diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
35 files changed, 672 insertions, 346 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index fdee98fae2ba..3b8e1ee8c475 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -47,12 +47,6 @@ static struct { spinlock_t mem_limit_lock; } kfd_mem_limit; -/* Struct used for amdgpu_amdkfd_bo_validate */ -struct amdgpu_vm_parser { - uint32_t domain; - bool wait; -}; - static const char * const domain_bit_to_string[] = { "CPU", "GTT", @@ -348,11 +342,9 @@ validate_fail: return ret; } -static int amdgpu_amdkfd_validate(void *param, struct amdgpu_bo *bo) +static int amdgpu_amdkfd_validate_vm_bo(void *_unused, struct amdgpu_bo *bo) { - struct amdgpu_vm_parser *p = param; - - return amdgpu_amdkfd_bo_validate(bo, p->domain, p->wait); + return amdgpu_amdkfd_bo_validate(bo, bo->allowed_domains, false); } /* vm_validate_pt_pd_bos - Validate page table and directory BOs @@ -364,28 +356,23 @@ static int amdgpu_amdkfd_validate(void *param, struct amdgpu_bo *bo) */ static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm) { - struct amdgpu_bo *pd = vm->root.base.bo; + struct amdgpu_bo *pd = vm->root.bo; struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev); - struct amdgpu_vm_parser param; int ret; - param.domain = AMDGPU_GEM_DOMAIN_VRAM; - param.wait = false; - - ret = amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_amdkfd_validate, - ¶m); + ret = amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_amdkfd_validate_vm_bo, NULL); if (ret) { pr_err("failed to validate PT BOs\n"); return ret; } - ret = amdgpu_amdkfd_validate(¶m, pd); + ret = amdgpu_amdkfd_validate_vm_bo(NULL, pd); if (ret) { pr_err("failed to validate PD\n"); return ret; } - vm->pd_phys_addr = amdgpu_gmc_pd_addr(vm->root.base.bo); + vm->pd_phys_addr = amdgpu_gmc_pd_addr(vm->root.bo); if (vm->use_cpu_for_update) { ret = amdgpu_bo_kmap(pd, NULL); @@ -400,7 +387,7 @@ static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm) static int vm_update_pds(struct amdgpu_vm *vm, struct amdgpu_sync *sync) { - struct amdgpu_bo *pd = vm->root.base.bo; + struct amdgpu_bo *pd = vm->root.bo; struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev); int ret; @@ -652,7 +639,7 @@ kfd_mem_attach_dmabuf(struct amdgpu_device *adev, struct kgd_mem *mem, } } - gobj = amdgpu_gem_prime_import(&adev->ddev, mem->dmabuf); + gobj = amdgpu_gem_prime_import(adev_to_drm(adev), mem->dmabuf); if (IS_ERR(gobj)) return PTR_ERR(gobj); @@ -1166,7 +1153,7 @@ static int process_sync_pds_resv(struct amdkfd_process_info *process_info, list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) { - struct amdgpu_bo *pd = peer_vm->root.base.bo; + struct amdgpu_bo *pd = peer_vm->root.bo; ret = amdgpu_sync_resv(NULL, sync, pd->tbo.base.resv, AMDGPU_SYNC_NE_OWNER, @@ -1233,7 +1220,7 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, vm->process_info = *process_info; /* Validate page directory and attach eviction fence */ - ret = amdgpu_bo_reserve(vm->root.base.bo, true); + ret = amdgpu_bo_reserve(vm->root.bo, true); if (ret) goto reserve_pd_fail; ret = vm_validate_pt_pd_bos(vm); @@ -1241,16 +1228,16 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, pr_err("validate_pt_pd_bos() failed\n"); goto validate_pd_fail; } - ret = amdgpu_bo_sync_wait(vm->root.base.bo, + ret = amdgpu_bo_sync_wait(vm->root.bo, AMDGPU_FENCE_OWNER_KFD, false); if (ret) goto wait_pd_fail; - ret = dma_resv_reserve_shared(vm->root.base.bo->tbo.base.resv, 1); + ret = dma_resv_reserve_shared(vm->root.bo->tbo.base.resv, 1); if (ret) goto reserve_shared_fail; - amdgpu_bo_fence(vm->root.base.bo, + amdgpu_bo_fence(vm->root.bo, &vm->process_info->eviction_fence->base, true); - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); /* Update process info */ mutex_lock(&vm->process_info->lock); @@ -1264,7 +1251,7 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, reserve_shared_fail: wait_pd_fail: validate_pd_fail: - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); reserve_pd_fail: vm->process_info = NULL; if (info) { @@ -1319,7 +1306,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct amdkfd_process_info *process_info = vm->process_info; - struct amdgpu_bo *pd = vm->root.base.bo; + struct amdgpu_bo *pd = vm->root.bo; if (!process_info) return; @@ -1375,7 +1362,7 @@ void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv) uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv) { struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); - struct amdgpu_bo *pd = avm->root.base.bo; + struct amdgpu_bo *pd = avm->root.bo; struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev); if (adev->asic_type < CHIP_VEGA10) @@ -2402,7 +2389,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef) /* Attach eviction fence to PD / PT BOs */ list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) { - struct amdgpu_bo *bo = peer_vm->root.base.bo; + struct amdgpu_bo *bo = peer_vm->root.bo; amdgpu_bo_fence(bo, &process_info->eviction_fence->base, true); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 1476236f5c7c..76fe5b71e35d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -832,7 +832,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p) if (r) return r; - p->job->vm_pd_addr = amdgpu_gmc_pd_addr(vm->root.base.bo); + p->job->vm_pd_addr = amdgpu_gmc_pd_addr(vm->root.bo); if (amdgpu_vm_debug) { /* Invalidate all BOs to test for userspace bugs */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index a9bbb0034e1e..536005bff24a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -1304,11 +1304,11 @@ static int amdgpu_debugfs_vm_info_show(struct seq_file *m, void *unused) seq_printf(m, "pid:%d\tProcess:%s ----------\n", vm->task_info.pid, vm->task_info.process_name); - r = amdgpu_bo_reserve(vm->root.base.bo, true); + r = amdgpu_bo_reserve(vm->root.bo, true); if (r) break; amdgpu_debugfs_vm_bo_info(vm, m); - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); } mutex_unlock(&dev->filelist_mutex); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 6a242ec3f7ef..130a9adf09ef 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4124,6 +4124,7 @@ static int amdgpu_device_recover_vram(struct amdgpu_device *adev) { struct dma_fence *fence = NULL, *next = NULL; struct amdgpu_bo *shadow; + struct amdgpu_bo_vm *vmbo; long r = 1, tmo; if (amdgpu_sriov_runtime(adev)) @@ -4133,8 +4134,8 @@ static int amdgpu_device_recover_vram(struct amdgpu_device *adev) dev_info(adev->dev, "recover vram bo from shadow start\n"); mutex_lock(&adev->shadow_list_lock); - list_for_each_entry(shadow, &adev->shadow_list, shadow_list) { - + list_for_each_entry(vmbo, &adev->shadow_list, shadow_list) { + shadow = &vmbo->bo; /* No need to recover an evicted BO */ if (shadow->tbo.resource->mem_type != TTM_PL_TT || shadow->tbo.resource->start == AMDGPU_BO_INVALID_OFFSET || diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index e1b6f5891759..43e7b61d1c5c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -325,7 +325,7 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) return 0; } -int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, +int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance, int *major, int *minor, int *revision) { struct binary_header *bhdr; @@ -357,7 +357,7 @@ int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, for (j = 0; j < num_ips; j++) { ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); - if (le16_to_cpu(ip->hw_id) == hw_id) { + if ((le16_to_cpu(ip->hw_id) == hw_id) && (ip->number_instance == number_instance)) { if (major) *major = ip->major; if (minor) @@ -373,6 +373,14 @@ int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, return -EINVAL; } + +int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance, + int *major, int *minor, int *revision) +{ + return amdgpu_discovery_get_ip_version(adev, VCN_HWID, + vcn_instance, major, minor, revision); +} + void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) { struct binary_header *bhdr; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h index 1b1ae21b1037..48e6b88cfdfe 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h @@ -30,8 +30,11 @@ void amdgpu_discovery_fini(struct amdgpu_device *adev); int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev); void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev); -int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, +int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance, int *major, int *minor, int *revision); + +int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance, + int *major, int *minor, int *revision); int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev); #endif /* __AMDGPU_DISCOVERY__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index 2b6a66c849d4..8e5a7ac8c36f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -1046,11 +1046,12 @@ int amdgpu_display_gem_fb_init(struct drm_device *dev, rfb->base.obj[0] = obj; drm_helper_mode_fill_fb_struct(dev, &rfb->base, mode_cmd); - ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs); + + ret = amdgpu_display_framebuffer_init(dev, rfb, mode_cmd, obj); if (ret) goto err; - ret = amdgpu_display_framebuffer_init(dev, rfb, mode_cmd, obj); + ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs); if (ret) goto err; @@ -1070,9 +1071,6 @@ int amdgpu_display_gem_fb_verify_and_init( rfb->base.obj[0] = obj; drm_helper_mode_fill_fb_struct(dev, &rfb->base, mode_cmd); - ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs); - if (ret) - goto err; /* Verify that the modifier is supported. */ if (!drm_any_plane_has_format(dev, mode_cmd->pixel_format, mode_cmd->modifier[0])) { @@ -1088,6 +1086,10 @@ int amdgpu_display_gem_fb_verify_and_init( if (ret) goto err; + ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs); + if (ret) + goto err; + return 0; err: drm_dbg_kms(dev, "Failed to verify and init gem fb: %d\n", ret); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c index c3053b83b80c..a3daaa89330c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c @@ -448,7 +448,7 @@ amdgpu_dma_buf_move_notify(struct dma_buf_attachment *attach) for (bo_base = bo->vm_bo; bo_base; bo_base = bo_base->next) { struct amdgpu_vm *vm = bo_base->vm; - struct dma_resv *resv = vm->root.base.bo->tbo.base.resv; + struct dma_resv *resv = vm->root.bo->tbo.base.resv; if (ticket) { /* When we get an error here it means that somebody diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index c080ba15ae77..6f30c525caac 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -837,8 +837,23 @@ module_param_named(tmz, amdgpu_tmz, int, 0444); /** * DOC: freesync_video (uint) - * Enabled the optimization to adjust front porch timing to achieve seamless mode change experience - * when setting a freesync supported mode for which full modeset is not needed. + * Enable the optimization to adjust front porch timing to achieve seamless + * mode change experience when setting a freesync supported mode for which full + * modeset is not needed. + * + * The Display Core will add a set of modes derived from the base FreeSync + * video mode into the corresponding connector's mode list based on commonly + * used refresh rates and VRR range of the connected display, when users enable + * this feature. From the userspace perspective, they can see a seamless mode + * change experience when the change between different refresh rates under the + * same resolution. Additionally, userspace applications such as Video playback + * can read this modeset list and change the refresh rate based on the video + * frame rate. Finally, the userspace can also derive an appropriate mode for a + * particular refresh rate based on the FreeSync Mode and add it to the + * connector's mode list. + * + * Note: This is an experimental feature. + * * The default value: 0 (off). */ MODULE_PARM_DESC( diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c index dbebbe16e3b3..d94c5419ec25 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c @@ -69,13 +69,13 @@ void amdgpu_show_fdinfo(struct seq_file *m, struct file *f) dev = PCI_SLOT(adev->pdev->devfn); fn = PCI_FUNC(adev->pdev->devfn); - ret = amdgpu_bo_reserve(fpriv->vm.root.base.bo, false); + ret = amdgpu_bo_reserve(fpriv->vm.root.bo, false); if (ret) { DRM_ERROR("Fail to reserve bo\n"); return; } amdgpu_vm_get_memory(&fpriv->vm, &vram_mem, >t_mem, &cpu_mem); - amdgpu_bo_unreserve(fpriv->vm.root.base.bo); + amdgpu_bo_unreserve(fpriv->vm.root.bo); seq_printf(m, "pdev:\t%04x:%02x:%02x.%d\npasid:\t%u\n", domain, bus, dev, fn, fpriv->vm.pasid); seq_printf(m, "vram mem:\t%llu kB\n", vram_mem/1024UL); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index 1313784605b0..b36405170ff3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -300,7 +300,6 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset, * @adev: amdgpu_device pointer * @offset: offset into the GPU's gart aperture * @pages: number of pages to bind - * @pagelist: pages to bind * @dma_addr: DMA addresses of pages * @flags: page table entry flags * @@ -309,7 +308,7 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset, * Returns 0 for success, -EINVAL for failure. */ int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset, - int pages, struct page **pagelist, dma_addr_t *dma_addr, + int pages, dma_addr_t *dma_addr, uint64_t flags) { if (!adev->gart.ready) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h index f53f6a7b9b60..78895413cf9f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h @@ -64,7 +64,6 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset, int pages, dma_addr_t *dma_addr, uint64_t flags, void *dst); int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset, - int pages, struct page **pagelist, - dma_addr_t *dma_addr, uint64_t flags); + int pages, dma_addr_t *dma_addr, uint64_t flags); void amdgpu_gart_invalidate_tlb(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 9cf4beaf646c..b3404c43a911 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -170,7 +170,7 @@ static int amdgpu_gem_object_open(struct drm_gem_object *obj, return -EPERM; if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID && - abo->tbo.base.resv != vm->root.base.bo->tbo.base.resv) + abo->tbo.base.resv != vm->root.bo->tbo.base.resv) return -EPERM; r = amdgpu_bo_reserve(abo, false); @@ -320,11 +320,11 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, } if (flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) { - r = amdgpu_bo_reserve(vm->root.base.bo, false); + r = amdgpu_bo_reserve(vm->root.bo, false); if (r) return r; - resv = vm->root.base.bo->tbo.base.resv; + resv = vm->root.bo->tbo.base.resv; } initial_domain = (u32)(0xffffffff & args->in.domains); @@ -353,9 +353,9 @@ retry: if (!r) { struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj); - abo->parent = amdgpu_bo_ref(vm->root.base.bo); + abo->parent = amdgpu_bo_ref(vm->root.bo); } - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); } if (r) return r; @@ -841,7 +841,7 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data, } for (base = robj->vm_bo; base; base = base->next) if (amdgpu_xgmi_same_hive(amdgpu_ttm_adev(robj->tbo.bdev), - amdgpu_ttm_adev(base->vm->root.base.bo->tbo.bdev))) { + amdgpu_ttm_adev(base->vm->root.bo->tbo.bdev))) { r = -EINVAL; amdgpu_bo_unreserve(robj); goto out; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 425596c2f6f8..96ef3f1051d8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -124,6 +124,22 @@ void amdgpu_register_gpu_instance(struct amdgpu_device *adev) mutex_unlock(&mgpu_info.mutex); } +static void amdgpu_get_audio_func(struct amdgpu_device *adev) +{ + struct pci_dev *p = NULL; + + p = pci_get_domain_bus_and_slot(pci_domain_nr(adev->pdev->bus), + adev->pdev->bus->number, 1); + if (p) { + pm_runtime_get_sync(&p->dev); + + pm_runtime_mark_last_busy(&p->dev); + pm_runtime_put_autosuspend(&p->dev); + + pci_dev_put(p); + } +} + /** * amdgpu_driver_load_kms - Main load function for KMS. * @@ -213,9 +229,35 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags) DPM_FLAG_MAY_SKIP_RESUME); pm_runtime_use_autosuspend(dev->dev); pm_runtime_set_autosuspend_delay(dev->dev, 5000); + pm_runtime_allow(dev->dev); + pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); + + /* + * For runpm implemented via BACO, PMFW will handle the + * timing for BACO in and out: + * - put ASIC into BACO state only when both video and + * audio functions are in D3 state. + * - pull ASIC out of BACO state when either video or + * audio function is in D0 state. + * Also, at startup, PMFW assumes both functions are in + * D0 state. + * + * So if snd driver was loaded prior to amdgpu driver + * and audio function was put into D3 state, there will + * be no PMFW-aware D-state transition(D0->D3) on runpm + * suspend. Thus the BACO will be not correctly kicked in. + * + * Via amdgpu_get_audio_func(), the audio dev is put + * into D0 state. Then there will be a PMFW-aware D-state + * transition(D0->D3) on runpm suspend. + */ + if (amdgpu_device_supports_baco(dev) && + !