diff options
Diffstat (limited to 'drivers/gpu/drm')
72 files changed, 1198 insertions, 520 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index aa2dcf578dd6..668939a14206 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -98,6 +98,9 @@ extern int amdgpu_sched_hw_submission; #define AMDGPU_MAX_COMPUTE_RINGS 8 #define AMDGPU_MAX_VCE_RINGS 2 +/* max number of IP instances */ +#define AMDGPU_MAX_SDMA_INSTANCES 2 + /* number of hw syncs before falling back on blocking */ #define AMDGPU_NUM_SYNCS 4 @@ -262,7 +265,7 @@ struct amdgpu_buffer_funcs { unsigned fill_num_dw; /* used for buffer clearing */ - void (*emit_fill_buffer)(struct amdgpu_ring *ring, + void (*emit_fill_buffer)(struct amdgpu_ib *ib, /* value to write to memory */ uint32_t src_data, /* dst addr in bytes */ @@ -340,6 +343,8 @@ struct amdgpu_ring_funcs { int (*test_ring)(struct amdgpu_ring *ring); int (*test_ib)(struct amdgpu_ring *ring); bool (*is_lockup)(struct amdgpu_ring *ring); + /* insert NOP packets */ + void (*insert_nop)(struct amdgpu_ring *ring, uint32_t count); }; /* @@ -440,12 +445,11 @@ int amdgpu_fence_wait_next(struct amdgpu_ring *ring); int amdgpu_fence_wait_empty(struct amdgpu_ring *ring); unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring); -signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev, - struct fence **array, - uint32_t count, - bool wait_all, - bool intr, - signed long t); +signed long amdgpu_fence_wait_any(struct amdgpu_device *adev, + struct fence **array, + uint32_t count, + bool intr, + signed long t); struct amdgpu_fence *amdgpu_fence_ref(struct amdgpu_fence *fence); void amdgpu_fence_unref(struct amdgpu_fence **fence); @@ -717,6 +721,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, void *owner); int amdgpu_sync_rings(struct amdgpu_sync *sync, struct amdgpu_ring *ring); +struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync); int amdgpu_sync_wait(struct amdgpu_sync *sync); void amdgpu_sync_free(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct fence *fence); @@ -1214,6 +1219,7 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev); void amdgpu_ring_free_size(struct amdgpu_ring *ring); int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw); int amdgpu_ring_lock(struct amdgpu_ring *ring, unsigned ndw); +void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count); void amdgpu_ring_commit(struct amdgpu_ring *ring); void amdgpu_ring_unlock_commit(struct amdgpu_ring *ring); void amdgpu_ring_undo(struct amdgpu_ring *ring); @@ -1665,7 +1671,6 @@ struct amdgpu_uvd { struct amdgpu_bo *vcpu_bo; void *cpu_addr; uint64_t gpu_addr; - void *saved_bo; atomic_t handles[AMDGPU_MAX_UVD_HANDLES]; struct drm_file *filp[AMDGPU_MAX_UVD_HANDLES]; struct delayed_work idle_work; @@ -1709,6 +1714,7 @@ struct amdgpu_sdma { uint32_t feature_version; struct amdgpu_ring ring; + bool burst_nop; }; /* @@ -2057,7 +2063,7 @@ struct amdgpu_device { struct amdgpu_gfx gfx; /* sdma */ - struct amdgpu_sdma sdma[2]; + struct amdgpu_sdma sdma[AMDGPU_MAX_SDMA_INSTANCES]; struct amdgpu_irq_src sdma_trap_irq; struct amdgpu_irq_src sdma_illegal_inst_irq; @@ -2196,6 +2202,21 @@ static inline void amdgpu_ring_write(struct amdgpu_ring *ring, uint32_t v) ring->ring_free_dw--; } +static inline struct amdgpu_sdma * amdgpu_get_sdma_instance(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + int i; + + for (i = 0; i < AMDGPU_MAX_SDMA_INSTANCES; i++) + if (&adev->sdma[i].ring == ring) + break; + + if (i < AMDGPU_MAX_SDMA_INSTANCES) + return &adev->sdma[i]; + else + return NULL; +} + /* * ASICs macro. */ @@ -2248,7 +2269,7 @@ static inline void amdgpu_ring_write(struct amdgpu_ring *ring, uint32_t v) #define amdgpu_display_stop_mc_access(adev, s) (adev)->mode_info.funcs->stop_mc_access((adev), (s)) #define amdgpu_display_resume_mc_access(adev, s) (adev)->mode_info.funcs->resume_mc_access((adev), (s)) #define amdgpu_emit_copy_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_copy_buffer((ib), (s), (d), (b)) -#define amdgpu_emit_fill_buffer(adev, r, s, d, b) (adev)->mman.buffer_funcs->emit_fill_buffer((r), (s), (d), (b)) +#define amdgpu_emit_fill_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_fill_buffer((ib), (s), (d), (b)) #define amdgpu_dpm_get_temperature(adev) (adev)->pm.funcs->get_temperature((adev)) #define amdgpu_dpm_pre_set_power_state(adev) (adev)->pm.funcs->pre_set_power_state((adev)) #define amdgpu_dpm_set_power_state(adev) (adev)->pm.funcs->set_power_state((adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 6a206f15635f..3b355aeb62fd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -354,7 +354,7 @@ int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p) * into account. We don't want to disallow buffer moves * completely. */ - if (current_domain != AMDGPU_GEM_DOMAIN_CPU && + if ((lobj->allowed_domains & current_domain) != 0 && (domain & current_domain) == 0 && /* will be moved */ bytes_moved > bytes_moved_threshold) { /* don't move it */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 42d1a22c1199..6ff6ae945794 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -244,7 +244,8 @@ static int amdgpu_vram_scratch_init(struct amdgpu_device *adev) if (adev->vram_scratch.robj == NULL) { r = amdgpu_bo_create(adev, AMDGPU_GPU_PAGE_SIZE, - PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_VRAM, 0, + PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, NULL, &adev->vram_scratch.robj); if (r) { return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c index 81b821247dde..8a122b1b7786 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c @@ -126,8 +126,8 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev, aligned_size = ALIGN(size, PAGE_SIZE); ret = amdgpu_gem_object_create(adev, aligned_size, 0, AMDGPU_GEM_DOMAIN_VRAM, - 0, true, - &gobj); + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + true, &gobj); if (ret) { printk(KERN_ERR "failed to allocate framebuffer (%d)\n", aligned_size); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index f446bf2fedc9..1be2bd6d07ea 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -851,22 +851,6 @@ static bool amdgpu_test_signaled_any(struct fence **fences, uint32_t count) return false; } -static bool amdgpu_test_signaled_all(struct fence **fences, uint32_t count) -{ - int idx; - struct fence *fence; - - for (idx = 0; idx < count; ++idx) { - fence = fences[idx]; - if (fence) { - if (!test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->flags)) - return false; - } - } - - return true; -} - struct amdgpu_wait_cb { struct fence_cb base; struct task_struct *task; @@ -885,7 +869,7 @@ static signed long amdgpu_fence_default_wait(struct fence *f, bool intr, struct amdgpu_fence *fence = to_amdgpu_fence(f); struct amdgpu_device *adev = fence->ring->adev; - return amdgpu_fence_wait_multiple(adev, &f, 1, false, intr, t); + return amdgpu_fence_wait_any(adev, &f, 1, intr, t); } /** @@ -894,23 +878,18 @@ static signed long amdgpu_fence_default_wait(struct fence *f, bool intr, * @adev: amdgpu device * @array: the fence array with amdgpu fence pointer * @count: the number of the fence array - * @wait_all: the flag of wait all(true) or wait any(false) * @intr: when sleep, set the current task interruptable or not * @t: timeout to wait * - * If wait_all is true, it will return when all fences are signaled or timeout. - * If wait_all is false, it will return when any fence is signaled or timeout. + * It will return when any fence is signaled or timeout. */ -signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev, - struct fence **array, - uint32_t count, - bool wait_all, - bool intr, - signed long t) -{ - long idx = 0; +signed long amdgpu_fence_wait_any(struct amdgpu_device *adev, + struct fence **array, uint32_t count, + bool intr, signed long t) +{ struct amdgpu_wait_cb *cb; struct fence *fence; + unsigned idx; BUG_ON(!array); @@ -927,10 +906,7 @@ signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev, if (fence_add_callback(fence, &cb[idx].base, amdgpu_fence_wait_cb)) { /* The fence is already signaled */ - if (wait_all) - continue; - else - goto fence_rm_cb; + goto fence_rm_cb; } } } @@ -945,9 +921,7 @@ signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev, * amdgpu_test_signaled_any must be called after * set_current_state to prevent a race with wake_up_process */ - if (!wait_all && amdgpu_test_signaled_any(array, count)) - break; - if (wait_all && amdgpu_test_signaled_all(array, count)) + if (amdgpu_test_signaled_any(array, count)) break; if (adev->needs_reset) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index e02db0b2e839..cbd3a486c5c2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -125,7 +125,8 @@ int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev) if (adev->gart.robj == NULL) { r = amdgpu_bo_create(adev, adev->gart.table_size, - PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_VRAM, 0, + PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, NULL, &adev->gart.robj); if (r) { return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 4afc507820c0..5839fab374bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -615,6 +615,7 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data, info.alignment = robj->tbo.mem.page_alignment << PAGE_SHIFT; info.domains = robj->initial_domain; info.domain_flags = robj->flags; + amdgpu_bo_unreserve(robj); if (copy_to_user(out, &info, sizeof(info))) r = -EFAULT; break; @@ -622,17 +623,19 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data, case AMDGPU_GEM_OP_SET_PLACEMENT: if (amdgpu_ttm_tt_has_userptr(robj->tbo.ttm)) { r = -EPERM; + amdgpu_bo_unreserve(robj); break; } robj->initial_domain = args->value & (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT | AMDGPU_GEM_DOMAIN_CPU); + amdgpu_bo_unreserve(robj); break; default: + amdgpu_bo_unreserve(robj); r = -EINVAL; } - amdgpu_bo_unreserve(robj); out: drm_gem_object_unreference_unlocked(gobj); return r; @@ -653,7 +656,8 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, r = amdgpu_gem_object_create(adev, args->size, 0, AMDGPU_GEM_DOMAIN_VRAM, - 0, ttm_bo_type_device, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + ttm_bo_type_device, &gobj); if (r) return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 57adcad2f7ba..08b09d55b96f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -127,7 +127,7 @@ static void amdgpu_ttm_placement_init(struct amdgpu_device *adev, placements[c].fpfn = adev->mc.visible_vram_size >> PAGE_SHIFT; placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | - TTM_PL_FLAG_VRAM; + TTM_PL_FLAG_VRAM | TTM_PL_FLAG_TOPDOWN; } placements[c].fpfn = 0; placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 7d442c51063e..9bec91484c24 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -131,6 +131,21 @@ int amdgpu_ring_lock(struct amdgpu_ring *ring, unsigned ndw) return 0; } +/** amdgpu_ring_insert_nop - insert NOP packets + * + * @ring: amdgpu_ring structure holding ring information + * @count: the number of NOP packets to insert + * + * This is the generic insert_nop function for rings except SDMA + */ +void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) +{ + int i; + + for (i = 0; i < count; i++) + amdgpu_ring_write(ring, ring->nop); +} + /** * amdgpu_ring_commit - tell the GPU to execute the new * commands on the ring buffer @@ -143,10 +158,13 @@ int amdgpu_ring_lock(struct amdgpu_ring *ring, unsigned ndw) */ void amdgpu_ring_commit(struct amdgpu_ring *ring) { + uint32_t count; + /* We pad to match fetch size */ - while (ring->wptr & ring->align_mask) { - amdgpu_ring_write(ring, ring->nop); - } + count = ring->align_mask + 1 - (ring->wptr & ring->align_mask); + count %= ring->align_mask + 1; + ring->funcs->insert_nop(ring, count); + mb(); amdgpu_ring_set_wptr(ring); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c index b92525329d6c..74dad270362c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c @@ -367,8 +367,8 @@ int amdgpu_sa_bo_new(struct amdgpu_device *adev, } while (amdgpu_sa_bo_next_hole(sa_manager, fences, tries)); spin_unlock(&sa_manager->wq.lock); - t = amdgpu_fence_wait_multiple(adev, fences, AMDGPU_MAX_RINGS, false, false, - MAX_SCHEDULE_TIMEOUT); + t = amdgpu_fence_wait_any(adev, fences, AMDGPU_MAX_RINGS, + false, MAX_SCHEDULE_TIMEOUT); r = (t > 0) ? 0 : t; spin_lock(&sa_manager->wq.lock); /* if we have nothing to wait for block */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c index f93fb3541488..de98fbd2971e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c @@ -27,6 +27,12 @@ #include <drm/drmP.h> #include "amdgpu.h" +static struct fence *amdgpu_sched_dependency(struct amd_sched_job *job) +{ + struct amdgpu_job *sched_job = (struct amdgpu_job *)job; + return amdgpu_sync_get_fence(&sched_job->ibs->sync); +} + static struct fence *amdgpu_sched_run_job(struct amd_sched_job *job) { struct amdgpu_job *sched_job; @@ -75,6 +81,7 @@ static void amdgpu_sched_process_job(struct amd_sched_job *job) } struct amd_sched_backend_ops amdgpu_sched_ops = { + .dependency = amdgpu_sched_dependency, .run_job = amdgpu_sched_run_job, .process_job = amdgpu_sched_process_job }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index 4fffb2539331..068aeaff7183 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -142,6 +142,18 @@ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, return 0; } +static void *amdgpu_sync_get_owner(struct fence *f) +{ + struct amdgpu_fence *a_fence = to_amdgpu_fence(f); + struct amd_sched_fence *s_fence = to_amd_sched_fence(f); + + if (s_fence) + return s_fence->owner; + else if (a_fence) + return a_fence->owner; + return AMDGPU_FENCE_OWNER_UNDEFINED; +} + /** * amdgpu_sync_resv - use the semaphores to sync to a reservation object * @@ -158,7 +170,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, { struct reservation_object_list *flist; struct fence *f; - struct amdgpu_fence *fence; + void *fence_owner; unsigned i; int r = 0; @@ -176,22 +188,22 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, for (i = 0; i < flist->shared_count; ++i) { f = rcu_dereference_protected(flist->shared[i], reservation_object_held(resv)); - fence = f ? to_amdgpu_fence(f) : NULL; - if (fence && fence->ring->adev == adev) { + if (amdgpu_sync_same_dev(adev, f)) { /* VM updates are only interesting * for other VM updates and moves. */ + fence_owner = amdgpu_sync_get_owner(f); if ((owner != AMDGPU_FENCE_OWNER_MOVE) && - (fence->owner != AMDGPU_FENCE_OWNER_MOVE) && + (fence_owner != AMDGPU_FENCE_OWNER_MOVE) && ((owner == AMDGPU_FENCE_OWNER_VM) != - (fence->owner == AMDGPU_FENCE_OWNER_VM))) + (fence_owner == AMDGPU_FENCE_OWNER_VM))) continue; /* Ignore fence from the same owner as * long as it isn't undefined. */ if (owner != AMDGPU_FENCE_OWNER_UNDEFINED && - fence->owner == owner) + fence_owner == owner) continue; } @@ -202,6 +214,28 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, return r; } +struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync) +{ + struct amdgpu_sync_entry *e; + struct hlist_node *tmp; + struct fence *f; + int i; + + hash_for_each_safe(sync->fences, i, tmp, e, node) { + + f = e->fence; + + hash_del(&e->node); + kfree(e); + + if (!fence_is_signaled(f)) + return f; + + fence_put(f); + } + return NULL; +} + int amdgpu_sync_wait(struct amdgpu_sync *sync) { struct amdgpu_sync_entry *e; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 399143541d8a..b5abd5cde413 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -859,7 +859,8 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) amdgpu_ttm_set_active_vram_size(adev, adev->mc.visible_vram_size); r = amdgpu_bo_create(adev, 256 * 1024, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, NULL, &adev->stollen_vga_memory); if (r) { return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index b87355ccfb1d..2cf6c6b06e3b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -154,7 +154,9 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) bo_size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8) + AMDGPU_UVD_STACK_SIZE + AMDGPU_UVD_HEAP_SIZE; r = amdgpu_bo_create(adev, bo_size, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &adev->uvd.vcpu_bo); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &adev->uvd.vcpu_bo); if (r) { dev_err(adev->dev, "(%d) failed to allocate UVD bo\n", r); return r; @@ -221,31 +223,32 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev) int amdgpu_uvd_suspend(struct amdgpu_device *adev) { - unsigned size; - void *ptr; - const struct common_firmware_header *hdr; - int i; + struct amdgpu_ring *ring = &adev->uvd.ring; + int i, r; if (adev->uvd.vcpu_bo == NULL) return 0; - for (i = 0; i < AMDGPU_MAX_UVD_HANDLES; ++i) - if (atomic_read(&adev->uvd.handles[i])) - break; - - if (i == AMDGPU_MAX_UVD_HANDLES) - return 0; + for (i = 0; i < AMDGPU_MAX_UVD_HANDLES; ++i) { + uint32_t handle = atomic_read(&adev->uvd.handles[i]); + if (handle != 0) { + struct fence *fence; - hdr = (const struct common_firmware_header *)adev->uvd.fw->data; + amdgpu_uvd_note_usage(adev); - size = amdgpu_bo_size(adev->uvd.vcpu_bo); - size -= le32_to_cpu(hdr->ucode_size_bytes); + r = amdgpu_uvd_get_destroy_msg(ring, handle, &fence); + if (r) { + DRM_ERROR("Error destroying UVD (%d)!