	const struct nvkm_fifo_chan_func *func;
 	struct nvkm_fifo *fifo;
-	u64 engines;
+	u32 engm;
 	struct nvkm_object object;
 
 	struct list_head head;
@@ -29,7 +30,7 @@ struct nvkm_fifo_chan {
 	u64 addr;
 	u32 size;
 
-	struct nvkm_fifo_engn engn[NVKM_SUBDEV_NR];
+	struct nvkm_fifo_engn engn[NVKM_FIFO_ENGN_NR];
 };
 
 struct nvkm_fifo {
@@ -40,6 +41,7 @@ struct nvkm_fifo {
 	int nr;
 	struct list_head chan;
 	spinlock_t lock;
+	struct mutex mutex;
 
 	struct nvkm_event uevent; /* async user trigger */
 	struct nvkm_event cevent; /* channel creation event */
@@ -57,22 +59,22 @@ nvkm_fifo_chan_inst(struct nvkm_fifo *, u64 inst, unsigned long *flags);
 struct nvkm_fifo_chan *
 nvkm_fifo_chan_chid(struct nvkm_fifo *, int chid, unsigned long *flags);
 
-int nv04_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int nv10_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int nv17_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int nv40_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int nv50_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int g84_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gf100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gk104_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gk110_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gk208_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gk20a_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gm107_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gm200_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gm20b_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gp100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gp10b_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int gv100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
-int tu102_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
+int nv04_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int nv10_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int nv17_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int nv40_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int nv50_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int g84_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gf100_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gk104_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gk110_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gk208_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gk20a_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gm107_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gm200_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gm20b_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gp100_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gp10b_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int gv100_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
+int tu102_fifo_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fifo **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h
index 1530c81f86a2..b28b752ffaa2 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h
@@ -14,44 +14,44 @@ int nvkm_gr_ctxsw_pause(struct nvkm_device *);
 int nvkm_gr_ctxsw_resume(struct nvkm_device *);
 u32 nvkm_gr_ctxsw_inst(struct nvkm_device *);
 
-int nv04_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv10_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv15_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv17_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv20_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv25_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv2a_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv30_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv34_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv35_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv40_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv44_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int nv50_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int g84_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gt200_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int mcp79_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gt215_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int mcp89_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gf100_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gf104_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gf108_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gf110_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gf117_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gf119_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gk104_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gk110_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gk110b_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gk208_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gk20a_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gm107_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gm200_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gm20b_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gp100_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gp102_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gp104_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gp107_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gp108_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gp10b_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int gv100_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
-int tu102_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
+int nv04_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv10_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv15_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv17_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv20_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv25_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv2a_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv30_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv34_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv35_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv40_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv44_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int nv50_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int g84_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gt200_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int mcp79_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gt215_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int mcp89_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gf100_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gf104_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gf108_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gf110_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gf117_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gf119_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gk104_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gk110_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gk110b_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gk208_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gk20a_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gm107_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gm200_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gm20b_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gp100_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gp102_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gp104_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gp107_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gp108_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gp10b_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int gv100_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
+int tu102_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/mpeg.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/mpeg.h
index 8585a31f5943..f137f277935f 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/mpeg.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/mpeg.h
@@ -2,9 +2,9 @@
 #ifndef __NVKM_MPEG_H__
 #define __NVKM_MPEG_H__
 #include <core/engine.h>
-int nv31_mpeg_new(struct nvkm_device *, int index, struct nvkm_engine **);
-int nv40_mpeg_new(struct nvkm_device *, int index, struct nvkm_engine **);
-int nv44_mpeg_new(struct nvkm_device *, int index, struct nvkm_engine **);
-int nv50_mpeg_new(struct nvkm_device *, int index, struct nvkm_engine **);
-int g84_mpeg_new(struct nvkm_device *, int index, struct nvkm_engine **);
+int nv31_mpeg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int nv40_mpeg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int nv44_mpeg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int nv50_mpeg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int g84_mpeg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/mspdec.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/mspdec.h
index 83bb2fcb2cbf..ac8f08ce183c 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/mspdec.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/mspdec.h
@@ -2,8 +2,8 @@
 #ifndef __NVKM_MSPDEC_H__
 #define __NVKM_MSPDEC_H__
 #include <engine/falcon.h>
-int g98_mspdec_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gt215_mspdec_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gf100_mspdec_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gk104_mspdec_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g98_mspdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gt215_mspdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gf100_mspdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gk104_mspdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/msppp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/msppp.h
index 69e09fd96e0c..81c2b6f0ad84 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/msppp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/msppp.h
@@ -2,7 +2,7 @@
 #ifndef __NVKM_MSPPP_H__
 #define __NVKM_MSPPP_H__
 #include <engine/falcon.h>
-int g98_msppp_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gt215_msppp_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gf100_msppp_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g98_msppp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gt215_msppp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gf100_msppp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/msvld.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/msvld.h
index 9e11cefc9649..2d5fa961ba66 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/msvld.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/msvld.h
@@ -2,9 +2,9 @@
 #ifndef __NVKM_MSVLD_H__
 #define __NVKM_MSVLD_H__
 #include <engine/falcon.h>
-int g98_msvld_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gt215_msvld_new(struct nvkm_device *, int, struct nvkm_engine **);
-int mcp89_msvld_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gf100_msvld_new(struct nvkm_device *, int, struct nvkm_engine **);
-int gk104_msvld_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g98_msvld_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gt215_msvld_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int mcp89_msvld_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gf100_msvld_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
+int gk104_msvld_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h
index 1b3183e31606..97bd3092f68a 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h
@@ -11,5 +11,5 @@ struct nvkm_nvdec {
 	struct nvkm_falcon falcon;
 };
 
-int gm107_nvdec_new(struct nvkm_device *, int, struct nvkm_nvdec **);
+int gm107_nvdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvdec **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h
index 33e6ba8adc8d..1a259c5c9a71 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h
@@ -11,5 +11,5 @@ struct nvkm_nvenc {
 	struct nvkm_falcon falcon;
 };
 
-int gm107_nvenc_new(struct nvkm_device *, int, struct nvkm_nvenc **);
+int gm107_nvenc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvenc **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/pm.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/pm.h
index 4d754e7650d9..af89d46ea360 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/pm.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/pm.h
@@ -7,20 +7,23 @@ struct nvkm_pm {
 	const struct nvkm_pm_func *func;
 	struct nvkm_engine engine;
 
-	struct nvkm_object *perfmon;
+	struct {
+		spinlock_t lock;
+		struct nvkm_object *object;
+	} client;
 
 	struct list_head domains;
 	struct list_head sources;
 	u32 sequence;
 };
 
-int nv40_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int nv50_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int g84_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int gt200_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int gt215_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int gf100_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int gf108_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int gf117_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
-int gk104_pm_new(struct nvkm_device *, int, struct nvkm_pm **);
+int nv40_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int nv50_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int g84_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int gt200_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int gt215_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int gf100_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int gf108_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int gf117_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
+int gk104_pm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pm **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/sec.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/sec.h
index f14e98a8a0ca..37ed7ab8d050 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/sec.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/sec.h
@@ -2,5 +2,5 @@
 #ifndef __NVKM_SEC_H__
 #define __NVKM_SEC_H__
 #include <engine/falcon.h>
-int g98_sec_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g98_sec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/sec2.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/sec2.h
index 34dc765648d5..06264c840eae 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/sec2.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/sec2.h
@@ -18,7 +18,7 @@ struct nvkm_sec2 {
 	bool initmsg_received;
 };
 
-int gp102_sec2_new(struct nvkm_device *, int, struct nvkm_sec2 **);
-int gp108_sec2_new(struct nvkm_device *, int, struct nvkm_sec2 **);
-int tu102_sec2_new(struct nvkm_device *, int, struct nvkm_sec2 **);
+int gp102_sec2_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sec2 **);
+int gp108_sec2_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sec2 **);
+int tu102_sec2_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sec2 **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/sw.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/sw.h
index 2e91769e3ee2..b1a53ffbfdef 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/sw.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/sw.h
@@ -12,8 +12,8 @@ struct nvkm_sw {
 
 bool nvkm_sw_mthd(struct nvkm_sw *sw, int chid, int subc, u32 mthd, u32 data);
 
-int nv04_sw_new(struct nvkm_device *, int, struct nvkm_sw **);
-int nv10_sw_new(struct nvkm_device *, int, struct nvkm_sw **);
-int nv50_sw_new(struct nvkm_device *, int, struct nvkm_sw **);
-int gf100_sw_new(struct nvkm_device *, int, struct nvkm_sw **);
+int nv04_sw_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sw **);
+int nv10_sw_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sw **);
+int nv50_sw_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sw **);
+int gf100_sw_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_sw **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/vp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/vp.h
index 8984415b2a3d..1bab26858538 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/vp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/vp.h
@@ -2,5 +2,5 @@
 #ifndef __NVKM_VP_H__
 #define __NVKM_VP_H__
 #include <engine/xtensa.h>
-int g84_vp_new(struct nvkm_device *, int, struct nvkm_engine **);
+int g84_vp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/xtensa.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/xtensa.h
index fbf27b2293a9..3083a5866a55 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/xtensa.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/xtensa.h
@@ -13,7 +13,7 @@ struct nvkm_xtensa {
 };
 
 int nvkm_xtensa_new_(const struct nvkm_xtensa_func *, struct nvkm_device *,
-		     int index, bool enable, u32 addr, struct nvkm_engine **);
+		     enum nvkm_subdev_type, int, bool enable, u32 addr, struct nvkm_engine **);
 
 struct nvkm_xtensa_func {
 	u32 fifo_val;
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/acr.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/acr.h
index 836d8b932822..c0b254f7f0b5 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/acr.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/acr.h
@@ -59,12 +59,12 @@ struct nvkm_acr {
 bool nvkm_acr_managed_falcon(struct nvkm_device *, enum nvkm_acr_lsf_id);
 int nvkm_acr_bootstrap_falcons(struct nvkm_device *, unsigned long mask);
 
-int gm200_acr_new(struct nvkm_device *, int, struct nvkm_acr **);
-int gm20b_acr_new(struct nvkm_device *, int, struct nvkm_acr **);
-int gp102_acr_new(struct nvkm_device *, int, struct nvkm_acr **);
-int gp108_acr_new(struct nvkm_device *, int, struct nvkm_acr **);
-int gp10b_acr_new(struct nvkm_device *, int, struct nvkm_acr **);
-int tu102_acr_new(struct nvkm_device *, int, struct nvkm_acr **);
+int gm200_acr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_acr **);
+int gm20b_acr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_acr **);
+int gp102_acr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_acr **);
+int gp108_acr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_acr **);
+int gp10b_acr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_acr **);
+int tu102_acr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_acr **);
 
 struct nvkm_acr_lsfw {
 	const struct nvkm_acr_lsf_func *func;
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h
index 14b09f7e46a5..4f07836ab984 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h
@@ -23,11 +23,11 @@ void nvkm_bar_bar2_reset(struct nvkm_device *);
 struct nvkm_vmm *nvkm_bar_bar2_vmm(struct nvkm_device *);
 void nvkm_bar_flush(struct nvkm_bar *);
 
-int nv50_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
-int g84_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
-int gf100_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
-int gk20a_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
-int gm107_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
-int gm20b_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
-int tu102_bar_new(struct nvkm_device *, int, struct nvkm_bar **);
+int nv50_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
+int g84_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
+int gf100_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
+int gk20a_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
+int gm107_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
+int gm20b_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
+int tu102_bar_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bar **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h
index f2860f8e0c2e..b61cfb077533 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h
@@ -30,5 +30,5 @@ u8  nvbios_rd08(struct nvkm_bios *, u32 addr);
 u16 nvbios_rd16(struct nvkm_bios *, u32 addr);
 u32 nvbios_rd32(struct nvkm_bios *, u32 addr);
 
-int nvkm_bios_new(struct nvkm_device *, int, struct nvkm_bios **);
+int nvkm_bios_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_bios **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h
index f5f59261ea81..d1beaad0c82b 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h
@@ -14,6 +14,7 @@ enum dcb_connector_type {
 	DCB_CONNECTOR_LVDS_SPWG = 0x41,
 	DCB_CONNECTOR_DP = 0x46,
 	DCB_CONNECTOR_eDP = 0x47,
+	DCB_CONNECTOR_mDP = 0x48,
 	DCB_CONNECTOR_HDMI_0 = 0x60,
 	DCB_CONNECTOR_HDMI_1 = 0x61,
 	DCB_CONNECTOR_HDMI_C = 0x63,
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bus.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bus.h
index ae9ad6c034fb..2ac03bbc6133 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bus.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bus.h
@@ -18,9 +18,9 @@ void nvkm_hwsq_wait(struct nvkm_hwsq *, u8 flag, u8 data);
 void nvkm_hwsq_wait_vblank(struct nvkm_hwsq *);
 void nvkm_hwsq_nsec(struct nvkm_hwsq *, u32 nsec);
 
-int nv04_bus_new(struct nvkm_device *, int, struct nvkm_bus **);
-int nv31_bus_new(struct nvkm_device *, int, struct nvkm_bus **);
-int nv50_bus_new(struct nvkm_device *, int, struct nvkm_bus **);
-int g94_bus_new(struct nvkm_device *, int, struct nvkm_bus **);
-int gf100_bus_new(struct nvkm_device *, int, struct nvkm_bus **);
+int nv04_bus_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bus **);
+int nv31_bus_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bus **);
+int nv50_bus_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bus **);
+int g94_bus_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bus **);
+int gf100_bus_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_bus **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/clk.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/clk.h
index bf937e7dfd77..05b99c9e9a26 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/clk.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/clk.h
@@ -125,14 +125,14 @@ int nvkm_clk_astate(struct nvkm_clk *, int req, int rel, bool wait);
 int nvkm_clk_dstate(struct nvkm_clk *, int req, int rel);
 int nvkm_clk_tstate(struct nvkm_clk *, u8 temperature);
 
-int nv04_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int nv40_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int nv50_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int g84_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int mcp77_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int gt215_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int gf100_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int gk104_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int gk20a_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
-int gm20b_clk_new(struct nvkm_device *, int, struct nvkm_clk **);
+int nv04_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int nv40_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int nv50_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int g84_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int mcp77_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int gt215_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int gf100_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int gk104_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int gk20a_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
+int gm20b_clk_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_clk **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h
index 1a39e52e09e3..848b5d9ce705 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h
@@ -14,22 +14,22 @@ struct nvkm_devinit {
 u32 nvkm_devinit_mmio(struct nvkm_devinit *, u32 addr);
 int nvkm_devinit_pll_set(struct nvkm_devinit *, u32 type, u32 khz);
 void nvkm_devinit_meminit(struct nvkm_devinit *);
-u64 nvkm_devinit_disable(struct nvkm_devinit *);
-int nvkm_devinit_post(struct nvkm_devinit *, u64 *disable);
+int nvkm_devinit_post(struct nvkm_devinit *);
 
-int nv04_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int nv05_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int nv10_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int nv1a_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int nv20_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int nv50_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int g84_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int g98_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int gt215_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int mcp89_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int gf100_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int gm107_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int gm200_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int gv100_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
-int tu102_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **);
+int nv04_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int nv05_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int nv10_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int nv1a_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int nv20_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int nv50_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int g84_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int g98_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int gt215_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int mcp89_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int gf100_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int gm107_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int gm200_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int gv100_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int tu102_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
+int ga100_devinit_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_devinit **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h
index a513c16ab105..581458ad38e0 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h
@@ -30,8 +30,8 @@ struct nvkm_fault_data {
 	u8 reason;
 };
 
-int gp100_fault_new(struct nvkm_device *, int, struct nvkm_fault **);
-int gp10b_fault_new(struct nvkm_device *, int, struct nvkm_fault **);
-int gv100_fault_new(struct nvkm_device *, int, struct nvkm_fault **);
-int tu102_fault_new(struct nvkm_device *, int, struct nvkm_fault **);
+int gp100_fault_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fault **);
+int gp10b_fault_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fault **);
+int gv100_fault_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fault **);
+int tu102_fault_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fault **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h
index 34b56b10218a..ef6a6297148c 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h
@@ -36,7 +36,11 @@ struct nvkm_fb {
 	struct nvkm_blob vpr_scrubber;
 
 	struct nvkm_ram *ram;
-	struct nvkm_mm tags;
+
+	struct {
+		struct mutex mutex; /* protects mm and nvkm_memory::tags */
+		struct nvkm_mm mm;
+	} tags;
 
 	struct {
 		struct nvkm_fb_tile region[16];
@@ -54,38 +58,40 @@ void nvkm_fb_tile_init(struct nvkm_fb *, int region, u32 addr, u32 size,
 void nvkm_fb_tile_fini(struct nvkm_fb *, int region, struct nvkm_fb_tile *);
 void nvkm_fb_tile_prog(struct nvkm_fb *, int region, struct nvkm_fb_tile *);
 
-int nv04_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv10_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv1a_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv20_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv25_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv30_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv35_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv36_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv40_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv41_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv44_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv46_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv47_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv49_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv4e_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int nv50_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int g84_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gt215_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int mcp77_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int mcp89_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gf100_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gf108_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gk104_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gk110_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gk20a_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gm107_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gm200_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gm20b_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gp100_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gp102_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gp10b_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
-int gv100_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
+int nv04_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv10_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv1a_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv20_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv25_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv30_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv35_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv36_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv40_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv41_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv44_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv46_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv47_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv49_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv4e_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int nv50_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int g84_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gt215_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int mcp77_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int mcp89_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gf100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gf108_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gk104_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gk110_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gk20a_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gm107_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gm200_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gm20b_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gp100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gp102_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gp10b_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int gv100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int ga100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
+int ga102_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **);
 
 #include <subdev/bios.h>
 #include <subdev/bios/ramcfg.h>
@@ -126,6 +132,7 @@ struct nvkm_ram {
 #define NVKM_RAM_MM_MIXED  (NVKM_MM_HEAP_ANY + 3)
 	struct nvkm_mm vram;
 	u64 stolen;
+	struct mutex mutex;
 
 	int ranks;
 	int parts;
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fuse.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fuse.h
index 00111c34311e..dabbef0ac96c 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fuse.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fuse.h
@@ -11,7 +11,7 @@ struct nvkm_fuse {
 
 u32 nvkm_fuse_read(struct nvkm_fuse *, u32 addr);
 
-int nv50_fuse_new(struct nvkm_device *, int, struct nvkm_fuse **);
-int gf100_fuse_new(struct nvkm_device *, int, struct nvkm_fuse **);
-int gm107_fuse_new(struct nvkm_device *, int, struct nvkm_fuse **);
+int nv50_fuse_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fuse **);
+int gf100_fuse_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fuse **);
+int gm107_fuse_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fuse **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/gpio.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/gpio.h
index eaacf8d80527..0e46ea1fe972 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/gpio.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/gpio.h
@@ -32,9 +32,10 @@ int nvkm_gpio_find(struct nvkm_gpio *, int idx, u8 tag, u8 line,
 int nvkm_gpio_set(struct nvkm_gpio *, int idx, u8 tag, u8 line, int state);
 int nvkm_gpio_get(struct nvkm_gpio *, int idx, u8 tag, u8 line);
 
-int nv10_gpio_new(struct nvkm_device *, int, struct nvkm_gpio **);
-int nv50_gpio_new(struct nvkm_device *, int, struct nvkm_gpio **);
-int g94_gpio_new(struct nvkm_device *, int, struct nvkm_gpio **);
-int gf119_gpio_new(struct nvkm_device *, int, struct nvkm_gpio **);
-int gk104_gpio_new(struct nvkm_device *, int, struct nvkm_gpio **);
+int nv10_gpio_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gpio **);
+int nv50_gpio_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gpio **);
+int g94_gpio_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gpio **);
+int gf119_gpio_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gpio **);
+int gk104_gpio_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gpio **);
+int ga102_gpio_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gpio **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h
index 06db67610a50..cf42a59d4e58 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h
@@ -9,5 +9,5 @@ struct nvkm_gsp {
 	struct nvkm_falcon falcon;
 };
 
-int gv100_gsp_new(struct nvkm_device *, int, struct nvkm_gsp **);
+int gv100_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/i2c.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/i2c.h
index 81b977319640..146e13292203 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/i2c.h
@@ -85,14 +85,15 @@ struct nvkm_i2c {
 struct nvkm_i2c_bus *nvkm_i2c_bus_find(struct nvkm_i2c *, int);
 struct nvkm_i2c_aux *nvkm_i2c_aux_find(struct nvkm_i2c *, int);
 
-int nv04_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int nv4e_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int nv50_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int g94_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int gf117_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int gf119_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int gk104_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
-int gm200_i2c_new(struct nvkm_device *, int, struct nvkm_i2c **);
+int nv04_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int nv4e_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int nv50_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int g94_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int gf117_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int gf119_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int gk104_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int gk110_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
+int gm200_i2c_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_i2c **);
 
 static inline int
 nvkm_rdi2cr(struct i2c_adapter *adap, u8 addr, u8 reg)
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ibus.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ibus.h
deleted file mode 100644
index db791411eaa8..000000000000
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ibus.h
+++ /dev/null
@@ -1,12 +0,0 @@
-/* SPDX-License-Identifier: MIT */
-#ifndef __NVKM_IBUS_H__
-#define __NVKM_IBUS_H__
-#include <core/subdev.h>
-
-int gf100_ibus_new(struct nvkm_device *, int, struct nvkm_subdev **);
-int gf117_ibus_new(struct nvkm_device *, int, struct nvkm_subdev **);
-int gk104_ibus_new(struct nvkm_device *, int, struct nvkm_subdev **);
-int gk20a_ibus_new(struct nvkm_device *, int, struct nvkm_subdev **);
-int gm200_ibus_new(struct nvkm_device *, int, struct nvkm_subdev **);
-int gp10b_ibus_new(struct nvkm_device *, int, struct nvkm_subdev **);
-#endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/iccsense.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/iccsense.h
index f483dcd7cd1c..7400d62dcbec 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/iccsense.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/iccsense.h
@@ -14,6 +14,6 @@ struct nvkm_iccsense {
 	u32 power_w_crit;
 };
 
-int gf100_iccsense_new(struct nvkm_device *, int index, struct nvkm_iccsense **);
+int gf100_iccsense_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_iccsense **);
 int nvkm_iccsense_read_all(struct nvkm_iccsense *iccsense);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h
index c74ab7c31d05..f967b97d163c 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h
@@ -13,6 +13,11 @@ struct nvkm_instmem {
 	struct list_head boot;
 	u32 reserved;
 
+	/* <=nv4x: protects NV_PRAMIN/BAR2 MM
+	 * >=nv50: protects BAR2 MM & LRU
+	 */
+	struct mutex mutex;
+
 	struct nvkm_memory *vbios;
 	struct nvkm_ramht  *ramht;
 	struct nvkm_memory *ramro;
@@ -25,8 +30,8 @@ int nvkm_instobj_new(struct nvkm_instmem *, u32 size, u32 align, bool zero,
 		     struct nvkm_memory **);
 
 
-int nv04_instmem_new(struct nvkm_device *, int, struct nvkm_instmem **);
-int nv40_instmem_new(struct nvkm_device *, int, struct nvkm_instmem **);
-int nv50_instmem_new(struct nvkm_device *, int, struct nvkm_instmem **);
-int gk20a_instmem_new(struct nvkm_device *, int, struct nvkm_instmem **);
+int nv04_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **);
+int nv40_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **);
+int nv50_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **);
+int gk20a_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h
index d76f60d7d29a..d32a326a9290 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h
@@ -13,6 +13,7 @@ struct nvkm_ltc {
 	u32 ltc_nr;
 	u32 lts_nr;
 
+	struct mutex mutex; /* serialises CBC operations */
 	u32 num_tags;
 	u32 tag_base;
 	struct nvkm_memory *tag_ram;
@@ -33,12 +34,11 @@ int nvkm_ltc_zbc_stencil_get(struct nvkm_ltc *, int index, const u32);
 void nvkm_ltc_invalidate(struct nvkm_ltc *);
 void nvkm_ltc_flush(struct nvkm_ltc *);
 
-int gf100_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gk104_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gk20a_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gm107_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gm200_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gp100_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gp102_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
-int gp10b_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
+int gf100_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
+int gk104_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
+int gm107_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
+int gm200_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
+int gp100_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
+int gp102_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
+int gp10b_ltc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_ltc **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h
index 6641fe4c252c..cb86a56e68d4 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h
@@ -8,28 +8,29 @@ struct nvkm_mc {
 	struct nvkm_subdev subdev;
 };
 
-void nvkm_mc_enable(struct nvkm_device *, enum nvkm_devidx);
-void nvkm_mc_disable(struct nvkm_device *, enum nvkm_devidx);
-bool nvkm_mc_enabled(struct nvkm_device *, enum nvkm_devidx);
-void nvkm_mc_reset(struct nvkm_device *, enum nvkm_devidx);
+void nvkm_mc_enable(struct nvkm_device *, enum nvkm_subdev_type, int);
+void nvkm_mc_disable(struct nvkm_device *, enum nvkm_subdev_type, int);
+bool nvkm_mc_enabled(struct nvkm_device *, enum nvkm_subdev_type, int);
+void nvkm_mc_reset(struct nvkm_device *, enum nvkm_subdev_type, int);
 void nvkm_mc_intr(struct nvkm_device *, bool *handled);
 void nvkm_mc_intr_unarm(struct nvkm_device *);
 void nvkm_mc_intr_rearm(struct nvkm_device *);
-void nvkm_mc_intr_mask(struct nvkm_device *, enum nvkm_devidx, bool enable);
+void nvkm_mc_intr_mask(struct nvkm_device *, enum nvkm_subdev_type, int, bool enable);
 void nvkm_mc_unk260(struct nvkm_device *, u32 data);
 
-int nv04_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int nv11_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int nv17_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int nv44_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int nv50_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int g84_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int g98_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int gt215_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int gf100_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int gk104_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int gk20a_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int gp100_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int gp10b_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
-int tu102_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
+int nv04_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int nv11_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int nv17_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int nv44_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int nv50_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int g84_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int g98_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int gt215_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int gf100_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int gk104_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int gk20a_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int gp100_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int gp10b_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int tu102_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
+int ga100_mc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mc **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h
index 54cdcb017518..0911e73f7424 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h
@@ -117,22 +117,24 @@ struct nvkm_mmu {
 		struct list_head list;
 	} ptc, ptp;
 
+	struct mutex mutex; /* serialises mmu invalidations */
+
 	struct nvkm_device_oclass user;
 };
 
-int nv04_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int nv41_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int nv44_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int nv50_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int g84_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int mcp77_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gf100_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gk104_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gk20a_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gm200_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gm20b_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gp100_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gp10b_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int gv100_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
-int tu102_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **);
+int nv04_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int nv41_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int nv44_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int nv50_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int g84_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int mcp77_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gf100_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gk104_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gk20a_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gm200_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gm20b_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gp100_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gp10b_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int gv100_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
+int tu102_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mxm.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mxm.h
index 78df1e9def05..7d4132a17d0f 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mxm.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mxm.h
@@ -3,5 +3,5 @@
 #define __NVKM_MXM_H__
 #include <core/subdev.h>
 
-int nv50_mxm_new(struct nvkm_device *, int, struct nvkm_subdev **);
+int nv50_mxm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h
index 4803a4fad4a2..74c19bdfb757 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h
@@ -39,17 +39,17 @@ void nvkm_pci_wr32(struct nvkm_pci *, u16 addr, u32 data);
 u32 nvkm_pci_mask(struct nvkm_pci *, u16 addr, u32 mask, u32 value);
 void nvkm_pci_rom_shadow(struct nvkm_pci *, bool shadow);
 
-int nv04_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int nv40_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int nv46_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int nv4c_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int g84_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int g92_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int g94_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int gf100_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int gf106_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int gk104_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
-int gp100_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
+int nv04_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int nv40_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int nv46_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int nv4c_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int g84_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int g92_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int g94_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int gf100_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int gf106_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int gk104_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
+int gp100_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **);
 
 /* pcie functions */
 int nvkm_pcie_set_link(struct nvkm_pci *, enum nvkm_pcie_speed, u8 width);
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pmu.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pmu.h
index 5ff6d1f8985a..f57a3a5a288d 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pmu.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pmu.h
@@ -18,6 +18,7 @@ struct nvkm_pmu {
 	struct completion wpr_ready;
 
 	struct {
+		struct mutex mutex;
 		u32 base;
 		u32 size;
 	} send;
@@ -39,18 +40,18 @@ int nvkm_pmu_send(struct nvkm_pmu *, u32 reply[2], u32 process,
 void nvkm_pmu_pgob(struct nvkm_pmu *, bool enable);
 bool nvkm_pmu_fan_controlled(struct nvkm_device *);
 
-int gt215_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gf100_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gf119_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gk104_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gk110_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gk208_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gk20a_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gm107_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gm200_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gm20b_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gp102_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
-int gp10b_pmu_new(struct nvkm_device *, int, struct nvkm_pmu **);
+int gt215_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gf100_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gf119_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gk104_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gk110_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gk208_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gk20a_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gm107_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gm200_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gm20b_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gp102_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
+int gp10b_pmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pmu **);
 
 /* interface to MEMX process running on PMU */
 struct nvkm_memx;
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/privring.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/privring.h
new file mode 100644
index 000000000000..e1399f8a90ad
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/privring.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: MIT */
+#ifndef __NVKM_PRIVRING_H__
+#define __NVKM_PRIVRING_H__
+#include <core/subdev.h>
+
+int gf100_privring_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
+int gf117_privring_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
+int gk104_privring_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
+int gk20a_privring_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
+int gm200_privring_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
+int gp10b_privring_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_subdev **);
+#endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/therm.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/therm.h
index 62c34f98c930..bd04f49272a6 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/therm.h
@@ -107,13 +107,13 @@ void nvkm_therm_clkgate_init(struct nvkm_therm *,
 void nvkm_therm_clkgate_enable(struct nvkm_therm *);
 void nvkm_therm_clkgate_fini(struct nvkm_therm *, bool);
 
-int nv40_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int nv50_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int g84_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int gt215_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int gf119_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int gk104_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int gm107_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int gm200_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
-int gp100_therm_new(struct nvkm_device *, int, struct nvkm_therm **);
+int nv40_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int nv50_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int g84_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int gt215_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int gf119_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int gk104_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int gm107_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int gm200_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
+int gp100_therm_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_therm **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h
index d06dcbe1faa6..439a3f72b0d7 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h
@@ -76,8 +76,8 @@ s64 nvkm_timer_wait_test(struct nvkm_timer_wait *);
 #define nvkm_wait_msec(d,m,addr,mask,data)                                     \
 	nvkm_wait_usec((d), (m) * 1000, (addr), (mask), (data))
 
-int nv04_timer_new(struct nvkm_device *, int, struct nvkm_timer **);
-int nv40_timer_new(struct nvkm_device *, int, struct nvkm_timer **);
-int nv41_timer_new(struct nvkm_device *, int, struct nvkm_timer **);
-int gk20a_timer_new(struct nvkm_device *, int, struct nvkm_timer **);
+int nv04_timer_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_timer **);
+int nv40_timer_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_timer **);
+int nv41_timer_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_timer **);
+int gk20a_timer_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_timer **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h
index 7be0e7e7bd77..ee75c5524c43 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h
@@ -9,13 +9,24 @@ struct nvkm_top {
 	struct list_head device;
 };
 
-u32 nvkm_top_addr(struct nvkm_device *, enum nvkm_devidx);
-u32 nvkm_top_reset(struct nvkm_device *, enum nvkm_devidx);
-u32 nvkm_top_intr(struct nvkm_device *, u32 intr, u64 *subdevs);
-u32 nvkm_top_intr_mask(struct nvkm_device *, enum nvkm_devidx);
-int nvkm_top_fault_id(struct nvkm_device *, enum nvkm_devidx);
-enum nvkm_devidx nvkm_top_fault(struct nvkm_device *, int fault);
-enum nvkm_devidx nvkm_top_engine(struct nvkm_device *, int, int *runl, int *engn);
+struct nvkm_top_device {
+	enum nvkm_subdev_type type;
+	int inst;
+	u32 addr;
+	int fault;
+	int engine;
+	int runlist;
+	int reset;
+	int intr;
+	struct list_head head;
+};
+
+u32 nvkm_top_addr(struct nvkm_device *, enum nvkm_subdev_type, int);
+u32 nvkm_top_reset(struct nvkm_device *, enum nvkm_subdev_type, int);
+u32 nvkm_top_intr_mask(struct nvkm_device *, enum nvkm_subdev_type, int);
+int nvkm_top_fault_id(struct nvkm_device *, enum nvkm_subdev_type, int);
+struct nvkm_subdev *nvkm_top_fault(struct nvkm_device *, int fault);
 
-int gk104_top_new(struct nvkm_device *, int, struct nvkm_top **);
+int gk104_top_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_top **);
+int ga100_top_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_top **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h
index 45053a280930..0be86d5f0158 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h
@@ -36,10 +36,10 @@ int nvkm_volt_get(struct nvkm_volt *);
 int nvkm_volt_set_id(struct nvkm_volt *, u8 id, u8 min_id, u8 temp,
 		     int condition);
 
-int nv40_volt_new(struct nvkm_device *, int, struct nvkm_volt **);
-int gf100_volt_new(struct nvkm_device *, int, struct nvkm_volt **);
-int gf117_volt_new(struct nvkm_device *, int, struct nvkm_volt **);
-int gk104_volt_new(struct nvkm_device *, int, struct nvkm_volt **);
-int gk20a_volt_new(struct nvkm_device *, int, struct nvkm_volt **);
-int gm20b_volt_new(struct nvkm_device *, int, struct nvkm_volt **);
+int nv40_volt_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_volt **);
+int gf100_volt_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_volt **);
+int gf117_volt_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_volt **);
+int gk104_volt_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_volt **);
+int gk20a_volt_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_volt **);
+int gm20b_volt_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_volt **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index f08b31d84d4d..0a9334deffe2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -269,19 +269,19 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	if (device->info.family >= NV_DEVICE_INFO_V0_KEPLER) {
 		if (init->fb_ctxdma_handle == ~0) {
 			switch (init->tt_ctxdma_handle) {
-			case 0x01: engine = NV_DEVICE_INFO_ENGINE_GR    ; break;
-			case 0x02: engine = NV_DEVICE_INFO_ENGINE_MSPDEC; break;
-			case 0x04: engine = NV_DEVICE_INFO_ENGINE_MSPPP ; break;
-			case 0x08: engine = NV_DEVICE_INFO_ENGINE_MSVLD ; break;
-			case 0x30: engine = NV_DEVICE_INFO_ENGINE_CE    ; break;
+			case 0x01: engine = NV_DEVICE_HOST_RUNLIST_ENGINES_GR    ; break;
+			case 0x02: engine = NV_DEVICE_HOST_RUNLIST_ENGINES_MSPDEC; break;
+			case 0x04: engine = NV_DEVICE_HOST_RUNLIST_ENGINES_MSPPP ; break;
+			case 0x08: engine = NV_DEVICE_HOST_RUNLIST_ENGINES_MSVLD ; break;
+			case 0x30: engine = NV_DEVICE_HOST_RUNLIST_ENGINES_CE    ; break;
 			default:
 				return nouveau_abi16_put(abi16, -ENOSYS);
 			}
 		} else {
-			engine = NV_DEVICE_INFO_ENGINE_GR;
+			engine = NV_DEVICE_HOST_RUNLIST_ENGINES_GR;
 		}
 
-		if (engine != NV_DEVICE_INFO_ENGINE_CE)
+		if (engine != NV_DEVICE_HOST_RUNLIST_ENGINES_CE)
 			engine = nvif_fifo_runlist(device, engine);
 		else
 			engine = nvif_fifo_runlist_ce(device);
diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c
index c7a94c94dbf3..72f35a2babcb 100644
--- a/drivers/gpu/drm/nouveau/nouveau_backlight.c
+++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c
@@ -256,6 +256,7 @@ nouveau_backlight_init(struct drm_connector *connector)
 	case NV_DEVICE_INFO_V0_PASCAL:
 	case NV_DEVICE_INFO_V0_VOLTA:
 	case NV_DEVICE_INFO_V0_TURING:
+	case NV_DEVICE_INFO_V0_AMPERE: //XXX: not confirmed
 		ret = nv50_backlight_init(nv_encoder, &props, &ops);
 		break;
 	default:
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index ca2a8ae1938e..281e9ed13989 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -544,7 +544,7 @@ nouveau_bo_sync_for_device(struct nouveau_bo *nvbo)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_tt *ttm_dma = (struct ttm_tt *)nvbo->bo.ttm;
-	int i;
+	int i, j;
 
 	if (!ttm_dma)
 		return;
@@ -553,10 +553,22 @@ nouveau_bo_sync_for_device(struct nouveau_bo *nvbo)
 	if (nvbo->force_coherent)
 		return;
 
-	for (i = 0; i < ttm_dma->num_pages; i++)
+	i = 0;
+	while (i < ttm_dma->num_pages) {
+		struct page *p = ttm_dma->pages[i];
+		size_t num_pages = 1;
+
+		for (j = i + 1; j < ttm_dma->num_pages; ++j) {
+			if (++p != ttm_dma->pages[j])
+				break;
+
+			++num_pages;
+		}
 		dma_sync_single_for_device(drm->dev->dev,
 					   ttm_dma->dma_address[i],
-					   PAGE_SIZE, DMA_TO_DEVICE);
+					   num_pages * PAGE_SIZE, DMA_TO_DEVICE);
+		i += num_pages;
+	}
 }
 
 void
@@ -564,7 +576,7 @@ nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_tt *ttm_dma = (struct ttm_tt *)nvbo->bo.ttm;
-	int i;
+	int i, j;
 
 	if (!ttm_dma)
 		return;
@@ -573,9 +585,22 @@ nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo)
 	if (nvbo->force_coherent)
 		return;
 
-	for (i = 0; i < ttm_dma->num_pages; i++)
+	i = 0;
+	while (i < ttm_dma->num_pages) {
+		struct page *p = ttm_dma->pages[i];
+		size_t num_pages = 1;
+
+		for (j = i + 1; j < ttm_dma->num_pages; ++j) {
+			if (++p != ttm_dma->pages[j])
+				break;
+
+			++num_pages;
+		}
+
 		dma_sync_single_for_cpu(drm->dev->dev, ttm_dma->dma_address[i],
-					PAGE_SIZE, DMA_FROM_DEVICE);
+					num_pages * PAGE_SIZE, DMA_FROM_DEVICE);
+		i += num_pages;
+	}
 }
 
 void nouveau_bo_add_io_reserve_lru(struct ttm_buffer_object *bo)
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index 5d191e58edf1..7cfac265fd45 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -533,6 +533,7 @@ nouveau_channel_new(struct nouveau_drm *drm, struct nvif_device *device,
 	if (ret) {
 		NV_PRINTK(err, cli, "channel failed to initialise, %d\n", ret);
 		nouveau_channel_del(pchan);
+		goto done;
 	}
 
 	ret = nouveau_svmm_join((*pchan)->vmm->svmm, (*pchan)->inst);
@@ -555,7 +556,7 @@ nouveau_channels_init(struct nouveau_drm *drm)
 	} args = {
 		.m.version = 1,
 		.m.count = sizeof(args.v) / sizeof(args.v.channels),
-		.v.channels.mthd = NV_DEVICE_FIFO_CHANNELS,
+		.v.channels.mthd = NV_DEVICE_HOST_CHANNELS,
 	};
 	struct nvif_object *device = &drm->client.device.object;
 	int ret;
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 14c29e68db8f..61e6d7412505 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -1212,6 +1212,7 @@ drm_conntype_from_dcb(enum dcb_connector_type dcb)
 	case DCB_CONNECTOR_DMS59_DP0:
 	case DCB_CONNECTOR_DMS59_DP1:
 	case DCB_CONNECTOR_DP       :
+	case DCB_CONNECTOR_mDP      :
 	case DCB_CONNECTOR_USB_C    : return DRM_MODE_CONNECTOR_DisplayPort;
 	case DCB_CONNECTOR_eDP      : return DRM_MODE_CONNECTOR_eDP;
 	case DCB_CONNECTOR_HDMI_0   :
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 1b2169e9c295..885815ea917f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -344,7 +344,7 @@ nouveau_accel_gr_init(struct nouveau_drm *drm)
 
 	/* Allocate channel that has access to the graphics engine. */
 	if (device->info.family >= NV_DEVICE_INFO_V0_KEPLER) {
-		arg0 = nvif_fifo_runlist(device, NV_DEVICE_INFO_ENGINE_GR);
+		arg0 = nvif_fifo_runlist(device, NV_DEVICE_HOST_RUNLIST_ENGINES_GR);
 		arg1 = 1;
 	} else {
 		arg0 = NvDmaFB;
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index a491c2c1c56e..ba65f136cf48 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -220,6 +220,7 @@ struct nouveau_drm {
 
 	struct {
 		struct drm_audio_component *component;
+		struct mutex lock;
 		bool component_registered;
 	} audio;
 };
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index 21937f1c7dd9..1ffcc0a491fd 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -53,7 +53,12 @@ struct nouveau_encoder {
 	 * actually programmed on the hw, not the proposed crtc */
 	struct drm_crtc *crtc;
 	u32 ctrl;
-	bool audio;
+
+	/* Protected by nouveau_drm.audio.lock */
+	struct {
+		bool enabled;
+		struct drm_connector *connector;
+	} audio;
 
 	struct drm_display_mode mode;
 	int last_dpms;
@@ -141,11 +146,9 @@ enum drm_mode_status nv50_dp_mode_valid(struct drm_connector *,
 					unsigned *clock);
 
 struct nouveau_connector *
-nv50_outp_get_new_connector(struct nouveau_encoder *outp,
-			    struct drm_atomic_state *state);
+nv50_outp_get_new_connector(struct drm_atomic_state *state, struct nouveau_encoder *outp);
 struct nouveau_connector *
-nv50_outp_get_old_connector(struct nouveau_encoder *outp,
-			    struct drm_atomic_state *state);
+nv50_outp_get_old_connector(struct drm_atomic_state *state, struct nouveau_encoder *outp);
 
 int nv50_mstm_detect(struct nouveau_encoder *encoder);
 void nv50_mstm_remove(struct nv50_mstm *mstm);
diff --git a/drivers/gpu/drm/nouveau/nouveau_svm.c b/drivers/gpu/drm/nouveau/nouveau_svm.c
index 4f69e4c3dafd..1c3f890377d2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_svm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -315,6 +315,10 @@ nouveau_svmm_init(struct drm_device *dev, void *data,
 	struct drm_nouveau_svm_init *args = data;
 	int ret;
 
+	/* We need to fail if svm is disabled */
+	if (!cli->drm->svm)
+		return -ENOSYS;
+
 	/* Allocate tracking for SVM-enabled VMM. */
 	if (!(svmm = kzalloc(sizeof(*svmm), GFP_KERNEL)))
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/nouveau/nvif/disp.c b/drivers/gpu/drm/nouveau/nvif/disp.c
index 8d0d30e08f57..529cb60d5efb 100644
--- a/drivers/gpu/drm/nouveau/nvif/disp.c
+++ b/drivers/gpu/drm/nouveau/nvif/disp.c
@@ -35,6 +35,7 @@ nvif_disp_ctor(struct nvif_device *device, const char *name, s32 oclass,
 	       struct nvif_disp *disp)
 {
 	static const struct nvif_mclass disps[] = {
+		{ GA102_DISP, -1 },
 		{ TU102_DISP, -1 },
 		{ GV100_DISP, -1 },
 		{ GP102_DISP, -1 },
diff --git a/drivers/gpu/drm/nouveau/nvif/fifo.c b/drivers/gpu/drm/nouveau/nvif/fifo.c
index e84a2e2ff043..a463289962b2 100644
--- a/drivers/gpu/drm/nouveau/nvif/fifo.c
+++ b/drivers/gpu/drm/nouveau/nvif/fifo.c
@@ -41,9 +41,11 @@ nvif_fifo_runlists(struct nvif_device *device)
 		return -ENOMEM;
 	a->m.version = 1;
 	a->m.count = sizeof(a->v) / sizeof(a->v.runlists);
-	a->v.runlists.mthd = NV_DEVICE_FIFO_RUNLISTS;
-	for (i = 0; i < ARRAY_SIZE(a->v.runlist); i++)
-		a->v.runlist[i].mthd = NV_DEVICE_FIFO_RUNLIST_ENGINES(i);
+	a->v.runlists.mthd = NV_DEVICE_HOST_RUNLISTS;
+	for (i = 0; i < ARRAY_SIZE(a->v.runlist); i++) {
+		a->v.runlist[i].mthd = NV_DEVICE_HOST_RUNLIST_ENGINES;
+		a->v.runlist[i].data = i;
+	}
 
 	ret = nvif_object_mthd(object, NV_DEVICE_V0_INFO, a, sizeof(*a));
 	if (ret)
@@ -58,7 +60,7 @@ nvif_fifo_runlists(struct nvif_device *device)
 	}
 
 	for (i = 0; i < device->runlists; i++) {
-		if (a->v.runlists.data & BIT_ULL(i))
+		if (a->v.runlist[i].mthd != NV_DEVICE_INFO_INVALID)
 			device->runlist[i].engines = a->v.runlist[i].data;
 	}
 
@@ -70,29 +72,15 @@ done:
 u64
 nvif_fifo_runlist(struct nvif_device *device, u64 engine)
 {
-	struct nvif_object *object = &device->object;
-	struct {
-		struct nv_device_info_v1 m;
-		struct {
-			struct nv_device_info_v1_data engine;
-		} v;
-	} a = {
-		.m.version = 1,
-		.m.count = sizeof(a.v) / sizeof(a.v.engine),
-		.v.engine.mthd = engine,
-	};
 	u64 runm = 0;
 	int ret, i;
 
 	if ((ret = nvif_fifo_runlists(device)))
 		return runm;
 
-	ret = nvif_object_mthd(object, NV_DEVICE_V0_INFO, &a, sizeof(a));
-	if (ret == 0) {
-		for (i = 0; i < device->runlists; i++) {
-			if (device->runlist[i].engines & a.v.engine.data)
-				runm |= BIT_ULL(i);
-		}
+	for (i = 0; i < device->runlists; i++) {
+		if (device->runlist[i].engines & engine)
+			runm |= BIT_ULL(i);
 	}
 
 	return runm;
diff --git a/drivers/gpu/drm/nouveau/nvkm/core/engine.c b/drivers/gpu/drm/nouveau/nvkm/core/engine.c
index 1a47c40e171b..e41a39ae1597 100644
--- a/drivers/gpu/drm/nouveau/nvkm/core/engine.c
+++ b/drivers/gpu/drm/nouveau/nvkm/core/engine.c
@@ -40,10 +40,11 @@ nvkm_engine_unref(struct nvkm_engine **pengine)
 {
 	struct nvkm_engine *engine = *pengine;
 	if (engine) {
-		mutex_lock(&engine->subdev.mutex);
-		if (--engine->usecount == 0)
+		if (refcount_dec_and_mutex_lock(&engine->use.refcount, &engine->use.mutex)) {
 			nvkm_subdev_fini(&engine->subdev, false);
-		mutex_unlock(&engine->subdev.mutex);
+			engine->use.enabled = false;
+			mutex_unlock(&engine->use.mutex);
+		}
 		*pengine = NULL;
 	}
 }
@@ -51,17 +52,21 @@ nvkm_engine_unref(struct nvkm_engine **pengine)
 struct nvkm_engine *
 nvkm_engine_ref(struct nvkm_engine *engine)
 {
+	int ret;
 	if (engine) {
-		mutex_lock(&engine->subdev.mutex);
-		if (++engine->usecount == 1) {
-			int ret = nvkm_subdev_init(&engine->subdev);
-			if (ret) {
-				engine->usecount--;
-				mutex_unlock(&engine->subdev.mutex);
-				return ERR_PTR(ret);
+		if (!refcount_inc_not_zero(&engine->use.refcount)) {
+			mutex_lock(&engine->use.mutex);
+			if (!refcount_inc_not_zero(&engine->use.refcount)) {
+				engine->use.enabled = true;
+				if ((ret = nvkm_subdev_init(&engine->subdev))) {
+					engine->use.enabled = false;
+					mutex_unlock(&engine->use.mutex);
+					return ERR_PTR(ret);
+				}
+				refcount_set(&engine->use.refcount, 1);
 			}
+			mutex_unlock(&engine->use.mutex);
 		}
-		mutex_unlock(&engine->subdev.mutex);
 	}
 	return engine;
 }
@@ -114,7 +119,7 @@ nvkm_engine_init(struct nvkm_subdev *subdev)
 	int ret = 0, i;
 	s64 time;
 
-	if (!engine->usecount) {
+	if (!engine->use.enabled) {
 		nvkm_trace(subdev, "init skipped, engine has no users\n");
 		return ret;
 	}
@@ -156,11 +161,12 @@ nvkm_engine_dtor(struct nvkm_subdev *subdev)
 	struct nvkm_engine *engine = nvkm_engine(subdev);
 	if (engine->func->dtor)
 		return engine->func->dtor(engine);
+	mutex_destroy(&engine->use.mutex);
 	return engine;
 }
 
-static const struct nvkm_subdev_func
-nvkm_engine_func = {
+const struct nvkm_subdev_func
+nvkm_engine = {
 	.dtor = nvkm_engine_dtor,
 	.preinit = nvkm_engine_preinit,
 	.init = nvkm_engine_init,
@@ -170,14 +176,15 @@ nvkm_engine_func = {
 };
 
 int
-nvkm_engine_ctor(const struct nvkm_engine_func *func,
-		 struct nvkm_device *device, int index, bool enable,
-		 struct nvkm_engine *engine)
+nvkm_engine_ctor(const struct nvkm_engine_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, bool enable, struct nvkm_engine *engine)
 {
-	nvkm_subdev_ctor(&nvkm_engine_func, device, index, &engine->subdev);
+	nvkm_subdev_ctor(&nvkm_engine, device, type, inst, &engine->subdev);
 	engine->func = func;
+	refcount_set(&engine->use.refcount, 0);
+	mutex_init(&engine->use.mutex);
 
-	if (!nvkm_boolopt(device->cfgopt, nvkm_subdev_name[index], enable)) {
+	if (!nvkm_boolopt(device->cfgopt, engine->subdev.name, enable)) {
 		nvkm_debug(&engine->subdev, "disabled\n");
 		return -ENODEV;
 	}
@@ -187,11 +194,11 @@ nvkm_engine_ctor(const struct nvkm_engine_func *func,
 }
 
 int
-nvkm_engine_new_(const struct nvkm_engine_func *func,
-		 struct nvkm_device *device, int index, bool enable,
+nvkm_engine_new_(const struct nvkm_engine_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, bool enable,
 		 struct nvkm_engine **pengine)
 {
 	if (!(*pengine = kzalloc(sizeof(**pengine), GFP_KERNEL)))
 		return -ENOMEM;
-	return nvkm_engine_ctor(func, device, index, enable, *pengine);
+	return nvkm_engine_ctor(func, device, type, inst, enable, *pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/core/memory.c b/drivers/gpu/drm/nouveau/nvkm/core/memory.c
index 38130ef272d6..c69daac9bac7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/core/memory.c
+++ b/drivers/gpu/drm/nouveau/nvkm/core/memory.c
@@ -33,13 +33,13 @@ nvkm_memory_tags_put(struct nvkm_memory *memory, struct nvkm_device *device,
 	struct nvkm_fb *fb = device->fb;
 	struct nvkm_tags *tags = *ptags;
 	if (tags) {
-		mutex_lock(&fb->subdev.mutex);
+		mutex_lock(&fb->tags.mutex);
 		if (refcount_dec_and_test(&tags->refcount)) {
-			nvkm_mm_free(&fb->tags, &tags->mn);
+			nvkm_mm_free(&fb->tags.mm, &tags->mn);
 			kfree(memory->tags);
 			memory->tags = NULL;
 		}
-		mutex_unlock(&fb->subdev.mutex);
+		mutex_unlock(&fb->tags.mutex);
 		*ptags = NULL;
 	}
 }
@@ -52,29 +52,29 @@ nvkm_memory_tags_get(struct nvkm_memory *memory, struct nvkm_device *device,
 	struct nvkm_fb *fb = device->fb;
 	struct nvkm_tags *tags;
 
-	mutex_lock(&fb->subdev.mutex);
+	mutex_lock(&fb->tags.mutex);
 	if ((tags = memory->tags)) {
 		/* If comptags exist for the memory, but a different amount
 		 * than requested, the buffer is being mapped with settings
 		 * that are incompatible with existing mappings.
 		 */
 		if (tags->mn && tags->mn->length != nr) {
-			mutex_unlock(&fb->subdev.mutex);
+			mutex_unlock(&fb->tags.mutex);
 			return -EINVAL;
 		}
 
 		refcount_inc(&tags->refcount);
-		mutex_unlock(&fb->subdev.mutex);
+		mutex_unlock(&fb->tags.mutex);
 		*ptags = tags;
 		return 0;
 	}
 
 	if (!(tags = kmalloc(sizeof(*tags), GFP_KERNEL))) {
-		mutex_unlock(&fb->subdev.mutex);
+		mutex_unlock(&fb->tags.mutex);
 		return -ENOMEM;
 	}
 
-	if (!nvkm_mm_head(&fb->tags, 0, 1, nr, nr, 1, &tags->mn)) {
+	if (!nvkm_mm_head(&fb->tags.mm, 0, 1, nr, nr, 1, &tags->mn)) {
 		if (clr)
 			clr(device, tags->mn->offset, tags->mn->length);
 	} else {
@@ -92,7 +92,7 @@ nvkm_memory_tags_get(struct nvkm_memory *memory, struct nvkm_device *device,
 
 	refcount_set(&tags->refcount, 1);
 	*ptags = memory->tags = tags;
-	mutex_unlock(&fb->subdev.mutex);
+	mutex_unlock(&fb->tags.mutex);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c
index 49d468b45d3f..a74b7acb6832 100644
--- a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c
+++ b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c
@@ -26,69 +26,13 @@
 #include <core/option.h>
 #include <subdev/mc.h>
 
-static struct lock_class_key nvkm_subdev_lock_class[NVKM_SUBDEV_NR];
-
 const char *
-nvkm_subdev_name[NVKM_SUBDEV_NR] = {
-	[NVKM_SUBDEV_ACR     ] = "acr",
-	[NVKM_SUBDEV_BAR     ] = "bar",
-	[NVKM_SUBDEV_VBIOS   ] = "bios",
-	[NVKM_SUBDEV_BUS     ] = "bus",
-	[NVKM_SUBDEV_CLK     ] = "clk",
-	[NVKM_SUBDEV_DEVINIT ] = "devinit",
-	[NVKM_SUBDEV_FAULT   ] = "fault",
-	[NVKM_SUBDEV_FB      ] = "fb",
-	[NVKM_SUBDEV_FUSE    ] = "fuse",
-	[NVKM_SUBDEV_GPIO    ] = "gpio",
-	[NVKM_SUBDEV_GSP     ] = "gsp",
-	[NVKM_SUBDEV_I2C     ] = "i2c",
-	[NVKM_SUBDEV_IBUS    ] = "priv",
-	[NVKM_SUBDEV_ICCSENSE] = "iccsense",
-	[NVKM_SUBDEV_INSTMEM ] = "imem",
-	[NVKM_SUBDEV_LTC     ] = "ltc",
-	[NVKM_SUBDEV_MC      ] = "mc",
-	[NVKM_SUBDEV_MMU     ] = "mmu",
-	[NVKM_SUBDEV_MXM     ] = "mxm",
-	[NVKM_SUBDEV_PCI     ] = "pci",
-	[NVKM_SUBDEV_PMU     ] = "pmu",
-	[NVKM_SUBDEV_THERM   ] = "therm",
-	[NVKM_SUBDEV_TIMER   ] = "tmr",
-	[NVKM_SUBDEV_TOP     ] = "top",
-	[NVKM_SUBDEV_VOLT    ] = "volt",
-	[NVKM_ENGINE_BSP     ] = "bsp",
-	[NVKM_ENGINE_CE0     ] = "ce0",
-	[NVKM_ENGINE_CE1     ] = "ce1",
-	[NVKM_ENGINE_CE2     ] = "ce2",
-	[NVKM_ENGINE_CE3     ] = "ce3",
-	[NVKM_ENGINE_CE4     ] = "ce4",
-	[NVKM_ENGINE_CE5     ] = "ce5",
-	[NVKM_ENGINE_CE6     ] = "ce6",
-	[NVKM_ENGINE_CE7     ] = "ce7",
-	[NVKM_ENGINE_CE8     ] = "ce8",
-	[NVKM_ENGINE_CIPHER  ] = "cipher",
-	[NVKM_ENGINE_DISP    ] = "disp",
-	[NVKM_ENGINE_DMAOBJ  ] = "dma",
-	[NVKM_ENGINE_FIFO    ] = "fifo",
-	[NVKM_ENGINE_GR      ] = "gr",
-	[NVKM_ENGINE_IFB     ] = "ifb",
-	[NVKM_ENGINE_ME      ] = "me",
-	[NVKM_ENGINE_MPEG    ] = "mpeg",
-	[NVKM_ENGINE_MSENC   ] = "msenc",
-	[NVKM_ENGINE_MSPDEC  ] = "mspdec",
-	[NVKM_ENGINE_MSPPP   ] = "msppp",
-	[NVKM_ENGINE_MSVLD   ] = "msvld",
-	[NVKM_ENGINE_NVENC0  ] = "nvenc0",
-	[NVKM_ENGINE_NVENC1  ] = "nvenc1",
-	[NVKM_ENGINE_NVENC2  ] = "nvenc2",
-	[NVKM_ENGINE_NVDEC0  ] = "nvdec0",
-	[NVKM_ENGINE_NVDEC1  ] = "nvdec1",
-	[NVKM_ENGINE_NVDEC2  ] = "nvdec2",
-	[NVKM_ENGINE_PM      ] = "pm",
-	[NVKM_ENGINE_SEC     ] = "sec",
-	[NVKM_ENGINE_SEC2    ] = "sec2",
-	[NVKM_ENGINE_SW      ] = "sw",
-	[NVKM_ENGINE_VIC     ] = "vic",
-	[NVKM_ENGINE_VP      ] = "vp",
+nvkm_subdev_type[NVKM_SUBDEV_NR] = {
+#define NVKM_LAYOUT_ONCE(type,data,ptr,...) [type] = #ptr,
+#define NVKM_LAYOUT_INST(A...) NVKM_LAYOUT_ONCE(A)
+#include <core/layout.h>
+#undef NVKM_LAYOUT_ONCE
+#undef NVKM_LAYOUT_INST
 };
 
 void
@@ -125,7 +69,7 @@ nvkm_subdev_fini(struct nvkm_subdev *subdev, bool suspend)
 		}
 	}
 
-	nvkm_mc_reset(device, subdev->index);
+	nvkm_mc_reset(device, subdev->type, subdev->inst);
 
 	time = ktime_to_us(ktime_get()) - time;
 	nvkm_trace(subdev, "%s completed in %lldus\n", action, time);
@@ -199,6 +143,7 @@ nvkm_subdev_del(struct nvkm_subdev **psubdev)
 	if (subdev && !WARN_ON(!subdev->func)) {
 		nvkm_trace(subdev, "destroy running...\n");
 		time = ktime_to_us(ktime_get());
+		list_del(&subdev->head);
 		if (subdev->func->dtor)
 			*psubdev = subdev->func->dtor(subdev);
 		time = ktime_to_us(ktime_get()) - time;
@@ -209,26 +154,41 @@ nvkm_subdev_del(struct nvkm_subdev **psubdev)
 }
 
 void
-nvkm_subdev_ctor(const struct nvkm_subdev_func *func,
-		 struct nvkm_device *device, int index,
-		 struct nvkm_subdev *subdev)
+nvkm_subdev_disable(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
+{
+	struct nvkm_subdev *subdev;
+	list_for_each_entry(subdev, &device->subdev, head) {
+		if (subdev->type == type && subdev->inst == inst) {
+			*subdev->pself = NULL;
+			nvkm_subdev_del(&subdev);
+			break;
+		}
+	}
+}
+
+void
+nvkm_subdev_ctor(const struct nvkm_subdev_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, struct nvkm_subdev *subdev)
 {
-	const char *name = nvkm_subdev_name[index];
 	subdev->func = func;
 	subdev->device = device;
-	subdev->index = index;
-
-	__mutex_init(&subdev->mutex, name, &nvkm_subdev_lock_class[index]);
-	subdev->debug = nvkm_dbgopt(device->dbgopt, name);
+	subdev->type = type;
+	subdev->inst = inst < 0 ? 0 : inst;
+
+	if (inst >= 0)
+		snprintf(subdev->name, sizeof(subdev->name), "%s%d", nvkm_subdev_type[type], inst);
+	else
+		strscpy(subdev->name, nvkm_subdev_type[type], sizeof(subdev->name));
+	subdev->debug = nvkm_dbgopt(device->dbgopt, subdev->name);
+	list_add_tail(&subdev->head, &device->subdev);
 }
 
 int
-nvkm_subdev_new_(const struct nvkm_subdev_func *func,
-		 struct nvkm_device *device, int index,
-		 struct nvkm_subdev **psubdev)
+nvkm_subdev_new_(const struct nvkm_subdev_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, struct nvkm_subdev **psubdev)
 {
 	if (!(*psubdev = kzalloc(sizeof(**psubdev), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(func, device, index, *psubdev);
+	nvkm_subdev_ctor(func, device, type, inst, *psubdev);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/bsp/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/bsp/g84.c
index 44e116f7880d..39f6db269c7a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/bsp/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/bsp/g84.c
@@ -36,8 +36,9 @@ g84_bsp = {
 };
 
 int
-g84_bsp_new(struct nvkm_device *device, int index, struct nvkm_engine **pengine)
+g84_bsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_engine **pengine)
 {
-	return nvkm_xtensa_new_(&g84_bsp, device, index,
+	return nvkm_xtensa_new_(&g84_bsp, device, type, inst,
 				device->chipset != 0x92, 0x103000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gf100.c
index ad9f855c9a40..b9cc39565985 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gf100.c
@@ -29,9 +29,7 @@
 static void
 gf100_ce_init(struct nvkm_falcon *ce)
 {
-	struct nvkm_device *device = ce->engine.subdev.device;
-	const int index = ce->engine.subdev.index - NVKM_ENGINE_CE0;
-	nvkm_wr32(device, ce->addr + 0x084, index);
+	nvkm_wr32(ce->engine.subdev.device, ce->addr + 0x084, ce->engine.subdev.inst);
 }
 
 static const struct nvkm_falcon_func
@@ -63,16 +61,9 @@ gf100_ce1 = {
 };
 
 int
-gf100_ce_new(struct nvkm_device *device, int index,
+gf100_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	if (index == NVKM_ENGINE_CE0) {
-		return nvkm_falcon_new_(&gf100_ce0, device, index, true,
-					0x104000, pengine);
-	} else
-	if (index == NVKM_ENGINE_CE1) {
-		return nvkm_falcon_new_(&gf100_ce1, device, index, true,
-					0x105000, pengine);
-	}
-	return -ENODEV;
+	return nvkm_falcon_new_(inst ? &gf100_ce1 : &gf100_ce0, device, type, inst, true,
+				0x104000 + (inst * 0x1000), pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gk104.c
index 9e0b53a10f77..27f29eb0494d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gk104.c
@@ -58,9 +58,9 @@ gk104_ce_intr_launcherr(struct nvkm_engine *ce, const u32 base)
 void
 gk104_ce_intr(struct nvkm_engine *ce)
 {
-	const u32 base = (ce->subdev.index - NVKM_ENGINE_CE0) * 0x1000;
 	struct nvkm_subdev *subdev = &ce->subdev;
 	struct nvkm_device *device = subdev->device;
+	const u32 base = subdev->inst * 0x1000;
 	u32 mask = nvkm_rd32(device, 0x104904 + base);
 	u32 intr = nvkm_rd32(device, 0x104908 + base) & mask;
 	if (intr & 0x00000001) {
@@ -94,8 +94,8 @@ gk104_ce = {
 };
 
 int
-gk104_ce_new(struct nvkm_device *device, int index,
+gk104_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&gk104_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&gk104_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm107.c
index c0df7daa85e2..c3c476592c43 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm107.c
@@ -36,8 +36,8 @@ gm107_ce = {
 };
 
 int
-gm107_ce_new(struct nvkm_device *device, int index,
+gm107_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&gm107_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&gm107_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm200.c
index c6fa8b20737e..d2db61865371 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gm200.c
@@ -35,8 +35,8 @@ gm200_ce = {
 };
 
 int
-gm200_ce_new(struct nvkm_device *device, int index,
+gm200_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&gm200_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&gm200_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c
index c7710456bc30..a4f08a4472c9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c
@@ -59,9 +59,9 @@ gp100_ce_intr_launcherr(struct nvkm_engine *ce, const u32 base)
 void
 gp100_ce_intr(struct nvkm_engine *ce)
 {
-	const u32 base = (ce->subdev.index - NVKM_ENGINE_CE0) * 0x80;
 	struct nvkm_subdev *subdev = &ce->subdev;
 	struct nvkm_device *device = subdev->device;
+	const u32 base = subdev->inst * 0x80;
 	u32 mask = nvkm_rd32(device, 0x10440c + base);
 	u32 intr = nvkm_rd32(device, 0x104410 + base) & mask;
 	if (intr & 0x00000001) { //XXX: guess
@@ -95,8 +95,8 @@ gp100_ce = {
 };
 
 int
-gp100_ce_new(struct nvkm_device *device, int index,
+gp100_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&gp100_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&gp100_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp102.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp102.c
index 985c8f653874..180d497a95eb 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp102.c
@@ -37,8 +37,8 @@ gp102_ce = {
 };
 
 int
-gp102_ce_new(struct nvkm_device *device, int index,
+gp102_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&gp102_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&gp102_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c
index 63ac51a54fd3..704df0f2d1f1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c
@@ -44,7 +44,7 @@ gt215_ce_intr(struct nvkm_falcon *ce, struct nvkm_fifo_chan *chan)
 {
 	struct nvkm_subdev *subdev = &ce->engine.subdev;
 	struct nvkm_device *device = subdev->device;
-	const u32 base = (subdev->index - NVKM_ENGINE_CE0) * 0x1000;
+	const u32 base = subdev->inst * 0x1000;
 	u32 ssta = nvkm_rd32(device, 0x104040 + base) & 0x0000ffff;
 	u32 addr = nvkm_rd32(device, 0x104040 + base) >> 16;
 	u32 mthd = (addr & 0x07ff) << 2;
@@ -75,9 +75,9 @@ gt215_ce = {
 };
 
 int
-gt215_ce_new(struct nvkm_device *device, int index,
+gt215_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_falcon_new_(&gt215_ce, device, index,
+	return nvkm_falcon_new_(&gt215_ce, device, type, inst,
 				(device->chipset != 0xaf), 0x104000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gv100.c
index fcda3de45857..cd5e9cdca1cf 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gv100.c
@@ -33,8 +33,8 @@ gv100_ce = {
 };
 
 int
-gv100_ce_new(struct nvkm_device *device, int index,
+gv100_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&gv100_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&gv100_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c
index b4308e2d8c75..e5ff92d9364c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c
@@ -33,8 +33,8 @@ tu102_ce = {
 };
 
 int
-tu102_ce_new(struct nvkm_device *device, int index,
+tu102_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&tu102_ce, device, index, true, pengine);
+	return nvkm_engine_new_(&tu102_ce, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/cipher/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/cipher/g84.c
index 68ffb520531e..be2a7181dc15 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/cipher/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/cipher/g84.c
@@ -127,8 +127,8 @@ g84_cipher = {
 };
 
 int
-g84_cipher_new(struct nvkm_device *device, int index,
+g84_cipher_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	       struct nvkm_engine **pengine)
 {
-	return nvkm_engine_new_(&g84_cipher, device, index, true, pengine);
+	return nvkm_engine_new_(&g84_cipher, device, type, inst, true, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
index 7851bec5f0e5..b930f539feec 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
@@ -71,2585 +71,2557 @@ nvkm_device_list(u64 *name, int size)
 static const struct nvkm_device_chip
 null_chipset = {
 	.name = "NULL",
-	.bios = nvkm_bios_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
 };
 
 static const struct nvkm_device_chip
 nv4_chipset = {
 	.name = "NV04",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv04_devinit_new,
-	.fb = nv04_fb_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv04_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv04_fifo_new,
-	.gr = nv04_gr_new,
-	.sw = nv04_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv04_devinit_new },
+	.fb       = { 0x00000001, nv04_fb_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv04_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv04_fifo_new },
+	.gr       = { 0x00000001, nv04_gr_new },
+	.sw       = { 0x00000001, nv04_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv5_chipset = {
 	.name = "NV05",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv05_devinit_new,
-	.fb = nv04_fb_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv04_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv04_fifo_new,
-	.gr = nv04_gr_new,
-	.sw = nv04_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv05_devinit_new },
+	.fb       = { 0x00000001, nv04_fb_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv04_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv04_fifo_new },
+	.gr       = { 0x00000001, nv04_gr_new },
+	.sw       = { 0x00000001, nv04_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv10_chipset = {
 	.name = "NV10",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv10_devinit_new,
-	.fb = nv10_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv04_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.gr = nv10_gr_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv10_devinit_new },
+	.fb       = { 0x00000001, nv10_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv04_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.gr       = { 0x00000001, nv10_gr_new },
 };
 
 static const struct nvkm_device_chip
 nv11_chipset = {
 	.name = "NV11",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv10_devinit_new,
-	.fb = nv10_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv11_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv10_fifo_new,
-	.gr = nv15_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv10_devinit_new },
+	.fb       = { 0x00000001, nv10_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv11_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv10_fifo_new },
+	.gr       = { 0x00000001, nv15_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv15_chipset = {
 	.name = "NV15",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv10_devinit_new,
-	.fb = nv10_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv04_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv10_fifo_new,
-	.gr = nv15_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv10_devinit_new },
+	.fb       = { 0x00000001, nv10_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv04_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv10_fifo_new },
+	.gr       = { 0x00000001, nv15_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv17_chipset = {
 	.name = "NV17",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv10_devinit_new,
-	.fb = nv10_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv17_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv10_devinit_new },
+	.fb       = { 0x00000001, nv10_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv17_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv18_chipset = {
 	.name = "NV18",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv10_devinit_new,
-	.fb = nv10_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv17_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv10_devinit_new },
+	.fb       = { 0x00000001, nv10_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv17_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv1a_chipset = {
 	.name = "nForce",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv1a_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv04_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv10_fifo_new,
-	.gr = nv15_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv1a_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv04_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv10_fifo_new },
+	.gr       = { 0x00000001, nv15_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv1f_chipset = {
 	.name = "nForce2",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv1a_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv17_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv1a_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv17_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv20_chipset = {
 	.name = "NV20",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv20_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv20_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv20_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv20_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv25_chipset = {
 	.name = "NV25",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv25_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv25_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv25_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv25_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv28_chipset = {
 	.name = "NV28",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv25_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv25_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv25_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv25_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv2a_chipset = {
 	.name = "NV2A",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv25_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv2a_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv25_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv2a_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv30_chipset = {
 	.name = "NV30",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv30_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv30_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv30_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv30_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv31_chipset = {
 	.name = "NV31",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv30_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv30_gr_new,
-	.mpeg = nv31_mpeg_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv30_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv30_gr_new },
+	.mpeg     = { 0x00000001, nv31_mpeg_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv34_chipset = {
 	.name = "NV34",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv10_devinit_new,
-	.fb = nv10_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv34_gr_new,
-	.mpeg = nv31_mpeg_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv10_devinit_new },
+	.fb       = { 0x00000001, nv10_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv34_gr_new },
+	.mpeg     = { 0x00000001, nv31_mpeg_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv35_chipset = {
 	.name = "NV35",
-	.bios = nvkm_bios_new,
-	.bus = nv04_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv35_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv35_gr_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv04_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv35_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv35_gr_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv36_chipset = {
 	.name = "NV36",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv04_clk_new,
-	.devinit = nv20_devinit_new,
-	.fb = nv36_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv04_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv04_pci_new,
-	.timer = nv04_timer_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv17_fifo_new,
-	.gr = nv35_gr_new,
-	.mpeg = nv31_mpeg_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv04_clk_new },
+	.devinit  = { 0x00000001, nv20_devinit_new },
+	.fb       = { 0x00000001, nv36_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv04_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv04_pci_new },
+	.timer    = { 0x00000001, nv04_timer_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv17_fifo_new },
+	.gr       = { 0x00000001, nv35_gr_new },
+	.mpeg     = { 0x00000001, nv31_mpeg_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv40_chipset = {
 	.name = "NV40",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv40_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv40_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv40_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv40_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv40_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv40_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv41_chipset = {
 	.name = "NV41",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv41_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv41_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv40_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv41_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv41_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv40_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv42_chipset = {
 	.name = "NV42",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv41_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv41_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv40_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv41_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv41_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv40_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv43_chipset = {
 	.name = "NV43",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv41_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv41_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv40_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv41_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv41_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv40_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv44_chipset = {
 	.name = "NV44",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv44_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv44_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv45_chipset = {
 	.name = "NV45",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv40_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv40_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv46_chipset = {
 	.name = "G72",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv46_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv46_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv46_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv46_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv47_chipset = {
 	.name = "G70",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv47_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv41_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv47_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv41_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv49_chipset = {
 	.name = "G71",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv49_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv41_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv49_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv41_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv4a_chipset = {
 	.name = "NV44A",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv44_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv04_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv44_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv04_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv4b_chipset = {
 	.name = "G73",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv49_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv17_mc_new,
-	.mmu = nv41_mmu_new,
-	.pci = nv40_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv40_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv49_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv17_mc_new },
+	.mmu      = { 0x00000001, nv41_mmu_new },
+	.pci      = { 0x00000001, nv40_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv40_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv4c_chipset = {
 	.name = "C61",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv46_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv4c_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv46_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv4c_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv4e_chipset = {
 	.name = "C51",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv4e_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv4e_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv4c_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv4e_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv4e_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv4c_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv50_chipset = {
 	.name = "G80",
-	.bar = nv50_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = nv50_bus_new,
-	.clk = nv50_clk_new,
-	.devinit = nv50_devinit_new,
-	.fb = nv50_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = nv50_gpio_new,
-	.i2c = nv50_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = nv50_mc_new,
-	.mmu = nv50_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = nv46_pci_new,
-	.therm = nv50_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv50_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = nv50_fifo_new,
-	.gr = nv50_gr_new,
-	.mpeg = nv50_mpeg_new,
-	.pm = nv50_pm_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, nv50_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv50_bus_new },
+	.clk      = { 0x00000001, nv50_clk_new },
+	.devinit  = { 0x00000001, nv50_devinit_new },
+	.fb       = { 0x00000001, nv50_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, nv50_gpio_new },
+	.i2c      = { 0x00000001, nv50_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, nv50_mc_new },
+	.mmu      = { 0x00000001, nv50_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, nv46_pci_new },
+	.therm    = { 0x00000001, nv50_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv50_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, nv50_fifo_new },
+	.gr       = { 0x00000001, nv50_gr_new },
+	.mpeg     = { 0x00000001, nv50_mpeg_new },
+	.pm       = { 0x00000001, nv50_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv63_chipset = {
 	.name = "C73",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv46_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv4c_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv46_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv4c_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv67_chipset = {
 	.name = "C67",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv46_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv4c_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv46_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv4c_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv68_chipset = {
 	.name = "C68",
-	.bios = nvkm_bios_new,
-	.bus = nv31_bus_new,
-	.clk = nv40_clk_new,
-	.devinit = nv1a_devinit_new,
-	.fb = nv46_fb_new,
-	.gpio = nv10_gpio_new,
-	.i2c = nv04_i2c_new,
-	.imem = nv40_instmem_new,
-	.mc = nv44_mc_new,
-	.mmu = nv44_mmu_new,
-	.pci = nv4c_pci_new,
-	.therm = nv40_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = nv04_disp_new,
-	.dma = nv04_dma_new,
-	.fifo = nv40_fifo_new,
-	.gr = nv44_gr_new,
-	.mpeg = nv44_mpeg_new,
-	.pm = nv40_pm_new,
-	.sw = nv10_sw_new,
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv31_bus_new },
+	.clk      = { 0x00000001, nv40_clk_new },
+	.devinit  = { 0x00000001, nv1a_devinit_new },
+	.fb       = { 0x00000001, nv46_fb_new },
+	.gpio     = { 0x00000001, nv10_gpio_new },
+	.i2c      = { 0x00000001, nv04_i2c_new },
+	.imem     = { 0x00000001, nv40_instmem_new },
+	.mc       = { 0x00000001, nv44_mc_new },
+	.mmu      = { 0x00000001, nv44_mmu_new },
+	.pci      = { 0x00000001, nv4c_pci_new },
+	.therm    = { 0x00000001, nv40_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, nv04_disp_new },
+	.dma      = { 0x00000001, nv04_dma_new },
+	.fifo     = { 0x00000001, nv40_fifo_new },
+	.gr       = { 0x00000001, nv44_gr_new },
+	.mpeg     = { 0x00000001, nv44_mpeg_new },
+	.pm       = { 0x00000001, nv40_pm_new },
+	.sw       = { 0x00000001, nv10_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv84_chipset = {
 	.name = "G84",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = nv50_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g84_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = nv50_gpio_new,
-	.i2c = nv50_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g84_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g84_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.bsp = g84_bsp_new,
-	.cipher = g84_cipher_new,
-	.disp = g84_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = g84_gr_new,
-	.mpeg = g84_mpeg_new,
-	.pm = g84_pm_new,
-	.sw = nv50_sw_new,
-	.vp = g84_vp_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv50_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g84_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, nv50_gpio_new },
+	.i2c      = { 0x00000001, nv50_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g84_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g84_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.bsp      = { 0x00000001, g84_bsp_new },
+	.cipher   = { 0x00000001, g84_cipher_new },
+	.disp     = { 0x00000001, g84_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, g84_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
+	.vp       = { 0x00000001, g84_vp_new },
 };
 