(adev->flags & AMD_IS_APU) && + (adev->asic_type >= CHIP_NAVI10)) + amdgpu_get_audio_func(adev); } if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DRV_LOAD)) @@ -1220,7 +1262,7 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev, } pasid = fpriv->vm.pasid; - pd = amdgpu_bo_ref(fpriv->vm.root.base.bo); + pd = amdgpu_bo_ref(fpriv->vm.root.bo); amdgpu_ctx_mgr_fini(&fpriv->ctx_mgr); amdgpu_vm_fini(adev, &fpriv->vm); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index b7a2070d90af..795fa7445abe 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -54,29 +54,42 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo) { - struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev); struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo); - struct amdgpu_bo_user *ubo; amdgpu_bo_kunmap(bo); if (bo->tbo.base.import_attach) drm_prime_gem_destroy(&bo->tbo.base, bo->tbo.sg); drm_gem_object_release(&bo->tbo.base); + amdgpu_bo_unref(&bo->parent); + kvfree(bo); +} + +static void amdgpu_bo_user_destroy(struct ttm_buffer_object *tbo) +{ + struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo); + struct amdgpu_bo_user *ubo; + + ubo = to_amdgpu_bo_user(bo); + kfree(ubo->metadata); + amdgpu_bo_destroy(tbo); +} + +static void amdgpu_bo_vm_destroy(struct ttm_buffer_object *tbo) +{ + struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev); + struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo); + struct amdgpu_bo_vm *vmbo; + + vmbo = to_amdgpu_bo_vm(bo); /* in case amdgpu_device_recover_vram got NULL of bo->parent */ - if (!list_empty(&bo->shadow_list)) { + if (!list_empty(&vmbo->shadow_list)) { mutex_lock(&adev->shadow_list_lock); - list_del_init(&bo->shadow_list); + list_del_init(&vmbo->shadow_list); mutex_unlock(&adev->shadow_list_lock); } - amdgpu_bo_unref(&bo->parent); - - if (bo->tbo.type != ttm_bo_type_kernel) { - ubo = to_amdgpu_bo_user(bo); - kfree(ubo->metadata); - } - kvfree(bo); + amdgpu_bo_destroy(tbo); } /** @@ -91,8 +104,11 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo) */ bool amdgpu_bo_is_amdgpu_bo(struct ttm_buffer_object *bo) { - if (bo->destroy == &amdgpu_bo_destroy) + if (bo->destroy == &amdgpu_bo_destroy || + bo->destroy == &amdgpu_bo_user_destroy || + bo->destroy == &amdgpu_bo_vm_destroy) return true; + return false; } @@ -545,7 +561,6 @@ int amdgpu_bo_create(struct amdgpu_device *adev, if (bo == NULL) return -ENOMEM; drm_gem_private_object_init(adev_to_drm(adev), &bo->tbo.base, size); - INIT_LIST_HEAD(&bo->shadow_list); bo->vm_bo = NULL; bo->preferred_domains = bp->preferred_domain ? bp->preferred_domain : bp->domain; @@ -568,9 +583,12 @@ int amdgpu_bo_create(struct amdgpu_device *adev, if (bp->type == ttm_bo_type_kernel) bo->tbo.priority = 1; + if (!bp->destroy) + bp->destroy = &amdgpu_bo_destroy; + r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, bp->type, &bo->placement, page_align, &ctx, NULL, - bp->resv, &amdgpu_bo_destroy); + bp->resv, bp->destroy); if (unlikely(r != 0)) return r; @@ -634,6 +652,7 @@ int amdgpu_bo_create_user(struct amdgpu_device *adev, int r; bp->bo_ptr_size = sizeof(struct amdgpu_bo_user); + bp->destroy = &amdgpu_bo_user_destroy; r = amdgpu_bo_create(adev, bp, &bo_ptr); if (r) return r; @@ -665,11 +684,13 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev, * num of amdgpu_vm_pt entries. */ BUG_ON(bp->bo_ptr_size < sizeof(struct amdgpu_bo_vm)); + bp->destroy = &amdgpu_bo_vm_destroy; r = amdgpu_bo_create(adev, bp, &bo_ptr); if (r) return r; *vmbo_ptr = to_amdgpu_bo_vm(bo_ptr); + INIT_LIST_HEAD(&(*vmbo_ptr)->shadow_list); return r; } @@ -714,12 +735,12 @@ retry: * * Insert a BO to the shadow list. */ -void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo *bo) +void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo) { - struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + struct amdgpu_device *adev = amdgpu_ttm_adev(vmbo->bo.tbo.bdev); mutex_lock(&adev->shadow_list_lock); - list_add_tail(&bo->shadow_list, &adev->shadow_list); + list_add_tail(&vmbo->shadow_list, &adev->shadow_list); mutex_unlock(&adev->shadow_list_lock); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 126df03a7066..38c834d0f930 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -55,7 +55,8 @@ struct amdgpu_bo_param { u64 flags; enum ttm_bo_type type; bool no_wait_gpu; - struct dma_resv *resv; + struct dma_resv *resv; + void (*destroy)(struct ttm_buffer_object *bo); }; /* bo virtual addresses in a vm */ @@ -108,9 +109,6 @@ struct amdgpu_bo { #ifdef CONFIG_MMU_NOTIFIER struct mmu_interval_notifier notifier; #endif - - struct list_head shadow_list; - struct kgd_mem *kfd_bo; }; @@ -126,7 +124,8 @@ struct amdgpu_bo_user { struct amdgpu_bo_vm { struct amdgpu_bo bo; struct amdgpu_bo *shadow; - struct amdgpu_vm_pt entries[]; + struct list_head shadow_list; + struct amdgpu_vm_bo_base entries[]; }; static inline struct amdgpu_bo *ttm_to_amdgpu_bo(struct ttm_buffer_object *tbo) @@ -332,7 +331,7 @@ u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo); int amdgpu_bo_validate(struct amdgpu_bo *bo); void amdgpu_bo_get_memory(struct amdgpu_bo *bo, uint64_t *vram_mem, uint64_t *gtt_mem, uint64_t *cpu_mem); -void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo *bo); +void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo); int amdgpu_bo_restore_shadow(struct amdgpu_bo *shadow, struct dma_fence **fence); uint32_t amdgpu_bo_get_preferred_pin_domain(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 6046123d0562..3ec5099ffeb6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -171,11 +171,81 @@ Err_out: return ret; } +/* + * Helper funciton to query psp runtime database entry + * + * @adev: amdgpu_device pointer + * @entry_type: the type of psp runtime database entry + * @db_entry: runtime database entry pointer + * + * Return false if runtime database doesn't exit or entry is invalid + * or true if the specific database entry is found, and copy to @db_entry + */ +static bool psp_get_runtime_db_entry(struct amdgpu_device *adev, + enum psp_runtime_entry_type entry_type, + void *db_entry) +{ + uint64_t db_header_pos, db_dir_pos; + struct psp_runtime_data_header db_header = {0}; + struct psp_runtime_data_directory db_dir = {0}; + bool ret = false; + int i; + + db_header_pos = adev->gmc.mc_vram_size - PSP_RUNTIME_DB_OFFSET; + db_dir_pos = db_header_pos + sizeof(struct psp_runtime_data_header); + + /* read runtime db header from vram */ + amdgpu_device_vram_access(adev, db_header_pos, (uint32_t *)&db_header, + sizeof(struct psp_runtime_data_header), false); + + if (db_header.cookie != PSP_RUNTIME_DB_COOKIE_ID) { + /* runtime db doesn't exist, exit */ + dev_warn(adev->dev, "PSP runtime database doesn't exist\n"); + return false; + } + + /* read runtime database entry from vram */ + amdgpu_device_vram_access(adev, db_dir_pos, (uint32_t *)&db_dir, + sizeof(struct psp_runtime_data_directory), false); + + if (db_dir.entry_count >= PSP_RUNTIME_DB_DIAG_ENTRY_MAX_COUNT) { + /* invalid db entry count, exit */ + dev_warn(adev->dev, "Invalid PSP runtime database entry count\n"); + return false; + } + + /* look up for requested entry type */ + for (i = 0; i < db_dir.entry_count && !ret; i++) { + if (db_dir.entry_list[i].entry_type == entry_type) { + switch (entry_type) { + case PSP_RUNTIME_ENTRY_TYPE_BOOT_CONFIG: + if (db_dir.entry_list[i].size < sizeof(struct psp_runtime_boot_cfg_entry)) { + /* invalid db entry size */ + dev_warn(adev->dev, "Invalid PSP runtime database entry size\n"); + return false; + } + /* read runtime database entry */ + amdgpu_device_vram_access(adev, db_header_pos + db_dir.entry_list[i].offset, + (uint32_t *)db_entry, sizeof(struct psp_runtime_boot_cfg_entry), false); + ret = true; + break; + default: + ret = false; + break; + } + } + } + + return ret; +} + static int psp_sw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct psp_context *psp = &adev->psp; int ret; + struct psp_runtime_boot_cfg_entry boot_cfg_entry; + struct psp_memory_training_context *mem_training_ctx = &psp->mem_train_ctx; if (!amdgpu_sriov_vf(adev)) { ret = psp_init_microcode(psp); @@ -191,15 +261,39 @@ static int psp_sw_init(void *handle) } } - ret = psp_memory_training_init(psp); - if (ret) { - DRM_ERROR("Failed to initialize memory training!