\n", r); + continue; + } - ptr = adev->uvd.cpu_addr; - ptr += le32_to_cpu(hdr->ucode_size_bytes); + fence_wait(fence, false); + fence_put(fence); - adev->uvd.saved_bo = kmalloc(size, GFP_KERNEL); - memcpy(adev->uvd.saved_bo, ptr, size); + adev->uvd.filp[i] = NULL; + atomic_set(&adev->uvd.handles[i], 0); + } + } return 0; } @@ -270,12 +273,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev) ptr = adev->uvd.cpu_addr; ptr += le32_to_cpu(hdr->ucode_size_bytes); - if (adev->uvd.saved_bo != NULL) { - memcpy(ptr, adev->uvd.saved_bo, size); - kfree(adev->uvd.saved_bo); - adev->uvd.saved_bo = NULL; - } else - memset(ptr, 0, size); + memset(ptr, 0, size); return 0; } @@ -905,7 +903,9 @@ int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, int r, i; r = amdgpu_bo_create(adev, 1024, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &bo); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &bo); if (r) return r; @@ -952,7 +952,9 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, int r, i; r = amdgpu_bo_create(adev, 1024, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &bo); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &bo); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 1a984c934b1f..3cab96c42aa8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -141,7 +141,9 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) /* allocate firmware, stack and heap BO */ r = amdgpu_bo_create(adev, size, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &adev->vce.vcpu_bo); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &adev->vce.vcpu_bo); if (r) { dev_err(adev->dev, "(%d) failed to allocate VCE bo\n", r); return r; @@ -836,6 +838,10 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring) struct fence *fence = NULL; int r; + /* skip vce ring1 ib test for now, since it's not reliable */ + if (ring == &ring->adev->vce.ring[1]) + return 0; + r = amdgpu_vce_get_create_msg(ring, 1, NULL); if (r) { DRM_ERROR("amdgpu: failed to get create msg (%d).\n", r); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 83b7ce6f5f72..f68b7cdc370a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -627,9 +627,14 @@ static int amdgpu_vm_update_ptes(struct amdgpu_device *adev, { uint64_t mask = AMDGPU_VM_PTE_COUNT - 1; uint64_t last_pte = ~0, last_dst = ~0; + void *owner = AMDGPU_FENCE_OWNER_VM; unsigned count = 0; uint64_t addr; + /* sync to everything on unmapping */ + if (!(flags & AMDGPU_PTE_VALID)) + owner = AMDGPU_FENCE_OWNER_UNDEFINED; + /* walk over the address space and update the page tables */ for (addr = start; addr < end; ) { uint64_t pt_idx = addr >> amdgpu_vm_block_size; @@ -638,8 +643,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_device *adev, uint64_t pte; int r; - amdgpu_sync_resv(adev, &ib->sync, pt->tbo.resv, - AMDGPU_FENCE_OWNER_VM); + amdgpu_sync_resv(adev, &ib->sync, pt->tbo.resv, owner); r = reservation_object_reserve_shared(pt->tbo.resv); if (r) return r; @@ -790,17 +794,6 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, ib->length_dw = 0; - if (!(flags & AMDGPU_PTE_VALID)) { - unsigned i; - - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - struct amdgpu_fence *f = vm->ids[i].last_id_use; - r = amdgpu_sync_fence(adev, &ib->sync, &f->base); - if (r) - return r; - } - } - r = amdgpu_vm_update_ptes(adev, vm, ib, mapping->it.start, mapping->it.last + 1, addr + mapping->offset, flags, gtt_flags); @@ -1106,7 +1099,9 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, r = amdgpu_bo_create(adev, AMDGPU_VM_PTE_COUNT * 8, AMDGPU_GPU_PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &pt); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_NO_CPU_ACCESS, + NULL, &pt); if (r) goto error_free; @@ -1306,7 +1301,8 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) vm->page_directory_fence = NULL; r = amdgpu_bo_create(adev, pd_size, align, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_NO_CPU_ACCESS, NULL, &vm->page_directory); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c index 9ba0a7d5bc8e..92b6acadfc52 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c +++ b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c @@ -139,7 +139,8 @@ amdgpu_atombios_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *m tx_buf[0] = msg->address & 0xff; tx_buf[1] = msg->address >> 8; - tx_buf[2] = msg->request << 4; + tx_buf[2] = (msg->request << 4) | + ((msg->address >> 16) & 0xf); tx_buf[3] = msg->size ? (msg->size - 1) : 0; switch (msg->request & ~DP_AUX_I2C_MOT) { diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 3920c1e346f8..9ea9de457da3 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -188,6 +188,19 @@ static void cik_sdma_ring_set_wptr(struct amdgpu_ring *ring) WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[me], (ring->wptr << 2) & 0x3fffc); } +static void cik_sdma_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) +{ + struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ring); + int i; + + for (i = 0; i < count; i++) + if (sdma && sdma->burst_nop && (i == 0)) + amdgpu_ring_write(ring, ring->nop | + SDMA_NOP_COUNT(count - 1)); + else + amdgpu_ring_write(ring, ring->nop); +} + /** * cik_sdma_ring_emit_ib - Schedule an IB on the DMA engine * @@ -213,8 +226,8 @@ static void cik_sdma_ring_emit_ib(struct amdgpu_ring *ring, amdgpu_ring_write(ring, next_rptr); /* IB packet must end on a 8 DW boundary */ - while ((ring->wptr & 7) != 4) - amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); + cik_sdma_ring_insert_nop(ring, (12 - (ring->wptr & 7)) % 8); + amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_INDIRECT_BUFFER, 0, extra_bits)); amdgpu_ring_write(ring, ib->gpu_addr & 0xffffffe0); /* base must be 32 byte aligned */ amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xffffffff); @@ -501,6 +514,8 @@ static int cik_sdma_load_microcode(struct amdgpu_device *adev) fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; adev->sdma[i].fw_version = le32_to_cpu(hdr->header.ucode_version); adev->sdma[i].feature_version = le32_to_cpu(hdr->ucode_feature_version); + if (adev->sdma[i].feature_version >= 20) + adev->sdma[i].burst_nop = true; fw_data = (const __le32 *) (adev->sdma[i].fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], 0); @@ -815,8 +830,19 @@ static void cik_sdma_vm_set_pte_pde(struct amdgpu_ib *ib, */ static void cik_sdma_vm_pad_ib(struct amdgpu_ib *ib) { - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0); + struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ib->ring); + u32 pad_count; + int i; + + pad_count = (8 - (ib->length_dw & 0x7)) % 8; + for (i = 0; i < pad_count; i++) + if (sdma && sdma->burst_nop && (i == 0)) + ib->ptr[ib->length_dw++] = + SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0) | + SDMA_NOP_COUNT(pad_count - 1); + else + ib->ptr[ib->length_dw++] = + SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0); } /** @@ -1303,6 +1329,7 @@ static const struct amdgpu_ring_funcs cik_sdma_ring_funcs = { .test_ring = cik_sdma_ring_test_ring, .test_ib = cik_sdma_ring_test_ib, .is_lockup = cik_sdma_ring_is_lockup, + .insert_nop = cik_sdma_ring_insert_nop, }; static void cik_sdma_set_ring_funcs(struct amdgpu_device *adev) @@ -1363,16 +1390,16 @@ static void cik_sdma_emit_copy_buffer(struct amdgpu_ib *ib, * * Fill GPU buffers using the DMA engine (CIK). */ -static void cik_sdma_emit_fill_buffer(struct amdgpu_ring *ring, +static void cik_sdma_emit_fill_buffer(struct amdgpu_ib *ib, uint32_t src_data, uint64_t dst_offset, uint32_t byte_count) { - amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_CONSTANT_FILL, 0, 0)); - amdgpu_ring_write(ring, lower_32_bits(dst_offset)); - amdgpu_ring_write(ring, upper_32_bits(dst_offset)); - amdgpu_ring_write(ring, src_data); - amdgpu_ring_write(ring, byte_count); + ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_CONSTANT_FILL, 0, 0); + ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset); + ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset); + ib->ptr[ib->length_dw++] = src_data; + ib->ptr[ib->length_dw++] = byte_count; } static const struct amdgpu_buffer_funcs cik_sdma_buffer_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/cikd.h b/drivers/gpu/drm/amd/amdgpu/cikd.h index a3e3dfaa01a4..7f6d457f250a 100644 --- a/drivers/gpu/drm/amd/amdgpu/cikd.h +++ b/drivers/gpu/drm/amd/amdgpu/cikd.h @@ -487,6 +487,7 @@ (((op) & 0xFF) << 0)) /* sDMA opcodes */ #define SDMA_OPCODE_NOP 0 +# define SDMA_NOP_COUNT(x) (((x) & 0x3FFF) << 16) #define SDMA_OPCODE_COPY 1 # define SDMA_COPY_SUB_OPCODE_LINEAR 0 # define SDMA_COPY_SUB_OPCODE_TILED 1 diff --git a/drivers/gpu/drm/amd/amdgpu/cz_dpm.c b/drivers/gpu/drm/amd/amdgpu/cz_dpm.c index ace870afc7d4..44fa96ad4709 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_dpm.c @@ -1596,9 +1596,9 @@ static int cz_dpm_update_low_memory_pstate(struct amdgpu_device *adev) if (pi->sys_info.nb_dpm_enable) { if (ps->force_high) - cz_dpm_nbdpm_lm_pstate_enable(adev, true); - else cz_dpm_nbdpm_lm_pstate_enable(adev, false); + else + cz_dpm_nbdpm_lm_pstate_enable(adev, true); } return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index 4b255ac3043c..e4d101b1252a 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -1353,7 +1353,7 @@ static void dce_v10_0_program_watermarks(struct amdgpu_device *adev, tmp = REG_SET_FIELD(wm_mask, DPG_WATERMARK_MASK_CONTROL, URGENCY_WATERMARK_MASK, 2); WREG32(mmDPG_WATERMARK_MASK_CONTROL + amdgpu_crtc->crtc_offset, tmp); tmp = RREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset); - tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_LOW_WATERMARK, latency_watermark_a); + tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_LOW_WATERMARK, latency_watermark_b); tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_HIGH_WATERMARK, line_time); WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset, tmp); /* restore original selection */ diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 70eee807421f..6411e8244671 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -1329,7 +1329,7 @@ static void dce_v11_0_program_watermarks(struct amdgpu_device *adev, tmp = REG_SET_FIELD(wm_mask, DPG_WATERMARK_MASK_CONTROL, URGENCY_WATERMARK_MASK, 2); WREG32(mmDPG_WATERMARK_MASK_CONTROL + amdgpu_crtc->crtc_offset, tmp); tmp = RREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset); - tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_LOW_WATERMARK, latency_watermark_a); + tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_LOW_WATERMARK, latency_watermark_b); tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_HIGH_WATERMARK, line_time); WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset, tmp); /* restore original selection */ diff --git a/drivers/gpu/drm/amd/amdgpu/fiji_smc.c b/drivers/gpu/drm/amd/amdgpu/fiji_smc.c index 493c8c9c7faa..322edea65857 100644 --- a/drivers/gpu/drm/amd/amdgpu/fiji_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/fiji_smc.c @@ -762,7 +762,9 @@ int fiji_smu_init(struct amdgpu_device *adev) /* Allocate FW image data structure and header buffer */ ret = amdgpu_bo_create(adev, image_size, PAGE_SIZE, - true, AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, toc_buf); + true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, toc_buf); if (ret) { DRM_ERROR("Failed to allocate memory for TOC buffer\n"); return -ENOMEM; @@ -770,7 +772,9 @@ int fiji_smu_init(struct amdgpu_device *adev) /* Allocate buffer for SMU internal buffer */ ret = amdgpu_bo_create(adev, smu_internal_buffer_size, PAGE_SIZE, - true, AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, smu_buf); + true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, smu_buf); if (ret) { DRM_ERROR("Failed to allocate memory for SMU internal buffer\n"); return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index fab7b236f37f..4bd1e5cf65ca 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -3786,7 +3786,9 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev) /* save restore block */ if (adev->gfx.rlc.save_restore_obj == NULL) { r = amdgpu_bo_create(adev, dws * 4, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &adev->gfx.rlc.save_restore_obj); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &adev->gfx.rlc.save_restore_obj); if (r) { dev_warn(adev->dev, "(%d) create RLC sr bo failed\n", r); return r; @@ -3827,7 +3829,9 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev) if (adev->gfx.rlc.clear_state_obj == NULL) { r = amdgpu_bo_create(adev, dws * 4, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &adev->gfx.rlc.clear_state_obj); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &adev->gfx.rlc.clear_state_obj); if (r) { dev_warn(adev->dev, "(%d) create RLC c bo failed\n", r); gfx_v7_0_rlc_fini(adev); @@ -3864,7 +3868,9 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev) if (adev->gfx.rlc.cp_table_size) { if (adev->gfx.rlc.cp_table_obj == NULL) { r = amdgpu_bo_create(adev, adev->gfx.rlc.cp_table_size, PAGE_SIZE, true, - AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, &adev->gfx.rlc.cp_table_obj); + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, &adev->gfx.rlc.cp_table_obj); if (r) { dev_warn(adev->dev, "(%d) create RLC cp table bo failed\n", r); gfx_v7_0_rlc_fini(adev); @@ -5598,6 +5604,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = { .test_ring = gfx_v7_0_ring_test_ring, .test_ib = gfx_v7_0_ring_test_ib, .is_lockup = gfx_v7_0_ring_is_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { @@ -5614,6 +5621,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { .test_ring = gfx_v7_0_ring_test_ring, .test_ib = gfx_v7_0_ring_test_ib, .is_lockup = gfx_v7_0_ring_is_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void gfx_v7_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 818edb37fa9c..53f07439a512 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -2005,7 +2005,7 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev, } /** - * gmc_v8_0_init_compute_vmid - gart enable + * gfx_v8_0_init_compute_vmid - gart enable * * @rdev: amdgpu_device pointer * @@ -2015,7 +2015,7 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev, #define DEFAULT_SH_MEM_BASES (0x6000) #define FIRST_COMPUTE_VMID (8) #define LAST_COMPUTE_VMID (16) -static void gmc_v8_0_init_compute_vmid(struct amdgpu_device *adev) +static void gfx_v8_0_init_compute_vmid(struct amdgpu_device *adev) { int i; uint32_t sh_mem_config; @@ -2282,7 +2282,7 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev) vi_srbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); - gmc_v8_0_init_compute_vmid(adev); + gfx_v8_0_init_compute_vmid(adev); mutex_lock(&adev->grbm_idx_mutex); /* @@ -3240,7 +3240,8 @@ static int gfx_v8_0_cp_compute_resume(struct amdgpu_device *adev) /* enable the doorbell if requested */ if (use_doorbell) { - if (adev->asic_type == CHIP_CARRIZO) { + if ((adev->asic_type == CHIP_CARRIZO) || + (adev->asic_type == CHIP_FIJI)) { WREG32(mmCP_MEC_DOORBELL_RANGE_LOWER, AMDGPU_DOORBELL_KIQ << 2); WREG32(mmCP_MEC_DOORBELL_RANGE_UPPER, @@ -4378,6 +4379,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = { .test_ring = gfx_v8_0_ring_test_ring, .test_ib = gfx_v8_0_ring_test_ib, .is_lockup = gfx_v8_0_ring_is_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { @@ -4394,6 +4396,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { .test_ring = gfx_v8_0_ring_test_ring, .test_ib = gfx_v8_0_ring_test_ib, .is_lockup = gfx_v8_0_ring_is_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void gfx_v8_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 10218828face..774528ab8704 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -523,17 +523,11 @@ static int gmc_v7_0_gart_enable(struct amdgpu_device *adev) tmp = RREG32(mmVM_CONTEXT1_CNTL); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, VALID_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, READ_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_BLOCK_SIZE, amdgpu_vm_block_size - 9); @@ -852,6 +846,13 @@ static int gmc_v7_0_early_init(void *handle) return 0; } +static int gmc_v7_0_late_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); +} + static int gmc_v7_0_sw_init(void *handle) { int r; @@ -976,6 +977,7 @@ static int gmc_v7_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_irq_put(adev, &adev->mc.vm_fault, 0); gmc_v7_0_gart_disable(adev); return 0; @@ -1301,7 +1303,7 @@ static int gmc_v7_0_set_powergating_state(void *handle, const struct amd_ip_funcs gmc_v7_0_ip_funcs = { .early_init = gmc_v7_0_early_init, - .late_init = NULL, + .late_init = gmc_v7_0_late_init, .sw_init = gmc_v7_0_sw_init, .sw_fini = gmc_v7_0_sw_fini, .hw_init = gmc_v7_0_hw_init, diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 78109b750d29..9a07742620d0 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -653,19 +653,12 @@ static int gmc_v8_0_gart_enable(struct amdgpu_device *adev) tmp = RREG32(mmVM_CONTEXT1_CNTL); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, VALID_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, READ_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_BLOCK_SIZE, amdgpu_vm_block_size - 9); @@ -852,6 +845,13 @@ static int gmc_v8_0_early_init(void *handle) return 0; } +static int gmc_v8_0_late_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); +} + static int gmc_v8_0_sw_init(void *handle) { int r; @@ -978,6 +978,7 @@ static int gmc_v8_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_irq_put(adev, &adev->mc.vm_fault, 0); gmc_v8_0_gart_disable(adev); return 0; @@ -1288,7 +1289,7 @@ static int gmc_v8_0_set_powergating_state(void *handle, const struct amd_ip_funcs gmc_v8_0_ip_funcs = { .early_init = gmc_v8_0_early_init, - .late_init = NULL, + .late_init = gmc_v8_0_late_init, .sw_init = gmc_v8_0_sw_init, .sw_fini = gmc_v8_0_sw_fini, .hw_init = gmc_v8_0_hw_init, diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_sdma_pkt_open.h b/drivers/gpu/drm/amd/amdgpu/iceland_sdma_pkt_open.h index c723602c7b0c..ee6a041cb288 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_sdma_pkt_open.h +++ b/drivers/gpu/drm/amd/amdgpu/iceland_sdma_pkt_open.h @@ -2163,5 +2163,10 @@ #define SDMA_PKT_NOP_HEADER_sub_op_shift 8 #define SDMA_PKT_NOP_HEADER_SUB_OP(x) (((x) & SDMA_PKT_NOP_HEADER_sub_op_mask) << SDMA_PKT_NOP_HEADER_sub_op_shift) +/*define for count field*/ +#define SDMA_PKT_NOP_HEADER_count_offset 0 +#define SDMA_PKT_NOP_HEADER_count_mask 0x00003FFF +#define SDMA_PKT_NOP_HEADER_count_shift 16 +#define SDMA_PKT_NOP_HEADER_COUNT(x) (((x) & SDMA_PKT_NOP_HEADER_count_mask) << SDMA_PKT_NOP_HEADER_count_shift) #endif /* __ICELAND_SDMA_PKT_OPEN_H_ */ diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c index c6f1e2f12b5f..c900aa942ade 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c @@ -623,7 +623,9 @@ int iceland_smu_init(struct amdgpu_device *adev) /* Allocate FW image data structure and header buffer */ ret = amdgpu_bo_create(adev, image_size, PAGE_SIZE, - true, AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, toc_buf); + true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, toc_buf); if (ret) { DRM_ERROR("Failed to allocate memory for TOC buffer\n"); return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index 715e02d3bfba..14e87234171a 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -146,6 +146,8 @@ static int sdma_v2_4_init_microcode(struct amdgpu_device *adev) hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma[i].fw->data; adev->sdma[i].fw_version = le32_to_cpu(hdr->header.ucode_version); adev->sdma[i].feature_version = le32_to_cpu(hdr->ucode_feature_version); + if (adev->sdma[i].feature_version >= 20) + adev->sdma[i].burst_nop = true; if (adev->firmware.smu_load) { info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i]; @@ -218,6 +220,19 @@ static void sdma_v2_4_ring_set_wptr(struct amdgpu_ring *ring) WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[me], ring->wptr << 2); } +static void sdma_v2_4_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) +{ + struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ring); + int i; + + for (i = 0; i < count; i++) + if (sdma && sdma->burst_nop && (i == 0)) + amdgpu_ring_write(ring, ring->nop | + SDMA_PKT_NOP_HEADER_COUNT(count - 1)); + else + amdgpu_ring_write(ring, ring->nop); +} + /** * sdma_v2_4_ring_emit_ib - Schedule an IB on the DMA engine * @@ -245,8 +260,8 @@ static void sdma_v2_4_ring_emit_ib(struct amdgpu_ring *ring, amdgpu_ring_write(ring, next_rptr); /* IB packet must end on a 8 DW boundary */ - while ((ring->wptr & 7) != 2) - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_NOP)); + sdma_v2_4_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8); + amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) | SDMA_PKT_INDIRECT_HEADER_VMID(vmid)); /* base must be 32 byte aligned */ @@ -879,8 +894,19 @@ static void sdma_v2_4_vm_set_pte_pde(struct amdgpu_ib *ib, */ static void sdma_v2_4_vm_pad_ib(struct amdgpu_ib *ib) { - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_NOP); + struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ib->ring); + u32 pad_count; + int i; + + pad_count = (8 - (ib->length_dw & 0x7)) % 8; + for (i = 0; i < pad_count; i++) + if (sdma && sdma->burst_nop && (i == 0)) + ib->ptr[ib->length_dw++] = + SDMA_PKT_HEADER_OP(SDMA_OP_NOP) | + SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1); + else + ib->ptr[ib->length_dw++] = + SDMA_PKT_HEADER_OP(SDMA_OP_NOP); } /** @@ -1314,6 +1340,7 @@ static const struct amdgpu_ring_funcs sdma_v2_4_ring_funcs = { .test_ring = sdma_v2_4_ring_test_ring, .test_ib = sdma_v2_4_ring_test_ib, .is_lockup = sdma_v2_4_ring_is_lockup, + .insert_nop = sdma_v2_4_ring_insert_nop, }; static void sdma_v2_4_set_ring_funcs(struct amdgpu_device *adev) @@ -1375,16 +1402,16 @@ static void sdma_v2_4_emit_copy_buffer(struct amdgpu_ib *ib, * * Fill GPU buffers using the DMA engine (VI). */ -static void sdma_v2_4_emit_fill_buffer(struct amdgpu_ring *ring, +static void sdma_v2_4_emit_fill_buffer(struct amdgpu_ib *ib, uint32_t src_data, uint64_t dst_offset, uint32_t byte_count) { - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_CONST_FILL)); - amdgpu_ring_write(ring, lower_32_bits(dst_offset)); - amdgpu_ring_write(ring, upper_32_bits(dst_offset)); - amdgpu_ring_write(ring, src_data); - amdgpu_ring_write(ring, byte_count); + ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_CONST_FILL); + ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset); + ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset); + ib->ptr[ib->length_dw++] = src_data; + ib->ptr[ib->length_dw++] = byte_count; } static const struct amdgpu_buffer_funcs sdma_v2_4_buffer_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index 67128c8e78b8..9bfe92df15f7 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -218,6 +218,8 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev) hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma[i].fw->data; adev->sdma[i].fw_version = le32_to_cpu(hdr->header.ucode_version); adev->sdma[i].feature_version = le32_to_cpu(hdr->ucode_feature_version); + if (adev->sdma[i].feature_version >= 20) + adev->sdma[i].burst_nop = true; if (adev->firmware.smu_load) { info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i]; @@ -304,6 +306,19 @@ static void sdma_v3_0_ring_set_wptr(struct amdgpu_ring *ring) } } +static void sdma_v3_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) +{ + struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ring); + int i; + + for (i = 0; i < count; i++) + if (sdma && sdma->burst_nop && (i == 0)) + amdgpu_ring_write(ring, ring->nop | + SDMA_PKT_NOP_HEADER_COUNT(count - 1)); + else + amdgpu_ring_write(ring, ring->nop); +} + /** * sdma_v3_0_ring_emit_ib - Schedule an IB on the DMA engine * @@ -330,8 +345,7 @@ static void sdma_v3_0_ring_emit_ib(struct amdgpu_ring *ring, amdgpu_ring_write(ring, next_rptr); /* IB packet must end on a 8 DW boundary */ - while ((ring->wptr & 7) != 2) - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_NOP)); + sdma_v3_0_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8); amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) | SDMA_PKT_INDIRECT_HEADER_VMID(vmid)); @@ -999,8 +1013,19 @@ static void sdma_v3_0_vm_set_pte_pde(struct amdgpu_ib *ib, */ static void sdma_v3_0_vm_pad_ib(struct amdgpu_ib *ib) { - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_NOP); + struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ib->ring); + u32 pad_count; + int i; + + pad_count = (8 - (ib->length_dw & 0x7)) % 8; + for (i = 0; i < pad_count; i++) + if (sdma && sdma->burst_nop && (i == 0)) + ib->ptr[ib->length_dw++] = + SDMA_PKT_HEADER_OP(SDMA_OP_NOP) | + SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1); + else + ib->ptr[ib->length_dw++] = + SDMA_PKT_HEADER_OP(SDMA_OP_NOP); } /** @@ -1438,6 +1463,7 @@ static const struct amdgpu_ring_funcs sdma_v3_0_ring_funcs = { .test_ring = sdma_v3_0_ring_test_ring, .test_ib = sdma_v3_0_ring_test_ib, .is_lockup = sdma_v3_0_ring_is_lockup, + .insert_nop = sdma_v3_0_ring_insert_nop, }; static void sdma_v3_0_set_ring_funcs(struct amdgpu_device *adev) @@ -1499,16 +1525,16 @@ static void sdma_v3_0_emit_copy_buffer(struct amdgpu_ib *ib, * * Fill GPU buffers using the DMA engine (VI). */ -static void sdma_v3_0_emit_fill_buffer(struct amdgpu_ring *ring, +static void sdma_v3_0_emit_fill_buffer(struct amdgpu_ib *ib, uint32_t src_data, uint64_t dst_offset, uint32_t byte_count) { - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_CONST_FILL)); - amdgpu_ring_write(ring, lower_32_bits(dst_offset)); - amdgpu_ring_write(ring, upper_32_bits(dst_offset)); - amdgpu_ring_write(ring, src_data); - amdgpu_ring_write(ring, byte_count); + ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_CONST_FILL); + ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset); + ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset); + ib->ptr[ib->length_dw++] = src_data; + ib->ptr[ib->length_dw++] = byte_count; } static const struct amdgpu_buffer_funcs sdma_v3_0_buffer_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_sdma_pkt_open.h b/drivers/gpu/drm/amd/amdgpu/tonga_sdma_pkt_open.h index 099b7b56113c..e5ebd084288d 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_sdma_pkt_open.h +++ b/drivers/gpu/drm/amd/amdgpu/tonga_sdma_pkt_open.h @@ -2236,5 +2236,10 @@ #define SDMA_PKT_NOP_HEADER_sub_op_shift 8 #define SDMA_PKT_NOP_HEADER_SUB_OP(x) (((x) & SDMA_PKT_NOP_HEADER_sub_op_mask) << SDMA_PKT_NOP_HEADER_sub_op_shift) +/*define for count field*/ +#define SDMA_PKT_NOP_HEADER_count_offset 0 +#define SDMA_PKT_NOP_HEADER_count_mask 0x00003FFF +#define SDMA_PKT_NOP_HEADER_count_shift 16 +#define SDMA_PKT_NOP_HEADER_COUNT(x) (((x) & SDMA_PKT_NOP_HEADER_count_mask) << SDMA_PKT_NOP_HEADER_count_shift) #endif /* __TONGA_SDMA_PKT_OPEN_H_ */ diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_smc.c b/drivers/gpu/drm/amd/amdgpu/tonga_smc.c index 5fc53a40c7ac..1f5ac941a610 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_smc.c @@ -761,7 +761,9 @@ int tonga_smu_init(struct amdgpu_device *adev) /* Allocate FW image data structure and header buffer */ ret = amdgpu_bo_create(adev, image_size, PAGE_SIZE, - true, AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, toc_buf); + true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, toc_buf); if (ret) { DRM_ERROR("Failed to allocate memory for TOC buffer\n"); return -ENOMEM; @@ -769,7 +771,9 @@ int tonga_smu_init(struct amdgpu_device *adev) /* Allocate buffer for SMU internal buffer */ ret = amdgpu_bo_create(adev, smu_internal_buffer_size, PAGE_SIZE, - true, AMDGPU_GEM_DOMAIN_VRAM, 0, NULL, smu_buf); + true, AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, smu_buf); if (ret) { DRM_ERROR("Failed to allocate memory for SMU internal buffer\n"); return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 9ac383bc6c1f..5fac5da694f0 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -886,6 +886,7 @@ static const struct amdgpu_ring_funcs uvd_v4_2_ring_funcs = { .test_ring = uvd_v4_2_ring_test_ring, .test_ib = uvd_v4_2_ring_test_ib, .is_lockup = amdgpu_ring_test_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index de4b3f57902d..2d5c59c318af 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -825,6 +825,7 @@ static const struct amdgpu_ring_funcs uvd_v5_0_ring_funcs = { .test_ring = uvd_v5_0_ring_test_ring, .test_ib = uvd_v5_0_ring_test_ib, .is_lockup = amdgpu_ring_test_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index 66c975870e97..d9f553fce531 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -805,6 +805,7 @@ static const struct amdgpu_ring_funcs uvd_v6_0_ring_funcs = { .test_ring = uvd_v6_0_ring_test_ring, .test_ib = uvd_v6_0_ring_test_ib, .is_lockup = amdgpu_ring_test_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void uvd_v6_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c index 303d961d57bd..cd16df543f64 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c @@ -643,6 +643,7 @@ static const struct amdgpu_ring_funcs vce_v2_0_ring_funcs = { .test_ring = amdgpu_vce_ring_test_ring, .test_ib = amdgpu_vce_ring_test_ib, .is_lockup = amdgpu_ring_test_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void vce_v2_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index 4349658081ff..f0656dfb53f3 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -32,8 +32,8 @@ #include "vid.h" #include "vce/vce_3_0_d.h" #include "vce/vce_3_0_sh_mask.h" -#include "oss/oss_2_0_d.h" -#include "oss/oss_2_0_sh_mask.h" +#include "oss/oss_3_0_d.h" +#include "oss/oss_3_0_sh_mask.h" #include "gca/gfx_8_0_d.h" #include "smu/smu_7_1_2_d.h" #include "smu/smu_7_1_2_sh_mask.h" @@ -426,17 +426,41 @@ static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx) static bool vce_v3_0_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + u32 mask = 0; + int idx; - return !