 static const struct nvkm_device_chip
 nv86_chipset = {
 	.name = "G86",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = nv50_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g84_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = nv50_gpio_new,
-	.i2c = nv50_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g84_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g84_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.bsp = g84_bsp_new,
-	.cipher = g84_cipher_new,
-	.disp = g84_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = g84_gr_new,
-	.mpeg = g84_mpeg_new,
-	.pm = g84_pm_new,
-	.sw = nv50_sw_new,
-	.vp = g84_vp_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv50_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g84_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, nv50_gpio_new },
+	.i2c      = { 0x00000001, nv50_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g84_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g84_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.bsp      = { 0x00000001, g84_bsp_new },
+	.cipher   = { 0x00000001, g84_cipher_new },
+	.disp     = { 0x00000001, g84_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, g84_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
+	.vp       = { 0x00000001, g84_vp_new },
 };
 
 static const struct nvkm_device_chip
 nv92_chipset = {
 	.name = "G92",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = nv50_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g84_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = nv50_gpio_new,
-	.i2c = nv50_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g84_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g92_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.bsp = g84_bsp_new,
-	.cipher = g84_cipher_new,
-	.disp = g84_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = g84_gr_new,
-	.mpeg = g84_mpeg_new,
-	.pm = g84_pm_new,
-	.sw = nv50_sw_new,
-	.vp = g84_vp_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, nv50_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g84_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, nv50_gpio_new },
+	.i2c      = { 0x00000001, nv50_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g84_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g92_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.bsp      = { 0x00000001, g84_bsp_new },
+	.cipher   = { 0x00000001, g84_cipher_new },
+	.disp     = { 0x00000001, g84_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, g84_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
+	.vp       = { 0x00000001, g84_vp_new },
 };
 
 static const struct nvkm_device_chip
 nv94_chipset = {
 	.name = "G94",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g84_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g84_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.bsp = g84_bsp_new,
-	.cipher = g84_cipher_new,
-	.disp = g94_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = g84_gr_new,
-	.mpeg = g84_mpeg_new,
-	.pm = g84_pm_new,
-	.sw = nv50_sw_new,
-	.vp = g84_vp_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g84_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g84_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.bsp      = { 0x00000001, g84_bsp_new },
+	.cipher   = { 0x00000001, g84_cipher_new },
+	.disp     = { 0x00000001, g94_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, g84_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
+	.vp       = { 0x00000001, g84_vp_new },
 };
 
 static const struct nvkm_device_chip
 nv96_chipset = {
 	.name = "G96",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g84_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g84_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.bsp = g84_bsp_new,
-	.cipher = g84_cipher_new,
-	.disp = g94_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = g84_gr_new,
-	.mpeg = g84_mpeg_new,
-	.pm = g84_pm_new,
-	.sw = nv50_sw_new,
-	.vp = g84_vp_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g84_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g84_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.bsp      = { 0x00000001, g84_bsp_new },
+	.cipher   = { 0x00000001, g84_cipher_new },
+	.disp     = { 0x00000001, g94_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, g84_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
+	.vp       = { 0x00000001, g84_vp_new },
 };
 
 static const struct nvkm_device_chip
 nv98_chipset = {
 	.name = "G98",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g98_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g98_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = g94_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = g84_gr_new,
-	.mspdec = g98_mspdec_new,
-	.msppp = g98_msppp_new,
-	.msvld = g98_msvld_new,
-	.pm = g84_pm_new,
-	.sec = g98_sec_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g98_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g98_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, g94_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, g84_gr_new },
+	.mspdec   = { 0x00000001, g98_mspdec_new },
+	.msppp    = { 0x00000001, g98_msppp_new },
+	.msvld    = { 0x00000001, g98_msvld_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sec      = { 0x00000001, g98_sec_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nva0_chipset = {
 	.name = "GT200",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = g84_clk_new,
-	.devinit = g84_devinit_new,
-	.fb = g84_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = nv50_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g84_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.bsp = g84_bsp_new,
-	.cipher = g84_cipher_new,
-	.disp = gt200_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = gt200_gr_new,
-	.mpeg = g84_mpeg_new,
-	.pm = gt200_pm_new,
-	.sw = nv50_sw_new,
-	.vp = g84_vp_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, g84_clk_new },
+	.devinit  = { 0x00000001, g84_devinit_new },
+	.fb       = { 0x00000001, g84_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, nv50_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g84_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.bsp      = { 0x00000001, g84_bsp_new },
+	.cipher   = { 0x00000001, g84_cipher_new },
+	.disp     = { 0x00000001, gt200_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, gt200_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.pm       = { 0x00000001, gt200_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
+	.vp       = { 0x00000001, g84_vp_new },
 };
 
 static const struct nvkm_device_chip
 nva3_chipset = {
 	.name = "GT215",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = gt215_clk_new,
-	.devinit = gt215_devinit_new,
-	.fb = gt215_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = gt215_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.pmu = gt215_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.ce[0] = gt215_ce_new,
-	.disp = gt215_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = gt215_gr_new,
-	.mpeg = g84_mpeg_new,
-	.mspdec = gt215_mspdec_new,
-	.msppp = gt215_msppp_new,
-	.msvld = gt215_msvld_new,
-	.pm = gt215_pm_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, gt215_clk_new },
+	.devinit  = { 0x00000001, gt215_devinit_new },
+	.fb       = { 0x00000001, gt215_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, gt215_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.pmu      = { 0x00000001, gt215_pmu_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.ce       = { 0x00000001, gt215_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, gt215_gr_new },
+	.mpeg     = { 0x00000001, g84_mpeg_new },
+	.mspdec   = { 0x00000001, gt215_mspdec_new },
+	.msppp    = { 0x00000001, gt215_msppp_new },
+	.msvld    = { 0x00000001, gt215_msvld_new },
+	.pm       = { 0x00000001, gt215_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nva5_chipset = {
 	.name = "GT216",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = gt215_clk_new,
-	.devinit = gt215_devinit_new,
-	.fb = gt215_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = gt215_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.pmu = gt215_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.ce[0] = gt215_ce_new,
-	.disp = gt215_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = gt215_gr_new,
-	.mspdec = gt215_mspdec_new,
-	.msppp = gt215_msppp_new,
-	.msvld = gt215_msvld_new,
-	.pm = gt215_pm_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, gt215_clk_new },
+	.devinit  = { 0x00000001, gt215_devinit_new },
+	.fb       = { 0x00000001, gt215_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, gt215_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.pmu      = { 0x00000001, gt215_pmu_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.ce       = { 0x00000001, gt215_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, gt215_gr_new },
+	.mspdec   = { 0x00000001, gt215_mspdec_new },
+	.msppp    = { 0x00000001, gt215_msppp_new },
+	.msvld    = { 0x00000001, gt215_msvld_new },
+	.pm       = { 0x00000001, gt215_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nva8_chipset = {
 	.name = "GT218",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = gt215_clk_new,
-	.devinit = gt215_devinit_new,
-	.fb = gt215_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = gt215_mc_new,
-	.mmu = g84_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.pmu = gt215_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.ce[0] = gt215_ce_new,
-	.disp = gt215_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = gt215_gr_new,
-	.mspdec = gt215_mspdec_new,
-	.msppp = gt215_msppp_new,
-	.msvld = gt215_msvld_new,
-	.pm = gt215_pm_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, gt215_clk_new },
+	.devinit  = { 0x00000001, gt215_devinit_new },
+	.fb       = { 0x00000001, gt215_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, gt215_mc_new },
+	.mmu      = { 0x00000001, g84_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.pmu      = { 0x00000001, gt215_pmu_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.ce       = { 0x00000001, gt215_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, gt215_gr_new },
+	.mspdec   = { 0x00000001, gt215_mspdec_new },
+	.msppp    = { 0x00000001, gt215_msppp_new },
+	.msvld    = { 0x00000001, gt215_msvld_new },
+	.pm       = { 0x00000001, gt215_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvaa_chipset = {
 	.name = "MCP77/MCP78",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = mcp77_clk_new,
-	.devinit = g98_devinit_new,
-	.fb = mcp77_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g98_mc_new,
-	.mmu = mcp77_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = mcp77_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = gt200_gr_new,
-	.mspdec = g98_mspdec_new,
-	.msppp = g98_msppp_new,
-	.msvld = g98_msvld_new,
-	.pm = g84_pm_new,
-	.sec = g98_sec_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, mcp77_clk_new },
+	.devinit  = { 0x00000001, g98_devinit_new },
+	.fb       = { 0x00000001, mcp77_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g98_mc_new },
+	.mmu      = { 0x00000001, mcp77_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, mcp77_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, gt200_gr_new },
+	.mspdec   = { 0x00000001, g98_mspdec_new },
+	.msppp    = { 0x00000001, g98_msppp_new },
+	.msvld    = { 0x00000001, g98_msvld_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sec      = { 0x00000001, g98_sec_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvac_chipset = {
 	.name = "MCP79/MCP7A",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = mcp77_clk_new,
-	.devinit = g98_devinit_new,
-	.fb = mcp77_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = g98_mc_new,
-	.mmu = mcp77_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.therm = g84_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.disp = mcp77_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = mcp79_gr_new,
-	.mspdec = g98_mspdec_new,
-	.msppp = g98_msppp_new,
-	.msvld = g98_msvld_new,
-	.pm = g84_pm_new,
-	.sec = g98_sec_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, mcp77_clk_new },
+	.devinit  = { 0x00000001, g98_devinit_new },
+	.fb       = { 0x00000001, mcp77_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, g98_mc_new },
+	.mmu      = { 0x00000001, mcp77_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.therm    = { 0x00000001, g84_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.disp     = { 0x00000001, mcp77_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, mcp79_gr_new },
+	.mspdec   = { 0x00000001, g98_mspdec_new },
+	.msppp    = { 0x00000001, g98_msppp_new },
+	.msvld    = { 0x00000001, g98_msvld_new },
+	.pm       = { 0x00000001, g84_pm_new },
+	.sec      = { 0x00000001, g98_sec_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvaf_chipset = {
 	.name = "MCP89",
-	.bar = g84_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = g94_bus_new,
-	.clk = gt215_clk_new,
-	.devinit = mcp89_devinit_new,
-	.fb = mcp89_fb_new,
-	.fuse = nv50_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.imem = nv50_instmem_new,
-	.mc = gt215_mc_new,
-	.mmu = mcp77_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = g94_pci_new,
-	.pmu = gt215_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = nv40_volt_new,
-	.ce[0] = gt215_ce_new,
-	.disp = mcp89_disp_new,
-	.dma = nv50_dma_new,
-	.fifo = g84_fifo_new,
-	.gr = mcp89_gr_new,
-	.mspdec = gt215_mspdec_new,
-	.msppp = gt215_msppp_new,
-	.msvld = mcp89_msvld_new,
-	.pm = gt215_pm_new,
-	.sw = nv50_sw_new,
+	.bar      = { 0x00000001, g84_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, g94_bus_new },
+	.clk      = { 0x00000001, gt215_clk_new },
+	.devinit  = { 0x00000001, mcp89_devinit_new },
+	.fb       = { 0x00000001, mcp89_fb_new },
+	.fuse     = { 0x00000001, nv50_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, gt215_mc_new },
+	.mmu      = { 0x00000001, mcp77_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, g94_pci_new },
+	.pmu      = { 0x00000001, gt215_pmu_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, nv40_volt_new },
+	.ce       = { 0x00000001, gt215_ce_new },
+	.disp     = { 0x00000001, mcp89_disp_new },
+	.dma      = { 0x00000001, nv50_dma_new },
+	.fifo     = { 0x00000001, g84_fifo_new },
+	.gr       = { 0x00000001, mcp89_gr_new },
+	.mspdec   = { 0x00000001, gt215_mspdec_new },
+	.msppp    = { 0x00000001, gt215_msppp_new },
+	.msvld    = { 0x00000001, mcp89_msvld_new },
+	.pm       = { 0x00000001, gt215_pm_new },
+	.sw       = { 0x00000001, nv50_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvc0_chipset = {
 	.name = "GF100",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf100_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.ce[1] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf100_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf100_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf100_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000003, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf100_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf100_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvc1_chipset = {
 	.name = "GF108",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf108_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf106_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf108_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf108_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf108_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf106_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000001, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf108_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf108_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvc3_chipset = {
 	.name = "GF106",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf106_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf104_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf100_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf106_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000001, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf104_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf100_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvc4_chipset = {
 	.name = "GF104",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf100_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.ce[1] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf104_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf100_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf100_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000003, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf104_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf100_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvc8_chipset = {
 	.name = "GF110",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf100_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.ce[1] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf110_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf100_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf100_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000003, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf110_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf100_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvce_chipset = {
 	.name = "GF114",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf100_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.ce[1] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf104_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf100_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf100_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000003, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf104_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf100_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvcf_chipset = {
 	.name = "GF116",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = g94_gpio_new,
-	.i2c = g94_i2c_new,
-	.ibus = gf100_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf106_pci_new,
-	.pmu = gf100_pmu_new,
-	.therm = gt215_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.disp = gt215_disp_new,
-	.dma = gf100_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf104_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf100_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, g94_gpio_new },
+	.i2c      = { 0x00000001, g94_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf106_pci_new },
+	.pmu      = { 0x00000001, gf100_pmu_new },
+	.privring = { 0x00000001, gf100_privring_new },
+	.therm    = { 0x00000001, gt215_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000001, gf100_ce_new },
+	.disp     = { 0x00000001, gt215_disp_new },
+	.dma      = { 0x00000001, gf100_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf104_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf100_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvd7_chipset = {
 	.name = "GF117",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gf119_gpio_new,
-	.i2c = gf117_i2c_new,
-	.ibus = gf117_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf106_pci_new,
-	.therm = gf119_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf117_volt_new,
-	.ce[0] = gf100_ce_new,
-	.disp = gf119_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf117_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf117_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gf119_gpio_new },
+	.i2c      = { 0x00000001, gf117_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf106_pci_new },
+	.privring = { 0x00000001, gf117_privring_new },
+	.therm    = { 0x00000001, gf119_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf117_volt_new },
+	.ce       = { 0x00000001, gf100_ce_new },
+	.disp     = { 0x00000001, gf119_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf117_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf117_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvd9_chipset = {
 	.name = "GF119",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gf100_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gf100_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gf119_gpio_new,
-	.i2c = gf119_i2c_new,
-	.ibus = gf117_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gf100_ltc_new,
-	.mc = gf100_mc_new,
-	.mmu = gf100_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gf106_pci_new,
-	.pmu = gf119_pmu_new,
-	.therm = gf119_therm_new,
-	.timer = nv41_timer_new,
-	.volt = gf100_volt_new,
-	.ce[0] = gf100_ce_new,
-	.disp = gf119_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gf100_fifo_new,
-	.gr = gf119_gr_new,
-	.mspdec = gf100_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gf100_msvld_new,
-	.pm = gf117_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gf100_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gf100_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gf119_gpio_new },
+	.i2c      = { 0x00000001, gf119_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gf100_ltc_new },
+	.mc       = { 0x00000001, gf100_mc_new },
+	.mmu      = { 0x00000001, gf100_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gf106_pci_new },
+	.pmu      = { 0x00000001, gf119_pmu_new },
+	.privring = { 0x00000001, gf117_privring_new },
+	.therm    = { 0x00000001, gf119_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.volt     = { 0x00000001, gf100_volt_new },
+	.ce       = { 0x00000001, gf100_ce_new },
+	.disp     = { 0x00000001, gf119_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gf100_fifo_new },
+	.gr       = { 0x00000001, gf119_gr_new },
+	.mspdec   = { 0x00000001, gf100_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gf100_msvld_new },
+	.pm       = { 0x00000001, gf117_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nve4_chipset = {
 	.name = "GK104",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk104_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk104_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk104_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk104_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk104_fifo_new,
-	.gr = gk104_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.pm = gk104_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk104_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk104_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk104_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk104_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk104_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk104_fifo_new },
+	.gr       = { 0x00000001, gk104_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.pm       = { 0x00000001, gk104_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nve6_chipset = {
 	.name = "GK106",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk104_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk104_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk104_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk104_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk104_fifo_new,
-	.gr = gk104_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.pm = gk104_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk104_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk104_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk104_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk104_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk104_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk104_fifo_new },
+	.gr       = { 0x00000001, gk104_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.pm       = { 0x00000001, gk104_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nve7_chipset = {
 	.name = "GK107",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk104_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk104_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk104_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk104_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk104_fifo_new,
-	.gr = gk104_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.pm = gk104_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk104_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk104_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk104_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk104_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk104_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk104_fifo_new },
+	.gr       = { 0x00000001, gk104_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.pm       = { 0x00000001, gk104_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvea_chipset = {
 	.name = "GK20A",
-	.bar = gk20a_bar_new,
-	.bus = gf100_bus_new,
-	.clk = gk20a_clk_new,
-	.fb = gk20a_fb_new,
-	.fuse = gf100_fuse_new,
-	.ibus = gk20a_ibus_new,
-	.imem = gk20a_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gk20a_mmu_new,
-	.pmu = gk20a_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.volt = gk20a_volt_new,
-	.ce[2] = gk104_ce_new,
-	.dma = gf119_dma_new,
-	.fifo = gk20a_fifo_new,
-	.gr = gk20a_gr_new,
-	.pm = gk104_pm_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gk20a_bar_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk20a_clk_new },
+	.fb       = { 0x00000001, gk20a_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.imem     = { 0x00000001, gk20a_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gk20a_mmu_new },
+	.pmu      = { 0x00000001, gk20a_pmu_new },
+	.privring = { 0x00000001, gk20a_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk20a_volt_new },
+	.ce       = { 0x00000004, gk104_ce_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk20a_fifo_new },
+	.gr       = { 0x00000001, gk20a_gr_new },
+	.pm       = { 0x00000001, gk104_pm_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvf0_chipset = {
 	.name = "GK110",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk110_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk104_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk110_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk110_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk110_fifo_new,
-	.gr = gk110_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk110_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk110_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk104_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk110_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk110_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk110_fifo_new },
+	.gr       = { 0x00000001, gk110_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nvf1_chipset = {
 	.name = "GK110B",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk110_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk104_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk110_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk110_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk110_fifo_new,
-	.gr = gk110b_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk110_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk110_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk104_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk110_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk110_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk110_fifo_new },
+	.gr       = { 0x00000001, gk110b_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv106_chipset = {
 	.name = "GK208B",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk110_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk208_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk110_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk208_fifo_new,
-	.gr = gk208_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk110_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk110_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk208_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk110_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk208_fifo_new },
+	.gr       = { 0x00000001, gk208_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv108_chipset = {
 	.name = "GK208",
-	.bar = gf100_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gf100_devinit_new,
-	.fb = gk110_fb_new,
-	.fuse = gf100_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gk104_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gk208_pmu_new,
-	.therm = gk104_therm_new,
-	.timer = nv41_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gk104_ce_new,
-	.ce[1] = gk104_ce_new,
-	.ce[2] = gk104_ce_new,
-	.disp = gk110_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gk208_fifo_new,
-	.gr = gk208_gr_new,
-	.mspdec = gk104_mspdec_new,
-	.msppp = gf100_msppp_new,
-	.msvld = gk104_msvld_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gf100_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gf100_devinit_new },
+	.fb       = { 0x00000001, gk110_fb_new },
+	.fuse     = { 0x00000001, gf100_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk110_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gk104_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gk208_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gk104_therm_new },
+	.timer    = { 0x00000001, nv41_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gk104_ce_new },
+	.disp     = { 0x00000001, gk110_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gk208_fifo_new },
+	.gr       = { 0x00000001, gk208_gr_new },
+	.mspdec   = { 0x00000001, gk104_mspdec_new },
+	.msppp    = { 0x00000001, gf100_msppp_new },
+	.msvld    = { 0x00000001, gk104_msvld_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv117_chipset = {
 	.name = "GM107",
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gm107_devinit_new,
-	.fb = gm107_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gm107_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gm107_pmu_new,
-	.therm = gm107_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gm107_ce_new,
-	.ce[2] = gm107_ce_new,
-	.disp = gm107_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gm107_fifo_new,
-	.gr = gm107_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gm107_devinit_new },
+	.fb       = { 0x00000001, gm107_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk110_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gm107_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gm107_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gm107_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000005, gm107_ce_new },
+	.disp     = { 0x00000001, gm107_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gm107_fifo_new },
+	.gr       = { 0x00000001, gm107_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv118_chipset = {
 	.name = "GM108",
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.clk = gk104_clk_new,
-	.devinit = gm107_devinit_new,
-	.fb = gm107_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gk104_i2c_new,
-	.ibus = gk104_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gm107_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gk104_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gm107_pmu_new,
-	.therm = gm107_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gm107_ce_new,
-	.ce[2] = gm107_ce_new,
-	.disp = gm107_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gm107_fifo_new,
-	.gr = gm107_gr_new,
-	.sw = gf100_sw_new,
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gk104_clk_new },
+	.devinit  = { 0x00000001, gm107_devinit_new },
+	.fb       = { 0x00000001, gm107_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gk110_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gm107_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gk104_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gm107_pmu_new },
+	.privring = { 0x00000001, gk104_privring_new },
+	.therm    = { 0x00000001, gm107_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000005, gm107_ce_new },
+	.disp     = { 0x00000001, gm107_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gm107_fifo_new },
+	.gr       = { 0x00000001, gm107_gr_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv120_chipset = {
 	.name = "GM200",
-	.acr = gm200_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fb = gm200_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gm200_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gm200_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gm200_pmu_new,
-	.therm = gm200_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gm200_ce_new,
-	.ce[1] = gm200_ce_new,
-	.ce[2] = gm200_ce_new,
-	.disp = gm200_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gm200_fifo_new,
-	.gr = gm200_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gm200_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fb       = { 0x00000001, gm200_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gm200_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gm200_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gm200_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gm200_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gm200_ce_new },
+	.disp     = { 0x00000001, gm200_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gm200_fifo_new },
+	.gr       = { 0x00000001, gm200_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000003, gm107_nvenc_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv124_chipset = {
 	.name = "GM204",
-	.acr = gm200_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fb = gm200_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gm200_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gm200_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gm200_pmu_new,
-	.therm = gm200_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gm200_ce_new,
-	.ce[1] = gm200_ce_new,
-	.ce[2] = gm200_ce_new,
-	.disp = gm200_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gm200_fifo_new,
-	.gr = gm200_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gm200_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fb       = { 0x00000001, gm200_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gm200_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gm200_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gm200_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gm200_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gm200_ce_new },
+	.disp     = { 0x00000001, gm200_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gm200_fifo_new },
+	.gr       = { 0x00000001, gm200_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000003, gm107_nvenc_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv126_chipset = {
 	.name = "GM206",
-	.acr = gm200_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fb = gm200_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.iccsense = gf100_iccsense_new,
-	.imem = nv50_instmem_new,
-	.ltc = gm200_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gm200_mmu_new,
-	.mxm = nv50_mxm_new,
-	.pci = gk104_pci_new,
-	.pmu = gm200_pmu_new,
-	.therm = gm200_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.volt = gk104_volt_new,
-	.ce[0] = gm200_ce_new,
-	.ce[1] = gm200_ce_new,
-	.ce[2] = gm200_ce_new,
-	.disp = gm200_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gm200_fifo_new,
-	.gr = gm200_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gm200_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fb       = { 0x00000001, gm200_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.iccsense = { 0x00000001, gf100_iccsense_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gm200_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gm200_mmu_new },
+	.mxm      = { 0x00000001, nv50_mxm_new },
+	.pci      = { 0x00000001, gk104_pci_new },
+	.pmu      = { 0x00000001, gm200_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gm200_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gk104_volt_new },
+	.ce       = { 0x00000007, gm200_ce_new },
+	.disp     = { 0x00000001, gm200_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gm200_fifo_new },
+	.gr       = { 0x00000001, gm200_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv12b_chipset = {
 	.name = "GM20B",
-	.acr = gm20b_acr_new,
-	.bar = gm20b_bar_new,
-	.bus = gf100_bus_new,
-	.clk = gm20b_clk_new,
-	.fb = gm20b_fb_new,
-	.fuse = gm107_fuse_new,
-	.ibus = gk20a_ibus_new,
-	.imem = gk20a_instmem_new,
-	.ltc = gm200_ltc_new,
-	.mc = gk20a_mc_new,
-	.mmu = gm20b_mmu_new,
-	.pmu = gm20b_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[2] = gm200_ce_new,
-	.volt = gm20b_volt_new,
-	.dma = gf119_dma_new,
-	.fifo = gm20b_fifo_new,
-	.gr = gm20b_gr_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gm20b_acr_new },
+	.bar      = { 0x00000001, gm20b_bar_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.clk      = { 0x00000001, gm20b_clk_new },
+	.fb       = { 0x00000001, gm20b_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.imem     = { 0x00000001, gk20a_instmem_new },
+	.ltc      = { 0x00000001, gm200_ltc_new },
+	.mc       = { 0x00000001, gk20a_mc_new },
+	.mmu      = { 0x00000001, gm20b_mmu_new },
+	.pmu      = { 0x00000001, gm20b_pmu_new },
+	.privring = { 0x00000001, gk20a_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.volt     = { 0x00000001, gm20b_volt_new },
+	.ce       = { 0x00000004, gm200_ce_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gm20b_fifo_new },
+	.gr       = { 0x00000001, gm20b_gr_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv130_chipset = {
 	.name = "GP100",
-	.acr = gm200_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fault = gp100_fault_new,
-	.fb = gp100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp100_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gp100_mmu_new,
-	.therm = gp100_therm_new,
-	.pci = gp100_pci_new,
-	.pmu = gm200_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp100_ce_new,
-	.ce[1] = gp100_ce_new,
-	.ce[2] = gp100_ce_new,
-	.ce[3] = gp100_ce_new,
-	.ce[4] = gp100_ce_new,
-	.ce[5] = gp100_ce_new,
-	.dma = gf119_dma_new,
-	.disp = gp100_disp_new,
-	.fifo = gp100_fifo_new,
-	.gr = gp100_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.nvenc[2] = gm107_nvenc_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gm200_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fault    = { 0x00000001, gp100_fault_new },
+	.fb       = { 0x00000001, gp100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp100_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gp100_mmu_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gm200_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000003f, gp100_ce_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.disp     = { 0x00000001, gp100_disp_new },
+	.fifo     = { 0x00000001, gp100_fifo_new },
+	.gr       = { 0x00000001, gp100_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000007, gm107_nvenc_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv132_chipset = {
 	.name = "GP102",
-	.acr = gp102_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fault = gp100_fault_new,
-	.fb = gp102_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gp100_mmu_new,
-	.therm = gp100_therm_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp102_ce_new,
-	.ce[1] = gp102_ce_new,
-	.ce[2] = gp102_ce_new,
-	.ce[3] = gp102_ce_new,
-	.disp = gp102_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gp100_fifo_new,
-	.gr = gp102_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.sec2 = gp102_sec2_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gp102_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fault    = { 0x00000001, gp100_fault_new },
+	.fb       = { 0x00000001, gp102_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gp100_mmu_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000000f, gp102_ce_new },
+	.disp     = { 0x00000001, gp102_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gp100_fifo_new },
+	.gr       = { 0x00000001, gp102_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000003, gm107_nvenc_new },
+	.sec2     = { 0x00000001, gp102_sec2_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv134_chipset = {
 	.name = "GP104",
-	.acr = gp102_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fault = gp100_fault_new,
-	.fb = gp102_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gp100_mmu_new,
-	.therm = gp100_therm_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp102_ce_new,
-	.ce[1] = gp102_ce_new,
-	.ce[2] = gp102_ce_new,
-	.ce[3] = gp102_ce_new,
-	.disp = gp102_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gp100_fifo_new,
-	.gr = gp104_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.sec2 = gp102_sec2_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gp102_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fault    = { 0x00000001, gp100_fault_new },
+	.fb       = { 0x00000001, gp102_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gp100_mmu_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000000f, gp102_ce_new },
+	.disp     = { 0x00000001, gp102_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gp100_fifo_new },
+	.gr       = { 0x00000001, gp104_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000003, gm107_nvenc_new },
+	.sec2     = { 0x00000001, gp102_sec2_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv136_chipset = {
 	.name = "GP106",
-	.acr = gp102_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fault = gp100_fault_new,
-	.fb = gp102_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gp100_mmu_new,
-	.therm = gp100_therm_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp102_ce_new,
-	.ce[1] = gp102_ce_new,
-	.ce[2] = gp102_ce_new,
-	.ce[3] = gp102_ce_new,
-	.disp = gp102_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gp100_fifo_new,
-	.gr = gp104_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sec2 = gp102_sec2_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gp102_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fault    = { 0x00000001, gp100_fault_new },
+	.fb       = { 0x00000001, gp102_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gp100_mmu_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000000f, gp102_ce_new },
+	.disp     = { 0x00000001, gp102_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gp100_fifo_new },
+	.gr       = { 0x00000001, gp104_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sec2     = { 0x00000001, gp102_sec2_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv137_chipset = {
 	.name = "GP107",
-	.acr = gp102_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fault = gp100_fault_new,
-	.fb = gp102_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gp100_mmu_new,
-	.therm = gp100_therm_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp102_ce_new,
-	.ce[1] = gp102_ce_new,
-	.ce[2] = gp102_ce_new,
-	.ce[3] = gp102_ce_new,
-	.disp = gp102_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gp100_fifo_new,
-	.gr = gp107_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.sec2 = gp102_sec2_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gp102_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fault    = { 0x00000001, gp100_fault_new },
+	.fb       = { 0x00000001, gp102_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gp100_mmu_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000000f, gp102_ce_new },
+	.disp     = { 0x00000001, gp102_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gp100_fifo_new },
+	.gr       = { 0x00000001, gp107_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000003, gm107_nvenc_new },
+	.sec2     = { 0x00000001, gp102_sec2_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv138_chipset = {
 	.name = "GP108",
-	.acr = gp108_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gm200_devinit_new,
-	.fault = gp100_fault_new,
-	.fb = gp102_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gp100_mmu_new,
-	.therm = gp100_therm_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp102_ce_new,
-	.ce[1] = gp102_ce_new,
-	.ce[2] = gp102_ce_new,
-	.ce[3] = gp102_ce_new,
-	.disp = gp102_disp_new,
-	.dma = gf119_dma_new,
-	.fifo = gp100_fifo_new,
-	.gr = gp108_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.sec2 = gp108_sec2_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gp108_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gm200_devinit_new },
+	.fault    = { 0x00000001, gp100_fault_new },
+	.fb       = { 0x00000001, gp102_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gp100_mmu_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000000f, gp102_ce_new },
+	.disp     = { 0x00000001, gp102_disp_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gp100_fifo_new },
+	.gr       = { 0x00000001, gp108_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.sec2     = { 0x00000001, gp108_sec2_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv13b_chipset = {
 	.name = "GP10B",
-	.acr = gp10b_acr_new,
-	.bar = gm20b_bar_new,
-	.bus = gf100_bus_new,
-	.fault = gp10b_fault_new,
-	.fb = gp10b_fb_new,
-	.fuse = gm107_fuse_new,
-	.ibus = gp10b_ibus_new,
-	.imem = gk20a_instmem_new,
-	.ltc = gp10b_ltc_new,
-	.mc = gp10b_mc_new,
-	.mmu = gp10b_mmu_new,
-	.pmu = gp10b_pmu_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = gp100_ce_new,
-	.dma = gf119_dma_new,
-	.fifo = gp10b_fifo_new,
-	.gr = gp10b_gr_new,
-	.sw = gf100_sw_new,
+	.acr      = { 0x00000001, gp10b_acr_new },
+	.bar      = { 0x00000001, gm20b_bar_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.fault    = { 0x00000001, gp10b_fault_new },
+	.fb       = { 0x00000001, gp10b_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.imem     = { 0x00000001, gk20a_instmem_new },
+	.ltc      = { 0x00000001, gp10b_ltc_new },
+	.mc       = { 0x00000001, gp10b_mc_new },
+	.mmu      = { 0x00000001, gp10b_mmu_new },
+	.pmu      = { 0x00000001, gp10b_pmu_new },
+	.privring = { 0x00000001, gp10b_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x00000001, gp100_ce_new },
+	.dma      = { 0x00000001, gf119_dma_new },
+	.fifo     = { 0x00000001, gp10b_fifo_new },
+	.gr       = { 0x00000001, gp10b_gr_new },
+	.sw       = { 0x00000001, gf100_sw_new },
 };
 
 static const struct nvkm_device_chip
 nv140_chipset = {
 	.name = "GV100",
-	.acr = gp108_acr_new,
-	.bar = gm107_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = gv100_devinit_new,
-	.fault = gv100_fault_new,
-	.fb = gv100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.gsp = gv100_gsp_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = gp100_mc_new,
-	.mmu = gv100_mmu_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.therm = gp100_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.disp = gv100_disp_new,
-	.ce[0] = gv100_ce_new,
-	.ce[1] = gv100_ce_new,
-	.ce[2] = gv100_ce_new,
-	.ce[3] = gv100_ce_new,
-	.ce[4] = gv100_ce_new,
-	.ce[5] = gv100_ce_new,
-	.ce[6] = gv100_ce_new,
-	.ce[7] = gv100_ce_new,
-	.ce[8] = gv100_ce_new,
-	.dma = gv100_dma_new,
-	.fifo = gv100_fifo_new,
-	.gr = gv100_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.nvenc[1] = gm107_nvenc_new,
-	.nvenc[2] = gm107_nvenc_new,
-	.sec2 = gp108_sec2_new,
+	.acr      = { 0x00000001, gp108_acr_new },
+	.bar      = { 0x00000001, gm107_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, gv100_devinit_new },
+	.fault    = { 0x00000001, gv100_fault_new },
+	.fb       = { 0x00000001, gv100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.gsp      = { 0x00000001, gv100_gsp_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, gp100_mc_new },
+	.mmu      = { 0x00000001, gv100_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x000001ff, gv100_ce_new },
+	.disp     = { 0x00000001, gv100_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+	.fifo     = { 0x00000001, gv100_fifo_new },
+	.gr       = { 0x00000001, gv100_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000007, gm107_nvenc_new },
+	.sec2     = { 0x00000001, gp108_sec2_new },
 };
 
 static const struct nvkm_device_chip
 nv162_chipset = {
 	.name = "TU102",
-	.acr = tu102_acr_new,
-	.bar = tu102_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = tu102_devinit_new,
-	.fault = tu102_fault_new,
-	.fb = gv100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.gsp = gv100_gsp_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = tu102_mc_new,
-	.mmu = tu102_mmu_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.therm = gp100_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = tu102_ce_new,
-	.ce[1] = tu102_ce_new,
-	.ce[2] = tu102_ce_new,
-	.ce[3] = tu102_ce_new,
-	.ce[4] = tu102_ce_new,
-	.disp = tu102_disp_new,
-	.dma = gv100_dma_new,
-	.fifo = tu102_fifo_new,
-	.gr = tu102_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sec2 = tu102_sec2_new,
+	.acr      = { 0x00000001, tu102_acr_new },
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, tu102_devinit_new },
+	.fault    = { 0x00000001, tu102_fault_new },
+	.fb       = { 0x00000001, gv100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.gsp      = { 0x00000001, gv100_gsp_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, tu102_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000001f, tu102_ce_new },
+	.disp     = { 0x00000001, tu102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+	.fifo     = { 0x00000001, tu102_fifo_new },
+	.gr       = { 0x00000001, tu102_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sec2     = { 0x00000001, tu102_sec2_new },
 };
 
 static const struct nvkm_device_chip
 nv164_chipset = {
 	.name = "TU104",
-	.acr = tu102_acr_new,
-	.bar = tu102_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = tu102_devinit_new,
-	.fault = tu102_fault_new,
-	.fb = gv100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.gsp = gv100_gsp_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = tu102_mc_new,
-	.mmu = tu102_mmu_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.therm = gp100_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = tu102_ce_new,
-	.ce[1] = tu102_ce_new,
-	.ce[2] = tu102_ce_new,
-	.ce[3] = tu102_ce_new,
-	.ce[4] = tu102_ce_new,
-	.disp = tu102_disp_new,
-	.dma = gv100_dma_new,
-	.fifo = tu102_fifo_new,
-	.gr = tu102_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvdec[1] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sec2 = tu102_sec2_new,
+	.acr      = { 0x00000001, tu102_acr_new },
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, tu102_devinit_new },
+	.fault    = { 0x00000001, tu102_fault_new },
+	.fb       = { 0x00000001, gv100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.gsp      = { 0x00000001, gv100_gsp_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, tu102_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000001f, tu102_ce_new },
+	.disp     = { 0x00000001, tu102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+	.fifo     = { 0x00000001, tu102_fifo_new },
+	.gr       = { 0x00000001, tu102_gr_new },
+	.nvdec    = { 0x00000003, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sec2     = { 0x00000001, tu102_sec2_new },
 };
 
 static const struct nvkm_device_chip
 nv166_chipset = {
 	.name = "TU106",
-	.acr = tu102_acr_new,
-	.bar = tu102_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = tu102_devinit_new,
-	.fault = tu102_fault_new,
-	.fb = gv100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.gsp = gv100_gsp_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = tu102_mc_new,
-	.mmu = tu102_mmu_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.therm = gp100_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = tu102_ce_new,
-	.ce[1] = tu102_ce_new,
-	.ce[2] = tu102_ce_new,
-	.ce[3] = tu102_ce_new,
-	.ce[4] = tu102_ce_new,
-	.disp = tu102_disp_new,
-	.dma = gv100_dma_new,
-	.fifo = tu102_fifo_new,
-	.gr = tu102_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvdec[1] = gm107_nvdec_new,
-	.nvdec[2] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sec2 = tu102_sec2_new,
+	.acr      = { 0x00000001, tu102_acr_new },
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, tu102_devinit_new },
+	.fault    = { 0x00000001, tu102_fault_new },
+	.fb       = { 0x00000001, gv100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.gsp      = { 0x00000001, gv100_gsp_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, tu102_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000001f, tu102_ce_new },
+	.disp     = { 0x00000001, tu102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+	.fifo     = { 0x00000001, tu102_fifo_new },
+	.gr       = { 0x00000001, tu102_gr_new },
+	.nvdec    = { 0x00000007, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sec2     = { 0x00000001, tu102_sec2_new },
 };
 
 static const struct nvkm_device_chip
 nv167_chipset = {
 	.name = "TU117",
-	.acr = tu102_acr_new,
-	.bar = tu102_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = tu102_devinit_new,
-	.fault = tu102_fault_new,
-	.fb = gv100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.gsp = gv100_gsp_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = tu102_mc_new,
-	.mmu = tu102_mmu_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.therm = gp100_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = tu102_ce_new,
-	.ce[1] = tu102_ce_new,
-	.ce[2] = tu102_ce_new,
-	.ce[3] = tu102_ce_new,
-	.ce[4] = tu102_ce_new,
-	.disp = tu102_disp_new,
-	.dma = gv100_dma_new,
-	.fifo = tu102_fifo_new,
-	.gr = tu102_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sec2 = tu102_sec2_new,
+	.acr      = { 0x00000001, tu102_acr_new },
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, tu102_devinit_new },
+	.fault    = { 0x00000001, tu102_fault_new },
+	.fb       = { 0x00000001, gv100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.gsp      = { 0x00000001, gv100_gsp_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, tu102_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000001f, tu102_ce_new },
+	.disp     = { 0x00000001, tu102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+	.fifo     = { 0x00000001, tu102_fifo_new },
+	.gr       = { 0x00000001, tu102_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sec2     = { 0x00000001, tu102_sec2_new },
 };
 
 static const struct nvkm_device_chip
 nv168_chipset = {
 	.name = "TU116",
-	.acr = tu102_acr_new,
-	.bar = tu102_bar_new,
-	.bios = nvkm_bios_new,
-	.bus = gf100_bus_new,
-	.devinit = tu102_devinit_new,
-	.fault = tu102_fault_new,
-	.fb = gv100_fb_new,
-	.fuse = gm107_fuse_new,
-	.gpio = gk104_gpio_new,
-	.gsp = gv100_gsp_new,
-	.i2c = gm200_i2c_new,
-	.ibus = gm200_ibus_new,
-	.imem = nv50_instmem_new,
-	.ltc = gp102_ltc_new,
-	.mc = tu102_mc_new,
-	.mmu = tu102_mmu_new,
-	.pci = gp100_pci_new,
-	.pmu = gp102_pmu_new,
-	.therm = gp100_therm_new,
-	.timer = gk20a_timer_new,
-	.top = gk104_top_new,
-	.ce[0] = tu102_ce_new,
-	.ce[1] = tu102_ce_new,
-	.ce[2] = tu102_ce_new,
-	.ce[3] = tu102_ce_new,
-	.ce[4] = tu102_ce_new,
-	.disp = tu102_disp_new,
-	.dma = gv100_dma_new,
-	.fifo = tu102_fifo_new,
-	.gr = tu102_gr_new,
-	.nvdec[0] = gm107_nvdec_new,
-	.nvenc[0] = gm107_nvenc_new,
-	.sec2 = tu102_sec2_new,
+	.acr      = { 0x00000001, tu102_acr_new },
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.bus      = { 0x00000001, gf100_bus_new },
+	.devinit  = { 0x00000001, tu102_devinit_new },
+	.fault    = { 0x00000001, tu102_fault_new },
+	.fb       = { 0x00000001, gv100_fb_new },
+	.fuse     = { 0x00000001, gm107_fuse_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.gsp      = { 0x00000001, gv100_gsp_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.ltc      = { 0x00000001, gp102_ltc_new },
+	.mc       = { 0x00000001, tu102_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.pmu      = { 0x00000001, gp102_pmu_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.therm    = { 0x00000001, gp100_therm_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, gk104_top_new },
+	.ce       = { 0x0000001f, tu102_ce_new },
+	.disp     = { 0x00000001, tu102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+	.fifo     = { 0x00000001, tu102_fifo_new },
+	.gr       = { 0x00000001, tu102_gr_new },
+	.nvdec    = { 0x00000001, gm107_nvdec_new },
+	.nvenc    = { 0x00000001, gm107_nvenc_new },
+	.sec2     = { 0x00000001, tu102_sec2_new },
+};
+
+static const struct nvkm_device_chip
+nv170_chipset = {
+	.name = "GA100",
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.devinit  = { 0x00000001, ga100_devinit_new },
+	.fb       = { 0x00000001, ga100_fb_new },
+	.gpio     = { 0x00000001, gk104_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, ga100_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, ga100_top_new },
+};
+
+static const struct nvkm_device_chip
+nv172_chipset = {
+	.name = "GA102",
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.devinit  = { 0x00000001, ga100_devinit_new },
+	.fb       = { 0x00000001, ga102_fb_new },
+	.gpio     = { 0x00000001, ga102_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, ga100_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, ga100_top_new },
+	.disp     = { 0x00000001, ga102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
+};
+
+static const struct nvkm_device_chip
+nv174_chipset = {
+	.name = "GA104",
+	.bar      = { 0x00000001, tu102_bar_new },
+	.bios     = { 0x00000001, nvkm_bios_new },
+	.devinit  = { 0x00000001, ga100_devinit_new },
+	.fb       = { 0x00000001, ga102_fb_new },
+	.gpio     = { 0x00000001, ga102_gpio_new },
+	.i2c      = { 0x00000001, gm200_i2c_new },
+	.imem     = { 0x00000001, nv50_instmem_new },
+	.mc       = { 0x00000001, ga100_mc_new },
+	.mmu      = { 0x00000001, tu102_mmu_new },
+	.pci      = { 0x00000001, gp100_pci_new },
+	.privring = { 0x00000001, gm200_privring_new },
+	.timer    = { 0x00000001, gk20a_timer_new },
+	.top      = { 0x00000001, ga100_top_new },
+	.disp     = { 0x00000001, ga102_disp_new },
+	.dma      = { 0x00000001, gv100_dma_new },
 };
 
 static int
@@ -2671,97 +2643,24 @@ nvkm_device_event_func = {
 };
 
 struct nvkm_subdev *
-nvkm_device_subdev(struct nvkm_device *device, int index)
+nvkm_device_subdev(struct nvkm_device *device, int type, int inst)
 {
-	struct nvkm_engine *engine;
-
-	if (device->disable_mask & (1ULL << index))
-		return NULL;
-
-	switch (index) {
-#define _(n,p,m) case NVKM_SUBDEV_##n: if (p) return (m); break
-	_(ACR     , device->acr     , &device->acr->subdev);
-	_(BAR     , device->bar     , &device->bar->subdev);
-	_(VBIOS   , device->bios    , &device->bios->subdev);
-	_(BUS     , device->bus     , &device->bus->subdev);
-	_(CLK     , device->clk     , &device->clk->subdev);
-	_(DEVINIT , device->devinit , &device->devinit->subdev);
-	_(FAULT   , device->fault   , &device->fault->subdev);
-	_(FB      , device->fb      , &device->fb->subdev);
-	_(FUSE    , device->fuse    , &device->fuse->subdev);
-	_(GPIO    , device->gpio    , &device->gpio->subdev);
-	_(GSP     , device->gsp     , &device->gsp->subdev);
-	_(I2C     , device->i2c     , &device->i2c->subdev);
-	_(IBUS    , device->ibus    ,  device->ibus);
-	_(ICCSENSE, device->iccsense, &device->iccsense->subdev);
-	_(INSTMEM , device->imem    , &device->imem->subdev);
-	_(LTC     , device->ltc     , &device->ltc->subdev);
-	_(MC      , device->mc      , &device->mc->subdev);
-	_(MMU     , device->mmu     , &device->mmu->subdev);
-	_(MXM     , device->mxm     ,  device->mxm);
-	_(PCI     , device->pci     , &device->pci->subdev);
-	_(PMU     , device->pmu     , &device->pmu->subdev);
-	_(THERM   , device->therm   , &device->therm->subdev);
-	_(TIMER   , device->timer   , &device->timer->subdev);
-	_(TOP     , device->top     , &device->top->subdev);
-	_(VOLT    , device->volt    , &device->volt->subdev);
-#undef _
-	default:
-		engine = nvkm_device_engine(device, index);
-		if (engine)
-			return &engine->subdev;
-		break;
+	struct nvkm_subdev *subdev;
+
+	list_for_each_entry(subdev, &device->subdev, head) {
+		if (subdev->type == type && subdev->inst == inst)
+			return subdev;
 	}
+
 	return NULL;
 }
 
 struct nvkm_engine *
-nvkm_device_engine(struct nvkm_device *device, int index)
+nvkm_device_engine(struct nvkm_device *device, int type, int inst)
 {
-	if (device->disable_mask & (1ULL << index))
-		return NULL;
-
-	switch (index) {
-#define _(n,p,m) case NVKM_ENGINE_##n: if (p) return (m); break
-	_(BSP    , device->bsp     ,  device->bsp);
-	_(CE0    , device->ce[0]   ,  device->ce[0]);
-	_(CE1    , device->ce[1]   ,  device->ce[1]);
-	_(CE2    , device->ce[2]   ,  device->ce[2]);
-	_(CE3    , device->ce[3]   ,  device->ce[3]);
-	_(CE4    , device->ce[4]   ,  device->ce[4]);
-	_(CE5    , device->ce[5]   ,  device->ce[5]);
-	_(CE6    , device->ce[6]   ,  device->ce[6]);
-	_(CE7    , device->ce[7]   ,  device->ce[7]);
-	_(CE8    , device->ce[8]   ,  device->ce[8]);
-	_(CIPHER , device->cipher  ,  device->cipher);
-	_(DISP   , device->disp    , &device->disp->engine);
-	_(DMAOBJ , device->dma     , &device->dma->engine);
-	_(FIFO   , device->fifo    , &device->fifo->engine);
-	_(GR     , device->gr      , &device->gr->engine);
-	_(IFB    , device->ifb     ,  device->ifb);
-	_(ME     , device->me      ,  device->me);
-	_(MPEG   , device->mpeg    ,  device->mpeg);
-	_(MSENC  , device->msenc   ,  device->msenc);
-	_(MSPDEC , device->mspdec  ,  device->mspdec);
-	_(MSPPP  , device->msppp   ,  device->msppp);
-	_(MSVLD  , device->msvld   ,  device->msvld);
-	_(NVENC0 , device->nvenc[0], &device->nvenc[0]->engine);
-	_(NVENC1 , device->nvenc[1], &device->nvenc[1]->engine);
-	_(NVENC2 , device->nvenc[2], &device->nvenc[2]->engine);
-	_(NVDEC0 , device->nvdec[0], &device->nvdec[0]->engine);
-	_(NVDEC1 , device->nvdec[1], &device->nvdec[1]->engine);
-	_(NVDEC2 , device->nvdec[2], &device->nvdec[2]->engine);
-	_(PM     , device->pm      , &device->pm->engine);
-	_(SEC    , device->sec     ,  device->sec);
-	_(SEC2   , device->sec2    , &device->sec2->engine);
-	_(SW     , device->sw      , &device->sw->engine);
-	_(VIC    , device->vic     ,  device->vic);
-	_(VP     , device->vp      ,  device->vp);
-#undef _
-	default:
-		WARN_ON(1);
-		break;
-	}
+	struct nvkm_subdev *subdev = nvkm_device_subdev(device, type, inst);
+	if (subdev && subdev->func == &nvkm_engine)
+		return container_of(subdev, struct nvkm_engine, subdev);
 	return NULL;
 }
 
@@ -2770,7 +2669,7 @@ nvkm_device_fini(struct nvkm_device *device, bool suspend)
 {
 	const char *action = suspend ? "suspend" : "fini";
 	struct nvkm_subdev *subdev;
-	int ret, i;
+	int ret;
 	s64 time;
 
 	nvdev_trace(device, "%s running...\n", action);
@@ -2778,12 +2677,10 @@ nvkm_device_fini(struct nvkm_device *device, bool suspend)
 
 	nvkm_acpi_fini(device);
 
-	for (i = NVKM_SUBDEV_NR - 1; i >= 0; i--) {
-		if ((subdev = nvkm_device_subdev(device, i))) {
-			ret = nvkm_subdev_fini(subdev, suspend);
-			if (ret && suspend)
-				goto fail;
-		}
+	list_for_each_entry_reverse(subdev, &device->subdev, head) {
+		ret = nvkm_subdev_fini(subdev, suspend);
+		if (ret && suspend)
+			goto fail;
 	}
 
 	nvkm_therm_clkgate_fini(device->therm, suspend);
@@ -2796,13 +2693,11 @@ nvkm_device_fini(struct nvkm_device *device, bool suspend)
 	return 0;
 
 fail:
-	do {
-		if ((subdev = nvkm_device_subdev(device, i))) {
-			int rret = nvkm_subdev_init(subdev);
-			if (rret)
-				nvkm_fatal(subdev, "failed restart, %d\n", ret);
-		}
-	} while (++i < NVKM_SUBDEV_NR);
+	list_for_each_entry_from(subdev, &device->subdev, head) {
+		int rret = nvkm_subdev_init(subdev);
+		if (rret)
+			nvkm_fatal(subdev, "failed restart, %d\n", ret);
+	}
 
 	nvdev_trace(device, "%s failed with %d\n", action, ret);
 	return ret;
@@ -2812,7 +2707,7 @@ static int
 nvkm_device_preinit(struct nvkm_device *device)
 {
 	struct nvkm_subdev *subdev;
-	int ret, i;
+	int ret;
 	s64 time;
 
 	nvdev_trace(device, "preinit running...\n");
@@ -2824,15 +2719,13 @@ nvkm_device_preinit(struct nvkm_device *device)
 			goto fail;
 	}
 
-	for (i = 0; i < NVKM_SUBDEV_NR; i++) {
-		if ((subdev = nvkm_device_subdev(device, i))) {
-			ret = nvkm_subdev_preinit(subdev);
-			if (ret)
-				goto fail;
-		}
+	list_for_each_entry(subdev, &device->subdev, head) {
+		ret = nvkm_subdev_preinit(subdev);
+		if (ret)
+			goto fail;
 	}
 
-	ret = nvkm_devinit_post(device->devinit, &device->disable_mask);
+	ret = nvkm_devinit_post(device->devinit);
 	if (ret)
 		goto fail;
 
@@ -2849,7 +2742,7 @@ int
 nvkm_device_init(struct nvkm_device *device)
 {
 	struct nvkm_subdev *subdev;
-	int ret, i;
+	int ret;
 	s64 time;
 
 	ret = nvkm_device_preinit(device);
@@ -2867,12 +2760,10 @@ nvkm_device_init(struct nvkm_device *device)
 			goto fail;
 	}
 
-	for (i = 0; i < NVKM_SUBDEV_NR; i++) {
-		if ((subdev = nvkm_device_subdev(device, i))) {
-			ret = nvkm_subdev_init(subdev);
-			if (ret)
-				goto fail_subdev;
-		}
+	list_for_each_entry(subdev, &device->subdev, head) {
+		ret = nvkm_subdev_init(subdev);
+		if (ret)
+			goto fail_subdev;
 	}
 
 	nvkm_acpi_init(device);
@@ -2883,11 +2774,8 @@ nvkm_device_init(struct nvkm_device *device)
 	return 0;
 
 fail_subdev:
-	do {
-		if ((subdev = nvkm_device_subdev(device, i)))
-			nvkm_subdev_fini(subdev, false);
-	} while (--i >= 0);
-
+	list_for_each_entry_from(subdev, &device->subdev, head)
+		nvkm_subdev_fini(subdev, false);
 fail:
 	nvkm_device_fini(device, false);
 
@@ -2899,15 +2787,12 @@ void
 nvkm_device_del(struct nvkm_device **pdevice)
 {
 	struct nvkm_device *device = *pdevice;
-	int i;
+	struct nvkm_subdev *subdev, *subtmp;
 	if (device) {
 		mutex_lock(&nv_devices_mutex);
-		device->disable_mask = 0;
-		for (i = NVKM_SUBDEV_NR - 1; i >= 0; i--) {
-			struct nvkm_subdev *subdev =
-				nvkm_device_subdev(device, i);
+
+		list_for_each_entry_safe_reverse(subdev, subtmp, &device->subdev, head)
 			nvkm_subdev_del(&subdev);
-		}
 
 		nvkm_event_fini(&device->event);
 
@@ -2966,7 +2851,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 	struct nvkm_subdev *subdev;
 	u64 mmio_base, mmio_size;
 	u32 boot0, boot1, strap;
-	int ret = -EEXIST, i;
+	int ret = -EEXIST, j;
 	unsigned chipset;
 
 	mutex_lock(&nv_devices_mutex);
@@ -2983,6 +2868,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 	device->name = name;
 	list_add_tail(&device->head, &nv_devices);
 	device->debug = nvkm_dbgopt(device->dbgopt, "device");
+	INIT_LIST_HEAD(&device->subdev);
 
 	ret = nvkm_event_init(&nvkm_device_event_func, 1, 1, &device->event);
 	if (ret)
@@ -3063,6 +2949,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 			case 0x130: device->card_type = GP100; break;
 			case 0x140: device->card_type = GV100; break;
 			case 0x160: device->card_type = TU100; break;
+			case 0x170: device->card_type = GA100; break;
 			default:
 				break;
 			}
@@ -3160,10 +3047,23 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 		case 0x166: device->chip = &nv166_chipset; break;
 		case 0x167: device->chip = &nv167_chipset; break;
 		case 0x168: device->chip = &nv168_chipset; break;
+		case 0x172: device->chip = &nv172_chipset; break;
+		case 0x174: device->chip = &nv174_chipset; break;
 		default:
-			nvdev_error(device, "unknown chipset (%08x)\n", boot0);
-			ret = -ENODEV;
-			goto done;
+			if (nvkm_boolopt(device->cfgopt, "NvEnableUnsupportedChipsets", false)) {
+				switch (device->chipset) {
+				case 0x170: device->chip = &nv170_chipset; break;
+				default:
+					break;
+				}
+			}
+
+			if (!device->chip) {
+				nvdev_error(device, "unknown chipset (%08x)\n", boot0);
+				ret = -ENODEV;
+				goto done;
+			}
+			break;
 		}
 
 		nvdev_info(device, "NVIDIA %s (%08x)\n",
@@ -3202,88 +3102,46 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 
 	mutex_init(&device->mutex);
 
-	for (i = 0; i < NVKM_SUBDEV_NR; i++) {
-#define _(s,m) case s:                                                         \
-	if (device->chip->m && (subdev_mask & (1ULL << (s)))) {                \
-		ret = device->chip->m(device, (s), &device->m);                \
-		if (ret) {                                                     \
-			subdev = nvkm_device_subdev(device, (s));              \
-			nvkm_subdev_del(&subdev);                              \
-			device->m = NULL;                                      \
-			if (ret != -ENODEV) {                                  \
-				nvdev_error(device, "%s ctor failed, %d\n",    \
-					    nvkm_subdev_name[s], ret);         \
-				goto done;                                     \
-			}                                                      \
-		}                                                              \
-	}                                                                      \
-	break
-		switch (i) {
-		_(NVKM_SUBDEV_ACR     ,      acr);
-		_(NVKM_SUBDEV_BAR     ,      bar);
-		_(NVKM_SUBDEV_VBIOS   ,     bios);
-		_(NVKM_SUBDEV_BUS     ,      bus);
-		_(NVKM_SUBDEV_CLK     ,      clk);
-		_(NVKM_SUBDEV_DEVINIT ,  devinit);
-		_(NVKM_SUBDEV_FAULT   ,    fault);
-		_(NVKM_SUBDEV_FB      ,       fb);
-		_(NVKM_SUBDEV_FUSE    ,     fuse);
-		_(NVKM_SUBDEV_GPIO    ,     gpio);
-		_(NVKM_SUBDEV_GSP     ,      gsp);
-		_(NVKM_SUBDEV_I2C     ,      i2c);
-		_(NVKM_SUBDEV_IBUS    ,     ibus);
-		_(NVKM_SUBDEV_ICCSENSE, iccsense);
-		_(NVKM_SUBDEV_INSTMEM ,     imem);
-		_(NVKM_SUBDEV_LTC     ,      ltc);
-		_(NVKM_SUBDEV_MC      ,       mc);
-		_(NVKM_SUBDEV_MMU     ,      mmu);
-		_(NVKM_SUBDEV_MXM     ,      mxm);
-		_(NVKM_SUBDEV_PCI     ,      pci);
-		_(NVKM_SUBDEV_PMU     ,      pmu);
-		_(NVKM_SUBDEV_THERM   ,    therm);
-		_(NVKM_SUBDEV_TIMER   ,    timer);
-		_(NVKM_SUBDEV_TOP     ,      top);
-		_(NVKM_SUBDEV_VOLT    ,     volt);
-		_(NVKM_ENGINE_BSP     ,      bsp);
-		_(NVKM_ENGINE_CE0     ,    ce[0]);
-		_(NVKM_ENGINE_CE1     ,    ce[1]);
-		_(NVKM_ENGINE_CE2     ,    ce[2]);
-		_(NVKM_ENGINE_CE3     ,    ce[3]);
-		_(NVKM_ENGINE_CE4     ,    ce[4]);
-		_(NVKM_ENGINE_CE5     ,    ce[5]);
-		_(NVKM_ENGINE_CE6     ,    ce[6]);
-		_(NVKM_ENGINE_CE7     ,    ce[7]);
-		_(NVKM_ENGINE_CE8     ,    ce[8]);
-		_(NVKM_ENGINE_CIPHER  ,   cipher);
-		_(NVKM_ENGINE_DISP    ,     disp);
-		_(NVKM_ENGINE_DMAOBJ  ,      dma);
-		_(NVKM_ENGINE_FIFO    ,     fifo);
-		_(NVKM_ENGINE_GR      ,       gr);
-		_(NVKM_ENGINE_IFB     ,      ifb);
-		_(NVKM_ENGINE_ME      ,       me);
-		_(NVKM_ENGINE_MPEG    ,     mpeg);
-		_(NVKM_ENGINE_MSENC   ,    msenc);
-		_(NVKM_ENGINE_MSPDEC  ,   mspdec);
-		_(NVKM_ENGINE_MSPPP   ,    msppp);
-		_(NVKM_ENGINE_MSVLD   ,    msvld);
-		_(NVKM_ENGINE_NVENC0  , nvenc[0]);
-		_(NVKM_ENGINE_NVENC1  , nvenc[1]);
-		_(NVKM_ENGINE_NVENC2  , nvenc[2]);
-		_(NVKM_ENGINE_NVDEC0  , nvdec[0]);
-		_(NVKM_ENGINE_NVDEC1  , nvdec[1]);
-		_(NVKM_ENGINE_NVDEC2  , nvdec[2]);
-		_(NVKM_ENGINE_PM      ,       pm);
-		_(NVKM_ENGINE_SEC     ,      sec);
-		_(NVKM_ENGINE_SEC2    ,     sec2);
-		_(NVKM_ENGINE_SW      ,       sw);
-		_(NVKM_ENGINE_VIC     ,      vic);
-		_(NVKM_ENGINE_VP      ,       vp);
-		default:
-			WARN_ON(1);
-			continue;
-		}
-#undef _
+#define NVKM_LAYOUT_ONCE(type,data,ptr)                                                      \
+	if (device->chip->ptr.inst && (subdev_mask & (BIT_ULL(type)))) {                     \
+		WARN_ON(device->chip->ptr.inst != 0x00000001);                               \
+		ret = device->chip->ptr.ctor(device, (type), -1, &device->ptr);              \
+		subdev = nvkm_device_subdev(device, (type), 0);                              \
+		if (ret) {                                                                   \
+			nvkm_subdev_del(&subdev);                                            \
+			device->ptr = NULL;                                                  \
+			if (ret != -ENODEV) {                                                \
+				nvdev_error(device, "%s ctor failed: %d\n",                  \
+					    nvkm_subdev_type[(type)], ret);                  \
+				goto done;                                                   \
+			}                                                                    \
+		} else {                                                                     \
+			subdev->pself = (void **)&device->ptr;                               \
+		}                                                                            \
+	}
+#define NVKM_LAYOUT_INST(type,data,ptr,cnt)                                                  \
+	WARN_ON(device->chip->ptr.inst & ~((1 << ARRAY_SIZE(device->ptr)) - 1));             \
+	for (j = 0; device->chip->ptr.inst && j < ARRAY_SIZE(device->ptr); j++) {            \
+		if ((device->chip->ptr.inst & BIT(j)) && (subdev_mask & BIT_ULL(type))) {    \
+			int inst = (device->chip->ptr.inst == 1) ? -1 : (j);                 \
+			ret = device->chip->ptr.ctor(device, (type), inst, &device->ptr[j]); \
+			subdev = nvkm_device_subdev(device, (type), (j));                    \
+			if (ret) {                                                           \
+				nvkm_subdev_del(&subdev);                                    \
+				device->ptr[j] = NULL;                                       \
+				if (ret != -ENODEV) {                                        \
+					nvdev_error(device, "%s%d ctor failed: %d\n",        \
+						    nvkm_subdev_type[(type)], (j), ret);     \
+					goto done;                                           \
+				}                                                            \
+			} else {                                                             \
+				subdev->pself = (void **)&device->ptr[j];                    \
+			}                                                                    \
+		}                                                                            \
 	}
+#include <core/layout.h>
+#undef NVKM_LAYOUT_INST
+#undef NVKM_LAYOUT_ONCE
 
 	ret = 0;
 done:
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h
index 54eab5e04230..93949b3c7214 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h
@@ -15,7 +15,6 @@
 #include <subdev/gpio.h>
 #include <subdev/gsp.h>
 #include <subdev/i2c.h>
-#include <subdev/ibus.h>
 #include <subdev/iccsense.h>
 #include <subdev/instmem.h>
 #include <subdev/ltc.h>
@@ -24,6 +23,7 @@
 #include <subdev/mxm.h>
 #include <subdev/pci.h>
 #include <subdev/pmu.h>
+#include <subdev/privring.h>
 #include <subdev/therm.h>
 #include <subdev/timer.h>
 #include <subdev/top.h>
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c
index 03c6d9aef075..fea9d8f2b10c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c
@@ -43,15 +43,15 @@ static int
 nvkm_udevice_info_subdev(struct nvkm_device *device, u64 mthd, u64 *data)
 {
 	struct nvkm_subdev *subdev;
-	enum nvkm_devidx subidx;
+	enum nvkm_subdev_type type;
 
 	switch (mthd & NV_DEVICE_INFO_UNIT) {
-	case NV_DEVICE_FIFO(0): subidx = NVKM_ENGINE_FIFO; break;
+	case NV_DEVICE_HOST(0): type = NVKM_ENGINE_FIFO; break;
 	default:
 		return -EINVAL;
 	}
 
-	subdev = nvkm_device_subdev(device, subidx);
+	subdev = nvkm_device_subdev(device, type, 0);
 	if (subdev)
 		return nvkm_subdev_info(subdev, mthd, data);
 	return -ENODEV;
@@ -66,37 +66,7 @@ nvkm_udevice_info_v1(struct nvkm_device *device,
 			args->mthd = NV_DEVICE_INFO_INVALID;
 		return;
 	}
-
-	switch (args->mthd) {
-#define ENGINE__(A,B,C) NV_DEVICE_INFO_ENGINE_##A: { int _i;                   \
-	for (_i = (B), args->data = 0ULL; _i <= (C); _i++) {                   \
-		if (nvkm_device_engine(device, _i))                            \
-			args->data |= BIT_ULL(_i);                             \
-	}                                                                      \
-}
-#define ENGINE_A(A) ENGINE__(A, NVKM_ENGINE_##A   , NVKM_ENGINE_##A)
-#define ENGINE_B(A) ENGINE__(A, NVKM_ENGINE_##A##0, NVKM_ENGINE_##A##_LAST)
-	case ENGINE_A(SW    ); break;
-	case ENGINE_A(GR    ); break;
-	case ENGINE_A(MPEG  ); break;
-	case ENGINE_A(ME    ); break;
-	case ENGINE_A(CIPHER); break;
-	case ENGINE_A(BSP   ); break;
-	case ENGINE_A(VP    ); break;
-	case ENGINE_B(CE    ); break;
-	case ENGINE_A(SEC   ); break;
-	case ENGINE_A(MSVLD ); break;
-	case ENGINE_A(MSPDEC); break;
-	case ENGINE_A(MSPPP ); break;
-	case ENGINE_A(MSENC ); break;
-	case ENGINE_A(VIC   ); break;
-	case ENGINE_A(SEC2  ); break;
-	case ENGINE_B(NVDEC ); break;
-	case ENGINE_B(NVENC ); break;
-	default:
-		args->mthd = NV_DEVICE_INFO_INVALID;
-		break;
-	}
+	args->mthd = NV_DEVICE_INFO_INVALID;
 }
 
 static int
@@ -176,6 +146,7 @@ nvkm_udevice_info(struct nvkm_udevice *udev, void *data, u32 size)
 	case GP100: args->v0.family = NV_DEVICE_INFO_V0_PASCAL; break;
 	case GV100: args->v0.family = NV_DEVICE_INFO_V0_VOLTA; break;
 	case TU100: args->v0.family = NV_DEVICE_INFO_V0_TURING; break;
+	case GA100: args->v0.family = NV_DEVICE_INFO_V0_AMPERE; break;
 	default:
 		args->v0.family = 0;
 		break;
@@ -356,7 +327,7 @@ nvkm_udevice_child_get(struct nvkm_object *object, int index,
 	int i;
 
 	for (; i = __ffs64(mask), mask && !sclass; mask &= ~(1ULL << i)) {
-		if (!(engine = nvkm_device_engine(device, i)) ||
+		if (!(engine = nvkm_device_engine(device, i, 0)) ||
 		    !(engine->func->base.sclass))
 			continue;
 		oclass->engine = engine;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
index cf075311cdd2..b03f043efe26 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
@@ -17,6 +17,7 @@ nvkm-y += nvkm/engine/disp/gp100.o
 nvkm-y += nvkm/engine/disp/gp102.o
 nvkm-y += nvkm/engine/disp/gv100.o
 nvkm-y += nvkm/engine/disp/tu102.o
+nvkm-y += nvkm/engine/disp/ga102.o
 nvkm-y += nvkm/engine/disp/vga.o
 
 nvkm-y += nvkm/engine/disp/head.o
@@ -42,6 +43,7 @@ nvkm-y += nvkm/engine/disp/sorgm200.o
 nvkm-y += nvkm/engine/disp/sorgp100.o
 nvkm-y += nvkm/engine/disp/sorgv100.o
 nvkm-y += nvkm/engine/disp/sortu102.o
+nvkm-y += nvkm/engine/disp/sorga102.o
 
 nvkm-y += nvkm/engine/disp/outp.o
 nvkm-y += nvkm/engine/disp/dp.o
@@ -75,6 +77,7 @@ nvkm-y += nvkm/engine/disp/rootgp100.o
 nvkm-y += nvkm/engine/disp/rootgp102.o
 nvkm-y += nvkm/engine/disp/rootgv100.o
 nvkm-y += nvkm/engine/disp/roottu102.o
+nvkm-y += nvkm/engine/disp/rootga102.o
 
 nvkm-y += nvkm/engine/disp/capsgv100.o
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/base.c
index cbd33e87b799..5daa77755276 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/base.c
@@ -149,10 +149,10 @@ static void
 nvkm_disp_class_del(struct nvkm_oproxy *oproxy)
 {
 	struct nvkm_disp *disp = nvkm_disp(oproxy->base.engine);
-	mutex_lock(&disp->engine.subdev.mutex);
-	if (disp->client == oproxy)
-		disp->client = NULL;
-	mutex_unlock(&disp->engine.subdev.mutex);
+	spin_lock(&disp->client.lock);
+	if (disp->client.object == oproxy)
+		disp->client.object = NULL;
+	spin_unlock(&disp->client.lock);
 }
 
 static const struct nvkm_oproxy_func
@@ -175,13 +175,13 @@ nvkm_disp_class_new(struct nvkm_device *device,
 		return ret;
 	*pobject = &oproxy->base;
 
-	mutex_lock(&disp->engine.subdev.mutex);
-	if (disp->client) {
-		mutex_unlock(&disp->engine.subdev.mutex);
+	spin_lock(&disp->client.lock);
+	if (disp->client.object) {
+		spin_unlock(&disp->client.lock);
 		return -EBUSY;
 	}
-	disp->client = oproxy;
-	mutex_unlock(&disp->engine.subdev.mutex);
+	disp->client.object = oproxy;
+	spin_unlock(&disp->client.lock);
 
 	return sclass->ctor(disp, oclass, data, size, &oproxy->object);
 }
@@ -473,21 +473,22 @@ nvkm_disp = {
 
 int
 nvkm_disp_ctor(const struct nvkm_disp_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_disp *disp)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_disp *disp)
 {
 	disp->func = func;
 	INIT_LIST_HEAD(&disp->head);
 	INIT_LIST_HEAD(&disp->ior);
 	INIT_LIST_HEAD(&disp->outp);
 	INIT_LIST_HEAD(&disp->conn);
-	return nvkm_engine_ctor(&nvkm_disp, device, index, true, &disp->engine);
+	spin_lock_init(&disp->client.lock);
+	return nvkm_engine_ctor(&nvkm_disp, device, type, inst, true, &disp->engine);
 }
 
 int
 nvkm_disp_new_(const struct nvkm_disp_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_disp **pdisp)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_disp **pdisp)
 {
 	if (!(*pdisp = kzalloc(sizeof(**pdisp), GFP_KERNEL)))
 		return -ENOMEM;
-	return nvkm_disp_ctor(func, device, index, *pdisp);
+	return nvkm_disp_ctor(func, device, type, inst, *pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.c
index 50e3539f33d2..a7a7eb041515 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.c
@@ -278,7 +278,7 @@ nv50_disp_chan_child_get(struct nvkm_object *object, int index,
 	const struct nvkm_device_oclass *oclass = NULL;
 
 	if (chan->func->bind)
-		sclass->engine = nvkm_device_engine(device, NVKM_ENGINE_DMAOBJ);
+		sclass->engine = nvkm_device_engine(device, NVKM_ENGINE_DMAOBJ, 0);
 	else
 		sclass->engine = NULL;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.c
index 3800aeb507d0..55fbfe28c6dc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.c
@@ -33,6 +33,12 @@
 
 #include <nvif/event.h>
 
+/* IED scripts are no longer used by UEFI/RM from Ampere, but have been updated for
+ * the x86 option ROM.  However, the relevant VBIOS table versions weren't modified,
+ * so we're unable to detect this in a nice way.
+ */
+#define AMPERE_IED_HACK(disp) ((disp)->engine.subdev.device->card_type >= GA100)
+
 struct lt_state {
 	struct nvkm_dp *dp;
 	u8  stat[6];
@@ -238,6 +244,19 @@ nvkm_dp_train_links(struct nvkm_dp *dp)
 		dp->dpcd[DPCD_RC02] &= ~DPCD_RC02_TPS3_SUPPORTED;
 	lt.pc2 = dp->dpcd[DPCD_RC02] & DPCD_RC02_TPS3_SUPPORTED;
 
+	if (AMPERE_IED_HACK(disp) && (lnkcmp = lt.dp->info.script[0])) {
+		/* Execute BeforeLinkTraining script from DP Info table. */
+		while (ior->dp.bw < nvbios_rd08(bios, lnkcmp))
+			lnkcmp += 3;
+		lnkcmp = nvbios_rd16(bios, lnkcmp + 1);
+
+		nvbios_init(&dp->outp.disp->engine.subdev, lnkcmp,
+			init.outp = &dp->outp.info;
+			init.or   = ior->id;
+			init.link = ior->asy.link;
+		);
+	}
+
 	/* Set desired link configuration on the source. */
 	if ((lnkcmp = lt.dp->info.lnkcmp)) {
 		if (dp->version < 0x30) {
@@ -316,12 +335,14 @@ nvkm_dp_train_init(struct nvkm_dp *dp)
 		);
 	}
 
-	/* Execute BeforeLinkTraining script from DP Info table. */
-	nvbios_init(&dp->outp.disp->engine.subdev, dp->info.script[0],
-		init.outp = &dp->outp.info;
-		init.or   = dp->outp.ior->id;
-		init.link = dp->outp.ior->asy.link;
-	);
+	if (!AMPERE_IED_HACK(dp->outp.disp)) {
+		/* Execute BeforeLinkTraining script from DP Info table. */
+		nvbios_init(&dp->outp.disp->engine.subdev, dp->info.script[0],
+			init.outp = &dp->outp.info;
+			init.or   = dp->outp.ior->id;
+			init.link = dp->outp.ior->asy.link;
+		);
+	}
 }
 
 static const struct dp_rates {
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/g84.c
index 731f188fc1ee..156bbe8b2de3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/g84.c
@@ -41,7 +41,8 @@ g84_disp = {
 };
 
 int
-g84_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+g84_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&g84_disp, device, index, pdisp);
+	return nv50_disp_new_(&g84_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/g94.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/g94.c
index def54fe1951e..3425b5d3bc72 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/g94.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/g94.c
@@ -41,7 +41,8 @@ g94_disp = {
 };
 
 int
-g94_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+g94_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&g94_disp, device, index, pdisp);
+	return nv50_disp_new_(&g94_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ga102.c
new file mode 100644
index 000000000000..68aa52588d92
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ga102.c
@@ -0,0 +1,47 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "nv50.h"
+#include "head.h"
+#include "ior.h"
+#include "channv50.h"
+#include "rootnv50.h"
+
+static const struct nv50_disp_func
+ga102_disp = {
+	.init = tu102_disp_init,
+	.fini = gv100_disp_fini,
+	.intr = gv100_disp_intr,
+	.uevent = &gv100_disp_chan_uevent,
+	.super = gv100_disp_super,
+	.root = &ga102_disp_root_oclass,
+	.wndw = { .cnt = gv100_disp_wndw_cnt },
+	.head = { .cnt = gv100_head_cnt, .new = gv100_head_new },
+	.sor = { .cnt = gv100_sor_cnt, .new = ga102_sor_new },
+	.ramht_size = 0x2000,
+};
+
+int
+ga102_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
+{
+	return nv50_disp_new_(&ga102_disp, device, type, inst, pdisp);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c
index e675d9b9d5d7..a6bafe7fea1f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c
@@ -266,7 +266,8 @@ gf119_disp = {
 };
 
 int
-gf119_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gf119_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gf119_disp, device, index, pdisp);
+	return nv50_disp_new_(&gf119_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c
index 4c3439b1a62d..3b79cf233ac5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c
@@ -41,7 +41,8 @@ gk104_disp = {
 };
 
 int
-gk104_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gk104_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gk104_disp, device, index, pdisp);
+	return nv50_disp_new_(&gk104_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c
index bc6f4750c942..988eb12237a6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c
@@ -41,7 +41,8 @@ gk110_disp = {
 };
 
 int
-gk110_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gk110_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gk110_disp, device, index, pdisp);
+	return nv50_disp_new_(&gk110_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c
index 031cf6b03a76..5d8108feeacd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c
@@ -41,7 +41,8 @@ gm107_disp = {
 };
 
 int
-gm107_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gm107_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gm107_disp, device, index, pdisp);
+	return nv50_disp_new_(&gm107_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c
index ec9c33a5162d..f7bb66087476 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c
@@ -41,7 +41,8 @@ gm200_disp = {
 };
 
 int
-gm200_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gm200_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gm200_disp, device, index, pdisp);
+	return nv50_disp_new_(&gm200_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
index 8471de3f3b61..af0ca812a394 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
@@ -40,7 +40,8 @@ gp100_disp = {
 };
 
 int
-gp100_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gp100_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gp100_disp, device, index, pdisp);
+	return nv50_disp_new_(&gp100_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c
index a3779c5046ea..065fea1bdfd1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c
@@ -67,7 +67,8 @@ gp102_disp = {
 };
 
 int
-gp102_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gp102_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gp102_disp, device, index, pdisp);
+	return nv50_disp_new_(&gp102_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt200.c
index f80183701f44..22bc269df64a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt200.c
@@ -41,7 +41,8 @@ gt200_disp = {
 };
 
 int
-gt200_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gt200_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gt200_disp, device, index, pdisp);
+	return nv50_disp_new_(&gt200_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt215.c
index 7581efc1357e..63a912b174d7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gt215.c
@@ -41,7 +41,8 @@ gt215_disp = {
 };
 
 int
-gt215_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gt215_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gt215_disp, device, index, pdisp);
+	return nv50_disp_new_(&gt215_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c
index c1032527f791..53879d5271cf 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c
@@ -441,7 +441,8 @@ gv100_disp = {
 };
 
 int
-gv100_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+gv100_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&gv100_disp, device, index, pdisp);
+	return nv50_disp_new_(&gv100_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h
index 09f3038eff26..9f0bb7c6b010 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h
@@ -150,6 +150,8 @@ void gv100_sor_dp_audio(struct nvkm_ior *, int, bool);
 void gv100_sor_dp_audio_sym(struct nvkm_ior *, int, u16, u32);
 void gv100_sor_dp_watermark(struct nvkm_ior *, int, u8);
 