\n"); - return ret; + memset(&boot_cfg_entry, 0, sizeof(boot_cfg_entry)); + if (psp_get_runtime_db_entry(adev, + PSP_RUNTIME_ENTRY_TYPE_BOOT_CONFIG, + &boot_cfg_entry)) { + psp->boot_cfg_bitmask = boot_cfg_entry.boot_cfg_bitmask; + if ((psp->boot_cfg_bitmask) & + BOOT_CFG_FEATURE_TWO_STAGE_DRAM_TRAINING) { + /* If psp runtime database exists, then + * only enable two stage memory training + * when TWO_STAGE_DRAM_TRAINING bit is set + * in runtime database */ + mem_training_ctx->enable_mem_training = true; + } + + } else { + /* If psp runtime database doesn't exist or + * is invalid, force enable two stage memory + * training */ + mem_training_ctx->enable_mem_training = true; } - ret = psp_mem_training(psp, PSP_MEM_TRAIN_COLD_BOOT); - if (ret) { - DRM_ERROR("Failed to process memory training!\n"); - return ret; + + if (mem_training_ctx->enable_mem_training) { + ret = psp_memory_training_init(psp); + if (ret) { + DRM_ERROR("Failed to initialize memory training!\n"); + return ret; + } + + ret = psp_mem_training(psp, PSP_MEM_TRAIN_COLD_BOOT); + if (ret) { + DRM_ERROR("Failed to process memory training!\n"); + return ret; + } } if (adev->asic_type == CHIP_NAVI10 || adev->asic_type == CHIP_SIENNA_CICHLID) { @@ -551,7 +645,30 @@ int psp_get_fw_attestation_records_addr(struct psp_context *psp, return ret; } -static int psp_boot_config_set(struct amdgpu_device *adev) +static int psp_boot_config_get(struct amdgpu_device *adev, uint32_t *boot_cfg) +{ + struct psp_context *psp = &adev->psp; + struct psp_gfx_cmd_resp *cmd = psp->cmd; + int ret; + + if (amdgpu_sriov_vf(adev)) + return 0; + + memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp)); + + cmd->cmd_id = GFX_CMD_ID_BOOT_CFG; + cmd->cmd.boot_cfg.sub_cmd = BOOTCFG_CMD_GET; + + ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); + if (!ret) { + *boot_cfg = + (cmd->resp.uresp.boot_cfg.boot_cfg & BOOT_CONFIG_GECC) ? 1 : 0; + } + + return ret; +} + +static int psp_boot_config_set(struct amdgpu_device *adev, uint32_t boot_cfg) { struct psp_context *psp = &adev->psp; struct psp_gfx_cmd_resp *cmd = psp->cmd; @@ -563,8 +680,8 @@ static int psp_boot_config_set(struct amdgpu_device *adev) cmd->cmd_id = GFX_CMD_ID_BOOT_CFG; cmd->cmd.boot_cfg.sub_cmd = BOOTCFG_CMD_SET; - cmd->cmd.boot_cfg.boot_config = BOOT_CONFIG_GECC; - cmd->cmd.boot_cfg.boot_config_valid = BOOT_CONFIG_GECC; + cmd->cmd.boot_cfg.boot_config = boot_cfg; + cmd->cmd.boot_cfg.boot_config_valid = boot_cfg; return psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); } @@ -1212,19 +1329,62 @@ static int psp_ras_terminate(struct psp_context *psp) static int psp_ras_initialize(struct psp_context *psp) { int ret; + uint32_t boot_cfg = 0xFF; + struct amdgpu_device *adev = psp->adev; /* * TODO: bypass the initialize in sriov for now */ - if (amdgpu_sriov_vf(psp->adev)) + if (amdgpu_sriov_vf(adev)) return 0; - if (!psp->adev->psp.ta_ras_ucode_size || - !psp->adev->psp.ta_ras_start_addr) { - dev_info(psp->adev->dev, "RAS: optional ras ta ucode is not available\n"); + if (!adev->psp.ta_ras_ucode_size || + !adev->psp.ta_ras_start_addr) { + dev_info(adev->dev, "RAS: optional ras ta ucode is not available\n"); return 0; } + if (amdgpu_atomfirmware_dynamic_boot_config_supported(adev)) { + /* query GECC enablement status from boot config + * boot_cfg: 1: GECC is enabled or 0: GECC is disabled + */ + ret = psp_boot_config_get(adev, &boot_cfg); + if (ret) + dev_warn(adev->dev, "PSP get boot config failed\n"); + + if (!amdgpu_ras_is_supported(psp->adev, AMDGPU_RAS_BLOCK__UMC)) { + if (!boot_cfg) { + dev_info(adev->dev, "GECC is disabled\n"); + } else { + /* disable GECC in next boot cycle if ras is + * disabled by module parameter amdgpu_ras_enable + * and/or amdgpu_ras_mask, or boot_config_get call + * is failed + */ + ret = psp_boot_config_set(adev, 0); + if (ret) + dev_warn(adev->dev, "PSP set boot config failed\n"); + else + dev_warn(adev->dev, "GECC will be disabled in next boot cycle " + "if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n"); + } + } else { + if (1 == boot_cfg) { + dev_info(adev->dev, "GECC is enabled\n"); + } else { + /* enable GECC in next boot cycle if it is disabled + * in boot config, or force enable GECC if failed to + * get boot configuration + */ + ret = psp_boot_config_set(adev, BOOT_CONFIG_GECC); + if (ret) + dev_warn(adev->dev, "PSP set boot config failed\n"); + else + dev_warn(adev->dev, "GECC will be enabled in next boot cycle\n"); + } + } + } + if (!psp->ras.ras_initialized) { ret = psp_ras_init_shared_buf(psp); if (ret) @@ -1945,12 +2105,6 @@ static int psp_hw_start(struct psp_context *psp) return ret; } - if (amdgpu_atomfirmware_dynamic_boot_config_supported(adev)) { - ret = psp_boot_config_set(adev); - if (ret) - dev_warn(adev->dev, "PSP set boot config failed\n"); - } - ret = psp_tmr_init(psp); if (ret) { DRM_ERROR("PSP tmr init failed!\n"); @@ -2188,10 +2342,7 @@ static int psp_load_smu_fw(struct psp_context *psp) if ((amdgpu_in_reset(adev) && ras && adev->ras_enabled && (adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_VEGA20)) || - (adev->in_runpm && - adev->asic_type >= CHIP_NAVI10 && - adev->asic_type <= CHIP_NAVI12)) { + adev->asic_type == CHIP_VEGA20))) { ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD); if (ret) { DRM_WARN("Failed to set MP1 state prepare for reload\n"); @@ -2562,10 +2713,12 @@ static int psp_resume(void *handle) DRM_INFO("PSP is resuming...\n"); - ret = psp_mem_training(psp, PSP_MEM_TRAIN_RESUME); - if (ret) { - DRM_ERROR("Failed to process memory training!\n"); - return ret; + if (psp->mem_train_ctx.enable_mem_training) { + ret = psp_mem_training(psp, PSP_MEM_TRAIN_RESUME); + if (ret) { + DRM_ERROR("Failed to process memory training!\n"); + return ret; + } } mutex_lock(&adev->firmware.mutex); @@ -2749,7 +2902,7 @@ int psp_init_asd_microcode(struct psp_context *psp, asd_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.asd_fw->data; adev->psp.asd_fw_version = le32_to_cpu(asd_hdr->header.ucode_version); - adev->psp.asd_feature_version = le32_to_cpu(asd_hdr->ucode_feature_version); + adev->psp.asd_feature_version = le32_to_cpu(asd_hdr->sos.fw_version); adev->psp.asd_ucode_size = le32_to_cpu(asd_hdr->header.ucode_size_bytes); adev->psp.asd_start_addr = (uint8_t *)asd_hdr + le32_to_cpu(asd_hdr->header.ucode_array_offset_bytes); @@ -2785,7 +2938,7 @@ int psp_init_toc_microcode(struct psp_context *psp, toc_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.toc_fw->data; adev->psp.toc_fw_version = le32_to_cpu(toc_hdr->header.ucode_version); - adev->psp.toc_feature_version = le32_to_cpu(toc_hdr->ucode_feature_version); + adev->psp.toc_feature_version = le32_to_cpu(toc_hdr->sos.fw_version); adev->psp.toc_bin_size = le32_to_cpu(toc_hdr->header.ucode_size_bytes); adev->psp.toc_start_addr = (uint8_t *)toc_hdr + le32_to_cpu(toc_hdr->header.ucode_array_offset_bytes); @@ -2797,6 +2950,50 @@ out: return err; } +static int psp_init_sos_base_fw(struct amdgpu_device *adev) +{ + const struct psp_firmware_header_v1_0 *sos_hdr; + const struct psp_firmware_header_v1_3 *sos_hdr_v1_3; + uint8_t *ucode_array_start_addr; + + sos_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.sos_fw->data; + ucode_array_start_addr = (uint8_t *)sos_hdr + + le32_to_cpu(sos_hdr->header.ucode_array_offset_bytes); + + if (adev->gmc.xgmi.connected_to_cpu || (adev->asic_type != CHIP_ALDEBARAN)) { + adev->psp.sos_fw_version = le32_to_cpu(sos_hdr->header.ucode_version); + adev->psp.sos_feature_version = le32_to_cpu(sos_hdr->sos.fw_version); + + adev->psp.sys_bin_size = le32_to_cpu(sos_hdr->sos.offset_bytes); + adev->psp.sys_start_addr = ucode_array_start_addr; + + adev->psp.sos_bin_size = le32_to_cpu(sos_hdr->sos.size_bytes); + adev->psp.sos_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr->sos.offset_bytes); + } else { + /* Load alternate PSP SOS FW */ + sos_hdr_v1_3 = (const struct psp_firmware_header_v1_3 *)adev->psp.sos_fw->data; + + adev->psp.sos_fw_version = le32_to_cpu(sos_hdr_v1_3->sos_aux.fw_version); + adev->psp.sos_feature_version = le32_to_cpu(sos_hdr_v1_3->sos_aux.fw_version); + + adev->psp.sys_bin_size = le32_to_cpu(sos_hdr_v1_3->sys_drv_aux.size_bytes); + adev->psp.sys_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr_v1_3->sys_drv_aux.offset_bytes); + + adev->psp.sos_bin_size = le32_to_cpu(sos_hdr_v1_3->sos_aux.size_bytes); + adev->psp.sos_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr_v1_3->sos_aux.offset_bytes); + } + + if ((adev->psp.sys_bin_size == 0) || (adev->psp.sos_bin_size == 0)) { + dev_warn(adev->dev, "PSP SOS FW not available"); + return -EINVAL; + } + + return 0; +} + int psp_init_sos_microcode(struct psp_context *psp, const char *chip_name) { @@ -2807,6 +3004,7 @@ int psp_init_sos_microcode(struct psp_context *psp, const struct psp_firmware_header_v1_2 *sos_hdr_v1_2; const struct psp_firmware_header_v1_3 *sos_hdr_v1_3; int err = 0; + uint8_t *ucode_array_start_addr; if (!chip_name) { dev_err(adev->dev, "invalid chip name for sos microcode\n"); @@ -2823,47 +3021,45 @@ int psp_init_sos_microcode(struct psp_context *psp, goto out; sos_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.sos_fw->data; + ucode_array_start_addr = (uint8_t *)sos_hdr + + le32_to_cpu(sos_hdr->header.ucode_array_offset_bytes); amdgpu_ucode_print_psp_hdr(&sos_hdr->header); switch (sos_hdr->header.header_version_major) { case 1: - adev->psp.sos_fw_version = le32_to_cpu(sos_hdr->header.ucode_version); - adev->psp.sos_feature_version = le32_to_cpu(sos_hdr->ucode_feature_version); - adev->psp.sos_bin_size = le32_to_cpu(sos_hdr->sos_size_bytes); - adev->psp.sys_bin_size = le32_to_cpu(sos_hdr->sos_offset_bytes); - adev->psp.sys_start_addr = (uint8_t *)sos_hdr + - le32_to_cpu(sos_hdr->header.ucode_array_offset_bytes); - adev->psp.sos_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr->sos_offset_bytes); + err = psp_init_sos_base_fw(adev); + if (err) + goto out; + if (sos_hdr->header.header_version_minor == 1) { sos_hdr_v1_1 = (const struct psp_firmware_header_v1_1 *)adev->psp.sos_fw->data; - adev->psp.toc_bin_size = le32_to_cpu(sos_hdr_v1_1->toc_size_bytes); + adev->psp.toc_bin_size = le32_to_cpu(sos_hdr_v1_1->toc.size_bytes); adev->psp.toc_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_1->toc_offset_bytes); - adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_1->kdb_size_bytes); + le32_to_cpu(sos_hdr_v1_1->toc.offset_bytes); + adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_1->kdb.size_bytes); adev->psp.kdb_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_1->kdb_offset_bytes); + le32_to_cpu(sos_hdr_v1_1->kdb.offset_bytes); } if (sos_hdr->header.header_version_minor == 2) { sos_hdr_v1_2 = (const struct psp_firmware_header_v1_2 *)adev->psp.sos_fw->data; - adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_2->kdb_size_bytes); + adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_2->kdb.size_bytes); adev->psp.kdb_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_2->kdb_offset_bytes); + le32_to_cpu(sos_hdr_v1_2->kdb.offset_bytes); } if (sos_hdr->header.header_version_minor == 3) { sos_hdr_v1_3 = (const struct psp_firmware_header_v1_3 *)adev->psp.sos_fw->data; - adev->psp.toc_bin_size = le32_to_cpu(sos_hdr_v1_3->v1_1.toc_size_bytes); - adev->psp.toc_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_3->v1_1.toc_offset_bytes); - adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_3->v1_1.kdb_size_bytes); - adev->psp.kdb_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_3->v1_1.kdb_offset_bytes); - adev->psp.spl_bin_size = le32_to_cpu(sos_hdr_v1_3->spl_size_bytes); - adev->psp.spl_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_3->spl_offset_bytes); - adev->psp.rl_bin_size = le32_to_cpu(sos_hdr_v1_3->rl_size_bytes); - adev->psp.rl_start_addr = (uint8_t *)adev->psp.sys_start_addr + - le32_to_cpu(sos_hdr_v1_3->rl_offset_bytes); + adev->psp.toc_bin_size = le32_to_cpu(sos_hdr_v1_3->v1_1.toc.size_bytes); + adev->psp.toc_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr_v1_3->v1_1.toc.offset_bytes); + adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_3->v1_1.kdb.size_bytes); + adev->psp.kdb_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr_v1_3->v1_1.kdb.offset_bytes); + adev->psp.spl_bin_size = le32_to_cpu(sos_hdr_v1_3->spl.size_bytes); + adev->psp.spl_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr_v1_3->spl.offset_bytes); + adev->psp.rl_bin_size = le32_to_cpu(sos_hdr_v1_3->rl.size_bytes); + adev->psp.rl_start_addr = ucode_array_start_addr + + le32_to_cpu(sos_hdr_v1_3->rl.offset_bytes); } break; default: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index e5023f1de7fd..3030ec24eb3b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -225,6 +225,61 @@ struct psp_memory_training_context { enum psp_memory_training_init_flag init; u32 training_cnt; + bool enable_mem_training; +}; + +/** PSP runtime DB **/ +#define PSP_RUNTIME_DB_SIZE_IN_BYTES 0x10000 +#define PSP_RUNTIME_DB_OFFSET 0x100000 +#define PSP_RUNTIME_DB_COOKIE_ID 0x0ed5 +#define PSP_RUNTIME_DB_VER_1 0x0100 +#define PSP_RUNTIME_DB_DIAG_ENTRY_MAX_COUNT 0x40 + +enum psp_runtime_entry_type { + PSP_RUNTIME_ENTRY_TYPE_INVALID = 0x0, + PSP_RUNTIME_ENTRY_TYPE_TEST = 0x1, + PSP_RUNTIME_ENTRY_TYPE_MGPU_COMMON = 0x2, /* Common mGPU runtime data */ + PSP_RUNTIME_ENTRY_TYPE_MGPU_WAFL = 0x3, /* WAFL runtime data */ + PSP_RUNTIME_ENTRY_TYPE_MGPU_XGMI = 0x4, /* XGMI runtime data */ + PSP_RUNTIME_ENTRY_TYPE_BOOT_CONFIG = 0x5, /* Boot Config runtime data */ +}; + +/* PSP runtime DB header */ +struct psp_runtime_data_header { + /* determine the existence of runtime db */ + uint16_t cookie; + /* version of runtime db */ + uint16_t version; +}; + +/* PSP runtime DB entry */ +struct psp_runtime_entry { + /* type of runtime db entry */ + uint32_t entry_type; + /* offset of entry in bytes */ + uint16_t offset; + /* size of entry in bytes */ + uint16_t size; +}; + +/* PSP runtime DB directory */ +struct psp_runtime_data_directory { + /* number of valid entries */ + uint16_t entry_count; + /* db entries*/ + struct psp_runtime_entry entry_list[PSP_RUNTIME_DB_DIAG_ENTRY_MAX_COUNT]; +}; + +/* PSP runtime DB boot config feature bitmask */ +enum psp_runtime_boot_cfg_feature { + BOOT_CFG_FEATURE_GECC = 0x1, + BOOT_CFG_FEATURE_TWO_STAGE_DRAM_TRAINING = 0x2, +}; + +/* PSP runtime DB boot config entry */ +struct psp_runtime_boot_cfg_entry { + uint32_t boot_cfg_bitmask; + uint32_t reserved; }; struct psp_context @@ -325,6 +380,8 @@ struct psp_context struct psp_securedisplay_context securedisplay_context; struct mutex mutex; struct psp_memory_training_context mem_train_ctx; + + uint32_t boot_cfg_bitmask; }; struct amdgpu_psp_funcs { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 9dfc1ebb41a9..c13b02caf8c3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1984,6 +1984,9 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev) ret = amdgpu_ras_load_bad_pages(adev); if (ret) goto free; + + if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num) + adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.num_recs); } return 0; @@ -2063,7 +2066,9 @@ static void amdgpu_ras_get_quirks(struct amdgpu_device *adev) return; if (strnstr(ctx->vbios_version, "D16406", - sizeof(ctx->vbios_version))) + sizeof(ctx->vbios_version)) || + strnstr(ctx->vbios_version, "D36002", + sizeof(ctx->vbios_version))) adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__GFX); } @@ -2122,7 +2127,7 @@ static void amdgpu_ras_counte_dw(struct work_struct *work) struct amdgpu_ras *con = container_of(work, struct amdgpu_ras, ras_counte_delay_work.work); struct amdgpu_device *adev = con->adev; - struct drm_device *dev = &adev->ddev; + struct drm_device *dev = adev_to_drm(adev); unsigned long ce_count, ue_count; int res; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 80dff29f2bc7..6a214a4dfe04 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -841,7 +841,7 @@ static int amdgpu_ttm_gart_bind(struct amdgpu_device *adev, uint64_t page_idx = 1; r = amdgpu_gart_bind(adev, gtt->offset, page_idx, - ttm->pages, gtt->ttm.dma_address, flags); + gtt->ttm.dma_address, flags); if (r) goto gart_bind_fail; @@ -855,11 +855,10 @@ static int amdgpu_ttm_gart_bind(struct amdgpu_device *adev, r = amdgpu_gart_bind(adev, gtt->offset + (page_idx << PAGE_SHIFT), ttm->num_pages - page_idx, - &ttm->pages[page_idx], &(gtt->ttm.dma_address[page_idx]), flags); } else { r = amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages, - ttm->pages, gtt->ttm.dma_address, flags); + gtt->ttm.dma_address, flags); } gart_bind_fail: @@ -935,7 +934,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev, /* bind pages into GART page tables */ gtt->offset = (u64)bo_mem->start << PAGE_SHIFT; r = amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages, - ttm->pages, gtt->ttm.dma_address, flags); + gtt->ttm.dma_address, flags); if (r) DRM_ERROR("failed to bind %u pages at 0x%08llX\n", @@ -1129,8 +1128,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev, ttm->sg = kzalloc(sizeof(struct sg_table), GFP_KERNEL); if (!ttm->sg) return -ENOMEM; - - ttm->page_flags |= TTM_PAGE_FLAG_SG; return 0; } @@ -1156,7 +1153,6 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev, amdgpu_ttm_tt_set_user_pages(ttm, NULL); kfree(ttm->sg); ttm->sg = NULL; - ttm->page_flags &= ~TTM_PAGE_FLAG_SG; return; } @@ -1190,6 +1186,9 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo, return -ENOMEM; } + /* Set TTM_PAGE_FLAG_SG before populate but after create. */ + bo->ttm->page_flags |= TTM_PAGE_FLAG_SG; + gtt = (void *)bo->ttm; gtt->userptr = addr; gtt->userflags = flags; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index 5eb84de588eb..2834981f8c08 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -257,36 +257,36 @@ void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr) container_of(hdr, struct psp_firmware_header_v1_0, header); DRM_DEBUG("ucode_feature_version: %u\n", - le32_to_cpu(psp_hdr->ucode_feature_version)); + le32_to_cpu(psp_hdr->sos.fw_version)); DRM_DEBUG("sos_offset_bytes: %u\n", - le32_to_cpu(psp_hdr->sos_offset_bytes)); + le32_to_cpu(psp_hdr->sos.offset_bytes)); DRM_DEBUG("sos_size_bytes: %u\n", - le32_to_cpu(psp_hdr->sos_size_bytes)); + le32_to_cpu(psp_hdr->sos.size_bytes)); if (version_minor == 1) { const struct psp_firmware_header_v1_1 *psp_hdr_v1_1 = container_of(psp_hdr, struct psp_firmware_header_v1_1, v1_0); DRM_DEBUG("toc_header_version: %u\n", - le32_to_cpu(psp_hdr_v1_1->toc_header_version)); + le32_to_cpu(psp_hdr_v1_1->toc.fw_version)); DRM_DEBUG("toc_offset_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_1->toc_offset_bytes)); + le32_to_cpu(psp_hdr_v1_1->toc.offset_bytes)); DRM_DEBUG("toc_size_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_1->toc_size_bytes)); + le32_to_cpu(psp_hdr_v1_1->toc.size_bytes)); DRM_DEBUG("kdb_header_version: %u\n", - le32_to_cpu(psp_hdr_v1_1->kdb_header_version)); + le32_to_cpu(psp_hdr_v1_1->kdb.fw_version)); DRM_DEBUG("kdb_offset_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_1->kdb_offset_bytes)); + le32_to_cpu(psp_hdr_v1_1->kdb.offset_bytes)); DRM_DEBUG("kdb_size_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_1->kdb_size_bytes)); + le32_to_cpu(psp_hdr_v1_1->kdb.size_bytes)); } if (version_minor == 2) { const struct psp_firmware_header_v1_2 *psp_hdr_v1_2 = container_of(psp_hdr, struct psp_firmware_header_v1_2, v1_0); DRM_DEBUG("kdb_header_version: %u\n", - le32_to_cpu(psp_hdr_v1_2->kdb_header_version)); + le32_to_cpu(psp_hdr_v1_2->kdb.fw_version)); DRM_DEBUG("kdb_offset_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_2->kdb_offset_bytes)); + le32_to_cpu(psp_hdr_v1_2->kdb.offset_bytes)); DRM_DEBUG("kdb_size_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_2->kdb_size_bytes)); + le32_to_cpu(psp_hdr_v1_2->kdb.size_bytes)); } if (version_minor == 3) { const struct psp_firmware_header_v1_1 *psp_hdr_v1_1 = @@ -294,23 +294,23 @@ void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr) const struct psp_firmware_header_v1_3 *psp_hdr_v1_3 = container_of(psp_hdr_v1_1, struct psp_firmware_header_v1_3, v1_1); DRM_DEBUG("toc_header_version: %u\n", - le32_to_cpu(psp_hdr_v1_3->v1_1.toc_header_version)); + le32_to_cpu(psp_hdr_v1_3->v1_1.toc.fw_version)); DRM_DEBUG("toc_offset_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_3->v1_1.toc_offset_bytes)); + le32_to_cpu(psp_hdr_v1_3->v1_1.toc.offset_bytes)); DRM_DEBUG("toc_size_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_3->v1_1.toc_size_bytes)); + le32_to_cpu(psp_hdr_v1_3->v1_1.toc.size_bytes)); DRM_DEBUG("kdb_header_version: %u\n", - le32_to_cpu(psp_hdr_v1_3->v1_1.kdb_header_version)); + le32_to_cpu(psp_hdr_v1_3->v1_1.kdb.fw_version)); DRM_DEBUG("kdb_offset_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_3->v1_1.kdb_offset_bytes)); + le32_to_cpu(psp_hdr_v1_3->v1_1.kdb.offset_bytes)); DRM_DEBUG("kdb_size_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_3->v1_1.kdb_size_bytes)); + le32_to_cpu(psp_hdr_v1_3->v1_1.kdb.size_bytes)); DRM_DEBUG("spl_header_version: %u\n", - le32_to_cpu(psp_hdr_v1_3->spl_header_version)); + le32_to_cpu(psp_hdr_v1_3->spl.fw_version)); DRM_DEBUG("spl_offset_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_3->spl_offset_bytes)); + le32_to_cpu(psp_hdr_v1_3->spl.offset_bytes)); DRM_DEBUG("spl_size_bytes: %u\n", - le32_to_cpu(psp_hdr_v1_3->spl_size_bytes)); + le32_to_cpu(psp_hdr_v1_3->spl.size_bytes)); } } else { DRM_ERROR("Unknown PSP ucode version: %u.%u\n", diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h index 2c42874f7784..270309e7f5f5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h @@ -71,43 +71,39 @@ struct smc_firmware_header_v2_1 { uint32_t pptable_entry_offset; }; +struct psp_fw_bin_desc { + uint32_t fw_version; + uint32_t offset_bytes; + uint32_t size_bytes; +}; + /* version_major=1, version_minor=0 */ struct psp_firmware_header_v1_0 { struct common_firmware_header header; - uint32_t ucode_feature_version; - uint32_t sos_offset_bytes; - uint32_t sos_size_bytes; + struct psp_fw_bin_desc sos; }; /* version_major=1, version_minor=1 */ struct psp_firmware_header_v1_1 { struct psp_firmware_header_v1_0 v1_0; - uint32_t toc_header_version; - uint32_t toc_offset_bytes; - uint32_t toc_size_bytes; - uint32_t kdb_header_version; - uint32_t kdb_offset_bytes; - uint32_t kdb_size_bytes; + struct psp_fw_bin_desc toc; + struct psp_fw_bin_desc kdb; }; /* version_major=1, version_minor=2 */ struct psp_firmware_header_v1_2 { struct psp_firmware_header_v1_0 v1_0; - uint32_t reserve[3]; - uint32_t kdb_header_version; - uint32_t kdb_offset_bytes; - uint32_t kdb_size_bytes; + struct psp_fw_bin_desc res; + struct psp_fw_bin_desc kdb; }; /* version_major=1, version_minor=3 */ struct psp_firmware_header_v1_3 { struct psp_firmware_header_v1_1 v1_1; - uint32_t spl_header_version; - uint32_t spl_offset_bytes; - uint32_t spl_size_bytes; - uint32_t rl_header_version; - uint32_t rl_offset_bytes; - uint32_t rl_size_bytes; + struct psp_fw_bin_desc spl; + struct psp_fw_bin_desc rl; + struct psp_fw_bin_desc sys_drv_aux; + struct psp_fw_bin_desc sos_aux; }; /* version_major=1, version_minor=0 */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index ea6f99be070b..f4489773715e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -94,6 +94,7 @@ int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + struct amdgpu_ras *con = amdgpu_ras_get_context(adev); kgd2kfd_set_sram_ecc_flag(adev->kfd.dev); if (adev->umc.ras_funcs && @@ -131,6 +132,9 @@ int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev, amdgpu_ras_add_bad_pages(adev, err_data->err_addr, err_data->err_addr_cnt); amdgpu_ras_save_bad_pages(adev); + + if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num) + adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.num_recs); } amdgpu_ras_reset_gpu(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index bbcccf53080d..e5a75fb788dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -22,6 +22,11 @@ #define __AMDGPU_UMC_H__ /* + * (addr / 256) * 4096, the higher 26 bits in ErrorAddr + * is the index of 4KB block + */ +#define ADDR_OF_4KB_BLOCK(addr) (((addr) & ~0xffULL) << 4) +/* * (addr / 256) * 8192, the higher 26 bits in ErrorAddr * is the index of 8KB block */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 647d2c31e8bd..6780df0fb265 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -288,6 +288,29 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev) return 0; } +bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, enum vcn_ring_type type, uint32_t vcn_instance) +{ + bool ret = false; + + int major; + int minor; + int revision; + + /* if cannot find IP data, then this VCN does not exist */ + if (amdgpu_discovery_get_vcn_version(adev, vcn_instance, &major, &minor, &revision) != 0) + return true; + + if ((type == VCN_ENCODE_RING) && (revision & VCN_BLOCK_ENCODE_DISABLE_MASK)) { + ret = true; + } else if ((type == VCN_DECODE_RING) && (revision & VCN_BLOCK_DECODE_DISABLE_MASK)) { + ret = true; + } else if ((type == VCN_UNIFIED_RING) && (revision & VCN_BLOCK_QUEUE_DISABLE_MASK)) { + ret = true; + } + + return ret; +} + int amdgpu_vcn_suspend(struct amdgpu_device *adev) { unsigned size; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index bc76cab67697..d74c62b49795 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -280,6 +280,16 @@ struct amdgpu_vcn_decode_buffer { uint32_t pad[30]; }; +#define VCN_BLOCK_ENCODE_DISABLE_MASK 0x80 +#define VCN_BLOCK_DECODE_DISABLE_MASK 0x40 +#define VCN_BLOCK_QUEUE_DISABLE_MASK 0xC0 + +enum vcn_ring_type { + VCN_ENCODE_RING, + VCN_DECODE_RING, + VCN_UNIFIED_RING, +}; + int amdgpu_vcn_sw_init(struct amdgpu_device *adev); int amdgpu_vcn_sw_fini(struct amdgpu_device *adev); int amdgpu_vcn_suspend(struct amdgpu_device *adev); @@ -287,6 +297,9 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev); void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring); void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring); +bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, + enum vcn_ring_type type, uint32_t vcn_instance); + int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring); int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); int amdgpu_vcn_dec_sw_ring_test_ring(struct amdgpu_ring *ring); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 18246b5b6ee3..79cfa2d68487 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -332,7 +332,7 @@ static void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base, base->next = bo->vm_bo; bo->vm_bo = base; - if (bo->tbo.base.resv != vm->root.base.bo->tbo.base.resv) + if (bo->tbo.base.resv != vm->root.bo->tbo.base.resv) return; vm->bulk_moveable = false; @@ -361,14 +361,14 @@ static void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base, * Helper to get the parent entry for the child page table. NULL if we are at * the root page directory. */ -static struct amdgpu_vm_pt *amdgpu_vm_pt_parent(struct amdgpu_vm_pt *pt) +static struct amdgpu_vm_bo_base *amdgpu_vm_pt_parent(struct amdgpu_vm_bo_base *pt) { - struct amdgpu_bo *parent = pt->base.bo->parent; + struct amdgpu_bo *parent = pt->bo->parent; if (!parent) return NULL; - return container_of(parent->vm_bo, struct amdgpu_vm_pt, base); + return parent->vm_bo; } /* @@ -376,8 +376,8 @@ static struct amdgpu_vm_pt *amdgpu_vm_pt_parent(struct amdgpu_vm_pt *pt) */ struct amdgpu_vm_pt_cursor { uint64_t pfn; - struct amdgpu_vm_pt *parent; - struct amdgpu_vm_pt *entry; + struct amdgpu_vm_bo_base *parent; + struct amdgpu_vm_bo_base *entry; unsigned level; }; @@ -416,17 +416,17 @@ static bool amdgpu_vm_pt_descendant(struct amdgpu_device *adev, { unsigned mask, shift, idx; - if (!cursor->entry->entries) + if ((cursor->level == AMDGPU_VM_PTB) || !cursor->entry || + !cursor->entry->bo) return false; - BUG_ON(!cursor->entry->base.bo); mask = amdgpu_vm_entries_mask(adev, cursor->level); shift = amdgpu_vm_level_shift(adev, cursor->level); ++cursor->level; idx = (cursor->pfn >> shift) & mask; cursor->parent = cursor->entry; - cursor->entry = &cursor->entry->entries[idx]; + cursor->entry = &to_amdgpu_bo_vm(cursor->entry->bo)->entries[idx]; return true; } @@ -453,7 +453,7 @@ static bool amdgpu_vm_pt_sibling(struct amdgpu_device *adev, shift = amdgpu_vm_level_shift(adev, cursor->level - 1); num_entries = amdgpu_vm_num_entries(adev, cursor->level - 1); - if (cursor->entry == &cursor->parent->entries[num_entries - 1]) + if (cursor->entry == &to_amdgpu_bo_vm(cursor->parent->bo)->entries[num_entries - 1]) return false; cursor->pfn += 1ULL << shift; @@ -539,7 +539,7 @@ static void amdgpu_vm_pt_first_dfs(struct amdgpu_device *adev, * True when the search should continue, false otherwise. */ static bool amdgpu_vm_pt_continue_dfs(struct amdgpu_vm_pt_cursor *start, - struct amdgpu_vm_pt *entry) + struct amdgpu_vm_bo_base *entry) { return entry && (!start || entry != start->entry); } @@ -590,7 +590,7 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct amdgpu_bo_list_entry *entry) { entry->priority = 0; - entry->tv.bo = &vm->root.base.bo->tbo; + entry->tv.bo = &vm->root.bo->tbo; /* Two for VM updates, one for TTM and one for the CS job */ entry->tv.num_shared = 4; entry->user_pages = NULL; @@ -622,7 +622,7 @@ void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo) for (bo_base = abo->vm_bo; bo_base; bo_base = bo_base->next) { struct amdgpu_vm *vm = bo_base->vm; - if (abo->tbo.base.resv == vm->root.base.bo->tbo.base.resv) + if (abo->tbo.base.resv == vm->root.bo->tbo.base.resv) vm->bulk_moveable = false; } @@ -781,11 +781,11 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, entries -= ats_entries; } else { - struct amdgpu_vm_pt *pt; + struct amdgpu_vm_bo_base *pt; - pt = container_of(ancestor->vm_bo, struct amdgpu_vm_pt, base); + pt = ancestor->vm_bo; ats_entries = amdgpu_vm_num_ats_entries(adev); - if ((pt - vm->root.entries) >= ats_entries) { + if ((pt - to_amdgpu_bo_vm(vm->root.bo)->entries) >= ats_entries) { ats_entries = 0; } else { ats_entries = entries; @@ -902,8 +902,8 @@ static int amdgpu_vm_pt_create(struct amdgpu_device *adev, bp.type = ttm_bo_type_kernel; bp.no_wait_gpu = immediate; - if (vm->root.base.bo) - bp.resv = vm->root.base.bo->tbo.base.resv; + if (vm->root.bo) + bp.resv = vm->root.bo->tbo.base.resv; r = amdgpu_bo_create_vm(adev, &bp, vmbo); if (r) @@ -938,7 +938,7 @@ static int amdgpu_vm_pt_create(struct amdgpu_device *adev, } (*vmbo)->shadow->parent = amdgpu_bo_ref(bo); - amdgpu_bo_add_to_shadow_list((*vmbo)->shadow); + amdgpu_bo_add_to_shadow_list(*vmbo); return 0; } @@ -962,19 +962,13 @@ static int amdgpu_vm_alloc_pts(struct amdgpu_device *adev, struct amdgpu_vm_pt_cursor *cursor, bool immediate) { - struct amdgpu_vm_pt *entry = cursor->entry; + struct amdgpu_vm_bo_base *entry = cursor->entry; struct amdgpu_bo *pt_bo; struct amdgpu_bo_vm *pt; int r; - if (entry->base.bo) { - if (cursor->level < AMDGPU_VM_PTB) - entry->entries = - to_amdgpu_bo_vm(entry->base.bo)->entries; - else - entry->entries = NULL; + if (entry->bo) return 0; - } r = amdgpu_vm_pt_create(adev, vm, cursor->level, immediate, &pt); if (r) @@ -984,13 +978,8 @@ static int amdgpu_vm_alloc_pts(struct amdgpu_device *adev, * freeing them up in the wrong order. */ pt_bo = &pt->bo; - pt_bo->parent = amdgpu_bo_ref(cursor->parent->base.bo); - amdgpu_vm_bo_base_init(&entry->base, vm, pt_bo); - if (cursor->level < AMDGPU_VM_PTB) - entry->entries = pt->entries; - else - entry->entries = NULL; - + pt_bo->parent = amdgpu_bo_ref(cursor->parent->bo); + amdgpu_vm_bo_base_init(entry, vm, pt_bo); r = amdgpu_vm_clear_bo(adev, vm, pt, immediate); if (r) goto error_free_pt; @@ -1008,18 +997,17 @@ error_free_pt: * * @entry: PDE to free */ -static void amdgpu_vm_free_table(struct amdgpu_vm_pt *entry) +static void amdgpu_vm_free_table(struct amdgpu_vm_bo_base *entry) { struct amdgpu_bo *shadow; - if (entry->base.bo) { - shadow = amdgpu_bo_shadowed(entry->base.bo); - entry->base.bo->vm_bo = NULL; - list_del(&entry->base.vm_status); - amdgpu_bo_unref(&shadow); - amdgpu_bo_unref(&entry->base.bo); - } - entry->entries = NULL; + if (!entry->bo) + return; + shadow = amdgpu_bo_shadowed(entry->bo); + entry->bo->vm_bo = NULL; + list_del(&entry->vm_status); + amdgpu_bo_unref(&shadow); + amdgpu_bo_unref(&entry->bo); } /** @@ -1036,7 +1024,7 @@ static void amdgpu_vm_free_pts(struct amdgpu_device *adev, struct amdgpu_vm_pt_cursor *start) { struct amdgpu_vm_pt_cursor cursor; - struct amdgpu_vm_pt *entry; + struct amdgpu_vm_bo_base *entry; vm->bulk_moveable = false; @@ -1304,10 +1292,10 @@ uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr) */ static int amdgpu_vm_update_pde(struct amdgpu_vm_update_params *params, struct amdgpu_vm *vm, - struct amdgpu_vm_pt *entry) + struct amdgpu_vm_bo_base *entry) { - struct amdgpu_vm_pt *parent = amdgpu_vm_pt_parent(entry); - struct amdgpu_bo *bo = parent->base.bo, *pbo; + struct amdgpu_vm_bo_base *parent = amdgpu_vm_pt_parent(entry); + struct amdgpu_bo *bo = parent->bo, *pbo; uint64_t pde, pt, flags; unsigned level; @@ -1315,8 +1303,8 @@ static int amdgpu_vm_update_pde(struct amdgpu_vm_update_params *params, pbo = pbo->parent; level += params->adev->vm_manager.root_level; - amdgpu_gmc_get_pde_for_bo(entry->base.bo, level, &pt, &flags); - pde = (entry - parent->entries) * 8; + amdgpu_gmc_get_pde_for_bo(entry->bo, level, &pt, &flags); + pde = (entry - to_amdgpu_bo_vm(parent->bo)->entries) * 8; return vm->update_funcs->update(params, to_amdgpu_bo_vm(bo), pde, pt, 1, 0, flags); } @@ -1333,11 +1321,11 @@ static void amdgpu_vm_invalidate_pds(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct amdgpu_vm_pt_cursor