(RREG32(mmSRBM_STATUS2) & SRBM_STATUS2__VCE_BUSY_MASK); + for (idx = 0; idx < 2; ++idx) { + if (adev->vce.harvest_config & (1 << idx)) + continue; + + if (idx == 0) + mask |= SRBM_STATUS2__VCE0_BUSY_MASK; + else + mask |= SRBM_STATUS2__VCE1_BUSY_MASK; + } + + return !(RREG32(mmSRBM_STATUS2) & mask); } static int vce_v3_0_wait_for_idle(void *handle) { unsigned i; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + u32 mask = 0; + int idx; + + for (idx = 0; idx < 2; ++idx) { + if (adev->vce.harvest_config & (1 << idx)) + continue; + + if (idx == 0) + mask |= SRBM_STATUS2__VCE0_BUSY_MASK; + else + mask |= SRBM_STATUS2__VCE1_BUSY_MASK; + } for (i = 0; i < adev->usec_timeout; i++) { - if (!(RREG32(mmSRBM_STATUS2) & SRBM_STATUS2__VCE_BUSY_MASK)) + if (!(RREG32(mmSRBM_STATUS2) & mask)) return 0; } return -ETIMEDOUT; @@ -445,9 +469,21 @@ static int vce_v3_0_wait_for_idle(void *handle) static int vce_v3_0_soft_reset(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + u32 mask = 0; + int idx; + + for (idx = 0; idx < 2; ++idx) { + if (adev->vce.harvest_config & (1 << idx)) + continue; - WREG32_P(mmSRBM_SOFT_RESET, SRBM_SOFT_RESET__SOFT_RESET_VCE_MASK, - ~SRBM_SOFT_RESET__SOFT_RESET_VCE_MASK); + if (idx == 0) + mask |= SRBM_SOFT_RESET__SOFT_RESET_VCE0_MASK; + else + mask |= SRBM_SOFT_RESET__SOFT_RESET_VCE1_MASK; + } + WREG32_P(mmSRBM_SOFT_RESET, mask, + ~(SRBM_SOFT_RESET__SOFT_RESET_VCE0_MASK | + SRBM_SOFT_RESET__SOFT_RESET_VCE1_MASK)); mdelay(5); return vce_v3_0_start(adev); @@ -608,6 +644,7 @@ static const struct amdgpu_ring_funcs vce_v3_0_ring_funcs = { .test_ring = amdgpu_vce_ring_test_ring, .test_ib = amdgpu_vce_ring_test_ib, .is_lockup = amdgpu_ring_test_lockup, + .insert_nop = amdgpu_ring_insert_nop, }; static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c index d99fe90991dc..9259f1b6664c 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c +++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c @@ -27,6 +27,8 @@ #include <drm/drmP.h> #include "gpu_scheduler.h" +static struct amd_sched_job * +amd_sched_entity_pop_job(struct amd_sched_entity *entity); static void amd_sched_wakeup(struct amd_gpu_scheduler *sched); /* Initialize a given run queue struct */ @@ -56,34 +58,36 @@ static void amd_sched_rq_remove_entity(struct amd_sched_rq *rq, } /** - * Select next entity from a specified run queue with round robin policy. - * It could return the same entity as current one if current is the only - * available one in the queue. Return NULL if nothing available. + * Select next job from a specified run queue with round robin policy. + * Return NULL if nothing available. */ -static struct amd_sched_entity * -amd_sched_rq_select_entity(struct amd_sched_rq *rq) +static struct amd_sched_job * +amd_sched_rq_select_job(struct amd_sched_rq *rq) { struct amd_sched_entity *entity; + struct amd_sched_job *job; spin_lock(&rq->lock); entity = rq->current_entity; if (entity) { list_for_each_entry_continue(entity, &rq->entities, list) { - if (!kfifo_is_empty(&entity->job_queue)) { + job = amd_sched_entity_pop_job(entity); + if (job) { rq->current_entity = entity; spin_unlock(&rq->lock); - return rq->current_entity; + return job; } } } list_for_each_entry(entity, &rq->entities, list) { - if (!kfifo_is_empty(&entity->job_queue)) { + job = amd_sched_entity_pop_job(entity); + if (job) { rq->current_entity = entity; spin_unlock(&rq->lock); - return rq->current_entity; + return job; } if (entity == rq->current_entity) @@ -188,6 +192,39 @@ void amd_sched_entity_fini(struct amd_gpu_scheduler *sched, kfifo_free(&entity->job_queue); } +static void amd_sched_entity_wakeup(struct fence *f, struct fence_cb *cb) +{ + struct amd_sched_entity *entity = + container_of(cb, struct amd_sched_entity, cb); + entity->dependency = NULL; + fence_put(f); + amd_sched_wakeup(entity->scheduler); +} + +static struct amd_sched_job * +amd_sched_entity_pop_job(struct amd_sched_entity *entity) +{ + struct amd_gpu_scheduler *sched = entity->scheduler; + struct amd_sched_job *job; + + if (ACCESS_ONCE(entity->dependency)) + return NULL; + + if (!kfifo_out_peek(&entity->job_queue, &job, sizeof(job))) + return NULL; + + while ((entity->dependency = sched->ops->dependency(job))) { + + if (fence_add_callback(entity->dependency, &entity->cb, + amd_sched_entity_wakeup)) + fence_put(entity->dependency); + else + return NULL; + } + + return job; +} + /** * Helper to submit a job to the job queue * @@ -227,7 +264,6 @@ int amd_sched_entity_push_job(struct amd_sched_job *sched_job) struct amd_sched_entity *entity = sched_job->s_entity; struct amd_sched_fence *fence = amd_sched_fence_create( entity, sched_job->owner); - int r; if (!fence) return -ENOMEM; @@ -235,10 +271,10 @@ int amd_sched_entity_push_job(struct amd_sched_job *sched_job) fence_get(&fence->base); sched_job->s_fence = fence; - r = wait_event_interruptible(entity->scheduler->job_scheduled, - amd_sched_entity_in(sched_job)); + wait_event(entity->scheduler->job_scheduled, + amd_sched_entity_in(sched_job)); - return r; + return 0; } /** @@ -260,22 +296,22 @@ static void amd_sched_wakeup(struct amd_gpu_scheduler *sched) } /** - * Select next entity containing real IB submissions + * Select next to run */ -static struct amd_sched_entity * -amd_sched_select_context(struct amd_gpu_scheduler *sched) +static struct amd_sched_job * +amd_sched_select_job(struct amd_gpu_scheduler *sched) { - struct amd_sched_entity *tmp; + struct amd_sched_job *job; if (!amd_sched_ready(sched)) return NULL; /* Kernel run queue has higher priority than normal run queue*/ - tmp = amd_sched_rq_select_entity(&sched->kernel_rq); - if (tmp == NULL) - tmp = amd_sched_rq_select_entity(&sched->sched_rq); + job = amd_sched_rq_select_job(&sched->kernel_rq); + if (job == NULL) + job = amd_sched_rq_select_job(&sched->sched_rq); - return tmp; + return job; } static void amd_sched_process_job(struct fence *f, struct fence_cb *cb) @@ -296,27 +332,24 @@ static int amd_sched_main(void *param) { struct sched_param sparam = {.sched_priority = 1}; struct amd_gpu_scheduler *sched = (struct amd_gpu_scheduler *)param; - int r; + int r, count; sched_setscheduler(current, SCHED_FIFO, &sparam); while (!kthread_should_stop()) { - struct amd_sched_entity *c_entity = NULL; + struct amd_sched_entity *entity; struct amd_sched_job *job; struct fence *fence; wait_event_interruptible(sched->wake_up_worker, kthread_should_stop() || - (c_entity = amd_sched_select_context(sched))); + (job = amd_sched_select_job(sched))); - if (!c_entity) + if (!job) continue; - r = kfifo_out(&c_entity->job_queue, &job, sizeof(void *)); - if (r != sizeof(void *)) - continue; + entity = job->s_entity; atomic_inc(&sched->hw_rq_count); - fence = sched->ops->run_job(job); if (fence) { r = fence_add_callback(fence, &job->cb, @@ -328,6 +361,8 @@ static int amd_sched_main(void *param) fence_put(fence); } + count = kfifo_out(&entity->job_queue, &job, sizeof(job)); + WARN_ON(count != sizeof(job)); wake_up(&sched->job_scheduled); } return 0; diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h index e797796dcad7..2af0e4d4d817 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h +++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h @@ -45,6 +45,8 @@ struct amd_sched_entity { spinlock_t queue_lock; struct amd_gpu_scheduler *scheduler; uint64_t fence_context; + struct fence *dependency; + struct fence_cb cb; }; /** @@ -89,6 +91,7 @@ static inline struct amd_sched_fence *to_amd_sched_fence(struct fence *f) * these functions should be implemented in driver side */ struct amd_sched_backend_ops { + struct fence *(*dependency)(struct amd_sched_job *job); struct fence *(*run_job)(struct amd_sched_job *job); void (*process_job)(struct amd_sched_job *job); }; diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c index 59ebbe547290..084280859589 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fb.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c @@ -23,7 +23,6 @@ #include "exynos_drm_drv.h" #include "exynos_drm_fb.h" #include "exynos_drm_fbdev.h" -#include "exynos_drm_gem.h" #include "exynos_drm_iommu.h" #include "exynos_drm_crtc.h" @@ -33,12 +32,10 @@ * exynos specific framebuffer structure. * * @fb: drm framebuffer obejct. - * @buf_cnt: a buffer count to drm framebuffer. * @exynos_gem_obj: array of exynos specific gem object containing a gem object. */ struct exynos_drm_fb { struct drm_framebuffer fb; - unsigned int buf_cnt; struct exynos_drm_gem_obj *exynos_gem_obj[MAX_FB_BUFFER]; }; @@ -98,10 +95,6 @@ static int exynos_drm_fb_create_handle(struct drm_framebuffer *fb, { struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb); - /* This fb should have only one gem object. */ - if (WARN_ON(exynos_fb->buf_cnt != 1)) - return -EINVAL; - return drm_gem_handle_create(file_priv, &exynos_fb->exynos_gem_obj[0]->base, handle); } @@ -122,119 +115,77 @@ static struct drm_framebuffer_funcs exynos_drm_fb_funcs = { .dirty = exynos_drm_fb_dirty, }; -void exynos_drm_fb_set_buf_cnt(struct drm_framebuffer *fb, - unsigned int cnt) -{ - struct exynos_drm_fb *exynos_fb; - - exynos_fb = to_exynos_fb(fb); - - exynos_fb->buf_cnt = cnt; -} - -unsigned int exynos_drm_fb_get_buf_cnt(struct drm_framebuffer *fb) -{ - struct exynos_drm_fb *exynos_fb; - - exynos_fb = to_exynos_fb(fb); - - return exynos_fb->buf_cnt; -} - struct drm_framebuffer * exynos_drm_framebuffer_init(struct drm_device *dev, struct drm_mode_fb_cmd2 *mode_cmd, - struct drm_gem_object *obj) + struct exynos_drm_gem_obj **gem_obj, + int count) { struct exynos_drm_fb *exynos_fb; - struct exynos_drm_gem_obj *exynos_gem_obj; + int i; int ret; - exynos_gem_obj = to_exynos_gem_obj(obj); - - ret = check_fb_gem_memory_type(dev, exynos_gem_obj); - if (ret < 0) - return ERR_PTR(ret); - exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL); if (!exynos_fb) return ERR_PTR(-ENOMEM); + for (i = 0; i < count; i++) { + ret = check_fb_gem_memory_type(dev, gem_obj[i]); + if (ret < 0) + goto err; + + exynos_fb->exynos_gem_obj[i] = gem_obj[i]; + } + drm_helper_mode_fill_fb_struct(&exynos_fb->fb, mode_cmd); - exynos_fb->exynos_gem_obj[0] = exynos_gem_obj; ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs); - if (ret) { - kfree(exynos_fb); + if (ret < 0) { DRM_ERROR("failed to initialize framebuffer\n"); - return ERR_PTR(ret); + goto err; } return &exynos_fb->fb; + +err: + kfree(exynos_fb); + return ERR_PTR(ret); } static struct drm_framebuffer * exynos_user_fb_create(struct drm_device *dev, struct drm_file *file_priv, struct drm_mode_fb_cmd2 *mode_cmd) { + struct exynos_drm_gem_obj *gem_objs[MAX_FB_BUFFER]; struct drm_gem_object *obj; - struct exynos_drm_gem_obj *exynos_gem_obj; - struct exynos_drm_fb *exynos_fb; - int i, ret; - - exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL); - if (!exynos_fb) - return ERR_PTR(-ENOMEM); - - obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]); - if (!obj) { - DRM_ERROR("failed to lookup gem object\n"); - ret = -ENOENT; - goto err_free; - } - - drm_helper_mode_fill_fb_struct(&exynos_fb->fb, mode_cmd); - exynos_fb->exynos_gem_obj[0] = to_exynos_gem_obj(obj); - exynos_fb->buf_cnt = drm_format_num_planes(mode_cmd->pixel_format); - - DRM_DEBUG_KMS("buf_cnt = %d\n", exynos_fb->buf_cnt); + struct drm_framebuffer *fb; + int i; + int ret; - for (i = 1; i < exynos_fb->buf_cnt; i++) { + for (i = 0; i < drm_format_num_planes(mode_cmd->pixel_format); i++) { obj = drm_gem_object_lookup(dev, file_priv, - mode_cmd->handles[i]); + mode_cmd->handles[i]); if (!obj) { DRM_ERROR("failed to lookup gem object\n"); ret = -ENOENT; - exynos_fb->buf_cnt = i; - goto err_unreference; + goto err; } - exynos_gem_obj = to_exynos_gem_obj(obj); - exynos_fb->exynos_gem_obj[i] = exynos_gem_obj; - - ret = check_fb_gem_memory_type(dev, exynos_gem_obj); - if (ret < 0) - goto err_unreference; + gem_objs[i] = to_exynos_gem_obj(obj); } - ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs); - if (ret) { - DRM_ERROR("failed to init framebuffer.