+void tu102_sor_dp_vcpi(struct nvkm_ior *, int, u8, u8, u16, u16);
+
 void g84_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8);
 void gt215_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8);
 void gf119_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8);
@@ -207,4 +209,6 @@ int gv100_sor_cnt(struct nvkm_disp *, unsigned long *);
 int gv100_sor_new(struct nvkm_disp *, int);
 
 int tu102_sor_new(struct nvkm_disp *, int);
+
+int ga102_sor_new(struct nvkm_disp *, int);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp77.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp77.c
index cfdce23ab83a..762a59f24bbb 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp77.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp77.c
@@ -39,7 +39,8 @@ mcp77_disp = {
 };
 
 int
-mcp77_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+mcp77_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&mcp77_disp, device, index, pdisp);
+	return nv50_disp_new_(&mcp77_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp89.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp89.c
index 85d9329cfa0e..e5c58aae15de 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp89.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/mcp89.c
@@ -39,7 +39,8 @@ mcp89_disp = {
 };
 
 int
-mcp89_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+mcp89_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&mcp89_disp, device, index, pdisp);
+	return nv50_disp_new_(&mcp89_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv04.c
index b780ba1a3bc7..a12097db2c2a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv04.c
@@ -64,11 +64,12 @@ nv04_disp = {
 };
 
 int
-nv04_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+nv04_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_disp **pdisp)
 {
 	int ret, i;
 
-	ret = nvkm_disp_new_(&nv04_disp, device, index, pdisp);
+	ret = nvkm_disp_new_(&nv04_disp, device, type, inst, pdisp);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c
index e21556bf2cb1..3f20e49070ce 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c
@@ -154,7 +154,7 @@ nv50_disp_ = {
 
 int
 nv50_disp_new_(const struct nv50_disp_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_disp **pdisp)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_disp **pdisp)
 {
 	struct nv50_disp *disp;
 	int ret;
@@ -164,7 +164,7 @@ nv50_disp_new_(const struct nv50_disp_func *func, struct nvkm_device *device,
 	disp->func = func;
 	*pdisp = &disp->base;
 
-	ret = nvkm_disp_ctor(&nv50_disp_, device, index, &disp->base);
+	ret = nvkm_disp_ctor(&nv50_disp_, device, type, inst, &disp->base);
 	if (ret)
 		return ret;
 
@@ -769,7 +769,8 @@ nv50_disp = {
 };
 
 int
-nv50_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+nv50_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&nv50_disp, device, index, pdisp);
+	return nv50_disp_new_(&nv50_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h
index a677161c7f3a..025cacd7c3b0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h
@@ -47,8 +47,8 @@ void nv50_disp_super_2_1(struct nv50_disp *, struct nvkm_head *);
 void nv50_disp_super_2_2(struct nv50_disp *, struct nvkm_head *);
 void nv50_disp_super_3_0(struct nv50_disp *, struct nvkm_head *);
 
-int nv50_disp_new_(const struct nv50_disp_func *, struct nvkm_device *,
-		   int index, struct nvkm_disp **);
+int nv50_disp_new_(const struct nv50_disp_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_disp **);
 
 struct nv50_disp_func {
 	int (*init)(struct nv50_disp *);
@@ -86,6 +86,8 @@ void gv100_disp_intr(struct nv50_disp *);
 void gv100_disp_super(struct work_struct *);
 int gv100_disp_wndw_cnt(struct nvkm_disp *, unsigned long *);
 
+int tu102_disp_init(struct nv50_disp *);
+
 void nv50_disp_dptmds_war_2(struct nv50_disp *, struct dcb_output *);
 void nv50_disp_dptmds_war_3(struct nv50_disp *, struct dcb_output *);
 void nv50_disp_update_sppll1(struct nv50_disp *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/priv.h
index f815a5342880..ec57d8b6bce9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/priv.h
@@ -4,10 +4,10 @@
 #include <engine/disp.h>
 #include "outp.h"
 
-int nvkm_disp_ctor(const struct nvkm_disp_func *, struct nvkm_device *,
-		   int index, struct nvkm_disp *);
-int nvkm_disp_new_(const struct nvkm_disp_func *, struct nvkm_device *,
-		   int index, struct nvkm_disp **);
+int nvkm_disp_ctor(const struct nvkm_disp_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_disp *);
+int nvkm_disp_new_(const struct nvkm_disp_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_disp **);
 void nvkm_disp_vblank(struct nvkm_disp *, int head);
 
 struct nvkm_disp_func {
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootga102.c
new file mode 100644
index 000000000000..9af07c3cf9fc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootga102.c
@@ -0,0 +1,52 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "rootnv50.h"
+#include "channv50.h"
+
+#include <nvif/class.h>
+
+static const struct nv50_disp_root_func
+ga102_disp_root = {
+	.user = {
+		{{-1,-1,GV100_DISP_CAPS                }, gv100_disp_caps_new },
+		{{0,0,GA102_DISP_CURSOR                }, gv100_disp_curs_new },
+		{{0,0,GA102_DISP_WINDOW_IMM_CHANNEL_DMA}, gv100_disp_wimm_new },
+		{{0,0,GA102_DISP_CORE_CHANNEL_DMA      }, gv100_disp_core_new },
+		{{0,0,GA102_DISP_WINDOW_CHANNEL_DMA    }, gv100_disp_wndw_new },
+		{}
+	},
+};
+
+static int
+ga102_disp_root_new(struct nvkm_disp *disp, const struct nvkm_oclass *oclass,
+		    void *data, u32 size, struct nvkm_object **pobject)
+{
+	return nv50_disp_root_new_(&ga102_disp_root, disp, oclass, data, size, pobject);
+}
+
+const struct nvkm_disp_oclass
+ga102_disp_root_oclass = {
+	.base.oclass = GA102_DISP,
+	.base.minver = -1,
+	.base.maxver = -1,
+	.ctor = ga102_disp_root_new,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h
index 7070f5408d92..27bb170d0293 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h
@@ -41,4 +41,5 @@ extern const struct nvkm_disp_oclass gp100_disp_root_oclass;
 extern const struct nvkm_disp_oclass gp102_disp_root_oclass;
 extern const struct nvkm_disp_oclass gv100_disp_root_oclass;
 extern const struct nvkm_disp_oclass tu102_disp_root_oclass;
+extern const struct nvkm_disp_oclass ga102_disp_root_oclass;
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorga102.c
new file mode 100644
index 000000000000..033827de9116
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorga102.c
@@ -0,0 +1,140 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "ior.h"
+
+#include <subdev/timer.h>
+
+static int
+ga102_sor_dp_links(struct nvkm_ior *sor, struct nvkm_i2c_aux *aux)
+{
+	struct nvkm_device *device = sor->disp->engine.subdev.device;
+	const u32 soff = nv50_ior_base(sor);
+	const u32 loff = nv50_sor_link(sor);
+	u32 dpctrl = 0x00000000;
+	u32 clksor = 0x00000000;
+
+	switch (sor->dp.bw) {
+	case 0x06: clksor |= 0x00000000; break;
+	case 0x0a: clksor |= 0x00040000; break;
+	case 0x14: clksor |= 0x00080000; break;
+	case 0x1e: clksor |= 0x000c0000; break;
+	default:
+		WARN_ON(1);
+		return -EINVAL;
+	}
+
+	dpctrl |= ((1 << sor->dp.nr) - 1) << 16;
+	if (sor->dp.mst)
+		dpctrl |= 0x40000000;
+	if (sor->dp.ef)
+		dpctrl |= 0x00004000;
+
+	nvkm_mask(device, 0x612300 + soff, 0x007c0000, clksor);
+
+	/*XXX*/
+	nvkm_msec(device, 40, NVKM_DELAY);
+	nvkm_mask(device, 0x612300 + soff, 0x00030000, 0x00010000);
+	nvkm_mask(device, 0x61c10c + loff, 0x00000003, 0x00000001);
+
+	nvkm_mask(device, 0x61c10c + loff, 0x401f4000, dpctrl);
+	return 0;
+}
+
+static void
+ga102_sor_clock(struct nvkm_ior *sor)
+{
+	struct nvkm_device *device = sor->disp->engine.subdev.device;
+	u32 div2 = 0;
+	if (sor->asy.proto == TMDS) {
+		if (sor->tmds.high_speed)
+			div2 = 1;
+	}
+	nvkm_wr32(device, 0x00ec08 + (sor->id * 0x10), 0x00000000);
+	nvkm_wr32(device, 0x00ec04 + (sor->id * 0x10), div2);
+}
+
+static const struct nvkm_ior_func
+ga102_sor_hda = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gv100_sor_state,
+	.power = nv50_sor_power,
+	.clock = ga102_sor_clock,
+	.hdmi = {
+		.ctrl = gv100_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = ga102_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.vcpi = tu102_sor_dp_vcpi,
+		.audio = gv100_sor_dp_audio,
+		.audio_sym = gv100_sor_dp_audio_sym,
+		.watermark = gv100_sor_dp_watermark,
+	},
+	.hda = {
+		.hpd = gf119_hda_hpd,
+		.eld = gf119_hda_eld,
+		.device_entry = gv100_hda_device_entry,
+	},
+};
+
+static const struct nvkm_ior_func
+ga102_sor = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gv100_sor_state,
+	.power = nv50_sor_power,
+	.clock = ga102_sor_clock,
+	.hdmi = {
+		.ctrl = gv100_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = ga102_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.vcpi = tu102_sor_dp_vcpi,
+		.audio = gv100_sor_dp_audio,
+		.audio_sym = gv100_sor_dp_audio_sym,
+		.watermark = gv100_sor_dp_watermark,
+	},
+};
+
+int
+ga102_sor_new(struct nvkm_disp *disp, int id)
+{
+	struct nvkm_device *device = disp->engine.subdev.device;
+	u32 hda = nvkm_rd32(device, 0x08a15c);
+	if (hda & BIT(id))
+		return nvkm_ior_new_(&ga102_sor_hda, disp, SOR, id);
+	return nvkm_ior_new_(&ga102_sor, disp, SOR, id);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c
index 59865a934c4b..0cf9e8752d25 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c
@@ -23,7 +23,7 @@
 
 #include <subdev/timer.h>
 
-static void
+void
 tu102_sor_dp_vcpi(struct nvkm_ior *sor, int head,
 		  u8 slot, u8 slot_nr, u16 pbn, u16 aligned)
 {
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu102.c
index 883ae4151ff8..f5f8dc8e8f35 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu102.c
@@ -28,7 +28,7 @@
 #include <core/gpuobj.h>
 #include <subdev/timer.h>
 
-static int
+int
 tu102_disp_init(struct nv50_disp *disp)
 {
 	struct nvkm_device *device = disp->base.engine.subdev.device;
@@ -146,7 +146,8 @@ tu102_disp = {
 };
 
 int
-tu102_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+tu102_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_disp **pdisp)
 {
-	return nv50_disp_new_(&tu102_disp, device, index, pdisp);
+	return nv50_disp_new_(&tu102_disp, device, type, inst, pdisp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/dma/base.c
index 11b7b8fd5dda..425cde35f128 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/base.c
@@ -104,7 +104,7 @@ nvkm_dma = {
 
 int
 nvkm_dma_new_(const struct nvkm_dma_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_dma **pdma)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_dma **pdma)
 {
 	struct nvkm_dma *dma;
 
@@ -112,5 +112,5 @@ nvkm_dma_new_(const struct nvkm_dma_func *func, struct nvkm_device *device,
 		return -ENOMEM;
 	dma->func = func;
 
-	return nvkm_engine_ctor(&nvkm_dma, device, index, true, &dma->engine);
+	return nvkm_engine_ctor(&nvkm_dma, device, type, inst, true, &dma->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf100.c
index efec5d322179..99a1e07fa204 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf100.c
@@ -30,7 +30,8 @@ gf100_dma = {
 };
 
 int
-gf100_dma_new(struct nvkm_device *device, int index, struct nvkm_dma **pdma)
+gf100_dma_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_dma **pdma)
 {
-	return nvkm_dma_new_(&gf100_dma, device, index, pdma);
+	return nvkm_dma_new_(&gf100_dma, device, type, inst, pdma);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf119.c
index 34c766039aed..fd1d1fc22dc6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/gf119.c
@@ -30,7 +30,8 @@ gf119_dma = {
 };
 
 int
-gf119_dma_new(struct nvkm_device *device, int index, struct nvkm_dma **pdma)
+gf119_dma_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_dma **pdma)
 {
-	return nvkm_dma_new_(&gf119_dma, device, index, pdma);
+	return nvkm_dma_new_(&gf119_dma, device, type, inst, pdma);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/dma/gv100.c
index c65a4c2ea93d..a5af0df30663 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/gv100.c
@@ -28,7 +28,8 @@ gv100_dma = {
 };
 
 int
-gv100_dma_new(struct nvkm_device *device, int index, struct nvkm_dma **pdma)
+gv100_dma_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_dma **pdma)
 {
-	return nvkm_dma_new_(&gv100_dma, device, index, pdma);
+	return nvkm_dma_new_(&gv100_dma, device, type, inst, pdma);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv04.c
index 30747a0ce488..ea5a889f60c2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv04.c
@@ -30,7 +30,8 @@ nv04_dma = {
 };
 
 int
-nv04_dma_new(struct nvkm_device *device, int index, struct nvkm_dma **pdma)
+nv04_dma_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_dma **pdma)
 {
-	return nvkm_dma_new_(&nv04_dma, device, index, pdma);
+	return nvkm_dma_new_(&nv04_dma, device, type, inst, pdma);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv50.c
index 77aca7b71c83..6e8f79660014 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/nv50.c
@@ -30,7 +30,8 @@ nv50_dma = {
 };
 
 int
-nv50_dma_new(struct nvkm_device *device, int index, struct nvkm_dma **pdma)
+nv50_dma_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_dma **pdma)
 {
-	return nvkm_dma_new_(&nv50_dma, device, index, pdma);
+	return nvkm_dma_new_(&nv50_dma, device, type, inst, pdma);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/dma/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/dma/priv.h
index 0c9d9640a59d..d403bedb485a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/dma/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/dma/priv.h
@@ -9,8 +9,8 @@ struct nvkm_dmaobj_func {
 		    struct nvkm_gpuobj **);
 };
 
-int nvkm_dma_new_(const struct nvkm_dma_func *, struct nvkm_device *,
-		  int index, struct nvkm_dma **);
+int nvkm_dma_new_(const struct nvkm_dma_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_dma **);
 
 struct nvkm_dma_func {
 	int (*class_new)(struct nvkm_dma *, const struct nvkm_oclass *,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c b/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c
index 8675613e142b..43b7dec45179 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c
@@ -108,7 +108,7 @@ nvkm_falcon_fini(struct nvkm_engine *engine, bool suspend)
 		}
 	}
 
-	if (nvkm_mc_enabled(device, engine->subdev.index)) {
+	if (nvkm_mc_enabled(device, engine->subdev.type, engine->subdev.inst)) {
 		nvkm_mask(device, base + 0x048, 0x00000003, 0x00000000);
 		nvkm_wr32(device, base + 0x014, 0xffffffff);
 	}
@@ -335,9 +335,9 @@ nvkm_falcon = {
 };
 
 int
-nvkm_falcon_new_(const struct nvkm_falcon_func *func,
-		 struct nvkm_device *device, int index, bool enable,
-		 u32 addr, struct nvkm_engine **pengine)
+nvkm_falcon_new_(const struct nvkm_falcon_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, bool enable, u32 addr,
+		 struct nvkm_engine **pengine)
 {
 	struct nvkm_falcon *falcon;
 
@@ -351,6 +351,5 @@ nvkm_falcon_new_(const struct nvkm_falcon_func *func,
 	falcon->data.size = func->data.size;
 	*pengine = &falcon->engine;
 
-	return nvkm_engine_ctor(&nvkm_falcon, device, index,
-				enable, &falcon->engine);
+	return nvkm_engine_ctor(&nvkm_falcon, device, type, inst, enable, &falcon->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c
index c773caf21f6b..2ed4ff05d207 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c
@@ -292,7 +292,7 @@ nvkm_fifo_info(struct nvkm_engine *engine, u64 mthd, u64 *data)
 {
 	struct nvkm_fifo *fifo = nvkm_fifo(engine);
 	switch (mthd) {
-	case NV_DEVICE_FIFO_CHANNELS: *data = fifo->nr; return 0;
+	case NV_DEVICE_HOST_CHANNELS: *data = fifo->nr; return 0;
 	default:
 		if (fifo->func->info)
 			return fifo->func->info(fifo, mthd, data);
@@ -313,7 +313,7 @@ nvkm_fifo_oneinit(struct nvkm_engine *engine)
 static void
 nvkm_fifo_preinit(struct nvkm_engine *engine)
 {
-	nvkm_mc_reset(engine->subdev.device, NVKM_ENGINE_FIFO);
+	nvkm_mc_reset(engine->subdev.device, NVKM_ENGINE_FIFO, 0);
 }
 
 static int
@@ -334,6 +334,7 @@ nvkm_fifo_dtor(struct nvkm_engine *engine)
 	nvkm_event_fini(&fifo->kevent);
 	nvkm_event_fini(&fifo->cevent);
 	nvkm_event_fini(&fifo->uevent);
+	mutex_destroy(&fifo->mutex);
 	return data;
 }
 
@@ -351,13 +352,14 @@ nvkm_fifo = {
 
 int
 nvkm_fifo_ctor(const struct nvkm_fifo_func *func, struct nvkm_device *device,
-	       int index, int nr, struct nvkm_fifo *fifo)
+	       enum nvkm_subdev_type type, int inst, int nr, struct nvkm_fifo *fifo)
 {
 	int ret;
 
 	fifo->func = func;
 	INIT_LIST_HEAD(&fifo->chan);
 	spin_lock_init(&fifo->lock);
+	mutex_init(&fifo->mutex);
 
 	if (WARN_ON(fifo->nr > NVKM_FIFO_CHID_NR))
 		fifo->nr = NVKM_FIFO_CHID_NR;
@@ -365,7 +367,7 @@ nvkm_fifo_ctor(const struct nvkm_fifo_func *func, struct nvkm_device *device,
 		fifo->nr = nr;
 	bitmap_clear(fifo->mask, 0, fifo->nr);
 
-	ret = nvkm_engine_ctor(&nvkm_fifo, device, index, true, &fifo->engine);
+	ret = nvkm_engine_ctor(&nvkm_fifo, device, type, inst, true, &fifo->engine);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c
index d83485385934..8d957643940a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c
@@ -35,6 +35,15 @@ struct nvkm_fifo_chan_object {
 	int hash;
 };
 
+static struct nvkm_fifo_engn *
+nvkm_fifo_chan_engn(struct nvkm_fifo_chan *chan, struct nvkm_engine *engine)
+{
+	int engi = chan->fifo->func->engine_id(chan->fifo, engine);
+	if (engi >= 0)
+		return &chan->engn[engi];
+	return NULL;
+}
+
 static int
 nvkm_fifo_chan_child_fini(struct nvkm_oproxy *base, bool suspend)
 {
@@ -42,8 +51,8 @@ nvkm_fifo_chan_child_fini(struct nvkm_oproxy *base, bool suspend)
 		container_of(base, typeof(*object), oproxy);
 	struct nvkm_engine *engine  = object->oproxy.object->engine;
 	struct nvkm_fifo_chan *chan = object->chan;
-	struct nvkm_fifo_engn *engn = &chan->engn[engine->subdev.index];
-	const char *name = nvkm_subdev_name[engine->subdev.index];
+	struct nvkm_fifo_engn *engn = nvkm_fifo_chan_engn(chan, engine);
+	const char *name = engine->subdev.name;
 	int ret = 0;
 
 	if (--engn->usecount)
@@ -75,8 +84,8 @@ nvkm_fifo_chan_child_init(struct nvkm_oproxy *base)
 		container_of(base, typeof(*object), oproxy);
 	struct nvkm_engine *engine  = object->oproxy.object->engine;
 	struct nvkm_fifo_chan *chan = object->chan;
-	struct nvkm_fifo_engn *engn = &chan->engn[engine->subdev.index];
-	const char *name = nvkm_subdev_name[engine->subdev.index];
+	struct nvkm_fifo_engn *engn = nvkm_fifo_chan_engn(chan, engine);
+	const char *name = engine->subdev.name;
 	int ret;
 
 	if (engn->usecount++)
@@ -108,7 +117,7 @@ nvkm_fifo_chan_child_del(struct nvkm_oproxy *base)
 		container_of(base, typeof(*object), oproxy);
 	struct nvkm_engine *engine  = object->oproxy.base.engine;
 	struct nvkm_fifo_chan *chan = object->chan;
-	struct nvkm_fifo_engn *engn = &chan->engn[engine->subdev.index];
+	struct nvkm_fifo_engn *engn = nvkm_fifo_chan_engn(chan, engine);
 
 	if (chan->func->object_dtor)
 		chan->func->object_dtor(chan, object->hash);
@@ -118,7 +127,7 @@ nvkm_fifo_chan_child_del(struct nvkm_oproxy *base)
 			chan->func->engine_dtor(chan, engine);
 		nvkm_object_del(&engn->object);
 		if (chan->vmm)
-			atomic_dec(&chan->vmm->engref[engine->subdev.index]);
+			atomic_dec(&chan->vmm->engref[engine->subdev.type]);
 	}
 }
 
@@ -135,7 +144,7 @@ nvkm_fifo_chan_child_new(const struct nvkm_oclass *oclass, void *data, u32 size,
 {
 	struct nvkm_engine *engine = oclass->engine;
 	struct nvkm_fifo_chan *chan = nvkm_fifo_chan(oclass->parent);
-	struct nvkm_fifo_engn *engn = &chan->engn[engine->subdev.index];
+	struct nvkm_fifo_engn *engn = nvkm_fifo_chan_engn(chan, engine);
 	struct nvkm_fifo_chan_object *object;
 	int ret = 0;
 
@@ -152,7 +161,7 @@ nvkm_fifo_chan_child_new(const struct nvkm_oclass *oclass, void *data, u32 size,
 		};
 
 		if (chan->vmm)
-			atomic_inc(&chan->vmm->engref[engine->subdev.index]);
+			atomic_inc(&chan->vmm->engref[engine->subdev.type]);
 
 		if (engine->func->fifo.cclass) {
 			ret = engine->func->fifo.cclass(chan, &cclass,
@@ -203,13 +212,12 @@ nvkm_fifo_chan_child_get(struct nvkm_object *object, int index,
 {
 	struct nvkm_fifo_chan *chan = nvkm_fifo_chan(object);
 	struct nvkm_fifo *fifo = chan->fifo;
-	struct nvkm_device *device = fifo->engine.subdev.device;
 	struct nvkm_engine *engine;
-	u64 mask = chan->engines;
-	int ret, i, c;
+	u32 engm = chan->engm;
+	int engi, ret, c;
 
-	for (; c = 0, i = __ffs64(mask), mask; mask &= ~(1ULL << i)) {
-		if (!(engine = nvkm_device_engine(device, i)))
+	for (; c = 0, engi = __ffs(engm), engm; engm &= ~(1ULL << engi)) {
+		if (!(engine = fifo->func->id_engine(fifo, engi)))
 			continue;
 		oclass->engine = engine;
 		oclass->base.oclass = 0;
@@ -352,7 +360,7 @@ nvkm_fifo_chan_func = {
 int
 nvkm_fifo_chan_ctor(const struct nvkm_fifo_chan_func *func,
 		    struct nvkm_fifo *fifo, u32 size, u32 align, bool zero,
-		    u64 hvmm, u64 push, u64 engines, int bar, u32 base,
+		    u64 hvmm, u64 push, u32 engm, int bar, u32 base,
 		    u32 user, const struct nvkm_oclass *oclass,
 		    struct nvkm_fifo_chan *chan)
 {
@@ -365,7 +373,7 @@ nvkm_fifo_chan_ctor(const struct nvkm_fifo_chan_func *func,
 	nvkm_object_ctor(&nvkm_fifo_chan_func, oclass, &chan->object);
 	chan->func = func;
 	chan->fifo = fifo;
-	chan->engines = engines;
+	chan->engm = engm;
 	INIT_LIST_HEAD(&chan->head);
 
 	/* instance memory */
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h
index 177e10562600..e53504354841 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h
@@ -22,7 +22,7 @@ struct nvkm_fifo_chan_func {
 
 int nvkm_fifo_chan_ctor(const struct nvkm_fifo_chan_func *, struct nvkm_fifo *,
 			u32 size, u32 align, bool zero, u64 vm, u64 push,
-			u64 engines, int bar, u32 base, u32 user,
+			u32 engm, int bar, u32 base, u32 user,
 			const struct nvkm_oclass *, struct nvkm_fifo_chan *);
 
 struct nvkm_fifo_chan_oclass {
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chang84.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chang84.c
index a5c998fe4485..353b77d9b3dc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chang84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chang84.c
@@ -45,29 +45,9 @@ g84_fifo_chan_ntfy(struct nvkm_fifo_chan *chan, u32 type,
 }
 
 static int
-g84_fifo_chan_engine(struct nvkm_engine *engine)
-{
-	switch (engine->subdev.index) {
-	case NVKM_ENGINE_GR    : return 0;
-	case NVKM_ENGINE_MPEG  :
-	case NVKM_ENGINE_MSPPP : return 1;
-	case NVKM_ENGINE_CE0   : return 2;
-	case NVKM_ENGINE_VP    :
-	case NVKM_ENGINE_MSPDEC: return 3;
-	case NVKM_ENGINE_CIPHER:
-	case NVKM_ENGINE_SEC   : return 4;
-	case NVKM_ENGINE_BSP   :
-	case NVKM_ENGINE_MSVLD : return 5;
-	default:
-		WARN_ON(1);
-		return 0;
-	}
-}
-
-static int
 g84_fifo_chan_engine_addr(struct nvkm_engine *engine)
 {
-	switch (engine->subdev.index) {
+	switch (engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW    : return -1;
 	case NVKM_ENGINE_GR    : return 0x0020;
@@ -79,7 +59,7 @@ g84_fifo_chan_engine_addr(struct nvkm_engine *engine)
 	case NVKM_ENGINE_MSVLD : return 0x0080;
 	case NVKM_ENGINE_CIPHER:
 	case NVKM_ENGINE_SEC   : return 0x00a0;
-	case NVKM_ENGINE_CE0   : return 0x00c0;
+	case NVKM_ENGINE_CE    : return 0x00c0;
 	default:
 		WARN_ON(1);
 		return -1;
@@ -102,7 +82,7 @@ g84_fifo_chan_engine_fini(struct nvkm_fifo_chan *base,
 	if (offset < 0)
 		return 0;
 
-	engn = g84_fifo_chan_engine(engine);
+	engn = fifo->base.func->engine_id(&fifo->base, engine);
 	save = nvkm_mask(device, 0x002520, 0x0000003f, 1 << engn);
 	nvkm_wr32(device, 0x0032fc, chan->base.inst->addr >> 12);
 	done = nvkm_msec(device, 2000,
@@ -134,7 +114,7 @@ g84_fifo_chan_engine_init(struct nvkm_fifo_chan *base,
 			  struct nvkm_engine *engine)
 {
 	struct nv50_fifo_chan *chan = nv50_fifo_chan(base);
-	struct nvkm_gpuobj *engn = chan->engn[engine->subdev.index];
+	struct nvkm_gpuobj *engn = *nv50_fifo_chan_engine(chan, engine);
 	u64 limit, start;
 	int offset;
 
@@ -162,12 +142,11 @@ g84_fifo_chan_engine_ctor(struct nvkm_fifo_chan *base,
 			  struct nvkm_object *object)
 {
 	struct nv50_fifo_chan *chan = nv50_fifo_chan(base);
-	int engn = engine->subdev.index;
 
 	if (g84_fifo_chan_engine_addr(engine) < 0)
 		return 0;
 
-	return nvkm_object_bind(object, NULL, 0, &chan->engn[engn]);
+	return nvkm_object_bind(object, NULL, 0, nv50_fifo_chan_engine(chan, engine));
 }
 
 static int
@@ -178,14 +157,14 @@ g84_fifo_chan_object_ctor(struct nvkm_fifo_chan *base,
 	u32 handle = object->handle;
 	u32 context;
 
-	switch (object->engine->subdev.index) {
+	switch (object->engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW    : context = 0x00000000; break;
 	case NVKM_ENGINE_GR    : context = 0x00100000; break;
 	case NVKM_ENGINE_MPEG  :
 	case NVKM_ENGINE_MSPPP : context = 0x00200000; break;
 	case NVKM_ENGINE_ME    :
-	case NVKM_ENGINE_CE0   : context = 0x00300000; break;
+	case NVKM_ENGINE_CE    : context = 0x00300000; break;
 	case NVKM_ENGINE_VP    :
 	case NVKM_ENGINE_MSPDEC: context = 0x00400000; break;
 	case NVKM_ENGINE_CIPHER:
@@ -241,20 +220,20 @@ g84_fifo_chan_ctor(struct nv50_fifo *fifo, u64 vmm, u64 push,
 
 	ret = nvkm_fifo_chan_ctor(&g84_fifo_chan_func, &fifo->base,
 				  0x10000, 0x1000, false, vmm, push,
-				  (1ULL << NVKM_ENGINE_BSP) |
-				  (1ULL << NVKM_ENGINE_CE0) |
-				  (1ULL << NVKM_ENGINE_CIPHER) |
-				  (1ULL << NVKM_ENGINE_DMAOBJ) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_ME) |
-				  (1ULL << NVKM_ENGINE_MPEG) |
-				  (1ULL << NVKM_ENGINE_MSPDEC) |
-				  (1ULL << NVKM_ENGINE_MSPPP) |
-				  (1ULL << NVKM_ENGINE_MSVLD) |
-				  (1ULL << NVKM_ENGINE_SEC) |
-				  (1ULL << NVKM_ENGINE_SW) |
-				  (1ULL << NVKM_ENGINE_VIC) |
-				  (1ULL << NVKM_ENGINE_VP),
+				  BIT(G84_FIFO_ENGN_SW) |
+				  BIT(G84_FIFO_ENGN_GR) |
+				  BIT(G84_FIFO_ENGN_MPEG) |
+				  BIT(G84_FIFO_ENGN_MSPPP) |
+				  BIT(G84_FIFO_ENGN_ME) |
+				  BIT(G84_FIFO_ENGN_CE0) |
+				  BIT(G84_FIFO_ENGN_VP) |
+				  BIT(G84_FIFO_ENGN_MSPDEC) |
+				  BIT(G84_FIFO_ENGN_CIPHER) |
+				  BIT(G84_FIFO_ENGN_SEC) |
+				  BIT(G84_FIFO_ENGN_VIC) |
+				  BIT(G84_FIFO_ENGN_BSP) |
+				  BIT(G84_FIFO_ENGN_MSVLD) |
+				  BIT(G84_FIFO_ENGN_DMA),
 				  0, 0xc00000, 0x2000, oclass, &chan->base);
 	chan->fifo = fifo;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changf100.h
index 7c125a15f963..f7ac1061fa84 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changf100.h
@@ -12,10 +12,17 @@ struct gf100_fifo_chan {
 	struct list_head head;
 	bool killed;
 
-	struct {
+#define GF100_FIFO_ENGN_GR     0
+#define GF100_FIFO_ENGN_MSPDEC 1
+#define GF100_FIFO_ENGN_MSPPP  2
+#define GF100_FIFO_ENGN_MSVLD  3
+#define GF100_FIFO_ENGN_CE0    4
+#define GF100_FIFO_ENGN_CE1    5
+#define GF100_FIFO_ENGN_SW     15
+	struct gf100_fifo_engn {
 		struct nvkm_gpuobj *inst;
 		struct nvkm_vma *vma;
-	} engn[NVKM_SUBDEV_NR];
+	} engn[NVKM_FIFO_ENGN_NR];
 };
 
 extern const struct nvkm_fifo_chan_oclass gf100_fifo_gpfifo_oclass;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h
index 22698661aa85..cfbe096e604f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h
@@ -16,10 +16,11 @@ struct gk104_fifo_chan {
 
 	struct nvkm_memory *mthd;
 
-	struct {
+#define GK104_FIFO_ENGN_SW 15
+	struct gk104_fifo_engn {
 		struct nvkm_gpuobj *inst;
 		struct nvkm_vma *vma;
-	} engn[NVKM_SUBDEV_NR];
+	} engn[NVKM_FIFO_ENGN_NR];
 };
 
 extern const struct nvkm_fifo_chan_func gk104_fifo_gpfifo_func;
@@ -29,6 +30,7 @@ int gk104_fifo_gpfifo_new(struct gk104_fifo *, const struct nvkm_oclass *,
 void *gk104_fifo_gpfifo_dtor(struct nvkm_fifo_chan *);
 void gk104_fifo_gpfifo_init(struct nvkm_fifo_chan *);
 void gk104_fifo_gpfifo_fini(struct nvkm_fifo_chan *);
+struct gk104_fifo_engn *gk104_fifo_gpfifo_engine(struct gk104_fifo_chan *, struct nvkm_engine *);
 int gk104_fifo_gpfifo_engine_ctor(struct nvkm_fifo_chan *, struct nvkm_engine *,
 				  struct nvkm_object *);
 void gk104_fifo_gpfifo_engine_dtor(struct nvkm_fifo_chan *,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv04.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv04.h
index 60ca79465aff..727bc8976b40 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv04.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv04.h
@@ -9,7 +9,11 @@ struct nv04_fifo_chan {
 	struct nvkm_fifo_chan base;
 	struct nv04_fifo *fifo;
 	u32 ramfc;
-	struct nvkm_gpuobj *engn[NVKM_SUBDEV_NR];
+#define NV04_FIFO_ENGN_SW   0
+#define NV04_FIFO_ENGN_GR   1
+#define NV04_FIFO_ENGN_MPEG 2
+#define NV04_FIFO_ENGN_DMA  3
+	struct nvkm_gpuobj *engn[NVKM_FIFO_ENGN_NR];
 };
 
 extern const struct nvkm_fifo_chan_func nv04_fifo_dma_func;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.c
index 85f7dbf53c99..c44d7c81dd52 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.c
@@ -31,7 +31,7 @@
 static int
 nv50_fifo_chan_engine_addr(struct nvkm_engine *engine)
 {
-	switch (engine->subdev.index) {
+	switch (engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW    : return -1;
 	case NVKM_ENGINE_GR    : return 0x0000;
@@ -42,6 +42,15 @@ nv50_fifo_chan_engine_addr(struct nvkm_engine *engine)
 	}
 }
 
+struct nvkm_gpuobj **
+nv50_fifo_chan_engine(struct nv50_fifo_chan *chan, struct nvkm_engine *engine)
+{
+	int engi = chan->base.fifo->func->engine_id(chan->base.fifo, engine);
+	if (engi >= 0)
+		return &chan->engn[engi];
+	return NULL;
+}
+
 static int
 nv50_fifo_chan_engine_fini(struct nvkm_fifo_chan *base,
 			   struct nvkm_engine *engine, bool suspend)
@@ -103,7 +112,7 @@ nv50_fifo_chan_engine_init(struct nvkm_fifo_chan *base,
 			   struct nvkm_engine *engine)
 {
 	struct nv50_fifo_chan *chan = nv50_fifo_chan(base);
-	struct nvkm_gpuobj *engn = chan->engn[engine->subdev.index];
+	struct nvkm_gpuobj *engn = *nv50_fifo_chan_engine(chan, engine);
 	u64 limit, start;
 	int offset;
 
@@ -130,7 +139,7 @@ nv50_fifo_chan_engine_dtor(struct nvkm_fifo_chan *base,
 			   struct nvkm_engine *engine)
 {
 	struct nv50_fifo_chan *chan = nv50_fifo_chan(base);
-	nvkm_gpuobj_del(&chan->engn[engine->subdev.index]);
+	nvkm_gpuobj_del(nv50_fifo_chan_engine(chan, engine));
 }
 
 static int
@@ -139,12 +148,11 @@ nv50_fifo_chan_engine_ctor(struct nvkm_fifo_chan *base,
 			   struct nvkm_object *object)
 {
 	struct nv50_fifo_chan *chan = nv50_fifo_chan(base);
-	int engn = engine->subdev.index;
 
 	if (nv50_fifo_chan_engine_addr(engine) < 0)
 		return 0;
 
-	return nvkm_object_bind(object, NULL, 0, &chan->engn[engn]);
+	return nvkm_object_bind(object, NULL, 0, nv50_fifo_chan_engine(chan, engine));
 }
 
 void
@@ -162,7 +170,7 @@ nv50_fifo_chan_object_ctor(struct nvkm_fifo_chan *base,
 	u32 handle = object->handle;
 	u32 context;
 
-	switch (object->engine->subdev.index) {
+	switch (object->engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW    : context = 0x00000000; break;
 	case NVKM_ENGINE_GR    : context = 0x00100000; break;
@@ -240,10 +248,10 @@ nv50_fifo_chan_ctor(struct nv50_fifo *fifo, u64 vmm, u64 push,
 
 	ret = nvkm_fifo_chan_ctor(&nv50_fifo_chan_func, &fifo->base,
 				  0x10000, 0x1000, false, vmm, push,
-				  (1ULL << NVKM_ENGINE_DMAOBJ) |
-				  (1ULL << NVKM_ENGINE_SW) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_MPEG),
+				  BIT(NV50_FIFO_ENGN_SW) |
+				  BIT(NV50_FIFO_ENGN_GR) |
+				  BIT(NV50_FIFO_ENGN_MPEG) |
+				  BIT(NV50_FIFO_ENGN_DMA),
 				  0, 0xc00000, 0x2000, oclass, &chan->base);
 	chan->fifo = fifo;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.h
index 5735ff72a9d1..af8bdf275552 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/channv50.h
@@ -15,13 +15,33 @@ struct nv50_fifo_chan {
 	struct nvkm_gpuobj *pgd;
 	struct nvkm_ramht *ramht;
 
-	struct nvkm_gpuobj *engn[NVKM_SUBDEV_NR];
+#define NV50_FIFO_ENGN_SW   0
+#define NV50_FIFO_ENGN_GR   1
+#define NV50_FIFO_ENGN_MPEG 2
+#define NV50_FIFO_ENGN_DMA  3
+
+#define G84_FIFO_ENGN_SW     0
+#define G84_FIFO_ENGN_GR     1
+#define G84_FIFO_ENGN_MPEG   2
+#define G84_FIFO_ENGN_MSPPP  2
+#define G84_FIFO_ENGN_ME     3
+#define G84_FIFO_ENGN_CE0    3
+#define G84_FIFO_ENGN_VP     4
+#define G84_FIFO_ENGN_MSPDEC 4
+#define G84_FIFO_ENGN_CIPHER 5
+#define G84_FIFO_ENGN_SEC    5
+#define G84_FIFO_ENGN_VIC    5
+#define G84_FIFO_ENGN_BSP    6
+#define G84_FIFO_ENGN_MSVLD  6
+#define G84_FIFO_ENGN_DMA    7
+	struct nvkm_gpuobj *engn[NVKM_FIFO_ENGN_NR];
 };
 
 int nv50_fifo_chan_ctor(struct nv50_fifo *, u64 vmm, u64 push,
 			const struct nvkm_oclass *, struct nv50_fifo_chan *);
 void *nv50_fifo_chan_dtor(struct nvkm_fifo_chan *);
 void nv50_fifo_chan_fini(struct nvkm_fifo_chan *);
+struct nvkm_gpuobj **nv50_fifo_chan_engine(struct nv50_fifo_chan *, struct nvkm_engine *);
 void nv50_fifo_chan_engine_dtor(struct nvkm_fifo_chan *, struct nvkm_engine *);
 void nv50_fifo_chan_object_dtor(struct nvkm_fifo_chan *, int);
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv04.c
index c213122cf088..dbcdc5fab990 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv04.c
@@ -38,9 +38,9 @@ nv04_fifo_dma_object_dtor(struct nvkm_fifo_chan *base, int cookie)
 	struct nv04_fifo_chan *chan = nv04_fifo_chan(base);
 	struct nvkm_instmem *imem = chan->fifo->base.engine.subdev.device->imem;
 
-	mutex_lock(&chan->fifo->base.engine.subdev.mutex);
+	mutex_lock(&chan->fifo->base.mutex);
 	nvkm_ramht_remove(imem->ramht, cookie);
-	mutex_unlock(&chan->fifo->base.engine.subdev.mutex);
+	mutex_unlock(&chan->fifo->base.mutex);
 }
 
 static int
@@ -53,7 +53,7 @@ nv04_fifo_dma_object_ctor(struct nvkm_fifo_chan *base,
 	u32 handle  = object->handle;
 	int hash;
 
-	switch (object->engine->subdev.index) {
+	switch (object->engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW    : context |= 0x00000000; break;
 	case NVKM_ENGINE_GR    : context |= 0x00010000; break;
@@ -63,10 +63,10 @@ nv04_fifo_dma_object_ctor(struct nvkm_fifo_chan *base,
 		return -EINVAL;
 	}
 
-	mutex_lock(&chan->fifo->base.engine.subdev.mutex);
+	mutex_lock(&chan->fifo->base.mutex);
 	hash = nvkm_ramht_insert(imem->ramht, object, chan->base.chid, 4,
 				 handle, context);
-	mutex_unlock(&chan->fifo->base.engine.subdev.mutex);
+	mutex_unlock(&chan->fifo->base.mutex);
 	return hash;
 }
 
@@ -191,9 +191,9 @@ nv04_fifo_dma_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 
 	ret = nvkm_fifo_chan_ctor(&nv04_fifo_dma_func, &fifo->base,
 				  0x1000, 0x1000, false, 0, args->v0.pushbuf,
-				  (1ULL << NVKM_ENGINE_DMAOBJ) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_SW),
+				  BIT(NV04_FIFO_ENGN_SW) |
+				  BIT(NV04_FIFO_ENGN_GR) |
+				  BIT(NV04_FIFO_ENGN_DMA),
 				  0, 0x800000, 0x10000, oclass, &chan->base);
 	chan->fifo = fifo;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv10.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv10.c
index f5f355ff005d..07d80d54a07c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv10.c
@@ -62,9 +62,9 @@ nv10_fifo_dma_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 
 	ret = nvkm_fifo_chan_ctor(&nv04_fifo_dma_func, &fifo->base,
 				  0x1000, 0x1000, false, 0, args->v0.pushbuf,
-				  (1ULL << NVKM_ENGINE_DMAOBJ) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_SW),
+				  BIT(NV04_FIFO_ENGN_SW) |
+				  BIT(NV04_FIFO_ENGN_GR) |
+				  BIT(NV04_FIFO_ENGN_DMA),
 				  0, 0x800000, 0x10000, oclass, &chan->base);
 	chan->fifo = fifo;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv17.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv17.c
index 7edc6a564b5d..edd70a114218 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv17.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv17.c
@@ -62,10 +62,10 @@ nv17_fifo_dma_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 
 	ret = nvkm_fifo_chan_ctor(&nv04_fifo_dma_func, &fifo->base,
 				  0x1000, 0x1000, false, 0, args->v0.pushbuf,
-				  (1ULL << NVKM_ENGINE_DMAOBJ) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_MPEG) | /* NV31- */
-				  (1ULL << NVKM_ENGINE_SW),
+				  BIT(NV04_FIFO_ENGN_SW) |
+				  BIT(NV04_FIFO_ENGN_GR) |
+				  BIT(NV04_FIFO_ENGN_MPEG) | /* NV31- */
+				  BIT(NV04_FIFO_ENGN_DMA),
 				  0, 0x800000, 0x10000, oclass, &chan->base);
 	chan->fifo = fifo;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv40.c
index 5f722c6e8a2f..0411fb908457 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/dmanv40.c
@@ -35,7 +35,7 @@
 static bool
 nv40_fifo_dma_engine(struct nvkm_engine *engine, u32 *reg, u32 *ctx)
 {
-	switch (engine->subdev.index) {
+	switch (engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW:
 		return false;
@@ -55,6 +55,15 @@ nv40_fifo_dma_engine(struct nvkm_engine *engine, u32 *reg, u32 *ctx)
 	}
 }
 
+static struct nvkm_gpuobj **
+nv40_fifo_dma_engn(struct nv04_fifo_chan *chan, struct nvkm_engine *engine)
+{
+	int engi = chan->base.fifo->func->engine_id(chan->base.fifo, engine);
+	if (engi >= 0)
+		return &chan->engn[engi];
+	return NULL;
+}
+
 static int
 nv40_fifo_dma_engine_fini(struct nvkm_fifo_chan *base,
 			  struct nvkm_engine *engine, bool suspend)
@@ -99,7 +108,7 @@ nv40_fifo_dma_engine_init(struct nvkm_fifo_chan *base,
 
 	if (!nv40_fifo_dma_engine(engine, &reg, &ctx))
 		return 0;
-	inst = chan->engn[engine->subdev.index]->addr >> 4;
+	inst = (*nv40_fifo_dma_engn(chan, engine))->addr >> 4;
 
 	spin_lock_irqsave(&fifo->base.lock, flags);
 	nvkm_mask(device, 0x002500, 0x00000001, 0x00000000);
@@ -121,7 +130,7 @@ nv40_fifo_dma_engine_dtor(struct nvkm_fifo_chan *base,
 			  struct nvkm_engine *engine)
 {
 	struct nv04_fifo_chan *chan = nv04_fifo_chan(base);
-	nvkm_gpuobj_del(&chan->engn[engine->subdev.index]);
+	nvkm_gpuobj_del(nv40_fifo_dma_engn(chan, engine));
 }
 
 static int
@@ -130,13 +139,12 @@ nv40_fifo_dma_engine_ctor(struct nvkm_fifo_chan *base,
 			  struct nvkm_object *object)
 {
 	struct nv04_fifo_chan *chan = nv04_fifo_chan(base);
-	const int engn = engine->subdev.index;
 	u32 reg, ctx;
 
 	if (!nv40_fifo_dma_engine(engine, &reg, &ctx))
 		return 0;
 
-	return nvkm_object_bind(object, NULL, 0, &chan->engn[engn]);
+	return nvkm_object_bind(object, NULL, 0, nv40_fifo_dma_engn(chan, engine));
 }
 
 static int
@@ -149,7 +157,7 @@ nv40_fifo_dma_object_ctor(struct nvkm_fifo_chan *base,
 	u32 handle  = object->handle;
 	int hash;
 
-	switch (object->engine->subdev.index) {
+	switch (object->engine->subdev.type) {
 	case NVKM_ENGINE_DMAOBJ:
 	case NVKM_ENGINE_SW    : context |= 0x00000000; break;
 	case NVKM_ENGINE_GR    : context |= 0x00100000; break;
@@ -159,10 +167,10 @@ nv40_fifo_dma_object_ctor(struct nvkm_fifo_chan *base,
 		return -EINVAL;
 	}
 
-	mutex_lock(&chan->fifo->base.engine.subdev.mutex);
+	mutex_lock(&chan->fifo->base.mutex);
 	hash = nvkm_ramht_insert(imem->ramht, object, chan->base.chid, 4,
 				 handle, context);
-	mutex_unlock(&chan->fifo->base.engine.subdev.mutex);
+	mutex_unlock(&chan->fifo->base.mutex);
 	return hash;
 }
 
@@ -209,10 +217,10 @@ nv40_fifo_dma_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 
 	ret = nvkm_fifo_chan_ctor(&nv40_fifo_dma_func, &fifo->base,
 				  0x1000, 0x1000, false, 0, args->v0.pushbuf,
-				  (1ULL << NVKM_ENGINE_DMAOBJ) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_MPEG) |
-				  (1ULL << NVKM_ENGINE_SW),
+				  BIT(NV04_FIFO_ENGN_SW) |
+				  BIT(NV04_FIFO_ENGN_GR) |
+				  BIT(NV04_FIFO_ENGN_MPEG) |
+				  BIT(NV04_FIFO_ENGN_DMA),
 				  0, 0xc00000, 0x1000, oclass, &chan->base);
 	chan->fifo = fifo;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/g84.c
index ff7b529764fe..c0a7d0f21dac 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/g84.c
@@ -38,12 +38,82 @@ g84_fifo_uevent_init(struct nvkm_fifo *fifo)
 	nvkm_mask(device, 0x002140, 0x40000000, 0x40000000);
 }
 
+static struct nvkm_engine *
+g84_fifo_id_engine(struct nvkm_fifo *fifo, int engi)
+{
+	struct nvkm_device *device = fifo->engine.subdev.device;
+	struct nvkm_engine *engine;
+	enum nvkm_subdev_type type;
+
+	switch (engi) {
+	case G84_FIFO_ENGN_SW    : type = NVKM_ENGINE_SW; break;
+	case G84_FIFO_ENGN_GR    : type = NVKM_ENGINE_GR; break;
+	case G84_FIFO_ENGN_MPEG  :
+		if ((engine = nvkm_device_engine(device, NVKM_ENGINE_MSPPP, 0)))
+			return engine;
+		type = NVKM_ENGINE_MPEG;
+		break;
+	case G84_FIFO_ENGN_ME    :
+		if ((engine = nvkm_device_engine(device, NVKM_ENGINE_CE, 0)))
+			return engine;
+		type = NVKM_ENGINE_ME;
+		break;
+	case G84_FIFO_ENGN_VP    :
+		if ((engine = nvkm_device_engine(device, NVKM_ENGINE_MSPDEC, 0)))
+			return engine;
+		type = NVKM_ENGINE_VP;
+		break;
+	case G84_FIFO_ENGN_CIPHER:
+		if ((engine = nvkm_device_engine(device, NVKM_ENGINE_VIC, 0)))
+			return engine;
+		if ((engine = nvkm_device_engine(device, NVKM_ENGINE_SEC, 0)))
+			return engine;
+		type = NVKM_ENGINE_CIPHER;
+		break;
+	case G84_FIFO_ENGN_BSP   :
+		if ((engine = nvkm_device_engine(device, NVKM_ENGINE_MSVLD, 0)))
+			return engine;
+		type = NVKM_ENGINE_BSP;
+		break;
+	case G84_FIFO_ENGN_DMA   : type = NVKM_ENGINE_DMAOBJ; break;
+	default:
+		WARN_ON(1);
+		return NULL;
+	}
+
+	return nvkm_device_engine(fifo->engine.subdev.device, type, 0);
+}
+
+static int
+g84_fifo_engine_id(struct nvkm_fifo *base, struct nvkm_engine *engine)
+{
+	switch (engine->subdev.type) {
+	case NVKM_ENGINE_SW    : return G84_FIFO_ENGN_SW;
+	case NVKM_ENGINE_GR    : return G84_FIFO_ENGN_GR;
+	case NVKM_ENGINE_MPEG  :
+	case NVKM_ENGINE_MSPPP : return G84_FIFO_ENGN_MPEG;
+	case NVKM_ENGINE_CE    : return G84_FIFO_ENGN_CE0;
+	case NVKM_ENGINE_VP    :
+	case NVKM_ENGINE_MSPDEC: return G84_FIFO_ENGN_VP;
+	case NVKM_ENGINE_CIPHER:
+	case NVKM_ENGINE_SEC   : return G84_FIFO_ENGN_CIPHER;
+	case NVKM_ENGINE_BSP   :
+	case NVKM_ENGINE_MSVLD : return G84_FIFO_ENGN_BSP;
+	case NVKM_ENGINE_DMAOBJ: return G84_FIFO_ENGN_DMA;
+	default:
+		WARN_ON(1);
+		return -1;
+	}
+}
+
 static const struct nvkm_fifo_func
 g84_fifo = {
 	.dtor = nv50_fifo_dtor,
 	.oneinit = nv50_fifo_oneinit,
 	.init = nv50_fifo_init,
 	.intr = nv04_fifo_intr,
+	.engine_id = g84_fifo_engine_id,
+	.id_engine = g84_fifo_id_engine,
 	.pause = nv04_fifo_pause,
 	.start = nv04_fifo_start,
 	.uevent_init = g84_fifo_uevent_init,
@@ -56,7 +126,8 @@ g84_fifo = {
 };
 
 int
-g84_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+g84_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_fifo **pfifo)
 {
-	return nv50_fifo_new_(&g84_fifo, device, index, pfifo);
+	return nv50_fifo_new_(&g84_fifo, device, type, inst, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c
index 5a39e51d42d7..8b4f36b3e34b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c
@@ -57,7 +57,7 @@ gf100_fifo_runlist_commit(struct gf100_fifo *fifo)
 	int nr = 0;
 	int target;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&fifo->base.mutex);
 	cur = fifo->runlist.mem[fifo->runlist.active];
 	fifo->runlist.active = !fifo->runlist.active;
 
@@ -73,7 +73,7 @@ gf100_fifo_runlist_commit(struct gf100_fifo *fifo)
 	case NVKM_MEM_TARGET_VRAM: target = 0; break;
 	case NVKM_MEM_TARGET_NCOH: target = 3; break;
 	default:
-		mutex_unlock(&subdev->mutex);
+		mutex_unlock(&fifo->base.mutex);
 		WARN_ON(1);
 		return;
 	}
@@ -86,59 +86,61 @@ gf100_fifo_runlist_commit(struct gf100_fifo *fifo)
 			       !(nvkm_rd32(device, 0x00227c) & 0x00100000),
 			       msecs_to_jiffies(2000)) == 0)
 		nvkm_error(subdev, "runlist update timeout\n");
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
 void
 gf100_fifo_runlist_remove(struct gf100_fifo *fifo, struct gf100_fifo_chan *chan)
 {
-	mutex_lock(&fifo->base.engine.subdev.mutex);
+	mutex_lock(&fifo->base.mutex);
 	list_del_init(&chan->head);
-	mutex_unlock(&fifo->base.engine.subdev.mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
 void
 gf100_fifo_runlist_insert(struct gf100_fifo *fifo, struct gf100_fifo_chan *chan)
 {
-	mutex_lock(&fifo->base.engine.subdev.mutex);
+	mutex_lock(&fifo->base.mutex);
 	list_add_tail(&chan->head, &fifo->chan);
-	mutex_unlock(&fifo->base.engine.subdev.mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
-static inline int
-gf100_fifo_engidx(struct gf100_fifo *fifo, u32 engn)
+static struct nvkm_engine *
+gf100_fifo_id_engine(struct nvkm_fifo *fifo, int engi)
 {
-	switch (engn) {
-	case NVKM_ENGINE_GR    : engn = 0; break;
-	case NVKM_ENGINE_MSVLD : engn = 1; break;
-	case NVKM_ENGINE_MSPPP : engn = 2; break;
-	case NVKM_ENGINE_MSPDEC: engn = 3; break;
-	case NVKM_ENGINE_CE0   : engn = 4; break;
-	case NVKM_ENGINE_CE1   : engn = 5; break;
+	enum nvkm_subdev_type type;
+	int inst;
+
+	switch (engi) {
+	case GF100_FIFO_ENGN_GR    : type = NVKM_ENGINE_GR    ; inst = 0; break;
+	case GF100_FIFO_ENGN_MSPDEC: type = NVKM_ENGINE_MSPDEC; inst = 0; break;
+	case GF100_FIFO_ENGN_MSPPP : type = NVKM_ENGINE_MSPPP ; inst = 0; break;
+	case GF100_FIFO_ENGN_MSVLD : type = NVKM_ENGINE_MSVLD ; inst = 0; break;
+	case GF100_FIFO_ENGN_CE0   : type = NVKM_ENGINE_CE    ; inst = 0; break;
+	case GF100_FIFO_ENGN_CE1   : type = NVKM_ENGINE_CE    ; inst = 1; break;
+	case GF100_FIFO_ENGN_SW    : type = NVKM_ENGINE_SW    ; inst = 0; break;
 	default:
-		return -1;
+		WARN_ON(1);
+		return NULL;
 	}
 
-	return engn;
+	return nvkm_device_engine(fifo->engine.subdev.device, type, inst);
 }
 
-static inline struct nvkm_engine *
-gf100_fifo_engine(struct gf100_fifo *fifo, u32 engn)
+static int
+gf100_fifo_engine_id(struct nvkm_fifo *base, struct nvkm_engine *engine)
 {
-	struct nvkm_device *device = fifo->base.engine.subdev.device;
-
-	switch (engn) {
-	case 0: engn = NVKM_ENGINE_GR; break;
-	case 1: engn = NVKM_ENGINE_MSVLD; break;
-	case 2: engn = NVKM_ENGINE_MSPPP; break;
-	case 3: engn = NVKM_ENGINE_MSPDEC; break;
-	case 4: engn = NVKM_ENGINE_CE0; break;
-	case 5: engn = NVKM_ENGINE_CE1; break;
+	switch (engine->subdev.type) {
+	case NVKM_ENGINE_GR    : return GF100_FIFO_ENGN_GR;
+	case NVKM_ENGINE_MSPDEC: return GF100_FIFO_ENGN_MSPDEC;
+	case NVKM_ENGINE_MSPPP : return GF100_FIFO_ENGN_MSPPP;
+	case NVKM_ENGINE_MSVLD : return GF100_FIFO_ENGN_MSVLD;
+	case NVKM_ENGINE_CE    : return GF100_FIFO_ENGN_CE0 + engine->subdev.inst;
+	case NVKM_ENGINE_SW    : return GF100_FIFO_ENGN_SW;
 	default:
-		return NULL;
+		WARN_ON(1);
+		return -1;
 	}
-
-	return nvkm_device_engine(device, engn);
 }
 
 static void
@@ -148,20 +150,17 @@ gf100_fifo_recover_work(struct work_struct *w)
 	struct nvkm_device *device = fifo->base.engine.subdev.device;
 	struct nvkm_engine *engine;
 	unsigned long flags;
-	u32 engn, engm = 0;
-	u64 mask, todo;
+	u32 engm, engn, todo;
 
 	spin_lock_irqsave(&fifo->base.lock, flags);
-	mask = fifo->recover.mask;
+	engm = fifo->recover.mask;
 	fifo->recover.mask = 0ULL;
 	spin_unlock_irqrestore(&fifo->base.lock, flags);
 
-	for (todo = mask; engn = __ffs64(todo), todo; todo &= ~BIT_ULL(engn))
-		engm |= 1 << gf100_fifo_engidx(fifo, engn);
 	nvkm_mask(device, 0x002630, engm, engm);
 
-	for (todo = mask; engn = __ffs64(todo), todo; todo &= ~BIT_ULL(engn)) {
-		if ((engine = nvkm_device_engine(device, engn))) {
+	for (todo = engm; engn = __ffs(todo), todo; todo &= ~BIT_ULL(engn)) {
+		if ((engine = gf100_fifo_id_engine(&fifo->base, engn))) {
 			nvkm_subdev_fini(&engine->subdev, false);
 			WARN_ON(nvkm_subdev_init(&engine->subdev));
 		}
@@ -179,17 +178,18 @@ gf100_fifo_recover(struct gf100_fifo *fifo, struct nvkm_engine *engine,
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
 	struct nvkm_device *device = subdev->device;
 	u32 chid = chan->base.chid;
+	int engi = gf100_fifo_engine_id(&fifo->base, engine);
 
 	nvkm_error(subdev, "%s engine fault on channel %d, recovering...\n",
-		   nvkm_subdev_name[engine->subdev.index], chid);
+		   engine->subdev.name, chid);
 	assert_spin_locked(&fifo->base.lock);
 
 	nvkm_mask(device, 0x003004 + (chid * 0x08), 0x00000001, 0x00000000);
 	list_del_init(&chan->head);
 	chan->killed = true;
 
-	if (engine != &fifo->base.engine)
-		fifo->recover.mask |= 1ULL << engine->subdev.index;
+	if (engi >= 0 && engi != GF100_FIFO_ENGN_SW)
+		fifo->recover.mask |= BIT(engi);
 	schedule_work(&fifo->recover.work);
 	nvkm_fifo_kevent(&fifo->base, chid);
 }
@@ -205,8 +205,8 @@ gf100_fifo_fault_engine[] = {
 	{ 0x11, "PMSPPP", NULL, NVKM_ENGINE_MSPPP },
 	{ 0x13, "PCOUNTER" },
 	{ 0x14, "PMSPDEC", NULL, NVKM_ENGINE_MSPDEC },
-	{ 0x15, "PCE0", NULL, NVKM_ENGINE_CE0 },
-	{ 0x16, "PCE1", NULL, NVKM_ENGINE_CE1 },
+	{ 0x15, "PCE0", NULL, NVKM_ENGINE_CE, 0 },
+	{ 0x16, "PCE1", NULL, NVKM_ENGINE_CE, 1 },
 	{ 0x17, "PMU" },
 	{}
 };
@@ -286,7 +286,7 @@ gf100_fifo_fault(struct nvkm_fifo *base, struct nvkm_fault_data *info)
 			nvkm_mask(device, 0x001718, 0x00000000, 0x00000000);
 			break;
 		default:
-			engine = nvkm_device_engine(device, eu->data2);
+			engine = nvkm_device_engine(device, eu->data2, eu->inst);
 			break;
 		}
 	}
@@ -335,7 +335,7 @@ gf100_fifo_intr_sched_ctxsw(struct gf100_fifo *fifo)
 		if (busy && unk0 && unk1) {
 			list_for_each_entry(chan, &fifo->chan, head) {
 				if (chan->base.chid == chid) {
-					engine = gf100_fifo_engine(fifo, engn);
+					engine = gf100_fifo_id_engine(&fifo->base, engn);
 					if (!engine)
 						break;
 					gf100_fifo_recover(fifo, engine, chan);
@@ -673,6 +673,8 @@ gf100_fifo = {
 	.fini = gf100_fifo_fini,
 	.intr = gf100_fifo_intr,
 	.fault = gf100_fifo_fault,
+	.engine_id = gf100_fifo_engine_id,
+	.id_engine = gf100_fifo_id_engine,
 	.uevent_init = gf100_fifo_uevent_init,
 	.uevent_fini = gf100_fifo_uevent_fini,
 	.chan = {
@@ -682,7 +684,8 @@ gf100_fifo = {
 };
 
 int
-gf100_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gf100_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
 	struct gf100_fifo *fifo;
 
@@ -692,5 +695,5 @@ gf100_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
 	INIT_WORK(&fifo->recover.work, gf100_fifo_recover_work);
 	*pfifo = &fifo->base;
 
-	return nvkm_fifo_ctor(&gf100_fifo, device, index, 128, &fifo->base);
+	return nvkm_fifo_ctor(&gf100_fifo, device, type, inst, 128, &fifo->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c
index 5d4b695cab8e..e771bd519ee2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c
@@ -36,19 +36,7 @@
 #include <nvif/class.h>
 #include <nvif/cl0080.h>
 
-struct gk104_fifo_engine_status {
-	bool busy;
-	bool faulted;
-	bool chsw;
-	bool save;
-	bool load;
-	struct {
-		bool tsg;
-		u32 id;
-	} prev, next, *chan;
-};
-
-static void
+void
 gk104_fifo_engine_status(struct gk104_fifo *fifo, int engn,
 			 struct gk104_fifo_engine_status *status)
 {
@@ -95,7 +83,7 @@ gk104_fifo_engine_status(struct gk104_fifo *fifo, int engn,
 		   status->chan == &status->next ? "*" : " ");
 }
 
-static int
+int
 gk104_fifo_class_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 		     void *argv, u32 argc, struct nvkm_object **pobject)
 {
@@ -112,7 +100,7 @@ gk104_fifo_class_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 	return -EINVAL;
 }
 
-static int
+int
 gk104_fifo_class_get(struct nvkm_fifo *base, int index,
 		     struct nvkm_oclass *oclass)
 {
@@ -134,14 +122,14 @@ gk104_fifo_class_get(struct nvkm_fifo *base, int index,
 	return c;
 }
 
-static void
+void
 gk104_fifo_uevent_fini(struct nvkm_fifo *fifo)
 {
 	struct nvkm_device *device = fifo->engine.subdev.device;
 	nvkm_mask(device, 0x002140, 0x80000000, 0x00000000);
 }
 
-static void
+void
 gk104_fifo_uevent_init(struct nvkm_fifo *fifo)
 {
 	struct nvkm_device *device = fifo->engine.subdev.device;
@@ -180,12 +168,11 @@ gk104_fifo_runlist_update(struct gk104_fifo *fifo, int runl)
 {
 	const struct gk104_fifo_runlist_func *func = fifo->func->runlist;
 	struct gk104_fifo_chan *chan;
-	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
 	struct nvkm_memory *mem;
 	struct nvkm_fifo_cgrp *cgrp;
 	int nr = 0;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&fifo->base.mutex);
 	mem = fifo->runlist[runl].mem[fifo->runlist[runl].next];
 	fifo->runlist[runl].next = !fifo->runlist[runl].next;
 
@@ -203,27 +190,27 @@ gk104_fifo_runlist_update(struct gk104_fifo *fifo, int runl)
 	nvkm_done(mem);
 
 	func->commit(fifo, runl, mem, nr);
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
 void
 gk104_fifo_runlist_remove(struct gk104_fifo *fifo, struct gk104_fifo_chan *chan)
 {
 	struct nvkm_fifo_cgrp *cgrp = chan->cgrp;
-	mutex_lock(&fifo->base.engine.subdev.mutex);
+	mutex_lock(&fifo->base.mutex);
 	if (!list_empty(&chan->head)) {
 		list_del_init(&chan->head);
 		if (cgrp && !--cgrp->chan_nr)
 			list_del_init(&cgrp->head);
 	}
-	mutex_unlock(&fifo->base.engine.subdev.mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
 void
 gk104_fifo_runlist_insert(struct gk104_fifo *fifo, struct gk104_fifo_chan *chan)
 {
 	struct nvkm_fifo_cgrp *cgrp = chan->cgrp;
-	mutex_lock(&fifo->base.engine.subdev.mutex);
+	mutex_lock(&fifo->base.mutex);
 	if (cgrp) {
 		if (!cgrp->chan_nr++)
 			list_add_tail(&cgrp->head, &fifo->runlist[chan->runl].cgrp);
@@ -231,7 +218,7 @@ gk104_fifo_runlist_insert(struct gk104_fifo *fifo, struct gk104_fifo_chan *chan)
 	} else {
 		list_add_tail(&chan->head, &fifo->runlist[chan->runl].chan);
 	}
-	mutex_unlock(&fifo->base.engine.subdev.mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
 void
@@ -271,6 +258,33 @@ gk104_fifo_pbdma = {
 	.init = gk104_fifo_pbdma_init,
 };
 
+struct nvkm_engine *
+gk104_fifo_id_engine(struct nvkm_fifo *base, int engi)
+{
+	if (engi == GK104_FIFO_ENGN_SW)
+		return nvkm_device_engine(base->engine.subdev.device, NVKM_ENGINE_SW, 0);
+
+	return gk104_fifo(base)->engine[engi].engine;
+}
+
+int
+gk104_fifo_engine_id(struct nvkm_fifo *base, struct nvkm_engine *engine)
+{
+	struct gk104_fifo *fifo = gk104_fifo(base);
+	int engn;
+
+	if (engine->subdev.type == NVKM_ENGINE_SW)
+		return GK104_FIFO_ENGN_SW;
+
+	for (engn = 0; engn < fifo->engine_nr && engine; engn++) {
+		if (fifo->engine[engn].engine == engine)
+			return engn;
+	}
+
+	WARN_ON(1);
+	return -1;
+}
+
 static void
 gk104_fifo_recover_work(struct work_struct *w)
 {
@@ -422,11 +436,12 @@ gk104_fifo_recover_engn(struct gk104_fifo *fifo, int engn)
 	 * called from the fault handler already.
 	 */
 	if (!status.faulted && engine) {
-		mmui = nvkm_top_fault_id(device, engine->subdev.index);
+		mmui = nvkm_top_fault_id(device, engine->subdev.type, engine->subdev.inst);
 		if (mmui < 0) {
 			const struct nvkm_enum *en = fifo->func->fault.engine;
 			for (; en && en->name; en++) {
-				if (en->data2 == engine->subdev.index) {
+				if (en->data2 == engine->subdev.type &&
+				    en->inst  == engine->subdev.inst) {
 					mmui = en->value;
 					break;
 				}
@@ -471,8 +486,8 @@ gk104_fifo_fault(struct nvkm_fifo *base, struct nvkm_fault_data *info)
 	struct nvkm_engine *engine = NULL;
 	struct nvkm_fifo_chan *chan;
 	unsigned long flags;
-	char ct[8] = "HUB/", en[16] = "";
-	int engn;
+	const char *en = "";
+	char ct[8] = "HUB/";
 
 	er = nvkm_enum_find(fifo->func->fault.reason, info->reason);
 	ee = nvkm_enum_find(fifo->func->fault.engine, info->engine);
@@ -496,23 +511,20 @@ gk104_fifo_fault(struct nvkm_fifo *base, struct nvkm_fault_data *info)
 			nvkm_mask(device, 0x001718, 0x00000000, 0x00000000);
 			break;
 		default:
-			engine = nvkm_device_engine(device, ee->data2);
+			engine = nvkm_device_engine(device, ee->data2, 0);
 			break;
 		}
 	}
 
 	if (ee == NULL) {
-		enum nvkm_devidx engidx = nvkm_top_fault(device, info->engine);
-		if (engidx < NVKM_SUBDEV_NR) {
-			const char *src = nvkm_subdev_name[engidx];
-			char *dst = en;
-			do {
-				*dst++ = toupper(*src++);
-			} while(*src);
-			engine = nvkm_device_engine(device, engidx);
+		struct nvkm_subdev *subdev = nvkm_top_fault(device, info->engine);
+		if (subdev) {
+			if (subdev->func == &nvkm_engine)
+				engine = container_of(subdev, typeof(*engine), subdev);
+			en = engine->subdev.name;
 		}
 	} else {
-		snprintf(en, sizeof(en), "%s", ee->name);
+		en = ee->name;
 	}
 
 	spin_lock_irqsave(&fifo->base.lock, flags);
@@ -535,11 +547,10 @@ gk104_fifo_fault(struct nvkm_fifo *base, struct nvkm_fault_data *info)
 	 * correct engine(s), but just in case we can't find the channel
 	 * information...
 	 */
-	for (engn = 0; engn < fifo->engine_nr && engine; engn++) {
-		if (fifo->engine[engn].engine == engine) {
+	if (engine) {
+		int engn = fifo->base.func->engine_id(&fifo->base, engine);
+		if (engn >= 0 && engn != GK104_FIFO_ENGN_SW)
 			gk104_fifo_recover_engn(fifo, engn);
-			break;
-		}
 	}
 
 	spin_unlock_irqrestore(&fifo->base.lock, flags);
@@ -556,7 +567,7 @@ gk104_fifo_bind_reason[] = {
 	{}
 };
 
-static void
+void
 gk104_fifo_intr_bind(struct gk104_fifo *fifo)
 {
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
@@ -627,7 +638,7 @@ gk104_fifo_intr_sched(struct gk104_fifo *fifo)
 	}
 }
 
-static void
+void
 gk104_fifo_intr_chsw(struct gk104_fifo *fifo)
 {
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
@@ -637,7 +648,7 @@ gk104_fifo_intr_chsw(struct gk104_fifo *fifo)
 	nvkm_wr32(device, 0x00256c, stat);
 }
 
-static void
+void
 gk104_fifo_intr_dropped_fault(struct gk104_fifo *fifo)
 {
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
@@ -680,7 +691,7 @@ static const struct nvkm_bitfield gk104_fifo_pbdma_intr_0[] = {
 	{}
 };
 
-static void
+void
 gk104_fifo_intr_pbdma_0(struct gk104_fifo *fifo, int unit)
 {
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
@@ -729,7 +740,7 @@ static const struct nvkm_bitfield gk104_fifo_pbdma_intr_1[] = {
 	{}
 };
 
-static void
+void
 gk104_fifo_intr_pbdma_1(struct gk104_fifo *fifo, int unit)
 {
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
@@ -750,7 +761,7 @@ gk104_fifo_intr_pbdma_1(struct gk104_fifo *fifo, int unit)
 	nvkm_wr32(device, 0x040148 + (unit * 0x2000), stat);
 }
 
-static void
+void
 gk104_fifo_intr_runlist(struct gk104_fifo *fifo)
 {
 	struct nvkm_device *device = fifo->base.engine.subdev.device;
@@ -763,7 +774,7 @@ gk104_fifo_intr_runlist(struct gk104_fifo *fifo)
 	}
 }
 
-static void
+void
 gk104_fifo_intr_engine(struct gk104_fifo *fifo)
 {
 	nvkm_fifo_uevent(&fifo->base);
@@ -861,7 +872,7 @@ gk104_fifo_intr(struct nvkm_fifo *base)
 	}
 }
 
-static void
+void
 gk104_fifo_fini(struct nvkm_fifo *base)
 {
 	struct gk104_fifo *fifo = gk104_fifo(base);
@@ -871,24 +882,46 @@ gk104_fifo_fini(struct nvkm_fifo *base)
 	nvkm_mask(device, 0x002140, 0x10000000, 0x10000000);
 }
 
-static int
+int
 gk104_fifo_info(struct nvkm_fifo *base, u64 mthd, u64 *data)
 {
 	struct gk104_fifo *fifo = gk104_fifo(base);
 	switch (mthd) {
-	case NV_DEVICE_FIFO_RUNLISTS:
+	case NV_DEVICE_HOST_RUNLISTS:
 		*data = (1ULL << fifo->runlist_nr) - 1;
 		return 0;
-	case NV_DEVICE_FIFO_RUNLIST_ENGINES(0)...
-	     NV_DEVICE_FIFO_RUNLIST_ENGINES(63): {
-		int runl = mthd - NV_DEVICE_FIFO_RUNLIST_ENGINES(0), engn;
-		if (runl < fifo->runlist_nr) {
-			unsigned long engm = fifo->runlist[runl].engm;
+	case NV_DEVICE_HOST_RUNLIST_ENGINES: {
+		if (*data < fifo->runlist_nr) {
+			unsigned long engm = fifo->runlist[*data].engm;
 			struct nvkm_engine *engine;
+			int engn;
 			*data = 0;
 			for_each_set_bit(engn, &engm, fifo->engine_nr) {
-				if ((engine = fifo->engine[engn].engine))
-					*data |= BIT_ULL(engine->subdev.index);
+				if ((engine = fifo->engine[engn].engine)) {
+#define CASE(n) case NVKM_ENGINE_##n: *data |= NV_DEVICE_HOST_RUNLIST_ENGINES_##n; break
+					switch (engine->subdev.type) {
+					CASE(SW    );
+					CASE(GR    );
+					CASE(MPEG  );
+					CASE(ME    );
+					CASE(CIPHER);
+					CASE(BSP   );
+					CASE(VP    );
+					CASE(CE    );
+					CASE(SEC   );
+					CASE(MSVLD );
+					CASE(MSPDEC);
+					CASE(MSPPP );
+					CASE(MSENC );
+					CASE(VIC   );
+					CASE(SEC2  );
+					CASE(NVDEC );
+					CASE(NVENC );
+					default:
+						WARN_ON(1);
+						break;
+					}
+				}
 			}
 			return 0;
 		}
@@ -899,15 +932,15 @@ gk104_fifo_info(struct nvkm_fifo *base, u64 mthd, u64 *data)
 	}
 }
 
-static int
+int
 gk104_fifo_oneinit(struct nvkm_fifo *base)
 {
 	struct gk104_fifo *fifo = gk104_fifo(base);
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
 	struct nvkm_device *device = subdev->device;
 	struct nvkm_vmm *bar = nvkm_bar_bar1_vmm(device);
-	int engn, runl, pbid, ret, i, j;
-	enum nvkm_devidx engidx;
+	struct nvkm_top_device *tdev;
+	int pbid, ret, i, j;
 	u32 *map;
 
 	fifo->pbdma_nr = fifo->func->pbdma->nr(fifo);
@@ -921,25 +954,41 @@ gk104_fifo_oneinit(struct nvkm_fifo *base)
 		map[i] = nvkm_rd32(device, 0x002390 + (i * 0x04));
 
 	/* Determine runlist configuration from topology device info. */
-	i = 0;
-	while ((int)(engidx = nvkm_top_engine(device, i++, &runl, &engn)) >= 0) {
+	list_for_each_entry(tdev, &device->top->device, head) {
+		const int engn = tdev->engine;
+		char _en[16], *en;
+
+		if (engn < 0)
+			continue;
+
 		/* Determine which PBDMA handles requests for this engine. */
 		for (j = 0, pbid = -1; j < fifo->pbdma_nr; j++) {
-			if (map[j] & (1 << runl)) {
+			if (map[j] & BIT(tdev->runlist)) {
 				pbid = j;
 				break;
 			}
 		}
 
+		fifo->engine[engn].engine = nvkm_device_engine(device, tdev->type, tdev->inst);
+		if (!fifo->engine[engn].engine) {
+			snprintf(_en, sizeof(_en), "%s, %d",
+				 nvkm_subdev_type[tdev->type], tdev->inst);
+			en = _en;
+		} else {
+			en = fifo->engine[engn].engine->subdev.name;
+		}
+
 		nvkm_debug(subdev, "engine %2d: runlist %2d pbdma %2d (%s)\n",
-			   engn, runl, pbid, nvkm_subdev_name[engidx]);
+			   tdev->engine, tdev->runlist, pbid, en);
 
