diff options
Diffstat (limited to 'drivers/gpu/drm')
113 files changed, 867 insertions, 554 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 77e263660288..41db030ddc4e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -141,11 +141,31 @@ static void amdgpu_amdkfd_reset_work(struct work_struct *work) static const struct drm_client_funcs kfd_client_funcs = { .unregister = drm_client_release, }; + +int amdgpu_amdkfd_drm_client_create(struct amdgpu_device *adev) +{ + int ret; + + if (!adev->kfd.init_complete) + return 0; + + ret = drm_client_init(&adev->ddev, &adev->kfd.client, "kfd", + &kfd_client_funcs); + if (ret) { + dev_err(adev->dev, "Failed to init DRM client: %d\n", + ret); + return ret; + } + + drm_client_register(&adev->kfd.client); + + return 0; +} + void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) { int i; int last_valid_bit; - int ret; amdgpu_amdkfd_gpuvm_init_mem_limits(); @@ -164,12 +184,6 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) .enable_mes = adev->enable_mes, }; - ret = drm_client_init(&adev->ddev, &adev->kfd.client, "kfd", &kfd_client_funcs); - if (ret) { - dev_err(adev->dev, "Failed to init DRM client: %d\n", ret); - return; - } - /* this is going to have a few of the MSBs set that we need to * clear */ @@ -208,10 +222,6 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) adev->kfd.init_complete = kgd2kfd_device_init(adev->kfd.dev, &gpu_resources); - if (adev->kfd.init_complete) - drm_client_register(&adev->kfd.client); - else - drm_client_release(&adev->kfd.client); amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index f262b9d89541..27c61c535e29 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -182,6 +182,8 @@ int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev, struct amdgpu_amdkfd_fence *amdgpu_amdkfd_fence_create(u64 context, struct mm_struct *mm, struct svm_range_bo *svm_bo); + +int amdgpu_amdkfd_drm_client_create(struct amdgpu_device *adev); #if defined(CONFIG_DEBUG_FS) int kfd_debugfs_kfd_mem_limits(struct seq_file *m, void *data); #endif @@ -301,7 +303,7 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv); int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu( struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv); -void amdgpu_amdkfd_gpuvm_dmaunmap_mem(struct kgd_mem *mem, void *drm_priv); +int amdgpu_amdkfd_gpuvm_dmaunmap_mem(struct kgd_mem *mem, void *drm_priv); int amdgpu_amdkfd_gpuvm_sync_memory( struct amdgpu_device *adev, struct kgd_mem *mem, bool intr); int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_mem *mem, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c index 899e31e3a5e8..3a3f3ce09f00 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c @@ -290,7 +290,7 @@ static int suspend_resume_compute_scheduler(struct amdgpu_device *adev, bool sus for (i = 0; i < adev->gfx.num_compute_rings; i++) { struct amdgpu_ring *ring = &adev->gfx.compute_ring[i]; - if (!(ring && drm_sched_wqueue_ready(&ring->sched))) + if (!amdgpu_ring_sched_ready(ring)) continue; /* stop secheduler and drain ring. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index f183d7faeeec..231fd927dcfb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -2085,21 +2085,35 @@ out: return ret; } -void amdgpu_amdkfd_gpuvm_dmaunmap_mem(struct kgd_mem *mem, void *drm_priv) +int amdgpu_amdkfd_gpuvm_dmaunmap_mem(struct kgd_mem *mem, void *drm_priv) { struct kfd_mem_attachment *entry; struct amdgpu_vm *vm; + int ret; vm = drm_priv_to_vm(drm_priv); mutex_lock(&mem->lock); + ret = amdgpu_bo_reserve(mem->bo, true); + if (ret) + goto out; + list_for_each_entry(entry, &mem->attachments, list) { - if (entry->bo_va->base.vm == vm) - kfd_mem_dmaunmap_attachment(mem, entry); + if (entry->bo_va->base.vm != vm) + continue; + if (entry->bo_va->base.bo->tbo.ttm && + !entry->bo_va->base.bo->tbo.ttm->sg) + continue; + + kfd_mem_dmaunmap_attachment(mem, entry); } + amdgpu_bo_unreserve(mem->bo); +out: mutex_unlock(&mem->lock); + + return ret; } int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu( diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index e485dd3357c6..1afbb2e932c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -1678,7 +1678,7 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused) for (i = 0; i < AMDGPU_MAX_RINGS; i++) { struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; drm_sched_wqueue_stop(&ring->sched); } @@ -1694,7 +1694,7 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused) for (i = 0; i < AMDGPU_MAX_RINGS; i++) { struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; drm_sched_wqueue_start(&ring->sched); } @@ -1916,8 +1916,8 @@ static int amdgpu_debugfs_ib_preempt(void *data, u64 val) ring = adev->rings[val]; - if (!ring || !ring->funcs->preempt_ib || - !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring) || + !ring->funcs->preempt_ib) return -EINVAL; /* the last preemption failed */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index b158d27d0a71..fdde7488d0ed 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4121,23 +4121,13 @@ int amdgpu_device_init(struct amdgpu_device *adev, } } } else { - switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) { - case IP_VERSION(13, 0, 0): - case IP_VERSION(13, 0, 7): - case IP_VERSION(13, 0, 10): - r = psp_gpu_reset(adev); - break; - default: - tmp = amdgpu_reset_method; - /* It should do a default reset when loading or reloading the driver, - * regardless of the module parameter reset_method. - */ - amdgpu_reset_method = AMD_RESET_METHOD_NONE; - r = amdgpu_asic_reset(adev); - amdgpu_reset_method = tmp; - break; - } - + tmp = amdgpu_reset_method; + /* It should do a default reset when loading or reloading the driver, + * regardless of the module parameter reset_method. + */ + amdgpu_reset_method = AMD_RESET_METHOD_NONE; + r = amdgpu_asic_reset(adev); + amdgpu_reset_method = tmp; if (r) { dev_err(adev->dev, "asic reset on init failed\n"); goto failed; @@ -5031,7 +5021,7 @@ bool amdgpu_device_has_job_running(struct amdgpu_device *adev) for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; spin_lock(&ring->sched.job_list_lock); @@ -5170,7 +5160,7 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; /* Clear job fence from fence drv to avoid force_completion @@ -5637,7 +5627,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = tmp_adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; drm_sched_stop(&ring->sched, job ? &job->base : NULL); @@ -5706,7 +5696,7 @@ skip_hw_reset: for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = tmp_adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; drm_sched_start(&ring->sched, true); @@ -6061,7 +6051,7 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; drm_sched_stop(&ring->sched, NULL); @@ -6189,7 +6179,7 @@ void amdgpu_pci_resume(struct pci_dev *pdev) for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !drm_sched_wqueue_ready(&ring->sched)) + if (!amdgpu_ring_sched_ready(ring)) continue; drm_sched_start(&ring->sched, true); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index cc69005f5b46..971acf01bea6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -2255,6 +2255,10 @@ retry_init: if (ret) goto err_pci; + ret = amdgpu_amdkfd_drm_client_create(adev); + if (ret) + goto err_pci; + /* * 1. don't init fbdev on hw without DCE * 2. don't init fbdev if there are no connectors diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index 73b8cca35bab..c623e23049d1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -121,6 +121,7 @@ int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev) struct amdgpu_bo_param bp; dma_addr_t dma_addr; struct page *p; + unsigned long x; int ret; if (adev->gart.bo != NULL) @@ -130,6 +131,10 @@ int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev) if (!p) return -ENOMEM; + /* assign pages to this device */ + for (x = 0; x < (1UL << order); x++) + p[x].mapping = adev->mman.bdev.dev_mapping; + /* If the hardware does not support UTCL2 snooping of the CPU caches * then set_memory_wc() could be used as a workaround to mark the pages * as write combine memory. @@ -223,6 +228,7 @@ void amdgpu_gart_table_ram_free(struct amdgpu_device *adev) unsigned int order = get_order(adev->gart.table_size); struct sg_table *sg = adev->gart.bo->tbo.sg; struct page *p; + unsigned long x; int ret; ret = amdgpu_bo_reserve(adev->gart.bo, false); @@ -234,6 +240,8 @@ void amdgpu_gart_table_ram_free(struct amdgpu_device *adev) sg_free_table(sg); kfree(sg); p = virt_to_page(adev->gart.ptr); + for (x = 0; x < (1UL << order); x++) + p[x].mapping = NULL; __free_pages(p, order); adev->gart.ptr = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 45424ebf9681..5505d646f43a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -635,6 +635,7 @@ int amdgpu_ring_test_helper(struct amdgpu_ring *ring) ring->name); ring->sched.ready = !r; + return r; } @@ -717,3 +718,14 @@ void amdgpu_ring_ib_on_emit_de(struct amdgpu_ring *ring) if (ring->is_sw_ring) amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_DE); } + +bool amdgpu_ring_sched_ready(struct amdgpu_ring *ring) +{ + if (!ring) + return false; + + if (ring->no_scheduler || !drm_sched_wqueue_ready(&ring->sched)) + return false; + + return true; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index bbb53720a018..fe1a61eb6e4c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -450,5 +450,5 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, int amdgpu_ib_pool_init(struct amdgpu_device *adev); void amdgpu_ib_pool_fini(struct amdgpu_device *adev); int amdgpu_ib_ring_tests(struct amdgpu_device *adev); - +bool amdgpu_ring_sched_ready(struct amdgpu_ring *ring); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c index 08916538a615..8db880244324 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c @@ -221,8 +221,23 @@ static struct attribute *amdgpu_vram_mgr_attributes[] = { NULL }; +static umode_t amdgpu_vram_attrs_is_visible(struct kobject *kobj, + struct attribute *attr, int i) +{ + struct device *dev = kobj_to_dev(kobj); + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + + if (attr == &dev_attr_mem_info_vram_vendor.attr && + !adev->gmc.vram_vendor) + return 0; + + return attr->mode; +} + const struct attribute_group amdgpu_vram_mgr_attr_group = { - .attrs = amdgpu_vram_mgr_attributes + .attrs = amdgpu_vram_mgr_attributes, + .is_visible = amdgpu_vram_attrs_is_visible }; /** diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c index 6f7c031dd197..f24e34dc33d1 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c @@ -204,6 +204,12 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev, tmp = RREG32(mmIH_RB_CNTL); tmp |= IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK; WREG32(mmIH_RB_CNTL, tmp); + + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp &= ~IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK; + WREG32(mmIH_RB_CNTL, tmp); } return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c index b8c47e0cf37a..c19681492efa 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c @@ -216,6 +216,11 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev, tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32(mmIH_RB_CNTL, tmp); + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32(mmIH_RB_CNTL, tmp); out: return (wptr & ih->ptr_mask); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index d63cab294883..dcdecb18b230 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -4027,8 +4027,6 @@ static int gfx_v10_0_init_microcode(struct amdgpu_device *adev) err = 0; adev->gfx.mec2_fw = NULL; } - amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2); - amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2_JT); gfx_v10_0_check_fw_write_wait(adev); out: @@ -6589,7 +6587,7 @@ static int gfx_v10_0_compute_mqd_init(struct amdgpu_device *adev, void *m, #ifdef __BIG_ENDIAN tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1); #endif - tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 0); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, prop->allow_tunneling); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 0ea0866c261f..4f3bfdc75b37 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -107,23 +107,6 @@ static const struct soc15_reg_golden golden_settings_gc_11_0_1[] = SOC15_REG_GOLDEN_VALUE(GC, 0, regTCP_CNTL2, 0xfcffffff, 0x0000000a) }; -static const struct soc15_reg_golden golden_settings_gc_11_5_0[] = { - SOC15_REG_GOLDEN_VALUE(GC, 0, regDB_DEBUG5, 0xffffffff, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x0c1807ff, 0x00000242), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGCR_GENERAL_CNTL, 0x1ff1ffff, 0x00000500), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2A_ADDR_MATCH_MASK, 0xffffffff, 0xfffffff3), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xfffffff3), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2C_CTRL, 0xffffffff, 0xf37fff3f), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2C_CTRL3, 0xfffffffb, 0x00f40188), - SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2C_CTRL4, 0xf0ffffff, 0x80009007), - SOC15_REG_GOLDEN_VALUE(GC, 0, regPA_CL_ENHANCE, 0xf1ffffff, 0x00880007), - SOC15_REG_GOLDEN_VALUE(GC, 0, regPC_CONFIG_CNTL_1, 0xffffffff, 0x00010000), - SOC15_REG_GOLDEN_VALUE(GC, 0, regTA_CNTL_AUX, 0xf7f7ffff, 0x01030000), - SOC15_REG_GOLDEN_VALUE(GC, 0, regTA_CNTL2, 0x007f0000, 0x00000000), - SOC15_REG_GOLDEN_VALUE(GC, 0, regTCP_CNTL2, 0xffcfffff, 0x0000200a), - SOC15_REG_GOLDEN_VALUE(GC, 0, regUTCL1_CTRL_2, 0xffffffff, 0x0000048f) -}; - #define DEFAULT_SH_MEM_CONFIG \ ((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \ (SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \ @@ -304,11 +287,6 @@ static void gfx_v11_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_11_0_1, (const u32)ARRAY_SIZE(golden_settings_gc_11_0_1)); break; - case IP_VERSION(11, 5, 0): - soc15_program_register_sequence(adev, - golden_settings_gc_11_5_0, - (const u32)ARRAY_SIZE(golden_settings_gc_11_5_0)); - break; default: break; } @@ -3846,7 +3824,7 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m, (order_base_2(prop->queue_size / 4) - 1)); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE, (order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1)); - tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 0); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, prop->allow_tunneling); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index 42e103d7077d..59d9215e5556 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -915,8 +915,8 @@ static int gmc_v6_0_hw_init(void *handle) if (amdgpu_emu_mode == 1) return amdgpu_gmc_vram_checking(adev); - else - return r; + + return 0; } static int gmc_v6_0_hw_fini(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index efc16e580f1e..45a2f8e031a2 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -1099,8 +1099,8 @@ static int gmc_v7_0_hw_init(void *handle) if (amdgpu_emu_mode == 1) return amdgpu_gmc_vram_checking(adev); - else - return r; + + return 0; } static int gmc_v7_0_hw_fini(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index ff4ae73d27ec..4422b27a3cc2 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1219,8 +1219,8 @@ static int gmc_v8_0_hw_init(void *handle) if (amdgpu_emu_mode == 1) return amdgpu_gmc_vram_checking(adev); - else - return r; + + return 0; } static int gmc_v8_0_hw_fini(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index f9039d64ff2d..40a00ea0009f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -1950,7 +1950,8 @@ static void gmc_v9_4_3_init_vram_info(struct amdgpu_device *adev) static const u32 regBIF_BIOS_SCRATCH_4 = 0x50; u32 vram_info; - if (!amdgpu_sriov_vf(adev)) { + /* Only for dGPU, vendor informaton is reliable */ + if (!amdgpu_sriov_vf(adev) && !