Merge drm/drm-next into drm-misc-next

Backmerging from drm/drm-next to the patches for AMD devices
for v5.14.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>

1 files changed, 35 insertions, 4 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index b066aa009b6f..dd1b1a464c28 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -26,10 +26,12 @@
 #include "kfd_priv.h"
 #include "kfd_device_queue_manager.h"
 #include "kfd_pm4_headers_vi.h"
+#include "kfd_pm4_headers_aldebaran.h"
 #include "cwsr_trap_handler.h"
 #include "kfd_iommu.h"
 #include "amdgpu_amdkfd.h"
 #include "kfd_smi_events.h"
+#include "kfd_migrate.h"
 
 #define MQD_SIZE_ALIGNED 768
 
@@ -80,6 +82,7 @@ static const struct kfd2kgd_calls *kfd2kgd_funcs[] = {
 	[CHIP_NAVY_FLOUNDER] = &gfx_v10_3_kfd2kgd,
 	[CHIP_VANGOGH] = &gfx_v10_3_kfd2kgd,
 	[CHIP_DIMGREY_CAVEFISH] = &gfx_v10_3_kfd2kgd,
+	[CHIP_BEIGE_GOBY] = &gfx_v10_3_kfd2kgd,
 };
 
 #ifdef KFD_SUPPORT_IOMMU_V2
@@ -556,6 +559,24 @@ static const struct kfd_device_info dimgrey_cavefish_device_info = {
 	.num_sdma_queues_per_engine = 8,
 };
 
+static const struct kfd_device_info beige_goby_device_info = {
+	.asic_family = CHIP_BEIGE_GOBY,
+	.asic_name = "beige_goby",
+	.max_pasid_bits = 16,
+	.max_no_of_hqd  = 24,
+	.doorbell_size  = 8,
+	.ih_ring_entry_size = 8 * sizeof(uint32_t),
+	.event_interrupt_class = &event_interrupt_class_v9,
+	.num_of_watch_points = 4,
+	.mqd_size_aligned = MQD_SIZE_ALIGNED,
+	.needs_iommu_device = false,
+	.supports_cwsr = true,
+	.needs_pci_atomics = true,
+	.num_sdma_engines = 1,
+	.num_xgmi_sdma_engines = 0,
+	.num_sdma_queues_per_engine = 8,
+};
+
 
 /* For each entry, [0] is regular and [1] is virtualisation device. */
 static const struct kfd_device_info *kfd_supported_devices[][2] = {
@@ -576,7 +597,7 @@ static const struct kfd_device_info *kfd_supported_devices[][2] = {
 	[CHIP_VEGA20] = {&vega20_device_info, NULL},
 	[CHIP_RENOIR] = {&renoir_device_info, NULL},
 	[CHIP_ARCTURUS] = {&arcturus_device_info, &arcturus_device_info},
-	[CHIP_ALDEBARAN] = {&aldebaran_device_info, NULL},
+	[CHIP_ALDEBARAN] = {&aldebaran_device_info, &aldebaran_device_info},
 	[CHIP_NAVI10] = {&navi10_device_info, NULL},
 	[CHIP_NAVI12] = {&navi12_device_info, &navi12_device_info},
 	[CHIP_NAVI14] = {&navi14_device_info, NULL},
@@ -584,6 +605,7 @@ static const struct kfd_device_info *kfd_supported_devices[][2] = {
 	[CHIP_NAVY_FLOUNDER] = {&navy_flounder_device_info, &navy_flounder_device_info},
 	[CHIP_VANGOGH] = {&vangogh_device_info, NULL},
 	[CHIP_DIMGREY_CAVEFISH] = {&dimgrey_cavefish_device_info, &dimgrey_cavefish_device_info},
+	[CHIP_BEIGE_GOBY] = {&beige_goby_device_info, &beige_goby_device_info},
 };
 
 static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
@@ -697,7 +719,9 @@ static int kfd_gws_init(struct kfd_dev *kfd)
 			&& kfd->device_info->asic_family <= CHIP_RAVEN
 			&& kfd->mec2_fw_version >= 0x1b3)
 		|| (kfd->device_info->asic_family == CHIP_ARCTURUS
-			&& kfd->mec2_fw_version >= 0x30))
+			&& kfd->mec2_fw_version >= 0x30)
+		|| (kfd->device_info->asic_family == CHIP_ALDEBARAN
+			&& kfd->mec2_fw_version >= 0x28))
 		ret = amdgpu_amdkfd_alloc_gws(kfd->kgd,
 				amdgpu_amdkfd_get_num_gws(kfd->kgd), &kfd->gws);
 
@@ -713,7 +737,7 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 			 struct drm_device *ddev,
 			 const struct kgd2kfd_shared_resources *gpu_resources)
 {
-	unsigned int size;
+	unsigned int size, map_process_packet_size;
 
 	kfd->ddev = ddev;
 	kfd->mec_fw_version = amdgpu_amdkfd_get_fw_version(kfd->kgd,
@@ -748,7 +772,11 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 	 * calculate max size of runlist packet.
 	 * There can be only 2 packets at once
 	 */
-	size += (KFD_MAX_NUM_OF_PROCESSES * sizeof(struct pm4_mes_map_process) +
+	map_process_packet_size =
+			kfd->device_info->asic_family == CHIP_ALDEBARAN ?
+				sizeof(struct pm4_mes_map_process_aldebaran) :
+					sizeof(struct pm4_mes_map_process);
+	size += (KFD_MAX_NUM_OF_PROCESSES * map_process_packet_size +
 		max_num_of_queues_per_device * sizeof(struct pm4_mes_map_queues)
 		+ sizeof(struct pm4_mes_runlist)) * 2;
 
@@ -814,6 +842,8 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 
 	kfd_cwsr_init(kfd);
 
+	svm_migrate_init((struct amdgpu_device *)kfd->kgd);
+
 	if (kfd_resume(kfd))
 		goto kfd_resume_error;
 
@@ -861,6 +891,7 @@ out:
 void kgd2kfd_device_exit(struct kfd_dev *kfd)
 {
 	if (kfd->init_complete) {
+		svm_migrate_fini((struct amdgpu_device *)kfd->kgd);
 		device_queue_manager_uninit(kfd->dqm);
 		kfd_interrupt_exit(kfd);
 		kfd_topology_remove_device(kfd);