-		fifo->engine[engn].engine = nvkm_device_engine(device, engidx);
-		fifo->engine[engn].runl = runl;
+		fifo->engine[engn].runl = tdev->runlist;
 		fifo->engine[engn].pbid = pbid;
 		fifo->engine_nr = max(fifo->engine_nr, engn + 1);
-		fifo->runlist[runl].engm |= 1 << engn;
-		fifo->runlist_nr = max(fifo->runlist_nr, runl + 1);
+		fifo->runlist[tdev->runlist].engm |= BIT(engn);
+		fifo->runlist[tdev->runlist].engm_sw |= BIT(engn);
+		if (tdev->type == NVKM_ENGINE_GR)
+			fifo->runlist[tdev->runlist].engm_sw |= BIT(GK104_FIFO_ENGN_SW);
+		fifo->runlist_nr = max(fifo->runlist_nr, tdev->runlist + 1);
 	}
 
 	kfree(map);
@@ -974,7 +1023,7 @@ gk104_fifo_oneinit(struct nvkm_fifo *base)
 	return nvkm_memory_map(fifo->user.mem, 0, bar, fifo->user.bar, NULL, 0);
 }
 
-static void
+void
 gk104_fifo_init(struct nvkm_fifo *base)
 {
 	struct gk104_fifo *fifo = gk104_fifo(base);
@@ -1006,7 +1055,7 @@ gk104_fifo_init(struct nvkm_fifo *base)
 	nvkm_wr32(device, 0x002140, 0x7fffffff);
 }
 
-static void *
+void *
 gk104_fifo_dtor(struct nvkm_fifo *base)
 {
 	struct gk104_fifo *fifo = gk104_fifo(base);
@@ -1033,6 +1082,8 @@ gk104_fifo_ = {
 	.fini = gk104_fifo_fini,
 	.intr = gk104_fifo_intr,
 	.fault = gk104_fifo_fault,
+	.engine_id = gk104_fifo_engine_id,
+	.id_engine = gk104_fifo_id_engine,
 	.uevent_init = gk104_fifo_uevent_init,
 	.uevent_fini = gk104_fifo_uevent_fini,
 	.recover_chan = gk104_fifo_recover_chan,
@@ -1042,7 +1093,7 @@ gk104_fifo_ = {
 
 int
 gk104_fifo_new_(const struct gk104_fifo_func *func, struct nvkm_device *device,
-		int index, int nr, struct nvkm_fifo **pfifo)
+		enum nvkm_subdev_type type, int inst, int nr, struct nvkm_fifo **pfifo)
 {
 	struct gk104_fifo *fifo;
 
@@ -1052,7 +1103,7 @@ gk104_fifo_new_(const struct gk104_fifo_func *func, struct nvkm_device *device,
 	INIT_WORK(&fifo->recover.work, gk104_fifo_recover_work);
 	*pfifo = &fifo->base;
 
-	return nvkm_fifo_ctor(&gk104_fifo_, device, index, nr, &fifo->base);
+	return nvkm_fifo_ctor(&gk104_fifo_, device, type, inst, nr, &fifo->base);
 }
 
 const struct nvkm_enum
@@ -1084,12 +1135,12 @@ gk104_fifo_fault_engine[] = {
 	{ 0x11, "MSPPP", NULL, NVKM_ENGINE_MSPPP },
 	{ 0x13, "PERF" },
 	{ 0x14, "MSPDEC", NULL, NVKM_ENGINE_MSPDEC },
-	{ 0x15, "CE0", NULL, NVKM_ENGINE_CE0 },
-	{ 0x16, "CE1", NULL, NVKM_ENGINE_CE1 },
+	{ 0x15, "CE0", NULL, NVKM_ENGINE_CE, 0 },
+	{ 0x16, "CE1", NULL, NVKM_ENGINE_CE, 1 },
 	{ 0x17, "PMU" },
 	{ 0x18, "PTP" },
 	{ 0x19, "MSENC", NULL, NVKM_ENGINE_MSENC },
-	{ 0x1b, "CE2", NULL, NVKM_ENGINE_CE2 },
+	{ 0x1b, "CE2", NULL, NVKM_ENGINE_CE, 2 },
 	{}
 };
 
@@ -1191,7 +1242,8 @@ gk104_fifo = {
 };
 
 int
-gk104_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gk104_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gk104_fifo, device, index, 4096, pfifo);
+	return gk104_fifo_new_(&gk104_fifo, device, type, inst, 4096, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h
index 6407a4a174cf..f2d12ae73944 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h
@@ -35,6 +35,7 @@ struct gk104_fifo {
 		struct list_head cgrp;
 		struct list_head chan;
 		u32 engm;
+		u32 engm_sw;
 	} runlist[16];
 	int runlist_nr;
 
@@ -87,11 +88,43 @@ struct gk104_fifo_func {
 	bool cgrp_force;
 };
 
-int gk104_fifo_new_(const struct gk104_fifo_func *, struct nvkm_device *,
+struct gk104_fifo_engine_status {
+	bool busy;
+	bool faulted;
+	bool chsw;
+	bool save;
+	bool load;
+	struct {
+		bool tsg;
+		u32 id;
+	} prev, next, *chan;
+};
+
+int gk104_fifo_new_(const struct gk104_fifo_func *, struct nvkm_device *, enum nvkm_subdev_type,
 		    int index, int nr, struct nvkm_fifo **);
 void gk104_fifo_runlist_insert(struct gk104_fifo *, struct gk104_fifo_chan *);
 void gk104_fifo_runlist_remove(struct gk104_fifo *, struct gk104_fifo_chan *);
 void gk104_fifo_runlist_update(struct gk104_fifo *, int runl);
+void gk104_fifo_engine_status(struct gk104_fifo *fifo, int engn,
+			      struct gk104_fifo_engine_status *status);
+void gk104_fifo_intr_bind(struct gk104_fifo *fifo);
+void gk104_fifo_intr_chsw(struct gk104_fifo *fifo);
+void gk104_fifo_intr_dropped_fault(struct gk104_fifo *fifo);
+void gk104_fifo_intr_pbdma_0(struct gk104_fifo *fifo, int unit);
+void gk104_fifo_intr_pbdma_1(struct gk104_fifo *fifo, int unit);
+void gk104_fifo_intr_runlist(struct gk104_fifo *fifo);
+void gk104_fifo_intr_engine(struct gk104_fifo *fifo);
+void *gk104_fifo_dtor(struct nvkm_fifo *base);
+int gk104_fifo_oneinit(struct nvkm_fifo *base);
+int gk104_fifo_info(struct nvkm_fifo *base, u64 mthd, u64 *data);
+void gk104_fifo_init(struct nvkm_fifo *base);
+void gk104_fifo_fini(struct nvkm_fifo *base);
+int gk104_fifo_class_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
+			 void *argv, u32 argc, struct nvkm_object **pobject);
+int gk104_fifo_class_get(struct nvkm_fifo *base, int index,
+			 struct nvkm_oclass *oclass);
+void gk104_fifo_uevent_fini(struct nvkm_fifo *fifo);
+void gk104_fifo_uevent_init(struct nvkm_fifo *fifo);
 
 extern const struct gk104_fifo_pbdma_func gk104_fifo_pbdma;
 int gk104_fifo_pbdma_nr(struct gk104_fifo *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c
index f820969e4405..915278c7e012 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c
@@ -60,7 +60,8 @@ gk110_fifo = {
 };
 
 int
-gk110_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gk110_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gk110_fifo, device, index, 4096, pfifo);
+	return gk104_fifo_new_(&gk110_fifo, device, type, inst, 4096, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c
index 2f54787b5fd0..cb703693de52 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c
@@ -57,7 +57,8 @@ gk208_fifo = {
 };
 
 int
-gk208_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gk208_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gk208_fifo, device, index, 1024, pfifo);
+	return gk104_fifo_new_(&gk208_fifo, device, type, inst, 1024, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c
index a814c4e0ed3e..6e35cf44c640 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c
@@ -38,7 +38,8 @@ gk20a_fifo = {
 };
 
 int
-gk20a_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gk20a_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gk20a_fifo, device, index, 128, pfifo);
+	return gk104_fifo_new_(&gk20a_fifo, device, type, inst, 128, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c
index c2a2e4572f6c..7af6e687d474 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c
@@ -106,7 +106,8 @@ gm107_fifo = {
 };
 
 int
-gm107_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gm107_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gm107_fifo, device, index, 2048, pfifo);
+	return gk104_fifo_new_(&gm107_fifo, device, type, inst, 2048, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c
index b8cfe3b28c4f..573658cb6c73 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c
@@ -54,7 +54,8 @@ gm200_fifo = {
 };
 
 int
-gm200_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gm200_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gm200_fifo, device, index, 4096, pfifo);
+	return gk104_fifo_new_(&gm200_fifo, device, type, inst, 4096, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c
index 70b4feebc1fa..556c97e54f14 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c
@@ -38,7 +38,8 @@ gm20b_fifo = {
 };
 
 int
-gm20b_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gm20b_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gm20b_fifo, device, index, 512, pfifo);
+	return gk104_fifo_new_(&gm20b_fifo, device, type, inst, 512, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c
index 2c7a0176b3c8..6b46b6b65b87 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c
@@ -91,7 +91,8 @@ gp100_fifo = {
 };
 
 int
-gp100_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gp100_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gp100_fifo, device, index, 4096, pfifo);
+	return gk104_fifo_new_(&gp100_fifo, device, type, inst, 4096, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c
index 8c65ad4feedb..7a5929cb4d29 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c
@@ -39,7 +39,8 @@ gp10b_fifo = {
 };
 
 int
-gp10b_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gp10b_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gp10b_fifo, device, index, 512, pfifo);
+	return gk104_fifo_new_(&gp10b_fifo, device, type, inst, 512, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogf100.c
index 75f9632789b3..4e78bbe3b94b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogf100.c
@@ -52,11 +52,10 @@ gf100_fifo_chan_ntfy(struct nvkm_fifo_chan *chan, u32 type,
 static u32
 gf100_fifo_gpfifo_engine_addr(struct nvkm_engine *engine)
 {
-	switch (engine->subdev.index) {
+	switch (engine->subdev.type) {
 	case NVKM_ENGINE_SW    : return 0;
 	case NVKM_ENGINE_GR    : return 0x0210;
-	case NVKM_ENGINE_CE0   : return 0x0230;
-	case NVKM_ENGINE_CE1   : return 0x0240;
+	case NVKM_ENGINE_CE    : return 0x0230 + (engine->subdev.inst * 0x10);
 	case NVKM_ENGINE_MSPDEC: return 0x0250;
 	case NVKM_ENGINE_MSPPP : return 0x0260;
 	case NVKM_ENGINE_MSVLD : return 0x0270;
@@ -66,6 +65,15 @@ gf100_fifo_gpfifo_engine_addr(struct nvkm_engine *engine)
 	}
 }
 
+static struct gf100_fifo_engn *
+gf100_fifo_gpfifo_engine(struct gf100_fifo_chan *chan, struct nvkm_engine *engine)
+{
+	int engi = chan->base.fifo->func->engine_id(chan->base.fifo, engine);
+	if (engi >= 0)
+		return &chan->engn[engi];
+	return NULL;
+}
+
 static int
 gf100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base,
 			      struct nvkm_engine *engine, bool suspend)
@@ -77,7 +85,7 @@ gf100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base,
 	struct nvkm_gpuobj *inst = chan->base.inst;
 	int ret = 0;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&chan->fifo->base.mutex);
 	nvkm_wr32(device, 0x002634, chan->base.chid);
 	if (nvkm_msec(device, 2000,
 		if (nvkm_rd32(device, 0x002634) == chan->base.chid)
@@ -87,7 +95,7 @@ gf100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base,
 			   chan->base.chid, chan->base.object.client->name);
 		ret = -ETIMEDOUT;
 	}
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&chan->fifo->base.mutex);
 
 	if (ret && suspend)
 		return ret;
@@ -108,13 +116,13 @@ gf100_fifo_gpfifo_engine_init(struct nvkm_fifo_chan *base,
 {
 	const u32 offset = gf100_fifo_gpfifo_engine_addr(engine);
 	struct gf100_fifo_chan *chan = gf100_fifo_chan(base);
+	struct gf100_fifo_engn *engn = gf100_fifo_gpfifo_engine(chan, engine);
 	struct nvkm_gpuobj *inst = chan->base.inst;
 
 	if (offset) {
-		u64 addr = chan->engn[engine->subdev.index].vma->addr;
 		nvkm_kmap(inst);
-		nvkm_wo32(inst, offset + 0x00, lower_32_bits(addr) | 4);
-		nvkm_wo32(inst, offset + 0x04, upper_32_bits(addr));
+		nvkm_wo32(inst, offset + 0x00, lower_32_bits(engn->vma->addr) | 4);
+		nvkm_wo32(inst, offset + 0x04, upper_32_bits(engn->vma->addr));
 		nvkm_done(inst);
 	}
 
@@ -126,8 +134,9 @@ gf100_fifo_gpfifo_engine_dtor(struct nvkm_fifo_chan *base,
 			      struct nvkm_engine *engine)
 {
 	struct gf100_fifo_chan *chan = gf100_fifo_chan(base);
-	nvkm_vmm_put(chan->base.vmm, &chan->engn[engine->subdev.index].vma);
-	nvkm_gpuobj_del(&chan->engn[engine->subdev.index].inst);
+	struct gf100_fifo_engn *engn = gf100_fifo_gpfifo_engine(chan, engine);
+	nvkm_vmm_put(chan->base.vmm, &engn->vma);
+	nvkm_gpuobj_del(&engn->inst);
 }
 
 static int
@@ -136,23 +145,21 @@ gf100_fifo_gpfifo_engine_ctor(struct nvkm_fifo_chan *base,
 			      struct nvkm_object *object)
 {
 	struct gf100_fifo_chan *chan = gf100_fifo_chan(base);
-	int engn = engine->subdev.index;
+	struct gf100_fifo_engn *engn = gf100_fifo_gpfifo_engine(chan, engine);
 	int ret;
 
 	if (!gf100_fifo_gpfifo_engine_addr(engine))
 		return 0;
 
-	ret = nvkm_object_bind(object, NULL, 0, &chan->engn[engn].inst);
+	ret = nvkm_object_bind(object, NULL, 0, &engn->inst);
 	if (ret)
 		return ret;
 
-	ret = nvkm_vmm_get(chan->base.vmm, 12, chan->engn[engn].inst->size,
-			   &chan->engn[engn].vma);
+	ret = nvkm_vmm_get(chan->base.vmm, 12, engn->inst->size, &engn->vma);
 	if (ret)
 		return ret;
 
-	return nvkm_memory_map(chan->engn[engn].inst, 0, chan->base.vmm,
-			       chan->engn[engn].vma, NULL, 0);
+	return nvkm_memory_map(engn->inst, 0, chan->base.vmm, engn->vma, NULL, 0);
 }
 
 static void
@@ -243,13 +250,13 @@ gf100_fifo_gpfifo_new(struct nvkm_fifo *base, const struct nvkm_oclass *oclass,
 
 	ret = nvkm_fifo_chan_ctor(&gf100_fifo_gpfifo_func, &fifo->base,
 				  0x1000, 0x1000, true, args->v0.vmm, 0,
-				  (1ULL << NVKM_ENGINE_CE0) |
-				  (1ULL << NVKM_ENGINE_CE1) |
-				  (1ULL << NVKM_ENGINE_GR) |
-				  (1ULL << NVKM_ENGINE_MSPDEC) |
-				  (1ULL << NVKM_ENGINE_MSPPP) |
-				  (1ULL << NVKM_ENGINE_MSVLD) |
-				  (1ULL << NVKM_ENGINE_SW),
+				  BIT(GF100_FIFO_ENGN_GR) |
+				  BIT(GF100_FIFO_ENGN_MSPDEC) |
+				  BIT(GF100_FIFO_ENGN_MSPPP) |
+				  BIT(GF100_FIFO_ENGN_MSVLD) |
+				  BIT(GF100_FIFO_ENGN_CE0) |
+				  BIT(GF100_FIFO_ENGN_CE1) |
+				  BIT(GF100_FIFO_ENGN_SW),
 				  1, fifo->user.bar->addr, 0x1000,
 				  oclass, &chan->base);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c
index 728a1edbf98c..b6900a52bcce 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c
@@ -65,19 +65,18 @@ int
 gk104_fifo_gpfifo_kick(struct gk104_fifo_chan *chan)
 {
 	int ret;
-	mutex_lock(&chan->base.fifo->engine.subdev.mutex);
+	mutex_lock(&chan->base.fifo->mutex);
 	ret = gk104_fifo_gpfifo_kick_locked(chan);
-	mutex_unlock(&chan->base.fifo->engine.subdev.mutex);
+	mutex_unlock(&chan->base.fifo->mutex);
 	return ret;
 }
 
 static u32
 gk104_fifo_gpfifo_engine_addr(struct nvkm_engine *engine)
 {
-	switch (engine->subdev.index) {
+	switch (engine->subdev.type) {
 	case NVKM_ENGINE_SW    :
-	case NVKM_ENGINE_CE0...NVKM_ENGINE_CE_LAST:
-		return 0;
+	case NVKM_ENGINE_CE    : return 0;
 	case NVKM_ENGINE_GR    : return 0x0210;
 	case NVKM_ENGINE_SEC   : return 0x0220;
 	case NVKM_ENGINE_MSPDEC: return 0x0250;
@@ -85,15 +84,26 @@ gk104_fifo_gpfifo_engine_addr(struct nvkm_engine *engine)
 	case NVKM_ENGINE_MSVLD : return 0x0270;
 	case NVKM_ENGINE_VIC   : return 0x0280;
 	case NVKM_ENGINE_MSENC : return 0x0290;
-	case NVKM_ENGINE_NVDEC0: return 0x02100270;
-	case NVKM_ENGINE_NVENC0: return 0x02100290;
-	case NVKM_ENGINE_NVENC1: return 0x0210;
+	case NVKM_ENGINE_NVDEC : return 0x02100270;
+	case NVKM_ENGINE_NVENC :
+		if (engine->subdev.inst)
+			return 0x0210;
+		return 0x02100290;
 	default:
 		WARN_ON(1);
 		return 0;
 	}
 }
 
+struct gk104_fifo_engn *
+gk104_fifo_gpfifo_engine(struct gk104_fifo_chan *chan, struct nvkm_engine *engine)
+{
+	int engi = chan->base.fifo->func->engine_id(chan->base.fifo, engine);
+	if (engi >= 0)
+		return &chan->engn[engi];
+	return NULL;
+}
+
 static int
 gk104_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base,
 			      struct nvkm_engine *engine, bool suspend)
@@ -126,13 +136,13 @@ gk104_fifo_gpfifo_engine_init(struct nvkm_fifo_chan *base,
 			      struct nvkm_engine *engine)
 {
 	struct gk104_fifo_chan *chan = gk104_fifo_chan(base);
+	struct gk104_fifo_engn *engn = gk104_fifo_gpfifo_engine(chan, engine);
 	struct nvkm_gpuobj *inst = chan->base.inst;
 	u32 offset = gk104_fifo_gpfifo_engine_addr(engine);
 
 	if (offset) {
-		u64   addr = chan->engn[engine->subdev.index].vma->addr;
-		u32 datalo = lower_32_bits(addr) | 0x00000004;
-		u32 datahi = upper_32_bits(addr);
+		u32 datalo = lower_32_bits(engn->vma->addr) | 0x00000004;
+		u32 datahi = upper_32_bits(engn->vma->addr);
 		nvkm_kmap(inst);
 		nvkm_wo32(inst, (offset & 0xffff) + 0x00, datalo);
 		nvkm_wo32(inst, (offset & 0xffff) + 0x04, datahi);
@@ -151,8 +161,9 @@ gk104_fifo_gpfifo_engine_dtor(struct nvkm_fifo_chan *base,
 			      struct nvkm_engine *engine)
 {
 	struct gk104_fifo_chan *chan = gk104_fifo_chan(base);
-	nvkm_vmm_put(chan->base.vmm, &chan->engn[engine->subdev.index].vma);
-	nvkm_gpuobj_del(&chan->engn[engine->subdev.index].inst);
+	struct gk104_fifo_engn *engn = gk104_fifo_gpfifo_engine(chan, engine);
+	nvkm_vmm_put(chan->base.vmm, &engn->vma);
+	nvkm_gpuobj_del(&engn->inst);
 }
 
 int
@@ -161,23 +172,21 @@ gk104_fifo_gpfifo_engine_ctor(struct nvkm_fifo_chan *base,
 			      struct nvkm_object *object)
 {
 	struct gk104_fifo_chan *chan = gk104_fifo_chan(base);
-	int engn = engine->subdev.index;
+	struct gk104_fifo_engn *engn = gk104_fifo_gpfifo_engine(chan, engine);
 	int ret;
 
 	if (!gk104_fifo_gpfifo_engine_addr(engine))
 		return 0;
 
-	ret = nvkm_object_bind(object, NULL, 0, &chan->engn[engn].inst);
+	ret = nvkm_object_bind(object, NULL, 0, &engn->inst);
 	if (ret)
 		return ret;
 
-	ret = nvkm_vmm_get(chan->base.vmm, 12, chan->engn[engn].inst->size,
-			   &chan->engn[engn].vma);
+	ret = nvkm_vmm_get(chan->base.vmm, 12, engn->inst->size, &engn->vma);
 	if (ret)
 		return ret;
 
-	return nvkm_memory_map(chan->engn[engn].inst, 0, chan->base.vmm,
-			       chan->engn[engn].vma, NULL, 0);
+	return nvkm_memory_map(engn->inst, 0, chan->base.vmm, engn->vma, NULL, 0);
 }
 
 void
@@ -247,23 +256,12 @@ gk104_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid,
 {
 	struct gk104_fifo_chan *chan;
 	int runlist = ffs(*runlists) -1, ret, i;
-	unsigned long engm;
-	u64 subdevs = 0;
 	u64 usermem;
 
 	if (!vmm || runlist < 0 || runlist >= fifo->runlist_nr)
 		return -EINVAL;
 	*runlists = BIT_ULL(runlist);
 
-	engm = fifo->runlist[runlist].engm;
-	for_each_set_bit(i, &engm, fifo->engine_nr) {
-		if (fifo->engine[i].engine)
-			subdevs |= BIT_ULL(fifo->engine[i].engine->subdev.index);
-	}
-
-	if (subdevs & BIT_ULL(NVKM_ENGINE_GR))
-		subdevs |= BIT_ULL(NVKM_ENGINE_SW);
-
 	/* Allocate the channel. */
 	if (!(chan = kzalloc(sizeof(*chan), GFP_KERNEL)))
 		return -ENOMEM;
@@ -273,7 +271,7 @@ gk104_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid,
 	INIT_LIST_HEAD(&chan->head);
 
 	ret = nvkm_fifo_chan_ctor(&gk104_fifo_gpfifo_func, &fifo->base,
-				  0x1000, 0x1000, true, vmm, 0, subdevs,
+				  0x1000, 0x1000, true, vmm, 0, fifo->runlist[runlist].engm_sw,
 				  1, fifo->user.bar->addr, 0x200,
 				  oclass, &chan->base);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c
index a7462cf59d65..ee4967b706a7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c
@@ -44,7 +44,7 @@ gv100_fifo_gpfifo_engine_valid(struct gk104_fifo_chan *chan, bool ce, bool valid
 	int ret;
 
 	/* Block runlist to prevent the channel from being rescheduled. */
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&chan->fifo->base.mutex);
 	nvkm_mask(device, 0x002630, BIT(chan->runl), BIT(chan->runl));
 
 	/* Preempt the channel. */
@@ -58,7 +58,7 @@ gv100_fifo_gpfifo_engine_valid(struct gk104_fifo_chan *chan, bool ce, bool valid
 
 	/* Resume runlist. */
 	nvkm_mask(device, 0x002630, BIT(chan->runl), 0);
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&chan->fifo->base.mutex);
 	return ret;
 }
 
@@ -70,8 +70,7 @@ gv100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base,
 	struct nvkm_gpuobj *inst = chan->base.inst;
 	int ret;
 
-	if (engine->subdev.index >= NVKM_ENGINE_CE0 &&
-	    engine->subdev.index <= NVKM_ENGINE_CE_LAST)
+	if (engine->subdev.type == NVKM_ENGINE_CE)
 		return gk104_fifo_gpfifo_kick(chan);
 
 	ret = gv100_fifo_gpfifo_engine_valid(chan, false, false);
@@ -90,17 +89,15 @@ gv100_fifo_gpfifo_engine_init(struct nvkm_fifo_chan *base,
 			      struct nvkm_engine *engine)
 {
 	struct gk104_fifo_chan *chan = gk104_fifo_chan(base);
+	struct gk104_fifo_engn *engn = gk104_fifo_gpfifo_engine(chan, engine);
 	struct nvkm_gpuobj *inst = chan->base.inst;
-	u64 addr;
 
-	if (engine->subdev.index >= NVKM_ENGINE_CE0 &&
-	    engine->subdev.index <= NVKM_ENGINE_CE_LAST)
+	if (engine->subdev.type == NVKM_ENGINE_CE)
 		return 0;
 
-	addr = chan->engn[engine->subdev.index].vma->addr;
 	nvkm_kmap(inst);
-	nvkm_wo32(inst, 0x210, lower_32_bits(addr) | 0x00000004);
-	nvkm_wo32(inst, 0x214, upper_32_bits(addr));
+	nvkm_wo32(inst, 0x210, lower_32_bits(engn->vma->addr) | 0x00000004);
+	nvkm_wo32(inst, 0x214, upper_32_bits(engn->vma->addr));
 	nvkm_done(inst);
 
 	return gv100_fifo_gpfifo_engine_valid(chan, false, true);
@@ -129,8 +126,6 @@ gv100_fifo_gpfifo_new_(const struct nvkm_fifo_chan_func *func,
 	struct nvkm_device *device = fifo->base.engine.subdev.device;
 	struct gk104_fifo_chan *chan;
 	int runlist = ffs(*runlists) -1, ret, i;
-	unsigned long engm;
-	u64 subdevs = 0;
 	u64 usermem, mthd;
 	u32 size;
 
@@ -138,12 +133,6 @@ gv100_fifo_gpfifo_new_(const struct nvkm_fifo_chan_func *func,
 		return -EINVAL;
 	*runlists = BIT_ULL(runlist);
 
-	engm = fifo->runlist[runlist].engm;
-	for_each_set_bit(i, &engm, fifo->engine_nr) {
-		if (fifo->engine[i].engine)
-			subdevs |= BIT_ULL(fifo->engine[i].engine->subdev.index);
-	}
-
 	/* Allocate the channel. */
 	if (!(chan = kzalloc(sizeof(*chan), GFP_KERNEL)))
 		return -ENOMEM;
@@ -153,7 +142,7 @@ gv100_fifo_gpfifo_new_(const struct nvkm_fifo_chan_func *func,
 	INIT_LIST_HEAD(&chan->head);
 
 	ret = nvkm_fifo_chan_ctor(func, &fifo->base, 0x1000, 0x1000, true, vmm,
-				  0, subdevs, 1, fifo->user.bar->addr, 0x200,
+				  0, fifo->runlist[runlist].engm, 1, fifo->user.bar->addr, 0x200,
 				  oclass, &chan->base);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c
index 6ee1bb32a071..70e16a91ac12 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c
@@ -301,7 +301,8 @@ gv100_fifo = {
 };
 
 int
-gv100_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+gv100_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&gv100_fifo, device, index, 4096, pfifo);
+	return gk104_fifo_new_(&gv100_fifo, device, type, inst, 4096, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c
index c1d1b1aa5bc6..c6730c124769 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c
@@ -94,6 +94,38 @@ __releases(fifo->base.lock)
 	spin_unlock_irqrestore(&fifo->base.lock, flags);
 }
 
+struct nvkm_engine *
+nv04_fifo_id_engine(struct nvkm_fifo *fifo, int engi)
+{
+	enum nvkm_subdev_type type;
+
+	switch (engi) {
+	case NV04_FIFO_ENGN_SW  : type = NVKM_ENGINE_SW; break;
+	case NV04_FIFO_ENGN_GR  : type = NVKM_ENGINE_GR; break;
+	case NV04_FIFO_ENGN_MPEG: type = NVKM_ENGINE_MPEG; break;
+	case NV04_FIFO_ENGN_DMA : type = NVKM_ENGINE_DMAOBJ; break;
+	default:
+		WARN_ON(1);
+		return NULL;
+	}
+
+	return nvkm_device_engine(fifo->engine.subdev.device, type, 0);
+}
+
+int
+nv04_fifo_engine_id(struct nvkm_fifo *base, struct nvkm_engine *engine)
+{
+	switch (engine->subdev.type) {
+	case NVKM_ENGINE_SW    : return NV04_FIFO_ENGN_SW;
+	case NVKM_ENGINE_GR    : return NV04_FIFO_ENGN_GR;
+	case NVKM_ENGINE_MPEG  : return NV04_FIFO_ENGN_MPEG;
+	case NVKM_ENGINE_DMAOBJ: return NV04_FIFO_ENGN_DMA;
+	default:
+		WARN_ON(1);
+		return 0;
+	}
+}
+
 static const char *
 nv_dma_state_err(u32 state)
 {
@@ -326,7 +358,7 @@ nv04_fifo_init(struct nvkm_fifo *base)
 
 int
 nv04_fifo_new_(const struct nvkm_fifo_func *func, struct nvkm_device *device,
-	       int index, int nr, const struct nv04_fifo_ramfc *ramfc,
+	       enum nvkm_subdev_type type, int inst, int nr, const struct nv04_fifo_ramfc *ramfc,
 	       struct nvkm_fifo **pfifo)
 {
 	struct nv04_fifo *fifo;
@@ -337,7 +369,7 @@ nv04_fifo_new_(const struct nvkm_fifo_func *func, struct nvkm_device *device,
 	fifo->ramfc = ramfc;
 	*pfifo = &fifo->base;
 
-	ret = nvkm_fifo_ctor(func, device, index, nr, &fifo->base);
+	ret = nvkm_fifo_ctor(func, device, type, inst, nr, &fifo->base);
 	if (ret)
 		return ret;
 
@@ -349,6 +381,8 @@ static const struct nvkm_fifo_func
 nv04_fifo = {
 	.init = nv04_fifo_init,
 	.intr = nv04_fifo_intr,
+	.engine_id = nv04_fifo_engine_id,
+	.id_engine = nv04_fifo_id_engine,
 	.pause = nv04_fifo_pause,
 	.start = nv04_fifo_start,
 	.chan = {
@@ -358,8 +392,8 @@ nv04_fifo = {
 };
 
 int
-nv04_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+nv04_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_fifo **pfifo)
 {
-	return nv04_fifo_new_(&nv04_fifo, device, index, 16,
-			      nv04_fifo_ramfc, pfifo);
+	return nv04_fifo_new_(&nv04_fifo, device, type, inst, 16, nv04_fifo_ramfc, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.h
index e5ecceee77ae..3f23bcde4a54 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.h
@@ -17,8 +17,7 @@ struct nv04_fifo {
 	const struct nv04_fifo_ramfc *ramfc;
 };
 
-int nv04_fifo_new_(const struct nvkm_fifo_func *, struct nvkm_device *,
-		   int index, int nr, const struct nv04_fifo_ramfc *,
-		   struct nvkm_fifo **);
+int nv04_fifo_new_(const struct nvkm_fifo_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   int nr, const struct nv04_fifo_ramfc *, struct nvkm_fifo **);
 void nv04_fifo_init(struct nvkm_fifo *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv10.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv10.c
index f9a87deb2b3d..f8887f0f2f82 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv10.c
@@ -43,6 +43,8 @@ static const struct nvkm_fifo_func
 nv10_fifo = {
 	.init = nv04_fifo_init,
 	.intr = nv04_fifo_intr,
+	.engine_id = nv04_fifo_engine_id,
+	.id_engine = nv04_fifo_id_engine,
 	.pause = nv04_fifo_pause,
 	.start = nv04_fifo_start,
 	.chan = {
@@ -52,8 +54,8 @@ nv10_fifo = {
 };
 
 int
-nv10_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+nv10_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_fifo **pfifo)
 {
-	return nv04_fifo_new_(&nv10_fifo, device, index, 32,
-			      nv10_fifo_ramfc, pfifo);
+	return nv04_fifo_new_(&nv10_fifo, device, type, inst, 32, nv10_fifo_ramfc, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv17.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv17.c
index f6d383a21222..3f94c7b5b054 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv17.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv17.c
@@ -81,6 +81,8 @@ static const struct nvkm_fifo_func
 nv17_fifo = {
 	.init = nv17_fifo_init,
 	.intr = nv04_fifo_intr,
+	.engine_id = nv04_fifo_engine_id,
+	.id_engine = nv04_fifo_id_engine,
 	.pause = nv04_fifo_pause,
 	.start = nv04_fifo_start,
 	.chan = {
@@ -90,8 +92,8 @@ nv17_fifo = {
 };
 
 int
-nv17_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+nv17_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_fifo **pfifo)
 {
-	return nv04_fifo_new_(&nv17_fifo, device, index, 32,
-			      nv17_fifo_ramfc, pfifo);
+	return nv04_fifo_new_(&nv17_fifo, device, type, inst, 32, nv17_fifo_ramfc, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c
index 2d61fd832ddb..f9ea46809bc0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c
@@ -112,6 +112,8 @@ static const struct nvkm_fifo_func
 nv40_fifo = {
 	.init = nv40_fifo_init,
 	.intr = nv04_fifo_intr,
+	.engine_id = nv04_fifo_engine_id,
+	.id_engine = nv04_fifo_id_engine,
 	.pause = nv04_fifo_pause,
 	.start = nv04_fifo_start,
 	.chan = {
@@ -121,8 +123,8 @@ nv40_fifo = {
 };
 
 int
-nv40_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+nv40_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_fifo **pfifo)
 {
-	return nv04_fifo_new_(&nv40_fifo, device, index, 32,
-			      nv40_fifo_ramfc, pfifo);
+	return nv04_fifo_new_(&nv40_fifo, device, type, inst, 32, nv40_fifo_ramfc, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c
index fa6e094d8068..be94156ea248 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c
@@ -51,9 +51,9 @@ nv50_fifo_runlist_update_locked(struct nv50_fifo *fifo)
 void
 nv50_fifo_runlist_update(struct nv50_fifo *fifo)
 {
-	mutex_lock(&fifo->base.engine.subdev.mutex);
+	mutex_lock(&fifo->base.mutex);
 	nv50_fifo_runlist_update_locked(fifo);
-	mutex_unlock(&fifo->base.engine.subdev.mutex);
+	mutex_unlock(&fifo->base.mutex);
 }
 
 int
@@ -107,7 +107,7 @@ nv50_fifo_dtor(struct nvkm_fifo *base)
 
 int
 nv50_fifo_new_(const struct nvkm_fifo_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_fifo **pfifo)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_fifo **pfifo)
 {
 	struct nv50_fifo *fifo;
 	int ret;
@@ -116,7 +116,7 @@ nv50_fifo_new_(const struct nvkm_fifo_func *func, struct nvkm_device *device,
 		return -ENOMEM;
 	*pfifo = &fifo->base;
 
-	ret = nvkm_fifo_ctor(func, device, index, 128, &fifo->base);
+	ret = nvkm_fifo_ctor(func, device, type, inst, 128, &fifo->base);
 	if (ret)
 		return ret;
 
@@ -131,6 +131,8 @@ nv50_fifo = {
 	.oneinit = nv50_fifo_oneinit,
 	.init = nv50_fifo_init,
 	.intr = nv04_fifo_intr,
+	.engine_id = nv04_fifo_engine_id,
+	.id_engine = nv04_fifo_id_engine,
 	.pause = nv04_fifo_pause,
 	.start = nv04_fifo_start,
 	.chan = {
@@ -141,7 +143,8 @@ nv50_fifo = {
 };
 
 int
-nv50_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+nv50_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_fifo **pfifo)
 {
-	return nv50_fifo_new_(&nv50_fifo, device, index, pfifo);
+	return nv50_fifo_new_(&nv50_fifo, device, type, inst, pfifo);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.h
index 87d30b6bd2ea..0111e7e5a4e3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.h
@@ -10,8 +10,8 @@ struct nv50_fifo {
 	int cur_runlist;
 };
 
-int nv50_fifo_new_(const struct nvkm_fifo_func *, struct nvkm_device *,
-		   int index, struct nvkm_fifo **);
+int nv50_fifo_new_(const struct nvkm_fifo_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_fifo **);
 
 void *nv50_fifo_dtor(struct nvkm_fifo *);
 int nv50_fifo_oneinit(struct nvkm_fifo *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h
index 0ef8baab513e..899272801a8b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h
@@ -4,8 +4,8 @@
 #define nvkm_fifo(p) container_of((p), struct nvkm_fifo, engine)
 #include <engine/fifo.h>
 
-int nvkm_fifo_ctor(const struct nvkm_fifo_func *, struct nvkm_device *,
-		   int index, int nr, struct nvkm_fifo *);
+int nvkm_fifo_ctor(const struct nvkm_fifo_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   int nr, struct nvkm_fifo *);
 void nvkm_fifo_uevent(struct nvkm_fifo *);
 void nvkm_fifo_cevent(struct nvkm_fifo *);
 void nvkm_fifo_kevent(struct nvkm_fifo *, int chid);
@@ -23,6 +23,8 @@ struct nvkm_fifo_func {
 	void (*fini)(struct nvkm_fifo *);
 	void (*intr)(struct nvkm_fifo *);
 	void (*fault)(struct nvkm_fifo *, struct nvkm_fault_data *);
+	int (*engine_id)(struct nvkm_fifo *, struct nvkm_engine *);
+	struct nvkm_engine *(*id_engine)(struct nvkm_fifo *, int engi);
 	void (*pause)(struct nvkm_fifo *, unsigned long *);
 	void (*start)(struct nvkm_fifo *, unsigned long *);
 	void (*uevent_init)(struct nvkm_fifo *);
@@ -35,8 +37,13 @@ struct nvkm_fifo_func {
 };
 
 void nv04_fifo_intr(struct nvkm_fifo *);
+int nv04_fifo_engine_id(struct nvkm_fifo *, struct nvkm_engine *);
+struct nvkm_engine *nv04_fifo_id_engine(struct nvkm_fifo *, int);
 void nv04_fifo_pause(struct nvkm_fifo *, unsigned long *);
 void nv04_fifo_start(struct nvkm_fifo *, unsigned long *);
 
 void gf100_fifo_intr_fault(struct nvkm_fifo *, int);
+
+int gk104_fifo_engine_id(struct nvkm_fifo *, struct nvkm_engine *);
+struct nvkm_engine *gk104_fifo_id_engine(struct nvkm_fifo *, int);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c
index 005f3e1729b9..e417044cc347 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c
@@ -24,7 +24,13 @@
 #include "changk104.h"
 #include "user.h"
 
+#include <core/client.h>
 #include <core/gpuobj.h>
+#include <subdev/bar.h>
+#include <subdev/fault.h>
+#include <subdev/top.h>
+#include <subdev/timer.h>
+#include <engine/sw.h>
 
 #include <nvif/class.h>
 
@@ -109,8 +115,363 @@ tu102_fifo = {
 	.cgrp_force = true,
 };
 
+static void
+tu102_fifo_recover_work(struct work_struct *w)
+{
+	struct gk104_fifo *fifo = container_of(w, typeof(*fifo), recover.work);
+	struct nvkm_device *device = fifo->base.engine.subdev.device;
+	struct nvkm_engine *engine;
+	unsigned long flags;
+	u32 engm, runm, todo;
+	int engn, runl;
+
+	spin_lock_irqsave(&fifo->base.lock, flags);
+	runm = fifo->recover.runm;
+	engm = fifo->recover.engm;
+	fifo->recover.engm = 0;
+	fifo->recover.runm = 0;
+	spin_unlock_irqrestore(&fifo->base.lock, flags);
+
+	nvkm_mask(device, 0x002630, runm, runm);
+
+	for (todo = engm; engn = __ffs(todo), todo; todo &= ~BIT(engn)) {
+		if ((engine = fifo->engine[engn].engine)) {
+			nvkm_subdev_fini(&engine->subdev, false);
+			WARN_ON(nvkm_subdev_init(&engine->subdev));
+		}
+	}
+
+	for (todo = runm; runl = __ffs(todo), todo; todo &= ~BIT(runl))
+		gk104_fifo_runlist_update(fifo, runl);
+
+	nvkm_mask(device, 0x002630, runm, 0x00000000);
+}
+
+static void tu102_fifo_recover_engn(struct gk104_fifo *fifo, int engn);
+
+static void
+tu102_fifo_recover_runl(struct gk104_fifo *fifo, int runl)
+{
+	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	const u32 runm = BIT(runl);
+
+	assert_spin_locked(&fifo->base.lock);
+	if (fifo->recover.runm & runm)
+		return;
+	fifo->recover.runm |= runm;
+
+	/* Block runlist to prevent channel assignment(s) from changing. */
+	nvkm_mask(device, 0x002630, runm, runm);
+
+	/* Schedule recovery. */
+	nvkm_warn(subdev, "runlist %d: scheduled for recovery\n", runl);
+	schedule_work(&fifo->recover.work);
+}
+
+static struct gk104_fifo_chan *
+tu102_fifo_recover_chid(struct gk104_fifo *fifo, int runl, int chid)
+{
+	struct gk104_fifo_chan *chan;
+	struct nvkm_fifo_cgrp *cgrp;
+
+	list_for_each_entry(chan, &fifo->runlist[runl].chan, head) {
+		if (chan->base.chid == chid) {
+			list_del_init(&chan->head);
+			return chan;
+		}
+	}
+
+	list_for_each_entry(cgrp, &fifo->runlist[runl].cgrp, head) {
+		if (cgrp->id == chid) {
+			chan = list_first_entry(&cgrp->chan, typeof(*chan), head);
+			list_del_init(&chan->head);
+			if (!--cgrp->chan_nr)
+				list_del_init(&cgrp->head);
+			return chan;
+		}
+	}
+
+	return NULL;
+}
+
+static void
+tu102_fifo_recover_chan(struct nvkm_fifo *base, int chid)
+{
+	struct gk104_fifo *fifo = gk104_fifo(base);
+	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	const u32  stat = nvkm_rd32(device, 0x800004 + (chid * 0x08));
+	const u32  runl = (stat & 0x000f0000) >> 16;
+	const bool used = (stat & 0x00000001);
+	unsigned long engn, engm = fifo->runlist[runl].engm;
+	struct gk104_fifo_chan *chan;
+
+	assert_spin_locked(&fifo->base.lock);
+	if (!used)
+		return;
+
+	/* Lookup SW state for channel, and mark it as dead. */
+	chan = tu102_fifo_recover_chid(fifo, runl, chid);
+	if (chan) {
+		chan->killed = true;
+		nvkm_fifo_kevent(&fifo->base, chid);
+	}
+
+	/* Disable channel. */
+	nvkm_wr32(device, 0x800004 + (chid * 0x08), stat | 0x00000800);
+	nvkm_warn(subdev, "channel %d: killed\n", chid);
+
+	/* Block channel assignments from changing during recovery. */
+	tu102_fifo_recover_runl(fifo, runl);
+
+	/* Schedule recovery for any engines the channel is on. */
+	for_each_set_bit(engn, &engm, fifo->engine_nr) {
+		struct gk104_fifo_engine_status status;
+
+		gk104_fifo_engine_status(fifo, engn, &status);
+		if (!status.chan || status.chan->id != chid)
+			continue;
+		tu102_fifo_recover_engn(fifo, engn);
+	}
+}
+
+static void
+tu102_fifo_recover_engn(struct gk104_fifo *fifo, int engn)
+{
+	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	const u32 runl = fifo->engine[engn].runl;
+	const u32 engm = BIT(engn);
+	struct gk104_fifo_engine_status status;
+
+	assert_spin_locked(&fifo->base.lock);
+	if (fifo->recover.engm & engm)
+		return;
+	fifo->recover.engm |= engm;
+
+	/* Block channel assignments from changing during recovery. */
+	tu102_fifo_recover_runl(fifo, runl);
+
+	/* Determine which channel (if any) is currently on the engine. */
+	gk104_fifo_engine_status(fifo, engn, &status);
+	if (status.chan) {
+		/* The channel is not longer viable, kill it. */
+		tu102_fifo_recover_chan(&fifo->base, status.chan->id);
+	}
+
+	/* Preempt the runlist */
+	nvkm_wr32(device, 0x2638, BIT(runl));
+
+	/* Schedule recovery. */
+	nvkm_warn(subdev, "engine %d: scheduled for recovery\n", engn);
+	schedule_work(&fifo->recover.work);
+}
+
+static void
+tu102_fifo_fault(struct nvkm_fifo *base, struct nvkm_fault_data *info)
+{
+	struct gk104_fifo *fifo = gk104_fifo(base);
+	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	const struct nvkm_enum *er, *ee, *ec, *ea;
+	struct nvkm_engine *engine = NULL;
+	struct nvkm_fifo_chan *chan;
+	unsigned long flags;
+	const char *en = "";
+	char ct[8] = "HUB/";
+	int engn;
+
+	er = nvkm_enum_find(fifo->func->fault.reason, info->reason);
+	ee = nvkm_enum_find(fifo->func->fault.engine, info->engine);
+	if (info->hub) {
+		ec = nvkm_enum_find(fifo->func->fault.hubclient, info->client);
+	} else {
+		ec = nvkm_enum_find(fifo->func->fault.gpcclient, info->client);
+		snprintf(ct, sizeof(ct), "GPC%d/", info->gpc);
+	}
+	ea = nvkm_enum_find(fifo->func->fault.access, info->access);
+
+	if (ee && ee->data2) {
+		switch (ee->data2) {
+		case NVKM_SUBDEV_BAR:
+			nvkm_bar_bar1_reset(device);
+			break;
+		case NVKM_SUBDEV_INSTMEM:
+			nvkm_bar_bar2_reset(device);
+			break;
+		case NVKM_ENGINE_IFB:
+			nvkm_mask(device, 0x001718, 0x00000000, 0x00000000);
+			break;
+		default:
+			engine = nvkm_device_engine(device, ee->data2, 0);
+			break;
+		}
+	}
+
+	if (ee == NULL) {
+		struct nvkm_subdev *subdev = nvkm_top_fault(device, info->engine);
+		if (subdev) {
+			if (subdev->func == &nvkm_engine)
+				engine = container_of(subdev, typeof(*engine), subdev);
+			en = engine->subdev.name;
+		}
+	} else {
+		en = ee->name;
+	}
+
+	spin_lock_irqsave(&fifo->base.lock, flags);
+	chan = nvkm_fifo_chan_inst_locked(&fifo->base, info->inst);
+
+	nvkm_error(subdev,
+		   "fault %02x [%s] at %016llx engine %02x [%s] client %02x "
+		   "[%s%s] reason %02x [%s] on channel %d [%010llx %s]\n",
+		   info->access, ea ? ea->name : "", info->addr,
+		   info->engine, ee ? ee->name : en,
+		   info->client, ct, ec ? ec->name : "",
+		   info->reason, er ? er->name : "", chan ? chan->chid : -1,
+		   info->inst, chan ? chan->object.client->name : "unknown");
+
+	/* Kill the channel that caused the fault. */
+	if (chan)
+		tu102_fifo_recover_chan(&fifo->base, chan->chid);
+
+	/* Channel recovery will probably have already done this for the
+	 * correct engine(s), but just in case we can't find the channel
+	 * information...
+	 */
+	for (engn = 0; engn < fifo->engine_nr && engine; engn++) {
+		if (fifo->engine[engn].engine == engine) {
+			tu102_fifo_recover_engn(fifo, engn);
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&fifo->base.lock, flags);
+}
+
+static void
+tu102_fifo_intr_ctxsw_timeout(struct gk104_fifo *fifo)
+{
+	struct nvkm_device *device = fifo->base.engine.subdev.device;
+	unsigned long flags, engm;
+	u32 engn;
+
+	spin_lock_irqsave(&fifo->base.lock, flags);
+
+	engm = nvkm_rd32(device, 0x2a30);
+	nvkm_wr32(device, 0x2a30, engm);
+
+	for_each_set_bit(engn, &engm, 32)
+		tu102_fifo_recover_engn(fifo, engn);
+
+	spin_unlock_irqrestore(&fifo->base.lock, flags);
+}
+
+static void
+tu102_fifo_intr_sched(struct gk104_fifo *fifo)
+{
+	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 intr = nvkm_rd32(device, 0x00254c);
+	u32 code = intr & 0x000000ff;
+
+	nvkm_error(subdev, "SCHED_ERROR %02x\n", code);
+}
+
+static void
+tu102_fifo_intr(struct nvkm_fifo *base)
+{
+	struct gk104_fifo *fifo = gk104_fifo(base);
+	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 mask = nvkm_rd32(device, 0x002140);
+	u32 stat = nvkm_rd32(device, 0x002100) & mask;
+
+	if (stat & 0x00000001) {
+		gk104_fifo_intr_bind(fifo);
+		nvkm_wr32(device, 0x002100, 0x00000001);
+		stat &= ~0x00000001;
+	}
+
+	if (stat & 0x00000002) {
+		tu102_fifo_intr_ctxsw_timeout(fifo);
+		stat &= ~0x00000002;
+	}
+
+	if (stat & 0x00000100) {
+		tu102_fifo_intr_sched(fifo);
+		nvkm_wr32(device, 0x002100, 0x00000100);
+		stat &= ~0x00000100;
+	}
+
+	if (stat & 0x00010000) {
+		gk104_fifo_intr_chsw(fifo);
+		nvkm_wr32(device, 0x002100, 0x00010000);
+		stat &= ~0x00010000;
+	}
+
+	if (stat & 0x20000000) {
+		u32 mask = nvkm_rd32(device, 0x0025a0);
+
+		while (mask) {
+			u32 unit = __ffs(mask);
+
+			gk104_fifo_intr_pbdma_0(fifo, unit);
+			gk104_fifo_intr_pbdma_1(fifo, unit);
+			nvkm_wr32(device, 0x0025a0, (1 << unit));
+			mask &= ~(1 << unit);
+		}
+		stat &= ~0x20000000;
+	}
+
+	if (stat & 0x40000000) {
+		gk104_fifo_intr_runlist(fifo);
+		stat &= ~0x40000000;
+	}
+
+	if (stat & 0x80000000) {
+		nvkm_wr32(device, 0x002100, 0x80000000);
+		gk104_fifo_intr_engine(fifo);
+		stat &= ~0x80000000;
+	}
+
+	if (stat) {
+		nvkm_error(subdev, "INTR %08x\n", stat);
+		nvkm_mask(device, 0x002140, stat, 0x00000000);
+		nvkm_wr32(device, 0x002100, stat);
+	}
+}
+
+static const struct nvkm_fifo_func
+tu102_fifo_ = {
+	.dtor = gk104_fifo_dtor,
+	.oneinit = gk104_fifo_oneinit,
+	.info = gk104_fifo_info,
+	.init = gk104_fifo_init,
+	.fini = gk104_fifo_fini,
+	.intr = tu102_fifo_intr,
+	.fault = tu102_fifo_fault,
+	.engine_id = gk104_fifo_engine_id,
+	.id_engine = gk104_fifo_id_engine,
+	.uevent_init = gk104_fifo_uevent_init,
+	.uevent_fini = gk104_fifo_uevent_fini,
+	.recover_chan = tu102_fifo_recover_chan,
+	.class_get = gk104_fifo_class_get,
+	.class_new = gk104_fifo_class_new,
+};
+
 int
-tu102_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+tu102_fifo_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fifo **pfifo)
 {
-	return gk104_fifo_new_(&tu102_fifo, device, index, 4096, pfifo);
+	struct gk104_fifo *fifo;
+
+	if (!(fifo = kzalloc(sizeof(*fifo), GFP_KERNEL)))
+		return -ENOMEM;
+	fifo->func = &tu102_fifo;
+	INIT_WORK(&fifo->recover.work, tu102_fifo_recover_work);
+	*pfifo = &fifo->base;
+
+	return nvkm_fifo_ctor(&tu102_fifo_, device, type, inst, 4096, &fifo->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/base.c
index d41fb94524e9..61759f54406e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/base.c
@@ -175,8 +175,8 @@ nvkm_gr = {
 
 int
 nvkm_gr_ctor(const struct nvkm_gr_func *func, struct nvkm_device *device,
-	     int index, bool enable, struct nvkm_gr *gr)
+	     enum nvkm_subdev_type type, int inst, bool enable, struct nvkm_gr *gr)
 {
 	gr->func = func;
-	return nvkm_engine_ctor(&nvkm_gr, device, index, enable, &gr->engine);
+	return nvkm_engine_ctor(&nvkm_gr, device, type, inst, enable, &gr->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/g84.c
index da1ba74682b4..65c332118fd6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/g84.c
@@ -192,7 +192,7 @@ g84_gr = {
 };
 
 int
-g84_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+g84_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv50_gr_new_(&g84_gr, device, index, pgr);
+	return nv50_gr_new_(&g84_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c
index 749f73fc45a8..397ff4fe9df8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c
@@ -2087,8 +2087,8 @@ gf100_gr_flcn = {
 };
 
 int
-gf100_gr_new_(const struct gf100_gr_fwif *fwif,
-	      struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf100_gr_new_(const struct gf100_gr_fwif *fwif, struct nvkm_device *device,
+	      enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
 	struct gf100_gr *gr;
 	int ret;
@@ -2097,7 +2097,7 @@ gf100_gr_new_(const struct gf100_gr_fwif *fwif,
 		return -ENOMEM;
 	*pgr = &gr->base;
 
-	ret = nvkm_gr_ctor(&gf100_gr_, device, index, true, &gr->base);
+	ret = nvkm_gr_ctor(&gf100_gr_, device, type, inst, true, &gr->base);
 	if (ret)
 		return ret;
 
@@ -2483,7 +2483,7 @@ gf100_gr_fwif[] = {
 };
 
 int
-gf100_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf100_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gf100_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gf100_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h
index dfd5dd74f0d5..c0038f906135 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h
@@ -416,6 +416,6 @@ void gm20b_gr_acr_bld_patch(struct nvkm_acr *, u32, s64);
 extern const struct nvkm_acr_lsf_func gp108_gr_gpccs_acr;
 extern const struct nvkm_acr_lsf_func gp108_gr_fecs_acr;
 
-int gf100_gr_new_(const struct gf100_gr_fwif *, struct nvkm_device *, int,
+int gf100_gr_new_(const struct gf100_gr_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  struct nvkm_gr **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf104.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf104.c
index 0536fe8b2b92..3acd99c306f2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf104.c
@@ -152,7 +152,7 @@ gf104_gr_fwif[] = {
 };
 
 int
-gf104_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf104_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gf104_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gf104_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf108.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf108.c
index 14284b06112f..030640bb3dca 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf108.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf108.c
@@ -151,7 +151,7 @@ gf108_gr_fwif[] = {
 };
 
 int
-gf108_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf108_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gf108_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gf108_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf110.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf110.c
index 280752551a3a..616e2def1865 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf110.c
@@ -127,7 +127,7 @@ gf110_gr_fwif[] = {
 };
 
 int
-gf110_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf110_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gf110_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gf110_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf117.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf117.c
index 235c3fbe4b95..669e7536970e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf117.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf117.c
@@ -192,7 +192,7 @@ gf117_gr_fwif[] = {
 };
 
 int
-gf117_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf117_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gf117_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gf117_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf119.c
index 7eac385ece97..5b09bda8110c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf119.c
@@ -218,7 +218,7 @@ gf119_gr_fwif[] = {
 };
 
 int
-gf119_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gf119_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gf119_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gf119_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk104.c
index 89f51d76082b..b680eaa0f350 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk104.c
@@ -497,7 +497,7 @@ gk104_gr_fwif[] = {
 };
 
 int
-gk104_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gk104_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gk104_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gk104_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110.c
index 735f05e54d62..103e06a77e65 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110.c
@@ -393,7 +393,7 @@ gk110_gr_fwif[] = {
 };
 
 int
-gk110_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gk110_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gk110_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gk110_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110b.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110b.c
index adc971be8f3b..034d0b11a17d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk110b.c
@@ -144,7 +144,8 @@ gk110b_gr_fwif[] = {
 };
 
 int
-gk110b_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gk110b_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gk110b_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gk110b_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk208.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk208.c
index aa0eff6795ac..116d682f9f96 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk208.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk208.c
@@ -202,7 +202,7 @@ gk208_gr_fwif[] = {
 };
 
 int
-gk208_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gk208_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gk208_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gk208_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
index 6d4d72851610..be0b2cefd8e8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
@@ -357,7 +357,7 @@ gk20a_gr_fwif[] = {
 };
 
 int
-gk20a_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gk20a_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gk20a_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gk20a_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm107.c
index 09bb78ba9d00..310987174cb5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm107.c
@@ -437,7 +437,7 @@ gm107_gr_fwif[] = {
 };
 
 int
-gm107_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gm107_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gm107_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gm107_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c
index 815137047518..5c38ff0fe7f9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c
@@ -288,7 +288,7 @@ gm200_gr_fwif[] = {
 };
 
 int
-gm200_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gm200_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gm200_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gm200_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm20b.c
index 1aab691fa71c..ec1c46e47e00 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm20b.c
@@ -181,7 +181,7 @@ gm20b_gr_fwif[] = {
 };
 
 int
-gm20b_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gm20b_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gm20b_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gm20b_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c
index ddba7ce937c7..0550dd6f46f1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c
@@ -156,7 +156,7 @@ gp100_gr_fwif[] = {
 };
 
 int
-gp100_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gp100_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gp100_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gp100_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp102.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp102.c
index c083f3757ff7..5b001f374be0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp102.c
@@ -152,7 +152,7 @@ gp102_gr_fwif[] = {
 };
 
 int
-gp102_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gp102_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gp102_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gp102_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp104.c
index f6a31e9a8cc8..2655574ec63b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp104.c
@@ -93,7 +93,7 @@ gp104_gr_fwif[] = {
 };
 
 int
-gp104_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gp104_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gp104_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gp104_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp107.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp107.c
index 2c80c6a75b56..adabc04d4f3a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp107.c
@@ -82,7 +82,7 @@ gp107_gr_fwif[] = {
 };
 
 int
-gp107_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gp107_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gp107_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gp107_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp108.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp108.c
index 2be8f416dd6f..7310f0466bb7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp108.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp108.c
@@ -92,7 +92,7 @@ gp108_gr_fwif[] = {
 };
 
 int
-gp108_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gp108_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gp108_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gp108_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp10b.c
index 6edc4bc7ed44..e13683b6e7b1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp10b.c
@@ -94,7 +94,7 @@ gp10b_gr_fwif[] = {
 };
 
 int
-gp10b_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gp10b_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gp10b_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gp10b_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt200.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt200.c
index c711a55ce392..1dfc65d45b52 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt200.c
@@ -43,7 +43,7 @@ gt200_gr = {
 };
 
 int
-gt200_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gt200_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv50_gr_new_(&gt200_gr, device, index, pgr);
+	return nv50_gr_new_(&gt200_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt215.c
index fa103df32ec7..fcb5ead345a3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gt215.c
@@ -44,7 +44,7 @@ gt215_gr = {
 };
 
 int
-gt215_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gt215_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv50_gr_new_(&gt215_gr, device, index, pgr);
+	return nv50_gr_new_(&gt215_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gv100.c
index 2189a8f4e644..4d043c1173ea 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gv100.c
@@ -141,7 +141,7 @@ gv100_gr_fwif[] = {
 };
 
 int
-gv100_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+gv100_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(gv100_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(gv100_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp79.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp79.c
index eb1a90644752..cf782b64f62e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp79.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp79.c
@@ -42,7 +42,7 @@ mcp79_gr = {
 };
 
 int
-mcp79_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+mcp79_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv50_gr_new_(&mcp79_gr, device, index, pgr);
+	return nv50_gr_new_(&mcp79_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp89.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp89.c
index c91eb56e9327..6f90a6395453 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp89.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/mcp89.c
@@ -44,7 +44,7 @@ mcp89_gr = {
 };
 
 int
-mcp89_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+mcp89_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv50_gr_new_(&mcp89_gr, device, index, pgr);
+	return nv50_gr_new_(&mcp89_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv04.c
index 9c2e985dc079..0bc1a238de43 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv04.c
@@ -1413,7 +1413,7 @@ nv04_gr = {
 };
 
 int
-nv04_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv04_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
 	struct nv04_gr *gr;
 
@@ -1422,5 +1422,5 @@ nv04_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
 	spin_lock_init(&gr->lock);
 	*pgr = &gr->base;
 
-	return nvkm_gr_ctor(&nv04_gr, device, index, true, &gr->base);
+	return nvkm_gr_ctor(&nv04_gr, device, type, inst, true, &gr->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.c
index 4ebbfbdd8240..942450b33bc6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.c
@@ -1173,7 +1173,7 @@ nv10_gr_init(struct nvkm_gr *base)
 
 int
 nv10_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_gr **pgr)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
 	struct nv10_gr *gr;
 
@@ -1182,7 +1182,7 @@ nv10_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
 	spin_lock_init(&gr->lock);
 	*pgr = &gr->base;
 
-	return nvkm_gr_ctor(func, device, index, true, &gr->base);
+	return nvkm_gr_ctor(func, device, type, inst, true, &gr->base);
 }
 
 static const struct nvkm_gr_func
@@ -1215,7 +1215,7 @@ nv10_gr = {
 };
 
 int
-nv10_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv10_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv10_gr_new_(&nv10_gr, device, index, pgr);
+	return nv10_gr_new_(&nv10_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.h
index 4327baea02af..5cfe927c9123 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv10.h
@@ -3,7 +3,7 @@
 #define __NV10_GR_H__
 #include "priv.h"
 
-int nv10_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, int index,
+int nv10_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		 struct nvkm_gr **);
 int nv10_gr_init(struct nvkm_gr *);
 void nv10_gr_intr(struct nvkm_gr *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv15.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv15.c
index 3e2c6856b4c4..69ece259df86 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv15.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv15.c
@@ -53,7 +53,7 @@ nv15_gr = {
 };
 
 int
-nv15_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv15_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv10_gr_new_(&nv15_gr, device, index, pgr);
+	return nv10_gr_new_(&nv15_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv17.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv17.c
index 12437d085a73..e39dfc7d4077 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv17.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv17.c
@@ -53,7 +53,7 @@ nv17_gr = {
 };
 
 int
-nv17_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv17_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv10_gr_new_(&nv17_gr, device, index, pgr);
+	return nv10_gr_new_(&nv17_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c
index d837630a3625..6bff10cee71b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c
@@ -330,7 +330,7 @@ nv20_gr_dtor(struct nvkm_gr *base)
 
 int
 nv20_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_gr **pgr)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
 	struct nv20_gr *gr;
 
@@ -338,7 +338,7 @@ nv20_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
 		return -ENOMEM;
 	*pgr = &gr->base;
 
-	return nvkm_gr_ctor(func, device, index, true, &gr->base);
+	return nvkm_gr_ctor(func, device, type, inst, true, &gr->base);
 }
 
 static const struct nvkm_gr_func
@@ -370,7 +370,7 @@ nv20_gr = {
 };
 
 int
-nv20_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv20_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv20_gr_new_(&nv20_gr, device, index, pgr);
+	return nv20_gr_new_(&nv20_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.h
index e57407a8a7c3..c0d2be53413e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.h
@@ -9,8 +9,8 @@ struct nv20_gr {
 	struct nvkm_memory *ctxtab;
 };
 
-int nv20_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *,
-		 int, struct nvkm_gr **);
+int nv20_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		 struct nvkm_gr **);
 void *nv20_gr_dtor(struct nvkm_gr *);
 int nv20_gr_oneinit(struct nvkm_gr *);
 int nv20_gr_init(struct nvkm_gr *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv25.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv25.c
index 32d29d3faee0..f3a56f17d94a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv25.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv25.c
@@ -129,7 +129,7 @@ nv25_gr = {
 };
 
 int
-nv25_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv25_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv20_gr_new_(&nv25_gr, device, index, pgr);
+	return nv20_gr_new_(&nv25_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv2a.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv2a.c
index f941062c66f0..f268d2642d29 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv2a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv2a.c
@@ -120,7 +120,7 @@ nv2a_gr = {
 };
 
 int
-nv2a_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv2a_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv20_gr_new_(&nv2a_gr, device, index, pgr);
+	return nv20_gr_new_(&nv2a_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv30.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv30.c
index 785ec956df0f..e5737cdf2fa1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv30.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv30.c
@@ -194,7 +194,7 @@ nv30_gr = {
 };
 
 int
-nv30_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv30_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv20_gr_new_(&nv30_gr, device, index, pgr);
+	return nv20_gr_new_(&nv30_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv34.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv34.c
index bd610d75c677..1ab2da8ebf4e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv34.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv34.c
@@ -131,7 +131,7 @@ nv34_gr = {
 };
 
 int
-nv34_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv34_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv20_gr_new_(&nv34_gr, device, index, pgr);
+	return nv20_gr_new_(&nv34_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv35.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv35.c
index 89db7f523037..591260f5676b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv35.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv35.c
@@ -131,7 +131,7 @@ nv35_gr = {
 };
 
 int
-nv35_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv35_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv20_gr_new_(&nv35_gr, device, index, pgr);
+	return nv20_gr_new_(&nv35_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c
index 5f1ad8344ea9..67f3535ff97e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c
@@ -429,7 +429,7 @@ nv40_gr_init(struct nvkm_gr *base)
 
 int
 nv40_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_gr **pgr)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
 	struct nv40_gr *gr;
 
@@ -438,7 +438,7 @@ nv40_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
 	*pgr = &gr->base;
 	INIT_LIST_HEAD(&gr->chan);
 
-	return nvkm_gr_ctor(func, device, index, true, &gr->base);
+	return nvkm_gr_ctor(func, device, type, inst, true, &gr->base);
 }
 
 static const struct nvkm_gr_func
@@ -470,7 +470,7 @@ nv40_gr = {
 };
 
 int
-nv40_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv40_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv40_gr_new_(&nv40_gr, device, index, pgr);
+	return nv40_gr_new_(&nv40_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.h
index e6128791b2d2..f3d3d3a5ae5b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.h
@@ -10,7 +10,7 @@ struct nv40_gr {
 	struct list_head chan;
 };
 
-int nv40_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, int index,
+int nv40_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		 struct nvkm_gr **);
 int nv40_gr_init(struct nvkm_gr *);
 void nv40_gr_intr(struct nvkm_gr *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv44.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv44.c
index 45ff80254eb4..22b6a38a7031 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv44.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv44.c
@@ -102,7 +102,7 @@ nv44_gr = {
 };
 
 int
-nv44_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv44_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv40_gr_new_(&nv44_gr, device, index, pgr);
+	return nv40_gr_new_(&nv44_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.c
index df16ffda1749..563a10097e95 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.c
@@ -761,7 +761,7 @@ nv50_gr_init(struct nvkm_gr *base)
 
 int
 nv50_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_gr **pgr)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
 	struct nv50_gr *gr;
 
@@ -770,7 +770,7 @@ nv50_gr_new_(const struct nvkm_gr_func *func, struct nvkm_device *device,
 	spin_lock_init(&gr->lock);
 	*pgr = &gr->base;
 
-	return nvkm_gr_ctor(func, device, index, true, &gr->base);
+	return nvkm_gr_ctor(func, device, type, inst, true, &gr->base);
 }
 
 static const struct nvkm_gr_func
@@ -790,7 +790,7 @@ nv50_gr = {
 };
 
 int
-nv50_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+nv50_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return nv50_gr_new_(&nv50_gr, device, index, pgr);
+	return nv50_gr_new_(&nv50_gr, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.h
index 465f4da0ddfc..84388c42e5c6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv50.h
@@ -11,7 +11,7 @@ struct nv50_gr {
 	u32 size;
 };
 
-int nv50_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, int index,
+int nv50_gr_new_(const struct nvkm_gr_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		 struct nvkm_gr **);
 int nv50_gr_init(struct nvkm_gr *);
 void nv50_gr_intr(struct nvkm_gr *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/priv.h
index 3b30f24032cc..9b2c66e8be90 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/priv.h
@@ -7,8 +7,8 @@
 struct nvkm_fb_tile;
 struct nvkm_fifo_chan;
 
-int nvkm_gr_ctor(const struct nvkm_gr_func *, struct nvkm_device *,
-		 int index, bool enable, struct nvkm_gr *);
+int nvkm_gr_ctor(const struct nvkm_gr_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		 bool enable, struct nvkm_gr *);
 
 bool nv04_gr_idle(struct nvkm_gr *);
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c
index 6039f9948aa2..1a8a21844e12 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c
@@ -198,7 +198,7 @@ tu102_gr_fwif[] = {
 };
 
 int
-tu102_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+tu102_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr)
 {
-	return gf100_gr_new_(tu102_gr_fwif, device, index, pgr);
+	return gf100_gr_new_(tu102_gr_fwif, device, type, inst, pgr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/g84.c
index c0e11a071843..0fcc0ffa1e40 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/g84.c
@@ -37,7 +37,8 @@ g84_mpeg = {
 };
 
 int
-g84_mpeg_new(struct nvkm_device *device, int index, struct nvkm_engine **pmpeg)
+g84_mpeg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_engine **pmpeg)
 {
-	return nvkm_engine_new_(&g84_mpeg, device, index, true, pmpeg);
+	return nvkm_engine_new_(&g84_mpeg, device, type, inst, true, pmpeg);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.c b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.c
index 7fea7d45202f..b1054db4c1b8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.c
@@ -274,7 +274,7 @@ nv31_mpeg_ = {
 
 int
 nv31_mpeg_new_(const struct nv31_mpeg_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_engine **pmpeg)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_engine **pmpeg)
 {
 	struct nv31_mpeg *mpeg;
 
@@ -283,8 +283,7 @@ nv31_mpeg_new_(const struct nv31_mpeg_func *func, struct nvkm_device *device,
 	mpeg->func = func;
 	*pmpeg = &mpeg->engine;
 
-	return nvkm_engine_ctor(&nv31_mpeg_, device, index,
-				true, &mpeg->engine);
+	return nvkm_engine_ctor(&nv31_mpeg_, device, type, inst, true, &mpeg->engine);
 }
 
 static const struct nv31_mpeg_func
@@ -293,7 +292,8 @@ nv31_mpeg = {
 };
 
 int
-nv31_mpeg_new(struct nvkm_device *device, int index, struct nvkm_engine **pmpeg)
+nv31_mpeg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_engine **pmpeg)
 {
-	return nv31_mpeg_new_(&nv31_mpeg, device, index, pmpeg);
+	return nv31_mpeg_new_(&nv31_mpeg, device, type, inst, pmpeg);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.h b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.h
index b3e131538858..9f30aaaf809e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv31.h
@@ -11,8 +11,8 @@ struct nv31_mpeg {
 	struct nv31_mpeg_chan *chan;
 };
 
-int nv31_mpeg_new_(const struct nv31_mpeg_func *, struct nvkm_device *,
-		   int index, struct nvkm_engine **);
+int nv31_mpeg_new_(const struct nv31_mpeg_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_engine **);
 
 struct nv31_mpeg_func {
 	bool (*mthd_dma)(struct nvkm_device *, u32 mthd, u32 data);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv40.c
index b5ec7c504dc6..179167484ef1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv40.c
@@ -76,7 +76,8 @@ nv40_mpeg = {
 };
 
 int
-nv40_mpeg_new(struct nvkm_device *device, int index, struct nvkm_engine **pmpeg)
+nv40_mpeg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_engine **pmpeg)
 {
-	return nv31_mpeg_new_(&nv40_mpeg, device, index, pmpeg);
+	return nv31_mpeg_new_(&nv40_mpeg, device, type, inst, pmpeg);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv44.c b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv44.c
index c3cf02ed468e..521ce43a2871 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv44.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv44.c
@@ -203,7 +203,8 @@ nv44_mpeg = {
 };
 
 int
-nv44_mpeg_new(struct nvkm_device *device, int index, struct nvkm_engine **pmpeg)
+nv44_mpeg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_engine **pmpeg)
 {
 	struct nv44_mpeg *mpeg;
 
@@ -212,5 +213,5 @@ nv44_mpeg_new(struct nvkm_device *device, int index, struct nvkm_engine **pmpeg)
 	INIT_LIST_HEAD(&mpeg->chan);
 	*pmpeg = &mpeg->engine;
 
-	return nvkm_engine_ctor(&nv44_mpeg, device, index, true, &mpeg->engine);
+	return nvkm_engine_ctor(&nv44_mpeg, device, type, inst, true, &mpeg->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv50.c
index 6df880a39019..e6374f36961c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mpeg/nv50.c
@@ -129,7 +129,8 @@ nv50_mpeg = {
 };
 
 int
-nv50_mpeg_new(struct nvkm_device *device, int index, struct nvkm_engine **pmpeg)
+nv50_mpeg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_engine **pmpeg)
 {
-	return nvkm_engine_new_(&nv50_mpeg, device, index, true, pmpeg);
+	return nvkm_engine_new_(&nv50_mpeg, device, type, inst, true, pmpeg);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/base.c
index 80211f76093b..842fcfbd28b8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/base.c
@@ -24,9 +24,8 @@
 #include "priv.h"
 
 int
-nvkm_mspdec_new_(const struct nvkm_falcon_func *func,
-		 struct nvkm_device *device, int index,
-		 struct nvkm_engine **pengine)
+nvkm_mspdec_new_(const struct nvkm_falcon_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine)
 {
-	return nvkm_falcon_new_(func, device, index, true, 0x085000, pengine);
+	return nvkm_falcon_new_(func, device, type, inst, true, 0x085000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/g98.c b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/g98.c
index f30cf1dcfb30..ecb06d68f544 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/g98.c
@@ -43,8 +43,8 @@ g98_mspdec = {
 };
 
 int
-g98_mspdec_new(struct nvkm_device *device, int index,
-	     struct nvkm_engine **pengine)
+g98_mspdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_engine **pengine)
 {
-	return nvkm_mspdec_new_(&g98_mspdec, device, index, pengine);
+	return nvkm_mspdec_new_(&g98_mspdec, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gf100.c
index cfe1aa81bd14..0a69bd767d69 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gf100.c
@@ -43,8 +43,8 @@ gf100_mspdec = {
 };
 
 int
-gf100_mspdec_new(struct nvkm_device *device, int index,
+gf100_mspdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_engine **pengine)
 {
-	return nvkm_mspdec_new_(&gf100_mspdec, device, index, pengine);
+	return nvkm_mspdec_new_(&gf100_mspdec, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gk104.c
index 24272b4927bc..a08991dca428 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gk104.c
@@ -35,8 +35,8 @@ gk104_mspdec = {
 };
 
 int
-gk104_mspdec_new(struct nvkm_device *device, int index,
+gk104_mspdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_engine **pengine)
 {
-	return nvkm_mspdec_new_(&gk104_mspdec, device, index, pengine);
+	return nvkm_mspdec_new_(&gk104_mspdec, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gt215.c
index cf6e59ad6ee2..791fb03a32ad 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/gt215.c
@@ -35,8 +35,8 @@ gt215_mspdec = {
 };
 
 int
-gt215_mspdec_new(struct nvkm_device *device, int index,
-	     struct nvkm_engine **pengine)
+gt215_mspdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		 struct nvkm_engine **pengine)
 {
-	return nvkm_mspdec_new_(&gt215_mspdec, device, index, pengine);
+	return nvkm_mspdec_new_(&gt215_mspdec, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/priv.h
index 86445a2600d0..2bc5537d40a3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/mspdec/priv.h
@@ -3,8 +3,8 @@
 #define __NVKM_MSPDEC_PRIV_H__
 #include <engine/mspdec.h>
 
-int nvkm_mspdec_new_(const struct nvkm_falcon_func *, struct nvkm_device *,
-		     int index, struct nvkm_engine **);
+int nvkm_mspdec_new_(const struct nvkm_falcon_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		     int, struct nvkm_engine **);
 
 void g98_mspdec_init(struct nvkm_falcon *);
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/base.c
index bfae5e60e925..45a9411ab2e2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/base.c
@@ -25,7 +25,7 @@
 
 int
 nvkm_msppp_new_(const struct nvkm_falcon_func *func, struct nvkm_device *device,
-		int index, struct nvkm_engine **pengine)
+		enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine)
 {
-	return nvkm_falcon_new_(func, device, index, true, 0x086000, pengine);
+	return nvkm_falcon_new_(func, device, type, inst, true, 0x086000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/g98.c b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/g98.c
index c45dbf79d1f9..160120b9bd64 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/g98.c
@@ -43,8 +43,8 @@ g98_msppp = {
 };
 
 int
-g98_msppp_new(struct nvkm_device *device, int index,
+g98_msppp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	      struct nvkm_engine **pengine)
 {
-	return nvkm_msppp_new_(&g98_msppp, device, index, pengine);
+	return nvkm_msppp_new_(&g98_msppp, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gf100.c
index 803c62ab516e..debed9ae8731 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gf100.c
@@ -43,8 +43,8 @@ gf100_msppp = {
 };
 
 int
-gf100_msppp_new(struct nvkm_device *device, int index,
+gf100_msppp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_engine **pengine)
 {
-	return nvkm_msppp_new_(&gf100_msppp, device, index, pengine);
+	return nvkm_msppp_new_(&gf100_msppp, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gt215.c
index 49cbf72cee4b..a2fd736fef94 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/gt215.c
@@ -35,8 +35,8 @@ gt215_msppp = {
 };
 
 int
-gt215_msppp_new(struct nvkm_device *device, int index,
-	      struct nvkm_engine **pengine)
+gt215_msppp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		struct nvkm_engine **pengine)
 {
-	return nvkm_msppp_new_(&gt215_msppp, device, index, pengine);
+	return nvkm_msppp_new_(&gt215_msppp, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/priv.h
index f20b10915db2..582ab8ce1425 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msppp/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msppp/priv.h
@@ -3,8 +3,8 @@
 #define __NVKM_MSPPP_PRIV_H__
 #include <engine/msppp.h>
 