(adev->flags & AMD_IS_APU)) { vram_info = RREG32(regBIF_BIOS_SCRATCH_4); adev->gmc.vram_vendor = vram_info & 0xF; } @@ -2340,8 +2341,8 @@ static int gmc_v9_0_hw_init(void *handle) if (amdgpu_emu_mode == 1) return amdgpu_gmc_vram_checking(adev); - else - return r; + + return 0; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c index aecad530b10a..2c02ae69883d 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c @@ -215,6 +215,11 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev, tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32(mmIH_RB_CNTL, tmp); + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32(mmIH_RB_CNTL, tmp); out: return (wptr & ih->ptr_mask); diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c index d9ed7332d805..ad4ad39f128f 100644 --- a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c @@ -418,6 +418,12 @@ static u32 ih_v6_0_get_wptr(struct amdgpu_device *adev, tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); out: return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c index 8fb05eae340a..b8da0fc29378 100644 --- a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c @@ -418,6 +418,13 @@ static u32 ih_v6_1_get_wptr(struct amdgpu_device *adev, tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + out: return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c index e64b33115848..de93614726c9 100644 --- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c @@ -442,6 +442,12 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev, tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); out: return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c index 9a24f17a5750..cada9f300a7f 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c @@ -119,6 +119,12 @@ static u32 si_ih_get_wptr(struct amdgpu_device *adev, tmp = RREG32(IH_RB_CNTL); tmp |= IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK; WREG32(IH_RB_CNTL, tmp); + + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp &= ~IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK; + WREG32(IH_RB_CNTL, tmp); } return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c index 917707bba7f3..450b6e831509 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c @@ -219,6 +219,12 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev, tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32(mmIH_RB_CNTL, tmp); + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32(mmIH_RB_CNTL, tmp); + out: return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 169ed400ee7b..8ab01ae919d2 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -2018,22 +2018,6 @@ static int vcn_v4_0_set_powergating_state(void *handle, enum amd_powergating_sta } /** - * vcn_v4_0_set_interrupt_state - set VCN block interrupt state - * - * @adev: amdgpu_device pointer - * @source: interrupt sources - * @type: interrupt types - * @state: interrupt states - * - * Set VCN block interrupt state - */ -static int vcn_v4_0_set_interrupt_state(struct amdgpu_device *adev, struct amdgpu_irq_src *source, - unsigned type, enum amdgpu_interrupt_state state) -{ - return 0; -} - -/** * vcn_v4_0_set_ras_interrupt_state - set VCN block RAS interrupt state * * @adev: amdgpu_device pointer @@ -2097,7 +2081,6 @@ static int vcn_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_ } static const struct amdgpu_irq_src_funcs vcn_v4_0_irq_funcs = { - .set = vcn_v4_0_set_interrupt_state, .process = vcn_v4_0_process_interrupt, }; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index 2eda30e78f61..49e4c3c09aca 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -269,8 +269,6 @@ static int vcn_v4_0_5_hw_fini(void *handle) vcn_v4_0_5_set_powergating_state(adev, AMD_PG_STATE_GATE); } } - - amdgpu_irq_put(adev, &adev->vcn.inst[i].irq, 0); } return 0; @@ -1669,22 +1667,6 @@ static int vcn_v4_0_5_set_powergating_state(void *handle, enum amd_powergating_s } /** - * vcn_v4_0_5_set_interrupt_state - set VCN block interrupt state - * - * @adev: amdgpu_device pointer - * @source: interrupt sources - * @type: interrupt types - * @state: interrupt states - * - * Set VCN block interrupt state - */ -static int vcn_v4_0_5_set_interrupt_state(struct amdgpu_device *adev, struct amdgpu_irq_src *source, - unsigned type, enum amdgpu_interrupt_state state) -{ - return 0; -} - -/** * vcn_v4_0_5_process_interrupt - process VCN block interrupt * * @adev: amdgpu_device pointer @@ -1726,7 +1708,6 @@ static int vcn_v4_0_5_process_interrupt(struct amdgpu_device *adev, struct amdgp } static const struct amdgpu_irq_src_funcs vcn_v4_0_5_irq_funcs = { - .set = vcn_v4_0_5_set_interrupt_state, .process = vcn_v4_0_5_process_interrupt, }; diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index d364c6dd152c..bf68e18e3824 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -373,6 +373,12 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev, tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + out: return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c index ddfc6941f9d5..db66e6cccaf2 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c @@ -421,6 +421,12 @@ static u32 vega20_ih_get_wptr(struct amdgpu_device *adev, tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + /* Unset the CLEAR_OVERFLOW bit immediately so new overflows + * can be detected. + */ + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0); + WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp); + out: return (wptr & ih->ptr_mask); } diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h index df75863393fc..d1caaf0e6a7c 100644 --- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h +++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h @@ -674,7 +674,7 @@ static const uint32_t cwsr_trap_gfx9_hex[] = { 0x86ea6a6a, 0x8f6e837a, 0xb96ee0c2, 0xbf800002, 0xb97a0002, 0xbf8a0000, - 0xbe801f6c, 0xbf810000, + 0xbe801f6c, 0xbf9b0000, }; static const uint32_t cwsr_trap_nv1x_hex[] = { @@ -1091,7 +1091,7 @@ static const uint32_t cwsr_trap_nv1x_hex[] = { 0xb9eef807, 0x876dff6d, 0x0000ffff, 0x87fe7e7e, 0x87ea6a6a, 0xb9faf802, - 0xbe80226c, 0xbf810000, + 0xbe80226c, 0xbf9b0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0x00000000, @@ -1574,7 +1574,7 @@ static const uint32_t cwsr_trap_arcturus_hex[] = { 0x86ea6a6a, 0x8f6e837a, 0xb96ee0c2, 0xbf800002, 0xb97a0002, 0xbf8a0000, - 0xbe801f6c, 0xbf810000, + 0xbe801f6c, 0xbf9b0000, }; static const uint32_t cwsr_trap_aldebaran_hex[] = { @@ -2065,7 +2065,7 @@ static const uint32_t cwsr_trap_aldebaran_hex[] = { 0x86ea6a6a, 0x8f6e837a, 0xb96ee0c2, 0xbf800002, 0xb97a0002, 0xbf8a0000, - 0xbe801f6c, 0xbf810000, + 0xbe801f6c, 0xbf9b0000, }; static const uint32_t cwsr_trap_gfx10_hex[] = { @@ -2500,7 +2500,7 @@ static const uint32_t cwsr_trap_gfx10_hex[] = { 0x876dff6d, 0x0000ffff, 0x87fe7e7e, 0x87ea6a6a, 0xb9faf802, 0xbe80226c, - 0xbf810000, 0xbf9f0000, + 0xbf9b0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, }; @@ -2944,7 +2944,7 @@ static const uint32_t cwsr_trap_gfx11_hex[] = { 0xb8eef802, 0xbf0d866e, 0xbfa20002, 0xb97af802, 0xbe80486c, 0xb97af802, - 0xbe804a6c, 0xbfb00000, + 0xbe804a6c, 0xbfb10000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0x00000000, @@ -3436,5 +3436,5 @@ static const uint32_t cwsr_trap_gfx9_4_3_hex[] = { 0x86ea6a6a, 0x8f6e837a, 0xb96ee0c2, 0xbf800002, 0xb97a0002, 0xbf8a0000, - 0xbe801f6c, 0xbf810000, + 0xbe801f6c, 0xbf9b0000, }; diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm index e0140df0b0ec..71b3dc0c7363 100644 --- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm +++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm @@ -1104,7 +1104,7 @@ L_RETURN_WITHOUT_PRIV: s_rfe_b64 s_restore_pc_lo //Return to the main shader program and resume execution L_END_PGM: - s_endpgm + s_endpgm_saved end function write_hwreg_to_mem(s, s_rsrc, s_mem_offset) diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm index e506411ad28a..bb26338204f4 100644 --- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm +++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm @@ -921,7 +921,7 @@ L_RESTORE: /* the END */ /**************************************************************************/ L_END_PGM: - s_endpgm + s_endpgm_saved end diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index ce4c52ec34d8..80e90fdef291 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -1442,7 +1442,9 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, kfd_flush_tlb(peer_pdd, TLB_FLUSH_HEAVYWEIGHT); /* Remove dma mapping after tlb flush to avoid IO_PAGE_FAULT */ - amdgpu_amdkfd_gpuvm_dmaunmap_mem(mem, peer_pdd->drm_priv); + err = amdgpu_amdkfd_gpuvm_dmaunmap_mem(mem, peer_pdd->drm_priv); + if (err) + goto sync_memory_failed; } mutex_unlock(&p->mutex); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c index f856901055d3..bdc01ca9609a 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c @@ -574,7 +574,7 @@ svm_migrate_copy_to_ram(struct amdgpu_device *adev, struct svm_range *prange, pr_debug("svms 0x%p [0x%lx 0x%lx]\n", prange->svms, prange->start, prange->last); - addr = prange->start << PAGE_SHIFT; + addr = migrate->start; src = (uint64_t *)(scratch + npages); dst = scratch; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c index 8b7fed913526..22cbfa1bdadd 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c @@ -170,6 +170,7 @@ static void update_mqd(struct mqd_manager *mm, void *mqd, m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT; m->cp_hqd_pq_control |= ffs(q->queue_size / sizeof(unsigned int)) - 1 - 1; + m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__UNORD_DISPATCH_MASK; pr_debug("cp_hqd_pq_control 0x%x\n", m->cp_hqd_pq_control); m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c index 15277f1d5cf0..d722cbd31783 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c @@ -224,6 +224,7 @@ static void update_mqd(struct mqd_manager *mm, void *mqd, m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT; m->cp_hqd_pq_control |= ffs(q->queue_size / sizeof(unsigned int)) - 1 - 1; + m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__UNORD_DISPATCH_MASK; pr_debug("cp_hqd_pq_control 0x%x\n", m->cp_hqd_pq_control); m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index 17fbedbf3651..677281c0793e 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -1488,10 +1488,15 @@ void kfd_dec_compute_active(struct kfd_node *dev); /* Cgroup Support */ /* Check with device cgroup if @kfd device is accessible */ -static inline int kfd_devcgroup_check_permission(struct kfd_node *kfd) +static inline int kfd_devcgroup_check_permission(struct kfd_node *node) { #if defined(CONFIG_CGROUP_DEVICE) || defined(CONFIG_CGROUP_BPF) - struct drm_device *ddev = adev_to_drm(kfd->adev); + struct drm_device *ddev; + + if (node->xcp) + ddev = node->xcp->ddev; + else + ddev = adev_to_drm(node->adev); return devcgroup_check_permission(DEVCG_DEV_CHAR, DRM_MAJOR, ddev->render->index, diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index d4f525b66a09..d292f290cd6e 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -272,6 +272,7 @@ static int dm_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc, { u32 v_blank_start, v_blank_end, h_position, v_position; struct amdgpu_crtc *acrtc = NULL; + struct dc *dc = adev->dm.dc; if ((crtc < 0) || (crtc >= adev->mode_info.num_crtc)) return -EINVAL; @@ -284,6 +285,9 @@ static int dm_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc, return 0; } + if (dc && dc->caps.ips_support && dc->idle_optimizations_allowed) + dc_allow_idle_optimizations(dc, false); + /* * TODO rework base driver to use values directly. * for now parse it back into reg-format @@ -1715,7 +1719,10 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) init_data.nbio_reg_offsets = adev->reg_offset[NBIO_HWIP][0]; init_data.clk_reg_offsets = adev->reg_offset[CLK_HWIP][0]; - init_data.flags.disable_ips = DMUB_IPS_DISABLE_ALL; + if (amdgpu_dc_debug_mask & DC_DISABLE_IPS) + init_data.flags.disable_ips = DMUB_IPS_DISABLE_ALL; + + init_data.flags.disable_ips_in_vpb = 1; /* Enable DWB for tested platforms only */ if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 0, 0)) @@ -8976,16 +8983,8 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) trace_amdgpu_dm_atomic_commit_tail_begin(state); - if (dm->dc->caps.ips_support) { - for_each_oldnew_connector_in_state(state, connector, old_con_state, new_con_state, i) { - if (new_con_state->crtc && - new_con_state->crtc->state->active && - drm_atomic_crtc_needs_modeset(new_con_state->crtc->state)) { - dc_dmub_srv_apply_idle_power_optimizations(dm->dc, false); - break; - } - } - } + if (dm->dc->caps.ips_support && dm->dc->idle_optimizations_allowed) + dc_allow_idle_optimizations(dm->dc, false); drm_atomic_helper_update_legacy_modeset_state(dev, state); drm_dp_mst_atomic_wait_for_dependencies(state); @@ -9188,6 +9187,10 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) * To fix this, DC should permit updating only stream properties. */ dummy_updates = kzalloc(sizeof(struct dc_surface_update) * MAX_SURFACES, GFP_ATOMIC); + if (!dummy_updates) { + DRM_ERROR("Failed to allocate memory for dummy_updates.