\n"); - goto err_unreference; + fb = exynos_drm_framebuffer_init(dev, mode_cmd, gem_objs, i); + if (IS_ERR(fb)) { + ret = PTR_ERR(fb); + goto err; } - return &exynos_fb->fb; + return fb; -err_unreference: - for (i = 0; i < exynos_fb->buf_cnt; i++) { - struct drm_gem_object *obj; +err: + while (i--) + drm_gem_object_unreference_unlocked(&gem_objs[i]->base); - obj = &exynos_fb->exynos_gem_obj[i]->base; - if (obj) - drm_gem_object_unreference_unlocked(obj); - } -err_free: - kfree(exynos_fb); return ERR_PTR(ret); } diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.h b/drivers/gpu/drm/exynos/exynos_drm_fb.h index 1c9e27c32cd1..85e4445b920e 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fb.h +++ b/drivers/gpu/drm/exynos/exynos_drm_fb.h @@ -14,10 +14,13 @@ #ifndef _EXYNOS_DRM_FB_H_ #define _EXYNOS_DRM_FB_H +#include "exynos_drm_gem.h" + struct drm_framebuffer * exynos_drm_framebuffer_init(struct drm_device *dev, struct drm_mode_fb_cmd2 *mode_cmd, - struct drm_gem_object *obj); + struct exynos_drm_gem_obj **gem_obj, + int count); /* get gem object of a drm framebuffer */ struct exynos_drm_gem_obj *exynos_drm_fb_gem_obj(struct drm_framebuffer *fb, @@ -25,11 +28,4 @@ struct exynos_drm_gem_obj *exynos_drm_fb_gem_obj(struct drm_framebuffer *fb, void exynos_drm_mode_config_init(struct drm_device *dev); -/* set a buffer count to drm framebuffer. */ -void exynos_drm_fb_set_buf_cnt(struct drm_framebuffer *fb, - unsigned int cnt); - -/* get a buffer count to drm framebuffer. */ -unsigned int exynos_drm_fb_get_buf_cnt(struct drm_framebuffer *fb); - #endif diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c index 624595afbce0..a221f753ad9c 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c @@ -21,7 +21,6 @@ #include "exynos_drm_drv.h" #include "exynos_drm_fb.h" #include "exynos_drm_fbdev.h" -#include "exynos_drm_gem.h" #include "exynos_drm_iommu.h" #define MAX_CONNECTOR 4 @@ -32,7 +31,7 @@ struct exynos_drm_fbdev { struct drm_fb_helper drm_fb_helper; - struct exynos_drm_gem_obj *exynos_gem_obj; + struct exynos_drm_gem_obj *obj; }; static int exynos_drm_fb_mmap(struct fb_info *info, @@ -40,7 +39,7 @@ static int exynos_drm_fb_mmap(struct fb_info *info, { struct drm_fb_helper *helper = info->par; struct exynos_drm_fbdev *exynos_fbd = to_exynos_fbdev(helper); - struct exynos_drm_gem_obj *obj = exynos_fbd->exynos_gem_obj; + struct exynos_drm_gem_obj *obj = exynos_fbd->obj; unsigned long vm_size; int ret; @@ -75,37 +74,38 @@ static struct fb_ops exynos_drm_fb_ops = { }; static int exynos_drm_fbdev_update(struct drm_fb_helper *helper, - struct drm_fb_helper_surface_size *sizes, - struct drm_framebuffer *fb) + struct drm_fb_helper_surface_size *sizes, + struct exynos_drm_gem_obj *obj) { - struct fb_info *fbi = helper->fbdev; - struct exynos_drm_gem_obj *obj; + struct fb_info *fbi; + struct drm_framebuffer *fb = helper->fb; unsigned int size = fb->width * fb->height * (fb->bits_per_pixel >> 3); unsigned int nr_pages; unsigned long offset; + fbi = drm_fb_helper_alloc_fbi(helper); + if (IS_ERR(fbi)) { + DRM_ERROR("failed to allocate fb info.\n"); + return PTR_ERR(fbi); + } + + fbi->par = helper; + fbi->flags = FBINFO_FLAG_DEFAULT; + fbi->fbops = &exynos_drm_fb_ops; + drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth); drm_fb_helper_fill_var(fbi, helper, sizes->fb_width, sizes->fb_height); - /* RGB formats use only one buffer */ - obj = exynos_drm_fb_gem_obj(fb, 0); - if (!obj) { - DRM_DEBUG_KMS("gem object is null.\n"); - return -EFAULT; - } - nr_pages = obj->size >> PAGE_SHIFT; obj->kvaddr = (void __iomem *) vmap(obj->pages, nr_pages, VM_MAP, pgprot_writecombine(PAGE_KERNEL)); if (!obj->kvaddr) { DRM_ERROR("failed to map pages to kernel space.\n"); + drm_fb_helper_release_fbi(helper); return -EIO; } - /* buffer count to framebuffer always is 1 at booting time. */ - exynos_drm_fb_set_buf_cnt(fb, 1); - offset = fbi->var.xoffset * (fb->bits_per_pixel >> 3); offset += fbi->var.yoffset * fb->pitches[0]; @@ -120,9 +120,8 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes) { struct exynos_drm_fbdev *exynos_fbdev = to_exynos_fbdev(helper); - struct exynos_drm_gem_obj *exynos_gem_obj; + struct exynos_drm_gem_obj *obj; struct drm_device *dev = helper->dev; - struct fb_info *fbi; struct drm_mode_fb_cmd2 mode_cmd = { 0 }; struct platform_device *pdev = dev->platformdev; unsigned long size; @@ -140,47 +139,34 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, mutex_lock(&dev->struct_mutex); - fbi = drm_fb_helper_alloc_fbi(helper); - if (IS_ERR(fbi)) { - DRM_ERROR("failed to allocate fb info.\n"); - ret = PTR_ERR(fbi); - goto out; - } - size = mode_cmd.pitches[0] * mode_cmd.height; - exynos_gem_obj = exynos_drm_gem_create(dev, EXYNOS_BO_CONTIG, size); + obj = exynos_drm_gem_create(dev, EXYNOS_BO_CONTIG, size); /* * If physically contiguous memory allocation fails and if IOMMU is * supported then try to get buffer from non physically contiguous * memory area. */ - if (IS_ERR(exynos_gem_obj) && is_drm_iommu_supported(dev)) { + if (IS_ERR(obj) && is_drm_iommu_supported(dev)) { dev_warn(&pdev->dev, "contiguous FB allocation failed, falling back to non-contiguous\n"); - exynos_gem_obj = exynos_drm_gem_create(dev, EXYNOS_BO_NONCONTIG, - size); + obj = exynos_drm_gem_create(dev, EXYNOS_BO_NONCONTIG, size); } - if (IS_ERR(exynos_gem_obj)) { - ret = PTR_ERR(exynos_gem_obj); - goto err_release_fbi; + if (IS_ERR(obj)) { + ret = PTR_ERR(obj); + goto out; } - exynos_fbdev->exynos_gem_obj = exynos_gem_obj; + exynos_fbdev->obj = obj; - helper->fb = exynos_drm_framebuffer_init(dev, &mode_cmd, - &exynos_gem_obj->base); + helper->fb = exynos_drm_framebuffer_init(dev, &mode_cmd, &obj, 1); if (IS_ERR(helper->fb)) { DRM_ERROR("failed to create drm framebuffer.\n"); ret = PTR_ERR(helper->fb); goto err_destroy_gem; } - fbi->par = helper; - fbi->flags = FBINFO_FLAG_DEFAULT; - fbi->fbops = &exynos_drm_fb_ops; - - ret = exynos_drm_fbdev_update(helper, sizes, helper->fb); + ret = exynos_drm_fbdev_update(helper, sizes, obj); if (ret < 0) goto err_destroy_framebuffer; @@ -190,9 +176,7 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, err_destroy_framebuffer: drm_framebuffer_cleanup(helper->fb); err_destroy_gem: - exynos_drm_gem_destroy(exynos_gem_obj); -err_release_fbi: - drm_fb_helper_release_fbi(helper); + exynos_drm_gem_destroy(obj); /* * if failed, all resources allocated above would be released by @@ -285,11 +269,11 @@ static void exynos_drm_fbdev_destroy(struct drm_device *dev, struct drm_fb_helper *fb_helper) { struct exynos_drm_fbdev *exynos_fbd = to_exynos_fbdev(fb_helper); - struct exynos_drm_gem_obj *exynos_gem_obj = exynos_fbd->exynos_gem_obj; + struct exynos_drm_gem_obj *obj = exynos_fbd->obj; struct drm_framebuffer *fb; - if (exynos_gem_obj->kvaddr) - vunmap(exynos_gem_obj->kvaddr); + if (obj->kvaddr) + vunmap(obj->kvaddr); /* release drm framebuffer and real buffer */ if (fb_helper->fb && fb_helper->fb->funcs) { diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c index ba008391a2fc..535b4ad6c4b1 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c +++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c @@ -48,11 +48,13 @@ /* registers for base address */ #define G2D_SRC_BASE_ADDR 0x0304 +#define G2D_SRC_STRIDE_REG 0x0308 #define G2D_SRC_COLOR_MODE 0x030C #define G2D_SRC_LEFT_TOP 0x0310 #define G2D_SRC_RIGHT_BOTTOM 0x0314 #define G2D_SRC_PLANE2_BASE_ADDR 0x0318 #define G2D_DST_BASE_ADDR 0x0404 +#define G2D_DST_STRIDE_REG 0x0408 #define G2D_DST_COLOR_MODE 0x040C #define G2D_DST_LEFT_TOP 0x0410 #define G2D_DST_RIGHT_BOTTOM 0x0414 @@ -148,6 +150,7 @@ struct g2d_cmdlist { * A structure of buffer description * * @format: color format + * @stride: buffer stride/pitch in bytes * @left_x: the x coordinates of left top corner * @top_y: the y coordinates of left top corner * @right_x: the x coordinates of right bottom corner @@ -156,6 +159,7 @@ struct g2d_cmdlist { */ struct g2d_buf_desc { unsigned int format; + unsigned int stride; unsigned int left_x; unsigned int top_y; unsigned int right_x; @@ -589,6 +593,7 @@ static enum g2d_reg_type g2d_get_reg_type(int reg_offset) switch (reg_offset) { case G2D_SRC_BASE_ADDR: + case G2D_SRC_STRIDE_REG: case G2D_SRC_COLOR_MODE: case G2D_SRC_LEFT_TOP: case G2D_SRC_RIGHT_BOTTOM: @@ -598,6 +603,7 @@ static enum g2d_reg_type g2d_get_reg_type(int reg_offset) reg_type = REG_TYPE_SRC_PLANE2; break; case G2D_DST_BASE_ADDR: + case G2D_DST_STRIDE_REG: case G2D_DST_COLOR_MODE: case G2D_DST_LEFT_TOP: case G2D_DST_RIGHT_BOTTOM: @@ -652,8 +658,8 @@ static bool g2d_check_buf_desc_is_valid(struct g2d_buf_desc *buf_desc, enum g2d_reg_type reg_type, unsigned long size) { - unsigned int width, height; - unsigned long area; + int width, height; + unsigned long bpp, last_pos; /* * check source and destination buffers only. @@ -662,22 +668,37 @@ static bool g2d_check_buf_desc_is_valid(struct g2d_buf_desc *buf_desc, if (reg_type != REG_TYPE_SRC && reg_type != REG_TYPE_DST) return true; - width = buf_desc->right_x - buf_desc->left_x; + /* This check also makes sure that right_x > left_x. */ + width = (int)buf_desc->right_x - (int)buf_desc->left_x; if (width < G2D_LEN_MIN || width > G2D_LEN_MAX) { - DRM_ERROR("width[%u] is out of range!\n", width); + DRM_ERROR("width[%d] is out of range!\n", width); return false; } - height = buf_desc->bottom_y - buf_desc->top_y; + /* This check also makes sure that bottom_y > top_y. */ + height = (int)buf_desc->bottom_y - (int)buf_desc->top_y; if (height < G2D_LEN_MIN || height > G2D_LEN_MAX) { - DRM_ERROR("height[%u] is out of range!\n", height); + DRM_ERROR("height[%d] is out of range!\n", height); return false; } - area = (unsigned long)width * (unsigned long)height * - g2d_get_buf_bpp(buf_desc->format); - if (area > size) { - DRM_ERROR("area[%lu] is out of range[%lu]!\n", area, size); + bpp = g2d_get_buf_bpp(buf_desc->format); + + /* Compute the position of the last byte that the engine accesses. */ + last_pos = ((unsigned long)buf_desc->bottom_y - 1) * + (unsigned long)buf_desc->stride + + (unsigned long)buf_desc->right_x * bpp - 1; + + /* + * Since right_x > left_x and bottom_y > top_y we already know + * that the first_pos < last_pos (first_pos being the position + * of the first byte the engine accesses), it just remains to + * check if last_pos is smaller then the buffer size. + */ + + if (last_pos >= size) { + DRM_ERROR("last engine access position [%lu] " + "is out of range [%lu]!\n", last_pos, size); return false; } @@ -973,8 +994,6 @@ static int g2d_check_reg_offset(struct device *dev, goto err; reg_type = g2d_get_reg_type(reg_offset); - if (reg_type == REG_TYPE_NONE) - goto err; /* check userptr buffer type. */ if ((cmdlist->data[index] & ~0x7fffffff) >> 31) { @@ -983,14 +1002,22 @@ static int g2d_check_reg_offset(struct device *dev, } else buf_info->types[reg_type] = BUF_TYPE_GEM; break; + case G2D_SRC_STRIDE_REG: + case G2D_DST_STRIDE_REG: + if (for_addr) + goto err; + + reg_type = g2d_get_reg_type(reg_offset); + + buf_desc = &buf_info->descs[reg_type]; + buf_desc->stride = cmdlist->data[index + 1]; + break; case G2D_SRC_COLOR_MODE: case G2D_DST_COLOR_MODE: if (for_addr) goto err; reg_type = g2d_get_reg_type(reg_offset); - if (reg_type == REG_TYPE_NONE) - goto err; buf_desc = &buf_info->descs[reg_type]; value = cmdlist->data[index + 1]; @@ -1003,8 +1030,6 @@ static int g2d_check_reg_offset(struct device *dev, goto err; reg_type = g2d_get_reg_type(reg_offset); - if (reg_type == REG_TYPE_NONE) - goto err; buf_desc = &buf_info->descs[reg_type]; value = cmdlist->data[index + 1]; @@ -1018,8 +1043,6 @@ static int g2d_check_reg_offset(struct device *dev, goto err; reg_type = g2d_get_reg_type(reg_offset); - if (reg_type == REG_TYPE_NONE) - goto err; buf_desc = &buf_info->descs[reg_type]; value = cmdlist->data[index + 1]; diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c index 865d6eb0c845..714822441467 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_plane.c +++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c @@ -126,7 +126,7 @@ static int exynos_plane_atomic_check(struct drm_plane *plane, if (!state->fb) return 0; - nr = exynos_drm_fb_get_buf_cnt(state->fb); + nr = drm_format_num_planes(state->fb->pixel_format); for (i = 0; i < nr; i++) { struct exynos_drm_gem_obj *obj = exynos_drm_fb_gem_obj(state->fb, i); diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 33aabc79813b..e3ec9049081f 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2562,6 +2562,8 @@ static const char *power_domain_str(enum intel_display_power_domain domain) return "PORT_DDI_D_2_LANES"; case POWER_DOMAIN_PORT_DDI_D_4_LANES: return "PORT_DDI_D_4_LANES"; + case POWER_DOMAIN_PORT_DDI_E_2_LANES: + return "PORT_DDI_E_2_LANES"; case POWER_DOMAIN_PORT_DSI: return "PORT_DSI"; case POWER_DOMAIN_PORT_CRT: diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 1d887459e37f..8edcec8ae592 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -662,15 +662,18 @@ static int i915_drm_suspend_late(struct drm_device *drm_dev, bool hibernation) pci_disable_device(drm_dev->pdev); /* - * During hibernation on some GEN4 platforms the BIOS may try to access + * During hibernation on some platforms the BIOS may try to access * the device even though it's already in D3 and hang the machine. So * leave the device in D0 on those platforms and hope the BIOS will - * power down the device properly. Platforms where this was seen: - * Lenovo Thinkpad X301, X61s + * power down the device properly. The issue was seen on multiple old + * GENs with different BIOS vendors, so having an explicit blacklist + * is inpractical; apply the workaround on everything pre GEN6. The + * platforms where the issue was seen: + * Lenovo Thinkpad X301, X61s, X60, T60, X41 + * Fujitsu FSC S7110 + * Acer Aspire 1830T */ - if (!(hibernation && - drm_dev->pdev->subsystem_vendor == PCI_VENDOR_ID_LENOVO && - INTEL_INFO(dev_priv)->gen == 4)) + if (!