-int nvkm_msppp_new_(const struct nvkm_falcon_func *, struct nvkm_device *,
-		    int index, struct nvkm_engine **);
+int nvkm_msppp_new_(const struct nvkm_falcon_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		    int, struct nvkm_engine **);
 
 void g98_msppp_init(struct nvkm_falcon *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/base.c
index 745bbb653dc0..7be42b980e57 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/base.c
@@ -25,7 +25,7 @@
 
 int
 nvkm_msvld_new_(const struct nvkm_falcon_func *func, struct nvkm_device *device,
-		int index, struct nvkm_engine **pengine)
+		enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine)
 {
-	return nvkm_falcon_new_(func, device, index, true, 0x084000, pengine);
+	return nvkm_falcon_new_(func, device, type, inst, true, 0x084000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/g98.c b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/g98.c
index 4a2a9f0494af..cfa2065319a6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/g98.c
@@ -43,8 +43,8 @@ g98_msvld = {
 };
 
 int
-g98_msvld_new(struct nvkm_device *device, int index,
+g98_msvld_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	      struct nvkm_engine **pengine)
 {
-	return nvkm_msvld_new_(&g98_msvld, device, index, pengine);
+	return nvkm_msvld_new_(&g98_msvld, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gf100.c
index 1695e532c081..8d58ad8e04d3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gf100.c
@@ -43,8 +43,8 @@ gf100_msvld = {
 };
 
 int
-gf100_msvld_new(struct nvkm_device *device, int index,
+gf100_msvld_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_engine **pengine)
 {
-	return nvkm_msvld_new_(&gf100_msvld, device, index, pengine);
+	return nvkm_msvld_new_(&gf100_msvld, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gk104.c
index b640cd63ebe8..b28be28046f1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gk104.c
@@ -35,8 +35,8 @@ gk104_msvld = {
 };
 
 int
-gk104_msvld_new(struct nvkm_device *device, int index,
+gk104_msvld_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_engine **pengine)
 {
-	return nvkm_msvld_new_(&gk104_msvld, device, index, pengine);
+	return nvkm_msvld_new_(&gk104_msvld, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gt215.c
index 201e8ef3519e..d7489f972c99 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/gt215.c
@@ -35,8 +35,8 @@ gt215_msvld = {
 };
 
 int
-gt215_msvld_new(struct nvkm_device *device, int index,
-	      struct nvkm_engine **pengine)
+gt215_msvld_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		struct nvkm_engine **pengine)
 {
-	return nvkm_msvld_new_(&gt215_msvld, device, index, pengine);
+	return nvkm_msvld_new_(&gt215_msvld, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/mcp89.c b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/mcp89.c
index a0f540ef257b..16c30b62ab09 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/mcp89.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/mcp89.c
@@ -35,8 +35,8 @@ mcp89_msvld = {
 };
 
 int
-mcp89_msvld_new(struct nvkm_device *device, int index,
-	      struct nvkm_engine **pengine)
+mcp89_msvld_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		struct nvkm_engine **pengine)
 {
-	return nvkm_msvld_new_(&mcp89_msvld, device, index, pengine);
+	return nvkm_msvld_new_(&mcp89_msvld, device, type, inst, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/priv.h
index 5cd1e83badbb..f729d919b054 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/msvld/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/msvld/priv.h
@@ -3,8 +3,8 @@
 #define __NVKM_MSVLD_PRIV_H__
 #include <engine/msvld.h>
 
-int nvkm_msvld_new_(const struct nvkm_falcon_func *, struct nvkm_device *,
-		    int index, struct nvkm_engine **);
+int nvkm_msvld_new_(const struct nvkm_falcon_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		    int, struct nvkm_engine **);
 
 void g98_msvld_init(struct nvkm_falcon *);
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/base.c
index 9b23c1b70ebf..b0181cc5953b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/base.c
@@ -37,7 +37,7 @@ nvkm_nvdec = {
 
 int
 nvkm_nvdec_new_(const struct nvkm_nvdec_fwif *fwif, struct nvkm_device *device,
-		int index, struct nvkm_nvdec **pnvdec)
+		enum nvkm_subdev_type type, int inst, struct nvkm_nvdec **pnvdec)
 {
 	struct nvkm_nvdec *nvdec;
 	int ret;
@@ -45,7 +45,7 @@ nvkm_nvdec_new_(const struct nvkm_nvdec_fwif *fwif, struct nvkm_device *device,
 	if (!(nvdec = *pnvdec = kzalloc(sizeof(*nvdec), GFP_KERNEL)))
 		return -ENOMEM;
 
-	ret = nvkm_engine_ctor(&nvkm_nvdec, device, index, true,
+	ret = nvkm_engine_ctor(&nvkm_nvdec, device, type, inst, true,
 			       &nvdec->engine);
 	if (ret)
 		return ret;
@@ -57,5 +57,5 @@ nvkm_nvdec_new_(const struct nvkm_nvdec_fwif *fwif, struct nvkm_device *device,
 	nvdec->func = fwif->func;
 
 	return nvkm_falcon_ctor(nvdec->func->flcn, &nvdec->engine.subdev,
-				nvkm_subdev_name[index], 0, &nvdec->falcon);
+				nvdec->engine.subdev.name, 0, &nvdec->falcon);
 };
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/gm107.c
index 0ab27ab4d8ee..8c44ce44a6d7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/gm107.c
@@ -56,8 +56,8 @@ gm107_nvdec_fwif[] = {
 };
 
 int
-gm107_nvdec_new(struct nvkm_device *device, int index,
+gm107_nvdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_nvdec **pnvdec)
 {
-	return nvkm_nvdec_new_(gm107_nvdec_fwif, device, index, pnvdec);
+	return nvkm_nvdec_new_(gm107_nvdec_fwif, device, type, inst, pnvdec);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h
index e14da8b000d0..0920f6a887e2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h
@@ -14,6 +14,6 @@ struct nvkm_nvdec_fwif {
 	const struct nvkm_nvdec_func *func;
 };
 
-int nvkm_nvdec_new_(const struct nvkm_nvdec_fwif *fwif,
-		    struct nvkm_device *, int, struct nvkm_nvdec **);
+int nvkm_nvdec_new_(const struct nvkm_nvdec_fwif *fwif, struct nvkm_device *,
+		    enum nvkm_subdev_type, int, struct nvkm_nvdec **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c
index 484100e15668..c39e797dc7c9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c
@@ -39,7 +39,7 @@ nvkm_nvenc = {
 
 int
 nvkm_nvenc_new_(const struct nvkm_nvenc_fwif *fwif, struct nvkm_device *device,
-		int index, struct nvkm_nvenc **pnvenc)
+		enum nvkm_subdev_type type, int inst, struct nvkm_nvenc **pnvenc)
 {
 	struct nvkm_nvenc *nvenc;
 	int ret;
@@ -47,7 +47,7 @@ nvkm_nvenc_new_(const struct nvkm_nvenc_fwif *fwif, struct nvkm_device *device,
 	if (!(nvenc = *pnvenc = kzalloc(sizeof(*nvenc), GFP_KERNEL)))
 		return -ENOMEM;
 
-	ret = nvkm_engine_ctor(&nvkm_nvenc, device, index, true,
+	ret = nvkm_engine_ctor(&nvkm_nvenc, device, type, inst, true,
 			       &nvenc->engine);
 	if (ret)
 		return ret;
@@ -59,5 +59,5 @@ nvkm_nvenc_new_(const struct nvkm_nvenc_fwif *fwif, struct nvkm_device *device,
 	nvenc->func = fwif->func;
 
 	return nvkm_falcon_ctor(nvenc->func->flcn, &nvenc->engine.subdev,
-				nvkm_subdev_name[index], 0, &nvenc->falcon);
+				nvenc->engine.subdev.name, 0, &nvenc->falcon);
 };
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/gm107.c
index d249c8ffb2d5..f44d41bf2034 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/gm107.c
@@ -56,8 +56,8 @@ gm107_nvenc_fwif[] = {
 };
 
 int
-gm107_nvenc_new(struct nvkm_device *device, int index,
+gm107_nvenc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_nvenc **pnvenc)
 {
-	return nvkm_nvenc_new_(gm107_nvenc_fwif, device, index, pnvenc);
+	return nvkm_nvenc_new_(gm107_nvenc_fwif, device, type, inst, pnvenc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h
index 100fa5ebbeef..4130a2bfbb4f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h
@@ -14,6 +14,6 @@ struct nvkm_nvenc_fwif {
 	const struct nvkm_nvenc_func *func;
 };
 
-int nvkm_nvenc_new_(const struct nvkm_nvenc_fwif *, struct nvkm_device *,
+int nvkm_nvenc_new_(const struct nvkm_nvenc_fwif *, struct nvkm_device *, enum nvkm_subdev_type,
 		    int, struct nvkm_nvenc **pnvenc);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/base.c
index b2785bee418e..8fe0444f761e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/base.c
@@ -628,10 +628,10 @@ nvkm_perfmon_dtor(struct nvkm_object *object)
 {
 	struct nvkm_perfmon *perfmon = nvkm_perfmon(object);
 	struct nvkm_pm *pm = perfmon->pm;
-	mutex_lock(&pm->engine.subdev.mutex);
-	if (pm->perfmon == &perfmon->object)
-		pm->perfmon = NULL;
-	mutex_unlock(&pm->engine.subdev.mutex);
+	spin_lock(&pm->client.lock);
+	if (pm->client.object == &perfmon->object)
+		pm->client.object = NULL;
+	spin_unlock(&pm->client.lock);
 	return perfmon;
 }
 
@@ -671,11 +671,11 @@ nvkm_pm_oclass_new(struct nvkm_device *device, const struct nvkm_oclass *oclass,
 	if (ret)
 		return ret;
 
-	mutex_lock(&pm->engine.subdev.mutex);
-	if (pm->perfmon == NULL)
-		pm->perfmon = *pobject;
-	ret = (pm->perfmon == *pobject) ? 0 : -EBUSY;
-	mutex_unlock(&pm->engine.subdev.mutex);
+	spin_lock(&pm->client.lock);
+	if (pm->client.object == NULL)
+		pm->client.object = *pobject;
+	ret = (pm->client.object == *pobject) ? 0 : -EBUSY;
+	spin_unlock(&pm->client.lock);
 	return ret;
 }
 
@@ -858,10 +858,11 @@ nvkm_pm = {
 
 int
 nvkm_pm_ctor(const struct nvkm_pm_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_pm *pm)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_pm *pm)
 {
 	pm->func = func;
 	INIT_LIST_HEAD(&pm->domains);
 	INIT_LIST_HEAD(&pm->sources);
-	return nvkm_engine_ctor(&nvkm_pm, device, index, true, &pm->engine);
+	spin_lock_init(&pm->client.lock);
+	return nvkm_engine_ctor(&nvkm_pm, device, type, inst, true, &pm->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/g84.c
index 6e441ddafd86..0086d00eb162 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/g84.c
@@ -159,7 +159,7 @@ g84_pm[] = {
 };
 
 int
-g84_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+g84_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return nv40_pm_new_(g84_pm, device, index, ppm);
+	return nv40_pm_new_(g84_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.c
index fe2532ee4145..8e02701def8e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.c
@@ -187,7 +187,7 @@ gf100_pm_ = {
 
 int
 gf100_pm_new_(const struct gf100_pm_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_pm **ppm)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
 	struct nvkm_pm *pm;
 	u32 mask;
@@ -196,7 +196,7 @@ gf100_pm_new_(const struct gf100_pm_func *func, struct nvkm_device *device,
 	if (!(pm = *ppm = kzalloc(sizeof(*pm), GFP_KERNEL)))
 		return -ENOMEM;
 
-	ret = nvkm_pm_ctor(&gf100_pm_, device, index, pm);
+	ret = nvkm_pm_ctor(&gf100_pm_, device, type, inst, pm);
 	if (ret)
 		return ret;
 
@@ -237,7 +237,7 @@ gf100_pm = {
 };
 
 int
-gf100_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+gf100_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return gf100_pm_new_(&gf100_pm, device, index, ppm);
+	return gf100_pm_new_(&gf100_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.h
index 461bb219b1c0..bc4b014c4e8e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf100.h
@@ -9,8 +9,8 @@ struct gf100_pm_func {
 	const struct nvkm_specdom *doms_part;
 };
 
-int gf100_pm_new_(const struct gf100_pm_func *, struct nvkm_device *,
-		  int index, struct nvkm_pm **);
+int gf100_pm_new_(const struct gf100_pm_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_pm **);
 
 extern const struct nvkm_funcdom gf100_perfctr_func;
 extern const struct nvkm_specdom gf100_pm_gpc[];
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf108.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf108.c
index 49b24c98a7f7..505565866b59 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf108.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf108.c
@@ -60,7 +60,7 @@ gf108_pm = {
 };
 
 int
-gf108_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+gf108_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return gf100_pm_new_(&gf108_pm, device, index, ppm);
+	return gf100_pm_new_(&gf108_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf117.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf117.c
index 9170025fc988..c61e8c010bb3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf117.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gf117.c
@@ -74,7 +74,7 @@ gf117_pm = {
 };
 
 int
-gf117_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+gf117_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return gf100_pm_new_(&gf117_pm, device, index, ppm);
+	return gf100_pm_new_(&gf117_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gk104.c
index 07f946d26ac6..75bf3df1cb18 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gk104.c
@@ -178,7 +178,7 @@ gk104_pm = {
 };
 
 int
-gk104_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+gk104_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return gf100_pm_new_(&gk104_pm, device, index, ppm);
+	return gf100_pm_new_(&gk104_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt200.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt200.c
index 5cf5dd536fd0..25874c541486 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt200.c
@@ -151,7 +151,7 @@ gt200_pm[] = {
 };
 
 int
-gt200_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+gt200_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return nv40_pm_new_(gt200_pm, device, index, ppm);
+	return nv40_pm_new_(gt200_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt215.c
index c9227ad41b04..54c23e2b6645 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/gt215.c
@@ -132,7 +132,7 @@ gt215_pm[] = {
 };
 
 int
-gt215_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+gt215_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return nv40_pm_new_(gt215_pm, device, index, ppm);
+	return nv40_pm_new_(gt215_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.c
index 3fda594700e0..eba5b3b79340 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.c
@@ -80,7 +80,7 @@ nv40_pm_ = {
 
 int
 nv40_pm_new_(const struct nvkm_specdom *doms, struct nvkm_device *device,
-	     int index, struct nvkm_pm **ppm)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
 	struct nv40_pm *pm;
 	int ret;
@@ -89,7 +89,7 @@ nv40_pm_new_(const struct nvkm_specdom *doms, struct nvkm_device *device,
 		return -ENOMEM;
 	*ppm = &pm->base;
 
-	ret = nvkm_pm_ctor(&nv40_pm_, device, index, &pm->base);
+	ret = nvkm_pm_ctor(&nv40_pm_, device, type, inst, &pm->base);
 	if (ret)
 		return ret;
 
@@ -117,7 +117,7 @@ nv40_pm[] = {
 };
 
 int
-nv40_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+nv40_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return nv40_pm_new_(nv40_pm, device, index, ppm);
+	return nv40_pm_new_(nv40_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.h b/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.h
index 8ed19320fda1..afb79843723d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv40.h
@@ -9,7 +9,7 @@ struct nv40_pm {
 	u32 sequence;
 };
 
-int nv40_pm_new_(const struct nvkm_specdom *, struct nvkm_device *,
-		 int index, struct nvkm_pm **);
+int nv40_pm_new_(const struct nvkm_specdom *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		 struct nvkm_pm **);
 extern const struct nvkm_funcdom nv40_perfctr_func;
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv50.c
index cc5a41d4c6f2..bbd3404901f9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/nv50.c
@@ -169,7 +169,7 @@ nv50_pm[] = {
 };
 
 int
-nv50_pm_new(struct nvkm_device *device, int index, struct nvkm_pm **ppm)
+nv50_pm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_pm **ppm)
 {
-	return nv40_pm_new_(nv50_pm, device, index, ppm);
+	return nv40_pm_new_(nv50_pm, device, type, inst, ppm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/pm/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/pm/priv.h
index cd6f8f79b235..6ae25d3e7f45 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/pm/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/pm/priv.h
@@ -4,8 +4,8 @@
 #define nvkm_pm(p) container_of((p), struct nvkm_pm, engine)
 #include <engine/pm.h>
 
-int nvkm_pm_ctor(const struct nvkm_pm_func *, struct nvkm_device *,
-		 int index, struct nvkm_pm *);
+int nvkm_pm_ctor(const struct nvkm_pm_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		 struct nvkm_pm *);
 
 struct nvkm_pm_func {
 	void (*fini)(struct nvkm_pm *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sec/g98.c b/drivers/gpu/drm/nouveau/nvkm/engine/sec/g98.c
index 6d2a7f0afbb5..1b87df03c823 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sec/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sec/g98.c
@@ -74,9 +74,8 @@ g98_sec = {
 };
 
 int
-g98_sec_new(struct nvkm_device *device, int index,
+g98_sec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	    struct nvkm_engine **pengine)
 {
-	return nvkm_falcon_new_(&g98_sec, device, index,
-				true, 0x087000, pengine);
+	return nvkm_falcon_new_(&g98_sec, device, type, inst, true, 0x087000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/base.c
index 41318aa0d481..092c6d0b8e01 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/base.c
@@ -85,7 +85,7 @@ nvkm_sec2 = {
 
 int
 nvkm_sec2_new_(const struct nvkm_sec2_fwif *fwif, struct nvkm_device *device,
-	       int index, u32 addr, struct nvkm_sec2 **psec2)
+	       enum nvkm_subdev_type type, int inst, u32 addr, struct nvkm_sec2 **psec2)
 {
 	struct nvkm_sec2 *sec2;
 	int ret;
@@ -93,7 +93,7 @@ nvkm_sec2_new_(const struct nvkm_sec2_fwif *fwif, struct nvkm_device *device,
 	if (!(sec2 = *psec2 = kzalloc(sizeof(*sec2), GFP_KERNEL)))
 		return -ENOMEM;
 
-	ret = nvkm_engine_ctor(&nvkm_sec2, device, index, true, &sec2->engine);
+	ret = nvkm_engine_ctor(&nvkm_sec2, device, type, inst, true, &sec2->engine);
 	if (ret)
 		return ret;
 
@@ -104,7 +104,7 @@ nvkm_sec2_new_(const struct nvkm_sec2_fwif *fwif, struct nvkm_device *device,
 	sec2->func = fwif->func;
 
 	ret = nvkm_falcon_ctor(sec2->func->flcn, &sec2->engine.subdev,
-			       nvkm_subdev_name[index], addr, &sec2->falcon);
+			       sec2->engine.subdev.name, addr, &sec2->falcon);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp102.c b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp102.c
index bccf7acb7f98..44e39f5743d5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp102.c
@@ -343,7 +343,8 @@ gp102_sec2_fwif[] = {
 };
 
 int
-gp102_sec2_new(struct nvkm_device *device, int index, struct nvkm_sec2 **psec2)
+gp102_sec2_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_sec2 **psec2)
 {
-	return nvkm_sec2_new_(gp102_sec2_fwif, device, index, 0, psec2);
+	return nvkm_sec2_new_(gp102_sec2_fwif, device, type, inst, 0, psec2);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp108.c b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp108.c
index e770c9497871..3e9f5c842f3c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp108.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/gp108.c
@@ -36,7 +36,8 @@ gp108_sec2_fwif[] = {
 };
 
 int
-gp108_sec2_new(struct nvkm_device *device, int index, struct nvkm_sec2 **psec2)
+gp108_sec2_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_sec2 **psec2)
 {
-	return nvkm_sec2_new_(gp108_sec2_fwif, device, index, 0, psec2);
+	return nvkm_sec2_new_(gp108_sec2_fwif, device, type, inst, 0, psec2);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/priv.h
index 8cbc0b7d0a27..af19229e885d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/priv.h
@@ -25,6 +25,6 @@ int gp102_sec2_load(struct nvkm_sec2 *, int, const struct nvkm_sec2_fwif *);
 extern const struct nvkm_sec2_func gp102_sec2;
 extern const struct nvkm_acr_lsf_func gp102_sec2_acr_1;
 
-int nvkm_sec2_new_(const struct nvkm_sec2_fwif *, struct nvkm_device *,
+int nvkm_sec2_new_(const struct nvkm_sec2_fwif *, struct nvkm_device *, enum nvkm_subdev_type,
 		   int, u32 addr, struct nvkm_sec2 **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/tu102.c
index a231c1c6c0a5..f3faeb705575 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sec2/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sec2/tu102.c
@@ -72,10 +72,11 @@ tu102_sec2_fwif[] = {
 };
 
 int
-tu102_sec2_new(struct nvkm_device *device, int index, struct nvkm_sec2 **psec2)
+tu102_sec2_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_sec2 **psec2)
 {
 	/* TOP info wasn't updated on Turing to reflect the PRI
 	 * address change for some reason.  We override it here.
 	 */
-	return nvkm_sec2_new_(tu102_sec2_fwif, device, index, 0x840000, psec2);
+	return nvkm_sec2_new_(tu102_sec2_fwif, device, type, inst, 0x840000, psec2);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sw/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/sw/base.c
index 7be3198e11de..14871d0bd746 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sw/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sw/base.c
@@ -97,7 +97,7 @@ nvkm_sw = {
 
 int
 nvkm_sw_new_(const struct nvkm_sw_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_sw **psw)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_sw **psw)
 {
 	struct nvkm_sw *sw;
 
@@ -106,5 +106,5 @@ nvkm_sw_new_(const struct nvkm_sw_func *func, struct nvkm_device *device,
 	INIT_LIST_HEAD(&sw->chan);
 	sw->func = func;
 
-	return nvkm_engine_ctor(&nvkm_sw, device, index, true, &sw->engine);
+	return nvkm_engine_ctor(&nvkm_sw, device, type, inst, true, &sw->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sw/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/sw/gf100.c
index ea8f4247b628..55abf839f29d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sw/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sw/gf100.c
@@ -149,7 +149,7 @@ gf100_sw = {
 };
 
 int
-gf100_sw_new(struct nvkm_device *device, int index, struct nvkm_sw **psw)
+gf100_sw_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_sw **psw)
 {
-	return nvkm_sw_new_(&gf100_sw, device, index, psw);
+	return nvkm_sw_new_(&gf100_sw, device, type, inst, psw);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv04.c
index b6675fe1b0ce..4aa57573869c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv04.c
@@ -133,7 +133,7 @@ nv04_sw = {
 };
 
 int
-nv04_sw_new(struct nvkm_device *device, int index, struct nvkm_sw **psw)
+nv04_sw_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_sw **psw)
 {
-	return nvkm_sw_new_(&nv04_sw, device, index, psw);
+	return nvkm_sw_new_(&nv04_sw, device, type, inst, psw);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv10.c b/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv10.c
index 09d22fcd194c..e79e640ae535 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv10.c
@@ -62,7 +62,7 @@ nv10_sw = {
 };
 
 int
-nv10_sw_new(struct nvkm_device *device, int index, struct nvkm_sw **psw)
+nv10_sw_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_sw **psw)
 {
-	return nvkm_sw_new_(&nv10_sw, device, index, psw);
+	return nvkm_sw_new_(&nv10_sw, device, type, inst, psw);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv50.c
index 01573d187f2c..1fdd094c8b7e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sw/nv50.c
@@ -142,7 +142,7 @@ nv50_sw = {
 };
 
 int
-nv50_sw_new(struct nvkm_device *device, int index, struct nvkm_sw **psw)
+nv50_sw_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_sw **psw)
 {
-	return nvkm_sw_new_(&nv50_sw, device, index, psw);
+	return nvkm_sw_new_(&nv50_sw, device, type, inst, psw);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/sw/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/sw/priv.h
index 6d18fc6180f2..d9d83b1b8849 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/sw/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/sw/priv.h
@@ -5,8 +5,8 @@
 #include <engine/sw.h>
 struct nvkm_sw_chan;
 
-int nvkm_sw_new_(const struct nvkm_sw_func *, struct nvkm_device *,
-		 int index, struct nvkm_sw **);
+int nvkm_sw_new_(const struct nvkm_sw_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		 struct nvkm_sw **);
 
 struct nvkm_sw_chan_sclass {
 	int (*ctor)(struct nvkm_sw_chan *, const struct nvkm_oclass *,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/vp/g84.c b/drivers/gpu/drm/nouveau/nvkm/engine/vp/g84.c
index 7a96178786c4..b502266c76fd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/vp/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/vp/g84.c
@@ -36,8 +36,8 @@ g84_vp = {
 };
 
 int
-g84_vp_new(struct nvkm_device *device, int index, struct nvkm_engine **pengine)
+g84_vp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	   struct nvkm_engine **pengine)
 {
-	return nvkm_xtensa_new_(&g84_vp, device, index,
-				true, 0x00f000, pengine);
+	return nvkm_xtensa_new_(&g84_vp, device, type, inst, true, 0x00f000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/xtensa.c b/drivers/gpu/drm/nouveau/nvkm/engine/xtensa.c
index 70549381e082..f7d3ba0afb55 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/xtensa.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/xtensa.c
@@ -175,9 +175,9 @@ nvkm_xtensa = {
 };
 
 int
-nvkm_xtensa_new_(const struct nvkm_xtensa_func *func,
-		 struct nvkm_device *device, int index, bool enable,
-		 u32 addr, struct nvkm_engine **pengine)
+nvkm_xtensa_new_(const struct nvkm_xtensa_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst, bool enable, u32 addr,
+		 struct nvkm_engine **pengine)
 {
 	struct nvkm_xtensa *xtensa;
 
@@ -187,6 +187,5 @@ nvkm_xtensa_new_(const struct nvkm_xtensa_func *func,
 	xtensa->addr = addr;
 	*pengine = &xtensa->engine;
 
-	return nvkm_engine_ctor(&nvkm_xtensa, device, index,
-				enable, &xtensa->engine);
+	return nvkm_engine_ctor(&nvkm_xtensa, device, type, inst, enable, &xtensa->engine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/falcon/base.c b/drivers/gpu/drm/nouveau/nvkm/falcon/base.c
index c6a3448180d6..262641a014b0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/falcon/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/falcon/base.c
@@ -88,13 +88,12 @@ int
 nvkm_falcon_enable(struct nvkm_falcon *falcon)
 {
 	struct nvkm_device *device = falcon->owner->device;
-	enum nvkm_devidx id = falcon->owner->index;
 	int ret;
 
-	nvkm_mc_enable(device, id);
+	nvkm_mc_enable(device, falcon->owner->type, falcon->owner->inst);
 	ret = falcon->func->enable(falcon);
 	if (ret) {
-		nvkm_mc_disable(device, id);
+		nvkm_mc_disable(device, falcon->owner->type, falcon->owner->inst);
 		return ret;
 	}
 
@@ -105,15 +104,14 @@ void
 nvkm_falcon_disable(struct nvkm_falcon *falcon)
 {
 	struct nvkm_device *device = falcon->owner->device;
-	enum nvkm_devidx id = falcon->owner->index;
 
 	/* already disabled, return or wait_idle will timeout */
-	if (!nvkm_mc_enabled(device, id))
+	if (!nvkm_mc_enabled(device, falcon->owner->type, falcon->owner->inst))
 		return;
 
 	falcon->func->disable(falcon);
 
-	nvkm_mc_disable(device, id);
+	nvkm_mc_disable(device, falcon->owner->type, falcon->owner->inst);
 }
 
 int
@@ -143,7 +141,7 @@ nvkm_falcon_oneinit(struct nvkm_falcon *falcon)
 	u32 reg;
 
 	if (!falcon->addr) {
-		falcon->addr = nvkm_top_addr(subdev->device, subdev->index);
+		falcon->addr = nvkm_top_addr(subdev->device, subdev->type, subdev->inst);
 		if (WARN_ON(!falcon->addr))
 			return -ENODEV;
 	}
@@ -188,7 +186,7 @@ nvkm_falcon_get(struct nvkm_falcon *falcon, const struct nvkm_subdev *user)
 	mutex_lock(&falcon->mutex);
 	if (falcon->user) {
 		nvkm_error(user, "%s falcon already acquired by %s!\n",
-			   falcon->name, nvkm_subdev_name[falcon->user->index]);
+			   falcon->name, falcon->user->name);
 		mutex_unlock(&falcon->mutex);
 		return -EBUSY;
 	}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild
index fb4fff1222af..2cb24fff7e32 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild
@@ -11,7 +11,6 @@ include $(src)/nvkm/subdev/fuse/Kbuild
 include $(src)/nvkm/subdev/gpio/Kbuild
 include $(src)/nvkm/subdev/gsp/Kbuild
 include $(src)/nvkm/subdev/i2c/Kbuild
-include $(src)/nvkm/subdev/ibus/Kbuild
 include $(src)/nvkm/subdev/iccsense/Kbuild
 include $(src)/nvkm/subdev/instmem/Kbuild
 include $(src)/nvkm/subdev/ltc/Kbuild
@@ -20,6 +19,7 @@ include $(src)/nvkm/subdev/mmu/Kbuild
 include $(src)/nvkm/subdev/mxm/Kbuild
 include $(src)/nvkm/subdev/pci/Kbuild
 include $(src)/nvkm/subdev/pmu/Kbuild
+include $(src)/nvkm/subdev/privring/Kbuild
 include $(src)/nvkm/subdev/therm/Kbuild
 include $(src)/nvkm/subdev/timer/Kbuild
 include $(src)/nvkm/subdev/top/Kbuild
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/base.c
index c962df9910dd..af6cac696d43 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/base.c
@@ -410,14 +410,14 @@ nvkm_acr_ctor_wpr(struct nvkm_acr *acr, int ver)
 
 int
 nvkm_acr_new_(const struct nvkm_acr_fwif *fwif, struct nvkm_device *device,
-	      int index, struct nvkm_acr **pacr)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_acr **pacr)
 {
 	struct nvkm_acr *acr;
 	long wprfw;
 
 	if (!(acr = *pacr = kzalloc(sizeof(*acr), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_acr, device, index, &acr->subdev);
+	nvkm_subdev_ctor(&nvkm_acr, device, type, inst, &acr->subdev);
 	INIT_LIST_HEAD(&acr->hsfw);
 	INIT_LIST_HEAD(&acr->lsfw);
 	INIT_LIST_HEAD(&acr->hsf);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm200.c
index cd41b2e6cc87..cdb1ead26d84 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm200.c
@@ -262,7 +262,7 @@ gm200_acr_hsfw_boot(struct nvkm_acr *acr, struct nvkm_acr_hsf *hsf,
 	hsf->func->bld(acr, hsf);
 
 	/* Boot the falcon. */
-	nvkm_mc_intr_mask(device, falcon->owner->index, false);
+	nvkm_mc_intr_mask(device, falcon->owner->type, falcon->owner->inst, false);
 
 	nvkm_falcon_wr32(falcon, 0x040, 0xdeada5a5);
 	nvkm_falcon_set_start_addr(falcon, hsf->imem_tag << 8);
@@ -279,7 +279,7 @@ gm200_acr_hsfw_boot(struct nvkm_acr *acr, struct nvkm_acr_hsf *hsf,
 		return -EIO;
 
 	nvkm_falcon_clear_interrupt(falcon, intr_clear);
-	nvkm_mc_intr_mask(device, falcon->owner->index, true);
+	nvkm_mc_intr_mask(device, falcon->owner->type, falcon->owner->inst, true);
 	return ret;
 }
 
@@ -478,7 +478,8 @@ gm200_acr_fwif[] = {
 };
 
 int
-gm200_acr_new(struct nvkm_device *device, int index, struct nvkm_acr **pacr)
+gm200_acr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_acr **pacr)
 {
-	return nvkm_acr_new_(gm200_acr_fwif, device, index, pacr);
+	return nvkm_acr_new_(gm200_acr_fwif, device, type, inst, pacr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm20b.c
index b1ecc58152cc..54e996f2f630 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gm20b.c
@@ -129,7 +129,8 @@ gm20b_acr_fwif[] = {
 };
 
 int
-gm20b_acr_new(struct nvkm_device *device, int index, struct nvkm_acr **pacr)
+gm20b_acr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_acr **pacr)
 {
-	return nvkm_acr_new_(gm20b_acr_fwif, device, index, pacr);
+	return nvkm_acr_new_(gm20b_acr_fwif, device, type, inst, pacr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp102.c
index 80eb9d8dbc80..fb9132a39bb1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp102.c
@@ -276,7 +276,8 @@ gp102_acr_fwif[] = {
 };
 
 int
-gp102_acr_new(struct nvkm_device *device, int index, struct nvkm_acr **pacr)
+gp102_acr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_acr **pacr)
 {
-	return nvkm_acr_new_(gp102_acr_fwif, device, index, pacr);
+	return nvkm_acr_new_(gp102_acr_fwif, device, type, inst, pacr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp108.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp108.c
index 67a7c141004b..373d638a2177 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp108.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp108.c
@@ -106,7 +106,8 @@ gp108_acr_fwif[] = {
 };
 
 int
-gp108_acr_new(struct nvkm_device *device, int index, struct nvkm_acr **pacr)
+gp108_acr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_acr **pacr)
 {
-	return nvkm_acr_new_(gp108_acr_fwif, device, index, pacr);
+	return nvkm_acr_new_(gp108_acr_fwif, device, type, inst, pacr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp10b.c
index 8249f0d2d81d..f03ba028867b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/gp10b.c
@@ -52,7 +52,8 @@ gp10b_acr_fwif[] = {
 };
 
 int
-gp10b_acr_new(struct nvkm_device *device, int index, struct nvkm_acr **pacr)
+gp10b_acr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_acr **pacr)
 {
-	return nvkm_acr_new_(gp10b_acr_fwif, device, index, pacr);
+	return nvkm_acr_new_(gp10b_acr_fwif, device, type, inst, pacr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/priv.h
index d71af17a169a..c30b841c9d35 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/priv.h
@@ -135,8 +135,8 @@ int gp102_acr_load_load(struct nvkm_acr *, struct nvkm_acr_hsfw *);
 extern const struct nvkm_acr_hsf_func gp108_acr_unload_0;
 void gp108_acr_hsfw_bld(struct nvkm_acr *, struct nvkm_acr_hsf *);
 
-int nvkm_acr_new_(const struct nvkm_acr_fwif *, struct nvkm_device *, int,
-		  struct nvkm_acr **);
+int nvkm_acr_new_(const struct nvkm_acr_fwif *, struct nvkm_device *, enum nvkm_subdev_type,
+		  int inst, struct nvkm_acr **);
 int nvkm_acr_hsf_boot(struct nvkm_acr *, const char *name);
 
 struct nvkm_acr_lsf {
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/tu102.c
index c4981bce9a2b..05a87e77525f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/acr/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/acr/tu102.c
@@ -224,7 +224,8 @@ tu102_acr_fwif[] = {
 };
 
 int
-tu102_acr_new(struct nvkm_device *device, int index, struct nvkm_acr **pacr)
+tu102_acr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_acr **pacr)
 {
-	return nvkm_acr_new_(tu102_acr_fwif, device, index, pacr);
+	return nvkm_acr_new_(tu102_acr_fwif, device, type, inst, pacr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c
index 209a6a40834a..d017a1b5e5dd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c
@@ -134,9 +134,9 @@ nvkm_bar = {
 
 void
 nvkm_bar_ctor(const struct nvkm_bar_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_bar *bar)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_bar *bar)
 {
-	nvkm_subdev_ctor(&nvkm_bar, device, index, &bar->subdev);
+	nvkm_subdev_ctor(&nvkm_bar, device, type, inst, &bar->subdev);
 	bar->func = func;
 	spin_lock_init(&bar->lock);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/g84.c
index 87f26f54b481..77a41bcf860e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/g84.c
@@ -56,7 +56,8 @@ g84_bar_func = {
 };
 
 int
-g84_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+g84_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_bar **pbar)
 {
-	return nv50_bar_new_(&g84_bar_func, device, index, 0x200, pbar);
+	return nv50_bar_new_(&g84_bar_func, device, type, inst, 0x200, pbar);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c
index a3dcb09a40ee..51070b7dda85 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c
@@ -162,12 +162,12 @@ gf100_bar_dtor(struct nvkm_bar *base)
 
 int
 gf100_bar_new_(const struct nvkm_bar_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_bar **pbar)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_bar **pbar)
 {
 	struct gf100_bar *bar;
 	if (!(bar = kzalloc(sizeof(*bar), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_bar_ctor(func, device, index, &bar->base);
+	nvkm_bar_ctor(func, device, type, inst, &bar->base);
 	bar->bar2_halve = nvkm_boolopt(device->cfgopt, "NvBar2Halve", false);
 	*pbar = &bar->base;
 	return 0;
@@ -189,7 +189,8 @@ gf100_bar_func = {
 };
 
 int
-gf100_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+gf100_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bar **pbar)
 {
-	return gf100_bar_new_(&gf100_bar_func, device, index, pbar);
+	return gf100_bar_new_(&gf100_bar_func, device, type, inst, pbar);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.h b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.h
index 4ae4c7145712..328a68b418d9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.h
@@ -15,7 +15,7 @@ struct gf100_bar {
 	struct gf100_barN bar[2];
 };
 
-int gf100_bar_new_(const struct nvkm_bar_func *, struct nvkm_device *,
+int gf100_bar_new_(const struct nvkm_bar_func *, struct nvkm_device *, enum nvkm_subdev_type,
 		   int, struct nvkm_bar **);
 void *gf100_bar_dtor(struct nvkm_bar *);
 int gf100_bar_oneinit(struct nvkm_bar *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gk20a.c
index 35878fb538f2..eead8ab88393 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gk20a.c
@@ -32,9 +32,10 @@ gk20a_bar_func = {
 };
 
 int
-gk20a_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+gk20a_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bar **pbar)
 {
-	int ret = gf100_bar_new_(&gk20a_bar_func, device, index, pbar);
+	int ret = gf100_bar_new_(&gk20a_bar_func, device, type, inst, pbar);
 	if (ret == 0)
 		(*pbar)->iomap_uncached = true;
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm107.c
index 3ddf9222d935..da95307a7912 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm107.c
@@ -59,7 +59,8 @@ gm107_bar_func = {
 };
 
 int
-gm107_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+gm107_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bar **pbar)
 {
-	return gf100_bar_new_(&gm107_bar_func, device, index, pbar);
+	return gf100_bar_new_(&gm107_bar_func, device, type, inst, pbar);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm20b.c
index 1ed6170891c4..4acdb4fb0107 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gm20b.c
@@ -32,9 +32,10 @@ gm20b_bar_func = {
 };
 
 int
-gm20b_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+gm20b_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bar **pbar)
 {
-	int ret = gf100_bar_new_(&gm20b_bar_func, device, index, pbar);
+	int ret = gf100_bar_new_(&gm20b_bar_func, device, type, inst, pbar);
 	if (ret == 0)
 		(*pbar)->iomap_uncached = true;
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c
index f23a0ccc2bec..27d8a1be43e4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c
@@ -220,12 +220,12 @@ nv50_bar_dtor(struct nvkm_bar *base)
 
 int
 nv50_bar_new_(const struct nvkm_bar_func *func, struct nvkm_device *device,
-	      int index, u32 pgd_addr, struct nvkm_bar **pbar)
+	      enum nvkm_subdev_type type, int inst, u32 pgd_addr, struct nvkm_bar **pbar)
 {
 	struct nv50_bar *bar;
 	if (!(bar = kzalloc(sizeof(*bar), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_bar_ctor(func, device, index, &bar->base);
+	nvkm_bar_ctor(func, device, type, inst, &bar->base);
 	bar->pgd_addr = pgd_addr;
 	*pbar = &bar->base;
 	return 0;
@@ -248,7 +248,8 @@ nv50_bar_func = {
 };
 
 int
-nv50_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+nv50_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_bar **pbar)
 {
-	return nv50_bar_new_(&nv50_bar_func, device, index, 0x1400, pbar);
+	return nv50_bar_new_(&nv50_bar_func, device, type, inst, 0x1400, pbar);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.h b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.h
index e4193deb2e51..dedee9394079 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.h
@@ -16,7 +16,7 @@ struct nv50_bar {
 	struct nvkm_gpuobj *bar2;
 };
 
-int nv50_bar_new_(const struct nvkm_bar_func *, struct nvkm_device *,
+int nv50_bar_new_(const struct nvkm_bar_func *, struct nvkm_device *, enum nvkm_subdev_type,
 		  int, u32 pgd_addr, struct nvkm_bar **);
 void *nv50_bar_dtor(struct nvkm_bar *);
 int nv50_bar_oneinit(struct nvkm_bar *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/priv.h
index 869ad184f923..daebfc991c76 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/priv.h
@@ -5,7 +5,7 @@
 #include <subdev/bar.h>
 
 void nvkm_bar_ctor(const struct nvkm_bar_func *, struct nvkm_device *,
-		   int, struct nvkm_bar *);
+		   enum nvkm_subdev_type, int, struct nvkm_bar *);
 
 struct nvkm_bar_func {
 	void *(*dtor)(struct nvkm_bar *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu102.c
index 798f65ec3a86..c25ab407b85d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu102.c
@@ -92,7 +92,8 @@ tu102_bar = {
 };
 
 int
-tu102_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar)
+tu102_bar_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bar **pbar)
 {
-	return gf100_bar_new_(&tu102_bar, device, index, pbar);
+	return gf100_bar_new_(&tu102_bar, device, type, inst, pbar);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c
index f3c30b2a788e..d0f52d59fc2f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c
@@ -140,7 +140,8 @@ nvkm_bios = {
 };
 
 int
-nvkm_bios_new(struct nvkm_device *device, int index, struct nvkm_bios **pbios)
+nvkm_bios_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bios **pbios)
 {
 	struct nvkm_bios *bios;
 	struct nvbios_image image;
@@ -149,7 +150,7 @@ nvkm_bios_new(struct nvkm_device *device, int index, struct nvkm_bios **pbios)
 
 	if (!(bios = *pbios = kzalloc(sizeof(*bios), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_bios, device, index, &bios->subdev);
+	nvkm_subdev_ctor(&nvkm_bios, device, type, inst, &bios->subdev);
 
 	ret = nvbios_shadow(bios);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadow.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadow.c
index 7deb81b6dbac..4b571cc6bc70 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadow.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadow.c
@@ -75,7 +75,7 @@ shadow_image(struct nvkm_bios *bios, int idx, u32 offset, struct shadow *mthd)
 	nvkm_debug(subdev, "%08x: type %02x, %d bytes\n",
 		   image.base, image.type, image.size);
 
-	if (!shadow_fetch(bios, mthd, image.size)) {
+	if (!shadow_fetch(bios, mthd, image.base + image.size)) {
 		nvkm_debug(subdev, "%08x: fetch failed\n", image.base);
 		return 0;
 	}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadowramin.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadowramin.c
index 3634cd0630b8..023ddc7c5399 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadowramin.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/shadowramin.c
@@ -64,6 +64,9 @@ pramin_init(struct nvkm_bios *bios, const char *name)
 		return NULL;
 
 	/* we can't get the bios image pointer without PDISP */
+	if (device->card_type >= GA100)
+		addr = device->chipset == 0x170; /*XXX: find the fuse reg for this */
+	else
 	if (device->card_type >= GM100)
 		addr = nvkm_rd32(device, 0x021c04);
 	else
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/base.c
index 52ad73bce5fe..0e5a46db52ea 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/base.c
@@ -53,12 +53,12 @@ nvkm_bus = {
 
 int
 nvkm_bus_new_(const struct nvkm_bus_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_bus **pbus)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_bus **pbus)
 {
 	struct nvkm_bus *bus;
 	if (!(bus = *pbus = kzalloc(sizeof(*bus), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_bus, device, index, &bus->subdev);
+	nvkm_subdev_ctor(&nvkm_bus, device, type, inst, &bus->subdev);
 	bus->func = func;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/g94.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/g94.c
index 9700b5c01cc6..a0d6e2d3f804 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/g94.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/g94.c
@@ -58,7 +58,8 @@ g94_bus = {
 };
 
 int
-g94_bus_new(struct nvkm_device *device, int index, struct nvkm_bus **pbus)
+g94_bus_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_bus **pbus)
 {
-	return nvkm_bus_new_(&g94_bus, device, index, pbus);
+	return nvkm_bus_new_(&g94_bus, device, type, inst, pbus);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/gf100.c
index e0930d5fdfb1..53a6651ac225 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/gf100.c
@@ -40,7 +40,7 @@ gf100_bus_intr(struct nvkm_bus *bus)
 			   (addr & 0x00000002) ? "write" : "read", data,
 			   (addr & 0x00fffffc),
 			   (stat & 0x00000002) ? "!ENGINE " : "",
-			   (stat & 0x00000004) ? "IBUS " : "",
+			   (stat & 0x00000004) ? "PRIVRING " : "",
 			   (stat & 0x00000008) ? "TIMEOUT " : "");
 
 		nvkm_wr32(device, 0x009084, 0x00000000);
@@ -69,7 +69,8 @@ gf100_bus = {
 };
 
 int
-gf100_bus_new(struct nvkm_device *device, int index, struct nvkm_bus **pbus)
+gf100_bus_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_bus **pbus)
 {
-	return nvkm_bus_new_(&gf100_bus, device, index, pbus);
+	return nvkm_bus_new_(&gf100_bus, device, type, inst, pbus);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv04.c
index 2b44ba5cf4b0..cfed17c062ea 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv04.c
@@ -68,7 +68,8 @@ nv04_bus = {
 };
 
 int
-nv04_bus_new(struct nvkm_device *device, int index, struct nvkm_bus **pbus)
+nv04_bus_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_bus **pbus)
 {
-	return nvkm_bus_new_(&nv04_bus, device, index, pbus);
+	return nvkm_bus_new_(&nv04_bus, device, type, inst, pbus);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv31.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv31.c
index 5153d89e1f0b..ad8da523bb22 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv31.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv31.c
@@ -82,7 +82,8 @@ nv31_bus = {
 };
 
 int
-nv31_bus_new(struct nvkm_device *device, int index, struct nvkm_bus **pbus)
+nv31_bus_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_bus **pbus)
 {
-	return nvkm_bus_new_(&nv31_bus, device, index, pbus);
+	return nvkm_bus_new_(&nv31_bus, device, type, inst, pbus);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv50.c
index 19e10fdc9291..3a1e45adeedc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/nv50.c
@@ -99,7 +99,8 @@ nv50_bus = {
 };
 
 int
-nv50_bus_new(struct nvkm_device *device, int index, struct nvkm_bus **pbus)
+nv50_bus_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_bus **pbus)
 {
-	return nvkm_bus_new_(&nv50_bus, device, index, pbus);
+	return nvkm_bus_new_(&nv50_bus, device, type, inst, pbus);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/priv.h
index 76f7ba1c6494..2e9345b17cf8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bus/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bus/priv.h
@@ -11,7 +11,7 @@ struct nvkm_bus_func {
 	u32 hwsq_size;
 };
 
-int nvkm_bus_new_(const struct nvkm_bus_func *, struct nvkm_device *, int,
+int nvkm_bus_new_(const struct nvkm_bus_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  struct nvkm_bus **);
 
 void nv50_bus_init(struct nvkm_bus *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c
index dc184e857f85..57199be082fd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c
@@ -649,7 +649,7 @@ nvkm_clk = {
 
 int
 nvkm_clk_ctor(const struct nvkm_clk_func *func, struct nvkm_device *device,
-	      int index, bool allow_reclock, struct nvkm_clk *clk)
+	      enum nvkm_subdev_type type, int inst, bool allow_reclock, struct nvkm_clk *clk)
 {
 	struct nvkm_subdev *subdev = &clk->subdev;
 	struct nvkm_bios *bios = device->bios;
@@ -657,7 +657,7 @@ nvkm_clk_ctor(const struct nvkm_clk_func *func, struct nvkm_device *device,
 	const char *mode;
 	struct nvbios_vpstate_header h;
 
-	nvkm_subdev_ctor(&nvkm_clk, device, index, subdev);
+	nvkm_subdev_ctor(&nvkm_clk, device, type, inst, subdev);
 
 	if (bios && !nvbios_vpstate_parse(bios, &h)) {
 		struct nvbios_vpstate_entry base, boost;
@@ -716,9 +716,9 @@ nvkm_clk_ctor(const struct nvkm_clk_func *func, struct nvkm_device *device,
 
 int
 nvkm_clk_new_(const struct nvkm_clk_func *func, struct nvkm_device *device,
-	      int index, bool allow_reclock, struct nvkm_clk **pclk)
+	      enum nvkm_subdev_type type, int inst, bool allow_reclock, struct nvkm_clk **pclk)
 {
 	if (!(*pclk = kzalloc(sizeof(**pclk), GFP_KERNEL)))
 		return -ENOMEM;
-	return nvkm_clk_ctor(func, device, index, allow_reclock, *pclk);
+	return nvkm_clk_ctor(func, device, type, inst, allow_reclock, *pclk);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/g84.c
index f97e3ec196bb..07157cf53c9e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/g84.c
@@ -41,8 +41,8 @@ g84_clk = {
 };
 
 int
-g84_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+g84_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_clk **pclk)
 {
-	return nv50_clk_new_(&g84_clk, device, index,
-			     (device->chipset >= 0x94), pclk);
+	return nv50_clk_new_(&g84_clk, device, type, inst, (device->chipset >= 0x94), pclk);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c
index 7f67f9f5a550..6eea11aefb70 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c
@@ -468,7 +468,8 @@ gf100_clk = {
 };
 
 int
-gf100_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+gf100_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_clk **pclk)
 {
 	struct gf100_clk *clk;
 
@@ -476,5 +477,5 @@ gf100_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	return nvkm_clk_ctor(&gf100_clk, device, index, false, &clk->base);
+	return nvkm_clk_ctor(&gf100_clk, device, type, inst, false, &clk->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c
index 0b37e3da7feb..0d8e2ddcc5ee 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c
@@ -504,7 +504,8 @@ gk104_clk = {
 };
 
 int
-gk104_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+gk104_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_clk **pclk)
 {
 	struct gk104_clk *clk;
 
@@ -512,5 +513,5 @@ gk104_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	return nvkm_clk_ctor(&gk104_clk, device, index, true, &clk->base);
+	return nvkm_clk_ctor(&gk104_clk, device, type, inst, true, &clk->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
index 218893e3e5f9..d573fb0917fc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
@@ -610,10 +610,9 @@ gk20a_clk = {
 };
 
 int
-gk20a_clk_ctor(struct nvkm_device *device, int index,
-		const struct nvkm_clk_func *func,
-		const struct gk20a_clk_pllg_params *params,
-		struct gk20a_clk *clk)
+gk20a_clk_ctor(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       const struct nvkm_clk_func *func, const struct gk20a_clk_pllg_params *params,
+	       struct gk20a_clk *clk)
 {
 	struct nvkm_device_tegra *tdev = device->func->tegra(device);
 	int ret;
@@ -628,7 +627,7 @@ gk20a_clk_ctor(struct nvkm_device *device, int index,
 	clk->params = params;
 	clk->parent_rate = clk_get_rate(tdev->clk);
 
-	ret = nvkm_clk_ctor(func, device, index, true, &clk->base);
+	ret = nvkm_clk_ctor(func, device, type, inst, true, &clk->base);
 	if (ret)
 		return ret;
 
@@ -639,7 +638,8 @@ gk20a_clk_ctor(struct nvkm_device *device, int index,
 }
 
 int
-gk20a_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+gk20a_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_clk **pclk)
 {
 	struct gk20a_clk *clk;
 	int ret;
@@ -649,11 +649,9 @@ gk20a_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	ret = gk20a_clk_ctor(device, index, &gk20a_clk, &gk20a_pllg_params,
-			      clk);
+	ret = gk20a_clk_ctor(device, type, inst, &gk20a_clk, &gk20a_pllg_params, clk);
 
 	clk->pl_to_div = pl_to_div;
 	clk->div_to_pl = div_to_pl;
-
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h
index 0d1450972162..286413ff4a9e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h
@@ -146,8 +146,8 @@ gk20a_pllg_n_lo(struct gk20a_clk *clk, struct gk20a_pll *pll)
 			    clk->parent_rate / KHZ);
 }
 
-int gk20a_clk_ctor(struct nvkm_device *, int, const struct nvkm_clk_func *,
-		    const struct gk20a_clk_pllg_params *, struct gk20a_clk *);
+int gk20a_clk_ctor(struct nvkm_device *, enum nvkm_subdev_type, int, const struct nvkm_clk_func *,
+		   const struct gk20a_clk_pllg_params *, struct gk20a_clk *);
 void gk20a_clk_fini(struct nvkm_clk *);
 int gk20a_clk_read(struct nvkm_clk *, enum nv_clk_src);
 int gk20a_clk_calc(struct nvkm_clk *, struct nvkm_cstate *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c
index b284e949f732..a139dafffe06 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c
@@ -908,7 +908,7 @@ gm20b_clk = {
 };
 
 static int
-gm20b_clk_new_speedo0(struct nvkm_device *device, int index,
+gm20b_clk_new_speedo0(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		      struct nvkm_clk **pclk)
 {
 	struct gk20a_clk *clk;
@@ -919,12 +919,9 @@ gm20b_clk_new_speedo0(struct nvkm_device *device, int index,
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	ret = gk20a_clk_ctor(device, index, &gm20b_clk_speedo0,
-			     &gm20b_pllg_params, clk);
-
+	ret = gk20a_clk_ctor(device, type, inst, &gm20b_clk_speedo0, &gm20b_pllg_params, clk);
 	clk->pl_to_div = pl_to_div;
 	clk->div_to_pl = div_to_pl;
-
 	return ret;
 }
 
@@ -1014,7 +1011,8 @@ gm20b_clk_init_safe_fmax(struct gm20b_clk *clk)
 }
 
 int
-gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+gm20b_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_clk **pclk)
 {
 	struct nvkm_device_tegra *tdev = device->func->tegra(device);
 	struct gm20b_clk *clk;
@@ -1024,7 +1022,7 @@ gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 
 	/* Speedo 0 GPUs cannot use noise-aware PLL */
 	if (tdev->gpu_speedo_id == 0)
-		return gm20b_clk_new_speedo0(device, index, pclk);
+		return gm20b_clk_new_speedo0(device, type, inst, pclk);
 
 	/* Speedo >= 1, use NAPLL */
 	clk = kzalloc(sizeof(*clk) + sizeof(*clk_params), GFP_KERNEL);
@@ -1036,8 +1034,7 @@ gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 	/* duplicate the clock parameters since we will patch them below */
 	clk_params = (void *) (clk + 1);
 	*clk_params = gm20b_pllg_params;
-	ret = gk20a_clk_ctor(device, index, &gm20b_clk, clk_params,
-			     &clk->base);
+	ret = gk20a_clk_ctor(device, type, inst, &gm20b_clk, clk_params, &clk->base);
 	if (ret)
 		return ret;
 
@@ -1050,7 +1047,7 @@ gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 	if (clk_params->max_m == 0) {
 		nvkm_warn(subdev, "cannot use NAPLL, using legacy clock...\n");
 		kfree(clk);
-		return gm20b_clk_new_speedo0(device, index, pclk);
+		return gm20b_clk_new_speedo0(device, type, inst, pclk);
 	}
 
 	clk->base.pl_to_div = pl_to_div;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gt215.c
index f0a26881d9b9..b5f3969727a2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gt215.c
@@ -537,7 +537,8 @@ gt215_clk = {
 };
 
 int
-gt215_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+gt215_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_clk **pclk)
 {
 	struct gt215_clk *clk;
 
@@ -545,5 +546,5 @@ gt215_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	return nvkm_clk_ctor(&gt215_clk, device, index, true, &clk->base);
+	return nvkm_clk_ctor(&gt215_clk, device, type, inst, true, &clk->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/mcp77.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/mcp77.c
index 4884eb4a9221..81f103f88dc8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/mcp77.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/mcp77.c
@@ -409,7 +409,8 @@ mcp77_clk = {
 };
 
 int
-mcp77_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+mcp77_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_clk **pclk)
 {
 	struct mcp77_clk *clk;
 
@@ -417,5 +418,5 @@ mcp77_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	return nvkm_clk_ctor(&mcp77_clk, device, index, true, &clk->base);
+	return nvkm_clk_ctor(&mcp77_clk, device, type, inst, true, &clk->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv04.c
index b280f85e8827..ca13598c2caa 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv04.c
@@ -72,9 +72,10 @@ nv04_clk = {
 };
 
 int
-nv04_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+nv04_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_clk **pclk)
 {
-	int ret = nvkm_clk_new_(&nv04_clk, device, index, false, pclk);
+	int ret = nvkm_clk_new_(&nv04_clk, device, type, inst, false, pclk);
 	if (ret == 0) {
 		(*pclk)->pll_calc = nv04_clk_pll_calc;
 		(*pclk)->pll_prog = nv04_clk_pll_prog;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv40.c
index 2ab9b9b84018..7ddd8cecb805 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv40.c
@@ -218,7 +218,8 @@ nv40_clk = {
 };
 
 int
-nv40_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+nv40_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_clk **pclk)
 {
 	struct nv40_clk *clk;
 
@@ -228,5 +229,5 @@ nv40_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 	clk->base.pll_prog = nv04_clk_pll_prog;
 	*pclk = &clk->base;
 
-	return nvkm_clk_ctor(&nv40_clk, device, index, true, &clk->base);
+	return nvkm_clk_ctor(&nv40_clk, device, type, inst, true, &clk->base);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.c
index da1770e47490..83067763c0ec 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.c
@@ -507,14 +507,14 @@ nv50_clk_tidy(struct nvkm_clk *base)
 
 int
 nv50_clk_new_(const struct nvkm_clk_func *func, struct nvkm_device *device,
-	      int index, bool allow_reclock, struct nvkm_clk **pclk)
+	      enum nvkm_subdev_type type, int inst, bool allow_reclock, struct nvkm_clk **pclk)
 {
 	struct nv50_clk *clk;
 	int ret;
 
 	if (!(clk = kzalloc(sizeof(*clk), GFP_KERNEL)))
 		return -ENOMEM;
-	ret = nvkm_clk_ctor(func, device, index, allow_reclock, &clk->base);
+	ret = nvkm_clk_ctor(func, device, type, inst, allow_reclock, &clk->base);
 	*pclk = &clk->base;
 	if (ret)
 		return ret;
@@ -555,7 +555,8 @@ nv50_clk = {
 };
 
 int
-nv50_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+nv50_clk_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_clk **pclk)
 {
-	return nv50_clk_new_(&nv50_clk, device, index, false, pclk);
+	return nv50_clk_new_(&nv50_clk, device, type, inst, false, pclk);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.h b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.h
index 7c7713238ec4..5b4cb7e5cff6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/nv50.h
@@ -20,7 +20,7 @@ struct nv50_clk {
 	struct nv50_clk_hwsq hwsq;
 };
 
-int nv50_clk_new_(const struct nvkm_clk_func *, struct nvkm_device *, int,
+int nv50_clk_new_(const struct nvkm_clk_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  bool, struct nvkm_clk **);
 int nv50_clk_read(struct nvkm_clk *, enum nv_clk_src);
 int nv50_clk_calc(struct nvkm_clk *, struct nvkm_cstate *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/priv.h
index 81dfb37480ae..810cc572cd30 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/priv.h
@@ -16,9 +16,9 @@ struct nvkm_clk_func {
 	struct nvkm_domain domains[];
 };
 
-int nvkm_clk_ctor(const struct nvkm_clk_func *, struct nvkm_device *, int,
+int nvkm_clk_ctor(const struct nvkm_clk_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  bool allow_reclock, struct nvkm_clk *);
-int nvkm_clk_new_(const struct nvkm_clk_func *, struct nvkm_device *, int,
+int nvkm_clk_new_(const struct nvkm_clk_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  bool allow_reclock, struct nvkm_clk **);
 
 int nv04_clk_pll_calc(struct nvkm_clk *, struct nvbios_pll *, int clk,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild
index b3429371ed82..d1abb64841da 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild
@@ -15,3 +15,4 @@ nvkm-y += nvkm/subdev/devinit/gm107.o
 nvkm-y += nvkm/subdev/devinit/gm200.o
 nvkm-y += nvkm/subdev/devinit/gv100.o
 nvkm-y += nvkm/subdev/devinit/tu102.o
+nvkm-y += nvkm/subdev/devinit/ga100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/base.c
index 4756019ddf3f..dd4981708fe4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/base.c
@@ -56,12 +56,12 @@ nvkm_devinit_disable(struct nvkm_devinit *init)
 }
 
 int
-nvkm_devinit_post(struct nvkm_devinit *init, u64 *disable)
+nvkm_devinit_post(struct nvkm_devinit *init)
 {
 	int ret = 0;
 	if (init && init->func->post)
 		ret = init->func->post(init, init->post);
-	*disable = nvkm_devinit_disable(init);
+	nvkm_devinit_disable(init);
 	return ret;
 }
 
@@ -126,11 +126,10 @@ nvkm_devinit = {
 };
 
 void
-nvkm_devinit_ctor(const struct nvkm_devinit_func *func,
-		  struct nvkm_device *device, int index,
-		  struct nvkm_devinit *init)
+nvkm_devinit_ctor(const struct nvkm_devinit_func *func, struct nvkm_device *device,
+		  enum nvkm_subdev_type type, int inst, struct nvkm_devinit *init)
 {
-	nvkm_subdev_ctor(&nvkm_devinit, device, index, &init->subdev);
+	nvkm_subdev_ctor(&nvkm_devinit, device, type, inst, &init->subdev);
 	init->func = func;
 	init->force_post = nvkm_boolopt(device->cfgopt, "NvForcePost", false);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g84.c
index e895289bf3c1..c224702b7bed 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g84.c
@@ -35,18 +35,18 @@ g84_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0ULL;
 
 	if (!(r001540 & 0x40000000)) {
-		disable |= (1ULL << NVKM_ENGINE_MPEG);
-		disable |= (1ULL << NVKM_ENGINE_VP);
-		disable |= (1ULL << NVKM_ENGINE_BSP);
-		disable |= (1ULL << NVKM_ENGINE_CIPHER);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MPEG, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_VP, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_BSP, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CIPHER, 0);
 	}
 
 	if (!(r00154c & 0x00000004))
-		disable |= (1ULL << NVKM_ENGINE_DISP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_DISP, 0);
 	if (!(r00154c & 0x00000020))
-		disable |= (1ULL << NVKM_ENGINE_BSP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_BSP, 0);
 	if (!(r00154c & 0x00000040))
-		disable |= (1ULL << NVKM_ENGINE_CIPHER);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CIPHER, 0);
 
 	return disable;
 }
@@ -61,8 +61,8 @@ g84_devinit = {
 };
 
 int
-g84_devinit_new(struct nvkm_device *device, int index,
+g84_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&g84_devinit, device, index, pinit);
+	return nv50_devinit_new_(&g84_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g98.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g98.c
index a9d45844df5a..05729ca19e9a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g98.c
@@ -35,17 +35,17 @@ g98_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0ULL;
 
 	if (!(r001540 & 0x40000000)) {
-		disable |= (1ULL << NVKM_ENGINE_MSPDEC);
-		disable |= (1ULL << NVKM_ENGINE_MSVLD);
-		disable |= (1ULL << NVKM_ENGINE_MSPPP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPDEC, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSVLD, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPPP, 0);
 	}
 
 	if (!(r00154c & 0x00000004))
-		disable |= (1ULL << NVKM_ENGINE_DISP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_DISP, 0);
 	if (!(r00154c & 0x00000020))
-		disable |= (1ULL << NVKM_ENGINE_MSVLD);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSVLD, 0);
 	if (!(r00154c & 0x00000040))
-		disable |= (1ULL << NVKM_ENGINE_SEC);
+		nvkm_subdev_disable(device, NVKM_ENGINE_SEC, 0);
 
 	return disable;
 }
@@ -60,8 +60,8 @@ g98_devinit = {
 };
 
 int
-g98_devinit_new(struct nvkm_device *device, int index,
+g98_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&g98_devinit, device, index, pinit);
+	return nv50_devinit_new_(&g98_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/ga100.c
new file mode 100644
index 000000000000..6b280b05c4ca
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/ga100.c
@@ -0,0 +1,77 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "nv50.h"
+
+#include <subdev/bios.h>
+#include <subdev/bios/pll.h>
+#include <subdev/clk/pll.h>
+
+static int
+ga100_devinit_pll_set(struct nvkm_devinit *init, u32 type, u32 freq)
+{
+	struct nvkm_subdev *subdev = &init->subdev;
+	struct nvkm_device *device = subdev->device;
+	struct nvbios_pll info;
+	int head = type - PLL_VPLL0;
+	int N, fN, M, P;
+	int ret;
+
+	ret = nvbios_pll_parse(device->bios, type, &info);
+	if (ret)
+		return ret;
+
+	ret = gt215_pll_calc(subdev, &info, freq, &N, &fN, &M, &P);
+	if (ret < 0)
+		return ret;
+
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+	case PLL_VPLL2:
+	case PLL_VPLL3:
+		nvkm_wr32(device, 0x00ef00 + (head * 0x40), 0x02080004);
+		nvkm_wr32(device, 0x00ef18 + (head * 0x40), (N << 16) | fN);
+		nvkm_wr32(device, 0x00ef04 + (head * 0x40), (P << 16) | M);
+		nvkm_wr32(device, 0x00e9c0 + (head * 0x04), 0x00000001);
+		break;
+	default:
+		nvkm_warn(subdev, "%08x/%dKhz unimplemented\n", type, freq);
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static const struct nvkm_devinit_func
+ga100_devinit = {
+	.init = nv50_devinit_init,
+	.post = tu102_devinit_post,
+	.pll_set = ga100_devinit_pll_set,
+};
+
+int
+ga100_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
+{
+	return nv50_devinit_new_(&ga100_devinit, device, type, inst, pinit);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gf100.c
index 8b1b34c3ad26..051cfd6a5caf 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gf100.c
@@ -71,21 +71,21 @@ gf100_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0ULL;
 
 	if (r022500 & 0x00000001)
-		disable |= (1ULL << NVKM_ENGINE_DISP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_DISP, 0);
 
 	if (r022500 & 0x00000002) {
-		disable |= (1ULL << NVKM_ENGINE_MSPDEC);
-		disable |= (1ULL << NVKM_ENGINE_MSPPP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPDEC, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPPP, 0);
 	}
 
 	if (r022500 & 0x00000004)
-		disable |= (1ULL << NVKM_ENGINE_MSVLD);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSVLD, 0);
 	if (r022500 & 0x00000008)
-		disable |= (1ULL << NVKM_ENGINE_MSENC);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSENC, 0);
 	if (r022500 & 0x00000100)
-		disable |= (1ULL << NVKM_ENGINE_CE0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CE, 0);
 	if (r022500 & 0x00000200)
-		disable |= (1ULL << NVKM_ENGINE_CE1);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CE, 1);
 
 	return disable;
 }
@@ -114,8 +114,8 @@ gf100_devinit = {
 };
 
 int
-gf100_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+gf100_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&gf100_devinit, device, index, pinit);
+	return nv50_devinit_new_(&gf100_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm107.c
index 28ca01be3d38..4323732a3cb2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm107.c
@@ -35,11 +35,11 @@ gm107_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0ULL;
 
 	if (r021c00 & 0x00000001)
-		disable |= (1ULL << NVKM_ENGINE_CE0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CE, 0);
 	if (r021c00 & 0x00000004)
-		disable |= (1ULL << NVKM_ENGINE_CE2);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CE, 2);
 	if (r021c04 & 0x00000001)
-		disable |= (1ULL << NVKM_ENGINE_DISP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_DISP, 0);
 
 	return disable;
 }
@@ -54,8 +54,8 @@ gm107_devinit = {
 };
 
 int
-gm107_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+gm107_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&gm107_devinit, device, index, pinit);
+	return nv50_devinit_new_(&gm107_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c
index 59940dacc2ba..a308b9bde449 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c
@@ -179,8 +179,8 @@ gm200_devinit = {
 };
 
 int
-gm200_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+gm200_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&gm200_devinit, device, index, pinit);
+	return nv50_devinit_new_(&gm200_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gt215.c
index 9a8522fa9c65..dc026ac1b595 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gt215.c
@@ -71,16 +71,16 @@ gt215_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0ULL;
 
 	if (!(r001540 & 0x40000000)) {
-		disable |= (1ULL << NVKM_ENGINE_MSPDEC);
-		disable |= (1ULL << NVKM_ENGINE_MSPPP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPDEC, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPPP, 0);
 	}
 
 	if (!(r00154c & 0x00000004))
-		disable |= (1ULL << NVKM_ENGINE_DISP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_DISP, 0);
 	if (!(r00154c & 0x00000020))
-		disable |= (1ULL << NVKM_ENGINE_MSVLD);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSVLD, 0);
 	if (!(r00154c & 0x00000200))
-		disable |= (1ULL << NVKM_ENGINE_CE0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CE, 0);
 
 	return disable;
 }
@@ -146,8 +146,8 @@ gt215_devinit = {
 };
 
 int
-gt215_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+gt215_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&gt215_devinit, device, index, pinit);
+	return nv50_devinit_new_(&gt215_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gv100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gv100.c
index fbde6828bd38..b4d1688517d5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gv100.c
@@ -72,8 +72,8 @@ gv100_devinit = {
 };
 
 int
-gv100_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+gv100_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&gv100_devinit, device, index, pinit);
+	return nv50_devinit_new_(&gv100_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/mcp89.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/mcp89.c
index ce4f718e98a1..fb90d47e1225 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/mcp89.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/mcp89.c
@@ -35,18 +35,18 @@ mcp89_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0;
 
 	if (!(r001540 & 0x40000000)) {
-		disable |= (1ULL << NVKM_ENGINE_MSPDEC);
-		disable |= (1ULL << NVKM_ENGINE_MSPPP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPDEC, 0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSPPP, 0);
 	}
 
 	if (!(r00154c & 0x00000004))
-		disable |= (1ULL << NVKM_ENGINE_DISP);
+		nvkm_subdev_disable(device, NVKM_ENGINE_DISP, 0);
 	if (!(r00154c & 0x00000020))
-		disable |= (1ULL << NVKM_ENGINE_MSVLD);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MSVLD, 0);
 	if (!(r00154c & 0x00000040))
-		disable |= (1ULL << NVKM_ENGINE_VIC);
+		nvkm_subdev_disable(device, NVKM_ENGINE_VIC, 0);
 	if (!(r00154c & 0x00000200))
-		disable |= (1ULL << NVKM_ENGINE_CE0);
+		nvkm_subdev_disable(device, NVKM_ENGINE_CE, 0);
 
 	return disable;
 }
@@ -61,8 +61,8 @@ mcp89_devinit = {
 };
 
 int
-mcp89_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+mcp89_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&mcp89_devinit, device, index, pinit);
+	return nv50_devinit_new_(&mcp89_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.c
index 317ce9fb8225..88bc890f89a2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.c
@@ -434,9 +434,8 @@ nv04_devinit_dtor(struct nvkm_devinit *base)
 }
 
 int
-nv04_devinit_new_(const struct nvkm_devinit_func *func,
-		  struct nvkm_device *device, int index,
-		  struct nvkm_devinit **pinit)
+nv04_devinit_new_(const struct nvkm_devinit_func *func, struct nvkm_device *device,
+		  enum nvkm_subdev_type type, int inst, struct nvkm_devinit **pinit)
 {
 	struct nv04_devinit *init;
 
@@ -444,7 +443,7 @@ nv04_devinit_new_(const struct nvkm_devinit_func *func,
 		return -ENOMEM;
 	*pinit = &init->base;
 
-	nvkm_devinit_ctor(func, device, index, &init->base);
+	nvkm_devinit_ctor(func, device, type, inst, &init->base);
 	init->owner = -1;
 	return 0;
 }
@@ -459,8 +458,8 @@ nv04_devinit = {
 };
 
 int
-nv04_devinit_new(struct nvkm_device *device, int index,
+nv04_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_devinit **pinit)
 {
-	return nv04_devinit_new_(&nv04_devinit, device, index, pinit);
+	return nv04_devinit_new_(&nv04_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.h b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.h
index 15b029ddf6df..06ad8a606bb8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv04.h
@@ -11,7 +11,7 @@ struct nv04_devinit {
 };
 
 int nv04_devinit_new_(const struct nvkm_devinit_func *, struct nvkm_device *,
-		      int, struct nvkm_devinit **);
+		      enum nvkm_subdev_type, int, struct nvkm_devinit **);
 void *nv04_devinit_dtor(struct nvkm_devinit *);
 void nv04_devinit_preinit(struct nvkm_devinit *);
 void nv04_devinit_fini(struct nvkm_devinit *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv05.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv05.c
index 9891eadca1ce..1410befd2285 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv05.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv05.c
@@ -136,8 +136,8 @@ nv05_devinit = {
 };
 
 int
-nv05_devinit_new(struct nvkm_device *device, int index,
+nv05_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_devinit **pinit)
 {
-	return nv04_devinit_new_(&nv05_devinit, device, index, pinit);
+	return nv04_devinit_new_(&nv05_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv10.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv10.c
index 570822f83acf..a6aa8786d610 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv10.c
@@ -106,8 +106,8 @@ nv10_devinit = {
 };
 
 int
-nv10_devinit_new(struct nvkm_device *device, int index,
+nv10_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_devinit **pinit)
 {
-	return nv04_devinit_new_(&nv10_devinit, device, index, pinit);
+	return nv04_devinit_new_(&nv10_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv1a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv1a.c
index fefafec7e2a7..4cc5ef9a5a63 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv1a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv1a.c
@@ -35,8 +35,8 @@ nv1a_devinit = {
 };
 
 int
-nv1a_devinit_new(struct nvkm_device *device, int index,
+nv1a_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_devinit **pinit)
 {
-	return nv04_devinit_new_(&nv1a_devinit, device, index, pinit);
+	return nv04_devinit_new_(&nv1a_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv20.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv20.c
index 4ef04e0d8826..67f46df723e4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv20.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv20.c
@@ -72,8 +72,8 @@ nv20_devinit = {
 };
 
 int
-nv20_devinit_new(struct nvkm_device *device, int index,
+nv20_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_devinit **pinit)
 {
-	return nv04_devinit_new_(&nv20_devinit, device, index, pinit);
+	return nv04_devinit_new_(&nv20_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.c
index d7947c4391dc..380995d398b1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.c
@@ -85,7 +85,7 @@ nv50_devinit_disable(struct nvkm_devinit *init)
 	u64 disable = 0ULL;
 
 	if (!(r001540 & 0x40000000))
-		disable |= (1ULL << NVKM_ENGINE_MPEG);
+		nvkm_subdev_disable(device, NVKM_ENGINE_MPEG, 0);
 
 	return disable;
 }
@@ -101,8 +101,8 @@ nv50_devinit_preinit(struct nvkm_devinit *base)
 	 * missing, assume it's a secondary gpu which requires post
 	 */
 	if (!base->post) {
-		u64 disable = nvkm_devinit_disable(base);
-		if (disable & (1ULL << NVKM_ENGINE_DISP))
+		nvkm_devinit_disable(base);
+		if (!device->disp)
 			base->post = true;
 	}
 
@@ -148,9 +148,8 @@ nv50_devinit_init(struct nvkm_devinit *base)
 }
 
 int
-nv50_devinit_new_(const struct nvkm_devinit_func *func,
-		  struct nvkm_device *device, int index,
-		  struct nvkm_devinit **pinit)
+nv50_devinit_new_(const struct nvkm_devinit_func *func, struct nvkm_device *device,
+		  enum nvkm_subdev_type type, int inst, struct nvkm_devinit **pinit)
 {
 	struct nv50_devinit *init;
 
@@ -158,7 +157,7 @@ nv50_devinit_new_(const struct nvkm_devinit_func *func,
 		return -ENOMEM;
 	*pinit = &init->base;
 
-	nvkm_devinit_ctor(func, device, index, &init->base);
+	nvkm_devinit_ctor(func, device, type, inst, &init->base);
 	return 0;
 }
 
@@ -172,8 +171,8 @@ nv50_devinit = {
 };
 
 int
-nv50_devinit_new(struct nvkm_device *device, int index,
+nv50_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&nv50_devinit, device, index, pinit);
+	return nv50_devinit_new_(&nv50_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h
index e8d37a6145a2..987a7f478b84 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h
@@ -9,7 +9,7 @@ struct nv50_devinit {
 	u32 r001540;
 };
 
-int nv50_devinit_new_(const struct nvkm_devinit_func *, struct nvkm_device *,
+int nv50_devinit_new_(const struct nvkm_devinit_func *, struct nvkm_device *, enum nvkm_subdev_type,
 		      int, struct nvkm_devinit **);
 void nv50_devinit_preinit(struct nvkm_devinit *);
 void nv50_devinit_init(struct nvkm_devinit *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/priv.h
index 94723352137a..dd8b038a8cee 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/priv.h
@@ -16,7 +16,9 @@ struct nvkm_devinit_func {
 };
 
 void nvkm_devinit_ctor(const struct nvkm_devinit_func *, struct nvkm_device *,
-		       int index, struct nvkm_devinit *);
+		       enum nvkm_subdev_type, int inst, struct nvkm_devinit *);
+u64 nvkm_devinit_disable(struct nvkm_devinit *);
 
 int nv04_devinit_post(struct nvkm_devinit *, bool);
+int tu102_devinit_post(struct nvkm_devinit *, bool);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu102.c
index 397670e72fff..634f64f88fc8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu102.c
@@ -65,7 +65,7 @@ tu102_devinit_pll_set(struct nvkm_devinit *init, u32 type, u32 freq)
 	return ret;
 }
 