\n"); + continue; + } for (j = 0; j < status->plane_count; j++) dummy_updates[j].surface = status->plane_states[0]; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c index 58b880acb087..3390f0d8420a 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c @@ -711,7 +711,7 @@ static inline int dm_irq_state(struct amdgpu_device *adev, { bool st; enum dc_irq_source irq_source; - + struct dc *dc = adev->dm.dc; struct amdgpu_crtc *acrtc = adev->mode_info.crtcs[crtc_id]; if (!acrtc) { @@ -729,6 +729,9 @@ static inline int dm_irq_state(struct amdgpu_device *adev, st = (state == AMDGPU_IRQ_STATE_ENABLE); + if (dc && dc->caps.ips_support && dc->idle_optimizations_allowed) + dc_allow_idle_optimizations(dc, false); + dc_interrupt_set(adev->dm.dc, irq_source, st); return 0; } diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c index 9c660d1facc7..14cec1c7b718 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c @@ -437,32 +437,32 @@ static struct wm_table ddr5_wm_table = { .wm_inst = WM_A, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, { .wm_inst = WM_B, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, { .wm_inst = WM_C, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, { .wm_inst = WM_D, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, } @@ -474,32 +474,32 @@ static struct wm_table lpddr5_wm_table = { .wm_inst = WM_A, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, { .wm_inst = WM_B, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, { .wm_inst = WM_C, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, { .wm_inst = WM_D, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .valid = true, }, } diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 5d7aa882416b..c9317ea0258e 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -434,6 +434,7 @@ struct dc_config { bool EnableMinDispClkODM; bool enable_auto_dpm_test_logs; unsigned int disable_ips; + unsigned int disable_ips_in_vpb; }; enum visual_confirm { diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index b08ccb8c68bc..9900dda2eef5 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -1034,6 +1034,7 @@ enum replay_FW_Message_type { Replay_Msg_Not_Support = -1, Replay_Set_Timing_Sync_Supported, Replay_Set_Residency_Frameupdate_Timer, + Replay_Set_Pseudo_VTotal, }; union replay_error_status { @@ -1089,6 +1090,10 @@ struct replay_settings { uint16_t coasting_vtotal_table[PR_COASTING_TYPE_NUM]; /* Maximum link off frame count */ enum replay_link_off_frame_count_level link_off_frame_count_level; + /* Replay pseudo vtotal for abm + ips on full screen video which can improve ips residency */ + uint16_t abm_with_ips_on_full_screen_video_pseudo_vtotal; + /* Replay last pseudo vtotal set to DMUB */ + uint16_t last_pseudo_vtotal; }; /* To split out "global" and "per-panel" config settings. diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dio_link_encoder.c index 501388014855..d761b0df2878 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dio_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dio_link_encoder.c @@ -203,12 +203,12 @@ void dcn32_link_encoder_construct( enc10->base.hpd_source = init_data->hpd_source; enc10->base.connector = init_data->connector; - if (enc10->base.connector.id == CONNECTOR_ID_USBC) - enc10->base.features.flags.bits.DP_IS_USB_C = 1; enc10->base.preferred_engine = ENGINE_ID_UNKNOWN; enc10->base.features = *enc_features; + if (enc10->base.connector.id == CONNECTOR_ID_USBC) + enc10->base.features.flags.bits.DP_IS_USB_C = 1; enc10->base.transmitter = init_data->transmitter; diff --git a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_link_encoder.c index da94e5309fba..81e349d5835b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_link_encoder.c @@ -184,8 +184,6 @@ void dcn35_link_encoder_construct( enc10->base.hpd_source = init_data->hpd_source; enc10->base.connector = init_data->connector; - if (enc10->base.connector.id == CONNECTOR_ID_USBC) - enc10->base.features.flags.bits.DP_IS_USB_C = 1; enc10->base.preferred_engine = ENGINE_ID_UNKNOWN; @@ -240,6 +238,8 @@ void dcn35_link_encoder_construct( } enc10->base.features.flags.bits.HDMI_6GB_EN = 1; + if (enc10->base.connector.id == CONNECTOR_ID_USBC) + enc10->base.features.flags.bits.DP_IS_USB_C = 1; if (bp_funcs->get_connector_speed_cap_info) result = bp_funcs->get_connector_speed_cap_info(enc10->base.ctx->dc_bios, diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c index 9f37f717a1f8..dd781a20692e 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c @@ -1112,7 +1112,7 @@ struct pipe_slice_table { struct pipe_ctx *pri_pipe; struct dc_plane_state *plane; int slice_count; - } mpc_combines[MAX_SURFACES]; + } mpc_combines[MAX_PLANES]; int mpc_combine_count; }; @@ -2753,7 +2753,7 @@ static int build_synthetic_soc_states(bool disable_dc_mode_overwrite, struct clk struct _vcs_dpi_voltage_scaling_st entry = {0}; struct clk_limit_table_entry max_clk_data = {0}; - unsigned int min_dcfclk_mhz = 199, min_fclk_mhz = 299; + unsigned int min_dcfclk_mhz = 399, min_fclk_mhz = 599; static const unsigned int num_dcfclk_stas = 5; unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {199, 615, 906, 1324, 1564}; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c index 475c4ec43c01..7ea2bd5374d5 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c @@ -164,8 +164,8 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { }, }, .num_states = 5, - .sr_exit_time_us = 14.0, - .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_time_us = 28.0, + .sr_enter_plus_exit_time_us = 30.0, .sr_exit_z8_time_us = 210.0, .sr_enter_plus_exit_z8_time_us = 320.0, .fclk_change_latency_us = 24.0, diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c index 64d01a9cd68c..23a608274096 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c @@ -341,9 +341,6 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc, break; } - if (dml2->config.bbox_overrides.clks_table.num_states) - p->in_states->num_states = dml2->config.bbox_overrides.clks_table.num_states; - /* Override from passed values, if available */ for (i = 0; i < p->in_states->num_states; i++) { if (dml2->config.bbox_overrides.sr_exit_latency_us) { @@ -400,6 +397,7 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc, } /* Copy clocks tables entries, if available */ if (dml2->config.bbox_overrides.clks_table.num_states) { + p->in_states->num_states = dml2->config.bbox_overrides.clks_table.num_states; for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels; i++) { p->in_states->state_array[i].dcfclk_mhz = dml2->config.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz; @@ -793,35 +791,28 @@ static void populate_dml_surface_cfg_from_plane_state(enum dml_project_id dml2_p } } -/*TODO no support for mpc combine, need rework - should calculate scaling params based on plane+stream*/ -static struct scaler_data get_scaler_data_for_plane(const struct dc_plane_state *in, const struct dc_state *context) +static struct scaler_data get_scaler_data_for_plane(const struct dc_plane_state *in, struct dc_state *context) { int i; - struct scaler_data data = { 0 }; + struct pipe_ctx *temp_pipe = &context->res_ctx.temp_pipe; + + memset(temp_pipe, 0, sizeof(struct pipe_ctx)); for (i = 0; i < MAX_PIPES; i++) { const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; if (pipe->plane_state == in && !pipe->prev_odm_pipe) { - const struct pipe_ctx *next_pipe = pipe->next_odm_pipe; - - data = context->res_ctx.pipe_ctx[i].plane_res.scl_data; - while (next_pipe) { - data.h_active += next_pipe->plane_res.scl_data.h_active; - data.recout.width += next_pipe->plane_res.scl_data.recout.width; - if (in->rotation == ROTATION_ANGLE_0 || in->rotation == ROTATION_ANGLE_180) { - data.viewport.width += next_pipe->plane_res.scl_data.viewport.width; - } else { - data.viewport.height += next_pipe->plane_res.scl_data.viewport.height; - } - next_pipe = next_pipe->next_odm_pipe; - } + temp_pipe->stream = pipe->stream; + temp_pipe->plane_state = pipe->plane_state; + temp_pipe->plane_res.scl_data.taps = pipe->plane_res.scl_data.taps; + + resource_build_scaling_params(temp_pipe); break; } } ASSERT(i < MAX_PIPES); - return data; + return temp_pipe->plane_res.scl_data; } static void populate_dummy_dml_plane_cfg(struct dml_plane_cfg_st *out, unsigned int location, const struct dc_stream_state *in) @@ -866,7 +857,7 @@ static void populate_dummy_dml_plane_cfg(struct dml_plane_cfg_st *out, unsigned out->ScalerEnabled[location] = false; } -static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out, unsigned int location, const struct dc_plane_state *in, const struct dc_state *context) +static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out, unsigned int location, const struct dc_plane_state *in, struct dc_state *context) { const struct scaler_data scaler_data = get_scaler_data_for_plane(in, context); diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c index 5660f15da291..2352428bcea3 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c @@ -1183,9 +1183,9 @@ void dce110_disable_stream(struct pipe_ctx *pipe_ctx) dto_params.timing = &pipe_ctx->stream->timing; dp_hpo_inst = pipe_ctx->stream_res.hpo_dp_stream_enc->inst; if (dccg) { - dccg->funcs->set_dtbclk_dto(dccg, &dto_params); dccg->funcs->disable_symclk32_se(dccg, dp_hpo_inst); dccg->funcs->set_dpstreamclk(dccg, REFCLK, tg->inst, dp_hpo_inst); + dccg->funcs->set_dtbclk_dto(dccg, &dto_params); } } else if (dccg && dccg->funcs->disable_symclk_se) { dccg->funcs->disable_symclk_se(dccg, stream_enc->stream_enc_inst, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c index e931342fcf4c..4853ecac53f9 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c @@ -2790,18 +2790,17 @@ void dcn20_enable_stream(struct pipe_ctx *pipe_ctx) } if (dc->link_srv->dp_is_128b_132b_signal(pipe_ctx)) { - dp_hpo_inst = pipe_ctx->stream_res.hpo_dp_stream_enc->inst; - dccg->funcs->set_dpstreamclk(dccg, DTBCLK0, tg->inst, dp_hpo_inst); - - phyd32clk = get_phyd32clk_src(link); - dccg->funcs->enable_symclk32_se(dccg, dp_hpo_inst, phyd32clk); - dto_params.otg_inst = tg->inst; dto_params.pixclk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10; dto_params.num_odm_segments = get_odm_segment_count(pipe_ctx); dto_params.timing = &pipe_ctx->stream->timing; dto_params.ref_dtbclk_khz = dc->clk_mgr->funcs->get_dtb_ref_clk_frequency(dc->clk_mgr); dccg->funcs->set_dtbclk_dto(dccg, &dto_params); + dp_hpo_inst = pipe_ctx->stream_res.hpo_dp_stream_enc->inst; + dccg->funcs->set_dpstreamclk(dccg, DTBCLK0, tg->inst, dp_hpo_inst); + + phyd32clk = get_phyd32clk_src(link); + dccg->funcs->enable_symclk32_se(dccg, dp_hpo_inst, phyd32clk); } else { if (dccg->funcs->enable_symclk_se) dccg->funcs->enable_symclk_se(dccg, stream_enc->stream_enc_inst, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c index 9c806385ecbd..8b6c49622f3b 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c @@ -680,7 +680,7 @@ void dcn35_power_down_on_boot(struct dc *dc) bool dcn35_apply_idle_power_optimizations(struct dc *dc, bool enable) { struct dc_link *edp_links[MAX_NUM_EDP]; - int edp_num; + int i, edp_num; if (dc->debug.dmcub_emulation) return true; @@ -688,6 +688,13 @@ bool dcn35_apply_idle_power_optimizations(struct dc *dc, bool enable) dc_get_edp_links(dc, edp_links, &edp_num); if (edp_num == 0 || edp_num > 1) return false; + + for (i = 0; i < dc->current_state->stream_count; ++i) { + struct dc_stream_state *stream = dc->current_state->streams[i]; + + if (!stream->dpms_off && !dc_is_embedded_signal(stream->signal)) + return false; + } } // TODO: review other cases when idle optimization is allowed diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h index f74ae0d41d3c..3a6bf77a6873 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h @@ -469,6 +469,8 @@ struct resource_context { unsigned int hpo_dp_link_enc_to_link_idx[MAX_HPO_DP2_LINK_ENCODERS]; int hpo_dp_link_enc_ref_cnts[MAX_HPO_DP2_LINK_ENCODERS]; bool is_mpc_3dlut_acquired[MAX_PIPES]; + /* solely used for build scalar data in dml2 */ + struct pipe_ctx temp_pipe; }; struct dce_bw_output { diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c index dd0d2b206462..5491b707cec8 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c @@ -196,7 +196,7 @@ static int get_host_router_total_dp_tunnel_bw(const struct dc *dc, uint8_t hr_in struct dc_link *link_dpia_primary, *link_dpia_secondary; int total_bw = 0; - for (uint8_t i = 0; i < MAX_PIPES * 2; ++i) { + for (uint8_t i = 0; i < (MAX_PIPES * 2) - 1; ++i) { if (!dc->links[i] || dc->links[i]->ep_type != DISPLAY_ENDPOINT_USB4_DPIA) continue; diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dpcd.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dpcd.c index 5c9a30211c10..fc50931c2aec 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dpcd.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dpcd.c @@ -205,7 +205,7 @@ enum dc_status core_link_read_dpcd( uint32_t extended_size; /* size of the remaining partitioned address space */ uint32_t size_left_to_read; - enum dc_status status; + enum dc_status status = DC_ERROR_UNEXPECTED; /* size of the next partition to be read from */ uint32_t partition_size; uint32_t data_index = 0; @@ -234,7 +234,7 @@ enum dc_status core_link_write_dpcd( { uint32_t partition_size; uint32_t data_index = 0; - enum dc_status status; + enum dc_status status = DC_ERROR_UNEXPECTED; while (size) { partition_size = dpcd_get_next_partition_size(address, size); diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h index c64b6c848ef7..e699731ee68e 100644 --- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h +++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h @@ -2832,6 +2832,7 @@ struct dmub_rb_cmd_psr_set_power_opt { #define REPLAY_RESIDENCY_MODE_MASK (0x1 << REPLAY_RESIDENCY_MODE_SHIFT) # define REPLAY_RESIDENCY_MODE_PHY (0x0 << REPLAY_RESIDENCY_MODE_SHIFT) # define REPLAY_RESIDENCY_MODE_ALPM (0x1 << REPLAY_RESIDENCY_MODE_SHIFT) +# define REPLAY_RESIDENCY_MODE_IPS 0x10 #define REPLAY_RESIDENCY_ENABLE_MASK (0x1 << REPLAY_RESIDENCY_ENABLE_SHIFT) # define REPLAY_RESIDENCY_DISABLE (0x0 << REPLAY_RESIDENCY_ENABLE_SHIFT) @@ -2894,6 +2895,10 @@ enum dmub_cmd_replay_type { * Set Residency Frameupdate Timer. */ DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER = 6, + /** + * Set pseudo vtotal + */ + DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL = 7, }; /** @@ -3077,6 +3082,26 @@ struct dmub_cmd_replay_set_timing_sync_data { }; /** + * Data passed from driver to FW in a DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL command. + */ +struct dmub_cmd_replay_set_pseudo_vtotal { + /** + * Panel Instance. + * Panel isntance to identify which replay_state to use + * Currently the support is only for 0 or 1 + */ + uint8_t panel_inst; + /** + * Source Vtotal that Replay + IPS + ABM full screen video src vtotal + */ + uint16_t vtotal; + /** + * Explicit padding to 4 byte boundary. + */ + uint8_t pad; +}; + +/** * Definition of a DMUB_CMD__SET_REPLAY_POWER_OPT command. */ struct dmub_rb_cmd_replay_set_power_opt { @@ -3157,6 +3182,20 @@ struct dmub_rb_cmd_replay_set_timing_sync { }; /** + * Definition of a DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL command. + */ +struct dmub_rb_cmd_replay_set_pseudo_vtotal { + /** + * Command header. + */ + struct dmub_cmd_header header; + /** + * Definition of DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL command. + */ + struct dmub_cmd_replay_set_pseudo_vtotal data; +}; + +/** * Data passed from driver to FW in DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER command. */ struct dmub_cmd_replay_frameupdate_timer_data { @@ -3207,6 +3246,10 @@ union dmub_replay_cmd_set { * Definition of DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER command data. */ struct dmub_cmd_replay_frameupdate_timer_data timer_data; + /** + * Definition of DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL command data. + */ + struct dmub_cmd_replay_set_pseudo_vtotal pseudo_vtotal_data; }; /** @@ -4358,6 +4401,10 @@ union dmub_rb_cmd { * Definition of a DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER command. */ struct dmub_rb_cmd_replay_set_frameupdate_timer replay_set_frameupdate_timer; + /** + * Definition of a DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL command. + */ + struct dmub_rb_cmd_replay_set_pseudo_vtotal replay_set_pseudo_vtotal; }; /** diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c index ad98e504c00d..e304e8435fb8 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c @@ -980,6 +980,11 @@ void set_replay_coasting_vtotal(struct dc_link *link, link->replay_settings.coasting_vtotal_table[type] = vtotal; } +void set_replay_ips_full_screen_video_src_vtotal(struct dc_link *link, uint16_t vtotal) +{ + link->replay_settings.abm_with_ips_on_full_screen_video_pseudo_vtotal = vtotal; +} + void calculate_replay_link_off_frame_count(struct dc_link *link, uint16_t vtotal, uint16_t htotal) { diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.h b/drivers/gpu/drm/amd/display/modules/power/power_helpers.h index c17bbc6fb38c..bef4815e1703 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.h +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.h @@ -57,6 +57,7 @@ void init_replay_config(struct dc_link *link, struct replay_config *pr_config); void set_replay_coasting_vtotal(struct dc_link *link, enum replay_coasting_vtotal_type type, uint16_t vtotal); +void set_replay_ips_full_screen_video_src_vtotal(struct dc_link *link, uint16_t vtotal); void calculate_replay_link_off_frame_count(struct dc_link *link, uint16_t vtotal, uint16_t htotal); diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index 1dc5dd9b7bf7..df2c7ffe190f 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -258,6 +258,7 @@ enum DC_DEBUG_MASK { DC_ENABLE_DML2 = 0x100, DC_DISABLE_PSR_SU = 0x200, DC_DISABLE_REPLAY = 0x400, + DC_DISABLE_IPS = 0x800, }; enum amd_dpm_forced_level; diff --git a/drivers/gpu/drm/amd/include/amdgpu_reg_state.h b/drivers/gpu/drm/amd/include/amdgpu_reg_state.h index be519c8edf49..335980e2afbf 100644 --- a/drivers/gpu/drm/amd/include/amdgpu_reg_state.h +++ b/drivers/gpu/drm/amd/include/amdgpu_reg_state.h @@ -138,7 +138,7 @@ static inline size_t amdgpu_reginst_size(uint16_t num_inst, size_t inst_size, } #define amdgpu_asic_get_reg_state_supported(adev) \ - ((adev)->asic_funcs->get_reg_state ? 