(hibernation && INTEL_INFO(dev_priv)->gen < 6)) pci_set_power_state(drm_dev->pdev, PCI_D3hot); return 0; diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 089459b39771..b06e03080771 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -182,6 +182,7 @@ enum intel_display_power_domain { POWER_DOMAIN_PORT_DDI_C_4_LANES, POWER_DOMAIN_PORT_DDI_D_2_LANES, POWER_DOMAIN_PORT_DDI_D_4_LANES, + POWER_DOMAIN_PORT_DDI_E_2_LANES, POWER_DOMAIN_PORT_DSI, POWER_DOMAIN_PORT_CRT, POWER_DOMAIN_PORT_OTHER, @@ -1416,6 +1417,10 @@ enum modeset_restore { #define DP_AUX_C 0x20 #define DP_AUX_D 0x30 +#define DDC_PIN_B 0x05 +#define DDC_PIN_C 0x04 +#define DDC_PIN_D 0x06 + struct ddi_vbt_port_info { /* * This is an index in the HDMI/DVI DDI buffer translation table. @@ -1430,6 +1435,7 @@ struct ddi_vbt_port_info { uint8_t supports_dp:1; uint8_t alternate_aux_channel; + uint8_t alternate_ddc_pin; uint8_t dp_boost_level; uint8_t hdmi_boost_level; @@ -1922,6 +1928,8 @@ struct drm_i915_private { struct skl_wm_values skl_hw; struct vlv_wm_values vlv; }; + + uint8_t max_level; } wm; struct i915_runtime_pm pm; @@ -3377,13 +3385,13 @@ int intel_freq_opcode(struct drm_i915_private *dev_priv, int val); #define I915_READ64(reg) dev_priv->uncore.funcs.mmio_readq(dev_priv, (reg), true) #define I915_READ64_2x32(lower_reg, upper_reg) ({ \ - u32 upper, lower, tmp; \ - tmp = I915_READ(upper_reg); \ + u32 upper, lower, old_upper, loop = 0; \ + upper = I915_READ(upper_reg); \ do { \ - upper = tmp; \ + old_upper = upper; \ lower = I915_READ(lower_reg); \ - tmp = I915_READ(upper_reg); \ - } while (upper != tmp); \ + upper = I915_READ(upper_reg); \ + } while (upper != old_upper && loop++ < 2); \ (u64)upper << 32 | lower; }) #define POSTING_READ(reg) (void)I915_READ_NOTRACE(reg) diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 923a3c4bf0b7..a953d4975b8c 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -1032,6 +1032,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas, u32 old_read = obj->base.read_domains; u32 old_write = obj->base.write_domain; + obj->dirty = 1; /* be paranoid */ obj->base.write_domain = obj->base.pending_write_domain; if (obj->base.write_domain == 0) obj->base.pending_read_domains |= obj->base.read_domains; @@ -1039,7 +1040,6 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas, i915_vma_move_to_active(vma, req); if (obj->base.write_domain) { - obj->dirty = 1; i915_gem_request_assign(&obj->last_write_req, req); intel_fb_obj_invalidate(obj, ORIGIN_CS); diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index d94c92d842fb..a2bceb70a3fd 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1558,7 +1558,7 @@ static void i9xx_hpd_irq_handler(struct drm_device *dev) u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915; intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger, - hotplug_trigger, hpd_status_g4x, + hotplug_trigger, hpd_status_i915, i9xx_port_hotplug_long_detect); intel_hpd_irq_handler(dev, pin_mask, long_mask); } diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index be83b77aa018..b3e437b3bb54 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -905,23 +905,23 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, uint8_t hdmi_level_shift; int i, j; bool is_dvi, is_hdmi, is_dp, is_edp, is_crt; - uint8_t aux_channel; + uint8_t aux_channel, ddc_pin; /* Each DDI port can have more than one value on the "DVO Port" field, * so look for all the possible values for each port and abort if more * than one is found. */ - int dvo_ports[][2] = { - {DVO_PORT_HDMIA, DVO_PORT_DPA}, - {DVO_PORT_HDMIB, DVO_PORT_DPB}, - {DVO_PORT_HDMIC, DVO_PORT_DPC}, - {DVO_PORT_HDMID, DVO_PORT_DPD}, - {DVO_PORT_CRT, -1 /* Port E can only be DVO_PORT_CRT */ }, + int dvo_ports[][3] = { + {DVO_PORT_HDMIA, DVO_PORT_DPA, -1}, + {DVO_PORT_HDMIB, DVO_PORT_DPB, -1}, + {DVO_PORT_HDMIC, DVO_PORT_DPC, -1}, + {DVO_PORT_HDMID, DVO_PORT_DPD, -1}, + {DVO_PORT_CRT, DVO_PORT_HDMIE, DVO_PORT_DPE}, }; /* Find the child device to use, abort if more than one found. */ for (i = 0; i < dev_priv->vbt.child_dev_num; i++) { it = dev_priv->vbt.child_dev + i; - for (j = 0; j < 2; j++) { + for (j = 0; j < 3; j++) { if (dvo_ports[port][j] == -1) break; @@ -939,6 +939,7 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, return; aux_channel = child->raw[25]; + ddc_pin = child->common.ddc_pin; is_dvi = child->common.device_type & DEVICE_TYPE_TMDS_DVI_SIGNALING; is_dp = child->common.device_type & DEVICE_TYPE_DISPLAYPORT_OUTPUT; @@ -970,11 +971,27 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, DRM_DEBUG_KMS("Port %c is internal DP\n", port_name(port)); if (is_dvi) { - if (child->common.ddc_pin == 0x05 && port != PORT_B) + if (port == PORT_E) { + info->alternate_ddc_pin = ddc_pin; + /* if DDIE share ddc pin with other port, then + * dvi/hdmi couldn't exist on the shared port. + * Otherwise they share the same ddc bin and system + * couldn't communicate with them seperately. */ + if (ddc_pin == DDC_PIN_B) { + dev_priv->vbt.ddi_port_info[PORT_B].supports_dvi = 0; + dev_priv->vbt.ddi_port_info[PORT_B].supports_hdmi = 0; + } else if (ddc_pin == DDC_PIN_C) { + dev_priv->vbt.ddi_port_info[PORT_C].supports_dvi = 0; + dev_priv->vbt.ddi_port_info[PORT_C].supports_hdmi = 0; + } else if (ddc_pin == DDC_PIN_D) { + dev_priv->vbt.ddi_port_info[PORT_D].supports_dvi = 0; + dev_priv->vbt.ddi_port_info[PORT_D].supports_hdmi = 0; + } + } else if (ddc_pin == DDC_PIN_B && port != PORT_B) DRM_DEBUG_KMS("Unexpected DDC pin for port B\n"); - if (child->common.ddc_pin == 0x04 && port != PORT_C) + else if (ddc_pin == DDC_PIN_C && port != PORT_C) DRM_DEBUG_KMS("Unexpected DDC pin for port C\n"); - if (child->common.ddc_pin == 0x06 && port != PORT_D) + else if (ddc_pin == DDC_PIN_D && port != PORT_D) DRM_DEBUG_KMS("Unexpected DDC pin for port D\n"); } diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h index 06d0dbde2be6..46cd5c7ebacd 100644 --- a/drivers/gpu/drm/i915/intel_bios.h +++ b/drivers/gpu/drm/i915/intel_bios.h @@ -758,11 +758,6 @@ int intel_parse_bios(struct drm_device *dev); #define DVO_C 2 #define DVO_D 3 -/* define the PORT for DP output type */ -#define PORT_IDPB 7 -#define PORT_IDPC 8 -#define PORT_IDPD 9 - /* Possible values for the "DVO Port" field for versions >= 155: */ #define DVO_PORT_HDMIA 0 #define DVO_PORT_HDMIB 1 @@ -775,6 +770,8 @@ int intel_parse_bios(struct drm_device *dev); #define DVO_PORT_DPC 8 #define DVO_PORT_DPD 9 #define DVO_PORT_DPA 10 +#define DVO_PORT_DPE 11 +#define DVO_PORT_HDMIE 12 #define DVO_PORT_MIPIA 21 #define DVO_PORT_MIPIB 22 #define DVO_PORT_MIPIC 23 diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index ba1ae031e6fd..d0f1b8d833cd 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -350,7 +350,7 @@ static void finish_csr_load(const struct firmware *fw, void *context) } csr->mmio_count = dmc_header->mmio_count; for (i = 0; i < dmc_header->mmio_count; i++) { - if (dmc_header->mmioaddr[i] < CSR_MMIO_START_RANGE && + if (dmc_header->mmioaddr[i] < CSR_MMIO_START_RANGE || dmc_header->mmioaddr[i] > CSR_MMIO_END_RANGE) { DRM_ERROR(" Firmware has wrong mmio address 0x%x\n", dmc_header->mmioaddr[i]); diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index 19004557c868..61575f67a626 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -1554,17 +1554,14 @@ skl_ddi_pll_select(struct intel_crtc *intel_crtc, DPLL_CFGCR2_PDIV(wrpll_params.pdiv) | wrpll_params.central_freq; } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT) { - struct drm_encoder *encoder = &intel_encoder->base; - struct intel_dp *intel_dp = enc_to_intel_dp(encoder); - - switch (intel_dp->link_bw) { - case DP_LINK_BW_1_62: + switch (crtc_state->port_clock / 2) { + case 81000: ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, 0); break; - case DP_LINK_BW_2_7: + case 135000: ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1350, 0); break; - case DP_LINK_BW_5_4: + case 270000: ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2700, 0); break; } diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 53f5476bc4bb..8cc9264f7809 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -5150,7 +5150,6 @@ static enum intel_display_power_domain port_to_power_domain(enum port port) { switch (port) { case PORT_A: - case PORT_E: return POWER_DOMAIN_PORT_DDI_A_4_LANES; case PORT_B: return POWER_DOMAIN_PORT_DDI_B_4_LANES; @@ -5158,6 +5157,8 @@ static enum intel_display_power_domain port_to_power_domain(enum port port) return POWER_DOMAIN_PORT_DDI_C_4_LANES; case PORT_D: return POWER_DOMAIN_PORT_DDI_D_4_LANES; + case PORT_E: + return POWER_DOMAIN_PORT_DDI_E_2_LANES; default: WARN_ON_ONCE(1); return POWER_DOMAIN_PORT_OTHER; @@ -5712,16 +5713,13 @@ void skl_init_cdclk(struct drm_i915_private *dev_priv) /* enable PG1 and Misc I/O */ intel_display_power_get(dev_priv, POWER_DOMAIN_PLLS); - /* DPLL0 already enabed !? */ - if (I915_READ(LCPLL1_CTL) & LCPLL_PLL_ENABLE) { - DRM_DEBUG_DRIVER("DPLL0 already running\n"); - return; + /* DPLL0 not enabled (happens on early BIOS versions) */ + if (!(I915_READ(LCPLL1_CTL) & LCPLL_PLL_ENABLE)) { + /* enable DPLL0 */ + required_vco = skl_cdclk_get_vco(dev_priv->skl_boot_cdclk); + skl_dpll0_enable(dev_priv, required_vco); } - /* enable DPLL0 */ - required_vco = skl_cdclk_get_vco(dev_priv->skl_boot_cdclk); - skl_dpll0_enable(dev_priv, required_vco); - /* set CDCLK to the frequency the BIOS chose */ skl_set_cdclk(dev_priv, dev_priv->skl_boot_cdclk); @@ -6307,7 +6305,7 @@ static void intel_connector_check_state(struct intel_connector *connector) connector->base.name); if (connector->get_hw_state(connector)) { - struct drm_encoder *encoder = &connector->encoder->base; + struct intel_encoder *encoder = connector->encoder; struct drm_connector_state *conn_state = connector->base.state; I915_STATE_WARN(!crtc, @@ -6319,13 +6317,13 @@ static void intel_connector_check_state(struct intel_connector *connector) I915_STATE_WARN(!crtc->state->active, "connector is active, but attached crtc isn't\n"); - if (!encoder) + if (!encoder || encoder->type == INTEL_OUTPUT_DP_MST) return; - I915_STATE_WARN(conn_state->best_encoder != encoder, + I915_STATE_WARN(conn_state->best_encoder != &encoder->base, "atomic encoder doesn't match attached encoder\n"); - I915_STATE_WARN(conn_state->crtc != encoder->crtc, + I915_STATE_WARN(conn_state->crtc != encoder->base.crtc, "attached encoder crtc differs from connector crtc\n"); } else { I915_STATE_WARN(crtc && crtc->state->active, @@ -13963,6 +13961,15 @@ static void intel_setup_outputs(struct drm_device *dev) intel_ddi_init(dev, PORT_C); if (found & SFUSE_STRAP_DDID_DETECTED) intel_ddi_init(dev, PORT_D); + /* + * On SKL we don't have a way to detect DDI-E so we rely on VBT. + */ + if (IS_SKYLAKE(dev) && + (dev_priv->vbt.ddi_port_info[PORT_E].supports_dp || + dev_priv->vbt.ddi_port_info[PORT_E].supports_dvi || + dev_priv->vbt.ddi_port_info[PORT_E].supports_hdmi)) + intel_ddi_init(dev, PORT_E); + } else if (HAS_PCH_SPLIT(dev)) { int found; dpd_is_edp = intel_dp_is_edp(dev, PORT_D); @@ -14733,6 +14740,24 @@ void intel_modeset_init(struct drm_device *dev) if (INTEL_INFO(dev)->num_pipes == 0) return; + /* + * There may be no VBT; and if the BIOS enabled SSC we can + * just keep using it to avoid unnecessary flicker. Whereas if the + * BIOS isn't using it, don't assume it will work even if the VBT + * indicates as much. + */ + if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) { + bool bios_lvds_use_ssc = !!(I915_READ(PCH_DREF_CONTROL) & + DREF_SSC1_ENABLE); + + if (dev_priv->vbt.lvds_use_ssc != bios_lvds_use_ssc) { + DRM_DEBUG_KMS("SSC %sabled by BIOS, overriding VBT which says %sabled\n", + bios_lvds_use_ssc ? "en" : "dis", + dev_priv->vbt.lvds_use_ssc ? "en" : "dis"); + dev_priv->vbt.lvds_use_ssc = bios_lvds_use_ssc; + } + } + intel_init_display(dev); intel_init_audio(dev); @@ -15292,7 +15317,6 @@ err: void intel_modeset_gem_init(struct drm_device *dev) { - struct drm_i915_private *dev_priv = dev->dev_private; struct drm_crtc *c; struct drm_i915_gem_object *obj; int ret; @@ -15301,16 +15325,6 @@ void intel_modeset_gem_init(struct drm_device *dev) intel_init_gt_powersave(dev); mutex_unlock(&dev->struct_mutex); - /* - * There may be no VBT; and if the BIOS enabled SSC we can - * just keep using it to avoid unnecessary flicker. Whereas if the - * BIOS isn't using it, don't assume it will work even if the VBT - * indicates as much. - */ - if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) - dev_priv->vbt.lvds_use_ssc = !!(I915_READ(PCH_DREF_CONTROL) & - DREF_SSC1_ENABLE); - intel_modeset_init_hw(dev); intel_setup_overlay(dev); diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 3781cd3e358a..0a2e33fbf20d 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -48,28 +48,28 @@ #define INTEL_DP_RESOLUTION_FAILSAFE (3 << INTEL_DP_RESOLUTION_SHIFT_MASK) struct dp_link_dpll { - int link_bw; + int clock; struct dpll dpll; }; static const struct dp_link_dpll gen4_dpll[] = { - { DP_LINK_BW_1_62, + { 162000, { .p1 = 2, .p2 = 10, .n = 2, .m1 = 23, .m2 = 8 } }, - { DP_LINK_BW_2_7, + { 270000, { .p1 = 1, .p2 = 10, .n = 1, .m1 = 14, .m2 = 2 } } }; static const struct dp_link_dpll pch_dpll[] = { - { DP_LINK_BW_1_62, + { 162000, { .p1 = 2, .p2 = 10, .n = 1, .m1 = 12, .m2 = 9 } }, - { DP_LINK_BW_2_7, + { 270000, { .p1 = 1, .p2 = 10, .n = 2, .m1 = 14, .m2 = 8 } } }; static const struct dp_link_dpll vlv_dpll[] = { - { DP_LINK_BW_1_62, + { 162000, { .p1 = 3, .p2 = 2, .n = 5, .m1 = 3, .m2 = 81 } }, - { DP_LINK_BW_2_7, + { 270000, { .p1 = 2, .p2 = 2, .n = 1, .m1 = 2, .m2 = 27 } } }; @@ -83,11 +83,11 @@ static const struct dp_link_dpll chv_dpll[] = { * m2 is stored in fixed point format using formula below * (m2_int << 22) | m2_fraction */ - { DP_LINK_BW_1_62, /* m2_int = 32, m2_fraction = 1677722 */ + { 162000, /* m2_int = 32, m2_fraction = 1677722 */ { .p1 = 4, .p2 = 2, .n = 1, .m1 = 2, .m2 = 0x819999a } }, - { DP_LINK_BW_2_7, /* m2_int = 27, m2_fraction = 0 */ + { 270000, /* m2_int = 27, m2_fraction = 0 */ { .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } }, - { DP_LINK_BW_5_4, /* m2_int = 27, m2_fraction = 0 */ + { 540000, /* m2_int = 27, m2_fraction = 0 */ { .p1 = 2, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } } }; @@ -1130,7 +1130,7 @@ intel_dp_connector_unregister(struct intel_connector *intel_connector) } static void -skl_edp_set_pll_config(struct intel_crtc_state *pipe_config, int link_clock) +skl_edp_set_pll_config(struct intel_crtc_state *pipe_config) { u32 ctrl1; @@ -1142,7 +1142,7 @@ skl_edp_set_pll_config(struct intel_crtc_state *pipe_config, int link_clock) pipe_config->dpll_hw_state.cfgcr2 = 0; ctrl1 = DPLL_CTRL1_OVERRIDE(SKL_DPLL0); - switch (link_clock / 2) { + switch (pipe_config->port_clock / 2) { case 81000: ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, SKL_DPLL0); @@ -1175,20 +1175,20 @@ skl_edp_set_pll_config(struct intel_crtc_state *pipe_config, int link_clock) pipe_config->dpll_hw_state.ctrl1 = ctrl1; } -static void -hsw_dp_set_ddi_pll_sel(struct intel_crtc_state *pipe_config, int link_bw) +void +hsw_dp_set_ddi_pll_sel(struct intel_crtc_state *pipe_config) { memset(&pipe_config->dpll_hw_state, 0, sizeof(pipe_config->dpll_hw_state)); - switch (link_bw) { - case DP_LINK_BW_1_62: + switch (pipe_config->port_clock / 2) { + case 81000: pipe_config->ddi_pll_sel = PORT_CLK_SEL_LCPLL_810; break; - case DP_LINK_BW_2_7: + case 135000: pipe_config->ddi_pll_sel = PORT_CLK_SEL_LCPLL_1350; break; - case DP_LINK_BW_5_4: + case 270000: pipe_config->ddi_pll_sel = PORT_CLK_SEL_LCPLL_2700; break; } @@ -1245,7 +1245,7 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates) static void intel_dp_set_clock(struct intel_encoder *encoder, - struct intel_crtc_state *pipe_config, int link_bw) + struct intel_crtc_state *pipe_config) { struct drm_device *dev = encoder->base.dev; const struct dp_link_dpll *divisor = NULL; @@ -1267,7 +1267,7 @@ intel_dp_set_clock(struct intel_encoder *encoder, if (divisor && count) { for (i = 0; i < count; i++) { - if (link_bw == divisor[i].link_bw) { + if (pipe_config->port_clock == divisor[i].clock) { pipe_config->dpll = divisor[i].dpll; pipe_config->clock_set = true; break; @@ -1544,13 +1544,13 @@ found: } if (IS_SKYLAKE(dev) && is_edp(intel_dp)) - skl_edp_set_pll_config(pipe_config, common_rates[clock]); + skl_edp_set_pll_config(pipe_config); else if (IS_BROXTON(dev)) /* handled in ddi */; else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) - hsw_dp_set_ddi_pll_sel(pipe_config, intel_dp->link_bw); + hsw_dp_set_ddi_pll_sel(pipe_config); else - intel_dp_set_clock(encoder, pipe_config, intel_dp->link_bw); + intel_dp_set_clock(encoder, pipe_config); return true; } @@ -4961,9 +4961,12 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd) intel_dp_probe_oui(intel_dp); - if (!intel_dp_probe_mst(intel_dp)) + if (!intel_dp_probe_mst(intel_dp)) { + drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); + intel_dp_check_link_status(intel_dp); + drm_modeset_unlock(&dev->mode_config.connection_mutex); goto mst_fail; - + } } else { if (intel_dp->is_mst) { if (intel_dp_check_mst_status(intel_dp) == -EINVAL) @@ -4971,10 +4974,6 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd) } if (!intel_dp->is_mst) { - /* - * we'll