-static int
+int
 tu102_devinit_post(struct nvkm_devinit *base, bool post)
 {
 	struct nv50_devinit *init = nv50_devinit(base);
@@ -82,8 +82,8 @@ tu102_devinit = {
 };
 
 int
-tu102_devinit_new(struct nvkm_device *device, int index,
-		struct nvkm_devinit **pinit)
+tu102_devinit_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		  struct nvkm_devinit **pinit)
 {
-	return nv50_devinit_new_(&tu102_devinit, device, index, pinit);
+	return nv50_devinit_new_(&tu102_devinit, device, type, inst, pinit);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c
index f6dca97140d6..fd54fa504efa 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c
@@ -170,12 +170,12 @@ nvkm_fault = {
 
 int
 nvkm_fault_new_(const struct nvkm_fault_func *func, struct nvkm_device *device,
-		int index, struct nvkm_fault **pfault)
+		enum nvkm_subdev_type type, int inst, struct nvkm_fault **pfault)
 {
 	struct nvkm_fault *fault;
 	if (!(fault = *pfault = kzalloc(sizeof(*fault), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_fault, device, index, &fault->subdev);
+	nvkm_subdev_ctor(&nvkm_fault, device, type, inst, &fault->subdev);
 	fault->func = func;
 	fault->user.ctor = nvkm_ufault_new;
 	fault->user.base = func->user.base;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c
index f6b189cc4330..6af7959e02ea 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c
@@ -30,7 +30,7 @@ void
 gp100_fault_buffer_intr(struct nvkm_fault_buffer *buffer, bool enable)
 {
 	struct nvkm_device *device = buffer->fault->subdev.device;
-	nvkm_mc_intr_mask(device, NVKM_SUBDEV_FAULT, enable);
+	nvkm_mc_intr_mask(device, NVKM_SUBDEV_FAULT, 0, enable);
 }
 
 void
@@ -82,8 +82,8 @@ gp100_fault = {
 };
 
 int
-gp100_fault_new(struct nvkm_device *device, int index,
+gp100_fault_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_fault **pfault)
 {
-	return nvkm_fault_new_(&gp100_fault, device, index, pfault);
+	return nvkm_fault_new_(&gp100_fault, device, type, inst, pfault);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp10b.c
index 9e66d1f7654d..89e0bc96fb92 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp10b.c
@@ -46,8 +46,8 @@ gp10b_fault = {
 };
 
 int
-gp10b_fault_new(struct nvkm_device *device, int index,
+gp10b_fault_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_fault **pfault)
 {
-	return nvkm_fault_new_(&gp10b_fault, device, index, pfault);
+	return nvkm_fault_new_(&gp10b_fault, device, type, inst, pfault);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c
index 2707be4ffabc..cd9d2ade5ac7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c
@@ -228,8 +228,8 @@ gv100_fault = {
 };
 
 int
-gv100_fault_new(struct nvkm_device *device, int index,
+gv100_fault_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_fault **pfault)
 {
-	return nvkm_fault_new_(&gv100_fault, device, index, pfault);
+	return nvkm_fault_new_(&gv100_fault, device, type, inst, pfault);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h
index f6f1dd7eee1f..36681c347fb5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h
@@ -18,8 +18,8 @@ struct nvkm_fault_buffer {
 	u64 addr;
 };
 
-int nvkm_fault_new_(const struct nvkm_fault_func *, struct nvkm_device *,
-		    int index, struct nvkm_fault **);
+int nvkm_fault_new_(const struct nvkm_fault_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		    int inst, struct nvkm_fault **);
 
 struct nvkm_fault_func {
 	int (*oneinit)(struct nvkm_fault *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu102.c
index 45a6a68b9f48..91eb6729c84d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu102.c
@@ -22,6 +22,7 @@
 #include "priv.h"
 
 #include <core/memory.h>
+#include <subdev/mc.h>
 #include <subdev/mmu.h>
 #include <engine/fifo.h>
 
@@ -34,6 +35,9 @@ tu102_fault_buffer_intr(struct nvkm_fault_buffer *buffer, bool enable)
 	 *     which don't appear to actually work anymore, but newer
 	 *     versions of RM don't appear to touch anything at all..
 	 */
+	struct nvkm_device *device = buffer->fault->subdev.device;
+
+	nvkm_mc_intr_mask(device, NVKM_SUBDEV_FAULT, 0, enable);
 }
 
 static void
@@ -41,6 +45,11 @@ tu102_fault_buffer_fini(struct nvkm_fault_buffer *buffer)
 {
 	struct nvkm_device *device = buffer->fault->subdev.device;
 	const u32 foff = buffer->id * 0x20;
+
+	/* Disable the fault interrupts */
+	nvkm_wr32(device, 0xb81408, 0x1);
+	nvkm_wr32(device, 0xb81410, 0x10);
+
 	nvkm_mask(device, 0xb83010 + foff, 0x80000000, 0x00000000);
 }
 
@@ -50,6 +59,10 @@ tu102_fault_buffer_init(struct nvkm_fault_buffer *buffer)
 	struct nvkm_device *device = buffer->fault->subdev.device;
 	const u32 foff = buffer->id * 0x20;
 
+	/* Enable the fault interrupts */
+	nvkm_wr32(device, 0xb81208, 0x1);
+	nvkm_wr32(device, 0xb81210, 0x10);
+
 	nvkm_mask(device, 0xb83010 + foff, 0xc0000000, 0x40000000);
 	nvkm_wr32(device, 0xb83004 + foff, upper_32_bits(buffer->addr));
 	nvkm_wr32(device, 0xb83000 + foff, lower_32_bits(buffer->addr));
@@ -109,14 +122,20 @@ tu102_fault_intr(struct nvkm_fault *fault)
 	}
 
 	if (stat & 0x00000200) {
+		/* Clear the associated interrupt flag */
+		nvkm_wr32(device, 0xb81010, 0x10);
+
 		if (fault->buffer[0]) {
 			nvkm_event_send(&fault->event, 1, 0, NULL, 0);
 			stat &= ~0x00000200;
 		}
 	}
 
-	/*XXX: guess, can't confirm until we get fw... */
+	/* Replayable MMU fault */
 	if (stat & 0x00000100) {
+		/* Clear the associated interrupt flag */
+		nvkm_wr32(device, 0xb81008, 0x1);
+
 		if (fault->buffer[1]) {
 			nvkm_event_send(&fault->event, 1, 1, NULL, 0);
 			stat &= ~0x00000100;
@@ -162,8 +181,8 @@ tu102_fault = {
 };
 
 int
-tu102_fault_new(struct nvkm_device *device, int index,
+tu102_fault_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_fault **pfault)
 {
-	return nvkm_fault_new_(&tu102_fault, device, index, pfault);
+	return nvkm_fault_new_(&tu102_fault, device, type, inst, pfault);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild
index 43a42159a3d0..5d0bab8ecb43 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild
@@ -32,6 +32,8 @@ nvkm-y += nvkm/subdev/fb/gp100.o
 nvkm-y += nvkm/subdev/fb/gp102.o
 nvkm-y += nvkm/subdev/fb/gp10b.o
 nvkm-y += nvkm/subdev/fb/gv100.o
+nvkm-y += nvkm/subdev/fb/ga100.o
+nvkm-y += nvkm/subdev/fb/ga102.o
 
 nvkm-y += nvkm/subdev/fb/ram.o
 nvkm-y += nvkm/subdev/fb/ramnv04.o
@@ -52,6 +54,7 @@ nvkm-y += nvkm/subdev/fb/ramgk104.o
 nvkm-y += nvkm/subdev/fb/ramgm107.o
 nvkm-y += nvkm/subdev/fb/ramgm200.o
 nvkm-y += nvkm/subdev/fb/ramgp100.o
+nvkm-y += nvkm/subdev/fb/ramga102.o
 nvkm-y += nvkm/subdev/fb/sddr2.o
 nvkm-y += nvkm/subdev/fb/sddr3.o
 nvkm-y += nvkm/subdev/fb/gddr3.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c
index 5940e0dea2f8..6faaea948fc4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c
@@ -122,7 +122,7 @@ nvkm_fb_oneinit(struct nvkm_subdev *subdev)
 		nvkm_debug(subdev, "%d comptags\n", tags);
 	}
 
-	return nvkm_mm_init(&fb->tags, 0, 0, tags, 1);
+	return nvkm_mm_init(&fb->tags.mm, 0, 0, tags, 1);
 }
 
 static int
@@ -205,7 +205,9 @@ nvkm_fb_dtor(struct nvkm_subdev *subdev)
 	for (i = 0; i < fb->tile.regions; i++)
 		fb->func->tile.fini(fb, i, &fb->tile.region[i]);
 
-	nvkm_mm_fini(&fb->tags);
+	nvkm_mm_fini(&fb->tags.mm);
+	mutex_destroy(&fb->tags.mutex);
+
 	nvkm_ram_del(&fb->ram);
 
 	nvkm_blob_dtor(&fb->vpr_scrubber);
@@ -225,21 +227,21 @@ nvkm_fb = {
 
 void
 nvkm_fb_ctor(const struct nvkm_fb_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_fb *fb)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_fb *fb)
 {
-	nvkm_subdev_ctor(&nvkm_fb, device, index, &fb->subdev);
+	nvkm_subdev_ctor(&nvkm_fb, device, type, inst, &fb->subdev);
 	fb->func = func;
 	fb->tile.regions = fb->func->tile.regions;
-	fb->page = nvkm_longopt(device->cfgopt, "NvFbBigPage",
-				fb->func->default_bigpage);
+	fb->page = nvkm_longopt(device->cfgopt, "NvFbBigPage", fb->func->default_bigpage);
+	mutex_init(&fb->tags.mutex);
 }
 
 int
 nvkm_fb_new_(const struct nvkm_fb_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_fb **pfb)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
 	if (!(*pfb = kzalloc(sizeof(**pfb), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_fb_ctor(func, device, index, *pfb);
+	nvkm_fb_ctor(func, device, type, inst, *pfb);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/g84.c
index 06bf95c0c549..770a4ad39122 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/g84.c
@@ -32,7 +32,7 @@ g84_fb = {
 };
 
 int
-g84_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+g84_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nv50_fb_new_(&g84_fb, device, index, pfb);
+	return nv50_fb_new_(&g84_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga100.c
new file mode 100644
index 000000000000..b47bebfbc26f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga100.c
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "gf100.h"
+#include "ram.h"
+
+static const struct nvkm_fb_func
+ga100_fb = {
+	.dtor = gf100_fb_dtor,
+	.oneinit = gf100_fb_oneinit,
+	.init = gp100_fb_init,
+	.init_page = gv100_fb_init_page,
+	.init_unkn = gp100_fb_init_unkn,
+	.ram_new = gp100_ram_new,
+	.default_bigpage = 16,
+};
+
+int
+ga100_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
+{
+	return gp102_fb_new_(&ga100_fb, device, type, inst, pfb);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c
new file mode 100644
index 000000000000..6ea7908f0563
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "gf100.h"
+#include "ram.h"
+
+static const struct nvkm_fb_func
+ga102_fb = {
+	.dtor = gf100_fb_dtor,
+	.oneinit = gf100_fb_oneinit,
+	.init = gp100_fb_init,
+	.init_page = gv100_fb_init_page,
+	.init_unkn = gp100_fb_init_unkn,
+	.ram_new = ga102_ram_new,
+	.default_bigpage = 16,
+};
+
+int
+ga102_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
+{
+	return gp102_fb_new_(&ga102_fb, device, type, inst, pfb);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c
index e8dc4e913494..9dcc40f9ef79 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c
@@ -117,13 +117,13 @@ gf100_fb_dtor(struct nvkm_fb *base)
 
 int
 gf100_fb_new_(const struct nvkm_fb_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_fb **pfb)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
 	struct gf100_fb *fb;
 
 	if (!(fb = kzalloc(sizeof(*fb), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_fb_ctor(func, device, index, &fb->base);
+	nvkm_fb_ctor(func, device, type, inst, &fb->base);
 	*pfb = &fb->base;
 
 	return 0;
@@ -141,7 +141,7 @@ gf100_fb = {
 };
 
 int
-gf100_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gf100_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gf100_fb, device, index, pfb);
+	return gf100_fb_new_(&gf100_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h
index 2ed7cdaab37c..0cac7b06acc8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h
@@ -10,8 +10,8 @@ struct gf100_fb {
 	dma_addr_t r100c10;
 };
 
-int gf100_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *,
-		  int index, struct nvkm_fb **);
+int gf100_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_fb **);
 void *gf100_fb_dtor(struct nvkm_fb *);
 void gf100_fb_init(struct nvkm_fb *);
 void gf100_fb_intr(struct nvkm_fb *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf108.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf108.c
index 4a9f463745b5..76678dd60f93 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf108.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf108.c
@@ -36,7 +36,7 @@ gf108_fb = {
 };
 
 int
-gf108_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gf108_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gf108_fb, device, index, pfb);
+	return gf100_fb_new_(&gf108_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c
index 48fd98e08baa..f73442ccb424 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c
@@ -83,7 +83,7 @@ gk104_fb = {
 };
 
 int
-gk104_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gk104_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gk104_fb, device, index, pfb);
+	return gf100_fb_new_(&gk104_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk110.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk110.c
index 0695e5dd360e..45d6cdffafee 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk110.c
@@ -65,7 +65,7 @@ gk110_fb = {
 };
 
 int
-gk110_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gk110_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gk110_fb, device, index, pfb);
+	return gf100_fb_new_(&gk110_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c
index a7e29b125094..6bc42f89d8c4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c
@@ -34,7 +34,7 @@ gk20a_fb = {
 };
 
 int
-gk20a_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gk20a_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gk20a_fb, device, index, pfb);
+	return gf100_fb_new_(&gk20a_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c
index 69c876d5d1c1..de52462a92bf 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c
@@ -36,7 +36,7 @@ gm107_fb = {
 };
 
 int
-gm107_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gm107_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gm107_fb, device, index, pfb);
+	return gf100_fb_new_(&gm107_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c
index d3b8c3367152..5acf8d15d06f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c
@@ -67,7 +67,7 @@ gm200_fb = {
 };
 
 int
-gm200_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gm200_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gm200_fb, device, index, pfb);
+	return gf100_fb_new_(&gm200_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm20b.c
index 12db61e31128..86f61a3f2fea 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm20b.c
@@ -34,7 +34,7 @@ gm20b_fb = {
 };
 
 int
-gm20b_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gm20b_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gm20b_fb, device, index, pfb);
+	return gf100_fb_new_(&gm20b_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c
index 8205ce436b3e..09e943edc362 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c
@@ -71,7 +71,7 @@ gp100_fb = {
 };
 
 int
-gp100_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gp100_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gp100_fb, device, index, pfb);
+	return gf100_fb_new_(&gp100_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp102.c
index fc8c93aa3da5..0e78b3d734a0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp102.c
@@ -114,9 +114,9 @@ gp102_fb = {
 
 int
 gp102_fb_new_(const struct nvkm_fb_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_fb **pfb)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	int ret = gf100_fb_new_(func, device, index, pfb);
+	int ret = gf100_fb_new_(func, device, type, inst, pfb);
 	if (ret)
 		return ret;
 
@@ -126,9 +126,9 @@ gp102_fb_new_(const struct nvkm_fb_func *func, struct nvkm_device *device,
 }
 
 int
-gp102_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gp102_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gp102_fb_new_(&gp102_fb, device, index, pfb);
+	return gp102_fb_new_(&gp102_fb, device, type, inst, pfb);
 }
 
 MODULE_FIRMWARE("nvidia/gp102/nvdec/scrubber.bin");
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp10b.c
index af8e43979dc1..84c9815a6d48 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp10b.c
@@ -31,7 +31,7 @@ gp10b_fb = {
 };
 
 int
-gp10b_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gp10b_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gf100_fb_new_(&gp10b_fb, device, index, pfb);
+	return gf100_fb_new_(&gp10b_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gt215.c
index 9266559b45f9..c1ec9758617c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gt215.c
@@ -32,7 +32,7 @@ gt215_fb = {
 };
 
 int
-gt215_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gt215_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nv50_fb_new_(&gt215_fb, device, index, pfb);
+	return nv50_fb_new_(&gt215_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gv100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gv100.c
index 10ff5d053f7e..63daa83ae12d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gv100.c
@@ -22,7 +22,7 @@
 #include "gf100.h"
 #include "ram.h"
 
-static int
+int
 gv100_fb_init_page(struct nvkm_fb *fb)
 {
 	return (fb->page == 16) ? 0 : -EINVAL;
@@ -42,9 +42,9 @@ gv100_fb = {
 };
 
 int
-gv100_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+gv100_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return gp102_fb_new_(&gv100_fb, device, index, pfb);
+	return gp102_fb_new_(&gv100_fb, device, type, inst, pfb);
 }
 
 MODULE_FIRMWARE("nvidia/gv100/nvdec/scrubber.bin");
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp77.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp77.c
index 73b3b86a2826..70c7b08ee0a6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp77.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp77.c
@@ -31,7 +31,7 @@ mcp77_fb = {
 };
 
 int
-mcp77_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+mcp77_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nv50_fb_new_(&mcp77_fb, device, index, pfb);
+	return nv50_fb_new_(&mcp77_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp89.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp89.c
index 6d11e32ec7ad..308d955168e8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp89.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/mcp89.c
@@ -31,7 +31,7 @@ mcp89_fb = {
 };
 
 int
-mcp89_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+mcp89_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nv50_fb_new_(&mcp89_fb, device, index, pfb);
+	return nv50_fb_new_(&mcp89_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv04.c
index c886664533c8..8d5a007ecc47 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv04.c
@@ -44,7 +44,7 @@ nv04_fb = {
 };
 
 int
-nv04_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv04_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv04_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv04_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv10.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv10.c
index c998b7e96aa3..7d2c16b27032 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv10.c
@@ -64,7 +64,7 @@ nv10_fb = {
 };
 
 int
-nv10_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv10_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv10_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv10_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv1a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv1a.c
index 7b9f04f44af8..4bdad2abd56f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv1a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv1a.c
@@ -36,7 +36,7 @@ nv1a_fb = {
 };
 
 int
-nv1a_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv1a_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv1a_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv1a_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv20.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv20.c
index a021d21ff153..d254f27f9b37 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv20.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv20.c
@@ -45,7 +45,7 @@ nv20_fb_tile_comp(struct nvkm_fb *fb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / fb->ram->parts, 0x40);
-	if (!nvkm_mm_head(&fb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
+	if (!nvkm_mm_head(&fb->tags.mm, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00000000; /* Z16 */
 		else              tile->zcomp = 0x04000000; /* Z24S8 */
 		tile->zcomp |= tile->tag->offset;
@@ -63,7 +63,7 @@ nv20_fb_tile_fini(struct nvkm_fb *fb, int i, struct nvkm_fb_tile *tile)
 	tile->limit = 0;
 	tile->pitch = 0;
 	tile->zcomp = 0;
-	nvkm_mm_free(&fb->tags, &tile->tag);
+	nvkm_mm_free(&fb->tags.mm, &tile->tag);
 }
 
 void
@@ -96,7 +96,7 @@ nv20_fb = {
 };
 
 int
-nv20_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv20_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv20_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv20_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv25.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv25.c
index 7709f5fe9a45..47da66dea6e6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv25.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv25.c
@@ -32,7 +32,7 @@ nv25_fb_tile_comp(struct nvkm_fb *fb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / fb->ram->parts, 0x40);
-	if (!nvkm_mm_head(&fb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
+	if (!nvkm_mm_head(&fb->tags.mm, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00100000; /* Z16 */
 		else              tile->zcomp = 0x00200000; /* Z24S8 */
 		tile->zcomp |= tile->tag->offset;
@@ -54,7 +54,7 @@ nv25_fb = {
 };
 
 int
-nv25_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv25_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv25_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv25_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv30.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv30.c
index 8aa782666507..0f87efb636d5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv30.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv30.c
@@ -51,7 +51,7 @@ nv30_fb_tile_comp(struct nvkm_fb *fb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / fb->ram->parts, 0x40);
-	if (!nvkm_mm_head(&fb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
+	if (!nvkm_mm_head(&fb->tags.mm, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x01000000; /* Z16 */
 		else           tile->zcomp |= 0x02000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
@@ -127,7 +127,7 @@ nv30_fb = {
 };
 
 int
-nv30_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv30_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv30_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv30_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv35.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv35.c
index 6e83dcff72e0..0694dcfd107e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv35.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv35.c
@@ -32,7 +32,7 @@ nv35_fb_tile_comp(struct nvkm_fb *fb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / fb->ram->parts, 0x40);
-	if (!nvkm_mm_head(&fb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
+	if (!nvkm_mm_head(&fb->tags.mm, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x04000000; /* Z16 */
 		else           tile->zcomp |= 0x08000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
@@ -56,7 +56,7 @@ nv35_fb = {
 };
 
 int
-nv35_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv35_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv35_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv35_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv36.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv36.c
index 2a07617bb44c..1a39770372f1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv36.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv36.c
@@ -32,7 +32,7 @@ nv36_fb_tile_comp(struct nvkm_fb *fb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / fb->ram->parts, 0x40);
-	if (!nvkm_mm_head(&fb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
+	if (!nvkm_mm_head(&fb->tags.mm, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x10000000; /* Z16 */
 		else           tile->zcomp |= 0x20000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
@@ -56,7 +56,7 @@ nv36_fb = {
 };
 
 int
-nv36_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv36_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv36_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv36_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv40.c
index 955160778b5b..77dbb9d6ba48 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv40.c
@@ -33,7 +33,7 @@ nv40_fb_tile_comp(struct nvkm_fb *fb, int i, u32 size, u32 flags,
 	u32 tiles = DIV_ROUND_UP(size, 0x80);
 	u32 tags  = round_up(tiles / fb->ram->parts, 0x100);
 	if ( (flags & 2) &&
-	    !nvkm_mm_head(&fb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
+	    !nvkm_mm_head(&fb->tags.mm, 0, 1, tags, tags, 1, &tile->tag)) {
 		tile->zcomp  = 0x28000000; /* Z24S8_SPLIT_GRAD */
 		tile->zcomp |= ((tile->tag->offset           ) >> 8);
 		tile->zcomp |= ((tile->tag->offset + tags - 1) >> 8) << 13;
@@ -62,7 +62,7 @@ nv40_fb = {
 };
 
 int
-nv40_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv40_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv40_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv40_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv41.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv41.c
index b77f08d34cc3..0f9d9e48e7ad 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv41.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv41.c
@@ -56,7 +56,7 @@ nv41_fb = {
 };
 
 int
-nv41_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv41_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv41_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv41_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv44.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv44.c
index b59dc486083d..b1046ee9f0ea 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv44.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv44.c
@@ -65,7 +65,7 @@ nv44_fb = {
 };
 
 int
-nv44_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv44_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv44_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv44_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv46.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv46.c
index cab7d20fa039..0d78de422dfa 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv46.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv46.c
@@ -51,7 +51,7 @@ nv46_fb = {
 };
 
 int
-nv46_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv46_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv46_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv46_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv47.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv47.c
index a8b0ad4c871d..5cedde29c8ee 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv47.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv47.c
@@ -39,7 +39,7 @@ nv47_fb = {
 };
 
 int
-nv47_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv47_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv47_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv47_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv49.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv49.c
index d0b317bb0252..95cc099603d8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv49.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv49.c
@@ -39,7 +39,7 @@ nv49_fb = {
 };
 
 int
-nv49_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv49_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv49_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv49_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv4e.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv4e.c
index 6a6f0c086071..c9f3148f4e75 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv4e.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv4e.c
@@ -37,7 +37,7 @@ nv4e_fb = {
 };
 
 int
-nv4e_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv4e_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nvkm_fb_new_(&nv4e_fb, device, index, pfb);
+	return nvkm_fb_new_(&nv4e_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.c
index b2f5bf8144ea..95fd8f834010 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.c
@@ -262,16 +262,15 @@ nv50_fb_ = {
 
 int
 nv50_fb_new_(const struct nv50_fb_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_fb **pfb)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
 	struct nv50_fb *fb;
 
 	if (!(fb = kzalloc(sizeof(*fb), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_fb_ctor(&nv50_fb_, device, index, &fb->base);
+	nvkm_fb_ctor(&nv50_fb_, device, type, inst, &fb->base);
 	fb->func = func;
 	*pfb = &fb->base;
-
 	return 0;
 }
 
@@ -283,7 +282,7 @@ nv50_fb = {
 };
 
 int
-nv50_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+nv50_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb)
 {
-	return nv50_fb_new_(&nv50_fb, device, index, pfb);
+	return nv50_fb_new_(&nv50_fb, device, type, inst, pfb);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.h
index 5e2b0c9539ed..a5e673859a90 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/nv50.h
@@ -17,6 +17,6 @@ struct nv50_fb_func {
 	u32 trap;
 };
 
-int nv50_fb_new_(const struct nv50_fb_func *, struct nvkm_device *, int index,
+int nv50_fb_new_(const struct nv50_fb_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		 struct nvkm_fb **pfb);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h
index 5be9c563350d..3f1be9780c65 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h
@@ -38,9 +38,9 @@ struct nvkm_fb_func {
 };
 
 void nvkm_fb_ctor(const struct nvkm_fb_func *, struct nvkm_device *device,
-		  int index, struct nvkm_fb *);
+		  enum nvkm_subdev_type type, int inst, struct nvkm_fb *);
 int nvkm_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *device,
-		 int index, struct nvkm_fb **);
+		 enum nvkm_subdev_type type, int inst, struct nvkm_fb **);
 int nvkm_fb_bios_memtype(struct nvkm_bios *);
 
 void nv10_fb_tile_init(struct nvkm_fb *, int i, u32 addr, u32 size,
@@ -78,8 +78,10 @@ int gm200_fb_init_page(struct nvkm_fb *);
 void gp100_fb_init_remapper(struct nvkm_fb *);
 void gp100_fb_init_unkn(struct nvkm_fb *);
 
-int gp102_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *, int,
+int gp102_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  struct nvkm_fb **);
 bool gp102_fb_vpr_scrub_required(struct nvkm_fb *);
 int gp102_fb_vpr_scrub(struct nvkm_fb *);
+
+int gv100_fb_init_page(struct nvkm_fb *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c
index b11867f682cb..03b1bdb27770 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c
@@ -81,12 +81,12 @@ nvkm_vram_dtor(struct nvkm_memory *memory)
 	struct nvkm_vram *vram = nvkm_vram(memory);
 	struct nvkm_mm_node *next = vram->mn;
 	struct nvkm_mm_node *node;
-	mutex_lock(&vram->ram->fb->subdev.mutex);
+	mutex_lock(&vram->ram->mutex);
 	while ((node = next)) {
 		next = node->next;
 		nvkm_mm_free(&vram->ram->vram, &node);
 	}
-	mutex_unlock(&vram->ram->fb->subdev.mutex);
+	mutex_unlock(&vram->ram->mutex);
 	return vram;
 }
 
@@ -126,7 +126,7 @@ nvkm_ram_get(struct nvkm_device *device, u8 heap, u8 type, u8 rpage, u64 size,
 	vram->page = page;
 	*pmemory = &vram->memory;
 
-	mutex_lock(&ram->fb->subdev.mutex);
+	mutex_lock(&ram->mutex);
 	node = &vram->mn;
 	do {
 		if (back)
@@ -134,7 +134,7 @@ nvkm_ram_get(struct nvkm_device *device, u8 heap, u8 type, u8 rpage, u64 size,
 		else
 			ret = nvkm_mm_head(mm, heap, type, max, min, align, &r);
 		if (ret) {
-			mutex_unlock(&ram->fb->subdev.mutex);
+			mutex_unlock(&ram->mutex);
 			nvkm_memory_unref(pmemory);
 			return ret;
 		}
@@ -143,7 +143,7 @@ nvkm_ram_get(struct nvkm_device *device, u8 heap, u8 type, u8 rpage, u64 size,
 		node = &r->next;
 		max -= r->length;
 	} while (max);
-	mutex_unlock(&ram->fb->subdev.mutex);
+	mutex_unlock(&ram->mutex);
 	return 0;
 }
 
@@ -163,6 +163,7 @@ nvkm_ram_del(struct nvkm_ram **pram)
 		if (ram->func->dtor)
 			*pram = ram->func->dtor(ram);
 		nvkm_mm_fini(&ram->vram);
+		mutex_destroy(&ram->mutex);
 		kfree(*pram);
 		*pram = NULL;
 	}
@@ -196,6 +197,7 @@ nvkm_ram_ctor(const struct nvkm_ram_func *func, struct nvkm_fb *fb,
 	ram->fb = fb;
 	ram->type = type;
 	ram->size = size;
+	mutex_init(&ram->mutex);
 
 	if (!nvkm_mm_initialised(&ram->vram)) {
 		ret = nvkm_mm_init(&ram->vram, NVKM_RAM_MM_NORMAL, 0,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h
index d723a9b4e3c4..ea7d66f3dd82 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h
@@ -70,4 +70,5 @@ int gk104_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 int gm107_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 int gm200_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 int gp100_ram_new(struct nvkm_fb *, struct nvkm_ram **);
+int ga102_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramga102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramga102.c
new file mode 100644
index 000000000000..298c136cefe0
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramga102.c
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "ram.h"
+
+#include <subdev/bios.h>
+#include <subdev/bios/init.h>
+#include <subdev/bios/rammap.h>
+
+static const struct nvkm_ram_func
+ga102_ram = {
+};
+
+int
+ga102_ram_new(struct nvkm_fb *fb, struct nvkm_ram **pram)
+{
+	struct nvkm_device *device = fb->subdev.device;
+	enum nvkm_ram_type type = nvkm_fb_bios_memtype(device->bios);
+	u32 size = nvkm_rd32(device, 0x1183a4);
+
+	return nvkm_ram_new_(&ga102_ram, fb, type, (u64)size << 20, pram);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgk104.c
index d350d92852d2..2b678b60b4d3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgk104.c
@@ -260,7 +260,7 @@ gk104_ram_calc_gddr5(struct gk104_ram *ram, u32 freq)
 	ram_mask(fuc, 0x10f808, 0x40000000, 0x40000000);
 	ram_block(fuc);
 
-	if (nvkm_device_engine(ram->base.fb->subdev.device, NVKM_ENGINE_DISP))
+	if (ram->base.fb->subdev.device->disp)
 		ram_wr32(fuc, 0x62c000, 0x0f0f0000);
 
 	/* MR1: turn termination on early, for some reason.. */
@@ -661,7 +661,7 @@ gk104_ram_calc_gddr5(struct gk104_ram *ram, u32 freq)
 
 	ram_unblock(fuc);
 
-	if (nvkm_device_engine(ram->base.fb->subdev.device, NVKM_ENGINE_DISP))
+	if (ram->base.fb->subdev.device->disp)
 		ram_wr32(fuc, 0x62c000, 0x0f0f0f00);
 
 	if (next->bios.rammap_11_08_01)
@@ -711,7 +711,7 @@ gk104_ram_calc_sddr3(struct gk104_ram *ram, u32 freq)
 	ram_mask(fuc, 0x10f808, 0x40000000, 0x40000000);
 	ram_block(fuc);
 
-	if (nvkm_device_engine(ram->base.fb->subdev.device, NVKM_ENGINE_DISP))
+	if (ram->base.fb->subdev.device->disp)
 		ram_wr32(fuc, 0x62c000, 0x0f0f0000);
 
 	if (vc == 1 && ram_have(fuc, gpio2E)) {
@@ -943,7 +943,7 @@ gk104_ram_calc_sddr3(struct gk104_ram *ram, u32 freq)
 
 	ram_unblock(fuc);
 
-	if (nvkm_device_engine(ram->base.fb->subdev.device, NVKM_ENGINE_DISP))
+	if (ram->base.fb->subdev.device->disp)
 		ram_wr32(fuc, 0x62c000, 0x0f0f0f00);
 
 	if (next->bios.rammap_11_08_01)
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/base.c
index 1c3c18ea8ced..375dfce09f84 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/base.c
@@ -42,12 +42,12 @@ nvkm_fuse = {
 
 int
 nvkm_fuse_new_(const struct nvkm_fuse_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_fuse **pfuse)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_fuse **pfuse)
 {
 	struct nvkm_fuse *fuse;
 	if (!(fuse = *pfuse = kzalloc(sizeof(*fuse), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_fuse, device, index, &fuse->subdev);
+	nvkm_subdev_ctor(&nvkm_fuse, device, type, inst, &fuse->subdev);
 	fuse->func = func;
 	spin_lock_init(&fuse->lock);
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gf100.c
index 13671fedc805..01f770654b1d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gf100.c
@@ -47,7 +47,8 @@ gf100_fuse = {
 };
 
 int
-gf100_fuse_new(struct nvkm_device *device, int index, struct nvkm_fuse **pfuse)
+gf100_fuse_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fuse **pfuse)
 {
-	return nvkm_fuse_new_(&gf100_fuse, device, index, pfuse);
+	return nvkm_fuse_new_(&gf100_fuse, device, type, inst, pfuse);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gm107.c
index 9aff4ea04506..7dc99492f536 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/gm107.c
@@ -36,7 +36,8 @@ gm107_fuse = {
 };
 
 int
-gm107_fuse_new(struct nvkm_device *device, int index, struct nvkm_fuse **pfuse)
+gm107_fuse_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_fuse **pfuse)
 {
-	return nvkm_fuse_new_(&gm107_fuse, device, index, pfuse);
+	return nvkm_fuse_new_(&gm107_fuse, device, type, inst, pfuse);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/nv50.c
index 514c193db25d..2505e8e1c1d3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/nv50.c
@@ -45,7 +45,8 @@ nv50_fuse = {
 };
 
 int
-nv50_fuse_new(struct nvkm_device *device, int index, struct nvkm_fuse **pfuse)
+nv50_fuse_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_fuse **pfuse)
 {
-	return nvkm_fuse_new_(&nv50_fuse, device, index, pfuse);
+	return nvkm_fuse_new_(&nv50_fuse, device, type, inst, pfuse);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/priv.h
index 2edc612408dd..e83d0c30dff6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fuse/priv.h
@@ -8,6 +8,6 @@ struct nvkm_fuse_func {
 	u32 (*read)(struct nvkm_fuse *, u32 addr);
 };
 
-int nvkm_fuse_new_(const struct nvkm_fuse_func *, struct nvkm_device *,
-		   int index, struct nvkm_fuse **);
+int nvkm_fuse_new_(const struct nvkm_fuse_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_fuse **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/Kbuild
index b2ad5922a1c2..efbbaa080de5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/Kbuild
@@ -5,3 +5,4 @@ nvkm-y += nvkm/subdev/gpio/nv50.o
 nvkm-y += nvkm/subdev/gpio/g94.o
 nvkm-y += nvkm/subdev/gpio/gf119.o
 nvkm-y += nvkm/subdev/gpio/gk104.o
+nvkm-y += nvkm/subdev/gpio/ga102.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/base.c
index 914276410ef8..048bcc70c3f4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/base.c
@@ -241,14 +241,14 @@ nvkm_gpio = {
 
 int
 nvkm_gpio_new_(const struct nvkm_gpio_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_gpio **pgpio)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_gpio **pgpio)
 {
 	struct nvkm_gpio *gpio;
 
 	if (!(gpio = *pgpio = kzalloc(sizeof(*gpio), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_subdev_ctor(&nvkm_gpio, device, index, &gpio->subdev);
+	nvkm_subdev_ctor(&nvkm_gpio, device, type, inst, &gpio->subdev);
 	gpio->func = func;
 
 	return nvkm_event_init(&nvkm_gpio_intr_func, 2, func->lines,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/g94.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/g94.c
index 6dcda55fb865..114728ccdf8e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/g94.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/g94.c
@@ -68,7 +68,8 @@ g94_gpio = {
 };
 
 int
-g94_gpio_new(struct nvkm_device *device, int index, struct nvkm_gpio **pgpio)
+g94_gpio_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_gpio **pgpio)
 {
-	return nvkm_gpio_new_(&g94_gpio, device, index, pgpio);
+	return nvkm_gpio_new_(&g94_gpio, device, type, inst, pgpio);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/ga102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/ga102.c
new file mode 100644
index 000000000000..4a96f926b66d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/ga102.c
@@ -0,0 +1,119 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "priv.h"
+
+static void
+ga102_gpio_reset(struct nvkm_gpio *gpio, u8 match)
+{
+	struct nvkm_device *device = gpio->subdev.device;
+	struct nvkm_bios *bios = device->bios;
+	u8 ver, len;
+	u16 entry;
+	int ent = -1;
+
+	while ((entry = dcb_gpio_entry(bios, 0, ++ent, &ver, &len))) {
+		u32 data = nvbios_rd32(bios, entry);
+		u8  line =   (data & 0x0000003f);
+		u8  defs = !!(data & 0x00000080);
+		u8  func =   (data & 0x0000ff00) >> 8;
+		u8  unk0 =   (data & 0x00ff0000) >> 16;
+		u8  unk1 =   (data & 0x1f000000) >> 24;
+
+		if ( func  == DCB_GPIO_UNUSED ||
+		    (match != DCB_GPIO_UNUSED && match != func))
+			continue;
+
+		nvkm_gpio_set(gpio, 0, func, line, defs);
+
+		nvkm_mask(device, 0x021200 + (line * 4), 0xff, unk0);
+		if (unk1--)
+			nvkm_mask(device, 0x00d740 + (unk1 * 4), 0xff, line);
+	}
+}
+
+static int
+ga102_gpio_drive(struct nvkm_gpio *gpio, int line, int dir, int out)
+{
+	struct nvkm_device *device = gpio->subdev.device;
+	u32 data = ((dir ^ 1) << 13) | (out << 12);
+	nvkm_mask(device, 0x021200 + (line * 4), 0x00003000, data);
+	nvkm_mask(device, 0x00d604, 0x00000001, 0x00000001); /* update? */
+	return 0;
+}
+
+static int
+ga102_gpio_sense(struct nvkm_gpio *gpio, int line)
+{
+	struct nvkm_device *device = gpio->subdev.device;
+	return !!(nvkm_rd32(device, 0x021200 + (line * 4)) & 0x00004000);
+}
+
+static void
+ga102_gpio_intr_stat(struct nvkm_gpio *gpio, u32 *hi, u32 *lo)
+{
+	struct nvkm_device *device = gpio->subdev.device;
+	u32 intr0 = nvkm_rd32(device, 0x021640);
+	u32 intr1 = nvkm_rd32(device, 0x02164c);
+	u32 stat0 = nvkm_rd32(device, 0x021648) & intr0;
+	u32 stat1 = nvkm_rd32(device, 0x021654) & intr1;
+	*lo = (stat1 & 0xffff0000) | (stat0 >> 16);
+	*hi = (stat1 << 16) | (stat0 & 0x0000ffff);
+	nvkm_wr32(device, 0x021640, intr0);
+	nvkm_wr32(device, 0x02164c, intr1);
+}
+
+static void
+ga102_gpio_intr_mask(struct nvkm_gpio *gpio, u32 type, u32 mask, u32 data)
+{
+	struct nvkm_device *device = gpio->subdev.device;
+	u32 inte0 = nvkm_rd32(device, 0x021648);
+	u32 inte1 = nvkm_rd32(device, 0x021654);
+	if (type & NVKM_GPIO_LO)
+		inte0 = (inte0 & ~(mask << 16)) | (data << 16);
+	if (type & NVKM_GPIO_HI)
+		inte0 = (inte0 & ~(mask & 0xffff)) | (data & 0xffff);
+	mask >>= 16;
+	data >>= 16;
+	if (type & NVKM_GPIO_LO)
+		inte1 = (inte1 & ~(mask << 16)) | (data << 16);
+	if (type & NVKM_GPIO_HI)
+		inte1 = (inte1 & ~mask) | data;
+	nvkm_wr32(device, 0x021648, inte0);
+	nvkm_wr32(device, 0x021654, inte1);
+}
+
+static const struct nvkm_gpio_func
+ga102_gpio = {
+	.lines = 32,
+	.intr_stat = ga102_gpio_intr_stat,
+	.intr_mask = ga102_gpio_intr_mask,
+	.drive = ga102_gpio_drive,
+	.sense = ga102_gpio_sense,
+	.reset = ga102_gpio_reset,
+};
+
+int
+ga102_gpio_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_gpio **pgpio)
+{
+	return nvkm_gpio_new_(&ga102_gpio, device, type, inst, pgpio);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gf119.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gf119.c
index bb7400dfaef8..ecb19e4f5c48 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gf119.c
@@ -80,7 +80,8 @@ gf119_gpio = {
 };
 
 int
-gf119_gpio_new(struct nvkm_device *device, int index, struct nvkm_gpio **pgpio)
+gf119_gpio_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_gpio **pgpio)
 {
-	return nvkm_gpio_new_(&gf119_gpio, device, index, pgpio);
+	return nvkm_gpio_new_(&gf119_gpio, device, type, inst, pgpio);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gk104.c
index 2ead515b8530..c0e4cdb45520 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/gk104.c
@@ -68,7 +68,8 @@ gk104_gpio = {
 };
 
 int
-gk104_gpio_new(struct nvkm_device *device, int index, struct nvkm_gpio **pgpio)
+gk104_gpio_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_gpio **pgpio)
 {
-	return nvkm_gpio_new_(&gk104_gpio, device, index, pgpio);
+	return nvkm_gpio_new_(&gk104_gpio, device, type, inst, pgpio);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv10.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv10.c
index ae3499b48330..48ad29b5638f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv10.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv10.c
@@ -112,7 +112,8 @@ nv10_gpio = {
 };
 
 int
-nv10_gpio_new(struct nvkm_device *device, int index, struct nvkm_gpio **pgpio)
+nv10_gpio_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_gpio **pgpio)
 {
-	return nvkm_gpio_new_(&nv10_gpio, device, index, pgpio);
+	return nvkm_gpio_new_(&nv10_gpio, device, type, inst, pgpio);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv50.c
index 73923fd5f7f2..b86c49762f11 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/nv50.c
@@ -126,7 +126,8 @@ nv50_gpio = {
 };
 
 int
-nv50_gpio_new(struct nvkm_device *device, int index, struct nvkm_gpio **pgpio)
+nv50_gpio_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_gpio **pgpio)
 {
-	return nvkm_gpio_new_(&nv50_gpio, device, index, pgpio);
+	return nvkm_gpio_new_(&nv50_gpio, device, type, inst, pgpio);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/priv.h
index 59e39affe2a0..6590d81164e7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gpio/priv.h
@@ -28,8 +28,8 @@ struct nvkm_gpio_func {
 	void (*reset)(struct nvkm_gpio *, u8);
 };
 
-int nvkm_gpio_new_(const struct nvkm_gpio_func *, struct nvkm_device *,
-		   int index, struct nvkm_gpio **);
+int nvkm_gpio_new_(const struct nvkm_gpio_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_gpio **);
 
 void nv50_gpio_reset(struct nvkm_gpio *, u8);
 int  nv50_gpio_drive(struct nvkm_gpio *, int, int, int);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c
index 5a32df0f9992..22574886b819 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c
@@ -40,20 +40,18 @@ nvkm_gsp = {
 
 int
 nvkm_gsp_new_(const struct nvkm_gsp_fwif *fwif, struct nvkm_device *device,
-	      int index, struct nvkm_gsp **pgsp)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_gsp **pgsp)
 {
 	struct nvkm_gsp *gsp;
 
 	if (!(gsp = *pgsp = kzalloc(sizeof(*gsp), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_subdev_ctor(&nvkm_gsp, device, index, &gsp->subdev);
+	nvkm_subdev_ctor(&nvkm_gsp, device, type, inst, &gsp->subdev);
 
 	fwif = nvkm_firmware_load(&gsp->subdev, fwif, "Gsp", gsp);
 	if (IS_ERR(fwif))
 		return PTR_ERR(fwif);
 
-	return nvkm_falcon_ctor(fwif->flcn, &gsp->subdev,
-				nvkm_subdev_name[gsp->subdev.index], 0,
-				&gsp->falcon);
+	return nvkm_falcon_ctor(fwif->flcn, &gsp->subdev, gsp->subdev.name, 0, &gsp->falcon);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gv100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gv100.c
index 2114f9b00a28..2ac7fc934c09 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gv100.c
@@ -49,7 +49,8 @@ gv100_gsp[] = {
 };
 
 int
-gv100_gsp_new(struct nvkm_device *device, int index, struct nvkm_gsp **pgsp)
+gv100_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_gsp **pgsp)
 {
-	return nvkm_gsp_new_(gv100_gsp, device, index, pgsp);
+	return nvkm_gsp_new_(gv100_gsp, device, type, inst, pgsp);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h
index 92820fb997c1..19381ddd38d4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h
@@ -10,6 +10,6 @@ struct nvkm_gsp_fwif {
 	const struct nvkm_falcon_func *flcn;
 };
 
-int nvkm_gsp_new_(const struct nvkm_gsp_fwif *, struct nvkm_device *, int,
+int nvkm_gsp_new_(const struct nvkm_gsp_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  struct nvkm_gsp **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/Kbuild
index 723d0284caef..819703913a00 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/Kbuild
@@ -7,6 +7,7 @@ nvkm-y += nvkm/subdev/i2c/g94.o
 nvkm-y += nvkm/subdev/i2c/gf117.o
 nvkm-y += nvkm/subdev/i2c/gf119.o
 nvkm-y += nvkm/subdev/i2c/gk104.o
+nvkm-y += nvkm/subdev/i2c/gk110.o
 nvkm-y += nvkm/subdev/i2c/gm200.o
 
 nvkm-y += nvkm/subdev/i2c/pad.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/aux.h b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/aux.h
index 30b48896965e..f920eabf8628 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/aux.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/aux.h
@@ -3,6 +3,13 @@
 #define __NVKM_I2C_AUX_H__
 #include "pad.h"
 
+static inline void
+nvkm_i2c_aux_autodpcd(struct nvkm_i2c *i2c, int aux, bool enable)
+{
+	if (i2c->func->aux_autodpcd)
+		i2c->func->aux_autodpcd(i2c, aux, false);
+}
+
 struct nvkm_i2c_aux_func {
 	bool address_only;
 	int  (*xfer)(struct nvkm_i2c_aux *, bool retry, u8 type,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxg94.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxg94.c
index db7769cb33eb..47068f6f9c55 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxg94.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxg94.c
@@ -77,7 +77,8 @@ g94_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 		 u8 type, u32 addr, u8 *data, u8 *size)
 {
 	struct g94_i2c_aux *aux = g94_i2c_aux(obj);
-	struct nvkm_device *device = aux->base.pad->i2c->subdev.device;
+	struct nvkm_i2c *i2c = aux->base.pad->i2c;
+	struct nvkm_device *device = i2c->subdev.device;
 	const u32 base = aux->ch * 0x50;
 	u32 ctrl, stat, timeout, retries = 0;
 	u32 xbuf[4] = {};
@@ -96,6 +97,8 @@ g94_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 		goto out;
 	}
 
+	nvkm_i2c_aux_autodpcd(i2c, aux->ch, false);
+
 	if (!(type & 1)) {
 		memcpy(xbuf, data, *size);
 		for (i = 0; i < 16; i += 4) {
@@ -128,7 +131,7 @@ g94_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 			if (!timeout--) {
 				AUX_ERR(&aux->base, "timeout %08x", ctrl);
 				ret = -EIO;
-				goto out;
+				goto out_err;
 			}
 		} while (ctrl & 0x00010000);
 		ret = 0;
@@ -154,7 +157,8 @@ g94_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 		memcpy(data, xbuf, *size);
 		*size = stat & 0x0000001f;
 	}
-
+out_err:
+	nvkm_i2c_aux_autodpcd(i2c, aux->ch, true);
 out:
 	g94_i2c_aux_fini(aux);
 	return ret < 0 ? ret : (stat & 0x000f0000) >> 16;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxgm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxgm200.c
index edb6148cbca0..8bd1d442e465 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxgm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/auxgm200.c
@@ -33,7 +33,7 @@ static void
 gm200_i2c_aux_fini(struct gm200_i2c_aux *aux)
 {
 	struct nvkm_device *device = aux->base.pad->i2c->subdev.device;
-	nvkm_mask(device, 0x00d954 + (aux->ch * 0x50), 0x00310000, 0x00000000);
+	nvkm_mask(device, 0x00d954 + (aux->ch * 0x50), 0x00710000, 0x00000000);
 }
 
 static int
@@ -54,10 +54,10 @@ gm200_i2c_aux_init(struct gm200_i2c_aux *aux)
 			AUX_ERR(&aux->base, "begin idle timeout %08x", ctrl);
 			return -EBUSY;
 		}
-	} while (ctrl & 0x03010000);
+	} while (ctrl & 0x07010000);
 
 	/* set some magic, and wait up to 1ms for it to appear */
-	nvkm_mask(device, 0x00d954 + (aux->ch * 0x50), 0x00300000, ureq);
+	nvkm_mask(device, 0x00d954 + (aux->ch * 0x50), 0x00700000, ureq);
 	timeout = 1000;
 	do {
 		ctrl = nvkm_rd32(device, 0x00d954 + (aux->ch * 0x50));
@@ -67,7 +67,7 @@ gm200_i2c_aux_init(struct gm200_i2c_aux *aux)
 			gm200_i2c_aux_fini(aux);
 			return -EBUSY;
 		}
-	} while ((ctrl & 0x03000000) != urep);
+	} while ((ctrl & 0x07000000) != urep);
 
 	return 0;
 }
@@ -77,7 +77,8 @@ gm200_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 		   u8 type, u32 addr, u8 *data, u8 *size)
 {
 	struct gm200_i2c_aux *aux = gm200_i2c_aux(obj);
-	struct nvkm_device *device = aux->base.pad->i2c->subdev.device;
+	struct nvkm_i2c *i2c = aux->base.pad->i2c;
+	struct nvkm_device *device = i2c->subdev.device;
 	const u32 base = aux->ch * 0x50;
 	u32 ctrl, stat, timeout, retries = 0;
 	u32 xbuf[4] = {};
@@ -96,6 +97,8 @@ gm200_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 		goto out;
 	}
 
+	nvkm_i2c_aux_autodpcd(i2c, aux->ch, false);
+
 	if (!(type & 1)) {
 		memcpy(xbuf, data, *size);
 		for (i = 0; i < 16; i += 4) {
@@ -128,7 +131,7 @@ gm200_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 			if (!timeout--) {
 				AUX_ERR(&aux->base, "timeout %08x", ctrl);
 				ret = -EIO;
-				goto out;
+				goto out_err;
 			}
 		} while (ctrl & 0x00010000);
 		ret = 0;
@@ -155,6 +158,8 @@ gm200_i2c_aux_xfer(struct nvkm_i2c_aux *obj, bool retry,
 		*size = stat & 0x0000001f;
 	}
 
+out_err:
+	nvkm_i2c_aux_autodpcd(i2c, aux->ch, true);
 out:
 	gm200_i2c_aux_fini(aux);
 	return ret < 0 ? ret : (stat & 0x000f0000) >> 16;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/base.c
index 719345074711..cb5cb533d91c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/base.c
@@ -277,7 +277,7 @@ nvkm_i2c_drv[] = {
 
 int
 nvkm_i2c_new_(const struct nvkm_i2c_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_i2c **pi2c)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_i2c **pi2c)
 {
 	struct nvkm_bios *bios = device->bios;
 	struct nvkm_i2c *i2c;
@@ -289,7 +289,7 @@ nvkm_i2c_new_(const struct nvkm_i2c_func *func, struct nvkm_device *device,
 	if (!(i2c = *pi2c = kzalloc(sizeof(*i2c), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_subdev_ctor(&nvkm_i2c, device, index, &i2c->subdev);
+	nvkm_subdev_ctor(&nvkm_i2c, device, type, inst, &i2c->subdev);
 	i2c->func = func;
 	INIT_LIST_HEAD(&i2c->pad);
 	INIT_LIST_HEAD(&i2c->bus);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/g94.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/g94.c
index bb2a31d88161..e5bad085c06f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/g94.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/g94.c
@@ -66,7 +66,8 @@ g94_i2c = {
 };
 
 int
-g94_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+g94_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&g94_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&g94_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf117.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf117.c
index ae4aad3fcd2e..cda30ee6767d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf117.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf117.c
@@ -30,7 +30,8 @@ gf117_i2c = {
 };
 
 int
-gf117_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+gf117_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&gf117_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&gf117_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf119.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf119.c
index 6f2b02af42c8..e9c6a6cca09d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gf119.c
@@ -34,7 +34,8 @@ gf119_i2c = {
 };
 
 int
-gf119_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+gf119_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&gf119_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&gf119_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk104.c
index f9f6bf4b66c9..d35aa6fe3015 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk104.c
@@ -66,7 +66,8 @@ gk104_i2c = {
 };
 
 int
-gk104_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+gk104_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&gk104_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&gk104_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk110.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk110.c
new file mode 100644
index 000000000000..9fec6af56e07
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gk110.c
@@ -0,0 +1,46 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "priv.h"
+#include "pad.h"
+
+static void
+gk110_aux_autodpcd(struct nvkm_i2c *i2c, int aux, bool enable)
+{
+	nvkm_mask(i2c->subdev.device, 0x00e4f8 + (aux * 0x50), 0x00010000, enable << 16);
+}
+
+static const struct nvkm_i2c_func
+gk110_i2c = {
+	.pad_x_new = gf119_i2c_pad_x_new,
+	.pad_s_new = gf119_i2c_pad_s_new,
+	.aux = 4,
+	.aux_stat = gk104_aux_stat,
+	.aux_mask = gk104_aux_mask,
+	.aux_autodpcd = gk110_aux_autodpcd,
+};
+
+int
+gk110_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_i2c **pi2c)
+{
+	return nvkm_i2c_new_(&gk110_i2c, device, type, inst, pi2c);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gm200.c
index a23c5f315221..46917eb600f9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/gm200.c
@@ -24,6 +24,12 @@
 #include "priv.h"
 #include "pad.h"
 
+static void
+gm200_aux_autodpcd(struct nvkm_i2c *i2c, int aux, bool enable)
+{
+	nvkm_mask(i2c->subdev.device, 0x00d968 + (aux * 0x50), 0x00010000, enable << 16);
+}
+
 static const struct nvkm_i2c_func
 gm200_i2c = {
 	.pad_x_new = gf119_i2c_pad_x_new,
@@ -31,10 +37,12 @@ gm200_i2c = {
 	.aux = 8,
 	.aux_stat = gk104_aux_stat,
 	.aux_mask = gk104_aux_mask,
+	.aux_autodpcd = gm200_aux_autodpcd,
 };
 
 int
-gm200_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+gm200_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&gm200_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&gm200_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv04.c
index 18776f49355c..ecfcf147c789 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv04.c
@@ -30,7 +30,8 @@ nv04_i2c = {
 };
 
 int
-nv04_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+nv04_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&nv04_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&nv04_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv4e.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv4e.c
index 6b762f7cee9e..ad1d3fd2bcbc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv4e.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv4e.c
@@ -30,7 +30,8 @@ nv4e_i2c = {
 };
 
 int
-nv4e_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+nv4e_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&nv4e_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&nv4e_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv50.c
index 75640ab97d6a..2f94bed2c056 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/nv50.c
@@ -30,7 +30,8 @@ nv50_i2c = {
 };
 
 int
-nv50_i2c_new(struct nvkm_device *device, int index, struct nvkm_i2c **pi2c)
+nv50_i2c_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_i2c **pi2c)
 {
-	return nvkm_i2c_new_(&nv50_i2c, device, index, pi2c);
+	return nvkm_i2c_new_(&nv50_i2c, device, type, inst, pi2c);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/pad.h b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/pad.h
index 461016814f4f..44b7bb7d4777 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/pad.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/pad.h
@@ -1,7 +1,7 @@
 /* SPDX-License-Identifier: MIT */
 #ifndef __NVKM_I2C_PAD_H__
 #define __NVKM_I2C_PAD_H__
-#include <subdev/i2c.h>
+#include "priv.h"
 
 struct nvkm_i2c_pad {
 	const struct nvkm_i2c_pad_func *func;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/priv.h
index bd86bc298ebe..f9d79f72f7e7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/i2c/priv.h
@@ -4,8 +4,8 @@
 #define nvkm_i2c(p) container_of((p), struct nvkm_i2c, subdev)
 #include <subdev/i2c.h>
 
-int nvkm_i2c_new_(const struct nvkm_i2c_func *, struct nvkm_device *,
-		  int index, struct nvkm_i2c **);
+int nvkm_i2c_new_(const struct nvkm_i2c_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_i2c **);
 
 struct nvkm_i2c_func {
 	int (*pad_x_new)(struct nvkm_i2c *, int id, struct nvkm_i2c_pad **);
@@ -23,6 +23,10 @@ struct nvkm_i2c_func {
 	/* mask on/off interrupt types for a given set of auxch
 	 */
 	void (*aux_mask)(struct nvkm_i2c *, u32, u32, u32);
+
+	/* enable/disable HW-initiated DPCD reads
+	 */
+	void (*aux_autodpcd)(struct nvkm_i2c *, int aux, bool enable);
 };
 
 void g94_aux_stat(struct nvkm_i2c *, u32 *, u32 *, u32 *, u32 *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/Kbuild
deleted file mode 100644
index 127efb51f67d..000000000000
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/Kbuild
+++ /dev/null
@@ -1,7 +0,0 @@
-# SPDX-License-Identifier: MIT
-nvkm-y += nvkm/subdev/ibus/gf100.o
-nvkm-y += nvkm/subdev/ibus/gf117.o
-nvkm-y += nvkm/subdev/ibus/gk104.o
-nvkm-y += nvkm/subdev/ibus/gk20a.o
-nvkm-y += nvkm/subdev/ibus/gm200.o
-nvkm-y += nvkm/subdev/ibus/gp10b.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/priv.h
deleted file mode 100644
index 302d69e384d8..000000000000
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/priv.h
+++ /dev/null
@@ -1,9 +0,0 @@
-/* SPDX-License-Identifier: MIT */
-#ifndef __NVKM_IBUS_PRIV_H__
-#define __NVKM_IBUS_PRIV_H__
-
-#include <subdev/ibus.h>
-
-void gf100_ibus_intr(struct nvkm_subdev *);
-void gk104_ibus_intr(struct nvkm_subdev *);
-#endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c
index fecfa6afcf54..8f0ccd3664eb 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c
@@ -312,20 +312,20 @@ iccsense_func = {
 };
 
 void
-nvkm_iccsense_ctor(struct nvkm_device *device, int index,
+nvkm_iccsense_ctor(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		   struct nvkm_iccsense *iccsense)
 {
-	nvkm_subdev_ctor(&iccsense_func, device, index, &iccsense->subdev);
+	nvkm_subdev_ctor(&iccsense_func, device, type, inst, &iccsense->subdev);
 }
 
 int
-nvkm_iccsense_new_(struct nvkm_device *device, int index,
+nvkm_iccsense_new_(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		   struct nvkm_iccsense **iccsense)
 {
 	if (!(*iccsense = kzalloc(sizeof(**iccsense), GFP_KERNEL)))
 		return -ENOMEM;
 	INIT_LIST_HEAD(&(*iccsense)->sensors);
 	INIT_LIST_HEAD(&(*iccsense)->rails);
-	nvkm_iccsense_ctor(device, index, *iccsense);
+	nvkm_iccsense_ctor(device, type, inst, *iccsense);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/gf100.c
index cccff1c8a409..3eabf4944395 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/gf100.c
@@ -24,8 +24,8 @@
 #include "priv.h"
 
 int
-gf100_iccsense_new(struct nvkm_device *device, int index,
+gf100_iccsense_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		   struct nvkm_iccsense **piccsense)
 {
-	return nvkm_iccsense_new_(device, index, piccsense);
+	return nvkm_iccsense_new_(device, type, inst, piccsense);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/priv.h
index cc09c6c504af..c33441124241 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/priv.h
@@ -22,6 +22,6 @@ struct nvkm_iccsense_rail {
 	u8 mohm;
 };
 
-void nvkm_iccsense_ctor(struct nvkm_device *, int, struct nvkm_iccsense *);
-int nvkm_iccsense_new_(struct nvkm_device *, int, struct nvkm_iccsense **);
+void nvkm_iccsense_ctor(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_iccsense *);
+int nvkm_iccsense_new_(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_iccsense **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c
index 364ea4492acc..cd8163a52bb6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c
@@ -218,9 +218,11 @@ static void *
 nvkm_instmem_dtor(struct nvkm_subdev *subdev)
 {
 	struct nvkm_instmem *imem = nvkm_instmem(subdev);
+	void *data = imem;
 	if (imem->func->dtor)
-		return imem->func->dtor(imem);
-	return imem;
+		data = imem->func->dtor(imem);
+	mutex_destroy(&imem->mutex);
+	return data;
 }
 
 static const struct nvkm_subdev_func
@@ -232,13 +234,13 @@ nvkm_instmem = {
 };
 
 void
-nvkm_instmem_ctor(const struct nvkm_instmem_func *func,
-		  struct nvkm_device *device, int index,
-		  struct nvkm_instmem *imem)
+nvkm_instmem_ctor(const struct nvkm_instmem_func *func, struct nvkm_device *device,
+		  enum nvkm_subdev_type type, int inst, struct nvkm_instmem *imem)
 {
-	nvkm_subdev_ctor(&nvkm_instmem, device, index, &imem->subdev);
+	nvkm_subdev_ctor(&nvkm_instmem, device, type, inst, &imem->subdev);
 	imem->func = func;
 	spin_lock_init(&imem->lock);
 	INIT_LIST_HEAD(&imem->list);
 	INIT_LIST_HEAD(&imem->boot);
+	mutex_init(&imem->mutex);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c
index 13d4d7ac0697..648ecf5a8fbc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c
@@ -568,7 +568,7 @@ gk20a_instmem = {
 };
 
 int
-gk20a_instmem_new(struct nvkm_device *device, int index,
+gk20a_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		  struct nvkm_instmem **pimem)
 {
 	struct nvkm_device_tegra *tdev = device->func->tegra(device);
@@ -576,7 +576,7 @@ gk20a_instmem_new(struct nvkm_device *device, int index,
 
 	if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_instmem_ctor(&gk20a_instmem, device, index, &imem->base);
+	nvkm_instmem_ctor(&gk20a_instmem, device, type, inst, &imem->base);
 	mutex_init(&imem->lock);
 	*pimem = &imem->base;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv04.c
index 6bf0dad46919..25603b01d6f8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv04.c
@@ -99,9 +99,9 @@ static void *
 nv04_instobj_dtor(struct nvkm_memory *memory)
 {
 	struct nv04_instobj *iobj = nv04_instobj(memory);
-	mutex_lock(&iobj->imem->base.subdev.mutex);
+	mutex_lock(&iobj->imem->base.mutex);
 	nvkm_mm_free(&iobj->imem->heap, &iobj->node);
-	mutex_unlock(&iobj->imem->base.subdev.mutex);
+	mutex_unlock(&iobj->imem->base.mutex);
 	nvkm_instobj_dtor(&iobj->imem->base, &iobj->base);
 	return iobj;
 }
@@ -132,10 +132,9 @@ nv04_instobj_new(struct nvkm_instmem *base, u32 size, u32 align, bool zero,
 	iobj->base.memory.ptrs = &nv04_instobj_ptrs;
 	iobj->imem = imem;
 
-	mutex_lock(&imem->base.subdev.mutex);
-	ret = nvkm_mm_head(&imem->heap, 0, 1, size, size,
-			   align ? align : 1, &iobj->node);
-	mutex_unlock(&imem->base.subdev.mutex);
+	mutex_lock(&imem->base.mutex);
+	ret = nvkm_mm_head(&imem->heap, 0, 1, size, size, align ? align : 1, &iobj->node);
+	mutex_unlock(&imem->base.mutex);
 	return ret;
 }
 
@@ -218,14 +217,14 @@ nv04_instmem = {
 };
 
 int
-nv04_instmem_new(struct nvkm_device *device, int index,
+nv04_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_instmem **pimem)
 {
 	struct nv04_instmem *imem;
 
 	if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_instmem_ctor(&nv04_instmem, device, index, &imem->base);
+	nvkm_instmem_ctor(&nv04_instmem, device, type, inst, &imem->base);
 	*pimem = &imem->base;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c
index 086c118488ef..6b462f960922 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c
@@ -99,9 +99,9 @@ static void *
 nv40_instobj_dtor(struct nvkm_memory *memory)
 {
 	struct nv40_instobj *iobj = nv40_instobj(memory);
-	mutex_lock(&iobj->imem->base.subdev.mutex);
+	mutex_lock(&iobj->imem->base.mutex);
 	nvkm_mm_free(&iobj->imem->heap, &iobj->node);
-	mutex_unlock(&iobj->imem->base.subdev.mutex);
+	mutex_unlock(&iobj->imem->base.mutex);
 	nvkm_instobj_dtor(&iobj->imem->base, &iobj->base);
 	return iobj;
 }
@@ -132,10 +132,9 @@ nv40_instobj_new(struct nvkm_instmem *base, u32 size, u32 align, bool zero,
 	iobj->base.memory.ptrs = &nv40_instobj_ptrs;
 	iobj->imem = imem;
 
-	mutex_lock(&imem->base.subdev.mutex);
-	ret = nvkm_mm_head(&imem->heap, 0, 1, size, size,
-			   align ? align : 1, &iobj->node);
-	mutex_unlock(&imem->base.subdev.mutex);
+	mutex_lock(&imem->base.mutex);
+	ret = nvkm_mm_head(&imem->heap, 0, 1, size, size, align ? align : 1, &iobj->node);
+	mutex_unlock(&imem->base.mutex);
 	return ret;
 }
 
@@ -236,7 +235,7 @@ nv40_instmem = {
 };
 
 int
-nv40_instmem_new(struct nvkm_device *device, int index,
+nv40_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_instmem **pimem)
 {
 	struct nv40_instmem *imem;
@@ -244,7 +243,7 @@ nv40_instmem_new(struct nvkm_device *device, int index,
 
 	if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_instmem_ctor(&nv40_instmem, device, index, &imem->base);
+	nvkm_instmem_ctor(&nv40_instmem, device, type, inst, &imem->base);
 	*pimem = &imem->base;
 
 	/* map bar */
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
index 02c4eb28cef4..96aca0edfa3c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
@@ -133,12 +133,12 @@ nv50_instobj_kmap(struct nv50_instobj *iobj, struct nvkm_vmm *vmm)
 	 * into it.  The lock has to be dropped while doing this due
 	 * to the possibility of recursion for page table allocation.
 	 */
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&imem->base.mutex);
 	while ((ret = nvkm_vmm_get(vmm, 12, size, &bar))) {
 		/* Evict unused mappings, and keep retrying until we either
 		 * succeed,or there's no more objects left on the LRU.
 		 */
-		mutex_lock(&subdev->mutex);
+		mutex_lock(&imem->base.mutex);
 		eobj = list_first_entry_or_null(&imem->lru, typeof(*eobj), lru);
 		if (eobj) {
 			nvkm_debug(subdev, "evict %016llx %016llx @ %016llx\n",
@@ -151,7 +151,7 @@ nv50_instobj_kmap(struct nv50_instobj *iobj, struct nvkm_vmm *vmm)
 			emap = eobj->map;
 			eobj->map = NULL;
 		}
-		mutex_unlock(&subdev->mutex);
+		mutex_unlock(&imem->base.mutex);
 		if (!eobj)
 			break;
 		iounmap(emap);
@@ -160,12 +160,12 @@ nv50_instobj_kmap(struct nv50_instobj *iobj, struct nvkm_vmm *vmm)
 
 	if (ret == 0)
 		ret = nvkm_memory_map(memory, 0, vmm, bar, NULL, 0);
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&imem->base.mutex);
 	if (ret || iobj->bar) {
 		/* We either failed, or another thread beat us. */
-		mutex_unlock(&subdev->mutex);
+		mutex_unlock(&imem->base.mutex);
 		nvkm_vmm_put(vmm, &bar);
-		mutex_lock(&subdev->mutex);
+		mutex_lock(&imem->base.mutex);
 		return;
 	}
 
@@ -197,7 +197,7 @@ nv50_instobj_release(struct nvkm_memory *memory)
 	wmb();
 	nvkm_bar_flush(subdev->device->bar);
 
-	if (refcount_dec_and_mutex_lock(&iobj->maps, &subdev->mutex)) {
+	if (refcount_dec_and_mutex_lock(&iobj->maps, &imem->base.mutex)) {
 		/* Add the now-unused mapping to the LRU instead of directly
 		 * unmapping it here, in case we need to map it again later.
 		 */
@@ -208,7 +208,7 @@ nv50_instobj_release(struct nvkm_memory *memory)
 
 		/* Switch back to NULL accessors when last map is gone. */
 		iobj->base.memory.ptrs = NULL;
-		mutex_unlock(&subdev->mutex);
+		mutex_unlock(&imem->base.mutex);
 	}
 }
 
@@ -227,9 +227,9 @@ nv50_instobj_acquire(struct nvkm_memory *memory)
 	/* Take the lock, and re-check that another thread hasn't
 	 * already mapped the object in the meantime.
 	 */
-	mutex_lock(&imem->subdev.mutex);
+	mutex_lock(&imem->mutex);
 	if (refcount_inc_not_zero(&iobj->maps)) {
-		mutex_unlock(&imem->subdev.mutex);
+		mutex_unlock(&imem->mutex);
 		return iobj->map;
 	}
 
@@ -252,7 +252,7 @@ nv50_instobj_acquire(struct nvkm_memory *memory)
 		refcount_set(&iobj->maps, 1);
 	}
 
-	mutex_unlock(&imem->subdev.mutex);
+	mutex_unlock(&imem->mutex);
 	return map;
 }
 
@@ -265,7 +265,7 @@ nv50_instobj_boot(struct nvkm_memory *memory, struct nvkm_vmm *vmm)
 	/* Exclude bootstrapped objects (ie. the page tables for the
 	 * instmem BAR itself) from eviction.
 	 */
-	mutex_lock(&imem->subdev.mutex);
+	mutex_lock(&imem->mutex);
 	if (likely(iobj->lru.next)) {
 		list_del_init(&iobj->lru);
 		iobj->lru.next = NULL;
@@ -273,7 +273,7 @@ nv50_instobj_boot(struct nvkm_memory *memory, struct nvkm_vmm *vmm)
 
 	nv50_instobj_kmap(iobj, vmm);
 	nvkm_instmem_boot(imem);
-	mutex_unlock(&imem->subdev.mutex);
+	mutex_unlock(&imem->mutex);
 }
 
 static u64
@@ -315,12 +315,12 @@ nv50_instobj_dtor(struct nvkm_memory *memory)
 	struct nvkm_vma *bar;
 	void *map = map;
 
-	mutex_lock(&imem->subdev.mutex);
+	mutex_lock(&imem->mutex);
 	if (likely(iobj->lru.next))
 		list_del(&iobj->lru);
 	map = iobj->map;
 	bar = iobj->bar;
-	mutex_unlock(&imem->subdev.mutex);
+	mutex_unlock(&imem->mutex);
 
 	if (map) {
 		struct nvkm_vmm *vmm = nvkm_bar_bar2_vmm(imem->subdev.device);
@@ -386,14 +386,14 @@ nv50_instmem = {
 };
 
 int
-nv50_instmem_new(struct nvkm_device *device, int index,
+nv50_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		 struct nvkm_instmem **pimem)
 {
 	struct nv50_instmem *imem;
 
 	if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_instmem_ctor(&nv50_instmem, device, index, &imem->base);
+	nvkm_instmem_ctor(&nv50_instmem, device, type, inst, &imem->base);
 	INIT_LIST_HEAD(&imem->lru);
 	*pimem = &imem->base;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h
index f5da8fcbdde3..56c15e30a5dd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h
@@ -16,7 +16,7 @@ struct nvkm_instmem_func {
 };
 
 void nvkm_instmem_ctor(const struct nvkm_instmem_func *, struct nvkm_device *,
-		       int index, struct nvkm_instmem *);
+		       enum nvkm_subdev_type, int, struct nvkm_instmem *);
 void nvkm_instmem_boot(struct nvkm_instmem *);
 
 #include <core/memory.h>
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/base.c
index 23242179e600..fa683c190795 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/base.c
@@ -33,10 +33,10 @@ nvkm_ltc_tags_clear(struct nvkm_device *device, u32 first, u32 count)
 
 	BUG_ON((first > limit) || (limit >= ltc->num_tags));
 
-	mutex_lock(&ltc->subdev.mutex);
+	mutex_lock(&ltc->mutex);
 	ltc->func->cbc_clear(ltc, first, limit);
 	ltc->func->cbc_wait(ltc);
-	mutex_unlock(&ltc->subdev.mutex);
+	mutex_unlock(&ltc->mutex);
 }
 
 int
@@ -113,6 +113,7 @@ nvkm_ltc_dtor(struct nvkm_subdev *subdev)
 {
 	struct nvkm_ltc *ltc = nvkm_ltc(subdev);
 	nvkm_memory_unref(&ltc->tag_ram);
+	mutex_destroy(&ltc->mutex);
 	return ltc;
 }
 
@@ -126,15 +127,16 @@ nvkm_ltc = {
 
 int
 nvkm_ltc_new_(const struct nvkm_ltc_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_ltc **pltc)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_ltc **pltc)
 {
 	struct nvkm_ltc *ltc;
 
 	if (!(ltc = *pltc = kzalloc(sizeof(*ltc), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_subdev_ctor(&nvkm_ltc, device, index, &ltc->subdev);
+	nvkm_subdev_ctor(&nvkm_ltc, device, type, inst, &ltc->subdev);
 	ltc->func = func;
+	mutex_init(&ltc->mutex);
 	ltc->zbc_min = 1; /* reserve 0 for disabled */
 	ltc->zbc_max = min(func->zbc, NVKM_LTC_MAX_ZBC_CNT) - 1;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c
index a21ef45b8572..fd8aeafc812d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c
@@ -200,8 +200,8 @@ gf100_ltc_oneinit_tag_ram(struct nvkm_ltc *ltc)
 	}
 
 mm_init:
-	nvkm_mm_fini(&fb->tags);
-	return nvkm_mm_init(&fb->tags, 0, 0, ltc->num_tags, 1);
+	nvkm_mm_fini(&fb->tags.mm);
+	return nvkm_mm_init(&fb->tags.mm, 0, 0, ltc->num_tags, 1);
 }
 
 int
@@ -249,7 +249,8 @@ gf100_ltc = {
 };
 
 int
-gf100_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gf100_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gf100_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gf100_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gk104.c
index b4f6e0034d58..94aa09244d67 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gk104.c
@@ -50,7 +50,8 @@ gk104_ltc = {
 };
 
 int
-gk104_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gk104_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gk104_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gk104_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c
index ec0a3844b2d1..54d1d65d5a85 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c
@@ -145,7 +145,8 @@ gm107_ltc = {
 };
 
 int
-gm107_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gm107_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gm107_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gm107_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c
index e18e0dc19ec8..8cfdbbdd8e8d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c
@@ -57,7 +57,8 @@ gm200_ltc = {
 };
 
 int
-gm200_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gm200_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gm200_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gm200_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c
index e923ed76d37a..a4a6cd9b435a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c
@@ -69,7 +69,8 @@ gp100_ltc = {
 };
 
 int
-gp100_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gp100_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gp100_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gp100_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp102.c
index 601747ada655..ff05d617e7f4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp102.c
@@ -45,7 +45,8 @@ gp102_ltc = {
 };
 
 int
-gp102_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gp102_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gp102_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gp102_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp10b.c
index c0063c7caa50..dfebd796cb4b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp10b.c
@@ -59,7 +59,8 @@ gp10b_ltc = {
 };
 
 int
-gp10b_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+gp10b_ltc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_ltc **pltc)
 {
-	return nvkm_ltc_new_(&gp10b_ltc, device, index, pltc);
+	return nvkm_ltc_new_(&gp10b_ltc, device, type, inst, pltc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h
index eca5a711b1b8..2bebe139005d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h
@@ -5,8 +5,8 @@
 #include <subdev/ltc.h>
 #include <core/enum.h>
 
-int nvkm_ltc_new_(const struct nvkm_ltc_func *, struct nvkm_device *,
-		  int index, struct nvkm_ltc **);
+int nvkm_ltc_new_(const struct nvkm_ltc_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_ltc **);
 
 struct nvkm_ltc_func {
 	int  (*oneinit)(struct nvkm_ltc *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild
index 2585ef07532a..ac2b34e9ac6a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild
@@ -14,3 +14,4 @@ nvkm-y += nvkm/subdev/mc/gk20a.o
 nvkm-y += nvkm/subdev/mc/gp100.o
 nvkm-y += nvkm/subdev/mc/gp10b.o
 nvkm-y += nvkm/subdev/mc/tu102.o
+nvkm-y += nvkm/subdev/mc/ga100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c
index 0e57ab2a709f..21c4af3f81d5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c
@@ -35,14 +35,14 @@ nvkm_mc_unk260(struct nvkm_device *device, u32 data)
 }
 
 void
-nvkm_mc_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx, bool en)
+nvkm_mc_intr_mask(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, bool en)
 {
 	struct nvkm_mc *mc = device->mc;
 	const struct nvkm_mc_map *map;
 	if (likely(mc) && mc->func->intr_mask) {
-		u32 mask = nvkm_top_intr_mask(device, devidx);
+		u32 mask = nvkm_top_intr_mask(device, type, inst);
 		for (map = mc->func->intr; !mask && map->stat; map++) {
-			if (map->unit == devidx)
+			if (map->type == type && map->inst == inst)
 				mask = map->stat;
 		}
 		mc->func->intr_mask(mc, mask, en ? mask : 0);
@@ -78,27 +78,34 @@ void
 nvkm_mc_intr(struct nvkm_device *device, bool *handled)
 {
 	struct nvkm_mc *mc = device->mc;
+	struct nvkm_top *top = device->top;
+	struct nvkm_top_device *tdev;
 	struct nvkm_subdev *subdev;
 	const struct nvkm_mc_map *map;
 	u32 stat, intr;
-	u64 subdevs;
 
 	if (unlikely(!mc))
 		return;
 