1 : 0) + (((adev)->asic_funcs && (adev)->asic_funcs->get_reg_state) ? 1 : 0) #define amdgpu_asic_get_reg_state(adev, state, buf, size) \ ((adev)->asic_funcs->get_reg_state ? \ diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index c16703868e5c..0ad947df777a 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -24,6 +24,7 @@ #include <linux/firmware.h> #include <linux/pci.h> +#include <linux/power_supply.h> #include <linux/reboot.h> #include "amdgpu.h" @@ -733,7 +734,7 @@ static int smu_early_init(void *handle) smu->adev = adev; smu->pm_enabled = !!amdgpu_dpm; smu->is_apu = false; - smu->smu_baco.state = SMU_BACO_STATE_NONE; + smu->smu_baco.state = SMU_BACO_STATE_EXIT; smu->smu_baco.platform_support = false; smu->user_dpm_profile.fan_mode = -1; @@ -817,16 +818,8 @@ static int smu_late_init(void *handle) * handle the switch automatically. Driver involvement * is unnecessary. */ - if (!smu->dc_controlled_by_gpio) { - ret = smu_set_power_source(smu, - adev->pm.ac_power ? SMU_POWER_SOURCE_AC : - SMU_POWER_SOURCE_DC); - if (ret) { - dev_err(adev->dev, "Failed to switch to %s mode!\n", - adev->pm.ac_power ? "AC" : "DC"); - return ret; - } - } + adev->pm.ac_power = power_supply_is_system_supplied() > 0; + smu_set_ac_dc(smu); if ((amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 1)) || (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 3))) @@ -1961,31 +1954,10 @@ static int smu_smc_hw_cleanup(struct smu_context *smu) return 0; } -static int smu_reset_mp1_state(struct smu_context *smu) -{ - struct amdgpu_device *adev = smu->adev; - int ret = 0; - - if ((!adev->in_runpm) && (!adev->in_suspend) && - (!amdgpu_in_reset(adev))) - switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) { - case IP_VERSION(13, 0, 0): - case IP_VERSION(13, 0, 7): - case IP_VERSION(13, 0, 10): - ret = smu_set_mp1_state(smu, PP_MP1_STATE_UNLOAD); - break; - default: - break; - } - - return ret; -} - static int smu_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct smu_context *smu = adev->powerplay.pp_handle; - int ret; if (amdgpu_sriov_vf(adev) && !amdgpu_sriov_is_pp_one_vf(adev)) return 0; @@ -2003,15 +1975,7 @@ static int smu_hw_fini(void *handle) adev->pm.dpm_enabled = false; - ret = smu_smc_hw_cleanup(smu); - if (ret) - return ret; - - ret = smu_reset_mp1_state(smu); - if (ret) - return ret; - - return 0; + return smu_smc_hw_cleanup(smu); } static void smu_late_fini(void *handle) @@ -2710,6 +2674,7 @@ int smu_get_power_limit(void *handle, case SMU_PPT_LIMIT_CURRENT: switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) { case IP_VERSION(13, 0, 2): + case IP_VERSION(13, 0, 6): case IP_VERSION(11, 0, 7): case IP_VERSION(11, 0, 11): case IP_VERSION(11, 0, 12): diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h index 2aa4fea87314..66e84defd0b6 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h @@ -424,7 +424,6 @@ enum smu_reset_mode { enum smu_baco_state { SMU_BACO_STATE_ENTER = 0, SMU_BACO_STATE_EXIT, - SMU_BACO_STATE_NONE, }; struct smu_baco_context { diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c index 5a314d0316c1..c7bfa68bf00f 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c @@ -1442,10 +1442,12 @@ static int smu_v11_0_irq_process(struct amdgpu_device *adev, case 0x3: dev_dbg(adev->dev, "Switched to AC mode!\n"); schedule_work(&smu->interrupt_work); + adev->pm.ac_power = true; break; case 0x4: dev_dbg(adev->dev, "Switched to DC mode!\n"); schedule_work(&smu->interrupt_work); + adev->pm.ac_power = false; break; case 0x7: /* diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c index 771a3d457c33..c486182ff275 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c @@ -1379,10 +1379,12 @@ static int smu_v13_0_irq_process(struct amdgpu_device *adev, case 0x3: dev_dbg(adev->dev, "Switched to AC mode!\n"); smu_v13_0_ack_ac_dc_interrupt(smu); + adev->pm.ac_power = true; break; case 0x4: dev_dbg(adev->dev, "Switched to DC mode!\n"); smu_v13_0_ack_ac_dc_interrupt(smu); + adev->pm.ac_power = false; break; case 0x7: /* diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index a9b25faa63e4..a9954ffc02c5 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -2357,6 +2357,7 @@ static int smu_v13_0_0_get_power_limit(struct smu_context *smu, PPTable_t *pptable = table_context->driver_pptable; SkuTable_t *skutable = &pptable->SkuTable; uint32_t power_limit, od_percent_upper, od_percent_lower; + uint32_t msg_limit = skutable->MsgLimits.Power[PPT_THROTTLER_PPT0][POWER_SOURCE_AC]; if (smu_v13_0_get_current_power_limit(smu, &power_limit)) power_limit = smu->adev->pm.ac_power ? @@ -2380,7 +2381,7 @@ static int smu_v13_0_0_get_power_limit(struct smu_context *smu, od_percent_upper, od_percent_lower, power_limit); if (max_power_limit) { - *max_power_limit = power_limit * (100 + od_percent_upper); + *max_power_limit = msg_limit * (100 + od_percent_upper); *max_power_limit /= 100; } @@ -2747,13 +2748,7 @@ static int smu_v13_0_0_set_mp1_state(struct smu_context *smu, switch (mp1_state) { case PP_MP1_STATE_UNLOAD: - ret = smu_cmn_send_smc_msg_with_param(smu, - SMU_MSG_PrepareMp1ForUnload, - 0x55, NULL); - - if (!ret && smu->smu_baco.state == SMU_BACO_STATE_EXIT) - ret = smu_v13_0_disable_pmfw_state(smu); - + ret = smu_cmn_set_mp1_state(smu, mp1_state); break; default: /* Ignore others */ @@ -2949,7 +2944,7 @@ static bool smu_v13_0_0_wbrf_support_check(struct smu_context *smu) { struct amdgpu_device *adev = smu->adev; - switch (adev->ip_versions[MP1_HWIP][0]) { + switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) { case IP_VERSION(13, 0, 0): return smu->smc_fw_version >= 0x004e6300; case IP_VERSION(13, 0, 10): @@ -2959,6 +2954,55 @@ static bool smu_v13_0_0_wbrf_support_check(struct smu_context *smu) } } +static int smu_v13_0_0_set_power_limit(struct smu_context *smu, + enum smu_ppt_limit_type limit_type, + uint32_t limit) +{ + PPTable_t *pptable = smu->smu_table.driver_pptable; + SkuTable_t *skutable = &pptable->SkuTable; + uint32_t msg_limit = skutable->MsgLimits.Power[PPT_THROTTLER_PPT0][POWER_SOURCE_AC]; + struct smu_table_context *table_context = &smu->smu_table; + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)table_context->overdrive_table; + int ret = 0; + + if (limit_type != SMU_DEFAULT_PPT_LIMIT) + return -EINVAL; + + if (limit <= msg_limit) { + if (smu->current_power_limit > msg_limit) { + od_table->OverDriveTable.Ppt = 0; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_PPT_BIT; + + ret = smu_v13_0_0_upload_overdrive_table(smu, od_table); + if (ret) { + dev_err(smu->adev->dev, "Failed to upload overdrive table!\n"); + return ret; + } + } + return smu_v13_0_set_power_limit(smu, limit_type, limit); + } else if (smu->od_enabled) { + ret = smu_v13_0_set_power_limit(smu, limit_type, msg_limit); + if (ret) + return ret; + + od_table->OverDriveTable.Ppt = (limit * 100) / msg_limit - 100; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_PPT_BIT; + + ret = smu_v13_0_0_upload_overdrive_table(smu, od_table); + if (ret) { + dev_err(smu->adev->dev, "Failed to upload overdrive table!\n"); + return ret; + } + + smu->current_power_limit = limit; + } else { + return -EINVAL; + } + + return 0; +} + static const struct pptable_funcs smu_v13_0_0_ppt_funcs = { .get_allowed_feature_mask = smu_v13_0_0_get_allowed_feature_mask, .set_default_dpm_table = smu_v13_0_0_set_default_dpm_table, @@ -3013,7 +3057,7 @@ static const struct pptable_funcs smu_v13_0_0_ppt_funcs = { .set_fan_control_mode = smu_v13_0_set_fan_control_mode, .enable_mgpu_fan_boost = smu_v13_0_0_enable_mgpu_fan_boost, .get_power_limit = smu_v13_0_0_get_power_limit, - .set_power_limit = smu_v13_0_set_power_limit, + .set_power_limit = smu_v13_0_0_set_power_limit, .set_power_source = smu_v13_0_set_power_source, .get_power_profile_mode = smu_v13_0_0_get_power_profile_mode, .set_power_profile_mode = smu_v13_0_0_set_power_profile_mode, diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c index 3c98a8a0386a..7e1941cf1796 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c @@ -160,8 +160,8 @@ static const struct cmn2asic_msg_mapping smu_v13_0_6_message_map[SMU_MSG_MAX_COU MSG_MAP(GfxDriverResetRecovery, PPSMC_MSG_GfxDriverResetRecovery, 0), MSG_MAP(GetMinGfxclkFrequency, PPSMC_MSG_GetMinGfxDpmFreq, 1), MSG_MAP(GetMaxGfxclkFrequency, PPSMC_MSG_GetMaxGfxDpmFreq, 1), - MSG_MAP(SetSoftMinGfxclk, PPSMC_MSG_SetSoftMinGfxClk, 0), - MSG_MAP(SetSoftMaxGfxClk, PPSMC_MSG_SetSoftMaxGfxClk, 0), + MSG_MAP(SetSoftMinGfxclk, PPSMC_MSG_SetSoftMinGfxClk, 1), + MSG_MAP(SetSoftMaxGfxClk, PPSMC_MSG_SetSoftMaxGfxClk, 1), MSG_MAP(PrepareMp1ForUnload, PPSMC_MSG_PrepareForDriverUnload, 0), MSG_MAP(GetCTFLimit, PPSMC_MSG_GetCTFLimit, 0), MSG_MAP(GetThermalLimit, PPSMC_MSG_ReadThrottlerLimit, 0), diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index 59606a19e3d2..0ffdb58af74e 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -2321,6 +2321,7 @@ static int smu_v13_0_7_get_power_limit(struct smu_context *smu, PPTable_t *pptable = table_context->driver_pptable; SkuTable_t *skutable = &pptable->SkuTable; uint32_t power_limit, od_percent_upper, od_percent_lower; + uint32_t msg_limit = skutable->MsgLimits.Power[PPT_THROTTLER_PPT0][POWER_SOURCE_AC]; if (smu_v13_0_get_current_power_limit(smu, &power_limit)) power_limit = smu->adev->pm.ac_power ? @@ -2344,7 +2345,7 @@ static int smu_v13_0_7_get_power_limit(struct smu_context *smu, od_percent_upper, od_percent_lower, power_limit); if (max_power_limit) { - *max_power_limit = power_limit * (100 + od_percent_upper); + *max_power_limit = msg_limit * (100 + od_percent_upper); *max_power_limit /= 100; } @@ -2504,13 +2505,7 @@ static int smu_v13_0_7_set_mp1_state(struct smu_context *smu, switch (mp1_state) { case PP_MP1_STATE_UNLOAD: - ret = smu_cmn_send_smc_msg_with_param(smu, - SMU_MSG_PrepareMp1ForUnload, - 0x55, NULL); - - if (!ret && smu->smu_baco.state == SMU_BACO_STATE_EXIT) - ret = smu_v13_0_disable_pmfw_state(smu); - + ret = smu_cmn_set_mp1_state(smu, mp1_state); break; default: /* Ignore others */ @@ -2545,6 +2540,55 @@ static bool smu_v13_0_7_wbrf_support_check(struct smu_context *smu) return smu->smc_fw_version > 0x00524600; } +static int smu_v13_0_7_set_power_limit(struct smu_context *smu, + enum smu_ppt_limit_type limit_type, + uint32_t limit) +{ + PPTable_t *pptable = smu->smu_table.driver_pptable; + SkuTable_t *skutable = &pptable->SkuTable; + uint32_t msg_limit = skutable->MsgLimits.Power[PPT_THROTTLER_PPT0][POWER_SOURCE_AC]; + struct smu_table_context *table_context = &smu->smu_table; + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)table_context->overdrive_table; + int ret = 0; + + if (limit_type != SMU_DEFAULT_PPT_LIMIT) + return -EINVAL; + + if (limit <= msg_limit) { + if (smu->current_power_limit > msg_limit) { + od_table->OverDriveTable.Ppt = 0; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_PPT_BIT; + + ret = smu_v13_0_7_upload_overdrive_table(smu, od_table); + if (ret) { + dev_err(smu->adev->dev, "Failed to upload overdrive table!\n"); + return ret; + } + } + return smu_v13_0_set_power_limit(smu, limit_type, limit); + } else if (smu->od_enabled) { + ret = smu_v13_0_set_power_limit(smu, limit_type, msg_limit); + if (ret) + return ret; + + od_table->OverDriveTable.Ppt = (limit * 100) / msg_limit - 100; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_PPT_BIT; + + ret = smu_v13_0_7_upload_overdrive_table(smu, od_table); + if (ret) { + dev_err(smu->adev->dev, "Failed to upload overdrive table!\n"); + return ret; + } + + smu->current_power_limit = limit; + } else { + return -EINVAL; + } + + return 0; +} + static const struct pptable_funcs smu_v13_0_7_ppt_funcs = { .get_allowed_feature_mask = smu_v13_0_7_get_allowed_feature_mask, .set_default_dpm_table = smu_v13_0_7_set_default_dpm_table, @@ -2596,7 +2640,7 @@ static const struct pptable_funcs smu_v13_0_7_ppt_funcs = { .set_fan_control_mode = smu_v13_0_set_fan_control_mode, .enable_mgpu_fan_boost = smu_v13_0_7_enable_mgpu_fan_boost, .get_power_limit = smu_v13_0_7_get_power_limit, - .set_power_limit = smu_v13_0_set_power_limit, + .set_power_limit = smu_v13_0_7_set_power_limit, .set_power_source = smu_v13_0_set_power_source, .get_power_profile_mode = smu_v13_0_7_get_power_profile_mode, .set_power_profile_mode = smu_v13_0_7_set_power_profile_mode, diff --git a/drivers/gpu/drm/bridge/analogix/anx7625.c b/drivers/gpu/drm/bridge/analogix/anx7625.c index ef31033439bc..29d91493b101 100644 --- a/drivers/gpu/drm/bridge/analogix/anx7625.c +++ b/drivers/gpu/drm/bridge/analogix/anx7625.c @@ -1762,6 +1762,7 @@ static ssize_t anx7625_aux_transfer(struct drm_dp_aux *aux, u8 request = msg->request & ~DP_AUX_I2C_MOT; int ret = 0; + mutex_lock(&ctx->aux_lock); pm_runtime_get_sync(dev); msg->reply = 0; switch (request) { @@ -1778,6 +1779,7 @@ static ssize_t anx7625_aux_transfer(struct drm_dp_aux *aux, msg->size, msg->buffer); pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); + mutex_unlock(&ctx->aux_lock); return ret; } @@ -2474,7 +2476,9 @@ static void anx7625_bridge_atomic_disable(struct drm_bridge *bridge, ctx->connector = NULL; anx7625_dp_stop(ctx); - pm_runtime_put_sync(dev); + mutex_lock(&ctx->aux_lock); + pm_runtime_put_sync_suspend(dev); + mutex_unlock(&ctx->aux_lock); } static enum drm_connector_status @@ -2668,6 +2672,7 @@ static int anx7625_i2c_probe(struct i2c_client *client) mutex_init(&platform->lock); mutex_init(&platform->hdcp_wq_lock); + mutex_init(&platform->aux_lock); INIT_DELAYED_WORK(&platform->hdcp_work, hdcp_check_work_func); platform->hdcp_workqueue = create_workqueue("hdcp workqueue"); diff --git a/drivers/gpu/drm/bridge/analogix/anx7625.h b/drivers/gpu/drm/bridge/analogix/anx7625.h index 66ebee7f3d83..39ed35d33836 100644 --- a/drivers/gpu/drm/bridge/analogix/anx7625.h +++ b/drivers/gpu/drm/bridge/analogix/anx7625.h @@ -475,6 +475,8 @@ struct anx7625_data { struct workqueue_struct *hdcp_workqueue; /* Lock for hdcp work queue */ struct mutex hdcp_wq_lock; + /* Lock for aux transfer and disable */ + struct mutex aux_lock; char edid_block; struct display_timing dt; u8 display_timing_valid; diff --git a/drivers/gpu/drm/bridge/parade-ps8640.c b/drivers/gpu/drm/bridge/parade-ps8640.c index 541e4f5afc4c..14d4dcf239da 100644 --- a/drivers/gpu/drm/bridge/parade-ps8640.c +++ b/drivers/gpu/drm/bridge/parade-ps8640.c @@ -107,6 +107,7 @@ struct ps8640 { struct device_link *link; bool pre_enabled; bool need_post_hpd_delay; + struct mutex aux_lock; }; static const struct regmap_config ps8640_regmap_config[] = { @@ -345,11 +346,20 @@ static ssize_t ps8640_aux_transfer(struct drm_dp_aux *aux, struct device *dev = &ps_bridge->page[PAGE0_DP_CNTL]->dev; int ret; + mutex_lock(&ps_bridge->aux_lock); pm_runtime_get_sync(dev); + ret = _ps8640_wait_hpd_asserted(ps_bridge, 200 * 1000); + if (ret) { + pm_runtime_put_sync_suspend(dev); + goto exit; + } ret = ps8640_aux_transfer_msg(aux, msg); pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); +exit: + mutex_unlock(&ps_bridge->aux_lock); + return ret; } @@ -470,7 +480,18 @@ static void ps8640_atomic_post_disable(struct drm_bridge *bridge, ps_bridge->pre_enabled = false; ps8640_bridge_vdo_control(ps_bridge, DISABLE); + + /* + * The bridge seems to expect everything to be power cycled at the + * disable process, so grab a lock here to make sure + * ps8640_aux_transfer() is not holding a runtime PM reference and + * preventing the bridge from suspend. + */ + mutex_lock(&ps_bridge->aux_lock); + pm_runtime_put_sync_suspend(&ps_bridge->page[PAGE0_DP_CNTL]->dev); + + mutex_unlock(&ps_bridge->aux_lock); } static int ps8640_bridge_attach(struct drm_bridge *bridge, @@ -619,6 +640,8 @@ static int ps8640_probe(struct i2c_client *client) if (!ps_bridge) return -ENOMEM; + mutex_init(&ps_bridge->aux_lock); + ps_bridge->supplies[0].supply = "vdd12"; ps_bridge->supplies[1].supply = "vdd33"; ret = devm_regulator_bulk_get(dev, ARRAY_SIZE(ps_bridge->supplies), diff --git a/drivers/gpu/drm/bridge/samsung-dsim.c b/drivers/gpu/drm/bridge/samsung-dsim.c index be5914caa17d..63a1a0c88be4 100644 --- a/drivers/gpu/drm/bridge/samsung-dsim.c +++ b/drivers/gpu/drm/bridge/samsung-dsim.c @@ -969,10 +969,6 @@ static int samsung_dsim_init_link(struct samsung_dsim *dsi) reg = samsung_dsim_read(dsi, DSIM_ESCMODE_REG); reg &= ~DSIM_STOP_STATE_CNT_MASK; reg |= DSIM_STOP_STATE_CNT(driver_data->reg_values[STOP_STATE_CNT]); - - if (!samsung_dsim_hw_is_exynos(dsi->plat_data->hw_type)) - reg |= DSIM_FORCE_STOP_STATE; - samsung_dsim_write(dsi, DSIM_ESCMODE_REG, reg); reg = DSIM_BTA_TIMEOUT(0xff) | DSIM_LPDR_TIMEOUT(0xffff); @@ -1431,18 +1427,6 @@ static void samsung_dsim_disable_irq(struct samsung_dsim *dsi) disable_irq(dsi->irq); } -static void samsung_dsim_set_stop_state(struct samsung_dsim *dsi, bool enable) -{ - u32 reg = samsung_dsim_read(dsi, DSIM_ESCMODE_REG); - - if (enable) - reg |= DSIM_FORCE_STOP_STATE; - else - reg &= ~DSIM_FORCE_STOP_STATE; - - samsung_dsim_write(dsi, DSIM_ESCMODE_REG, reg); -} - static int samsung_dsim_init(struct samsung_dsim *dsi) { const struct samsung_dsim_driver_data *driver_data = dsi->driver_data; @@ -1492,9 +1476,6 @@ static void samsung_dsim_atomic_pre_enable(struct drm_bridge *bridge, ret = samsung_dsim_init(dsi); if (ret) return; - - samsung_dsim_set_display_mode(dsi); - samsung_dsim_set_display_enable(dsi, true); } } @@ -1503,12 +1484,8 @@ static void samsung_dsim_atomic_enable(struct drm_bridge *bridge, { struct samsung_dsim *dsi = bridge_to_dsi(bridge); - if (samsung_dsim_hw_is_exynos(dsi->plat_data->hw_type)) { - samsung_dsim_set_display_mode(dsi); - samsung_dsim_set_display_enable(dsi, true); - } else { - samsung_dsim_set_stop_state(dsi, false); - } + samsung_dsim_set_display_mode(dsi); + samsung_dsim_set_display_enable(dsi, true); dsi->state |= DSIM_STATE_VIDOUT_AVAILABLE; } @@ -1521,9 +1498,6 @@ static void samsung_dsim_atomic_disable(struct drm_bridge *bridge, if (!(dsi->state & DSIM_STATE_ENABLED)) return; - if (!samsung_dsim_hw_is_exynos(dsi->plat_data->hw_type)) - samsung_dsim_set_stop_state(dsi, true); - dsi->state &= ~DSIM_STATE_VIDOUT_AVAILABLE; } @@ -1828,8 +1802,6 @@ static ssize_t samsung_dsim_host_transfer(struct mipi_dsi_host *host, if (ret) return ret; - samsung_dsim_set_stop_state(dsi, false); - ret = mipi_dsi_create_packet(&xfer.packet, msg); if (ret < 0) return ret; diff --git a/drivers/gpu/drm/bridge/sii902x.c b/drivers/gpu/drm/bridge/sii902x.c index 2bdc5b439beb..4560ae9cbce1 100644 --- a/drivers/gpu/drm/bridge/sii902x.c +++ b/drivers/gpu/drm/bridge/sii902x.c @@ -1080,6 +1080,26 @@ static int sii902x_init(struct sii902x *sii902x) return ret; } + ret = sii902x_audio_codec_init(sii902x, dev); + if (ret) + return ret; + + i2c_set_clientdata(sii902x->i2c, sii902x); + + sii902x->i2cmux = i2c_mux_alloc(sii902x->i2c->adapter, dev, + 1, 0, I2C_MUX_GATE, + sii902x_i2c_bypass_select, + sii902x_i2c_bypass_deselect); + if (!sii902x->i2cmux) { + ret = -ENOMEM; + goto err_unreg_audio; + } + + sii902x->i2cmux->priv = sii902x; + ret = i2c_mux_add_adapter(sii902x->i2cmux, 0, 0, 0); + if (ret) + goto err_unreg_audio; + sii902x->bridge.funcs = &sii902x_bridge_funcs; sii902x->bridge.of_node = dev->of_node; sii902x->bridge.timings = &default_sii902x_timings; @@ -1090,19 +1110,13 @@ static int sii902x_init(struct sii902x *sii902x) drm_bridge_add(&sii902x->bridge); - sii902x_audio_codec_init(sii902x, dev); - - i2c_set_clientdata(sii902x->i2c, sii902x); + return 0; - sii902x->i2cmux = i2c_mux_alloc(sii902x->i2c->adapter, dev, - 1, 0, I2C_MUX_GATE, - sii902x_i2c_bypass_select, - sii902x_i2c_bypass_deselect); - if (!sii902x->i2cmux) - return -ENOMEM; +err_unreg_audio: + if (!PTR_ERR_OR_ZERO(sii902x->audio.pdev)) + platform_device_unregister(sii902x->audio.pdev); - sii902x->i2cmux->priv = sii902x; - return i2c_mux_add_adapter(sii902x->i2cmux, 0, 0, 0); + return ret; } static int sii902x_probe(struct i2c_client *client) @@ -1170,12 +1184,14 @@ static int sii902x_probe(struct i2c_client *client) } static void sii902x_remove(struct i2c_client *client) - { struct sii902x *sii902x = i2c_get_clientdata(client); - i2c_mux_del_adapters(sii902x->i2cmux); drm_bridge_remove(&sii902x->bridge); + i2c_mux_del_adapters(sii902x->i2cmux); + + if (!PTR_ERR_OR_ZERO(sii902x->audio.pdev)) + platform_device_unregister(sii902x->audio.pdev); } static const struct of_device_id sii902x_dt_ids[] = { diff --git a/drivers/gpu/drm/display/drm_dp_mst_topology.c b/drivers/gpu/drm/display/drm_dp_mst_topology.c index bd6c24d4213c..f7c6b60629c2 100644 --- a/drivers/gpu/drm/display/drm_dp_mst_topology.c +++ b/drivers/gpu/drm/display/drm_dp_mst_topology.c @@ -5491,6 +5491,7 @@ EXPORT_SYMBOL(drm_dp_mst_atomic_enable_dsc); * - 0 if the new state is valid * - %-ENOSPC, if the new state is invalid, because of BW limitation * @failing_port is set to: + * * - The non-root port where a BW limit check failed * with all the ports downstream of @failing_port passing * the BW limit check. @@ -5499,6 +5500,7 @@ EXPORT_SYMBOL(drm_dp_mst_atomic_enable_dsc); * - %NULL if the BW limit check failed at the root port * with all the ports downstream of the root port passing * the BW limit check. + * * - %-EINVAL, if the new state is invalid, because the root port has * too many payloads. */ diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index 776f2f0b602d..0ef7bc8848b0 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -319,9 +319,9 @@ static void decon_win_set_bldmod(struct decon_context *ctx, unsigned int win, static void decon_win_set_pixfmt(struct decon_context *ctx, unsigned int win, struct drm_framebuffer *fb) { - struct exynos_drm_plane plane = ctx->planes[win]; + struct exynos_drm_plane *plane = &ctx->planes[win]; struct exynos_drm_plane_state *state = - to_exynos_plane_state(plane.base.state); + to_exynos_plane_state(plane->base.state); unsigned int alpha = state->base.alpha; unsigned int pixel_alpha; unsigned long val; diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c index a9f1c5c05894..f2145227a1e0 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c @@ -480,7 +480,7 @@ static void fimd_commit(struct exynos_drm_crtc *crtc) struct fimd_context *ctx = crtc->ctx; struct drm_display_mode *mode = &crtc->base.state->adjusted_mode; const struct fimd_driver_data *driver_data = ctx->driver_data; - void *timing_base = ctx->regs + driver_data->timing_base; + void __iomem *timing_base = ctx->regs + driver_data->timing_base; u32 val; if (ctx->suspended) @@ -661,9 +661,9 @@ static void fimd_win_set_bldmod(struct fimd_context *ctx, unsigned int win, static void fimd_win_set_pixfmt(struct fimd_context *ctx, unsigned int win, struct drm_framebuffer *fb, int width) { - struct exynos_drm_plane plane = ctx->planes[win]; + struct exynos_drm_plane *plane = &ctx->planes[win]; struct exynos_drm_plane_state *state = - to_exynos_plane_state(plane.base.state); + to_exynos_plane_state(plane->base.state); uint32_t pixel_format = fb->format->format; unsigned int alpha = state->base.alpha; u32 val = WINCONx_ENWIN; diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c index e9a769590415..180507a47700 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c +++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c @@ -1341,7 +1341,7 @@ static int __maybe_unused gsc_runtime_resume(struct device *dev) for (i = 0; i < ctx->num_clocks; i++) { ret = clk_prepare_enable(ctx->clocks[i]); if (ret) { - while (--i > 0) + while (--i >= 0) clk_disable_unprepare(ctx->clocks[i]); return ret; } diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index e777686190ca..c13f14edb508 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -17,7 +17,6 @@ subdir-ccflags-y += $(call cc-option, -Wunused-const-variable) subdir-ccflags-y += $(call cc-option, -Wpacked-not-aligned) subdir-ccflags-y += $(call cc-option, -Wformat-overflow) subdir-ccflags-y += $(call cc-option, -Wformat-truncation) -subdir-ccflags-y += $(call cc-option, -Wstringop-overflow) subdir-ccflags-y += $(call cc-option, -Wstringop-truncation) # The following turn off the warnings enabled by -Wextra ifeq ($(findstring 2, $(KBUILD_EXTRA_WARN)),) diff --git a/drivers/gpu/drm/i915/display/icl_dsi.c b/drivers/gpu/drm/i915/display/icl_dsi.c index ac456a2275db..eda4a8b88590 100644 --- a/drivers/gpu/drm/i915/display/icl_dsi.c +++ b/drivers/gpu/drm/i915/display/icl_dsi.c @@ -1155,6 +1155,7 @@ static void gen11_dsi_powerup_panel(struct intel_encoder *encoder) } intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_INIT_OTP); + intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON); /* ensure all panel commands dispatched before enabling transcoder */ wait_for_cmds_dispatched_to_panel(encoder); @@ -1255,8 +1256,6 @@ static void gen11_dsi_enable(struct intel_atomic_state *state, /* step6d: enable dsi transcoder */ gen11_dsi_enable_transcoder(encoder); - intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON); - /* step7: enable backlight */ intel_backlight_enable(crtc_state, conn_state); intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_ON); diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index 8f702c3fc62d..57bbf3e3af92 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -1525,8 +1525,18 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, * can rely on frontbuffer tracking. */ mask = EDP_PSR_DEBUG_MASK_MEMUP | - EDP_PSR_DEBUG_MASK_HPD | - EDP_PSR_DEBUG_MASK_LPSP; + EDP_PSR_DEBUG_MASK_HPD; + + /* + * For some unknown reason on HSW non-ULT (or at least on + * Dell Latitude E6540) external displays start to flicker + * when PSR is enabled on the eDP. SR/PC6 residency is much + * higher than should be possible with an external display. + * As a workaround leave LPSP unmasked to prevent PSR entry + * when external displays are active. + */ + if (DISPLAY_VER(dev_priv) >= 8 || IS_HASWELL_ULT(dev_priv)) + mask |= EDP_PSR_DEBUG_MASK_LPSP; if (DISPLAY_VER(dev_priv) < 20) mask |= EDP_PSR_DEBUG_MASK_MAX_SLEEP; diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c index 5057d976fa57..93f08f9479d8 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -62,7 +62,7 @@ nouveau_fence_signal(struct nouveau_fence *fence) if (test_bit(DMA_FENCE_FLAG_USER_BITS, &fence->base.flags)) { struct nouveau_fence_chan *fctx = nouveau_fctx(fence); - if (atomic_dec_and_test(&fctx->notify_ref)) + if (!--fctx->notify_ref) drop = 1; } @@ -103,7 +103,7 @@ nouveau_fence_context_kill(struct nouveau_fence_chan *fctx, int error) void nouveau_fence_context_del(struct nouveau_fence_chan *fctx) { - cancel_work_sync(&fctx->allow_block_work); + cancel_work_sync(&fctx->uevent_work); nouveau_fence_context_kill(fctx, 0); nvif_event_dtor(&fctx->event); fctx->dead = 1; @@ -146,12 +146,13 @@ nouveau_fence_update(struct nouveau_channel *chan, struct nouveau_fence_chan *fc return drop; } -static int -nouveau_fence_wait_uevent_handler(struct nvif_event *event, void *repv, u32 repc) +static void +nouveau_fence_uevent_work(struct work_struct *work) { - struct nouveau_fence_chan *fctx = container_of(event, typeof(*fctx), event); + struct nouveau_fence_chan *fctx = container_of(work, struct nouveau_fence_chan, + uevent_work); unsigned long flags; - int ret = NVIF_EVENT_KEEP; + int drop = 0; spin_lock_irqsave(&fctx->lock, flags); if (!list_empty(&fctx->pending)) { @@ -161,23 +162,20 @@ nouveau_fence_wait_uevent_handler(struct nvif_event *event, void *repv, u32 repc fence = list_entry(fctx->pending.next, typeof(*fence), head); chan = rcu_dereference_protected(fence->channel, lockdep_is_held(&fctx->lock)); if (nouveau_fence_update(chan, fctx)) - ret = NVIF_EVENT_DROP; + drop = 1; } - spin_unlock_irqrestore(&fctx->lock, flags); + if (drop) + nvif_event_block(&fctx->event); - return ret; + spin_unlock_irqrestore(&fctx->lock, flags); } -static void -nouveau_fence_work_allow_block(struct work_struct *work) +static int +nouveau_fence_wait_uevent_handler(struct nvif_event *event, void *repv, u32 repc) { - struct nouveau_fence_chan *fctx = container_of(work, struct nouveau_fence_chan, - allow_block_work); - - if (atomic_read(&fctx->notify_ref) == 0) - nvif_event_block(&fctx->event); - else - nvif_event_allow(&fctx->event); + struct nouveau_fence_chan *fctx = container_of(event, typeof(*fctx), event); + schedule_work(&fctx->uevent_work); + return NVIF_EVENT_KEEP; } void @@ -191,7 +189,7 @@ nouveau_fence_context_new(struct nouveau_channel *chan, struct nouveau_fence_cha } args; int ret; - INIT_WORK(&fctx->allow_block_work, nouveau_fence_work_allow_block); + INIT_WORK(&fctx->uevent_work, nouveau_fence_uevent_work); INIT_LIST_HEAD(&fctx->flip); INIT_LIST_HEAD(&fctx->pending); spin_lock_init(&fctx->lock); @@ -535,19 +533,15 @@ static bool nouveau_fence_enable_signaling(struct dma_fence *f) struct nouveau_fence *fence = from_fence(f); struct nouveau_fence_chan *fctx = nouveau_fctx(fence); bool ret; - bool do_work; - if (atomic_inc_return(&fctx->notify_ref) == 0) - do_work = true; + if (!fctx->notify_ref++) + nvif_event_allow(&fctx->event); ret = nouveau_fence_no_signaling(f); if (ret) set_bit(DMA_FENCE_FLAG_USER_BITS, &fence->base.flags); - else if (atomic_dec_and_test(&fctx->notify_ref)) - do_work = true; - - if (do_work) - schedule_work(&fctx->allow_block_work); + else if (!