check the link status via the normal hot plug path later - - * but for short hpds we should check it now - */ drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); intel_dp_check_link_status(intel_dp); drm_modeset_unlock(&dev->mode_config.connection_mutex); @@ -5016,16 +5015,17 @@ intel_trans_dp_port_sel(struct drm_crtc *crtc) return -1; } -/* check the VBT to see whether the eDP is on DP-D port */ +/* check the VBT to see whether the eDP is on another port */ bool intel_dp_is_edp(struct drm_device *dev, enum port port) { struct drm_i915_private *dev_priv = dev->dev_private; union child_device_config *p_child; int i; static const short port_mapping[] = { - [PORT_B] = PORT_IDPB, - [PORT_C] = PORT_IDPC, - [PORT_D] = PORT_IDPD, + [PORT_B] = DVO_PORT_DPB, + [PORT_C] = DVO_PORT_DPC, + [PORT_D] = DVO_PORT_DPD, + [PORT_E] = DVO_PORT_DPE, }; if (port == PORT_A) diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c index 369f8b6b804f..3e4be5a3becd 100644 --- a/drivers/gpu/drm/i915/intel_dp_mst.c +++ b/drivers/gpu/drm/i915/intel_dp_mst.c @@ -33,6 +33,7 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { + struct drm_device *dev = encoder->base.dev; struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base); struct intel_digital_port *intel_dig_port = intel_mst->primary; struct intel_dp *intel_dp = &intel_dig_port->dp; @@ -97,6 +98,10 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder, &pipe_config->dp_m_n); pipe_config->dp_m_n.tu = slots; + + if (IS_HASWELL(dev) || IS_BROADWELL(dev)) + hsw_dp_set_ddi_pll_sel(pipe_config); + return true; } @@ -168,6 +173,11 @@ static void intel_mst_pre_enable_dp(struct intel_encoder *encoder) return; } + /* MST encoders are bound to a crtc, not to a connector, + * force the mapping here for get_hw_state. + */ + found->encoder = encoder; + DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links); intel_mst->port = found->port; @@ -395,7 +405,7 @@ static const struct drm_encoder_funcs intel_dp_mst_enc_funcs = { static bool intel_dp_mst_get_hw_state(struct intel_connector *connector) { - if (connector->encoder) { + if (connector->encoder && connector->base.state->crtc) { enum pipe pipe; if (!connector->encoder->get_hw_state(connector->encoder, &pipe)) return false; diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 93008fbb815d..2b9e6f9775c5 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1185,6 +1185,7 @@ void intel_edp_drrs_disable(struct intel_dp *intel_dp); void intel_edp_drrs_invalidate(struct drm_device *dev, unsigned frontbuffer_bits); void intel_edp_drrs_flush(struct drm_device *dev, unsigned frontbuffer_bits); +void hsw_dp_set_ddi_pll_sel(struct intel_crtc_state *pipe_config); /* intel_dp_mst.c */ int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port, int conn_id); diff --git a/drivers/gpu/drm/i915/intel_dsi.c b/drivers/gpu/drm/i915/intel_dsi.c index 4a601cf90f16..32a6c7184ca4 100644 --- a/drivers/gpu/drm/i915/intel_dsi.c +++ b/drivers/gpu/drm/i915/intel_dsi.c @@ -1048,11 +1048,7 @@ void intel_dsi_init(struct drm_device *dev) intel_connector->unregister = intel_connector_unregister; /* Pipe A maps to MIPI DSI port A, pipe B maps to MIPI DSI port C */ - if (dev_priv->vbt.dsi.config->dual_link) { - /* XXX: does dual link work on either pipe? */ - intel_encoder->crtc_mask = (1 << PIPE_A); - intel_dsi->ports = ((1 << PORT_A) | (1 << PORT_C)); - } else if (dev_priv->vbt.dsi.port == DVO_PORT_MIPIA) { + if (dev_priv->vbt.dsi.port == DVO_PORT_MIPIA) { intel_encoder->crtc_mask = (1 << PIPE_A); intel_dsi->ports = (1 << PORT_A); } else if (dev_priv->vbt.dsi.port == DVO_PORT_MIPIC) { @@ -1060,6 +1056,9 @@ void intel_dsi_init(struct drm_device *dev) intel_dsi->ports = (1 << PORT_C); } + if (dev_priv->vbt.dsi.config->dual_link) + intel_dsi->ports = ((1 << PORT_A) | (1 << PORT_C)); + /* Create a DSI host (and a device) for each port. */ for_each_dsi_port(port, intel_dsi->ports) { struct intel_dsi_host *host; diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 51cbea8247fe..dcd336bcdfe7 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -1958,6 +1958,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, struct drm_device *dev = intel_encoder->base.dev; struct drm_i915_private *dev_priv = dev->dev_private; enum port port = intel_dig_port->port; + uint8_t alternate_ddc_pin; drm_connector_init(dev, connector, &intel_hdmi_connector_funcs, DRM_MODE_CONNECTOR_HDMIA); @@ -1991,6 +1992,26 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, intel_hdmi->ddc_bus = GMBUS_PIN_DPD; intel_encoder->hpd_pin = HPD_PORT_D; break; + case PORT_E: + /* On SKL PORT E doesn't have seperate GMBUS pin + * We rely on VBT to set a proper alternate GMBUS pin. */ + alternate_ddc_pin = + dev_priv->vbt.ddi_port_info[PORT_E].alternate_ddc_pin; + switch (alternate_ddc_pin) { + case DDC_PIN_B: + intel_hdmi->ddc_bus = GMBUS_PIN_DPB; + break; + case DDC_PIN_C: + intel_hdmi->ddc_bus = GMBUS_PIN_DPC; + break; + case DDC_PIN_D: + intel_hdmi->ddc_bus = GMBUS_PIN_DPD; + break; + default: + MISSING_CASE(alternate_ddc_pin); + } + intel_encoder->hpd_pin = HPD_PORT_E; + break; case PORT_A: intel_encoder->hpd_pin = HPD_PORT_A; /* Internal port only for eDP. */ diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index fff0c22682ee..ddbb7ed0a193 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -955,8 +955,6 @@ enum vlv_wm_level { VLV_WM_LEVEL_PM2, VLV_WM_LEVEL_PM5, VLV_WM_LEVEL_DDR_DVFS, - CHV_WM_NUM_LEVELS, - VLV_WM_NUM_LEVELS = 1, }; /* latency must be in 0.1us units. */ @@ -982,9 +980,13 @@ static void vlv_setup_wm_latency(struct drm_device *dev) /* all latencies in usec */ dev_priv->wm.pri_latency[VLV_WM_LEVEL_PM2] = 3; + dev_priv->wm.max_level = VLV_WM_LEVEL_PM2; + if (IS_CHERRYVIEW(dev_priv)) { dev_priv->wm.pri_latency[VLV_WM_LEVEL_PM5] = 12; dev_priv->wm.pri_latency[VLV_WM_LEVEL_DDR_DVFS] = 33; + + dev_priv->wm.max_level = VLV_WM_LEVEL_DDR_DVFS; } } @@ -1137,10 +1139,7 @@ static void vlv_compute_wm(struct intel_crtc *crtc) memset(wm_state, 0, sizeof(*wm_state)); wm_state->cxsr = crtc->pipe != PIPE_C && crtc->wm.cxsr_allowed; - if (IS_CHERRYVIEW(dev)) - wm_state->num_levels = CHV_WM_NUM_LEVELS; - else - wm_state->num_levels = VLV_WM_NUM_LEVELS; + wm_state->num_levels = to_i915(dev)->wm.max_level + 1; wm_state->num_active_planes = 0; @@ -1220,7 +1219,7 @@ static void vlv_compute_wm(struct intel_crtc *crtc) } /* clear any (partially) filled invalid levels */ - for (level = wm_state->num_levels; level < CHV_WM_NUM_LEVELS; level++) { + for (level = wm_state->num_levels; level < to_i915(dev)->wm.max_level + 1; level++) { memset(&wm_state->wm[level], 0, sizeof(wm_state->wm[level])); memset(&wm_state->sr[level], 0, sizeof(wm_state->sr[level])); } @@ -1324,10 +1323,7 @@ static void vlv_merge_wm(struct drm_device *dev, struct intel_crtc *crtc; int num_active_crtcs = 0; - if (IS_CHERRYVIEW(dev)) - wm->level = VLV_WM_LEVEL_DDR_DVFS; - else - wm->level = VLV_WM_LEVEL_PM2; + wm->level = to_i915(dev)->wm.max_level; wm->cxsr = true; for_each_intel_crtc(dev, crtc) { @@ -4083,9 +4079,29 @@ void vlv_wm_get_hw_state(struct drm_device *dev) if (val & DSP_MAXFIFO_PM5_ENABLE) wm->level = VLV_WM_LEVEL_PM5; + /* + * If DDR DVFS is disabled in the BIOS, Punit + * will never ack the request. So if that happens + * assume we don't have to enable/disable DDR DVFS + * dynamically. To test that just set the REQ_ACK + * bit to poke the Punit, but don't change the + * HIGH/LOW bits so that we don't actually change + * the current state. + */ val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2); - if ((val & FORCE_DDR_HIGH_FREQ) == 0) - wm->level = VLV_WM_LEVEL_DDR_DVFS; + val |= FORCE_DDR_FREQ_REQ_ACK; + vlv_punit_write(dev_priv, PUNIT_REG_DDR_SETUP2, val); + + if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2) & + FORCE_DDR_FREQ_REQ_ACK) == 0, 3)) { + DRM_DEBUG_KMS("Punit not acking DDR DVFS request, " + "assuming DDR DVFS is disabled\n"); + dev_priv->wm.max_level = VLV_WM_LEVEL_PM5; + } else { + val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2); + if ((val & FORCE_DDR_HIGH_FREQ) == 0) + wm->level = VLV_WM_LEVEL_DDR_DVFS; + } mutex_unlock(&dev_priv->rps.hw_lock); } diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index 821644d1b544..af7fdb3bd663 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -297,6 +297,7 @@ static void hsw_set_power_well(struct drm_i915_private *dev_priv, BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) | \ BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) | \ BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) | \ + BIT(POWER_DOMAIN_PORT_DDI_E_2_LANES) | \ BIT(POWER_DOMAIN_AUX_B) | \ BIT(POWER_DOMAIN_AUX_C) | \ BIT(POWER_DOMAIN_AUX_D) | \ @@ -316,6 +317,7 @@ static void hsw_set_power_well(struct drm_i915_private *dev_priv, #define SKL_DISPLAY_DDI_A_E_POWER_DOMAINS ( \ BIT(POWER_DOMAIN_PORT_DDI_A_2_LANES) | \ BIT(POWER_DOMAIN_PORT_DDI_A_4_LANES) | \ + BIT(POWER_DOMAIN_PORT_DDI_E_2_LANES) | \ BIT(POWER_DOMAIN_INIT)) #define SKL_DISPLAY_DDI_B_POWER_DOMAINS ( \ BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) | \ diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c index a8dbb3ef4e3c..7c6225c84ba6 100644 --- a/drivers/gpu/drm/qxl/qxl_display.c +++ b/drivers/gpu/drm/qxl/qxl_display.c @@ -160,9 +160,35 @@ static int qxl_add_monitors_config_modes(struct drm_connector *connector, *pwidth = head->width; *pheight = head->height; drm_mode_probed_add(connector, mode); + /* remember the last custom size for mode validation */ + qdev->monitors_config_width = mode->hdisplay; + qdev->monitors_config_height = mode->vdisplay; return 1; } +static struct mode_size { + int w; + int h; +} common_modes[] = { + { 640, 480}, + { 720, 480}, + { 800, 600}, + { 848, 480}, + {1024, 768}, + {1152, 768}, + {1280, 720}, + {1280, 800}, + {1280, 854}, + {1280, 960}, + {1280, 1024}, + {1440, 900}, + {1400, 1050}, + {1680, 1050}, + {1600, 1200}, + {1920, 1080}, + {1920, 1200} +}; + static int qxl_add_common_modes(struct drm_connector *connector, unsigned pwidth, unsigned pheight) @@ -170,29 +196,6 @@ static int qxl_add_common_modes(struct drm_connector *connector, struct drm_device *dev = connector->dev; struct drm_display_mode *mode = NULL; int i; - struct mode_size { - int w; - int h; - } common_modes[] = { - { 640, 480}, - { 720, 480}, - { 800, 600}, - { 848, 480}, - {1024, 768}, - {1152, 768}, - {1280, 720}, - {1280, 800}, - {1280, 854}, - {1280, 960}, - {1280, 1024}, - {1440, 900}, - {1400, 1050}, - {1680, 1050}, - {1600, 1200}, - {1920, 1080}, - {1920, 1200} - }; - 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); @@ -823,11 +826,22 @@ static int qxl_conn_get_modes(struct drm_connector *connector) static int qxl_conn_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { + struct drm_device *ddev = connector->dev; + struct qxl_device *qdev = ddev->dev_private; + int i; + /* TODO: is this called for user defined modes? (xrandr --add-mode) * TODO: check that the mode fits in the framebuffer */ - DRM_DEBUG("%s: %dx%d status=%d\n", mode->name, mode->hdisplay, - mode->vdisplay, mode->status); - return MODE_OK; + + if(qdev->monitors_config_width == mode->hdisplay && + qdev->monitors_config_height == mode->vdisplay) + return MODE_OK; + + for (i = 0; i < ARRAY_SIZE(common_modes); i++) { + if (common_modes[i].w == mode->hdisplay && common_modes[i].h == mode->vdisplay) + return MODE_OK; + } + return MODE_BAD; } static struct drm_encoder *qxl_best_encoder(struct drm_connector *connector) diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h index d8549690801d..01a86948eb8c 100644 --- a/drivers/gpu/drm/qxl/qxl_drv.h +++ b/drivers/gpu/drm/qxl/qxl_drv.h @@ -325,6 +325,8 @@ struct qxl_device { struct work_struct fb_work; struct drm_property *hotplug_mode_update_property; + int monitors_config_width; + int monitors_config_height; }; /* forward declaration for QXL_INFO_IO */ diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c index f81e0d7d0232..9cd49c584263 100644 --- a/drivers/gpu/drm/radeon/atombios_dp.c +++ b/drivers/gpu/drm/radeon/atombios_dp.c @@ -171,8 +171,9 @@ radeon_dp_aux_transfer_atom(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) return -E2BIG; tx_buf[0] = msg->address & 0xff; - tx_buf[1] = msg->address >> 8; - tx_buf[2] = msg->request << 4; + tx_buf[1] = (msg->address >> 8) & 0xff; + tx_buf[2] = (msg->request << 4) | + ((msg->address >> 16) & 0xf); tx_buf[3] = msg->size ? (msg->size - 1) : 0; switch (msg->request & ~DP_AUX_I2C_MOT) { diff --git a/drivers/gpu/drm/radeon/radeon_audio.c b/drivers/gpu/drm/radeon/radeon_audio.c index fbc8d88d6e5d..2c02e99b5f95 100644 --- a/drivers/gpu/drm/radeon/radeon_audio.c +++ b/drivers/gpu/drm/radeon/radeon_audio.c @@ -522,13 +522,15 @@ static int radeon_audio_set_avi_packet(struct drm_encoder *encoder, return err; } - if (drm_rgb_quant_range_selectable(radeon_connector_edid(connector))) { - if (radeon_encoder->output_csc == RADEON_OUTPUT_CSC_TVRGB) - frame.quantization_range = HDMI_QUANTIZATION_RANGE_LIMITED; - else - frame.quantization_range = HDMI_QUANTIZATION_RANGE_FULL; - } else { - frame.quantization_range = HDMI_QUANTIZATION_RANGE_DEFAULT; + if (radeon_encoder->output_csc != RADEON_OUTPUT_CSC_BYPASS) { + if (drm_rgb_quant_range_selectable(radeon_connector_edid(connector))) { + if (radeon_encoder->output_csc == RADEON_OUTPUT_CSC_TVRGB) + frame.quantization_range = HDMI_QUANTIZATION_RANGE_LIMITED; + else + frame.quantization_range = HDMI_QUANTIZATION_RANGE_FULL; + } else { + frame.quantization_range = HDMI_QUANTIZATION_RANGE_DEFAULT; + } } err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer)); diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c index c097d3a82bda..a9b01bcf7d0a 100644 --- a/drivers/gpu/drm/radeon/radeon_combios.c +++ b/drivers/gpu/drm/radeon/radeon_combios.c @@ -3387,6 +3387,14 @@ void radeon_combios_asic_init(struct drm_device *dev) rdev->pdev->subsystem_device == 0x30ae) return; + /* quirk for rs4xx HP Compaq dc5750 Small Form Factor to make it resume + * - it hangs on resume inside the dynclk 1 table. + */ + if (rdev->family == CHIP_RS480 && + rdev->pdev->subsystem_vendor == 0x103c && + rdev->pdev->subsystem_device == 0x280a) + return; + /* DYN CLK 1 */ table = combios_get_table_offset(dev, COMBIOS_DYN_CLK_1_TABLE); if (table) diff --git a/drivers/gpu/drm/radeon/radeon_dp_auxch.c b/drivers/gpu/drm/radeon/radeon_dp_auxch.c index fcbd60bb0349..3b0c229d7dcd 100644 --- a/drivers/gpu/drm/radeon/radeon_dp_auxch.c +++ b/drivers/gpu/drm/radeon/radeon_dp_auxch.c @@ -116,8 +116,8 @@ radeon_dp_aux_transfer_native(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg AUX_SW_WR_BYTES(bytes)); /* write the data header into the registers */ - /* request, addres, msg size */ - byte = (msg->request << 4); + /* request, address, msg size */ + byte = (msg->request << 4) | ((msg->address >> 16) & 0xf); WREG32(AUX_SW_DATA + aux_offset[instance], AUX_SW_DATA_MASK(byte) | AUX_SW_AUTOINCREMENT_DISABLE); diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index 