-	intr = nvkm_mc_intr_stat(mc);
-	stat = nvkm_top_intr(device, intr, &subdevs);
-	while (subdevs) {
-		enum nvkm_devidx subidx = __ffs64(subdevs);
-		subdev = nvkm_device_subdev(device, subidx);
-		if (subdev)
-			nvkm_subdev_intr(subdev);
-		subdevs &= ~BIT_ULL(subidx);
+	stat = intr = nvkm_mc_intr_stat(mc);
+
+	if (top) {
+		list_for_each_entry(tdev, &top->device, head) {
+			if (tdev->intr >= 0 && (stat & BIT(tdev->intr))) {
+				subdev = nvkm_device_subdev(device, tdev->type, tdev->inst);
+				if (subdev) {
+					nvkm_subdev_intr(subdev);
+					stat &= ~BIT(tdev->intr);
+					if (!stat)
+						break;
+				}
+			}
+		}
 	}
 
 	for (map = mc->func->intr; map->stat; map++) {
 		if (intr & map->stat) {
-			subdev = nvkm_device_subdev(device, map->unit);
+			subdev = nvkm_device_subdev(device, map->type, map->inst);
 			if (subdev)
 				nvkm_subdev_intr(subdev);
 			stat &= ~map->stat;
@@ -108,23 +115,19 @@ nvkm_mc_intr(struct nvkm_device *device, bool *handled)
 	if (stat)
 		nvkm_error(&mc->subdev, "intr %08x\n", stat);
 	*handled = intr != 0;
-
-	if (mc->func->intr_hack)
-		mc->func->intr_hack(mc, handled);
 }
 
 static u32
-nvkm_mc_reset_mask(struct nvkm_device *device, bool isauto,
-		   enum nvkm_devidx devidx)
+nvkm_mc_reset_mask(struct nvkm_device *device, bool isauto, enum nvkm_subdev_type type, int inst)
 {
 	struct nvkm_mc *mc = device->mc;
 	const struct nvkm_mc_map *map;
 	u64 pmc_enable = 0;
 	if (likely(mc)) {
-		if (!(pmc_enable = nvkm_top_reset(device, devidx))) {
+		if (!(pmc_enable = nvkm_top_reset(device, type, inst))) {
 			for (map = mc->func->reset; map && map->stat; map++) {
 				if (!isauto || !map->noauto) {
-					if (map->unit == devidx) {
+					if (map->type == type && map->inst == inst) {
 						pmc_enable = map->stat;
 						break;
 					}
@@ -136,9 +139,9 @@ nvkm_mc_reset_mask(struct nvkm_device *device, bool isauto,
 }
 
 void
-nvkm_mc_reset(struct nvkm_device *device, enum nvkm_devidx devidx)
+nvkm_mc_reset(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
-	u64 pmc_enable = nvkm_mc_reset_mask(device, true, devidx);
+	u64 pmc_enable = nvkm_mc_reset_mask(device, true, type, inst);
 	if (pmc_enable) {
 		nvkm_mask(device, 0x000200, pmc_enable, 0x00000000);
 		nvkm_mask(device, 0x000200, pmc_enable, pmc_enable);
@@ -147,17 +150,17 @@ nvkm_mc_reset(struct nvkm_device *device, enum nvkm_devidx devidx)
 }
 
 void
-nvkm_mc_disable(struct nvkm_device *device, enum nvkm_devidx devidx)
+nvkm_mc_disable(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
-	u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx);
+	u64 pmc_enable = nvkm_mc_reset_mask(device, false, type, inst);
 	if (pmc_enable)
 		nvkm_mask(device, 0x000200, pmc_enable, 0x00000000);
 }
 
 void
-nvkm_mc_enable(struct nvkm_device *device, enum nvkm_devidx devidx)
+nvkm_mc_enable(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
-	u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx);
+	u64 pmc_enable = nvkm_mc_reset_mask(device, false, type, inst);
 	if (pmc_enable) {
 		nvkm_mask(device, 0x000200, pmc_enable, pmc_enable);
 		nvkm_rd32(device, 0x000200);
@@ -165,9 +168,9 @@ nvkm_mc_enable(struct nvkm_device *device, enum nvkm_devidx devidx)
 }
 
 bool
-nvkm_mc_enabled(struct nvkm_device *device, enum nvkm_devidx devidx)
+nvkm_mc_enabled(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
-	u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx);
+	u64 pmc_enable = nvkm_mc_reset_mask(device, false, type, inst);
 
 	return (pmc_enable != 0) &&
 	       ((nvkm_rd32(device, 0x000200) & pmc_enable) == pmc_enable);
@@ -206,19 +209,19 @@ nvkm_mc = {
 
 void
 nvkm_mc_ctor(const struct nvkm_mc_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_mc *mc)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_mc *mc)
 {
-	nvkm_subdev_ctor(&nvkm_mc, device, index, &mc->subdev);
+	nvkm_subdev_ctor(&nvkm_mc, device, type, inst, &mc->subdev);
 	mc->func = func;
 }
 
 int
 nvkm_mc_new_(const struct nvkm_mc_func *func, struct nvkm_device *device,
-	     int index, struct nvkm_mc **pmc)
+	     enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
 	struct nvkm_mc *mc;
 	if (!(mc = *pmc = kzalloc(sizeof(*mc), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_mc_ctor(func, device, index, *pmc);
+	nvkm_mc_ctor(func, device, type, inst, *pmc);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c
index 430a61c3df44..4cfc1c984006 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c
@@ -62,7 +62,7 @@ g84_mc = {
 };
 
 int
-g84_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+g84_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&g84_mc, device, index, pmc);
+	return nvkm_mc_new_(&g84_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c
index 93ad4982ce5f..b7e58d75d894 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c
@@ -62,7 +62,7 @@ g98_mc = {
 };
 
 int
-g98_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+g98_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&g98_mc, device, index, pmc);
+	return nvkm_mc_new_(&g98_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/ga100.c
new file mode 100644
index 000000000000..4105175dfccd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/ga100.c
@@ -0,0 +1,74 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "priv.h"
+
+static void
+ga100_mc_intr_unarm(struct nvkm_mc *mc)
+{
+	nvkm_wr32(mc->subdev.device, 0xb81610, 0x00000004);
+}
+
+static void
+ga100_mc_intr_rearm(struct nvkm_mc *mc)
+{
+	nvkm_wr32(mc->subdev.device, 0xb81608, 0x00000004);
+}
+
+static void
+ga100_mc_intr_mask(struct nvkm_mc *mc, u32 mask, u32 intr)
+{
+	nvkm_wr32(mc->subdev.device, 0xb81210,          mask & intr );
+	nvkm_wr32(mc->subdev.device, 0xb81410, mask & ~(mask & intr));
+}
+
+static u32
+ga100_mc_intr_stat(struct nvkm_mc *mc)
+{
+	u32 intr_top = nvkm_rd32(mc->subdev.device, 0xb81600), intr = 0x00000000;
+	if (intr_top & 0x00000004)
+		intr = nvkm_mask(mc->subdev.device, 0xb81010, 0x00000000, 0x00000000);
+	return intr;
+}
+
+static void
+ga100_mc_init(struct nvkm_mc *mc)
+{
+	nv50_mc_init(mc);
+	nvkm_wr32(mc->subdev.device, 0xb81210, 0xffffffff);
+}
+
+static const struct nvkm_mc_func
+ga100_mc = {
+	.init = ga100_mc_init,
+	.intr = gp100_mc_intr,
+	.intr_unarm = ga100_mc_intr_unarm,
+	.intr_rearm = ga100_mc_intr_rearm,
+	.intr_mask = ga100_mc_intr_mask,
+	.intr_stat = ga100_mc_intr_stat,
+	.reset = gk104_mc_reset,
+};
+
+int
+ga100_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
+{
+	return nvkm_mc_new_(&ga100_mc, device, type, inst, pmc);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c
index f93766418056..3a589c6f7fad 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c
@@ -27,11 +27,11 @@ static const struct nvkm_mc_map
 gf100_mc_reset[] = {
 	{ 0x00020000, NVKM_ENGINE_MSPDEC },
 	{ 0x00008000, NVKM_ENGINE_MSVLD },
-	{ 0x00002000, NVKM_SUBDEV_PMU, true },
+	{ 0x00002000, NVKM_SUBDEV_PMU, 0, true },
 	{ 0x00001000, NVKM_ENGINE_GR },
 	{ 0x00000100, NVKM_ENGINE_FIFO },
-	{ 0x00000080, NVKM_ENGINE_CE1 },
-	{ 0x00000040, NVKM_ENGINE_CE0 },
+	{ 0x00000080, NVKM_ENGINE_CE, 1 },
+	{ 0x00000040, NVKM_ENGINE_CE, 0 },
 	{ 0x00000002, NVKM_ENGINE_MSPPP },
 	{}
 };
@@ -43,10 +43,10 @@ gf100_mc_intr[] = {
 	{ 0x00008000, NVKM_ENGINE_MSVLD },
 	{ 0x00001000, NVKM_ENGINE_GR },
 	{ 0x00000100, NVKM_ENGINE_FIFO },
-	{ 0x00000040, NVKM_ENGINE_CE1 },
-	{ 0x00000020, NVKM_ENGINE_CE0 },
+	{ 0x00000040, NVKM_ENGINE_CE, 1 },
+	{ 0x00000020, NVKM_ENGINE_CE, 0 },
 	{ 0x00000001, NVKM_ENGINE_MSPPP },
-	{ 0x40000000, NVKM_SUBDEV_IBUS },
+	{ 0x40000000, NVKM_SUBDEV_PRIVRING },
 	{ 0x10000000, NVKM_SUBDEV_BUS },
 	{ 0x08000000, NVKM_SUBDEV_FB },
 	{ 0x02000000, NVKM_SUBDEV_LTC },
@@ -112,7 +112,7 @@ gf100_mc = {
 };
 
 int
-gf100_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+gf100_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&gf100_mc, device, index, pmc);
+	return nvkm_mc_new_(&gf100_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c
index 7b8c6ecad1a5..d9b9067fa93f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c
@@ -26,7 +26,7 @@
 const struct nvkm_mc_map
 gk104_mc_reset[] = {
 	{ 0x00000100, NVKM_ENGINE_FIFO },
-	{ 0x00002000, NVKM_SUBDEV_PMU, true },
+	{ 0x00002000, NVKM_SUBDEV_PMU, 0, true },
 	{}
 };
 
@@ -34,7 +34,7 @@ const struct nvkm_mc_map
 gk104_mc_intr[] = {
 	{ 0x04000000, NVKM_ENGINE_DISP },
 	{ 0x00000100, NVKM_ENGINE_FIFO },
-	{ 0x40000000, NVKM_SUBDEV_IBUS },
+	{ 0x40000000, NVKM_SUBDEV_PRIVRING },
 	{ 0x10000000, NVKM_SUBDEV_BUS },
 	{ 0x08000000, NVKM_SUBDEV_FB },
 	{ 0x02000000, NVKM_SUBDEV_LTC },
@@ -60,7 +60,7 @@ gk104_mc = {
 };
 
 int
-gk104_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+gk104_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&gk104_mc, device, index, pmc);
+	return nvkm_mc_new_(&gk104_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c
index ca1bf3279dbe..03590292749a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c
@@ -35,7 +35,7 @@ gk20a_mc = {
 };
 
 int
-gk20a_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+gk20a_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&gk20a_mc, device, index, pmc);
+	return nvkm_mc_new_(&gk20a_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c
index 43db245eec9a..5fd1a0595c33 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c
@@ -80,7 +80,7 @@ gp100_mc_intr[] = {
 	{ 0x04000000, NVKM_ENGINE_DISP },
 	{ 0x00000100, NVKM_ENGINE_FIFO },
 	{ 0x00000200, NVKM_SUBDEV_FAULT },
-	{ 0x40000000, NVKM_SUBDEV_IBUS },
+	{ 0x40000000, NVKM_SUBDEV_PRIVRING },
 	{ 0x10000000, NVKM_SUBDEV_BUS },
 	{ 0x08000000, NVKM_SUBDEV_FB },
 	{ 0x02000000, NVKM_SUBDEV_LTC },
@@ -106,13 +106,13 @@ gp100_mc = {
 
 int
 gp100_mc_new_(const struct nvkm_mc_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_mc **pmc)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
 	struct gp100_mc *mc;
 
 	if (!(mc = kzalloc(sizeof(*mc), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_mc_ctor(func, device, index, &mc->base);
+	nvkm_mc_ctor(func, device, type, inst, &mc->base);
 	*pmc = &mc->base;
 
 	spin_lock_init(&mc->lock);
@@ -122,7 +122,7 @@ gp100_mc_new_(const struct nvkm_mc_func *func, struct nvkm_device *device,
 }
 
 int
-gp100_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+gp100_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return gp100_mc_new_(&gp100_mc, device, index, pmc);
+	return gp100_mc_new_(&gp100_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp10b.c
index 45c62f5ef782..dd581d030ced 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp10b.c
@@ -43,7 +43,7 @@ gp10b_mc = {
 };
 
 int
-gp10b_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+gp10b_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return gp100_mc_new_(&gp10b_mc, device, index, pmc);
+	return gp100_mc_new_(&gp10b_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c
index 99d50a3d956f..1b4d43531dba 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c
@@ -27,7 +27,7 @@ static const struct nvkm_mc_map
 gt215_mc_reset[] = {
 	{ 0x04008000, NVKM_ENGINE_MSVLD },
 	{ 0x01020000, NVKM_ENGINE_MSPDEC },
-	{ 0x00802000, NVKM_ENGINE_CE0 },
+	{ 0x00802000, NVKM_ENGINE_CE, 0 },
 	{ 0x00400002, NVKM_ENGINE_MSPPP },
 	{ 0x00201000, NVKM_ENGINE_GR },
 	{ 0x00000100, NVKM_ENGINE_FIFO },
@@ -37,7 +37,7 @@ gt215_mc_reset[] = {
 static const struct nvkm_mc_map
 gt215_mc_intr[] = {
 	{ 0x04000000, NVKM_ENGINE_DISP },
-	{ 0x00400000, NVKM_ENGINE_CE0 },
+	{ 0x00400000, NVKM_ENGINE_CE, 0 },
 	{ 0x00020000, NVKM_ENGINE_MSPDEC },
 	{ 0x00008000, NVKM_ENGINE_MSVLD },
 	{ 0x00001000, NVKM_ENGINE_GR },
@@ -71,7 +71,7 @@ gt215_mc = {
 };
 
 int
-gt215_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+gt215_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&gt215_mc, device, index, pmc);
+	return nvkm_mc_new_(&gt215_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c
index 6509defd1460..bc0d09bafa99 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c
@@ -80,7 +80,7 @@ nv04_mc = {
 };
 
 int
-nv04_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+nv04_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&nv04_mc, device, index, pmc);
+	return nvkm_mc_new_(&nv04_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c
index 9213107901e6..ab59ca1ee068 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c
@@ -44,7 +44,7 @@ nv11_mc = {
 };
 
 int
-nv11_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+nv11_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&nv11_mc, device, index, pmc);
+	return nvkm_mc_new_(&nv11_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c
index 64bf5bbf8146..03d756e26e57 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c
@@ -53,7 +53,7 @@ nv17_mc = {
 };
 
 int
-nv17_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+nv17_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&nv17_mc, device, index, pmc);
+	return nvkm_mc_new_(&nv17_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c
index 65fa44a64b98..95f65766e8b0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c
@@ -48,7 +48,7 @@ nv44_mc = {
 };
 
 int
-nv44_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+nv44_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&nv44_mc, device, index, pmc);
+	return nvkm_mc_new_(&nv44_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c
index fe93b4fd7100..fce3613cdfa5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c
@@ -55,7 +55,7 @@ nv50_mc = {
 };
 
 int
-nv50_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+nv50_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return nvkm_mc_new_(&nv50_mc, device, index, pmc);
+	return nvkm_mc_new_(&nv50_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h
index 4aab753a6040..c8bcabb98f99 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h
@@ -4,14 +4,15 @@
 #define nvkm_mc(p) container_of((p), struct nvkm_mc, subdev)
 #include <subdev/mc.h>
 
-void nvkm_mc_ctor(const struct nvkm_mc_func *, struct nvkm_device *,
-		  int index, struct nvkm_mc *);
-int nvkm_mc_new_(const struct nvkm_mc_func *, struct nvkm_device *,
-		 int index, struct nvkm_mc **);
+void nvkm_mc_ctor(const struct nvkm_mc_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_mc *);
+int nvkm_mc_new_(const struct nvkm_mc_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		 struct nvkm_mc **);
 
 struct nvkm_mc_map {
 	u32 stat;
-	u32 unit;
+	enum nvkm_subdev_type type;
+	int inst;
 	bool noauto;
 };
 
@@ -26,7 +27,6 @@ struct nvkm_mc_func {
 	void (*intr_mask)(struct nvkm_mc *, u32 mask, u32 stat);
 	/* retrieve pending interrupt mask (NV_PMC_INTR) */
 	u32 (*intr_stat)(struct nvkm_mc *);
-	void (*intr_hack)(struct nvkm_mc *, bool *handled);
 	const struct nvkm_mc_map *reset;
 	void (*unk260)(struct nvkm_mc *, u32);
 };
@@ -53,7 +53,7 @@ void gf100_mc_unk260(struct nvkm_mc *, u32);
 void gp100_mc_intr_unarm(struct nvkm_mc *);
 void gp100_mc_intr_rearm(struct nvkm_mc *);
 void gp100_mc_intr_mask(struct nvkm_mc *, u32, u32);
-int gp100_mc_new_(const struct nvkm_mc_func *, struct nvkm_device *, int,
+int gp100_mc_new_(const struct nvkm_mc_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
 		  struct nvkm_mc **);
 
 extern const struct nvkm_mc_map gk104_mc_intr[];
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu102.c
index d098c44a4fcb..58db83ebadc5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu102.c
@@ -19,37 +19,118 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  */
+#define tu102_mc(p) container_of((p), struct tu102_mc, base)
 #include "priv.h"
 
+struct tu102_mc {
+	struct nvkm_mc base;
+	spinlock_t lock;
+	bool intr;
+	u32 mask;
+};
+
 static void
-tu102_mc_intr_hack(struct nvkm_mc *mc, bool *handled)
+tu102_mc_intr_update(struct tu102_mc *mc)
 {
-	struct nvkm_device *device = mc->subdev.device;
-	u32 stat = nvkm_rd32(device, 0xb81010);
-	if (stat & 0x00000050) {
-		struct nvkm_subdev *subdev =
-			nvkm_device_subdev(device, NVKM_SUBDEV_FAULT);
-		nvkm_wr32(device, 0xb81010, stat & 0x00000050);
-		if (subdev)
-			nvkm_subdev_intr(subdev);
-		*handled = true;
+	struct nvkm_device *device = mc->base.subdev.device;
+	u32 mask = mc->intr ? mc->mask : 0, i;
+
+	for (i = 0; i < 2; i++) {
+		nvkm_wr32(device, 0x000180 + (i * 0x04), ~mask);
+		nvkm_wr32(device, 0x000160 + (i * 0x04),  mask);
 	}
+
+	if (mask & 0x00000200)
+		nvkm_wr32(device, 0xb81608, 0x6);
+	else
+		nvkm_wr32(device, 0xb81610, 0x6);
 }
 
+void
+tu102_mc_intr_unarm(struct nvkm_mc *base)
+{
+	struct tu102_mc *mc = tu102_mc(base);
+	unsigned long flags;
+
+	spin_lock_irqsave(&mc->lock, flags);
+	mc->intr = false;
+	tu102_mc_intr_update(mc);
+	spin_unlock_irqrestore(&mc->lock, flags);
+}
+
+void
+tu102_mc_intr_rearm(struct nvkm_mc *base)
+{
+	struct tu102_mc *mc = tu102_mc(base);
+	unsigned long flags;
+
+	spin_lock_irqsave(&mc->lock, flags);
+	mc->intr = true;
+	tu102_mc_intr_update(mc);
+	spin_unlock_irqrestore(&mc->lock, flags);
+}
+
+void
+tu102_mc_intr_mask(struct nvkm_mc *base, u32 mask, u32 intr)
+{
+	struct tu102_mc *mc = tu102_mc(base);
+	unsigned long flags;
+
+	spin_lock_irqsave(&mc->lock, flags);
+	mc->mask = (mc->mask & ~mask) | intr;
+	tu102_mc_intr_update(mc);
+	spin_unlock_irqrestore(&mc->lock, flags);
+}
+
+static u32
+tu102_mc_intr_stat(struct nvkm_mc *mc)
+{
+	struct nvkm_device *device = mc->subdev.device;
+	u32 intr0 = nvkm_rd32(device, 0x000100);
+	u32 intr1 = nvkm_rd32(device, 0x000104);
+	u32 intr_top = nvkm_rd32(device, 0xb81600);
+
+	/* Turing and above route the MMU fault interrupts via a different
+	 * interrupt tree with different control registers. For the moment remap
+	 * them back to the old PMC vector.
+	 */
+	if (intr_top & 0x00000006)
+		intr0 |= 0x00000200;
+
+	return intr0 | intr1;
+}
+
+
 static const struct nvkm_mc_func
 tu102_mc = {
 	.init = nv50_mc_init,
 	.intr = gp100_mc_intr,
-	.intr_unarm = gp100_mc_intr_unarm,
-	.intr_rearm = gp100_mc_intr_rearm,
-	.intr_mask = gp100_mc_intr_mask,
-	.intr_stat = gf100_mc_intr_stat,
-	.intr_hack = tu102_mc_intr_hack,
+	.intr_unarm = tu102_mc_intr_unarm,
+	.intr_rearm = tu102_mc_intr_rearm,
+	.intr_mask = tu102_mc_intr_mask,
+	.intr_stat = tu102_mc_intr_stat,
 	.reset = gk104_mc_reset,
 };
 
+static int
+tu102_mc_new_(const struct nvkm_mc_func *func, struct nvkm_device *device,
+	      enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
+{
+	struct tu102_mc *mc;
+
+	if (!(mc = kzalloc(sizeof(*mc), GFP_KERNEL)))
+		return -ENOMEM;
+	nvkm_mc_ctor(func, device, type, inst, &mc->base);
+	*pmc = &mc->base;
+
+	spin_lock_init(&mc->lock);
+	mc->intr = false;
+	mc->mask = 0x7fffffff;
+	return 0;
+}
+
 int
-tu102_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+tu102_mc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_mc **pmc)
 {
-	return gp100_mc_new_(&tu102_mc, device, index, pmc);
+	return tu102_mc_new_(&tu102_mc, device, type, inst, pmc);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/base.c
index de91e9a26172..ad3b44a9e0e7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/base.c
@@ -316,9 +316,9 @@ nvkm_mmu_vram(struct nvkm_mmu *mmu)
 {
 	struct nvkm_device *device = mmu->subdev.device;
 	struct nvkm_mm *mm = &device->fb->ram->vram;
-	const u32 sizeN = nvkm_mm_heap_size(mm, NVKM_RAM_MM_NORMAL);
-	const u32 sizeU = nvkm_mm_heap_size(mm, NVKM_RAM_MM_NOMAP);
-	const u32 sizeM = nvkm_mm_heap_size(mm, NVKM_RAM_MM_MIXED);
+	const u64 sizeN = nvkm_mm_heap_size(mm, NVKM_RAM_MM_NORMAL);
+	const u64 sizeU = nvkm_mm_heap_size(mm, NVKM_RAM_MM_NOMAP);
+	const u64 sizeM = nvkm_mm_heap_size(mm, NVKM_RAM_MM_MIXED);
 	u8 type = NVKM_MEM_KIND * !!mmu->func->kind;
 	u8 heap = NVKM_MEM_VRAM;
 	int heapM, heapN, heapU;
@@ -402,6 +402,7 @@ nvkm_mmu_dtor(struct nvkm_subdev *subdev)
 	nvkm_vmm_unref(&mmu->vmm);
 
 	nvkm_mmu_ptc_fini(mmu);
+	mutex_destroy(&mmu->mutex);
 	return mmu;
 }
 
@@ -414,22 +415,23 @@ nvkm_mmu = {
 
 void
 nvkm_mmu_ctor(const struct nvkm_mmu_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_mmu *mmu)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_mmu *mmu)
 {
-	nvkm_subdev_ctor(&nvkm_mmu, device, index, &mmu->subdev);
+	nvkm_subdev_ctor(&nvkm_mmu, device, type, inst, &mmu->subdev);
 	mmu->func = func;
 	mmu->dma_bits = func->dma_bits;
 	nvkm_mmu_ptc_init(mmu);
+	mutex_init(&mmu->mutex);
 	mmu->user.ctor = nvkm_ummu_new;
 	mmu->user.base = func->mmu.user;
 }
 
 int
 nvkm_mmu_new_(const struct nvkm_mmu_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_mmu **pmmu)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_mmu **pmmu)
 {
 	if (!(*pmmu = kzalloc(sizeof(**pmmu), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_mmu_ctor(func, device, index, *pmmu);
+	nvkm_mmu_ctor(func, device, type, inst, *pmmu);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/g84.c
index 8accda5a772b..ce47a3b97be9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/g84.c
@@ -35,7 +35,8 @@ g84_mmu = {
 };
 
 int
-g84_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+g84_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&g84_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&g84_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gf100.c
index 2cd5ec81c0d0..7a28b1d49f7c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gf100.c
@@ -84,7 +84,8 @@ gf100_mmu = {
 };
 
 int
-gf100_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gf100_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&gf100_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&gf100_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk104.c
index 3d7d1eb1cff9..34c9b2b821f6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk104.c
@@ -35,7 +35,8 @@ gk104_mmu = {
 };
 
 int
-gk104_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gk104_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&gk104_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&gk104_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk20a.c
index ac74965a60d4..a7db29c429ee 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gk20a.c
@@ -35,7 +35,8 @@ gk20a_mmu = {
 };
 
 int
-gk20a_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gk20a_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&gk20a_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&gk20a_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm200.c
index 83990c83f9f8..e1696f637a68 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm200.c
@@ -90,9 +90,10 @@ gm200_mmu_fixed = {
 };
 
 int
-gm200_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gm200_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
 	if (device->fb->page)
-		return nvkm_mmu_new_(&gm200_mmu_fixed, device, index, pmmu);
-	return nvkm_mmu_new_(&gm200_mmu, device, index, pmmu);
+		return nvkm_mmu_new_(&gm200_mmu_fixed, device, type, inst, pmmu);
+	return nvkm_mmu_new_(&gm200_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm20b.c
index 7353a94b4091..e6e1a8ad701e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gm20b.c
@@ -47,9 +47,10 @@ gm20b_mmu_fixed = {
 };
 
 int
-gm20b_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gm20b_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
 	if (device->fb->page)
-		return nvkm_mmu_new_(&gm20b_mmu_fixed, device, index, pmmu);
-	return nvkm_mmu_new_(&gm20b_mmu, device, index, pmmu);
+		return nvkm_mmu_new_(&gm20b_mmu_fixed, device, type, inst, pmmu);
+	return nvkm_mmu_new_(&gm20b_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp100.c
index 65cb9d28e60e..daa5ab0f8711 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp100.c
@@ -37,9 +37,10 @@ gp100_mmu = {
 };
 
 int
-gp100_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gp100_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
 	if (!nvkm_boolopt(device->cfgopt, "GP100MmuLayout", true))
-		return gm200_mmu_new(device, index, pmmu);
-	return nvkm_mmu_new_(&gp100_mmu, device, index, pmmu);
+		return gm200_mmu_new(device, type, inst, pmmu);
+	return nvkm_mmu_new_(&gp100_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp10b.c
index 0a50be9a785a..edd0bf9a5cd8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gp10b.c
@@ -37,9 +37,10 @@ gp10b_mmu = {
 };
 
 int
-gp10b_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gp10b_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
 	if (!nvkm_boolopt(device->cfgopt, "GP100MmuLayout", true))
-		return gm20b_mmu_new(device, index, pmmu);
-	return nvkm_mmu_new_(&gp10b_mmu, device, index, pmmu);
+		return gm20b_mmu_new(device, type, inst, pmmu);
+	return nvkm_mmu_new_(&gp10b_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gv100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gv100.c
index e0997eedd6d9..fb8bdc88d566 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gv100.c
@@ -37,7 +37,8 @@ gv100_mmu = {
 };
 
 int
-gv100_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+gv100_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&gv100_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&gv100_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/mcp77.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/mcp77.c
index 0527b50730d9..514876d6411b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/mcp77.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/mcp77.c
@@ -35,7 +35,8 @@ mcp77_mmu = {
 };
 
 int
-mcp77_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+mcp77_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&mcp77_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&mcp77_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv04.c
index d201c887c2cd..0674aa8f68c8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv04.c
@@ -35,7 +35,8 @@ nv04_mmu = {
 };
 
 int
-nv04_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+nv04_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&nv04_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&nv04_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv41.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv41.c
index adca81895c09..909f92b72847 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv41.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv41.c
@@ -47,11 +47,12 @@ nv41_mmu = {
 };
 
 int
-nv41_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+nv41_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_mmu **pmmu)
 {
 	if (device->type == NVKM_DEVICE_AGP ||
 	    !nvkm_boolopt(device->cfgopt, "NvPCIE", true))
-		return nv04_mmu_new(device, index, pmmu);
+		return nv04_mmu_new(device, type, inst, pmmu);
 
-	return nvkm_mmu_new_(&nv41_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&nv41_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv44.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv44.c
index 598c53a27bde..dd2a8d461da3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv44.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv44.c
@@ -62,11 +62,12 @@ nv44_mmu = {
 };
 
 int
-nv44_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+nv44_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_mmu **pmmu)
 {
 	if (device->type == NVKM_DEVICE_AGP ||
 	    !nvkm_boolopt(device->cfgopt, "NvPCIE", true))
-		return nv04_mmu_new(device, index, pmmu);
+		return nv04_mmu_new(device, type, inst, pmmu);
 
-	return nvkm_mmu_new_(&nv44_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&nv44_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv50.c
index c0083ddda65a..78d46e35d0a9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/nv50.c
@@ -71,7 +71,8 @@ nv50_mmu = {
 };
 
 int
-nv50_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+nv50_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&nv50_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&nv50_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h
index 479b02344271..5265bf4d8366 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h
@@ -4,10 +4,10 @@
 #define nvkm_mmu(p) container_of((p), struct nvkm_mmu, subdev)
 #include <subdev/mmu.h>
 
-void nvkm_mmu_ctor(const struct nvkm_mmu_func *, struct nvkm_device *,
-		   int index, struct nvkm_mmu *);
-int nvkm_mmu_new_(const struct nvkm_mmu_func *, struct nvkm_device *,
-		  int index, struct nvkm_mmu **);
+void nvkm_mmu_ctor(const struct nvkm_mmu_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_mmu *);
+int nvkm_mmu_new_(const struct nvkm_mmu_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_mmu **);
 
 struct nvkm_mmu_func {
 	void (*init)(struct nvkm_mmu *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c
index 94081f35f967..8d060ce47f86 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c
@@ -51,7 +51,8 @@ tu102_mmu = {
 };
 
 int
-tu102_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu)
+tu102_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mmu **pmmu)
 {
-	return nvkm_mmu_new_(&tu102_mmu, device, index, pmmu);
+	return nvkm_mmu_new_(&tu102_mmu, device, type, inst, pmmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgf100.c
index 6a2d9eb8e1ea..5438384d9a67 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgf100.c
@@ -187,12 +187,11 @@ gf100_vmm_invalidate_pdb(struct nvkm_vmm *vmm, u64 addr)
 void
 gf100_vmm_invalidate(struct nvkm_vmm *vmm, u32 type)
 {
-	struct nvkm_subdev *subdev = &vmm->mmu->subdev;
-	struct nvkm_device *device = subdev->device;
+	struct nvkm_device *device = vmm->mmu->subdev.device;
 	struct nvkm_mmu_pt *pd = vmm->pd->pt[0];
 	u64 addr = 0;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&vmm->mmu->mutex);
 	/* Looks like maybe a "free flush slots" counter, the
 	 * faster you write to 0x100cbc to more it decreases.
 	 */
@@ -222,7 +221,7 @@ gf100_vmm_invalidate(struct nvkm_vmm *vmm, u32 type)
 		if (nvkm_rd32(device, 0x100c80) & 0x00008000)
 			break;
 	);
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&vmm->mmu->mutex);
 }
 
 void
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv41.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv41.c
index 1d3369683a21..31984671daf8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv41.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv41.c
@@ -80,17 +80,16 @@ nv41_vmm_desc_12[] = {
 static void
 nv41_vmm_flush(struct nvkm_vmm *vmm, int level)
 {
-	struct nvkm_subdev *subdev = &vmm->mmu->subdev;
-	struct nvkm_device *device = subdev->device;
+	struct nvkm_device *device = vmm->mmu->subdev.device;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&vmm->mmu->mutex);
 	nvkm_wr32(device, 0x100810, 0x00000022);
 	nvkm_msec(device, 2000,
 		if (nvkm_rd32(device, 0x100810) & 0x00000020)
 			break;
 	);
 	nvkm_wr32(device, 0x100810, 0x00000000);
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&vmm->mmu->mutex);
 }
 
 static const struct nvkm_vmm_func
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv50.c
index 2d89e27e8e9e..b7548dcd72c7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv50.c
@@ -184,7 +184,7 @@ nv50_vmm_flush(struct nvkm_vmm *vmm, int level)
 	struct nvkm_device *device = subdev->device;
 	int i, id;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&vmm->mmu->mutex);
 	for (i = 0; i < NVKM_SUBDEV_NR; i++) {
 		if (!atomic_read(&vmm->engref[i]))
 			continue;
@@ -207,7 +207,7 @@ nv50_vmm_flush(struct nvkm_vmm *vmm, int level)
 		case NVKM_ENGINE_MSVLD : id = 0x09; break;
 		case NVKM_ENGINE_CIPHER:
 		case NVKM_ENGINE_SEC   : id = 0x0a; break;
-		case NVKM_ENGINE_CE0   : id = 0x0d; break;
+		case NVKM_ENGINE_CE    : id = 0x0d; break;
 		default:
 			continue;
 		}
@@ -217,10 +217,9 @@ nv50_vmm_flush(struct nvkm_vmm *vmm, int level)
 			if (!(nvkm_rd32(device, 0x100c80) & 0x00000001))
 				break;
 		) < 0)
-			nvkm_error(subdev, "%s mmu invalidate timeout\n",
-				   nvkm_subdev_name[i]);
+			nvkm_error(subdev, "%s mmu invalidate timeout\n", nvkm_subdev_type[i]);
 	}
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&vmm->mmu->mutex);
 }
 
 int
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c
index b1294d0076c0..6cb5eefa45e9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c
@@ -26,15 +26,14 @@
 static void
 tu102_vmm_flush(struct nvkm_vmm *vmm, int depth)
 {
-	struct nvkm_subdev *subdev = &vmm->mmu->subdev;
-	struct nvkm_device *device = subdev->device;
+	struct nvkm_device *device = vmm->mmu->subdev.device;
 	u32 type = (5 /* CACHE_LEVEL_UP_TO_PDE3 */ - depth) << 24;
 
 	type |= 0x00000001; /* PAGE_ALL */
 	if (atomic_read(&vmm->engref[NVKM_SUBDEV_BAR]))
 		type |= 0x00000004; /* HUB_ONLY */
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&vmm->mmu->mutex);
 
 	nvkm_wr32(device, 0xb830a0, vmm->pd->pt[0]->addr >> 8);
 	nvkm_wr32(device, 0xb830a4, 0x00000000);
@@ -46,7 +45,7 @@ tu102_vmm_flush(struct nvkm_vmm *vmm, int depth)
 			break;
 	);
 
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&vmm->mmu->mutex);
 }
 
 static const struct nvkm_vmm_func
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/base.c
index f44682d62f75..c1acfe642da3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/base.c
@@ -230,7 +230,8 @@ nvkm_mxm = {
 };
 
 int
-nvkm_mxm_new_(struct nvkm_device *device, int index, struct nvkm_mxm **pmxm)
+nvkm_mxm_new_(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_mxm **pmxm)
 {
 	struct nvkm_bios *bios = device->bios;
 	struct nvkm_mxm *mxm;
@@ -240,7 +241,7 @@ nvkm_mxm_new_(struct nvkm_device *device, int index, struct nvkm_mxm **pmxm)
 	if (!(mxm = *pmxm = kzalloc(sizeof(*mxm), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_subdev_ctor(&nvkm_mxm, device, index, &mxm->subdev);
+	nvkm_subdev_ctor(&nvkm_mxm, device, type, inst, &mxm->subdev);
 
 	data = mxm_table(bios, &ver, &len);
 	if (!data || !(ver = nvbios_rd08(bios, data))) {
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/nv50.c
index 70e2c414bb7b..f3167904dcb0 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/nv50.c
@@ -201,12 +201,13 @@ mxm_dcb_sanitise(struct nvkm_mxm *mxm)
 }
 
 int
-nv50_mxm_new(struct nvkm_device *device, int index, struct nvkm_subdev **pmxm)
+nv50_mxm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_subdev **pmxm)
 {
 	struct nvkm_mxm *mxm;
 	int ret;
 
-	ret = nvkm_mxm_new_(device, index, &mxm);
+	ret = nvkm_mxm_new_(device, type, inst, &mxm);
 	if (mxm)
 		*pmxm = &mxm->subdev;
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/priv.h
index fc8f69e6fc64..fcacb6c6a7f7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mxm/priv.h
@@ -12,5 +12,5 @@ struct nvkm_mxm {
 	u8 *mxms;
 };
 
-int nvkm_mxm_new_(struct nvkm_device *, int index, struct nvkm_mxm **);
+int nvkm_mxm_new_(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_mxm **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
index ee2431a7804e..a7d42ea8ba28 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
@@ -183,13 +183,13 @@ nvkm_pci_func = {
 
 int
 nvkm_pci_new_(const struct nvkm_pci_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_pci **ppci)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_pci **ppci)
 {
 	struct nvkm_pci *pci;
 
 	if (!(pci = *ppci = kzalloc(sizeof(**ppci), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_pci_func, device, index, &pci->subdev);
+	nvkm_subdev_ctor(&nvkm_pci_func, device, type, inst, &pci->subdev);
 	pci->func = func;
 	pci->pdev = device->func->pci(device)->pdev;
 	pci->irq = -1;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c
index 62438d892f42..5b29aacedef3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c
@@ -150,7 +150,8 @@ g84_pci_func = {
 };
 
 int
-g84_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+g84_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&g84_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&g84_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c
index 48874359d5f6..a9e0674009c6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c
@@ -51,7 +51,8 @@ g92_pci_func = {
 };
 
 int
-g92_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+g92_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&g92_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&g92_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c
index 09adb37a5664..7bacd0693283 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c
@@ -43,7 +43,8 @@ g94_pci_func = {
 };
 
 int
-g94_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+g94_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	    struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&g94_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&g94_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c
index 00a5e7d3ee9d..099906092fe1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c
@@ -96,7 +96,8 @@ gf100_pci_func = {
 };
 
 int
-gf100_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+gf100_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&gf100_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&gf100_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c
index 11bf419afe3f..bcde609ba866 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c
@@ -43,7 +43,8 @@ gf106_pci_func = {
 };
 
 int
-gf106_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+gf106_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&gf106_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&gf106_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c
index e68030507d88..6be87ecffc89 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c
@@ -222,7 +222,8 @@ gk104_pci_func = {
 };
 
 int
-gk104_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+gk104_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&gk104_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&gk104_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c
index 82c5234a06ff..a5fafda0014d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c
@@ -38,7 +38,8 @@ gp100_pci_func = {
 };
 
 int
-gp100_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+gp100_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&gp100_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&gp100_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c
index 5b1ed42cb90b..9ab64194b185 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c
@@ -52,7 +52,8 @@ nv04_pci_func = {
 };
 
 int
-nv04_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+nv04_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&nv04_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&nv04_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c
index 6eb417765802..6a3c31cf0200 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c
@@ -59,7 +59,8 @@ nv40_pci_func = {
 };
 
 int
-nv40_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+nv40_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&nv40_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&nv40_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c
index fc617e4c0ab6..9cad17f178ec 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c
@@ -45,7 +45,8 @@ nv46_pci_func = {
 };
 
 int
-nv46_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+nv46_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&nv46_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&nv46_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c
index 1f1b26b5fa72..741e34bf307c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c
@@ -31,7 +31,8 @@ nv4c_pci_func = {
 };
 
 int
-nv4c_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+nv4c_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	     struct nvkm_pci **ppci)
 {
-	return nvkm_pci_new_(&nv4c_pci_func, device, index, ppci);
+	return nvkm_pci_new_(&nv4c_pci_func, device, type, inst, ppci);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h
index 7009aad86b6e..9b7583532962 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h
@@ -4,8 +4,8 @@
 #define nvkm_pci(p) container_of((p), struct nvkm_pci, subdev)
 #include <subdev/pci.h>
 
-int nvkm_pci_new_(const struct nvkm_pci_func *, struct nvkm_device *,
-		  int index, struct nvkm_pci **);
+int nvkm_pci_new_(const struct nvkm_pci_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_pci **);
 
 struct nvkm_pci_func {
 	void (*init)(struct nvkm_pci *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/base.c
index a0fe607c9c07..24382875fb4f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/base.c
@@ -148,6 +148,7 @@ nvkm_pmu_dtor(struct nvkm_subdev *subdev)
 	nvkm_falcon_cmdq_del(&pmu->hpq);
 	nvkm_falcon_qmgr_del(&pmu->qmgr);
 	nvkm_falcon_dtor(&pmu->falcon);
+	mutex_destroy(&pmu->send.mutex);
 	return nvkm_pmu(subdev);
 }
 
@@ -162,11 +163,13 @@ nvkm_pmu = {
 
 int
 nvkm_pmu_ctor(const struct nvkm_pmu_fwif *fwif, struct nvkm_device *device,
-	      int index, struct nvkm_pmu *pmu)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_pmu *pmu)
 {
 	int ret;
 
-	nvkm_subdev_ctor(&nvkm_pmu, device, index, &pmu->subdev);
+	nvkm_subdev_ctor(&nvkm_pmu, device, type, inst, &pmu->subdev);
+
+	mutex_init(&pmu->send.mutex);
 
 	INIT_WORK(&pmu->recv.work, nvkm_pmu_recv);
 	init_waitqueue_head(&pmu->recv.wait);
@@ -177,9 +180,8 @@ nvkm_pmu_ctor(const struct nvkm_pmu_fwif *fwif, struct nvkm_device *device,
 
 	pmu->func = fwif->func;
 
-	ret = nvkm_falcon_ctor(pmu->func->flcn, &pmu->subdev,
-			       nvkm_subdev_name[pmu->subdev.index], 0x10a000,
-			       &pmu->falcon);
+	ret = nvkm_falcon_ctor(pmu->func->flcn, &pmu->subdev, pmu->subdev.name,
+			       0x10a000, &pmu->falcon);
 	if (ret)
 		return ret;
 
@@ -195,10 +197,10 @@ nvkm_pmu_ctor(const struct nvkm_pmu_fwif *fwif, struct nvkm_device *device,
 
 int
 nvkm_pmu_new_(const struct nvkm_pmu_fwif *fwif, struct nvkm_device *device,
-	      int index, struct nvkm_pmu **ppmu)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_pmu **ppmu)
 {
 	struct nvkm_pmu *pmu;
 	if (!(pmu = *ppmu = kzalloc(sizeof(*pmu), GFP_KERNEL)))
 		return -ENOMEM;
-	return nvkm_pmu_ctor(fwif, device, index, *ppmu);
+	return nvkm_pmu_ctor(fwif, device, type, inst, *ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf100.c
index 3ecb3d9cbcf2..f725a3ec5479 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf100.c
@@ -30,14 +30,14 @@ void
 gf100_pmu_reset(struct nvkm_pmu *pmu)
 {
 	struct nvkm_device *device = pmu->subdev.device;
-	nvkm_mc_disable(device, NVKM_SUBDEV_PMU);
-	nvkm_mc_enable(device, NVKM_SUBDEV_PMU);
+	nvkm_mc_disable(device, NVKM_SUBDEV_PMU, 0);
+	nvkm_mc_enable(device, NVKM_SUBDEV_PMU, 0);
 }
 
 bool
 gf100_pmu_enabled(struct nvkm_pmu *pmu)
 {
-	return nvkm_mc_enabled(pmu->subdev.device, NVKM_SUBDEV_PMU);
+	return nvkm_mc_enabled(pmu->subdev.device, NVKM_SUBDEV_PMU, 0);
 }
 
 static const struct nvkm_pmu_func
@@ -69,7 +69,8 @@ gf100_pmu_fwif[] = {
 };
 
 int
-gf100_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gf100_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gf100_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gf100_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf119.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf119.c
index 8dd0271aaaee..0f4b6697a4e4 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gf119.c
@@ -47,7 +47,8 @@ gf119_pmu_fwif[] = {
 };
 
 int
-gf119_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gf119_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gf119_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gf119_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk104.c
index 8b70cc17a634..9e7631d7aa41 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk104.c
@@ -127,7 +127,8 @@ gk104_pmu_fwif[] = {
 };
 
 int
-gk104_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gk104_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gk104_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gk104_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk110.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk110.c
index 0081f2141b10..dbaefee53e1f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk110.c
@@ -106,7 +106,8 @@ gk110_pmu_fwif[] = {
 };
 
 int
-gk110_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gk110_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gk110_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gk110_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk208.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk208.c
index b227c701a5e7..a08fb049e6d6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk208.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk208.c
@@ -48,7 +48,8 @@ gk208_pmu_fwif[] = {
 };
 
 int
-gk208_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gk208_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gk208_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gk208_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk20a.c
index 26c1adf8f44c..a67a42e73f08 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gk20a.c
@@ -210,7 +210,8 @@ gk20a_pmu_fwif[] = {
 };
 
 int
-gk20a_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gk20a_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
 	struct gk20a_pmu *pmu;
 	int ret;
@@ -219,7 +220,7 @@ gk20a_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
 		return -ENOMEM;
 	*ppmu = &pmu->base;
 
-	ret = nvkm_pmu_ctor(gk20a_pmu_fwif, device, index, &pmu->base);
+	ret = nvkm_pmu_ctor(gk20a_pmu_fwif, device, type, inst, &pmu->base);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm107.c
index 5afb55e58b51..622ee637f97b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm107.c
@@ -49,7 +49,8 @@ gm107_pmu_fwif[] = {
 };
 
 int
-gm107_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gm107_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gm107_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gm107_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm200.c
index 383376addb41..5968c7696596 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm200.c
@@ -45,7 +45,8 @@ gm200_pmu_fwif[] = {
 };
 
 int
-gm200_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gm200_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gm200_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gm200_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm20b.c
index 8f6ed5373ea1..148706977eec 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gm20b.c
@@ -240,7 +240,8 @@ gm20b_pmu_fwif[] = {
 };
 
 int
-gm20b_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gm20b_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gm20b_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gm20b_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp102.c
index 3d8ce14dba7b..00da1b873ce8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp102.c
@@ -51,7 +51,8 @@ gp102_pmu_fwif[] = {
 };
 
 int
-gp102_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gp102_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gp102_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gp102_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp10b.c
index 9c237c426599..461f722656e2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gp10b.c
@@ -99,7 +99,8 @@ gp10b_pmu_fwif[] = {
 };
 
 int
-gp10b_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gp10b_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gp10b_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gp10b_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gt215.c
index 88b909913ff9..b0407b86bc10 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/gt215.c
@@ -34,7 +34,7 @@ gt215_pmu_send(struct nvkm_pmu *pmu, u32 reply[2],
 	struct nvkm_device *device = subdev->device;
 	u32 addr;
 
-	mutex_lock(&subdev->mutex);
+	mutex_lock(&pmu->send.mutex);
 	/* wait for a free slot in the fifo */
 	addr  = nvkm_rd32(device, 0x10a4a0);
 	if (nvkm_msec(device, 2000,
@@ -42,7 +42,7 @@ gt215_pmu_send(struct nvkm_pmu *pmu, u32 reply[2],
 		if (tmp != (addr ^ 8))
 			break;
 	) < 0) {
-		mutex_unlock(&subdev->mutex);
+		mutex_unlock(&pmu->send.mutex);
 		return -EBUSY;
 	}
 
@@ -79,7 +79,7 @@ gt215_pmu_send(struct nvkm_pmu *pmu, u32 reply[2],
 		reply[1] = pmu->recv.data[1];
 	}
 
-	mutex_unlock(&subdev->mutex);
+	mutex_unlock(&pmu->send.mutex);
 	return 0;
 }
 
@@ -282,7 +282,8 @@ gt215_pmu_fwif[] = {
 };
 
 int
-gt215_pmu_new(struct nvkm_device *device, int index, struct nvkm_pmu **ppmu)
+gt215_pmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_pmu **ppmu)
 {
-	return nvkm_pmu_new_(gt215_pmu_fwif, device, index, ppmu);
+	return nvkm_pmu_new_(gt215_pmu_fwif, device, type, inst, ppmu);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/priv.h
index 276b6d778e53..e7860d177353 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/priv.h
@@ -62,8 +62,8 @@ int gf100_pmu_nofw(struct nvkm_pmu *, int, const struct nvkm_pmu_fwif *);
 int gm200_pmu_nofw(struct nvkm_pmu *, int, const struct nvkm_pmu_fwif *);
 int gm20b_pmu_load(struct nvkm_pmu *, int, const struct nvkm_pmu_fwif *);
 
-int nvkm_pmu_ctor(const struct nvkm_pmu_fwif *, struct nvkm_device *,
-		  int index, struct nvkm_pmu *);
-int nvkm_pmu_new_(const struct nvkm_pmu_fwif *, struct nvkm_device *,
-		  int index, struct nvkm_pmu **);
+int nvkm_pmu_ctor(const struct nvkm_pmu_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_pmu *);
+int nvkm_pmu_new_(const struct nvkm_pmu_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_pmu **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/privring/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/Kbuild
new file mode 100644
index 000000000000..d47d1bdd0f2b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/Kbuild
@@ -0,0 +1,7 @@
+# SPDX-License-Identifier: MIT
+nvkm-y += nvkm/subdev/privring/gf100.o
+nvkm-y += nvkm/subdev/privring/gf117.o
+nvkm-y += nvkm/subdev/privring/gk104.o
+nvkm-y += nvkm/subdev/privring/gk20a.o
+nvkm-y += nvkm/subdev/privring/gm200.o
+nvkm-y += nvkm/subdev/privring/gp10b.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gf100.c
index 2340040942c9..ef7caca70372 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gf100.c
@@ -22,44 +22,42 @@
  * Authors: Ben Skeggs
  */
 #include "priv.h"
+#include <subdev/timer.h>
 
 static void
-gf100_ibus_intr_hub(struct nvkm_subdev *ibus, int i)
+gf100_privring_intr_hub(struct nvkm_subdev *privring, int i)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 addr = nvkm_rd32(device, 0x122120 + (i * 0x0400));
 	u32 data = nvkm_rd32(device, 0x122124 + (i * 0x0400));
 	u32 stat = nvkm_rd32(device, 0x122128 + (i * 0x0400));
-	nvkm_debug(ibus, "HUB%d: %06x %08x (%08x)\n", i, addr, data, stat);
-	nvkm_mask(device, 0x122128 + (i * 0x0400), 0x00000200, 0x00000000);
+	nvkm_debug(privring, "HUB%d: %06x %08x (%08x)\n", i, addr, data, stat);
 }
 
 static void
-gf100_ibus_intr_rop(struct nvkm_subdev *ibus, int i)
+gf100_privring_intr_rop(struct nvkm_subdev *privring, int i)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 addr = nvkm_rd32(device, 0x124120 + (i * 0x0400));
 	u32 data = nvkm_rd32(device, 0x124124 + (i * 0x0400));
 	u32 stat = nvkm_rd32(device, 0x124128 + (i * 0x0400));
-	nvkm_debug(ibus, "ROP%d: %06x %08x (%08x)\n", i, addr, data, stat);
-	nvkm_mask(device, 0x124128 + (i * 0x0400), 0x00000200, 0x00000000);
+	nvkm_debug(privring, "ROP%d: %06x %08x (%08x)\n", i, addr, data, stat);
 }
 
 static void
-gf100_ibus_intr_gpc(struct nvkm_subdev *ibus, int i)
+gf100_privring_intr_gpc(struct nvkm_subdev *privring, int i)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 addr = nvkm_rd32(device, 0x128120 + (i * 0x0400));
 	u32 data = nvkm_rd32(device, 0x128124 + (i * 0x0400));
 	u32 stat = nvkm_rd32(device, 0x128128 + (i * 0x0400));
-	nvkm_debug(ibus, "GPC%d: %06x %08x (%08x)\n", i, addr, data, stat);
-	nvkm_mask(device, 0x128128 + (i * 0x0400), 0x00000200, 0x00000000);
+	nvkm_debug(privring, "GPC%d: %06x %08x (%08x)\n", i, addr, data, stat);
 }
 
 void
-gf100_ibus_intr(struct nvkm_subdev *ibus)
+gf100_privring_intr(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 intr0 = nvkm_rd32(device, 0x121c58);
 	u32 intr1 = nvkm_rd32(device, 0x121c5c);
 	u32 hubnr = nvkm_rd32(device, 0x121c70);
@@ -70,7 +68,7 @@ gf100_ibus_intr(struct nvkm_subdev *ibus)
 	for (i = 0; (intr0 & 0x0000ff00) && i < hubnr; i++) {
 		u32 stat = 0x00000100 << i;
 		if (intr0 & stat) {
-			gf100_ibus_intr_hub(ibus, i);
+			gf100_privring_intr_hub(privring, i);
 			intr0 &= ~stat;
 		}
 	}
@@ -78,7 +76,7 @@ gf100_ibus_intr(struct nvkm_subdev *ibus)
 	for (i = 0; (intr0 & 0xffff0000) && i < ropnr; i++) {
 		u32 stat = 0x00010000 << i;
 		if (intr0 & stat) {
-			gf100_ibus_intr_rop(ibus, i);
+			gf100_privring_intr_rop(privring, i);
 			intr0 &= ~stat;
 		}
 	}
@@ -86,16 +84,22 @@ gf100_ibus_intr(struct nvkm_subdev *ibus)
 	for (i = 0; intr1 && i < gpcnr; i++) {
 		u32 stat = 0x00000001 << i;
 		if (intr1 & stat) {
-			gf100_ibus_intr_gpc(ibus, i);
+			gf100_privring_intr_gpc(privring, i);
 			intr1 &= ~stat;
 		}
 	}
+
+	nvkm_mask(device, 0x121c4c, 0x0000003f, 0x00000002);
+	nvkm_msec(device, 2000,
+		if (!(nvkm_rd32(device, 0x121c4c) & 0x0000003f))
+			break;
+	);
 }
 
 static int
-gf100_ibus_init(struct nvkm_subdev *ibus)
+gf100_privring_init(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	nvkm_mask(device, 0x122310, 0x0003ffff, 0x00000800);
 	nvkm_wr32(device, 0x12232c, 0x00100064);
 	nvkm_wr32(device, 0x122330, 0x00100064);
@@ -105,14 +109,14 @@ gf100_ibus_init(struct nvkm_subdev *ibus)
 }
 
 static const struct nvkm_subdev_func
-gf100_ibus = {
-	.init = gf100_ibus_init,
-	.intr = gf100_ibus_intr,
+gf100_privring = {
+	.init = gf100_privring_init,
+	.intr = gf100_privring_intr,
 };
 
 int
-gf100_ibus_new(struct nvkm_device *device, int index,
-	       struct nvkm_subdev **pibus)
+gf100_privring_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		   struct nvkm_subdev **pprivring)
 {
-	return nvkm_subdev_new_(&gf100_ibus, device, index, pibus);
+	return nvkm_subdev_new_(&gf100_privring, device, type, inst, pprivring);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gf117.c b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gf117.c
index 1124dadac145..c78721fcd729 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gf117.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gf117.c
@@ -24,9 +24,9 @@
 #include "priv.h"
 
 static int
-gf117_ibus_init(struct nvkm_subdev *ibus)
+gf117_privring_init(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	nvkm_mask(device, 0x122310, 0x0003ffff, 0x00000800);
 	nvkm_mask(device, 0x122348, 0x0003ffff, 0x00000100);
 	nvkm_mask(device, 0x1223b0, 0x0003ffff, 0x00000fff);
@@ -34,14 +34,14 @@ gf117_ibus_init(struct nvkm_subdev *ibus)
 }
 
 static const struct nvkm_subdev_func
-gf117_ibus = {
-	.init = gf117_ibus_init,
-	.intr = gf100_ibus_intr,
+gf117_privring = {
+	.init = gf117_privring_init,
+	.intr = gf100_privring_intr,
 };
 
 int
-gf117_ibus_new(struct nvkm_device *device, int index,
-	       struct nvkm_subdev **pibus)
+gf117_privring_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		   struct nvkm_subdev **pprivring)
 {
-	return nvkm_subdev_new_(&gf117_ibus, device, index, pibus);
+	return nvkm_subdev_new_(&gf117_privring, device, type, inst, pprivring);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gk104.c
index f3915f85838e..568a4c0997bd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gk104.c
@@ -22,44 +22,42 @@
  * Authors: Ben Skeggs
  */
 #include "priv.h"
+#include <subdev/timer.h>
 
 static void
-gk104_ibus_intr_hub(struct nvkm_subdev *ibus, int i)
+gk104_privring_intr_hub(struct nvkm_subdev *privring, int i)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 addr = nvkm_rd32(device, 0x122120 + (i * 0x0800));
 	u32 data = nvkm_rd32(device, 0x122124 + (i * 0x0800));
 	u32 stat = nvkm_rd32(device, 0x122128 + (i * 0x0800));
-	nvkm_debug(ibus, "HUB%d: %06x %08x (%08x)\n", i, addr, data, stat);
-	nvkm_mask(device, 0x122128 + (i * 0x0800), 0x00000200, 0x00000000);
+	nvkm_debug(privring, "HUB%d: %06x %08x (%08x)\n", i, addr, data, stat);
 }
 
 static void
-gk104_ibus_intr_rop(struct nvkm_subdev *ibus, int i)
+gk104_privring_intr_rop(struct nvkm_subdev *privring, int i)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 addr = nvkm_rd32(device, 0x124120 + (i * 0x0800));
 	u32 data = nvkm_rd32(device, 0x124124 + (i * 0x0800));
 	u32 stat = nvkm_rd32(device, 0x124128 + (i * 0x0800));
-	nvkm_debug(ibus, "ROP%d: %06x %08x (%08x)\n", i, addr, data, stat);
-	nvkm_mask(device, 0x124128 + (i * 0x0800), 0x00000200, 0x00000000);
+	nvkm_debug(privring, "ROP%d: %06x %08x (%08x)\n", i, addr, data, stat);
 }
 
 static void
-gk104_ibus_intr_gpc(struct nvkm_subdev *ibus, int i)
+gk104_privring_intr_gpc(struct nvkm_subdev *privring, int i)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 addr = nvkm_rd32(device, 0x128120 + (i * 0x0800));
 	u32 data = nvkm_rd32(device, 0x128124 + (i * 0x0800));
 	u32 stat = nvkm_rd32(device, 0x128128 + (i * 0x0800));
-	nvkm_debug(ibus, "GPC%d: %06x %08x (%08x)\n", i, addr, data, stat);
-	nvkm_mask(device, 0x128128 + (i * 0x0800), 0x00000200, 0x00000000);
+	nvkm_debug(privring, "GPC%d: %06x %08x (%08x)\n", i, addr, data, stat);
 }
 
 void
-gk104_ibus_intr(struct nvkm_subdev *ibus)
+gk104_privring_intr(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 intr0 = nvkm_rd32(device, 0x120058);
 	u32 intr1 = nvkm_rd32(device, 0x12005c);
 	u32 hubnr = nvkm_rd32(device, 0x120070);
@@ -70,7 +68,7 @@ gk104_ibus_intr(struct nvkm_subdev *ibus)
 	for (i = 0; (intr0 & 0x0000ff00) && i < hubnr; i++) {
 		u32 stat = 0x00000100 << i;
 		if (intr0 & stat) {
-			gk104_ibus_intr_hub(ibus, i);
+			gk104_privring_intr_hub(privring, i);
 			intr0 &= ~stat;
 		}
 	}
@@ -78,7 +76,7 @@ gk104_ibus_intr(struct nvkm_subdev *ibus)
 	for (i = 0; (intr0 & 0xffff0000) && i < ropnr; i++) {
 		u32 stat = 0x00010000 << i;
 		if (intr0 & stat) {
-			gk104_ibus_intr_rop(ibus, i);
+			gk104_privring_intr_rop(privring, i);
 			intr0 &= ~stat;
 		}
 	}
@@ -86,16 +84,22 @@ gk104_ibus_intr(struct nvkm_subdev *ibus)
 	for (i = 0; intr1 && i < gpcnr; i++) {
 		u32 stat = 0x00000001 << i;
 		if (intr1 & stat) {
-			gk104_ibus_intr_gpc(ibus, i);
+			gk104_privring_intr_gpc(privring, i);
 			intr1 &= ~stat;
 		}
 	}
+
+	nvkm_mask(device, 0x12004c, 0x0000003f, 0x00000002);
+	nvkm_msec(device, 2000,
+		if (!(nvkm_rd32(device, 0x12004c) & 0x0000003f))
+			break;
+	);
 }
 
 static int
-gk104_ibus_init(struct nvkm_subdev *ibus)
+gk104_privring_init(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	nvkm_mask(device, 0x122318, 0x0003ffff, 0x00001000);
 	nvkm_mask(device, 0x12231c, 0x0003ffff, 0x00000200);
 	nvkm_mask(device, 0x122310, 0x0003ffff, 0x00000800);
@@ -107,15 +111,15 @@ gk104_ibus_init(struct nvkm_subdev *ibus)
 }
 
 static const struct nvkm_subdev_func
-gk104_ibus = {
-	.preinit = gk104_ibus_init,
-	.init = gk104_ibus_init,
-	.intr = gk104_ibus_intr,
+gk104_privring = {
+	.preinit = gk104_privring_init,
+	.init = gk104_privring_init,
+	.intr = gk104_privring_intr,
 };
 
 int
-gk104_ibus_new(struct nvkm_device *device, int index,
-	       struct nvkm_subdev **pibus)
+gk104_privring_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		   struct nvkm_subdev **pprivring)
 {
-	return nvkm_subdev_new_(&gk104_ibus, device, index, pibus);
+	return nvkm_subdev_new_(&gk104_privring, device, type, inst, pprivring);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gk20a.c
index 187d544378b0..55e4a60d8770 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gk20a.c
@@ -19,13 +19,13 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  */
-#include <subdev/ibus.h>
+#include <subdev/privring.h>
 #include <subdev/timer.h>
 
 static void
-gk20a_ibus_init_ibus_ring(struct nvkm_subdev *ibus)
+gk20a_privring_init_privring_ring(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	nvkm_mask(device, 0x137250, 0x3f, 0);
 
 	nvkm_mask(device, 0x000200, 0x20, 0);
@@ -46,14 +46,14 @@ gk20a_ibus_init_ibus_ring(struct nvkm_subdev *ibus)
 }
 
 static void
-gk20a_ibus_intr(struct nvkm_subdev *ibus)
+gk20a_privring_intr(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 	u32 status0 = nvkm_rd32(device, 0x120058);
 
 	if (status0 & 0x7) {
-		nvkm_debug(ibus, "resetting ibus ring\n");
-		gk20a_ibus_init_ibus_ring(ibus);
+		nvkm_debug(privring, "resetting privring ring\n");
+		gk20a_privring_init_privring_ring(privring);
 	}
 
 	/* Acknowledge interrupt */
@@ -65,21 +65,21 @@ gk20a_ibus_intr(struct nvkm_subdev *ibus)
 }
 
 static int
-gk20a_ibus_init(struct nvkm_subdev *ibus)
+gk20a_privring_init(struct nvkm_subdev *privring)
 {
-	gk20a_ibus_init_ibus_ring(ibus);
+	gk20a_privring_init_privring_ring(privring);
 	return 0;
 }
 
 static const struct nvkm_subdev_func
-gk20a_ibus = {
-	.init = gk20a_ibus_init,
-	.intr = gk20a_ibus_intr,
+gk20a_privring = {
+	.init = gk20a_privring_init,
+	.intr = gk20a_privring_intr,
 };
 
 int
-gk20a_ibus_new(struct nvkm_device *device, int index,
-	       struct nvkm_subdev **pibus)
+gk20a_privring_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		   struct nvkm_subdev **pprivring)
 {
-	return nvkm_subdev_new_(&gk20a_ibus, device, index, pibus);
+	return nvkm_subdev_new_(&gk20a_privring, device, type, inst, pprivring);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gm200.c
index 0f1f0ad6377e..b4eaf6db36d7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gm200.c
@@ -24,13 +24,13 @@
 #include "priv.h"
 
 static const struct nvkm_subdev_func
-gm200_ibus = {
-	.intr = gk104_ibus_intr,
+gm200_privring = {
+	.intr = gk104_privring_intr,
 };
 
 int
-gm200_ibus_new(struct nvkm_device *device, int index,
-	       struct nvkm_subdev **pibus)
+gm200_privring_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		   struct nvkm_subdev **pprivring)
 {
-	return nvkm_subdev_new_(&gm200_ibus, device, index, pibus);
+	return nvkm_subdev_new_(&gm200_privring, device, type, inst, pprivring);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gp10b.c
index 0347b367cefe..4534111cf907 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ibus/gp10b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/gp10b.c
@@ -19,14 +19,14 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  */
-#include <subdev/ibus.h>
+#include <subdev/privring.h>
 
 #include "priv.h"
 
 static int
-gp10b_ibus_init(struct nvkm_subdev *ibus)
+gp10b_privring_init(struct nvkm_subdev *privring)
 {
-	struct nvkm_device *device = ibus->device;
+	struct nvkm_device *device = privring->device;
 
 	nvkm_wr32(device, 0x1200a8, 0x0);
 
@@ -42,14 +42,14 @@ gp10b_ibus_init(struct nvkm_subdev *ibus)
 }
 
 static const struct nvkm_subdev_func
-gp10b_ibus = {
-	.init = gp10b_ibus_init,
-	.intr = gk104_ibus_intr,
+gp10b_privring = {
+	.init = gp10b_privring_init,
+	.intr = gk104_privring_intr,
 };
 
 int
-gp10b_ibus_new(struct nvkm_device *device, int index,
-	       struct nvkm_subdev **pibus)
+gp10b_privring_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		   struct nvkm_subdev **pprivring)
 {
-	return nvkm_subdev_new_(&gp10b_ibus, device, index, pibus);
+	return nvkm_subdev_new_(&gp10b_privring, device, type, inst, pprivring);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/privring/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/priv.h
new file mode 100644
index 000000000000..b378c14bc8dc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/privring/priv.h
@@ -0,0 +1,8 @@
+/* SPDX-License-Identifier: MIT */
+#ifndef __NVKM_PRIVRING_PRIV_H__
+#define __NVKM_PRIVRING_PRIV_H__
+#include <subdev/privring.h>
+
+void gf100_privring_intr(struct nvkm_subdev *);
+void gk104_privring_intr(struct nvkm_subdev *);
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c
index 4a4d1e224126..fc5ee118e910 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c
@@ -421,10 +421,10 @@ nvkm_therm = {
 };
 
 void
-nvkm_therm_ctor(struct nvkm_therm *therm, struct nvkm_device *device,
-		int index, const struct nvkm_therm_func *func)
+nvkm_therm_ctor(struct nvkm_therm *therm, struct nvkm_device *device, enum nvkm_subdev_type type,
+		int inst, const struct nvkm_therm_func *func)
 {
-	nvkm_subdev_ctor(&nvkm_therm, device, index, &therm->subdev);
+	nvkm_subdev_ctor(&nvkm_therm, device, type, inst, &therm->subdev);
 	therm->func = func;
 
 	nvkm_alarm_init(&therm->alarm, nvkm_therm_alarm);
@@ -443,13 +443,13 @@ nvkm_therm_ctor(struct nvkm_therm *therm, struct nvkm_device *device,
 
 int
 nvkm_therm_new_(const struct nvkm_therm_func *func, struct nvkm_device *device,
-		int index, struct nvkm_therm **ptherm)
+		enum nvkm_subdev_type type, int inst, struct nvkm_therm **ptherm)
 {
 	struct nvkm_therm *therm;
 
 	if (!(therm = *ptherm = kzalloc(sizeof(*therm), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_therm_ctor(therm, device, index, func);
+	nvkm_therm_ctor(therm, device, type, inst, func);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/g84.c
index 96f8da40ac82..4af86f2d3e7e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/g84.c
@@ -223,12 +223,13 @@ g84_therm = {
 };
 
 int
-g84_therm_new(struct nvkm_device *device, int index, struct nvkm_therm **ptherm)
+g84_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_therm **ptherm)
 {
 	struct nvkm_therm *therm;
 	int ret;
 
-	ret = nvkm_therm_new_(&g84_therm, device, index, &therm);
+	ret = nvkm_therm_new_(&g84_therm, device, type, inst, &therm);
 	*ptherm = therm;
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gf119.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gf119.c
index 0981b02790e2..2b031d4eaeb6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gf119.c
@@ -146,8 +146,8 @@ gf119_therm = {
 };
 
 int
-gf119_therm_new(struct nvkm_device *device, int index,
-	       struct nvkm_therm **ptherm)
+gf119_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&gf119_therm, device, index, ptherm);
+	return nvkm_therm_new_(&gf119_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.c
index 4e03971d2e3d..45e295c271fb 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.c
@@ -35,8 +35,8 @@ gk104_clkgate_enable(struct nvkm_therm *base)
 	int i;
 
 	/* Program ENG_MANT, ENG_FILTER */
-	for (i = 0; order[i].engine != NVKM_SUBDEV_NR; i++) {
-		if (!nvkm_device_subdev(dev, order[i].engine))
+	for (i = 0; order[i].type != NVKM_SUBDEV_NR; i++) {
+		if (!nvkm_device_subdev(dev, order[i].type, order[i].inst))
 			continue;
 
 		nvkm_mask(dev, 0x20200 + order[i].offset, 0xff00, 0x4500);
@@ -47,8 +47,8 @@ gk104_clkgate_enable(struct nvkm_therm *base)
 	nvkm_wr32(dev, 0x02028c, therm->idle_filter->hubmmu);
 
 	/* Enable clockgating (ENG_CLK = RUN->AUTO) */
-	for (i = 0; order[i].engine != NVKM_SUBDEV_NR; i++) {
-		if (!nvkm_device_subdev(dev, order[i].engine))
+	for (i = 0; order[i].type != NVKM_SUBDEV_NR; i++) {
+		if (!nvkm_device_subdev(dev, order[i].type, order[i].inst))
 			continue;
 
 		nvkm_mask(dev, 0x20200 + order[i].offset, 0x00ff, 0x0045);
@@ -64,8 +64,8 @@ gk104_clkgate_fini(struct nvkm_therm *base, bool suspend)
 	int i;
 
 	/* ENG_CLK = AUTO->RUN, ENG_PWR = RUN->AUTO */
-	for (i = 0; order[i].engine != NVKM_SUBDEV_NR; i++) {
-		if (!nvkm_device_subdev(dev, order[i].engine))
+	for (i = 0; order[i].type != NVKM_SUBDEV_NR; i++) {
+		if (!nvkm_device_subdev(dev, order[i].type, order[i].inst))
 			continue;
 
 		nvkm_mask(dev, 0x20200 + order[i].offset, 0xff, 0x54);
@@ -73,15 +73,15 @@ gk104_clkgate_fini(struct nvkm_therm *base, bool suspend)
 }
 
 const struct gk104_clkgate_engine_info gk104_clkgate_engine_info[] = {
-	{ NVKM_ENGINE_GR,     0x00 },
-	{ NVKM_ENGINE_MSPDEC, 0x04 },
-	{ NVKM_ENGINE_MSPPP,  0x08 },
-	{ NVKM_ENGINE_MSVLD,  0x0c },
-	{ NVKM_ENGINE_CE0,    0x10 },
-	{ NVKM_ENGINE_CE1,    0x14 },
-	{ NVKM_ENGINE_MSENC,  0x18 },
-	{ NVKM_ENGINE_CE2,    0x1c },
-	{ NVKM_SUBDEV_NR, 0 },
+	{ NVKM_ENGINE_GR,     0, 0x00 },
+	{ NVKM_ENGINE_MSPDEC, 0, 0x04 },
+	{ NVKM_ENGINE_MSPPP,  0, 0x08 },
+	{ NVKM_ENGINE_MSVLD,  0, 0x0c },
+	{ NVKM_ENGINE_CE,     0, 0x10 },
+	{ NVKM_ENGINE_CE,     1, 0x14 },
+	{ NVKM_ENGINE_MSENC,  0, 0x18 },
+	{ NVKM_ENGINE_CE,     2, 0x1c },
+	{ NVKM_SUBDEV_NR },
 };
 
 const struct gf100_idle_filter gk104_idle_filter = {
@@ -106,9 +106,8 @@ gk104_therm_func = {
 };
 
 static int
-gk104_therm_new_(const struct nvkm_therm_func *func,
-		 struct nvkm_device *device,
-		 int index,
+gk104_therm_new_(const struct nvkm_therm_func *func, struct nvkm_device *device,
+		 enum nvkm_subdev_type type, int inst,
 		 const struct gk104_clkgate_engine_info *clkgate_order,
 		 const struct gf100_idle_filter *idle_filter,
 		 struct nvkm_therm **ptherm)
@@ -118,19 +117,17 @@ gk104_therm_new_(const struct nvkm_therm_func *func,
 	if (!therm)
 		return -ENOMEM;
 
-	nvkm_therm_ctor(&therm->base, device, index, func);
+	nvkm_therm_ctor(&therm->base, device, type, inst, func);
 	*ptherm = &therm->base;
 	therm->clkgate_order = clkgate_order;
 	therm->idle_filter = idle_filter;
-
 	return 0;
 }
 
 int
-gk104_therm_new(struct nvkm_device *device,
-		int index, struct nvkm_therm **ptherm)
+gk104_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_therm **ptherm)
 {
-	return gk104_therm_new_(&gk104_therm_func, device, index,
+	return gk104_therm_new_(&gk104_therm_func, device, type, inst,
 				gk104_clkgate_engine_info, &gk104_idle_filter,
 				ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.h b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.h
index 293e7743b19b..9a8641421038 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gk104.h
@@ -31,7 +31,8 @@
 #include "gf100.h"
 
 struct gk104_clkgate_engine_info {
-	enum nvkm_devidx engine;
+	enum nvkm_subdev_type type;
+	int inst;
 	u8 offset;
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm107.c
index 86848ece4d89..c845fd392f58 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm107.c
@@ -68,8 +68,8 @@ gm107_therm = {
 };
 
 int
-gm107_therm_new(struct nvkm_device *device, int index,
+gm107_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&gm107_therm, device, index, ptherm);
+	return nvkm_therm_new_(&gm107_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm200.c
index 73dc78093d5d..e0cdd12463ec 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gm200.c
@@ -32,8 +32,8 @@ gm200_therm = {
 };
 
 int
-gm200_therm_new(struct nvkm_device *device, int index,
+gm200_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&gm200_therm, device, index, ptherm);
+	return nvkm_therm_new_(&gm200_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gp100.c
index 9f0dea3f61dc..44f021392b95 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gp100.c
@@ -49,8 +49,8 @@ gp100_therm = {
 };
 
 int
-gp100_therm_new(struct nvkm_device *device, int index,
+gp100_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&gp100_therm, device, index, ptherm);
+	return nvkm_therm_new_(&gp100_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gt215.c
index c08097f2aff5..9e451bd9395c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/gt215.c
@@ -68,8 +68,8 @@ gt215_therm = {
 };
 
 int
-gt215_therm_new(struct nvkm_device *device, int index,
-	       struct nvkm_therm **ptherm)
+gt215_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&gt215_therm, device, index, ptherm);
+	return nvkm_therm_new_(&gt215_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv40.c
index 2c92ffb5f9d0..c13fee9734df 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv40.c
@@ -197,8 +197,8 @@ nv40_therm = {
 };
 
 int
-nv40_therm_new(struct nvkm_device *device, int index,
+nv40_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	       struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&nv40_therm, device, index, ptherm);
+	return nvkm_therm_new_(&nv40_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv50.c
index 9b57b433d4cf..9cf16a75a3cd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/nv50.c
@@ -169,8 +169,8 @@ nv50_therm = {
 };
 
 int
-nv50_therm_new(struct nvkm_device *device, int index,
+nv50_therm_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	       struct nvkm_therm **ptherm)
 {
-	return nvkm_therm_new_(&nv50_therm, device, index, ptherm);
+	return nvkm_therm_new_(&nv50_therm, device, type, inst, ptherm);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/priv.h
index 21659daf1864..54e960589411 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/priv.h
@@ -30,10 +30,10 @@
 #include <subdev/bios/gpio.h>
 #include <subdev/bios/perf.h>
 
-int nvkm_therm_new_(const struct nvkm_therm_func *, struct nvkm_device *,
-		    int index, struct nvkm_therm **);
-void nvkm_therm_ctor(struct nvkm_therm *therm, struct nvkm_device *device,
-		     int index, const struct nvkm_therm_func *func);
+int nvkm_therm_new_(const struct nvkm_therm_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		    int, struct nvkm_therm **);
+void nvkm_therm_ctor(struct nvkm_therm *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		     const struct nvkm_therm_func *);
 
 struct nvkm_fan {
 	struct nvkm_therm *parent;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c
index dd922033628c..8b0da0c06268 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c
@@ -183,14 +183,14 @@ nvkm_timer = {
 
 int
 nvkm_timer_new_(const struct nvkm_timer_func *func, struct nvkm_device *device,
-		int index, struct nvkm_timer **ptmr)
+		enum nvkm_subdev_type type, int inst, struct nvkm_timer **ptmr)
 {
 	struct nvkm_timer *tmr;
 
 	if (!(tmr = *ptmr = kzalloc(sizeof(*tmr), GFP_KERNEL)))
 		return -ENOMEM;
 
-	nvkm_subdev_ctor(&nvkm_timer, device, index, &tmr->subdev);
+	nvkm_subdev_ctor(&nvkm_timer, device, type, inst, &tmr->subdev);
 	tmr->func = func;
 	INIT_LIST_HEAD(&tmr->alarms);
 	spin_lock_init(&tmr->lock);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/gk20a.c
index 9ed5f64912d0..73c3776b6b83 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/gk20a.c
@@ -33,7 +33,8 @@ gk20a_timer = {
 };
 
 int
-gk20a_timer_new(struct nvkm_device *device, int index, struct nvkm_timer **ptmr)
+gk20a_timer_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+		struct nvkm_timer **ptmr)
 {
-	return nvkm_timer_new_(&gk20a_timer, device, index, ptmr);
+	return nvkm_timer_new_(&gk20a_timer, device, type, inst, ptmr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv04.c
index 7f48249f41de..0058e856b378 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv04.c
@@ -145,7 +145,8 @@ nv04_timer = {
 };
 
 int
-nv04_timer_new(struct nvkm_device *device, int index, struct nvkm_timer **ptmr)
+nv04_timer_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_timer **ptmr)
 {
-	return nvkm_timer_new_(&nv04_timer, device, index, ptmr);
+	return nvkm_timer_new_(&nv04_timer, device, type, inst, ptmr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv40.c
index bb99a152f26e..7e1f8c22f2a8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv40.c
@@ -82,7 +82,8 @@ nv40_timer = {
 };
 
 int
-nv40_timer_new(struct nvkm_device *device, int index, struct nvkm_timer **ptmr)
+nv40_timer_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_timer **ptmr)
 {
-	return nvkm_timer_new_(&nv40_timer, device, index, ptmr);
+	return nvkm_timer_new_(&nv40_timer, device, type, inst, ptmr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv41.c b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv41.c
index 3cf9ec1b1b57..c2b263721f10 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv41.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/nv41.c
@@ -79,7 +79,8 @@ nv41_timer = {
 };
 
 int
-nv41_timer_new(struct nvkm_device *device, int index, struct nvkm_timer **ptmr)
+nv41_timer_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_timer **ptmr)
 {
-	return nvkm_timer_new_(&nv41_timer, device, index, ptmr);
+	return nvkm_timer_new_(&nv41_timer, device, type, inst, ptmr);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/priv.h
index 89e97294b182..e6debe7e2fa9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/priv.h
@@ -4,8 +4,8 @@
 #define nvkm_timer(p) container_of((p), struct nvkm_timer, subdev)
 #include <subdev/timer.h>
 