--fctx->notify_ref) + nvif_event_block(&fctx->event); return ret; } diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h index 28f5cf013b89..8bc065acfe35 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.h +++ b/drivers/gpu/drm/nouveau/nouveau_fence.h @@ -3,7 +3,6 @@ #define __NOUVEAU_FENCE_H__ #include <linux/dma-fence.h> -#include <linux/workqueue.h> #include <nvif/event.h> struct nouveau_drm; @@ -45,10 +44,9 @@ struct nouveau_fence_chan { u32 context; char name[32]; + struct work_struct uevent_work; struct nvif_event event; - struct work_struct allow_block_work; - atomic_t notify_ref; - int dead, killed; + int notify_ref, dead, killed; }; struct nouveau_fence_priv { diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/r535.c index 9ee58e2a0eb2..5e1fa176aac4 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/r535.c @@ -1078,7 +1078,6 @@ r535_gsp_rpc_set_registry(struct nvkm_gsp *gsp) if (IS_ERR(rpc)) return PTR_ERR(rpc); - rpc->size = sizeof(*rpc); rpc->numEntries = NV_GSP_REG_NUM_ENTRIES; str_offset = offsetof(typeof(*rpc), entries[NV_GSP_REG_NUM_ENTRIES]); @@ -1094,6 +1093,7 @@ r535_gsp_rpc_set_registry(struct nvkm_gsp *gsp) strings += name_len; str_offset += name_len; } + rpc->size = str_offset; return nvkm_gsp_rpc_wr(gsp, rpc, false); } diff --git a/drivers/gpu/drm/panel/Kconfig b/drivers/gpu/drm/panel/Kconfig index dad938cf6dec..8f3783742208 100644 --- a/drivers/gpu/drm/panel/Kconfig +++ b/drivers/gpu/drm/panel/Kconfig @@ -539,6 +539,8 @@ config DRM_PANEL_RAYDIUM_RM692E5 depends on OF depends on DRM_MIPI_DSI depends on BACKLIGHT_CLASS_DEVICE + select DRM_DISPLAY_DP_HELPER + select DRM_DISPLAY_HELPER help Say Y here if you want to enable support for Raydium RM692E5-based display panels, such as the one found in the Fairphone 5 smartphone. diff --git a/drivers/gpu/drm/panel/panel-samsung-s6d7aa0.c b/drivers/gpu/drm/panel/panel-samsung-s6d7aa0.c index ea5a85779382..f23d8832a1ad 100644 --- a/drivers/gpu/drm/panel/panel-samsung-s6d7aa0.c +++ b/drivers/gpu/drm/panel/panel-samsung-s6d7aa0.c @@ -309,7 +309,7 @@ static const struct s6d7aa0_panel_desc s6d7aa0_lsl080al02_desc = { .off_func = s6d7aa0_lsl080al02_off, .drm_mode = &s6d7aa0_lsl080al02_mode, .mode_flags = MIPI_DSI_MODE_VSYNC_FLUSH | MIPI_DSI_MODE_VIDEO_NO_HFP, - .bus_flags = DRM_BUS_FLAG_DE_HIGH, + .bus_flags = 0, .has_backlight = false, .use_passwd3 = false, diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index 2214cb09678c..d493ee735c73 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -3948,6 +3948,7 @@ static const struct panel_desc tianma_tm070jdhg30 = { }, .bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_SPWG, .connector_type = DRM_MODE_CONNECTOR_LVDS, + .bus_flags = DRM_BUS_FLAG_DE_HIGH, }; static const struct panel_desc tianma_tm070jvhg33 = { @@ -3960,6 +3961,7 @@ static const struct panel_desc tianma_tm070jvhg33 = { }, .bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_SPWG, .connector_type = DRM_MODE_CONNECTOR_LVDS, + .bus_flags = DRM_BUS_FLAG_DE_HIGH, }; static const struct display_timing tianma_tm070rvhg71_timing = { diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 550492a7a031..85f082396d42 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -1178,21 +1178,20 @@ static void drm_sched_run_job_work(struct work_struct *w) struct drm_sched_entity *entity; struct dma_fence *fence; struct drm_sched_fence *s_fence; - struct drm_sched_job *sched_job; + struct drm_sched_job *sched_job = NULL; int r; if (READ_ONCE(sched->pause_submit)) return; - entity = drm_sched_select_entity(sched); + /* Find entity with a ready job */ + while (!sched_job && (entity = drm_sched_select_entity(sched))) { + sched_job = drm_sched_entity_pop_job(entity); + if (!sched_job) + complete_all(&entity->entity_idle); + } if (!entity) - return; - - sched_job = drm_sched_entity_pop_job(entity); - if (!sched_job) { - complete_all(&entity->entity_idle); return; /* No more work */ - } s_fence = sched_job->s_fence; diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index ff36171c8fb7..a73cff7a3070 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -960,7 +960,8 @@ int host1x_client_iommu_attach(struct host1x_client *client) * not the shared IOMMU domain, don't try to attach it to a different * domain. This allows using the IOMMU-backed DMA API. */ - if (domain && domain != tegra->domain) + if (domain && domain->type != IOMMU_DOMAIN_IDENTITY && + domain != tegra->domain) return 0; if (tegra->domain) { diff --git a/drivers/gpu/drm/tests/drm_mm_test.c b/drivers/gpu/drm/tests/drm_mm_test.c index 4e9247cf9977..1eb0c304f960 100644 --- a/drivers/gpu/drm/tests/drm_mm_test.c +++ b/drivers/gpu/drm/tests/drm_mm_test.c @@ -188,13 +188,13 @@ out: static void drm_test_mm_debug(struct kunit *test) { + struct drm_printer p = drm_debug_printer(test->name); struct drm_mm mm; struct drm_mm_node nodes[2]; /* Create a small drm_mm with a couple of nodes and a few holes, and * check that the debug iterator doesn't explode over a trivial drm_mm. */ - drm_mm_init(&mm, 0, 4096); memset(nodes, 0, sizeof(nodes)); @@ -209,6 +209,9 @@ static void drm_test_mm_debug(struct kunit *test) KUNIT_ASSERT_FALSE_MSG(test, drm_mm_reserve_node(&mm, &nodes[1]), "failed to reserve node[0] {start=%lld, size=%lld)\n", nodes[0].start, nodes[0].size); + + drm_mm_print(&mm, &p); + KUNIT_SUCCEED(test); } static bool expect_insert(struct kunit *test, struct drm_mm *mm, diff --git a/drivers/gpu/drm/ttm/ttm_device.c b/drivers/gpu/drm/ttm/ttm_device.c index f5187b384ae9..76027960054f 100644 --- a/drivers/gpu/drm/ttm/ttm_device.c +++ b/drivers/gpu/drm/ttm/ttm_device.c @@ -95,11 +95,17 @@ static int ttm_global_init(void) ttm_pool_mgr_init(num_pages); ttm_tt_mgr_init(num_pages, num_dma32); - glob->dummy_read_page = alloc_page(__GFP_ZERO | GFP_DMA32); + glob->dummy_read_page = alloc_page(__GFP_ZERO | GFP_DMA32 | + __GFP_NOWARN); + /* Retry without GFP_DMA32 for platforms DMA32 is not available */ if (unlikely(glob->dummy_read_page == NULL)) { - ret = -ENOMEM; - goto out; + glob->dummy_read_page = alloc_page(__GFP_ZERO); + if (unlikely(glob->dummy_read_page == NULL)) { + ret = -ENOMEM; + goto out; + } + pr_warn("Using GFP_DMA32 fallback for dummy_read_page\n"); } INIT_LIST_HEAD(&glob->device_list); @@ -195,7 +201,7 @@ int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *func bool use_dma_alloc, bool use_dma32) { struct ttm_global *glob = &ttm_glob; - int ret; + int ret, nid; if (WARN_ON(vma_manager == NULL)) return -EINVAL; @@ -215,7 +221,12 @@ int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *func ttm_sys_man_init(bdev); - ttm_pool_init(&bdev->pool, dev, dev_to_node(dev), use_dma_alloc, use_dma32); + if (dev) + nid = dev_to_node(dev); + else + nid = NUMA_NO_NODE; + + ttm_pool_init(&bdev->pool, dev, nid, use_dma_alloc, use_dma32); bdev->vma_manager = vma_manager; spin_lock_init(&bdev->lru_lock); diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index fcff41dd2315..88f63d526b22 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -147,6 +147,13 @@ v3d_job_allocate(void **container, size_t size) return 0; } +static void +v3d_job_deallocate(void **container) +{ + kfree(*container); + *container = NULL; +} + static int v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, struct v3d_job *job, void (*free)(struct kref *ref), @@ -273,8 +280,10 @@ v3d_setup_csd_jobs_and_bos(struct drm_file *file_priv, ret = v3d_job_init(v3d, file_priv, &(*job)->base, v3d_job_free, args->in_sync, se, V3D_CSD); - if (ret) + if (ret) { + v3d_job_deallocate((void *)job); return ret; + } ret = v3d_job_allocate((void *)clean_job, sizeof(**clean_job)); if (ret) @@ -282,8 +291,10 @@ v3d_setup_csd_jobs_and_bos(struct drm_file *file_priv, ret = v3d_job_init(v3d, file_priv, *clean_job, v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); - if (ret) + if (ret) { + v3d_job_deallocate((void *)clean_job); return ret; + } (*job)->args = *args; @@ -860,8 +871,10 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data, ret = v3d_job_init(v3d, file_priv, &render->base, v3d_render_job_free, args->in_sync_rcl, &se, V3D_RENDER); - if (ret) + if (ret) { + v3d_job_deallocate((void *)&render); goto fail; + } render->start = args->rcl_start; render->end = args->rcl_end; @@ -874,8 +887,10 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data, ret = v3d_job_init(v3d, file_priv, &bin->base, v3d_job_free, args->in_sync_bcl, &se, V3D_BIN); - if (ret) + if (ret) { + v3d_job_deallocate((void *)&bin); goto fail; + } bin->start = args->bcl_start; bin->end = args->bcl_end; @@ -892,8 +907,10 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data, ret = v3d_job_init(v3d, file_priv, clean_job, v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); - if (ret) + if (ret) { + v3d_job_deallocate((void *)&clean_job); goto fail; + } last_job = clean_job; } else { @@ -1015,8 +1032,10 @@ v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, ret = v3d_job_init(v3d, file_priv, &job->base, v3d_job_free, args->in_sync, &se, V3D_TFU); - if (ret) + if (ret) { + v3d_job_deallocate((void *)&job); goto fail; + } job->base.bo = kcalloc(ARRAY_SIZE(args->bo_handles), sizeof(*job->base.bo), GFP_KERNEL); @@ -1233,8 +1252,10 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, ret = v3d_job_init(v3d, file_priv, &cpu_job->base, v3d_job_free, 0, &se, V3D_CPU); - if (ret) + if (ret) { + v3d_job_deallocate((void *)&cpu_job); goto fail; + } clean_job = cpu_job->indirect_csd.clean_job; csd_job = cpu_job->indirect_csd.job; diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.c b/drivers/gpu/drm/virtio/virtgpu_drv.c index f8e9abe647b9..9539aa28937f 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.c +++ b/drivers/gpu/drm/virtio/virtgpu_drv.c @@ -94,6 +94,7 @@ static int virtio_gpu_probe(struct virtio_device *vdev) goto err_free; } + dma_set_max_seg_size(dev->dev, dma_max_mapping_size(dev->dev) ?: UINT_MAX); ret = virtio_gpu_init(vdev, dev); if (ret) goto err_free; diff --git a/drivers/gpu/drm/xe/abi/guc_actions_abi.h b/drivers/gpu/drm/xe/abi/guc_actions_abi.h index 3062e0e0d467..79ba98a169f9 100644 --- a/drivers/gpu/drm/xe/abi/guc_actions_abi.h +++ b/drivers/gpu/drm/xe/abi/guc_actions_abi.h @@ -50,8 +50,8 @@ #define HOST2GUC_SELF_CFG_REQUEST_MSG_LEN (GUC_HXG_REQUEST_MSG_MIN_LEN + 3u) #define HOST2GUC_SELF_CFG_REQUEST_MSG_0_MBZ GUC_HXG_REQUEST_MSG_0_DATA0 -#define HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_KEY (0xffff << 16) -#define HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_LEN (0xffff << 0) +#define HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_KEY (0xffffu << 16) +#define HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_LEN (0xffffu << 0) #define HOST2GUC_SELF_CFG_REQUEST_MSG_2_VALUE32 GUC_HXG_REQUEST_MSG_n_DATAn #define HOST2GUC_SELF_CFG_REQUEST_MSG_3_VALUE64 GUC_HXG_REQUEST_MSG_n_DATAn diff --git a/drivers/gpu/drm/xe/abi/guc_actions_slpc_abi.h b/drivers/gpu/drm/xe/abi/guc_actions_slpc_abi.h index 811add10c30d..c165e26c0976 100644 --- a/drivers/gpu/drm/xe/abi/guc_actions_slpc_abi.h +++ b/drivers/gpu/drm/xe/abi/guc_actions_slpc_abi.h @@ -242,8 +242,8 @@ struct slpc_shared_data { (HOST2GUC_PC_SLPC_REQUEST_REQUEST_MSG_MIN_LEN + \ HOST2GUC_PC_SLPC_EVENT_MAX_INPUT_ARGS) #define HOST2GUC_PC_SLPC_REQUEST_MSG_0_MBZ GUC_HXG_REQUEST_MSG_0_DATA0 -#define HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ID (0xff << 8) -#define HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ARGC (0xff << 0) +#define HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ID (0xffu << 8) +#define HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ARGC (0xffu << 0) #define HOST2GUC_PC_SLPC_REQUEST_MSG_N_EVENT_DATA_N GUC_HXG_REQUEST_MSG_n_DATAn #endif diff --git a/drivers/gpu/drm/xe/abi/guc_communication_ctb_abi.h b/drivers/gpu/drm/xe/abi/guc_communication_ctb_abi.h index 3b83f907ece4..0b1146d0c997 100644 --- a/drivers/gpu/drm/xe/abi/guc_communication_ctb_abi.h +++ b/drivers/gpu/drm/xe/abi/guc_communication_ctb_abi.h @@ -82,11 +82,11 @@ static_assert(sizeof(struct guc_ct_buffer_desc) == 64); #define GUC_CTB_HDR_LEN 1u #define GUC_CTB_MSG_MIN_LEN GUC_CTB_HDR_LEN #define GUC_CTB_MSG_MAX_LEN 256u -#define GUC_CTB_MSG_0_FENCE (0xffff << 16) -#define GUC_CTB_MSG_0_FORMAT (0xf << 12) +#define GUC_CTB_MSG_0_FENCE (0xffffu << 16) +#define GUC_CTB_MSG_0_FORMAT (0xfu << 12) #define GUC_CTB_FORMAT_HXG 0u -#define GUC_CTB_MSG_0_RESERVED (0xf << 8) -#define GUC_CTB_MSG_0_NUM_DWORDS (0xff << 0) +#define GUC_CTB_MSG_0_RESERVED (0xfu << 8) +#define GUC_CTB_MSG_0_NUM_DWORDS (0xffu << 0) /** * DOC: CTB HXG Message diff --git a/drivers/gpu/drm/xe/abi/guc_klvs_abi.h b/drivers/gpu/drm/xe/abi/guc_klvs_abi.h index 47094b9b044c..0400bc0fccdc 100644 --- a/drivers/gpu/drm/xe/abi/guc_klvs_abi.h +++ b/drivers/gpu/drm/xe/abi/guc_klvs_abi.h @@ -31,9 +31,9 @@ */ #define GUC_KLV_LEN_MIN 1u -#define GUC_KLV_0_KEY (0xffff << 16) -#define GUC_KLV_0_LEN (0xffff << 0) -#define GUC_KLV_n_VALUE (0xffffffff << 0) +#define GUC_KLV_0_KEY (0xffffu << 16) +#define GUC_KLV_0_LEN (0xffffu << 0) +#define GUC_KLV_n_VALUE (0xffffffffu << 0) /** * DOC: GuC Self Config KLVs diff --git a/drivers/gpu/drm/xe/abi/guc_messages_abi.h b/drivers/gpu/drm/xe/abi/guc_messages_abi.h index 3d199016cf88..29e414c82d56 100644 --- a/drivers/gpu/drm/xe/abi/guc_messages_abi.h +++ b/drivers/gpu/drm/xe/abi/guc_messages_abi.h @@ -40,18 +40,18 @@ */ #define GUC_HXG_MSG_MIN_LEN 1u -#define GUC_HXG_MSG_0_ORIGIN (0x1 << 31) +#define GUC_HXG_MSG_0_ORIGIN (0x1u << 31) #define GUC_HXG_ORIGIN_HOST 0u #define GUC_HXG_ORIGIN_GUC 1u -#define GUC_HXG_MSG_0_TYPE (0x7 << 28) +#define GUC_HXG_MSG_0_TYPE (0x7u << 28) #define GUC_HXG_TYPE_REQUEST 0u #define GUC_HXG_TYPE_EVENT 1u #define GUC_HXG_TYPE_NO_RESPONSE_BUSY 3u #define GUC_HXG_TYPE_NO_RESPONSE_RETRY 5u #define GUC_HXG_TYPE_RESPONSE_FAILURE 6u #define GUC_HXG_TYPE_RESPONSE_SUCCESS 7u -#define GUC_HXG_MSG_0_AUX (0xfffffff << 0) -#define GUC_HXG_MSG_n_PAYLOAD (0xffffffff << 0) +#define GUC_HXG_MSG_0_AUX (0xfffffffu << 0) +#define GUC_HXG_MSG_n_PAYLOAD (0xffffffffu << 0) /** * DOC: HXG Request @@ -85,8 +85,8 @@ */ #define GUC_HXG_REQUEST_MSG_MIN_LEN GUC_HXG_MSG_MIN_LEN -#define GUC_HXG_REQUEST_MSG_0_DATA0 (0xfff << 16) -#define GUC_HXG_REQUEST_MSG_0_ACTION (0xffff << 0) +#define GUC_HXG_REQUEST_MSG_0_DATA0 (0xfffu << 16) +#define GUC_HXG_REQUEST_MSG_0_ACTION (0xffffu << 0) #define GUC_HXG_REQUEST_MSG_n_DATAn GUC_HXG_MSG_n_PAYLOAD /** @@ -117,8 +117,8 @@ */ #define GUC_HXG_EVENT_MSG_MIN_LEN GUC_HXG_MSG_MIN_LEN -#define GUC_HXG_EVENT_MSG_0_DATA0 (0xfff << 16) -#define GUC_HXG_EVENT_MSG_0_ACTION (0xffff << 0) +#define GUC_HXG_EVENT_MSG_0_DATA0 (0xfffu << 16) +#define GUC_HXG_EVENT_MSG_0_ACTION (0xffffu << 0) #define GUC_HXG_EVENT_MSG_n_DATAn GUC_HXG_MSG_n_PAYLOAD /** @@ -188,8 +188,8 @@ */ #define GUC_HXG_FAILURE_MSG_LEN GUC_HXG_MSG_MIN_LEN -#define GUC_HXG_FAILURE_MSG_0_HINT (0xfff << 16) -#define GUC_HXG_FAILURE_MSG_0_ERROR (0xffff << 0) +#define GUC_HXG_FAILURE_MSG_0_HINT (0xfffu << 16) +#define GUC_HXG_FAILURE_MSG_0_ERROR (0xffffu << 0) /** * DOC: HXG Response diff --git a/drivers/gpu/drm/xe/compat-i915-headers/gem/i915_gem_object.h b/drivers/gpu/drm/xe/compat-i915-headers/gem/i915_gem_object.h index 5f19550cc845..68d9f6116bdf 100644 --- a/drivers/gpu/drm/xe/compat-i915-headers/gem/i915_gem_object.h +++ b/drivers/gpu/drm/xe/compat-i915-headers/gem/i915_gem_object.h @@ -35,12 +35,10 @@ static inline int i915_gem_object_read_from_page(struct xe_bo *bo, u32 ofs, u64 *ptr, u32 size) { struct ttm_bo_kmap_obj map; - void *virtual; + void *src; bool is_iomem; int ret; - XE_WARN_ON(size != 8); - ret = xe_bo_lock(bo, true); if (ret) return ret; @@ -50,11 +48,12 @@ static inline int i915_gem_object_read_from_page(struct xe_bo *bo, goto out_unlock; ofs &= ~PAGE_MASK; - virtual = ttm_kmap_obj_virtual(&map, &is_iomem); + src = ttm_kmap_obj_virtual(&map, &is_iomem); + src += ofs; if (is_iomem) - *ptr = readq((void __iomem *)(virtual + ofs)); + memcpy_fromio(ptr, (void __iomem *)src, size); else - *ptr = *(u64 *)(virtual + ofs); + memcpy(ptr, src, size); ttm_bo_kunmap(&map); out_unlock: diff --git a/drivers/gpu/drm/xe/tests/xe_wa_test.c b/drivers/gpu/drm/xe/tests/xe_wa_test.c index a53c22a19582..b4715b78ef3b 100644 --- a/drivers/gpu/drm/xe/tests/xe_wa_test.c +++ b/drivers/gpu/drm/xe/tests/xe_wa_test.c @@ -74,9 +74,6 @@ static const struct platform_test_case cases[] = { SUBPLATFORM_CASE(DG2, G11, B1), SUBPLATFORM_CASE(DG2, G12, A0), SUBPLATFORM_CASE(DG2, G12, A1), - PLATFORM_CASE(PVC, B0), - PLATFORM_CASE(PVC, B1), - PLATFORM_CASE(PVC, C0), GMDID_CASE(METEORLAKE, 