34b78e736532..5d8ae5e49c44 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -50,6 +50,8 @@ #define VOP_WIN_SET(x, win, name, v) \ REG_SET(x, win->base, win->phy->name, v, RELAXED) +#define VOP_SCL_SET(x, win, name, v) \ + REG_SET(x, win->base, win->phy->scl->name, v, RELAXED) #define VOP_CTRL_SET(x, name, v) \ REG_SET(x, 0, (x)->data->ctrl->name, v, NORMAL) @@ -164,7 +166,37 @@ struct vop_ctrl { struct vop_reg vpost_st_end; }; +struct vop_scl_regs { + struct vop_reg cbcr_vsd_mode; + struct vop_reg cbcr_vsu_mode; + struct vop_reg cbcr_hsd_mode; + struct vop_reg cbcr_ver_scl_mode; + struct vop_reg cbcr_hor_scl_mode; + struct vop_reg yrgb_vsd_mode; + struct vop_reg yrgb_vsu_mode; + struct vop_reg yrgb_hsd_mode; + struct vop_reg yrgb_ver_scl_mode; + struct vop_reg yrgb_hor_scl_mode; + struct vop_reg line_load_mode; + struct vop_reg cbcr_axi_gather_num; + struct vop_reg yrgb_axi_gather_num; + struct vop_reg vsd_cbcr_gt2; + struct vop_reg vsd_cbcr_gt4; + struct vop_reg vsd_yrgb_gt2; + struct vop_reg vsd_yrgb_gt4; + struct vop_reg bic_coe_sel; + struct vop_reg cbcr_axi_gather_en; + struct vop_reg yrgb_axi_gather_en; + + struct vop_reg lb_mode; + struct vop_reg scale_yrgb_x; + struct vop_reg scale_yrgb_y; + struct vop_reg scale_cbcr_x; + struct vop_reg scale_cbcr_y; +}; + struct vop_win_phy { + const struct vop_scl_regs *scl; const uint32_t *data_formats; uint32_t nformats; @@ -222,7 +254,36 @@ static const uint32_t formats_234[] = { DRM_FORMAT_BGR565, }; +static const struct vop_scl_regs win_full_scl = { + .cbcr_vsd_mode = VOP_REG(WIN0_CTRL1, 0x1, 31), + .cbcr_vsu_mode = VOP_REG(WIN0_CTRL1, 0x1, 30), + .cbcr_hsd_mode = VOP_REG(WIN0_CTRL1, 0x3, 28), + .cbcr_ver_scl_mode = VOP_REG(WIN0_CTRL1, 0x3, 26), + .cbcr_hor_scl_mode = VOP_REG(WIN0_CTRL1, 0x3, 24), + .yrgb_vsd_mode = VOP_REG(WIN0_CTRL1, 0x1, 23), + .yrgb_vsu_mode = VOP_REG(WIN0_CTRL1, 0x1, 22), + .yrgb_hsd_mode = VOP_REG(WIN0_CTRL1, 0x3, 20), + .yrgb_ver_scl_mode = VOP_REG(WIN0_CTRL1, 0x3, 18), + .yrgb_hor_scl_mode = VOP_REG(WIN0_CTRL1, 0x3, 16), + .line_load_mode = VOP_REG(WIN0_CTRL1, 0x1, 15), + .cbcr_axi_gather_num = VOP_REG(WIN0_CTRL1, 0x7, 12), + .yrgb_axi_gather_num = VOP_REG(WIN0_CTRL1, 0xf, 8), + .vsd_cbcr_gt2 = VOP_REG(WIN0_CTRL1, 0x1, 7), + .vsd_cbcr_gt4 = VOP_REG(WIN0_CTRL1, 0x1, 6), + .vsd_yrgb_gt2 = VOP_REG(WIN0_CTRL1, 0x1, 5), + .vsd_yrgb_gt4 = VOP_REG(WIN0_CTRL1, 0x1, 4), + .bic_coe_sel = VOP_REG(WIN0_CTRL1, 0x3, 2), + .cbcr_axi_gather_en = VOP_REG(WIN0_CTRL1, 0x1, 1), + .yrgb_axi_gather_en = VOP_REG(WIN0_CTRL1, 0x1, 0), + .lb_mode = VOP_REG(WIN0_CTRL0, 0x7, 5), + .scale_yrgb_x = VOP_REG(WIN0_SCL_FACTOR_YRGB, 0xffff, 0x0), + .scale_yrgb_y = VOP_REG(WIN0_SCL_FACTOR_YRGB, 0xffff, 16), + .scale_cbcr_x = VOP_REG(WIN0_SCL_FACTOR_CBR, 0xffff, 0x0), + .scale_cbcr_y = VOP_REG(WIN0_SCL_FACTOR_CBR, 0xffff, 16), +}; + static const struct vop_win_phy win01_data = { + .scl = &win_full_scl, .data_formats = formats_01, .nformats = ARRAY_SIZE(formats_01), .enable = VOP_REG(WIN0_CTRL0, 0x1, 0), @@ -279,6 +340,12 @@ static const struct vop_reg_data vop_init_reg_table[] = { {DSP_CTRL0, 0x00000000}, {WIN0_CTRL0, 0x00000080}, {WIN1_CTRL0, 0x00000080}, + /* TODO: Win2/3 support multiple area function, but we haven't found + * a suitable way to use it yet, so let's just use them as other windows + * with only area 0 enabled. + */ + {WIN2_CTRL0, 0x00000010}, + {WIN3_CTRL0, 0x00000010}, }; /* @@ -393,6 +460,18 @@ static enum vop_data_format vop_convert_format(uint32_t format) } } +static bool is_yuv_support(uint32_t format) +{ + switch (format) { + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV16: + case DRM_FORMAT_NV24: + return true; + default: + return false; + } +} + static bool is_alpha_support(uint32_t format) { switch (format) { @@ -404,6 +483,126 @@ static bool is_alpha_support(uint32_t format) } } +static uint16_t scl_vop_cal_scale(enum scale_mode mode, uint32_t src, + uint32_t dst, bool is_horizontal, + int vsu_mode, int *vskiplines) +{ + uint16_t val = 1 << SCL_FT_DEFAULT_FIXPOINT_SHIFT; + + if (is_horizontal) { + if (mode == SCALE_UP) + val = GET_SCL_FT_BIC(src, dst); + else if (mode == SCALE_DOWN) + val = GET_SCL_FT_BILI_DN(src, dst); + } else { + if (mode == SCALE_UP) { + if (vsu_mode == SCALE_UP_BIL) + val = GET_SCL_FT_BILI_UP(src, dst); + else + val = GET_SCL_FT_BIC(src, dst); + } else if (mode == SCALE_DOWN) { + if (vskiplines) { + *vskiplines = scl_get_vskiplines(src, dst); + val = scl_get_bili_dn_vskip(src, dst, + *vskiplines); + } else { + val = GET_SCL_FT_BILI_DN(src, dst); + } + } + } + + return val; +} + +static void scl_vop_cal_scl_fac(struct vop *vop, const struct vop_win_data *win, + uint32_t src_w, uint32_t src_h, uint32_t dst_w, + uint32_t dst_h, uint32_t pixel_format) +{ + uint16_t yrgb_hor_scl_mode, yrgb_ver_scl_mode; + uint16_t cbcr_hor_scl_mode = SCALE_NONE; + uint16_t cbcr_ver_scl_mode = SCALE_NONE; + int hsub = drm_format_horz_chroma_subsampling(pixel_format); + int vsub = drm_format_vert_chroma_subsampling(pixel_format); + bool is_yuv = is_yuv_support(pixel_format); + uint16_t cbcr_src_w = src_w / hsub; + uint16_t cbcr_src_h = src_h / vsub; + uint16_t vsu_mode; + uint16_t lb_mode; + uint32_t val; + int vskiplines; + + if (dst_w > 3840) { + DRM_ERROR("Maximum destination width (3840) exceeded\n"); + return; + } + + yrgb_hor_scl_mode = scl_get_scl_mode(src_w, dst_w); + yrgb_ver_scl_mode = scl_get_scl_mode(src_h, dst_h); + + if (is_yuv) { + cbcr_hor_scl_mode = scl_get_scl_mode(cbcr_src_w, dst_w); + cbcr_ver_scl_mode = scl_get_scl_mode(cbcr_src_h, dst_h); + if (cbcr_hor_scl_mode == SCALE_DOWN) + lb_mode = scl_vop_cal_lb_mode(dst_w, true); + else + lb_mode = scl_vop_cal_lb_mode(cbcr_src_w, true); + } else { + if (yrgb_hor_scl_mode == SCALE_DOWN) + lb_mode = scl_vop_cal_lb_mode(dst_w, false); + else + lb_mode = scl_vop_cal_lb_mode(src_w, false); + } + + VOP_SCL_SET(vop, win, lb_mode, lb_mode); + if (lb_mode == LB_RGB_3840X2) { + if (yrgb_ver_scl_mode != SCALE_NONE) { + DRM_ERROR("ERROR : not allow yrgb ver scale\n"); + return; + } + if (cbcr_ver_scl_mode != SCALE_NONE) { + DRM_ERROR("ERROR : not allow cbcr ver scale\n"); + return; + } + vsu_mode = SCALE_UP_BIL; + } else if (lb_mode == LB_RGB_2560X4) { + vsu_mode = SCALE_UP_BIL; + } else { + vsu_mode = SCALE_UP_BIC; + } + + val = scl_vop_cal_scale(yrgb_hor_scl_mode, src_w, dst_w, + true, 0, NULL); + VOP_SCL_SET(vop, win, scale_yrgb_x, val); + val = scl_vop_cal_scale(yrgb_ver_scl_mode, src_h, dst_h, + false, vsu_mode, &vskiplines); + VOP_SCL_SET(vop, win, scale_yrgb_y, val); + + VOP_SCL_SET(vop, win, vsd_yrgb_gt4, vskiplines == 4); + VOP_SCL_SET(vop, win, vsd_yrgb_gt2, vskiplines == 2); + + VOP_SCL_SET(vop, win, yrgb_hor_scl_mode, yrgb_hor_scl_mode); + VOP_SCL_SET(vop, win, yrgb_ver_scl_mode, yrgb_ver_scl_mode); + VOP_SCL_SET(vop, win, yrgb_hsd_mode, SCALE_DOWN_BIL); + VOP_SCL_SET(vop, win, yrgb_vsd_mode, SCALE_DOWN_BIL); + VOP_SCL_SET(vop, win, yrgb_vsu_mode, vsu_mode); + if (is_yuv) { + val = scl_vop_cal_scale(cbcr_hor_scl_mode, cbcr_src_w, + dst_w, true, 0, NULL); + VOP_SCL_SET(vop, win, scale_cbcr_x, val); + val = scl_vop_cal_scale(cbcr_ver_scl_mode, cbcr_src_h, + dst_h, false, vsu_mode, &vskiplines); + VOP_SCL_SET(vop, win, scale_cbcr_y, val); + + VOP_SCL_SET(vop, win, vsd_cbcr_gt4, vskiplines == 4); + VOP_SCL_SET(vop, win, vsd_cbcr_gt2, vskiplines == 2); + VOP_SCL_SET(vop, win, cbcr_hor_scl_mode, cbcr_hor_scl_mode); + VOP_SCL_SET(vop, win, cbcr_ver_scl_mode, cbcr_ver_scl_mode); + VOP_SCL_SET(vop, win, cbcr_hsd_mode, SCALE_DOWN_BIL); + VOP_SCL_SET(vop, win, cbcr_vsd_mode, SCALE_DOWN_BIL); + VOP_SCL_SET(vop, win, cbcr_vsu_mode, vsu_mode); + } +} + static void vop_dsp_hold_valid_irq_enable(struct vop *vop) { unsigned long flags; @@ -478,6 +677,7 @@ static void vop_enable(struct drm_crtc *crtc) goto err_disable_aclk; } + memcpy(vop->regs, vop->regsbak, vop->len); /* * At here, vop clock & iommu is enable, R/W vop regs would be safe. */ @@ -598,17 +798,22 @@ static int vop_update_plane_event(struct drm_plane *plane, struct vop *vop = to_vop(crtc); struct drm_gem_object *obj; struct rockchip_gem_object *rk_obj; + struct drm_gem_object *uv_obj; + struct rockchip_gem_object *rk_uv_obj; unsigned long offset; unsigned int actual_w; unsigned int actual_h; unsigned int dsp_stx; unsigned int dsp_sty; unsigned int y_vir_stride; + unsigned int uv_vir_stride = 0; dma_addr_t yrgb_mst; + dma_addr_t uv_mst = 0; enum vop_data_format format; uint32_t val; bool is_alpha; bool rb_swap; + bool is_yuv; bool visible; int ret; struct drm_rect dest = { @@ -629,11 +834,15 @@ static int vop_update_plane_event(struct drm_plane *plane, .y2 = crtc->mode.vdisplay, }; bool can_position = plane->type != DRM_PLANE_TYPE_PRIMARY; + int min_scale = win->phy->scl ? FRAC_16_16(1, 8) : + DRM_PLANE_HELPER_NO_SCALING; + int max_scale = win->phy->scl ? FRAC_16_16(8, 1) : + DRM_PLANE_HELPER_NO_SCALING; ret = drm_plane_helper_check_update(plane, crtc, fb, &src, &dest, &clip, - DRM_PLANE_HELPER_NO_SCALING, - DRM_PLANE_HELPER_NO_SCALING, + min_scale, + max_scale, can_position, false, &visible); if (ret) return ret; @@ -643,6 +852,8 @@ static int vop_update_plane_event(struct drm_plane *plane, is_alpha = is_alpha_support(fb->pixel_format); rb_swap = has_rb_swapped(fb->pixel_format); + is_yuv = is_yuv_support(fb->pixel_format); + format = vop_convert_format(fb->pixel_format); if (format < 0) return format; @@ -655,19 +866,46 @@ static int vop_update_plane_event(struct drm_plane *plane, rk_obj = to_rockchip_obj(obj); + if (is_yuv) { + /* + * Src.x1 can be odd when do clip, but yuv plane start point + * need align with 2 pixel. + */ + val = (src.x1 >> 16) % 2; + src.x1 += val << 16; + src.x2 += val << 16; + } + actual_w = (src.x2 - src.x1) >> 16; actual_h = (src.y2 - src.y1) >> 16; - crtc_x = max(0, crtc_x); - crtc_y = max(0, crtc_y); - dsp_stx = crtc_x + crtc->mode.htotal - crtc->mode.hsync_start; - dsp_sty = crtc_y + crtc->mode.vtotal - crtc->mode.vsync_start; + dsp_stx = dest.x1 + crtc->mode.htotal - crtc->mode.hsync_start; + dsp_sty = dest.y1 + crtc->mode.vtotal - crtc->mode.vsync_start; - offset = (src.x1 >> 16) * (fb->bits_per_pixel >> 3); + offset = (src.x1 >> 16) * drm_format_plane_cpp(fb->pixel_format, 0); offset += (src.y1 >> 16) * fb->pitches[0]; - yrgb_mst = rk_obj->dma_addr + offset; - y_vir_stride = fb->pitches[0] / (fb->bits_per_pixel >> 3); + yrgb_mst = rk_obj->dma_addr + offset + fb->offsets[0]; + y_vir_stride = fb->pitches[0] >> 2; + + if (is_yuv) { + int hsub = drm_format_horz_chroma_subsampling(fb->pixel_format); + int vsub = drm_format_vert_chroma_subsampling(fb->pixel_format); + int bpp = drm_format_plane_cpp(fb->pixel_format, 1); + + uv_obj = rockchip_fb_get_gem_obj(fb, 1); + if (!uv_obj) { + DRM_ERROR("fail to get uv object from framebuffer\n"); + return -EINVAL; + } + rk_uv_obj = to_rockchip_obj(uv_obj); + uv_vir_stride = fb->pitches[1] >> 2; + + offset = (src.x1 >> 16) * bpp / hsub; + offset += (src.y1 >> 16) * fb->pitches[1] / vsub; + + uv_mst = rk_uv_obj->dma_addr + offset + fb->offsets[1]; + } /* * If this plane update changes the plane's framebuffer, (or more @@ -704,9 +942,22 @@ static int vop_update_plane_event(struct drm_plane *plane, VOP_WIN_SET(vop, win, format, format); VOP_WIN_SET(vop, win, yrgb_vir, y_vir_stride); VOP_WIN_SET(vop, win, yrgb_mst, yrgb_mst); + if (is_yuv) { + VOP_WIN_SET(vop, win, uv_vir, uv_vir_stride); + VOP_WIN_SET(vop, win, uv_mst, uv_mst); + } + + if (win->phy->scl) + scl_vop_cal_scl_fac(vop, win, actual_w, actual_h, + dest.x2 - dest.x1, dest.y2 - dest.y1, + fb->pixel_format); + val = (actual_h - 1) << 16; val |= (actual_w - 1) & 0xffff; VOP_WIN_SET(vop, win, act_info, val); + + val = (dest.y2 - dest.y1 - 1) << 16; + val |= (dest.x2 - dest.x1 - 1) & 0xffff; VOP_WIN_SET(vop, win, dsp_info, val); val = (dsp_sty - 1) << 16; val |= (dsp_stx - 1) & 0xffff; diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h index 63e9b3a084c5..a2d4ddb896fa 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h @@ -198,4 +198,92 @@ enum factor_mode { ALPHA_SRC_GLOBAL, }; +enum scale_mode { + SCALE_NONE = 0x0, + SCALE_UP = 0x1, + SCALE_DOWN = 0x2 +}; + +enum lb_mode { + LB_YUV_3840X5 = 0x0, + LB_YUV_2560X8 = 0x1, + LB_RGB_3840X2 = 0x2, + LB_RGB_2560X4 = 0x3, + LB_RGB_1920X5 = 0x4, + LB_RGB_1280X8 = 0x5 +}; + +enum sacle_up_mode { + SCALE_UP_BIL = 0x0, + SCALE_UP_BIC = 0x1 +}; + +enum scale_down_mode { + SCALE_DOWN_BIL = 0x0, + SCALE_DOWN_AVG = 0x1 +}; + +#define FRAC_16_16(mult, div) (((mult) << 16) / (div)) +#define SCL_FT_DEFAULT_FIXPOINT_SHIFT 12 +#define SCL_MAX_VSKIPLINES 4 +#define MIN_SCL_FT_AFTER_VSKIP 1 + +static inline uint16_t scl_cal_scale(int src, int dst, int shift) +{ + return ((src * 2 - 3) << (shift - 1)) / (dst - 1); +} + +#define GET_SCL_FT_BILI_DN(src, dst) scl_cal_scale(src, dst, 12) +#define GET_SCL_FT_BILI_UP(src, dst) scl_cal_scale(src, dst, 16) +#define GET_SCL_FT_BIC(src, dst) scl_cal_scale(src, dst, 16) + +static inline uint16_t scl_get_bili_dn_vskip(int src_h, int dst_h, + int vskiplines) +{ + int act_height; + + act_height = (src_h + vskiplines - 1) / vskiplines; + + return GET_SCL_FT_BILI_DN(act_height, dst_h); +} + +static inline enum scale_mode scl_get_scl_mode(int src, int dst) +{ + if (src < dst) + return SCALE_UP; + else if (src > dst) + return SCALE_DOWN; + + return SCALE_NONE; +} + +static inline int scl_get_vskiplines(uint32_t srch, uint32_t dsth) +{ + uint32_t vskiplines; + + for (vskiplines = SCL_MAX_VSKIPLINES; vskiplines > 1; vskiplines /= 2) + if (srch >= vskiplines * dsth * MIN_SCL_FT_AFTER_VSKIP) + break; + + return vskiplines; +} + +static inline int scl_vop_cal_lb_mode(int width, bool is_yuv) +{ + int lb_mode; + + if (width > 2560) + lb_mode = LB_RGB_3840X2; + else if (width > 1920) + lb_mode = LB_RGB_2560X4; + else if (!is_yuv) + lb_mode = LB_RGB_1920X5; + else if (width > 1280) + lb_mode = LB_YUV_3840X5; + else + lb_mode = LB_YUV_2560X8; + + return lb_mode; +} + #endif /* _ROCKCHIP_DRM_VOP_H */ diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 03854d606d58..e13b20bd9908 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -1052,10 +1052,15 @@ static struct vmw_master *vmw_master_check(struct drm_device *dev, } /* - * Check if we were previously master, but now dropped. + * Check if we were previously master, but now dropped. In that + * case, allow at least render node functionality. */ if (vmw_fp->locked_master) { mutex_unlock(&dev->master_mutex); + + if (flags & DRM_RENDER_ALLOW) + return NULL; + DRM_ERROR("Dropped master trying to access ioctl that " "requires authentication.\n"); return ERR_PTR(-EACCES); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c index 5b8595b78429..3361769842f4 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c @@ -911,6 +911,12 @@ vmw_surface_handle_reference(struct vmw_private *dev_priv, "surface reference.\n"); return -EACCES; } + if (ACCESS_ONCE(vmw_fpriv(file_priv)->locked_master)) { + DRM_ERROR("Locked master refused legacy " + "surface reference.\n"); + return -EACCES; + } + handle = u_handle; } |