-int nvkm_timer_new_(const struct nvkm_timer_func *, struct nvkm_device *,
-		    int index, struct nvkm_timer **);
+int nvkm_timer_new_(const struct nvkm_timer_func *, struct nvkm_device *, enum nvkm_subdev_type,
+		    int, struct nvkm_timer **);
 
 struct nvkm_timer_func {
 	void (*init)(struct nvkm_timer *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/top/Kbuild
index 438d9d78ab52..d5db845195dc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/Kbuild
@@ -1,3 +1,4 @@
 # SPDX-License-Identifier: MIT
 nvkm-y += nvkm/subdev/top/base.o
 nvkm-y += nvkm/subdev/top/gk104.o
+nvkm-y += nvkm/subdev/top/ga100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c
index cce6e4e90ebf..28d0789f50fe 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c
@@ -28,7 +28,8 @@ nvkm_top_device_new(struct nvkm_top *top)
 {
 	struct nvkm_top_device *info = kmalloc(sizeof(*info), GFP_KERNEL);
 	if (info) {
-		info->index = NVKM_SUBDEV_NR;
+		info->type = NVKM_SUBDEV_NR;
+		info->inst = -1;
 		info->addr = 0;
 		info->fault = -1;
 		info->engine = -1;
@@ -41,14 +42,14 @@ nvkm_top_device_new(struct nvkm_top *top)
 }
 
 u32
-nvkm_top_addr(struct nvkm_device *device, enum nvkm_devidx index)
+nvkm_top_addr(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
 	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 
 	if (top) {
 		list_for_each_entry(info, &top->device, head) {
-			if (info->index == index)
+			if (info->type == type && info->inst == inst)
 				return info->addr;
 		}
 	}
@@ -57,14 +58,14 @@ nvkm_top_addr(struct nvkm_device *device, enum nvkm_devidx index)
 }
 
 u32
-nvkm_top_reset(struct nvkm_device *device, enum nvkm_devidx index)
+nvkm_top_reset(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
 	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 
 	if (top) {
 		list_for_each_entry(info, &top->device, head) {
-			if (info->index == index && info->reset >= 0)
+			if (info->type == type && info->inst == inst && info->reset >= 0)
 				return BIT(info->reset);
 		}
 	}
@@ -73,14 +74,14 @@ nvkm_top_reset(struct nvkm_device *device, enum nvkm_devidx index)
 }
 
 u32
-nvkm_top_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx)
+nvkm_top_intr_mask(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
 	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 
 	if (top) {
 		list_for_each_entry(info, &top->device, head) {
-			if (info->index == devidx && info->intr >= 0)
+			if (info->type == type && info->inst == inst && info->intr >= 0)
 				return BIT(info->intr);
 		}
 	}
@@ -88,44 +89,21 @@ nvkm_top_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx)
 	return 0;
 }
 
-u32
-nvkm_top_intr(struct nvkm_device *device, u32 intr, u64 *psubdevs)
-{
-	struct nvkm_top *top = device->top;
-	struct nvkm_top_device *info;
-	u64 subdevs = 0;
-	u32 handled = 0;
-
-	if (top) {
-		list_for_each_entry(info, &top->device, head) {
-			if (info->index != NVKM_SUBDEV_NR && info->intr >= 0) {
-				if (intr & BIT(info->intr)) {
-					subdevs |= BIT_ULL(info->index);
-					handled |= BIT(info->intr);
-				}
-			}
-		}
-	}
-
-	*psubdevs = subdevs;
-	return intr & ~handled;
-}
-
 int
-nvkm_top_fault_id(struct nvkm_device *device, enum nvkm_devidx devidx)
+nvkm_top_fault_id(struct nvkm_device *device, enum nvkm_subdev_type type, int inst)
 {
 	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 
 	list_for_each_entry(info, &top->device, head) {
-		if (info->index == devidx && info->fault >= 0)
+		if (info->type == type && info->inst == inst && info->fault >= 0)
 			return info->fault;
 	}
 
 	return -ENOENT;
 }
 
-enum nvkm_devidx
+struct nvkm_subdev *
 nvkm_top_fault(struct nvkm_device *device, int fault)
 {
 	struct nvkm_top *top = device->top;
@@ -133,28 +111,10 @@ nvkm_top_fault(struct nvkm_device *device, int fault)
 
 	list_for_each_entry(info, &top->device, head) {
 		if (info->fault == fault)
-			return info->index;
-	}
-
-	return NVKM_SUBDEV_NR;
-}
-
-enum nvkm_devidx
-nvkm_top_engine(struct nvkm_device *device, int index, int *runl, int *engn)
-{
-	struct nvkm_top *top = device->top;
-	struct nvkm_top_device *info;
-	int n = 0;
-
-	list_for_each_entry(info, &top->device, head) {
-		if (info->engine >= 0 && info->runlist >= 0 && n++ == index) {
-			*runl = info->runlist;
-			*engn = info->engine;
-			return info->index;
-		}
+			return nvkm_device_subdev(device, info->type, info->inst);
 	}
 
-	return -ENODEV;
+	return NULL;
 }
 
 static int
@@ -186,12 +146,12 @@ nvkm_top = {
 
 int
 nvkm_top_new_(const struct nvkm_top_func *func, struct nvkm_device *device,
-	      int index, struct nvkm_top **ptop)
+	      enum nvkm_subdev_type type, int inst, struct nvkm_top **ptop)
 {
 	struct nvkm_top *top;
 	if (!(top = *ptop = kzalloc(sizeof(*top), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_subdev_ctor(&nvkm_top, device, index, &top->subdev);
+	nvkm_subdev_ctor(&nvkm_top, device, type, inst, &top->subdev);
 	top->func = func;
 	INIT_LIST_HEAD(&top->device);
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/ga100.c
new file mode 100644
index 000000000000..31933f3e5a07
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/ga100.c
@@ -0,0 +1,107 @@
+/*
+ * Copyright 2021 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "priv.h"
+
+static int
+ga100_top_oneinit(struct nvkm_top *top)
+{
+	struct nvkm_subdev *subdev = &top->subdev;
+	struct nvkm_device *device = subdev->device;
+	struct nvkm_top_device *info = NULL;
+	u32 data, type, inst;
+	int i, n, size = nvkm_rd32(device, 0x0224fc) >> 20;
+
+	for (i = 0, n = 0; i < size; i++) {
+		if (!info) {
+			if (!(info = nvkm_top_device_new(top)))
+				return -ENOMEM;
+			type = ~0;
+			inst = 0;
+		}
+
+		data = nvkm_rd32(device, 0x022800 + (i * 0x04));
+		nvkm_trace(subdev, "%02x: %08x\n", i, data);
+		if (!data && n == 0)
+			continue;
+
+		switch (n++) {
+		case 0:
+			type	      = (data & 0x3f000000) >> 24;
+			inst	      = (data & 0x000f0000) >> 16;
+			info->fault   = (data & 0x0000007f);
+			break;
+		case 1:
+			info->addr    = (data & 0x00fff000);
+			info->reset   = (data & 0x0000001f);
+			break;
+		case 2:
+			info->runlist = (data & 0x0000fc00) >> 10;
+			info->engine  = (data & 0x00000003);
+			break;
+		default:
+			break;
+		}
+
+		if (data & 0x80000000)
+			continue;
+		n = 0;
+
+		/* Translate engine type to NVKM engine identifier. */
+#define I_(T,I) do { info->type = (T); info->inst = (I); } while(0)
+#define O_(T,I) do { WARN_ON(inst); I_(T, I); } while (0)
+		switch (type) {
+		case 0x00000000: O_(NVKM_ENGINE_GR    ,    0); break;
+		case 0x0000000d: O_(NVKM_ENGINE_SEC2  ,    0); break;
+		case 0x0000000e: I_(NVKM_ENGINE_NVENC , inst); break;
+		case 0x00000010: I_(NVKM_ENGINE_NVDEC , inst); break;
+		case 0x00000012: I_(NVKM_SUBDEV_IOCTRL, inst); break;
+		case 0x00000013: I_(NVKM_ENGINE_CE    , inst); break;
+		case 0x00000014: O_(NVKM_SUBDEV_GSP   ,    0); break;
+		case 0x00000015: O_(NVKM_ENGINE_NVJPG ,    0); break;
+		case 0x00000016: O_(NVKM_ENGINE_OFA   ,    0); break;
+		case 0x00000017: O_(NVKM_SUBDEV_FLA   ,    0); break;
+			break;
+		default:
+			break;
+		}
+
+		nvkm_debug(subdev, "%02x.%d (%8s): addr %06x fault %2d "
+				   "runlist %2d engine %2d reset %2d\n", type, inst,
+			   info->type == NVKM_SUBDEV_NR ? "????????" : nvkm_subdev_type[info->type],
+			   info->addr, info->fault, info->runlist, info->engine, info->reset);
+		info = NULL;
+	}
+
+	return 0;
+}
+
+static const struct nvkm_top_func
+ga100_top = {
+	.oneinit = ga100_top_oneinit,
+};
+
+int
+ga100_top_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_top **ptop)
+{
+	return nvkm_top_new_(&ga100_top, device, type, inst, ptop);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c
index 1156634533f9..4dcad97bd505 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c
@@ -70,26 +70,26 @@ gk104_top_oneinit(struct nvkm_top *top)
 			continue;
 
 		/* Translate engine type to NVKM engine identifier. */
-#define A_(A) if (inst == 0) info->index = NVKM_ENGINE_##A
-#define B_(A) if (inst + NVKM_ENGINE_##A##0 < NVKM_ENGINE_##A##_LAST + 1)      \
-		info->index = NVKM_ENGINE_##A##0 + inst
-#define C_(A) if (inst == 0) info->index = NVKM_SUBDEV_##A
+#define I_(T,I) do { info->type = (T); info->inst = (I); } while(0)
+#define O_(T,I) do { WARN_ON(inst); I_(T, I); } while (0)
 		switch (type) {
-		case 0x00000000: A_(GR    ); break;
-		case 0x00000001: A_(CE0   ); break;
-		case 0x00000002: A_(CE1   ); break;
-		case 0x00000003: A_(CE2   ); break;
-		case 0x00000008: A_(MSPDEC); break;
-		case 0x00000009: A_(MSPPP ); break;
-		case 0x0000000a: A_(MSVLD ); break;
-		case 0x0000000b: A_(MSENC ); break;
-		case 0x0000000c: A_(VIC   ); break;
-		case 0x0000000d: A_(SEC2  ); break;
-		case 0x0000000e: B_(NVENC ); break;
-		case 0x0000000f: A_(NVENC1); break;
-		case 0x00000010: B_(NVDEC ); break;
-		case 0x00000013: B_(CE    ); break;
-		case 0x00000014: C_(GSP   ); break;
+		case 0x00000000: O_(NVKM_ENGINE_GR    ,    0); break;
+		case 0x00000001: O_(NVKM_ENGINE_CE    ,    0); break;
+		case 0x00000002: O_(NVKM_ENGINE_CE    ,    1); break;
+		case 0x00000003: O_(NVKM_ENGINE_CE    ,    2); break;
+		case 0x00000008: O_(NVKM_ENGINE_MSPDEC,    0); break;
+		case 0x00000009: O_(NVKM_ENGINE_MSPPP ,    0); break;
+		case 0x0000000a: O_(NVKM_ENGINE_MSVLD ,    0); break;
+		case 0x0000000b: O_(NVKM_ENGINE_MSENC ,    0); break;
+		case 0x0000000c: O_(NVKM_ENGINE_VIC   ,    0); break;
+		case 0x0000000d: O_(NVKM_ENGINE_SEC2  ,    0); break;
+		case 0x0000000e: I_(NVKM_ENGINE_NVENC , inst); break;
+		case 0x0000000f: O_(NVKM_ENGINE_NVENC ,    1); break;
+		case 0x00000010: I_(NVKM_ENGINE_NVDEC , inst); break;
+		case 0x00000012: I_(NVKM_SUBDEV_IOCTRL, inst); break;
+		case 0x00000013: I_(NVKM_ENGINE_CE    , inst); break;
+		case 0x00000014: O_(NVKM_SUBDEV_GSP   ,    0); break;
+		case 0x00000015: O_(NVKM_ENGINE_NVJPG ,    0); break;
 		default:
 			break;
 		}
@@ -97,8 +97,7 @@ gk104_top_oneinit(struct nvkm_top *top)
 		nvkm_debug(subdev, "%02x.%d (%8s): addr %06x fault %2d "
 				   "engine %2d runlist %2d intr %2d "
 				   "reset %2d\n", type, inst,
-			   info->index == NVKM_SUBDEV_NR ? NULL :
-					  nvkm_subdev_name[info->index],
+			   info->type == NVKM_SUBDEV_NR ? "????????" : nvkm_subdev_type[info->type],
 			   info->addr, info->fault, info->engine, info->runlist,
 			   info->intr, info->reset);
 		info = NULL;
@@ -113,7 +112,8 @@ gk104_top = {
 };
 
 int
-gk104_top_new(struct nvkm_device *device, int index, struct nvkm_top **ptop)
+gk104_top_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_top **ptop)
 {
-	return nvkm_top_new_(&gk104_top, device, index, ptop);
+	return nvkm_top_new_(&gk104_top, device, type, inst, ptop);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/top/priv.h
index a16baa2941cf..8e103a836705 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/priv.h
@@ -8,19 +8,8 @@ struct nvkm_top_func {
 	int (*oneinit)(struct nvkm_top *);
 };
 
-int nvkm_top_new_(const struct nvkm_top_func *, struct nvkm_device *,
-		  int, struct nvkm_top **);
-
-struct nvkm_top_device {
-	enum nvkm_devidx index;
-	u32 addr;
-	int fault;
-	int engine;
-	int runlist;
-	int reset;
-	int intr;
-	struct list_head head;
-};
+int nvkm_top_new_(const struct nvkm_top_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		  struct nvkm_top **);
 
 struct nvkm_top_device *nvkm_top_device_new(struct nvkm_top *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c
index e344901cfdc7..a17a6dd8d3de 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c
@@ -281,12 +281,12 @@ nvkm_volt = {
 
 void
 nvkm_volt_ctor(const struct nvkm_volt_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_volt *volt)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_volt *volt)
 {
 	struct nvkm_bios *bios = device->bios;
 	int i;
 
-	nvkm_subdev_ctor(&nvkm_volt, device, index, &volt->subdev);
+	nvkm_subdev_ctor(&nvkm_volt, device, type, inst, &volt->subdev);
 	volt->func = func;
 
 	/* Assuming the non-bios device should build the voltage table later */
@@ -319,10 +319,10 @@ nvkm_volt_ctor(const struct nvkm_volt_func *func, struct nvkm_device *device,
 
 int
 nvkm_volt_new_(const struct nvkm_volt_func *func, struct nvkm_device *device,
-	       int index, struct nvkm_volt **pvolt)
+	       enum nvkm_subdev_type type, int inst, struct nvkm_volt **pvolt)
 {
 	if (!(*pvolt = kzalloc(sizeof(**pvolt), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_volt_ctor(func, device, index, *pvolt);
+	nvkm_volt_ctor(func, device, type, inst, *pvolt);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf100.c
index d9ed6925ca64..b47a1c0817be 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf100.c
@@ -56,12 +56,13 @@ gf100_volt = {
 };
 
 int
-gf100_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
+gf100_volt_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_volt **pvolt)
 {
 	struct nvkm_volt *volt;
 	int ret;
 
-	ret = nvkm_volt_new_(&gf100_volt, device, index, &volt);
+	ret = nvkm_volt_new_(&gf100_volt, device, type, inst, &volt);
 	*pvolt = volt;
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf117.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf117.c
index 547a58f0aeac..03c8a2c2916c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf117.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gf117.c
@@ -46,12 +46,13 @@ gf117_volt = {
 };
 
 int
-gf117_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
+gf117_volt_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_volt **pvolt)
 {
 	struct nvkm_volt *volt;
 	int ret;
 
-	ret = nvkm_volt_new_(&gf117_volt, device, index, &volt);
+	ret = nvkm_volt_new_(&gf117_volt, device, type, inst, &volt);
 	*pvolt = volt;
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk104.c
index 1c744e029454..d1ce4309cfb8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk104.c
@@ -95,7 +95,8 @@ gk104_volt_pwm = {
 };
 
 int
-gk104_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
+gk104_volt_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_volt **pvolt)
 {
 	const struct nvkm_volt_func *volt_func = &gk104_volt_gpio;
 	struct dcb_gpio_func gpio;
@@ -114,7 +115,7 @@ gk104_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
 
 	if (!(volt = kzalloc(sizeof(*volt), GFP_KERNEL)))
 		return -ENOMEM;
-	nvkm_volt_ctor(volt_func, device, index, &volt->base);
+	nvkm_volt_ctor(volt_func, device, type, inst, &volt->base);
 	*pvolt = &volt->base;
 	volt->bios = bios;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c
index ce5d83cdc7cf..8c2faa964511 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c
@@ -144,14 +144,14 @@ gk20a_volt = {
 };
 
 int
-gk20a_volt_ctor(struct nvkm_device *device, int index,
+gk20a_volt_ctor(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 		const struct cvb_coef *coefs, int nb_coefs,
 		int vmin, struct gk20a_volt *volt)
 {
 	struct nvkm_device_tegra *tdev = device->func->tegra(device);
 	int i, uv;
 
-	nvkm_volt_ctor(&gk20a_volt, device, index, &volt->base);
+	nvkm_volt_ctor(&gk20a_volt, device, type, inst, &volt->base);
 
 	uv = regulator_get_voltage(tdev->vdd);
 	nvkm_debug(&volt->base.subdev, "the default voltage is %duV\n", uv);
@@ -172,7 +172,7 @@ gk20a_volt_ctor(struct nvkm_device *device, int index,
 }
 
 int
-gk20a_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
+gk20a_volt_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_volt **pvolt)
 {
 	struct gk20a_volt *volt;
 
@@ -181,6 +181,6 @@ gk20a_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
 		return -ENOMEM;
 	*pvolt = &volt->base;
 
-	return gk20a_volt_ctor(device, index, gk20a_cvb_coef,
+	return gk20a_volt_ctor(device, type, inst, gk20a_cvb_coef,
 			       ARRAY_SIZE(gk20a_cvb_coef), 0, volt);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h
index 6a6c97f9684e..01f8a5fcf496 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h
@@ -37,7 +37,7 @@ struct gk20a_volt {
 	struct regulator *vdd;
 };
 
-int gk20a_volt_ctor(struct nvkm_device *device, int index,
+int gk20a_volt_ctor(struct nvkm_device *device, enum nvkm_subdev_type, int,
 		    const struct cvb_coef *coefs, int nb_coefs,
 		    int vmin, struct gk20a_volt *volt);
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c
index 2925b9cae681..c2e9694d333f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c
@@ -64,7 +64,8 @@ static const u32 speedo_to_vmin[] = {
 };
 
 int
-gm20b_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
+gm20b_volt_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	       struct nvkm_volt **pvolt)
 {
 	struct nvkm_device_tegra *tdev = device->func->tegra(device);
 	struct gk20a_volt *volt;
@@ -84,9 +85,9 @@ gm20b_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
 	vmin = speedo_to_vmin[tdev->gpu_speedo_id];
 
 	if (tdev->gpu_speedo_id >= 1)
-		return gk20a_volt_ctor(device, index, gm20b_na_cvb_coef,
-				     ARRAY_SIZE(gm20b_na_cvb_coef), vmin, volt);
+		return gk20a_volt_ctor(device, type, inst, gm20b_na_cvb_coef,
+				       ARRAY_SIZE(gm20b_na_cvb_coef), vmin, volt);
 	else
-		return gk20a_volt_ctor(device, index, gm20b_cvb_coef,
-					ARRAY_SIZE(gm20b_cvb_coef), vmin, volt);
+		return gk20a_volt_ctor(device, type, inst, gm20b_cvb_coef,
+				       ARRAY_SIZE(gm20b_cvb_coef), vmin, volt);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/nv40.c
index 23409387abb5..d6a587d6082d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/nv40.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/nv40.c
@@ -30,12 +30,13 @@ nv40_volt = {
 };
 
 int
-nv40_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
+nv40_volt_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
+	      struct nvkm_volt **pvolt)
 {
 	struct nvkm_volt *volt;
 	int ret;
 
-	ret = nvkm_volt_new_(&nv40_volt, device, index, &volt);
+	ret = nvkm_volt_new_(&nv40_volt, device, type, inst, &volt);
 	*pvolt = volt;
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/priv.h
index 75f13a34671f..24e2d16d1913 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/priv.h
@@ -4,10 +4,10 @@
 #define nvkm_volt(p) container_of((p), struct nvkm_volt, subdev)
 #include <subdev/volt.h>
 
-void nvkm_volt_ctor(const struct nvkm_volt_func *, struct nvkm_device *,
-		    int index, struct nvkm_volt *);
-int nvkm_volt_new_(const struct nvkm_volt_func *, struct nvkm_device *,
-		   int index, struct nvkm_volt **);
+void nvkm_volt_ctor(const struct nvkm_volt_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		    struct nvkm_volt *);
+int nvkm_volt_new_(const struct nvkm_volt_func *, struct nvkm_device *, enum nvkm_subdev_type, int,
+		   struct nvkm_volt **);
 
 struct nvkm_volt_func {
 	int (*oneinit)(struct nvkm_volt *);
diff --git a/drivers/gpu/drm/panel/panel-elida-kd35t133.c b/drivers/gpu/drm/panel/panel-elida-kd35t133.c
index bc36aa3c1123..fe5ac3ef9018 100644
--- a/drivers/gpu/drm/panel/panel-elida-kd35t133.c
+++ b/drivers/gpu/drm/panel/panel-elida-kd35t133.c
@@ -265,7 +265,8 @@ static int kd35t133_probe(struct mipi_dsi_device *dsi)
 	dsi->lanes = 1;
 	dsi->format = MIPI_DSI_FMT_RGB888;
 	dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_BURST |
-			  MIPI_DSI_MODE_LPM | MIPI_DSI_MODE_EOT_PACKET;
+			  MIPI_DSI_MODE_LPM | MIPI_DSI_MODE_EOT_PACKET |
+			  MIPI_DSI_CLOCK_NON_CONTINUOUS;
 
 	drm_panel_init(&ctx->panel, &dsi->dev, &kd35t133_funcs,
 		       DRM_MODE_CONNECTOR_DSI);
diff --git a/drivers/gpu/drm/pl111/pl111_drv.c b/drivers/gpu/drm/pl111/pl111_drv.c
index e4dcaef6c143..fa0a737e9dea 100644
--- a/drivers/gpu/drm/pl111/pl111_drv.c
+++ b/drivers/gpu/drm/pl111/pl111_drv.c
@@ -320,7 +320,7 @@ dev_put:
 	return ret;
 }
 
-static int pl111_amba_remove(struct amba_device *amba_dev)
+static void pl111_amba_remove(struct amba_device *amba_dev)
 {
 	struct device *dev = &amba_dev->dev;
 	struct drm_device *drm = amba_get_drvdata(amba_dev);
@@ -331,8 +331,6 @@ static int pl111_amba_remove(struct amba_device *amba_dev)
 		drm_panel_bridge_remove(priv->bridge);
 	drm_dev_put(drm);
 	of_reserved_mem_device_release(dev);
-
-	return 0;
 }
 
 /*
diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
index e09b5041b684..a7637e79cb42 100644
--- a/drivers/gpu/drm/qxl/qxl_display.c
+++ b/drivers/gpu/drm/qxl/qxl_display.c
@@ -328,6 +328,7 @@ static void qxl_crtc_update_monitors_config(struct drm_crtc *crtc,
 
 	head.id = i;
 	head.flags = 0;
+	head.surface_id = 0;
 	oldcount = qdev->monitors_config->count;
 	if (crtc->state->active) {
 		struct drm_display_mode *mode = &crtc->mode;
diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c
index fb5f6a5e81d7..1864467f1063 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.c
+++ b/drivers/gpu/drm/qxl/qxl_drv.c
@@ -141,7 +141,7 @@ static void qxl_drm_release(struct drm_device *dev)
 
 	/*
 	 * TODO: qxl_device_fini() call should be in qxl_pci_remove(),
-	 * reodering qxl_modeset_fini() + qxl_device_fini() calls is
+	 * reordering qxl_modeset_fini() + qxl_device_fini() calls is
 	 * non-trivial though.
 	 */
 	qxl_modeset_fini(qdev);
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index dc68e538d5a9..34b7c6f16479 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -220,7 +220,7 @@ int r600_fmt_get_nblocksx(u32 format, u32 w)
 	if (bw == 0)
 		return 0;
 
-	return (w + bw - 1) / bw;
+	return DIV_ROUND_UP(w, bw);
 }
 
 int r600_fmt_get_nblocksy(u32 format, u32 h)
@@ -234,7 +234,7 @@ int r600_fmt_get_nblocksy(u32 format, u32 h)
 	if (bh == 0)
 		return 0;
 
-	return (h + bh - 1) / bh;
+	return DIV_ROUND_UP(h, bh);
 }
 
 struct array_mode_checker {
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 449ba0095ed9..1aab2ccfed13 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -574,6 +574,8 @@ struct radeon_gem {
 	struct list_head	objects;
 };
 
+extern const struct drm_gem_object_funcs radeon_gem_object_funcs;
+
 int radeon_gem_init(struct radeon_device *rdev);
 void radeon_gem_fini(struct radeon_device *rdev);
 int radeon_gem_object_create(struct radeon_device *rdev, unsigned long size,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 8becbe09af2f..bfacf8fe5cc1 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -2478,6 +2478,9 @@ int radeon_asic_init(struct radeon_device *rdev)
 		if (rdev->family == CHIP_HAINAN) {
 			rdev->has_uvd = false;
 			rdev->has_vce = false;
+		} else if (rdev->family == CHIP_OLAND) {
+			rdev->has_uvd = true;
+			rdev->has_vce = false;
 		} else {
 			rdev->has_uvd = true;
 			rdev->has_vce = true;
diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c
index 500796dc5d74..33121655d50b 100644
--- a/drivers/gpu/drm/radeon/radeon_bios.c
+++ b/drivers/gpu/drm/radeon/radeon_bios.c
@@ -205,7 +205,7 @@ static bool radeon_atrm_get_bios(struct radeon_device *rdev)
 			continue;
 
 		status = acpi_get_handle(dhandle, "ATRM", &atrm_handle);
-		if (!ACPI_FAILURE(status)) {
+		if (ACPI_SUCCESS(status)) {
 			found = true;
 			break;
 		}
@@ -218,7 +218,7 @@ static bool radeon_atrm_get_bios(struct radeon_device *rdev)
 				continue;
 
 			status = acpi_get_handle(dhandle, "ATRM", &atrm_handle);
-			if (!ACPI_FAILURE(status)) {
+			if (ACPI_SUCCESS(status)) {
 				found = true;
 				break;
 			}
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 941826923247..db14a82a2e4b 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -43,7 +43,7 @@ struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj);
 int radeon_gem_prime_pin(struct drm_gem_object *obj);
 void radeon_gem_prime_unpin(struct drm_gem_object *obj);
 
-static const struct drm_gem_object_funcs radeon_gem_object_funcs;
+const struct drm_gem_object_funcs radeon_gem_object_funcs;
 
 static void radeon_gem_object_free(struct drm_gem_object *gobj)
 {
@@ -227,7 +227,7 @@ static int radeon_gem_handle_lockup(struct radeon_device *rdev, int r)
 	return r;
 }
 
-static const struct drm_gem_object_funcs radeon_gem_object_funcs = {
+const struct drm_gem_object_funcs radeon_gem_object_funcs = {
 	.free = radeon_gem_object_free,
 	.open = radeon_gem_object_open,
 	.close = radeon_gem_object_close,
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index ab29eb9e8667..42a87948e28c 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -56,6 +56,8 @@ struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
 	if (ret)
 		return ERR_PTR(ret);
 
+	bo->tbo.base.funcs = &radeon_gem_object_funcs;
+
 	mutex_lock(&rdev->gem.mutex);
 	list_add_tail(&bo->list, &rdev->gem.objects);
 	mutex_unlock(&rdev->gem.mutex);
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 4f05a3f95e1e..5fc8bae401af 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -727,9 +727,6 @@ int radeon_ttm_init(struct radeon_device *rdev)
 	}
 	rdev->mman.initialized = true;
 
-	ttm_pool_init(&rdev->mman.bdev.pool, rdev->dev, rdev->need_swiotlb,
-		      dma_addressing_limited(&rdev->pdev->dev));
-
 	r = radeon_ttm_init_vram(rdev);
 	if (r) {
 		DRM_ERROR("Failed initializing VRAM heap.\n");
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index 39c1c339be7b..dfa9fdbe98da 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -781,7 +781,7 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 	uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
 		RADEON_GPU_PAGE_SIZE;
 
-	uint32_t *msg = rdev->uvd.cpu_addr + offs;
+	uint32_t __iomem *msg = (void __iomem *)(rdev->uvd.cpu_addr + offs);
 	uint64_t addr = rdev->uvd.gpu_addr + offs;
 
 	int r, i;
@@ -791,19 +791,19 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 		return r;
 
 	/* stitch together an UVD create msg */
-	msg[0] = cpu_to_le32(0x00000de4);
-	msg[1] = cpu_to_le32(0x00000000);
-	msg[2] = cpu_to_le32(handle);
-	msg[3] = cpu_to_le32(0x00000000);
-	msg[4] = cpu_to_le32(0x00000000);
-	msg[5] = cpu_to_le32(0x00000000);
-	msg[6] = cpu_to_le32(0x00000000);
-	msg[7] = cpu_to_le32(0x00000780);
-	msg[8] = cpu_to_le32(0x00000440);
-	msg[9] = cpu_to_le32(0x00000000);
-	msg[10] = cpu_to_le32(0x01b37000);
+	writel(cpu_to_le32(0x00000de4), &msg[0]);
+	writel(0x0, (void __iomem *)&msg[1]);
+	writel(cpu_to_le32(handle), &msg[2]);
+	writel(0x0, &msg[3]);
+	writel(0x0, &msg[4]);
+	writel(0x0, &msg[5]);
+	writel(0x0, &msg[6]);
+	writel(cpu_to_le32(0x00000780), &msg[7]);
+	writel(cpu_to_le32(0x00000440), &msg[8]);
+	writel(0x0, &msg[9]);
+	writel(cpu_to_le32(0x01b37000), &msg[10]);
 	for (i = 11; i < 1024; ++i)
-		msg[i] = cpu_to_le32(0x0);
+		writel(0x0, &msg[i]);
 
 	r = radeon_uvd_send_msg(rdev, ring, addr, fence);
 	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
@@ -817,7 +817,7 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 	uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
 		RADEON_GPU_PAGE_SIZE;
 
-	uint32_t *msg = rdev->uvd.cpu_addr + offs;
+	uint32_t __iomem *msg = (void __iomem *)(rdev->uvd.cpu_addr + offs);
 	uint64_t addr = rdev->uvd.gpu_addr + offs;
 
 	int r, i;
@@ -827,12 +827,12 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 		return r;
 
 	/* stitch together an UVD destroy msg */
-	msg[0] = cpu_to_le32(0x00000de4);
-	msg[1] = cpu_to_le32(0x00000002);
-	msg[2] = cpu_to_le32(handle);
-	msg[3] = cpu_to_le32(0x00000000);
+	writel(cpu_to_le32(0x00000de4), &msg[0]);
+	writel(cpu_to_le32(0x00000002), &msg[1]);
+	writel(cpu_to_le32(handle), &msg[2]);
+	writel(0x0, &msg[3]);
 	for (i = 4; i < 1024; ++i)
-		msg[i] = cpu_to_le32(0x0);
+		writel(0x0, &msg[i]);
 
 	r = radeon_uvd_send_msg(rdev, ring, addr, fence);
 	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c
index a450497368b2..511a942e851d 100644
--- a/drivers/gpu/drm/radeon/radeon_vce.c
+++ b/drivers/gpu/drm/radeon/radeon_vce.c
@@ -68,7 +68,6 @@ int radeon_vce_init(struct radeon_device *rdev)
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
 	case CHIP_VERDE:
-	case CHIP_OLAND:
 	case CHIP_ARUBA:
 		fw_name = FIRMWARE_TAHITI;
 		break;
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index c296f94f9700..7bc302a89232 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -187,7 +187,7 @@ static void rs690_mc_init(struct radeon_device *rdev)
 		/* FastFB shall be used with UMA memory. Here it is simply disabled when sideport 
 		 * memory is present.
 		 */
-		if (rdev->mc.igp_sideport_enabled == false && radeon_fastfb == 1) {
+		if (!rdev->mc.igp_sideport_enabled && radeon_fastfb == 1) {
 			DRM_INFO("Direct mapping: aper base at 0x%llx, replaced by direct mapping base 0x%llx.\n", 
 					(unsigned long long)rdev->mc.aper_base, k8_addr);
 			rdev->mc.aper_base = (resource_size_t)k8_addr;
diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c
index 70c5da2141d7..bdfbcf14b864 100644
--- a/drivers/gpu/drm/radeon/vce_v1_0.c
+++ b/drivers/gpu/drm/radeon/vce_v1_0.c
@@ -169,7 +169,6 @@ int vce_v1_0_load_fw(struct radeon_device *rdev, uint32_t *data)
 		chip_id = 0x01000015;
 		break;
 	case CHIP_PITCAIRN:
-	case CHIP_OLAND:
 		chip_id = 0x01000016;
 		break;
 	case CHIP_ARUBA:
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
index 4a2099cb582e..857d97cdc67c 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
@@ -17,9 +17,20 @@
 
 #define NUM_YUV2YUV_COEFFICIENTS 12
 
+/* AFBC supports a number of configurable modes. Relevant to us is block size
+ * (16x16 or 32x8), storage modifiers (SPARSE, SPLIT), and the YUV-like
+ * colourspace transform (YTR). 16x16 SPARSE mode is always used. SPLIT mode
+ * could be enabled via the hreg_block_split register, but is not currently
+ * handled. The colourspace transform is implicitly always assumed by the
+ * decoder, so consumers must use this transform as well.
+ *
+ * Failure to match modifiers will cause errors displaying AFBC buffers
+ * produced by conformant AFBC producers, including Mesa.
+ */
 #define ROCKCHIP_AFBC_MOD \
 	DRM_FORMAT_MOD_ARM_AFBC( \
 		AFBC_FORMAT_MOD_BLOCK_SIZE_16x16 | AFBC_FORMAT_MOD_SPARSE \
+			| AFBC_FORMAT_MOD_YTR \
 	)
 
 enum vop_data_format {
diff --git a/drivers/gpu/drm/sun4i/sun4i_tcon.c b/drivers/gpu/drm/sun4i/sun4i_tcon.c
index 6b9af4c08cd6..9f06dec0fc61 100644
--- a/drivers/gpu/drm/sun4i/sun4i_tcon.c
+++ b/drivers/gpu/drm/sun4i/sun4i_tcon.c
@@ -672,6 +672,30 @@ static void sun4i_tcon1_mode_set(struct sun4i_tcon *tcon,
 		     SUN4I_TCON1_BASIC5_V_SYNC(vsync) |
 		     SUN4I_TCON1_BASIC5_H_SYNC(hsync));
 
+	/* Setup the polarity of multiple signals */
+	if (tcon->quirks->polarity_in_ch0) {
+		val = 0;
+
+		if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+			val |= SUN4I_TCON0_IO_POL_HSYNC_POSITIVE;
+
+		if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+			val |= SUN4I_TCON0_IO_POL_VSYNC_POSITIVE;
+
+		regmap_write(tcon->regs, SUN4I_TCON0_IO_POL_REG, val);
+	} else {
+		/* according to vendor driver, this bit must be always set */
+		val = SUN4I_TCON1_IO_POL_UNKNOWN;
+
+		if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+			val |= SUN4I_TCON1_IO_POL_HSYNC_POSITIVE;
+
+		if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+			val |= SUN4I_TCON1_IO_POL_VSYNC_POSITIVE;
+
+		regmap_write(tcon->regs, SUN4I_TCON1_IO_POL_REG, val);
+	}
+
 	/* Map output pins to channel 1 */
 	regmap_update_bits(tcon->regs, SUN4I_TCON_GCTL_REG,
 			   SUN4I_TCON_GCTL_IOMAP_MASK,
@@ -1500,6 +1524,7 @@ static const struct sun4i_tcon_quirks sun8i_a83t_tv_quirks = {
 
 static const struct sun4i_tcon_quirks sun8i_r40_tv_quirks = {
 	.has_channel_1		= true,
+	.polarity_in_ch0	= true,
 	.set_mux		= sun8i_r40_tcon_tv_set_mux,
 };
 
diff --git a/drivers/gpu/drm/sun4i/sun4i_tcon.h b/drivers/gpu/drm/sun4i/sun4i_tcon.h
index c5ac1b02482c..e624f6977eb8 100644
--- a/drivers/gpu/drm/sun4i/sun4i_tcon.h
+++ b/drivers/gpu/drm/sun4i/sun4i_tcon.h
@@ -154,6 +154,11 @@
 #define SUN4I_TCON1_BASIC5_V_SYNC(height)		(((height) - 1) & 0x3ff)
 
 #define SUN4I_TCON1_IO_POL_REG			0xf0
+/* there is no documentation about this bit */
+#define SUN4I_TCON1_IO_POL_UNKNOWN			BIT(26)
+#define SUN4I_TCON1_IO_POL_HSYNC_POSITIVE		BIT(25)
+#define SUN4I_TCON1_IO_POL_VSYNC_POSITIVE		BIT(24)
+
 #define SUN4I_TCON1_IO_TRI_REG			0xf4
 
 #define SUN4I_TCON_ECC_FIFO_REG			0xf8
@@ -236,6 +241,7 @@ struct sun4i_tcon_quirks {
 	bool	needs_de_be_mux; /* sun6i needs mux to select backend */
 	bool    needs_edp_reset; /* a80 edp reset needed for tcon0 access */
 	bool	supports_lvds;   /* Does the TCON support an LVDS output? */
+	bool	polarity_in_ch0; /* some tcon1 channels have polarity bits in tcon0 pol register */
 	u8	dclk_min_div;	/* minimum divider for TCON0 DCLK */
 
 	/* callback to handle tcon muxing options */
diff --git a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
index 92add2cef2e7..bbdfd5e26ec8 100644
--- a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
+++ b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
@@ -21,8 +21,7 @@ static void sun8i_dw_hdmi_encoder_mode_set(struct drm_encoder *encoder,
 {
 	struct sun8i_dw_hdmi *hdmi = encoder_to_sun8i_dw_hdmi(encoder);
 
-	if (hdmi->quirks->set_rate)
-		clk_set_rate(hdmi->clk_tmds, mode->crtc_clock * 1000);
+	clk_set_rate(hdmi->clk_tmds, mode->crtc_clock * 1000);
 }
 
 static const struct drm_encoder_helper_funcs
@@ -48,11 +47,9 @@ sun8i_dw_hdmi_mode_valid_h6(struct dw_hdmi *hdmi, void *data,
 {
 	/*
 	 * Controller support maximum of 594 MHz, which correlates to
-	 * 4K@60Hz 4:4:4 or RGB. However, for frequencies greater than
-	 * 340 MHz scrambling has to be enabled. Because scrambling is
-	 * not yet implemented, just limit to 340 MHz for now.
+	 * 4K@60Hz 4:4:4 or RGB.
 	 */
-	if (mode->clock > 340000)
+	if (mode->clock > 594000)
 		return MODE_CLOCK_HIGH;
 
 	return MODE_OK;
@@ -295,7 +292,6 @@ static int sun8i_dw_hdmi_remove(struct platform_device *pdev)
 
 static const struct sun8i_dw_hdmi_quirks sun8i_a83t_quirks = {
 	.mode_valid = sun8i_dw_hdmi_mode_valid_a83t,
-	.set_rate = true,
 };
 
 static const struct sun8i_dw_hdmi_quirks sun50i_h6_quirks = {
diff --git a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h
index d983746fa194..d4b55af0592f 100644
--- a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h
+++ b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h
@@ -179,7 +179,6 @@ struct sun8i_dw_hdmi_quirks {
 	enum drm_mode_status (*mode_valid)(struct dw_hdmi *hdmi, void *data,
 					   const struct drm_display_info *info,
 					   const struct drm_display_mode *mode);
-	unsigned int set_rate : 1;
 	unsigned int use_drm_infoframe : 1;
 };
 
diff --git a/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c b/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c
index 35c2133724e2..9994edf67509 100644
--- a/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c
+++ b/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c
@@ -104,29 +104,21 @@ static const struct dw_hdmi_mpll_config sun50i_h6_mpll_cfg[] = {
 
 static const struct dw_hdmi_curr_ctrl sun50i_h6_cur_ctr[] = {
 	/* pixelclk    bpp8    bpp10   bpp12 */
-	{ 25175000,  { 0x0000, 0x0000, 0x0000 }, },
 	{ 27000000,  { 0x0012, 0x0000, 0x0000 }, },
-	{ 59400000,  { 0x0008, 0x0008, 0x0008 }, },
-	{ 72000000,  { 0x0008, 0x0008, 0x001b }, },
-	{ 74250000,  { 0x0013, 0x0013, 0x0013 }, },
-	{ 90000000,  { 0x0008, 0x001a, 0x001b }, },
-	{ 118800000, { 0x001b, 0x001a, 0x001b }, },
-	{ 144000000, { 0x001b, 0x001a, 0x0034 }, },
-	{ 180000000, { 0x001b, 0x0033, 0x0034 }, },
-	{ 216000000, { 0x0036, 0x0033, 0x0034 }, },
-	{ 237600000, { 0x0036, 0x0033, 0x001b }, },
-	{ 288000000, { 0x0036, 0x001b, 0x001b }, },
-	{ 297000000, { 0x0019, 0x001b, 0x0019 }, },
-	{ 330000000, { 0x0036, 0x001b, 0x001b }, },
-	{ 594000000, { 0x003f, 0x001b, 0x001b }, },
+	{ 74250000,  { 0x0013, 0x001a, 0x001b }, },
+	{ 148500000, { 0x0019, 0x0033, 0x0034 }, },
+	{ 297000000, { 0x0019, 0x001b, 0x001b }, },
+	{ 594000000, { 0x0010, 0x001b, 0x001b }, },
 	{ ~0UL,      { 0x0000, 0x0000, 0x0000 }, }
 };
 
 static const struct dw_hdmi_phy_config sun50i_h6_phy_config[] = {
 	/*pixelclk   symbol   term   vlev*/
-	{ 74250000,  0x8009, 0x0004, 0x0232},
-	{ 148500000, 0x8029, 0x0004, 0x0273},
-	{ 594000000, 0x8039, 0x0004, 0x014a},
+	{ 27000000,  0x8009, 0x0007, 0x02b0 },
+	{ 74250000,  0x8009, 0x0006, 0x022d },
+	{ 148500000, 0x8029, 0x0006, 0x0270 },
+	{ 297000000, 0x8039, 0x0005, 0x01ab },
+	{ 594000000, 0x8029, 0x0000, 0x008a },
 	{ ~0UL,	     0x0000, 0x0000, 0x0000}
 };
 
diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 455d449191ab..da6afe7f0c7d 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -2198,7 +2198,7 @@ static int tegra_dc_runtime_resume(struct host1x_client *client)
 	struct device *dev = client->dev;
 	int err;
 
-	err = pm_runtime_get_sync(dev);
+	err = pm_runtime_resume_and_get(dev);
 	if (err < 0) {
 		dev_err(dev, "failed to get runtime PM: %d\n", err);
 		return err;
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index ef2b79a903b1..90709c38c993 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -1304,8 +1304,10 @@ static const struct of_device_id host1x_drm_subdevs[] = {
 	{ .compatible = "nvidia,tegra30-hdmi", },
 	{ .compatible = "nvidia,tegra30-gr2d", },
 	{ .compatible = "nvidia,tegra30-gr3d", },
+	{ .compatible = "nvidia,tegra114-dc", },
 	{ .compatible = "nvidia,tegra114-dsi", },
 	{ .compatible = "nvidia,tegra114-hdmi", },
+	{ .compatible = "nvidia,tegra114-gr2d", },
 	{ .compatible = "nvidia,tegra114-gr3d", },
 	{ .compatible = "nvidia,tegra124-dc", },
 	{ .compatible = "nvidia,tegra124-sor", },
diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c
index 5691ef1b0e58..f46d377f0c30 100644
--- a/drivers/gpu/drm/tegra/dsi.c
+++ b/drivers/gpu/drm/tegra/dsi.c
@@ -1111,7 +1111,7 @@ static int tegra_dsi_runtime_resume(struct host1x_client *client)
 	struct device *dev = client->dev;
 	int err;
 
-	err = pm_runtime_get_sync(dev);
+	err = pm_runtime_resume_and_get(dev);
 	if (err < 0) {
 		dev_err(dev, "failed to get runtime PM: %d\n", err);
 		return err;
diff --git a/drivers/gpu/drm/tegra/falcon.c b/drivers/gpu/drm/tegra/falcon.c
index 56edef06c48e..223ab2ceb7e6 100644
--- a/drivers/gpu/drm/tegra/falcon.c
+++ b/drivers/gpu/drm/tegra/falcon.c
@@ -72,7 +72,7 @@ static int falcon_parse_firmware_image(struct falcon *falcon)
 	struct falcon_fw_os_header_v1 *os;
 
 	/* endian problems would show up right here */
-	if (bin->magic != PCI_VENDOR_ID_NVIDIA) {
+	if (bin->magic != PCI_VENDOR_ID_NVIDIA && bin->magic != 0x10fe) {
 		dev_err(falcon->dev, "incorrect firmware magic\n");
 		return -EINVAL;
 	}
@@ -178,9 +178,10 @@ int falcon_boot(struct falcon *falcon)
 				  falcon->firmware.data.offset + offset,
 				  offset, FALCON_MEMORY_DATA);
 
-	/* copy the first code segment into Falcon internal memory */
-	falcon_copy_chunk(falcon, falcon->firmware.code.offset,
-			  0, FALCON_MEMORY_IMEM);
+	/* copy the code segment into Falcon internal memory */
+	for (offset = 0; offset < falcon->firmware.code.size; offset += 256)
+		falcon_copy_chunk(falcon, falcon->firmware.code.offset + offset,
+				  offset, FALCON_MEMORY_IMEM);
 
 	/* setup falcon interrupts */
 	falcon_writel(falcon, FALCON_IRQMSET_EXT(0xff) |
diff --git a/drivers/gpu/drm/tegra/gr2d.c b/drivers/gpu/drm/tegra/gr2d.c
index 1a0d3ba6e525..adbe2ddcda19 100644
--- a/drivers/gpu/drm/tegra/gr2d.c
+++ b/drivers/gpu/drm/tegra/gr2d.c
@@ -161,9 +161,14 @@ static const struct gr2d_soc tegra30_gr2d_soc = {
 	.version = 0x30,
 };
 
+static const struct gr2d_soc tegra114_gr2d_soc = {
+	.version = 0x35,
+};
+
 static const struct of_device_id gr2d_match[] = {
-	{ .compatible = "nvidia,tegra30-gr2d", .data = &tegra20_gr2d_soc },
-	{ .compatible = "nvidia,tegra20-gr2d", .data = &tegra30_gr2d_soc },
+	{ .compatible = "nvidia,tegra114-gr2d", .data = &tegra114_gr2d_soc },
+	{ .compatible = "nvidia,tegra30-gr2d", .data = &tegra30_gr2d_soc },
+	{ .compatible = "nvidia,tegra20-gr2d", .data = &tegra20_gr2d_soc },
 	{ },
 };
 MODULE_DEVICE_TABLE(of, gr2d_match);
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
index d09a24931c87..e5d2a4026028 100644
--- a/drivers/gpu/drm/tegra/hdmi.c
+++ b/drivers/gpu/drm/tegra/hdmi.c
@@ -1510,7 +1510,7 @@ static int tegra_hdmi_runtime_resume(struct host1x_client *client)
 	struct device *dev = client->dev;
 	int err;
 
-	err = pm_runtime_get_sync(dev);
+	err = pm_runtime_resume_and_get(dev);
 	if (err < 0) {
 		dev_err(dev, "failed to get runtime PM: %d\n", err);
 		return err;
diff --git a/drivers/gpu/drm/tegra/hub.c b/drivers/gpu/drm/tegra/hub.c
index 8118007cb6d7..8e6d329d062b 100644
--- a/drivers/gpu/drm/tegra/hub.c
+++ b/drivers/gpu/drm/tegra/hub.c
@@ -795,7 +795,7 @@ static int tegra_display_hub_runtime_resume(struct host1x_client *client)
 	unsigned int i;
 	int err;
 
-	err = pm_runtime_get_sync(dev);
+	err = pm_runtime_resume_and_get(dev);
 	if (err < 0) {
 		dev_err(dev, "failed to get runtime PM: %d\n", err);
 		return err;
diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c
index cc2aa2308a51..f02a035dda45 100644
--- a/drivers/gpu/drm/tegra/sor.c
+++ b/drivers/gpu/drm/tegra/sor.c
@@ -3218,7 +3218,7 @@ static int tegra_sor_runtime_resume(struct host1x_client *client)
 	struct device *dev = client->dev;
 	int err;
 
-	err = pm_runtime_get_sync(dev);
+	err = pm_runtime_resume_and_get(dev);
 	if (err < 0) {
 		dev_err(dev, "failed to get runtime PM: %d\n", err);
 		return err;
diff --git a/drivers/gpu/drm/tegra/vic.c b/drivers/gpu/drm/tegra/vic.c
index ade56b860cf9..77e128832920 100644
--- a/drivers/gpu/drm/tegra/vic.c
+++ b/drivers/gpu/drm/tegra/vic.c
@@ -117,7 +117,19 @@ static int vic_boot(struct vic *vic)
 		if (spec->num_ids > 0) {
 			value = spec->ids[0] & 0xffff;
 
+			/*
+			 * STREAMID0 is used for input/output buffers.
+			 * Initialize it to SID_VIC in case context isolation
+			 * is not enabled, and SID_VIC is used for both firmware
+			 * and data buffers.
+			 *
+			 * If context isolation is enabled, it will be
+			 * overridden by the SETSTREAMID opcode as part of
+			 * each job.
+			 */
 			vic_writel(vic, value, VIC_THI_STREAMID0);
+
+			/* STREAMID1 is used for firmware loading. */
 			vic_writel(vic, value, VIC_THI_STREAMID1);
 		}
 	}
@@ -135,16 +147,21 @@ static int vic_boot(struct vic *vic)
 
 	hdr = vic->falcon.firmware.virt;
 	fce_bin_data_offset = *(u32 *)(hdr + VIC_UCODE_FCE_DATA_OFFSET);
-	hdr = vic->falcon.firmware.virt +
-		*(u32 *)(hdr + VIC_UCODE_FCE_HEADER_OFFSET);
-	fce_ucode_size = *(u32 *)(hdr + FCE_UCODE_SIZE_OFFSET);
 
 	falcon_execute_method(&vic->falcon, VIC_SET_APPLICATION_ID, 1);
-	falcon_execute_method(&vic->falcon, VIC_SET_FCE_UCODE_SIZE,
-			      fce_ucode_size);
-	falcon_execute_method(&vic->falcon, VIC_SET_FCE_UCODE_OFFSET,
-			      (vic->falcon.firmware.iova + fce_bin_data_offset)
-				>> 8);
+
+	/* Old VIC firmware needs kernel help with setting up FCE microcode. */
+	if (fce_bin_data_offset != 0x0 && fce_bin_data_offset != 0xa5a5a5a5) {
+		hdr = vic->falcon.firmware.virt +
+			*(u32 *)(hdr + VIC_UCODE_FCE_HEADER_OFFSET);
+		fce_ucode_size = *(u32 *)(hdr + FCE_UCODE_SIZE_OFFSET);
+
+		falcon_execute_method(&vic->falcon, VIC_SET_FCE_UCODE_SIZE,
+				      fce_ucode_size);
+		falcon_execute_method(
+			&vic->falcon, VIC_SET_FCE_UCODE_OFFSET,
+			(vic->falcon.firmware.iova + fce_bin_data_offset) >> 8);
+	}
 
 	err = falcon_wait_idle(&vic->falcon);
 	if (err < 0) {
@@ -314,7 +331,7 @@ static int vic_open_channel(struct tegra_drm_client *client,
 	struct vic *vic = to_vic(client);
 	int err;
 
-	err = pm_runtime_get_sync(vic->dev);
+	err = pm_runtime_resume_and_get(vic->dev);
 	if (err < 0)
 		return err;
 
diff --git a/drivers/gpu/drm/tilcdc/Makefile b/drivers/gpu/drm/tilcdc/Makefile
index 662bf3a348c9..f5190477de72 100644
--- a/drivers/gpu/drm/tilcdc/Makefile
+++ b/drivers/gpu/drm/tilcdc/Makefile
@@ -1,5 +1,5 @@
 # SPDX-License-Identifier: GPL-2.0
-ifeq (, $(findstring -W,$(EXTRA_CFLAGS)))
+ifeq (, $(findstring -W,$(KCFLAGS)))
 	ccflags-y += -Werror
 endif
 
diff --git a/drivers/gpu/drm/tiny/gm12u320.c b/drivers/gpu/drm/tiny/gm12u320.c
index 0b4f4f2af1ef..a233c86d428b 100644
--- a/drivers/gpu/drm/tiny/gm12u320.c
+++ b/drivers/gpu/drm/tiny/gm12u320.c
@@ -84,6 +84,7 @@ MODULE_PARM_DESC(eco_mode, "Turn on Eco mode (less bright, more silent)");
 
 struct gm12u320_device {
 	struct drm_device	         dev;
+	struct device                   *dmadev;
 	struct drm_simple_display_pipe   pipe;
 	struct drm_connector	         conn;
 	unsigned char                   *cmd_buf;
@@ -599,6 +600,22 @@ static const uint64_t gm12u320_pipe_modifiers[] = {
 	DRM_FORMAT_MOD_INVALID
 };
 
+/*
+ * FIXME: Dma-buf sharing requires DMA support by the importing device.
+ *        This function is a workaround to make USB devices work as well.
+ *        See todo.rst for how to fix the issue in the dma-buf framework.
+ */
+static struct drm_gem_object *gm12u320_gem_prime_import(struct drm_device *dev,
+							struct dma_buf *dma_buf)
+{
+	struct gm12u320_device *gm12u320 = to_gm12u320(dev);
+
+	if (!gm12u320->dmadev)
+		return ERR_PTR(-ENODEV);
+
+	return drm_gem_prime_import_dev(dev, dma_buf, gm12u320->dmadev);
+}
+
 DEFINE_DRM_GEM_FOPS(gm12u320_fops);
 
 static const struct drm_driver gm12u320_drm_driver = {
@@ -612,6 +629,7 @@ static const struct drm_driver gm12u320_drm_driver = {
 
 	.fops		 = &gm12u320_fops,
 	DRM_GEM_SHMEM_DRIVER_OPS,
+	.gem_prime_import = gm12u320_gem_prime_import,
 };
 
 static const struct drm_mode_config_funcs gm12u320_mode_config_funcs = {
@@ -638,15 +656,18 @@ static int gm12u320_usb_probe(struct usb_interface *interface,
 				      struct gm12u320_device, dev);
 	if (IS_ERR(gm12u320))
 		return PTR_ERR(gm12u320);
+	dev = &gm12u320->dev;
+
+	gm12u320->dmadev = usb_intf_get_dma_device(to_usb_interface(dev->dev));
+	if (!gm12u320->dmadev)
+		drm_warn(dev, "buffer sharing not supported"); /* not an error */
 
 	INIT_DELAYED_WORK(&gm12u320->fb_update.work, gm12u320_fb_update_work);
 	mutex_init(&gm12u320->fb_update.lock);
 
-	dev = &gm12u320->dev;
-
 	ret = drmm_mode_config_init(dev);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
 	dev->mode_config.min_width = GM12U320_USER_WIDTH;
 	dev->mode_config.max_width = GM12U320_USER_WIDTH;
@@ -656,15 +677,15 @@ static int gm12u320_usb_probe(struct usb_interface *interface,
 
 	ret = gm12u320_usb_alloc(gm12u320);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
 	ret = gm12u320_set_ecomode(gm12u320);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
 	ret = gm12u320_conn_init(gm12u320);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
 	ret = drm_simple_display_pipe_init(&gm12u320->dev,
 					   &gm12u320->pipe,
@@ -674,24 +695,31 @@ static int gm12u320_usb_probe(struct usb_interface *interface,
 					   gm12u320_pipe_modifiers,
 					   &gm12u320->conn);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
 	drm_mode_config_reset(dev);
 
 	usb_set_intfdata(interface, dev);
 	ret = drm_dev_register(dev, 0);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
 	drm_fbdev_generic_setup(dev, 0);
 
 	return 0;
+
+err_put_device:
+	put_device(gm12u320->dmadev);
+	return ret;
 }
 
 static void gm12u320_usb_disconnect(struct usb_interface *interface)
 {
 	struct drm_device *dev = usb_get_intfdata(interface);
+	struct gm12u320_device *gm12u320 = to_gm12u320(dev);
 
+	put_device(gm12u320->dmadev);
+	gm12u320->dmadev = NULL;
 	drm_dev_unplug(dev);
 	drm_atomic_helper_shutdown(dev);
 }
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index a08dec7281fc..3c23e863a3f0 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -95,7 +95,8 @@ void ttm_bo_move_to_lru_tail(struct ttm_buffer_object *bo,
 	struct ttm_device *bdev = bo->bdev;
 	struct ttm_resource_manager *man;
 
-	dma_resv_assert_held(bo->base.resv);
+	if (!bo->deleted)
+		dma_resv_assert_held(bo->base.resv);
 
 	if (bo->pin_count) {
 		ttm_bo_del_from_lru(bo);
@@ -458,8 +459,11 @@ static void ttm_bo_release(struct kref *kref)
 		 * Make pinned bos immediately available to
 		 * shrinkers, now that they are queued for
 		 * destruction.
+		 *
+		 * FIXME: QXL is triggering this. Can be removed when the
+		 * driver is fixed.
 		 */
-		if (WARN_ON(bo->pin_count)) {
+		if (WARN_ON_ONCE(bo->pin_count)) {
 			bo->pin_count = 0;
 			ttm_bo_move_to_lru_tail(bo, &bo->mem, NULL);
 		}
@@ -908,8 +912,10 @@ static int ttm_bo_bounce_temp_buffer(struct ttm_buffer_object *bo,
 		return ret;
 	/* move to the bounce domain */
 	ret = ttm_bo_handle_move_mem(bo, &hop_mem, false, ctx, NULL);
-	if (ret)
+	if (ret) {
+		ttm_resource_free(bo, &hop_mem);
 		return ret;
+	}
 	return 0;
 }
 
@@ -940,18 +946,19 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	 * stop and the driver will be called to make
 	 * the second hop.
 	 */
-bounce:
 	ret = ttm_bo_mem_space(bo, placement, &mem, ctx);
 	if (ret)
 		return ret;
+bounce:
 	ret = ttm_bo_handle_move_mem(bo, &mem, false, ctx, &hop);
 	if (ret == -EMULTIHOP) {
 		ret = ttm_bo_bounce_temp_buffer(bo, &mem, ctx, &hop);
 		if (ret)
-			return ret;
+			goto out;
 		/* try and move to final place now. */
 		goto bounce;
 	}
+out:
 	if (ret)
 		ttm_resource_free(bo, &mem);
 	return ret;
diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c
index 6b0f957d63d5..cb38b1a17b09 100644
--- a/drivers/gpu/drm/ttm/ttm_pool.c
+++ b/drivers/gpu/drm/ttm/ttm_pool.c
@@ -33,6 +33,7 @@
 
 #include <linux/module.h>
 #include <linux/dma-mapping.h>
+#include <linux/highmem.h>
 #include <linux/sched/mm.h>
 
 #ifdef CONFIG_X86
@@ -69,7 +70,7 @@ static struct ttm_pool_type global_uncached[MAX_ORDER];
 static struct ttm_pool_type global_dma32_write_combined[MAX_ORDER];
 static struct ttm_pool_type global_dma32_uncached[MAX_ORDER];
 
-static spinlock_t shrinker_lock;
+static struct mutex shrinker_lock;
 static struct list_head shrinker_list;
 static struct shrinker mm_shrinker;
 
@@ -82,12 +83,13 @@ static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags,
 	struct page *p;
 	void *vaddr;
 
-	if (order) {
-		gfp_flags |= GFP_TRANSHUGE_LIGHT | __GFP_NORETRY |
+	/* Don't set the __GFP_COMP flag for higher order allocations.
+	 * Mapping pages directly into an userspace process and calling
+	 * put_page() on a TTM allocated page is illegal.
+	 */
+	if (order)
+		gfp_flags |= __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN |
 			__GFP_KSWAPD_RECLAIM;
-		gfp_flags &= ~__GFP_MOVABLE;
-		gfp_flags &= ~__GFP_COMP;
-	}
 
 	if (!pool->use_dma_alloc) {
 		p = alloc_pages(gfp_flags, order);
@@ -193,7 +195,7 @@ static int ttm_pool_map(struct ttm_pool *pool, unsigned int order,
 		size_t size = (1ULL << order) * PAGE_SIZE;
 
 		addr = dma_map_page(pool->dev, p, 0, size, DMA_BIDIRECTIONAL);
-		if (dma_mapping_error(pool->dev, **dma_addr))
+		if (dma_mapping_error(pool->dev, addr))
 			return -EFAULT;
 	}
 
@@ -220,6 +222,15 @@ static void ttm_pool_unmap(struct ttm_pool *pool, dma_addr_t dma_addr,
 /* Give pages into a specific pool_type */
 static void ttm_pool_type_give(struct ttm_pool_type *pt, struct page *p)
 {
+	unsigned int i, num_pages = 1 << pt->order;
+
+	for (i = 0; i < num_pages; ++i) {
+		if (PageHighMem(p))
+			clear_highpage(p + i);
+		else
+			clear_page(page_address(p + i));
+	}
+
 	spin_lock(&pt->lock);
 	list_add(&p->lru, &pt->pages);
 	spin_unlock(&pt->lock);
@@ -252,21 +263,21 @@ static void ttm_pool_type_init(struct ttm_pool_type *pt, struct ttm_pool *pool,
 	spin_lock_init(&pt->lock);
 	INIT_LIST_HEAD(&pt->pages);
 
-	spin_lock(&shrinker_lock);
+	mutex_lock(&shrinker_lock);
 	list_add_tail(&pt->shrinker_list, &shrinker_list);
-	spin_unlock(&shrinker_lock);
+	mutex_unlock(&shrinker_lock);
 }
 
 /* Remove a pool_type from the global shrinker list and free all pages */
 static void ttm_pool_type_fini(struct ttm_pool_type *pt)
 {
-	struct page *p, *tmp;
+	struct page *p;
 
-	spin_lock(&shrinker_lock);
+	mutex_lock(&shrinker_lock);
 	list_del(&pt->shrinker_list);
-	spin_unlock(&shrinker_lock);
+	mutex_unlock(&shrinker_lock);
 
-	list_for_each_entry_safe(p, tmp, &pt->pages, lru)
+	while ((p = ttm_pool_type_take(pt)))
 		ttm_pool_free_page(pt->pool, pt->caching, pt->order, p);
 }
 
@@ -305,7 +316,7 @@ static unsigned int ttm_pool_shrink(void)
 	unsigned int num_freed;
 	struct page *p;
 
-	spin_lock(&shrinker_lock);
+	mutex_lock(&shrinker_lock);
 	pt = list_first_entry(&shrinker_list, typeof(*pt), shrinker_list);
 
 	p = ttm_pool_type_take(pt);
@@ -317,7 +328,7 @@ static unsigned int ttm_pool_shrink(void)
 	}
 
 	list_move_tail(&pt->shrinker_list, &shrinker_list);
-	spin_unlock(&shrinker_lock);
+	mutex_unlock(&shrinker_lock);
 
 	return num_freed;
 }
@@ -501,7 +512,6 @@ void ttm_pool_init(struct ttm_pool *pool, struct device *dev,
 						   pool, i, j);
 	}
 }
-EXPORT_SYMBOL(ttm_pool_init);
 
 /**
  * ttm_pool_fini - Cleanup a pool
@@ -521,7 +531,6 @@ void ttm_pool_fini(struct ttm_pool *pool)
 				ttm_pool_type_fini(&pool->caching[i].orders[j]);
 	}
 }
-EXPORT_SYMBOL(ttm_pool_fini);
 
 /* As long as pages are available make sure to release at least one */
 static unsigned long ttm_pool_shrinker_scan(struct shrinker *shrink,
@@ -595,7 +604,7 @@ static int ttm_pool_debugfs_globals_show(struct seq_file *m, void *data)
 {
 	ttm_pool_debugfs_header(m);
 
-	spin_lock(&shrinker_lock);
+	mutex_lock(&shrinker_lock);
 	seq_puts(m, "wc\t:");
 	ttm_pool_debugfs_orders(global_write_combined, m);
 	seq_puts(m, "uc\t:");
@@ -604,7 +613,7 @@ static int ttm_pool_debugfs_globals_show(struct seq_file *m, void *data)
 	ttm_pool_debugfs_orders(global_dma32_write_combined, m);
 	seq_puts(m, "uc 32\t:");
 	ttm_pool_debugfs_orders(global_dma32_uncached, m);
-	spin_unlock(&shrinker_lock);
+	mutex_unlock(&shrinker_lock);
 
 	ttm_pool_debugfs_footer(m);
 
@@ -631,7 +640,7 @@ int ttm_pool_debugfs(struct ttm_pool *pool, struct seq_file *m)
 
 	ttm_pool_debugfs_header(m);
 
-	spin_lock(&shrinker_lock);
+	mutex_lock(&shrinker_lock);
 	for (i = 0; i < TTM_NUM_CACHING_TYPES; ++i) {
 		seq_puts(m, "DMA ");
 		switch (i) {
@@ -647,7 +656,7 @@ int ttm_pool_debugfs(struct ttm_pool *pool, struct seq_file *m)
 		}
 		ttm_pool_debugfs_orders(pool->caching[i].orders, m);
 	}
-	spin_unlock(&shrinker_lock);
+	mutex_unlock(&shrinker_lock);
 
 	ttm_pool_debugfs_footer(m);
 	return 0;
@@ -684,7 +693,7 @@ int ttm_pool_mgr_init(unsigned long num_pages)
 	if (!page_pool_size)
 		page_pool_size = num_pages;
 
-	spin_lock_init(&shrinker_lock);
+	mutex_init(&shrinker_lock);
 	INIT_LIST_HEAD(&shrinker_list);
 
 	for (i = 0; i < MAX_ORDER; ++i) {
diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c
index 9269092697d8..5703277c6f52 100644
--- a/drivers/gpu/drm/udl/udl_drv.c
+++ b/drivers/gpu/drm/udl/udl_drv.c
@@ -32,6 +32,22 @@ static int udl_usb_resume(struct usb_interface *interface)
 	return drm_mode_config_helper_resume(dev);
 }
 
+/*
+ * FIXME: Dma-buf sharing requires DMA support by the importing device.
+ *        This function is a workaround to make USB devices work as well.
+ *        See todo.rst for how to fix the issue in the dma-buf framework.
+ */
+static struct drm_gem_object *udl_driver_gem_prime_import(struct drm_device *dev,
+							  struct dma_buf *dma_buf)
+{
+	struct udl_device *udl = to_udl(dev);
+
+	if (!udl->dmadev)
+		return ERR_PTR(-ENODEV);
+
+	return drm_gem_prime_import_dev(dev, dma_buf, udl->dmadev);
+}
+
 DEFINE_DRM_GEM_FOPS(udl_driver_fops);
 
 static const struct drm_driver driver = {
@@ -40,6 +56,7 @@ static const struct drm_driver driver = {
 	/* GEM hooks */
 	.fops = &udl_driver_fops,
 	DRM_GEM_SHMEM_DRIVER_OPS,
+	.gem_prime_import = udl_driver_gem_prime_import,
 
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h
index 875e73551ae9..cc16a13316e4 100644
--- a/drivers/gpu/drm/udl/udl_drv.h
+++ b/drivers/gpu/drm/udl/udl_drv.h
@@ -50,6 +50,7 @@ struct urb_list {
 struct udl_device {
 	struct drm_device drm;
 	struct device *dev;
+	struct device *dmadev;
 
 	struct drm_simple_display_pipe display_pipe;
 
diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c
index 0e2a376cb075..853f147036f6 100644
--- a/drivers/gpu/drm/udl/udl_main.c
+++ b/drivers/gpu/drm/udl/udl_main.c
@@ -315,6 +315,10 @@ int udl_init(struct udl_device *udl)
 
 	DRM_DEBUG("\n");
 
+	udl->dmadev = usb_intf_get_dma_device(to_usb_interface(dev->dev));
+	if (!udl->dmadev)
+		drm_warn(dev, "buffer sharing not supported"); /* not an error */
+
 	mutex_init(&udl->gem_lock);
 
 	if (!udl_parse_vendor_descriptor(udl)) {
@@ -343,12 +347,18 @@ int udl_init(struct udl_device *udl)
 err:
 	if (udl->urbs.count)
 		udl_free_urb_list(dev);
+	put_device(udl->dmadev);
 	DRM_ERROR("%d\n", ret);
 	return ret;
 }
 
 int udl_drop_usb(struct drm_device *dev)
 {
+	struct udl_device *udl = to_udl(dev);
+
 	udl_free_urb_list(dev);
+	put_device(udl->dmadev);
+	udl->dmadev = NULL;
+
 	return 0;
 }
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c
index eee9751009c2..1fda574579af 100644
--- a/drivers/gpu/drm/vc4/vc4_hdmi.c
+++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
@@ -1434,6 +1434,7 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *vc4_hdmi)
 	card->dai_link = dai_link;
 	card->num_links = 1;
 	card->name = vc4_hdmi->variant->card_name;
+	card->driver_name = "vc4-hdmi";
 	card->dev = dev;
 	card->owner = THIS_MODULE;
 
diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index 2b3a597fa65f..c239045e05d6 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -622,11 +622,11 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 	 * for now we just allocate globally.
 	 */
 	if (!hvs->hvs5)
-		/* 96kB */
-		drm_mm_init(&hvs->lbm_mm, 0, 96 * 1024);
+		/* 48k words of 2x12-bit pixels */
+		drm_mm_init(&hvs->lbm_mm, 0, 48 * 1024);
 	else
-		/* 70k words */
-		drm_mm_init(&hvs->lbm_mm, 0, 70 * 2 * 1024);
+		/* 60k words of 4x12-bit pixels */
+		drm_mm_init(&hvs->lbm_mm, 0, 60 * 1024);
 
 	/* Upload filter kernels.  We only have the one for now, so we
 	 * keep it around for the lifetime of the driver.
diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index eb37d7f8a118..c76e73a452e0 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -220,7 +220,7 @@ static void vc4_plane_reset(struct drm_plane *plane)
 	__drm_atomic_helper_plane_reset(plane, &vc4_state->base);
 }
 
-static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val)
+static void vc4_dlist_counter_increment(struct vc4_plane_state *vc4_state)
 {
 	if (vc4_state->dlist_count == vc4_state->dlist_size) {
 		u32 new_size = max(4u, vc4_state->dlist_count * 2);
@@ -235,7 +235,15 @@ static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val)
 		vc4_state->dlist_size = new_size;
 	}
 
-	vc4_state->dlist[vc4_state->dlist_count++] = val;
+	vc4_state->dlist_count++;
+}
+
+static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val)
+{
+	unsigned int idx = vc4_state->dlist_count;
+
+	vc4_dlist_counter_increment(vc4_state);
+	vc4_state->dlist[idx] = val;
 }
 
 /* Returns the scl0/scl1 field based on whether the dimensions need to
@@ -437,6 +445,7 @@ static void vc4_write_ppf(struct vc4_plane_state *vc4_state, u32 src, u32 dst)
 static u32 vc4_lbm_size(struct drm_plane_state *state)
 {
 	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
+	struct vc4_dev *vc4 = to_vc4_dev(state->plane->dev);
 	u32 pix_per_line;
 	u32 lbm;
 
@@ -472,7 +481,11 @@ static u32 vc4_lbm_size(struct drm_plane_state *state)
 		lbm = pix_per_line * 16;
 	}
 
-	lbm = roundup(lbm, 32);
+	/* Align it to 64 or 128 (hvs5) bytes */
+	lbm = roundup(lbm, vc4->hvs->hvs5 ? 128 : 64);
+
+	/* Each "word" of the LBM memory contains 2 or 4 (hvs5) pixels */
+	lbm /= vc4->hvs->hvs5 ? 4 : 2;
 
 	return lbm;
 }
@@ -912,9 +925,9 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
 		if (!vc4_state->is_unity) {
 			vc4_dlist_write(vc4_state,
 					VC4_SET_FIELD(vc4_state->crtc_w,
-						      SCALER_POS1_SCL_WIDTH) |
+						      SCALER5_POS1_SCL_WIDTH) |
 					VC4_SET_FIELD(vc4_state->crtc_h,
-						      SCALER_POS1_SCL_HEIGHT));
+						      SCALER5_POS1_SCL_HEIGHT));
 		}
 
 		/* Position Word 2: Source Image Size */
@@ -973,8 +986,10 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
 		 * be set when calling vc4_plane_allocate_lbm().
 		 */
 		if (vc4_state->y_scaling[0] != VC4_SCALING_NONE ||
-		    vc4_state->y_scaling[1] != VC4_SCALING_NONE)
-			vc4_state->lbm_offset = vc4_state->dlist_count++;
+		    vc4_state->y_scaling[1] != VC4_SCALING_NONE) {
+			vc4_state->lbm_offset = vc4_state->dlist_count;
+			vc4_dlist_counter_increment(vc4_state);
+		}
 
 		if (num_planes > 1) {
 			/* Emit Cb/Cr as channel 0 and Y as channel
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
index 0a900afc66ff..45c9c6a7f1d6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
@@ -500,8 +500,6 @@ vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
 	vm_fault_t ret;
 	pgoff_t fault_page_size;
 	bool write = vmf->flags & FAULT_FLAG_WRITE;
-	bool is_cow_mapping =
-		(vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE;
 
 	switch (pe_size) {
 	case PE_SIZE_PMD:
@@ -518,7 +516,7 @@ vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
 	}
 
 	/* Always do write dirty-tracking and COW on PTE level. */
-	if (write && (READ_ONCE(vbo->dirty) || is_cow_mapping))
+	if (write && (READ_ONCE(vbo->dirty) || is_cow_mapping(vma->vm_flags)))
 		return VM_FAULT_FALLBACK;
 
 	ret = ttm_bo_vm_reserve(bo, vmf);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
index 3c03b1746661..cb9975889e2f 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
@@ -49,7 +49,7 @@ int vmw_mmap(struct file *filp, struct vm_area_struct *vma)
 	vma->vm_ops = &vmw_vm_ops;
 
 	/* Use VM_PFNMAP rather than VM_MIXEDMAP if not a COW mapping */
-	if ((vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) != VM_MAYWRITE)
+	if (!is_cow_mapping(vma->vm_flags))
 		vma->vm_flags = (vma->vm_flags & ~VM_MIXEDMAP) | VM_PFNMAP;
 
 	return 0;
diff --git a/drivers/gpu/drm/xlnx/zynqmp_disp.c b/drivers/gpu/drm/xlnx/zynqmp_disp.c
index d8dd99be7b92..109d627968ac 100644
--- a/drivers/gpu/drm/xlnx/zynqmp_disp.c
+++ b/drivers/gpu/drm/xlnx/zynqmp_disp.c
@@ -1403,19 +1403,11 @@ static void zynqmp_disp_enable(struct zynqmp_disp *disp)
  */
 static void zynqmp_disp_disable(struct zynqmp_disp *disp)
 {
-	struct drm_crtc *crtc = &disp->crtc;
-
 	zynqmp_disp_audio_disable(&disp->audio);
 
 	zynqmp_disp_avbuf_disable_audio(&disp->avbuf);
 	zynqmp_disp_avbuf_disable_channels(&disp->avbuf);
 	zynqmp_disp_avbuf_disable(&disp->avbuf);
-
-	/* Mark the flip is done as crtc is disabled anyway */
-	if (crtc->state->event) {
-		complete_all(crtc->state->event->base.completion);
-		crtc->state->event = NULL;
-	}
 }
 
 static inline struct zynqmp_disp *crtc_to_disp(struct drm_crtc *crtc)
@@ -1503,6 +1495,13 @@ zynqmp_disp_crtc_atomic_disable(struct drm_crtc *crtc,
 
 	drm_crtc_vblank_off(&disp->crtc);
 
+	spin_lock_irq(&crtc->dev->event_lock);
+	if (crtc->state->event) {
+		drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+	}
+	spin_unlock_irq(&crtc->dev->event_lock);
+
 	clk_disable_unprepare(disp->pclk);
 	pm_runtime_put_sync(disp->dev);
 }