1270, A0, 1300, A0), GMDID_CASE(METEORLAKE, 1271, A0, 1300, A0), GMDID_CASE(LUNARLAKE, 2004, A0, 2000, A0), diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index b8d8da546670..1f0b4b9ce84f 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -613,7 +613,7 @@ void xe_device_wmb(struct xe_device *xe) u32 xe_device_ccs_bytes(struct xe_device *xe, u64 size) { return xe_device_has_flat_ccs(xe) ? - DIV_ROUND_UP(size, NUM_BYTES_PER_CCS_BYTE(xe)) : 0; + DIV_ROUND_UP_ULL(size, NUM_BYTES_PER_CCS_BYTE(xe)) : 0; } bool xe_device_mem_access_ongoing(struct xe_device *xe) diff --git a/drivers/gpu/drm/xe/xe_dma_buf.c b/drivers/gpu/drm/xe/xe_dma_buf.c index 64ed303728fd..da2627ed6ae7 100644 --- a/drivers/gpu/drm/xe/xe_dma_buf.c +++ b/drivers/gpu/drm/xe/xe_dma_buf.c @@ -175,7 +175,7 @@ static int xe_dma_buf_begin_cpu_access(struct dma_buf *dma_buf, return 0; } -const struct dma_buf_ops xe_dmabuf_ops = { +static const struct dma_buf_ops xe_dmabuf_ops = { .attach = xe_dma_buf_attach, .detach = xe_dma_buf_detach, .pin = xe_dma_buf_pin, diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index b853feed9ccc..17f26952e665 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -111,7 +111,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) u64 addresses[XE_HW_ENGINE_MAX_INSTANCE]; struct drm_gpuvm_exec vm_exec = {.extra.fn = xe_exec_fn}; struct drm_exec *exec = &vm_exec.exec; - u32 i, num_syncs = 0; + u32 i, num_syncs = 0, num_ufence = 0; struct xe_sched_job *job; struct dma_fence *rebind_fence; struct xe_vm *vm; @@ -157,6 +157,14 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) SYNC_PARSE_FLAG_LR_MODE : 0)); if (err) goto err_syncs; + + if (xe_sync_is_ufence(&syncs[i])) + num_ufence++; + } + + if (XE_IOCTL_DBG(xe, num_ufence > 1)) { + err = -EINVAL; + goto err_syncs; } if (xe_exec_queue_is_parallel(q)) { diff --git a/drivers/gpu/drm/xe/xe_gt_mcr.c b/drivers/gpu/drm/xe/xe_gt_mcr.c index 77925b35cf8d..8546cd3cc50d 100644 --- a/drivers/gpu/drm/xe/xe_gt_mcr.c +++ b/drivers/gpu/drm/xe/xe_gt_mcr.c @@ -480,7 +480,7 @@ static bool xe_gt_mcr_get_nonterminated_steering(struct xe_gt *gt, * to synchronize with external clients (e.g., firmware), so a semaphore * register will also need to be taken. */ -static void mcr_lock(struct xe_gt *gt) +static void mcr_lock(struct xe_gt *gt) __acquires(>->mcr_lock) { struct xe_device *xe = gt_to_xe(gt); int ret = 0; @@ -500,7 +500,7 @@ static void mcr_lock(struct xe_gt *gt) drm_WARN_ON_ONCE(&xe->drm, ret == -ETIMEDOUT); } -static void mcr_unlock(struct xe_gt *gt) +static void mcr_unlock(struct xe_gt *gt) __releases(>->mcr_lock) { /* Release hardware semaphore - this is done by writing 1 to the register */ if (GRAPHICS_VERx100(gt_to_xe(gt)) >= 1270) diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c b/drivers/gpu/drm/xe/xe_gt_pagefault.c index 59a70d2e0a7a..9c2fe1697d6e 100644 --- a/drivers/gpu/drm/xe/xe_gt_pagefault.c +++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c @@ -165,7 +165,8 @@ retry_userptr: goto unlock_vm; } - if (!xe_vma_is_userptr(vma) || !xe_vma_userptr_check_repin(vma)) { + if (!xe_vma_is_userptr(vma) || + !xe_vma_userptr_check_repin(to_userptr_vma(vma))) { downgrade_write(&vm->lock); write_locked = false; } @@ -181,11 +182,13 @@ retry_userptr: /* TODO: Validate fault */ if (xe_vma_is_userptr(vma) && write_locked) { + struct xe_userptr_vma *uvma = to_userptr_vma(vma); + spin_lock(&vm->userptr.invalidated_lock); - list_del_init(&vma->userptr.invalidate_link); + list_del_init(&uvma->userptr.invalidate_link); spin_unlock(&vm->userptr.invalidated_lock); - ret = xe_vma_userptr_pin_pages(vma); + ret = xe_vma_userptr_pin_pages(uvma); if (ret) goto unlock_vm; @@ -220,7 +223,7 @@ retry_userptr: dma_fence_put(fence); if (xe_vma_is_userptr(vma)) - ret = xe_vma_userptr_check_repin(vma); + ret = xe_vma_userptr_check_repin(to_userptr_vma(vma)); vma->usm.tile_invalidated &= ~BIT(tile->id); unlock_dma_resv: diff --git a/drivers/gpu/drm/xe/xe_guc_pc.c b/drivers/gpu/drm/xe/xe_guc_pc.c index f71085228cb3..d91702592520 100644 --- a/drivers/gpu/drm/xe/xe_guc_pc.c +++ b/drivers/gpu/drm/xe/xe_guc_pc.c @@ -963,7 +963,9 @@ void xe_guc_pc_fini(struct xe_guc_pc *pc) struct xe_device *xe = pc_to_xe(pc); if (xe->info.skip_guc_pc) { + xe_device_mem_access_get(xe); xe_gt_idle_disable_c6(pc_to_gt(pc)); + xe_device_mem_access_put(xe); return; } diff --git a/drivers/gpu/drm/xe/xe_hw_fence.c b/drivers/gpu/drm/xe/xe_hw_fence.c index a6094c81f2ad..a5de3e7b0bd6 100644 --- a/drivers/gpu/drm/xe/xe_hw_fence.c +++ b/drivers/gpu/drm/xe/xe_hw_fence.c @@ -217,13 +217,13 @@ struct xe_hw_fence *xe_hw_fence_create(struct xe_hw_fence_ctx *ctx, if (!fence) return ERR_PTR(-ENOMEM); - dma_fence_init(&fence->dma, &xe_hw_fence_ops, &ctx->irq->lock, - ctx->dma_fence_ctx, ctx->next_seqno++); - fence->ctx = ctx; fence->seqno_map = seqno_map; INIT_LIST_HEAD(&fence->irq_link); + dma_fence_init(&fence->dma, &xe_hw_fence_ops, &ctx->irq->lock, + ctx->dma_fence_ctx, ctx->next_seqno++); + trace_xe_hw_fence_create(fence); return fence; diff --git a/drivers/gpu/drm/xe/xe_hwmon.c b/drivers/gpu/drm/xe/xe_hwmon.c index 6ef2aa1eae8b..174ed2185481 100644 --- a/drivers/gpu/drm/xe/xe_hwmon.c +++ b/drivers/gpu/drm/xe/xe_hwmon.c @@ -419,7 +419,7 @@ static int xe_hwmon_pcode_read_i1(struct xe_gt *gt, u32 *uval) return xe_pcode_read(gt, PCODE_MBOX(PCODE_POWER_SETUP, POWER_SETUP_SUBCOMMAND_READ_I1, 0), - uval, 0); + uval, NULL); } static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval) diff --git a/drivers/gpu/drm/xe/xe_lrc.c b/drivers/gpu/drm/xe/xe_lrc.c index b7fa3831b684..0ec5ad2539f1 100644 --- a/drivers/gpu/drm/xe/xe_lrc.c +++ b/drivers/gpu/drm/xe/xe_lrc.c @@ -21,10 +21,10 @@ #include "xe_map.h" #include "xe_vm.h" -#define CTX_VALID (1 << 0) -#define CTX_PRIVILEGE (1 << 8) -#define CTX_ADDRESSING_MODE_SHIFT 3 -#define LEGACY_64B_CONTEXT 3 +#define LRC_VALID (1 << 0) +#define LRC_PRIVILEGE (1 << 8) +#define LRC_ADDRESSING_MODE_SHIFT 3 +#define LRC_LEGACY_64B_CONTEXT 3 #define ENGINE_CLASS_SHIFT 61 #define ENGINE_INSTANCE_SHIFT 48 @@ -762,15 +762,15 @@ int xe_lrc_init(struct xe_lrc *lrc, struct xe_hw_engine *hwe, (q->usm.acc_notify << ACC_NOTIFY_S) | q->usm.acc_trigger); - lrc->desc = CTX_VALID; - lrc->desc |= LEGACY_64B_CONTEXT << CTX_ADDRESSING_MODE_SHIFT; + lrc->desc = LRC_VALID; + lrc->desc |= LRC_LEGACY_64B_CONTEXT << LRC_ADDRESSING_MODE_SHIFT; /* TODO: Priority */ /* While this appears to have something about privileged batches or * some such, it really just means PPGTT mode. */ if (vm) - lrc->desc |= CTX_PRIVILEGE; + lrc->desc |= LRC_PRIVILEGE; if (GRAPHICS_VERx100(xe) < 1250) { lrc->desc |= (u64)hwe->instance << ENGINE_INSTANCE_SHIFT; diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index e05e9e7282b6..5c6c54624252 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -472,7 +472,7 @@ static void emit_pte(struct xe_migrate *m, /* Indirect access needs compression enabled uncached PAT index */ if (GRAPHICS_VERx100(xe) >= 2000) pat_index = is_comp_pte ? xe->pat.idx[XE_CACHE_NONE_COMPRESSION] : - xe->pat.idx[XE_CACHE_NONE]; + xe->pat.idx[XE_CACHE_WB]; else pat_index = xe->pat.idx[XE_CACHE_WB]; @@ -760,14 +760,14 @@ struct dma_fence *xe_migrate_copy(struct xe_migrate *m, if (src_is_vram && xe_migrate_allow_identity(src_L0, &src_it)) xe_res_next(&src_it, src_L0); else - emit_pte(m, bb, src_L0_pt, src_is_vram, true, &src_it, src_L0, - src); + emit_pte(m, bb, src_L0_pt, src_is_vram, copy_system_ccs, + &src_it, src_L0, src); if (dst_is_vram && xe_migrate_allow_identity(src_L0, &dst_it)) xe_res_next(&dst_it, src_L0); else - emit_pte(m, bb, dst_L0_pt, dst_is_vram, true, &dst_it, src_L0, - dst); + emit_pte(m, bb, dst_L0_pt, dst_is_vram, copy_system_ccs, + &dst_it, src_L0, dst); if (copy_system_ccs) emit_pte(m, bb, ccs_pt, false, false, &ccs_it, ccs_size, src); @@ -1009,8 +1009,8 @@ struct dma_fence *xe_migrate_clear(struct xe_migrate *m, if (clear_vram && xe_migrate_allow_identity(clear_L0, &src_it)) xe_res_next(&src_it, clear_L0); else - emit_pte(m, bb, clear_L0_pt, clear_vram, true, &src_it, clear_L0, - dst); + emit_pte(m, bb, clear_L0_pt, clear_vram, clear_system_ccs, + &src_it, clear_L0, dst); bb->cs[bb->len++] = MI_BATCH_BUFFER_END; update_idx = bb->len; diff --git a/drivers/gpu/drm/xe/xe_mmio.c b/drivers/gpu/drm/xe/xe_mmio.c index c8c5d74b6e90..5f6b53ea5528 100644 --- a/drivers/gpu/drm/xe/xe_mmio.c +++ b/drivers/gpu/drm/xe/xe_mmio.c @@ -272,8 +272,8 @@ int xe_mmio_probe_vram(struct xe_device *xe) drm_info(&xe->drm, "VRAM[%u, %u]: Actual physical size %pa, usable size exclude stolen %pa, CPU accessible size %pa\n", id, tile->id, &tile->mem.vram.actual_physical_size, &tile->mem.vram.usable_size, &tile->mem.vram.io_size); drm_info(&xe->drm, "VRAM[%u, %u]: DPA range: [%pa-%llx], io range: [%pa-%llx]\n", id, tile->id, - &tile->mem.vram.dpa_base, tile->mem.vram.dpa_base + tile->mem.vram.actual_physical_size, - &tile->mem.vram.io_start, tile->mem.vram.io_start + tile->mem.vram.io_size); + &tile->mem.vram.dpa_base, tile->mem.vram.dpa_base + (u64)tile->mem.vram.actual_physical_size, + &tile->mem.vram.io_start, tile->mem.vram.io_start + (u64)tile->mem.vram.io_size); /* calculate total size using tile size to get the correct HW sizing */ total_size += tile_size; diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c index de1030a47588..e45b37c3f0c2 100644 --- a/drivers/gpu/drm/xe/xe_pt.c +++ b/drivers/gpu/drm/xe/xe_pt.c @@ -618,8 +618,8 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma, if (!xe_vma_is_null(vma)) { if (xe_vma_is_userptr(vma)) - xe_res_first_sg(vma->userptr.sg, 0, xe_vma_size(vma), - &curs); + xe_res_first_sg(to_userptr_vma(vma)->userptr.sg, 0, + xe_vma_size(vma), &curs); else if (xe_bo_is_vram(bo) || xe_bo_is_stolen(bo)) xe_res_first(bo->ttm.resource, xe_vma_bo_offset(vma), xe_vma_size(vma), &curs); @@ -906,17 +906,17 @@ static void xe_vm_dbg_print_entries(struct xe_device *xe, #ifdef CONFIG_DRM_XE_USERPTR_INVAL_INJECT -static int xe_pt_userptr_inject_eagain(struct xe_vma *vma) +static int xe_pt_userptr_inject_eagain(struct xe_userptr_vma *uvma) { - u32 divisor = vma->userptr.divisor ? vma->userptr.divisor : 2; + u32 divisor = uvma->userptr.divisor ? uvma->userptr.divisor : 2; static u32 count; if (count++ % divisor == divisor - 1) { - struct xe_vm *vm = xe_vma_vm(vma); + struct xe_vm *vm = xe_vma_vm(&uvma->vma); - vma->userptr.divisor = divisor << 1; + uvma->userptr.divisor = divisor << 1; spin_lock(&vm->userptr.invalidated_lock); - list_move_tail(&vma->userptr.invalidate_link, + list_move_tail(&uvma->userptr.invalidate_link, &vm->userptr.invalidated); spin_unlock(&vm->userptr.invalidated_lock); return true; @@ -927,7 +927,7 @@ static int xe_pt_userptr_inject_eagain(struct xe_vma *vma) #else -static bool xe_pt_userptr_inject_eagain(struct xe_vma *vma) +static bool xe_pt_userptr_inject_eagain(struct xe_userptr_vma *uvma) { return false; } @@ -1000,9 +1000,9 @@ static int xe_pt_userptr_pre_commit(struct xe_migrate_pt_update *pt_update) { struct xe_pt_migrate_pt_update *userptr_update = container_of(pt_update, typeof(*userptr_update), base); - struct xe_vma *vma = pt_update->vma; - unsigned long notifier_seq = vma->userptr.notifier_seq; - struct xe_vm *vm = xe_vma_vm(vma); + struct xe_userptr_vma *uvma = to_userptr_vma(pt_update->vma); + unsigned long notifier_seq = uvma->userptr.notifier_seq; + struct xe_vm *vm = xe_vma_vm(&uvma->vma); int err = xe_pt_vm_dependencies(pt_update->job, &vm->rftree[pt_update->tile_id], pt_update->start, @@ -1023,7 +1023,7 @@ static int xe_pt_userptr_pre_commit(struct xe_migrate_pt_update *pt_update) */ do { down_read(&vm->userptr.notifier_lock); - if (!mmu_interval_read_retry(&vma->userptr.notifier, + if (!mmu_interval_read_retry(&uvma->userptr.notifier, notifier_seq)) break; @@ -1032,11 +1032,11 @@ static int xe_pt_userptr_pre_commit(struct xe_migrate_pt_update *pt_update) if (userptr_update->bind) return -EAGAIN; - notifier_seq = mmu_interval_read_begin(&vma->userptr.notifier); + notifier_seq = mmu_interval_read_begin(&uvma->userptr.notifier); } while (true); /* Inject errors to test_whether they are handled correctly */ - if (userptr_update->bind && xe_pt_userptr_inject_eagain(vma)) { + if (userptr_update->bind && xe_pt_userptr_inject_eagain(uvma)) { up_read(&vm->userptr.notifier_lock); return -EAGAIN; } @@ -1297,7 +1297,7 @@ __xe_pt_bind_vma(struct xe_tile *tile, struct xe_vma *vma, struct xe_exec_queue vma->tile_present |= BIT(tile->id); if (bind_pt_update.locked) { - vma->userptr.initial_bind = true; + to_userptr_vma(vma)->userptr.initial_bind = true; up_read(&vm->userptr.notifier_lock); xe_bo_put_commit(&deferred); } @@ -1642,7 +1642,7 @@ __xe_pt_unbind_vma(struct xe_tile *tile, struct xe_vma *vma, struct xe_exec_queu if (!vma->tile_present) { spin_lock(&vm->userptr.invalidated_lock); - list_del_init(&vma->userptr.invalidate_link); + list_del_init(&to_userptr_vma(vma)->userptr.invalidate_link); spin_unlock(&vm->userptr.invalidated_lock); } up_read(&vm->userptr.notifier_lock); diff --git a/drivers/gpu/drm/xe/xe_query.c b/drivers/gpu/drm/xe/xe_query.c index 9b35673b286c..7e924faeeea0 100644 --- a/drivers/gpu/drm/xe/xe_query.c +++ b/drivers/gpu/drm/xe/xe_query.c @@ -459,21 +459,21 @@ static size_t calc_topo_query_size(struct xe_device *xe) sizeof_field(struct xe_gt, fuse_topo.eu_mask_per_dss)); } -static void __user *copy_mask(void __user *ptr, - struct drm_xe_query_topology_mask *topo, - void *mask, size_t mask_size) +static int copy_mask(void __user **ptr, + struct drm_xe_query_topology_mask *topo, + void *mask, size_t mask_size) { topo->num_bytes = mask_size; - if (copy_to_user(ptr, topo, sizeof(*topo))) - return ERR_PTR(-EFAULT); - ptr += sizeof(topo); + if (copy_to_user(*ptr, topo, sizeof(*topo))) + return -EFAULT; + *ptr += sizeof(topo); - if (copy_to_user(ptr, mask, mask_size)) - return ERR_PTR(-EFAULT); - ptr += mask_size; + if (copy_to_user(*ptr, mask, mask_size)) + return -EFAULT; + *ptr += mask_size; - return ptr; + return 0; } static int query_gt_topology(struct xe_device *xe, @@ -493,28 +493,28 @@ static int query_gt_topology(struct xe_device *xe, } for_each_gt(gt, xe, id) { + int err; + topo.gt_id = id; topo.type = DRM_XE_TOPO_DSS_GEOMETRY; - query_ptr = copy_mask(query_ptr, &topo, - gt->fuse_topo.g_dss_mask, - sizeof(gt->fuse_topo.g_dss_mask)); - if (IS_ERR(query_ptr)) - return PTR_ERR(query_ptr); + err = copy_mask(&query_ptr, &topo, gt->fuse_topo.g_dss_mask, + sizeof(gt->fuse_topo.g_dss_mask)); + if (err) + return err; topo.type = DRM_XE_TOPO_DSS_COMPUTE; - query_ptr = copy_mask(query_ptr, &topo, - gt->fuse_topo.c_dss_mask, - sizeof(gt->fuse_topo.c_dss_mask)); - if (IS_ERR(query_ptr)) - return PTR_ERR(query_ptr); + err = copy_mask(&query_ptr, &topo, gt->fuse_topo.c_dss_mask, + sizeof(gt->fuse_topo.c_dss_mask)); + if (err) + return err; topo.type = DRM_XE_TOPO_EU_PER_DSS; - query_ptr = copy_mask(query_ptr, &topo, - gt->fuse_topo.eu_mask_per_dss, - sizeof(gt->fuse_topo.eu_mask_per_dss)); - if (IS_ERR(query_ptr)) - return PTR_ERR(query_ptr); + err = copy_mask(&query_ptr, &topo, + gt->fuse_topo.eu_mask_per_dss, + sizeof(gt->fuse_topo.eu_mask_per_dss)); + if (err) + return err; } return 0; diff --git a/drivers/gpu/drm/xe/xe_sync.h b/drivers/gpu/drm/xe/xe_sync.h index d284afbe917c..f43cdcaca6c5 100644 --- a/drivers/gpu/drm/xe/xe_sync.h +++ b/drivers/gpu/drm/xe/xe_sync.h @@ -33,4 +33,9 @@ struct dma_fence * xe_sync_in_fence_get(struct xe_sync_entry *sync, int num_sync, struct xe_exec_queue *q, struct xe_vm *vm); +static inline bool xe_sync_is_ufence(struct xe_sync_entry *sync) +{ + return !!sync->ufence; +} + #endif diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 10b6995fbf29..30db264d34a3 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -46,7 +46,7 @@ static struct drm_gem_object *xe_vm_obj(struct xe_vm *vm) /** * xe_vma_userptr_check_repin() - Advisory check for repin needed - * @vma: The userptr vma + * @uvma: The userptr vma * * Check if the userptr vma has been invalidated since last successful * repin. The check is advisory only and can the function can be called @@ -56,15 +56,17 @@ static struct drm_gem_object *xe_vm_obj(struct xe_vm *vm) * * Return: 0 if userptr vma is valid, -EAGAIN otherwise; repin recommended. */ -int xe_vma_userptr_check_repin(struct xe_vma *vma) +int xe_vma_userptr_check_repin(struct xe_userptr_vma *uvma) { - return mmu_interval_check_retry(&vma->userptr.notifier, - vma->userptr.notifier_seq) ? + return mmu_interval_check_retry(&uvma->userptr.notifier, + uvma->userptr.notifier_seq) ? -EAGAIN : 0; } -int xe_vma_userptr_pin_pages(struct xe_vma *vma) +int xe_vma_userptr_pin_pages(struct xe_userptr_vma *uvma) { + struct xe_userptr *userptr = &uvma->userptr; + struct xe_vma *vma = &uvma->vma; struct xe_vm *vm = xe_vma_vm(vma); struct xe_device *xe = vm->xe; const unsigned long num_pages = xe_vma_size(vma) >> PAGE_SHIFT; @@ -80,30 +82,30 @@ retry: if (vma->gpuva.flags & XE_VMA_DESTROYED) return 0; - notifier_seq = mmu_interval_read_begin(&vma->userptr.notifier); - if (notifier_seq == vma->userptr.notifier_seq) + notifier_seq = mmu_interval_read_begin(&userptr->notifier); + if (notifier_seq == userptr->notifier_seq) return 0; pages = kvmalloc_array(num_pages, sizeof(*pages), GFP_KERNEL); if (!pages) return -ENOMEM; - if (vma->userptr.sg) { + if (userptr->sg) { dma_unmap_sgtable(xe->drm.dev, - vma->userptr.sg, + userptr->sg, read_only ? DMA_TO_DEVICE : DMA_BIDIRECTIONAL, 0); - sg_free_table(vma->userptr.sg); - vma->userptr.sg = NULL; + sg_free_table(userptr->sg); + userptr->sg = NULL; } pinned = ret = 0; if (in_kthread) { - if (!mmget_not_zero(vma->userptr.notifier.mm)) { + if (!mmget_not_zero(userptr->notifier.mm)) { ret = -EFAULT; goto mm_closed; } - kthread_use_mm(vma->userptr.notifier.mm); + kthread_use_mm(userptr->notifier.mm); } while (pinned < num_pages) { @@ -123,32 +125,32 @@ retry: } if (in_kthread) { - kthread_unuse_mm(vma->userptr.notifier.mm); - mmput(vma->userptr.notifier.mm); + kthread_unuse_mm(userptr->notifier.mm); + mmput(userptr->notifier.mm); } mm_closed: if (ret) goto out; - ret = sg_alloc_table_from_pages_segment(&vma->userptr.sgt, pages, + ret = sg_alloc_table_from_pages_segment(&userptr->sgt, pages, pinned, 0, (u64)pinned << PAGE_SHIFT, xe_sg_segment_size(xe->drm.dev), GFP_KERNEL); if (ret) { - vma->userptr.sg = NULL; + userptr->sg = NULL; goto out; } - vma->userptr.sg = &vma->userptr.sgt; + userptr->sg = &userptr->sgt; - ret = dma_map_sgtable(xe->drm.dev, vma->userptr.sg, + ret = dma_map_sgtable(xe->drm.dev, userptr->sg, read_only ? DMA_TO_DEVICE : DMA_BIDIRECTIONAL, DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_NO_KERNEL_MAPPING); if (ret) { - sg_free_table(vma->userptr.sg); - vma->userptr.sg = NULL; + sg_free_table(userptr->sg); + userptr->sg = NULL; goto out; } @@ -167,8 +169,8 @@ out: kvfree(pages); if (!(ret < 0)) { - vma->userptr.notifier_seq = notifier_seq; - if (xe_vma_userptr_check_repin(vma) == -EAGAIN) + userptr->notifier_seq = notifier_seq; + if (xe_vma_userptr_check_repin(uvma) == -EAGAIN) goto retry; } @@ -635,7 +637,9 @@ static bool vma_userptr_invalidate(struct mmu_interval_notifier *mni, const struct mmu_notifier_range *range, unsigned long cur_seq) { - struct xe_vma *vma = container_of(mni, struct xe_vma, userptr.notifier); + struct xe_userptr *userptr = container_of(mni, typeof(*userptr), notifier); + struct xe_userptr_vma *uvma = container_of(userptr, typeof(*uvma), userptr); + struct xe_vma *vma = &uvma->vma; struct xe_vm *vm = xe_vma_vm(vma); struct dma_resv_iter cursor; struct dma_fence *fence; @@ -651,7 +655,7 @@ static bool vma_userptr_invalidate(struct mmu_interval_notifier *mni, mmu_interval_set_seq(mni, cur_seq); /* No need to stop gpu access if the userptr is not yet bound. */ - if (!vma->userptr.initial_bind) { + if (!userptr->initial_bind) { up_write(&vm->userptr.notifier_lock); return true; } @@ -663,7 +667,7 @@ static bool vma_userptr_invalidate(struct mmu_interval_notifier *mni, if (!xe_vm_in_fault_mode(vm) && !(vma->gpuva.flags & XE_VMA_DESTROYED) && vma->tile_present) { spin_lock(&vm->userptr.invalidated_lock); - list_move_tail(&vma->userptr.invalidate_link, + list_move_tail(&userptr->invalidate_link, &vm->userptr.invalidated); spin_unlock(&vm->userptr.invalidated_lock); } @@ -703,7 +707,7 @@ static const struct mmu_interval_notifier_ops vma_userptr_notifier_ops = { int xe_vm_userptr_pin(struct xe_vm *vm) { - struct xe_vma *vma, *next; + struct xe_userptr_vma *uvma, *next; int err = 0; LIST_HEAD(tmp_evict); @@ -711,22 +715,23 @@ int xe_vm_userptr_pin(struct xe_vm *vm) /* Collect invalidated userptrs */ spin_lock(&vm->userptr.invalidated_lock); - list_for_each_entry_safe(vma, next, &vm->userptr.invalidated, + list_for_each_entry_safe(uvma, next, &vm->userptr.invalidated, userptr.invalidate_link) { - list_del_init(&vma->userptr.invalidate_link); - list_move_tail(&vma->combined_links.userptr, + list_del_init(&uvma->userptr.invalidate_link); + list_move_tail(&uvma->userptr.repin_link, &vm->userptr.repin_list); } spin_unlock(&vm->userptr.invalidated_lock); /* Pin and move to temporary list */ - list_for_each_entry_safe(vma, next, &vm->userptr.repin_list, - combined_links.userptr) { - err = xe_vma_userptr_pin_pages(vma); + list_for_each_entry_safe(uvma, next, &vm->userptr.repin_list, + userptr.repin_link) { + err = xe_vma_userptr_pin_pages(uvma); if (err < 0) return err; - list_move_tail(&vma->combined_links.userptr, &vm->rebind_list); + list_del_init(&uvma->userptr.repin_link); + list_move_tail(&uvma->vma.combined_links.rebind, &vm->rebind_list); } return 0; @@ -782,6 +787,14 @@ struct dma_fence *xe_vm_rebind(struct xe_vm *vm, bool rebind_worker) return fence; } +static void xe_vma_free(struct xe_vma *vma) +{ + if (xe_vma_is_userptr(vma)) + kfree(to_userptr_vma(vma)); + else + kfree(vma); +} + #define VMA_CREATE_FLAG_READ_ONLY BIT(0) #define VMA_CREATE_FLAG_IS_NULL BIT(1) @@ -800,14 +813,26 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, xe_assert(vm->xe, start < end); xe_assert(vm->xe, end < vm->size); - if (!bo && !is_null) /* userptr */ + /* + * Allocate and ensure that the xe_vma_is_userptr() return + * matches what was allocated. + */ + if (!bo && !is_null) { + struct xe_userptr_vma *uvma = kzalloc(sizeof(*uvma), GFP_KERNEL); + + if (!uvma) + return ERR_PTR(-ENOMEM); + + vma = &uvma->vma; + } else { vma = kzalloc(sizeof(*vma), GFP_KERNEL); - else - vma = kzalloc(sizeof(*vma) - sizeof(struct xe_userptr), - GFP_KERNEL); - if (!vma) { - vma = ERR_PTR(-ENOMEM); - return vma; + if (!vma) + return ERR_PTR(-ENOMEM); + + if (is_null) + vma->gpuva.flags |= DRM_GPUVA_SPARSE; + if (bo) + vma->gpuva.gem.obj = &bo->ttm.base; } INIT_LIST_HEAD(&vma->combined_links.rebind); @@ -818,8 +843,6 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, vma->gpuva.va.range = end - start + 1; if (read_only) vma->gpuva.flags |= XE_VMA_READ_ONLY; - if (is_null) - vma->gpuva.flags |= DRM_GPUVA_SPARSE; for_each_tile(tile, vm->xe, id) vma->tile_mask |= 0x1 << id; @@ -836,35 +859,35 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, vm_bo = drm_gpuvm_bo_obtain(vma->gpuva.vm, &bo->ttm.base); if (IS_ERR(vm_bo)) { - kfree(vma); + xe_vma_free(vma); return ERR_CAST(vm_bo); } drm_gpuvm_bo_extobj_add(vm_bo); drm_gem_object_get(&bo->ttm.base); - vma->gpuva.gem.obj = &bo->ttm.base; vma->gpuva.gem.offset = bo_offset_or_userptr; drm_gpuva_link(&vma->gpuva, vm_bo); drm_gpuvm_bo_put(vm_bo); } else /* userptr or null */ { if (!is_null) { + struct xe_userptr *userptr = &to_userptr_vma(vma)->userptr; u64 size = end - start + 1; int err; - INIT_LIST_HEAD(&vma->userptr.invalidate_link); + INIT_LIST_HEAD(&userptr->invalidate_link); + INIT_LIST_HEAD(&userptr->repin_link); vma->gpuva.gem.offset = bo_offset_or_userptr; - err = mmu_interval_notifier_insert(&vma->userptr.notifier, + err = mmu_interval_notifier_insert(&userptr->notifier, current->mm, xe_vma_userptr(vma), size, &vma_userptr_notifier_ops); if (err) { - kfree(vma); - vma = ERR_PTR(err); - return vma; + xe_vma_free(vma); + return ERR_PTR(err); } - vma->userptr.notifier_seq = LONG_MAX; + userptr->notifier_seq = LONG_MAX; } xe_vm_get(vm); @@ -880,13 +903,15 @@ static void xe_vma_destroy_late(struct xe_vma *vma) bool read_only = xe_vma_read_only(vma); if (xe_vma_is_userptr(vma)) { - if (vma->userptr.sg) { + struct xe_userptr *userptr = &to_userptr_vma(vma)->userptr; + + if (userptr->sg) { dma_unmap_sgtable(xe->drm.dev, - vma->userptr.sg, + userptr->sg, read_only ? DMA_TO_DEVICE : DMA_BIDIRECTIONAL, 0); - sg_free_table(vma->userptr.sg); - vma->userptr.sg = NULL; + sg_free_table(userptr->sg); + userptr->sg = NULL; } /* @@ -894,7 +919,7 @@ static void xe_vma_destroy_late(struct xe_vma *vma) * the notifer until we're sure the GPU is not accessing * them anymore */ - mmu_interval_notifier_remove(&vma->userptr.notifier); + mmu_interval_notifier_remove(&userptr->notifier); xe_vm_put(vm); } else if (xe_vma_is_null(vma)) { xe_vm_put(vm); @@ -902,7 +927,7 @@ static void xe_vma_destroy_late(struct xe_vma *vma) xe_bo_put(xe_vma_bo(vma)); } - kfree(vma); + xe_vma_free(vma); } static void vma_destroy_work_func(struct work_struct *w) @@ -933,7 +958,7 @@ static void xe_vma_destroy(struct xe_vma *vma, struct dma_fence *fence) xe_assert(vm->xe, vma->gpuva.flags & XE_VMA_DESTROYED); spin_lock(&vm->userptr.invalidated_lock); - list_del(&vma->userptr.invalidate_link); + list_del(&to_userptr_vma(vma)->userptr.invalidate_link); spin_unlock(&vm->userptr.invalidated_lock); } else if (!xe_vma_is_null(vma)) { xe_bo_assert_held(xe_vma_bo(vma)); @@ -1855,10 +1880,8 @@ int xe_vm_create_ioctl(struct drm_device *dev, void *data, mutex_lock(&xef->vm.lock); err = xa_alloc(&xef->vm.xa, &id, vm, xa_limit_32b, GFP_KERNEL); mutex_unlock(&xef->vm.lock); - if (err) { - xe_vm_close_and_put(vm); - return err; - } + if (err) + goto err_close_and_put; if (xe->info.has_asid) { mutex_lock(&xe->usm.lock); @@ -1866,11 +1889,9 @@ int xe_vm_create_ioctl(struct drm_device *dev, void *data, XA_LIMIT(1, XE_MAX_ASID - 1), &xe->usm.next_asid, GFP_KERNEL); mutex_unlock(&xe->usm.lock); - if (err < 0) { - xe_vm_close_and_put(vm); - return err; - } - err = 0; + if (err < 0) + goto err_free_id; + vm->usm.asid = asid; } @@ -1888,6 +1909,15 @@ int xe_vm_create_ioctl(struct drm_device *dev, void *data, #endif return 0; + +err_free_id: + mutex_lock(&xef->vm.lock); + xa_erase(&xef->vm.xa, id); + mutex_unlock(&xef->vm.lock); +err_close_and_put: + xe_vm_close_and_put(vm); + + return err; } int xe_vm_destroy_ioctl(struct drm_device *dev, void *data, @@ -2145,7 +2175,7 @@ static struct xe_vma *new_vma(struct xe_vm *vm, struct drm_gpuva_op_map *op, drm_exec_fini(&exec); if (xe_vma_is_userptr(vma)) { - err = xe_vma_userptr_pin_pages(vma); + err = xe_vma_userptr_pin_pages(to_userptr_vma(vma)); if (err) { prep_vma_destroy(vm, vma, false); xe_vma_destroy_unlocked(vma); @@ -2502,7 +2532,7 @@ retry_userptr: if (err == -EAGAIN && xe_vma_is_userptr(vma)) { lockdep_assert_held_write(&vm->lock); - err = xe_vma_userptr_pin_pages(vma); + err = xe_vma_userptr_pin_pages(to_userptr_vma(vma)); if (!err) goto retry_userptr; @@ -2846,7 +2876,7 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) struct drm_gpuva_ops **ops = NULL; struct xe_vm *vm; struct xe_exec_queue *q = NULL; - u32 num_syncs; + u32 num_syncs, num_ufence = 0; struct xe_sync_entry *syncs = NULL; struct drm_xe_vm_bind_op *bind_ops; LIST_HEAD(ops_list); @@ -2983,6 +3013,14 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) SYNC_PARSE_FLAG_DISALLOW_USER_FENCE : 0)); if (err) goto free_syncs; + + if (xe_sync_is_ufence(&syncs[num_syncs])) + num_ufence++; + } + + if (XE_IOCTL_DBG(xe, num_ufence > 1)) { + err = -EINVAL; + goto free_syncs; } if (!args->num_binds) { @@ -3125,8 +3163,8 @@ int xe_vm_invalidate_vma(struct xe_vma *vma) if (IS_ENABLED(CONFIG_PROVE_LOCKING)) { if (xe_vma_is_userptr(vma)) { WARN_ON_ONCE(!mmu_interval_check_retry - (&vma->userptr.notifier, - vma->userptr.notifier_seq)); + (&to_userptr_vma(vma)->userptr.notifier, + to_userptr_vma(vma)->userptr.notifier_seq)); WARN_ON_ONCE(!dma_resv_test_signaled(xe_vm_resv(xe_vma_vm(vma)), DMA_RESV_USAGE_BOOKKEEP)); @@ -3187,11 +3225,11 @@ int xe_analyze_vm(struct drm_printer *p, struct xe_vm *vm, int gt_id) if (is_null) { addr = 0; } else if (is_userptr) { + struct sg_table *sg = to_userptr_vma(vma)->userptr.sg; struct xe_res_cursor cur; - if (vma->userptr.sg) { - xe_res_first_sg(vma->userptr.sg, 0, XE_PAGE_SIZE, - &cur); + if (sg) { + xe_res_first_sg(sg, 0, XE_PAGE_SIZE, &cur); addr = xe_res_dma(&cur); } else { addr = 0; diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h index cf2f96e8c1ab..9654a0612fc2 100644 --- a/drivers/gpu/drm/xe/xe_vm.h +++ b/drivers/gpu/drm/xe/xe_vm.h @@ -160,6 +160,18 @@ static inline bool xe_vma_is_userptr(struct xe_vma *vma) return xe_vma_has_no_bo(vma) && !xe_vma_is_null(vma); } +/** + * to_userptr_vma() - Return a pointer to an embedding userptr vma + * @vma: Pointer to the embedded struct xe_vma + * + * Return: Pointer to the embedding userptr vma + */ +static inline struct xe_userptr_vma *to_userptr_vma(struct xe_vma *vma) +{ + xe_assert(xe_vma_vm(vma)->xe, xe_vma_is_userptr(vma)); + return container_of(vma, struct xe_userptr_vma, vma); +} + u64 xe_vm_pdp4_descriptor(struct xe_vm *vm, struct xe_tile *tile); int xe_vm_create_ioctl(struct drm_device *dev, void *data, @@ -224,9 +236,9 @@ static inline void xe_vm_reactivate_rebind(struct xe_vm *vm) } } -int xe_vma_userptr_pin_pages(struct xe_vma *vma); +int xe_vma_userptr_pin_pages(struct xe_userptr_vma *uvma); -int xe_vma_userptr_check_repin(struct xe_vma *vma); +int xe_vma_userptr_check_repin(struct xe_userptr_vma *uvma); bool xe_vm_validate_should_retry(struct drm_exec *exec, int err, ktime_t *end); diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h index 63e8a50b88e9..1fec66ae2eb2 100644 --- a/drivers/gpu/drm/xe/xe_vm_types.h +++ b/drivers/gpu/drm/xe/xe_vm_types.h @@ -37,6 +37,8 @@ struct xe_vm; struct xe_userptr { /** @invalidate_link: Link for the vm::userptr.invalidated list */ struct list_head invalidate_link; + /** @userptr: link into VM repin list if userptr. */ + struct list_head repin_link; /** * @notifier: MMU notifier for user pointer (invalidation call back) */ @@ -68,8 +70,6 @@ struct xe_vma { * resv. */ union { - /** @userptr: link into VM repin list if userptr. */ - struct list_head userptr; /** @rebind: link into VM if this VMA needs rebinding. */ struct list_head rebind; /** @destroy: link to contested list when VM is being closed. */ @@ -105,11 +105,15 @@ struct xe_vma { * @pat_index: The pat index to use when encoding the PTEs for this vma. */ u16 pat_index; +}; - /** - * @userptr: user pointer state, only allocated for VMAs that are - * user pointers - */ +/** + * struct xe_userptr_vma - A userptr vma subclass + * @vma: The vma. + * @userptr: Additional userptr information. + */ +struct xe_userptr_vma { + struct xe_vma vma; struct xe_userptr userptr; }; |