cursor; - struct amdgpu_vm_pt *entry; + struct amdgpu_vm_bo_base *entry; for_each_amdgpu_vm_pt_dfs_safe(adev, vm, NULL, cursor, entry) - if (entry->base.bo && !entry->base.moved) - amdgpu_vm_bo_relocated(&entry->base); + if (entry->bo && !entry->moved) + amdgpu_vm_bo_relocated(entry); } /** @@ -1371,11 +1359,12 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev, return r; while (!list_empty(&vm->relocated)) { - struct amdgpu_vm_pt *entry; + struct amdgpu_vm_bo_base *entry; - entry = list_first_entry(&vm->relocated, struct amdgpu_vm_pt, - base.vm_status); - amdgpu_vm_bo_idle(&entry->base); + entry = list_first_entry(&vm->relocated, + struct amdgpu_vm_bo_base, + vm_status); + amdgpu_vm_bo_idle(entry); r = amdgpu_vm_update_pde(¶ms, vm, entry); if (r) @@ -1555,7 +1544,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params, continue; } - pt = cursor.entry->base.bo; + pt = cursor.entry->bo; if (!pt) { /* We need all PDs and PTs for mapping something, */ if (flags & AMDGPU_PTE_VALID) @@ -1567,7 +1556,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params, if (!amdgpu_vm_pt_ancestor(&cursor)) return -EINVAL; - pt = cursor.entry->base.bo; + pt = cursor.entry->bo; shift = parent_shift; frag_end = max(frag_end, ALIGN(frag_start + 1, 1ULL << shift)); @@ -1622,7 +1611,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params, */ while (cursor.pfn < frag_start) { /* Make sure previous mapping is freed */ - if (cursor.entry->base.bo) { + if (cursor.entry->bo) { params->table_freed = true; amdgpu_vm_free_pts(adev, params->vm, &cursor); } @@ -1704,7 +1693,7 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, if (!unlocked && !dma_fence_is_signaled(vm->last_unlocked)) { struct dma_fence *tmp = dma_fence_get_stub(); - amdgpu_bo_fence(vm->root.base.bo, vm->last_unlocked, true); + amdgpu_bo_fence(vm->root.bo, vm->last_unlocked, true); swap(vm->last_unlocked, tmp); dma_fence_put(tmp); } @@ -1850,7 +1839,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va, if (clear || !bo) { mem = NULL; - resv = vm->root.base.bo->tbo.base.resv; + resv = vm->root.bo->tbo.base.resv; } else { struct drm_gem_object *obj = &bo->tbo.base; @@ -1881,7 +1870,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va, } if (clear || (bo && bo->tbo.base.resv == - vm->root.base.bo->tbo.base.resv)) + vm->root.bo->tbo.base.resv)) last_update = &vm->last_update; else last_update = &bo_va->last_pt_update; @@ -1923,7 +1912,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va, * the evicted list so that it gets validated again on the * next command submission. */ - if (bo && bo->tbo.base.resv == vm->root.base.bo->tbo.base.resv) { + if (bo && bo->tbo.base.resv == vm->root.bo->tbo.base.resv) { uint32_t mem_type = bo->tbo.resource->mem_type; if (!(bo->preferred_domains & @@ -2060,7 +2049,7 @@ static void amdgpu_vm_free_mapping(struct amdgpu_device *adev, */ static void amdgpu_vm_prt_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) { - struct dma_resv *resv = vm->root.base.bo->tbo.base.resv; + struct dma_resv *resv = vm->root.bo->tbo.base.resv; struct dma_fence *excl, **shared; unsigned i, shared_count; int r; @@ -2106,7 +2095,7 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct dma_fence **fence) { - struct dma_resv *resv = vm->root.base.bo->tbo.base.resv; + struct dma_resv *resv = vm->root.bo->tbo.base.resv; struct amdgpu_bo_va_mapping *mapping; uint64_t init_pte_value = 0; struct dma_fence *f = NULL; @@ -2265,7 +2254,7 @@ static void amdgpu_vm_bo_insert_map(struct amdgpu_device *adev, if (mapping->flags & AMDGPU_PTE_PRT) amdgpu_vm_prt_get(adev); - if (bo && bo->tbo.base.resv == vm->root.base.bo->tbo.base.resv && + if (bo && bo->tbo.base.resv == vm->root.bo->tbo.base.resv && !bo_va->base.moved) { list_move(&bo_va->base.vm_status, &vm->moved); } @@ -2627,7 +2616,7 @@ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev, struct amdgpu_vm_bo_base **base; if (bo) { - if (bo->tbo.base.resv == vm->root.base.bo->tbo.base.resv) + if (bo->tbo.base.resv == vm->root.bo->tbo.base.resv) vm->bulk_moveable = false; for (base = &bo_va->base.bo->vm_bo; *base; @@ -2721,7 +2710,7 @@ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev, for (bo_base = bo->vm_bo; bo_base; bo_base = bo_base->next) { struct amdgpu_vm *vm = bo_base->vm; - if (evicted && bo->tbo.base.resv == vm->root.base.bo->tbo.base.resv) { + if (evicted && bo->tbo.base.resv == vm->root.bo->tbo.base.resv) { amdgpu_vm_bo_evicted(bo_base); continue; } @@ -2732,7 +2721,7 @@ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev, if (bo->tbo.type == ttm_bo_type_kernel) amdgpu_vm_bo_relocated(bo_base); - else if (bo->tbo.base.resv == vm->root.base.bo->tbo.base.resv) + else if (bo->tbo.base.resv == vm->root.bo->tbo.base.resv) amdgpu_vm_bo_moved(bo_base); else amdgpu_vm_bo_invalidated(bo_base); @@ -2862,7 +2851,7 @@ void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t min_vm_size, */ long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout) { - timeout = dma_resv_wait_timeout(vm->root.base.bo->tbo.base.resv, true, + timeout = dma_resv_wait_timeout(vm->root.bo->tbo.base.resv, true, true, timeout); if (timeout <= 0) return timeout; @@ -2948,13 +2937,13 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, u32 pasid) if (r) goto error_unreserve; - amdgpu_vm_bo_base_init(&vm->root.base, vm, root_bo); + amdgpu_vm_bo_base_init(&vm->root, vm, root_bo); r = amdgpu_vm_clear_bo(adev, vm, root, false); if (r) goto error_unreserve; - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); if (pasid) { unsigned long flags; @@ -2974,12 +2963,12 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, u32 pasid) return 0; error_unreserve: - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); error_free_root: amdgpu_bo_unref(&root->shadow); amdgpu_bo_unref(&root_bo); - vm->root.base.bo = NULL; + vm->root.bo = NULL; error_free_delayed: dma_fence_put(vm->last_unlocked); @@ -3005,17 +2994,14 @@ error_free_immediate: * 0 if this VM is clean */ static int amdgpu_vm_check_clean_reserved(struct amdgpu_device *adev, - struct amdgpu_vm *vm) + struct amdgpu_vm *vm) { enum amdgpu_vm_level root = adev->vm_manager.root_level; unsigned int entries = amdgpu_vm_num_entries(adev, root); unsigned int i = 0; - if (!(vm->root.entries)) - return 0; - for (i = 0; i < entries; i++) { - if (vm->root.entries[i].base.bo) + if (to_amdgpu_bo_vm(vm->root.bo)->entries[i].bo) return -EINVAL; } @@ -3049,7 +3035,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, bool pte_support_ats = (adev->asic_type == CHIP_RAVEN); int r; - r = amdgpu_bo_reserve(vm->root.base.bo, true); + r = amdgpu_bo_reserve(vm->root.bo, true); if (r) return r; @@ -3077,7 +3063,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, if (pte_support_ats != vm->pte_support_ats) { vm->pte_support_ats = pte_support_ats; r = amdgpu_vm_clear_bo(adev, vm, - to_amdgpu_bo_vm(vm->root.base.bo), + to_amdgpu_bo_vm(vm->root.bo), false); if (r) goto free_idr; @@ -3094,7 +3080,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, if (vm->use_cpu_for_update) { /* Sync with last SDMA update/clear before switching to CPU */ - r = amdgpu_bo_sync_wait(vm->root.base.bo, + r = amdgpu_bo_sync_wait(vm->root.bo, AMDGPU_FENCE_OWNER_UNDEFINED, true); if (r) goto free_idr; @@ -3122,7 +3108,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, } /* Free the shadow bo for compute VM */ - amdgpu_bo_unref(&to_amdgpu_bo_vm(vm->root.base.bo)->shadow); + amdgpu_bo_unref(&to_amdgpu_bo_vm(vm->root.bo)->shadow); if (pasid) vm->pasid = pasid; @@ -3138,7 +3124,7 @@ free_idr: spin_unlock_irqrestore(&adev->vm_manager.pasid_lock, flags); } unreserve_bo: - amdgpu_bo_unreserve(vm->root.base.bo); + amdgpu_bo_unreserve(vm->root.bo); return r; } @@ -3181,7 +3167,7 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) amdgpu_amdkfd_gpuvm_destroy_cb(adev, vm); - root = amdgpu_bo_ref(vm->root.base.bo); + root = amdgpu_bo_ref(vm->root.bo); amdgpu_bo_reserve(root, true); if (vm->pasid) { unsigned long flags; @@ -3208,7 +3194,7 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) amdgpu_vm_free_pts(adev, vm, NULL); amdgpu_bo_unreserve(root); amdgpu_bo_unref(&root); - WARN_ON(vm->root.base.bo); + WARN_ON(vm->root.bo); drm_sched_entity_destroy(&vm->immediate); drm_sched_entity_destroy(&vm->delayed); @@ -3325,7 +3311,7 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) /* Wait vm idle to make sure the vmid set in SPM_VMID is * not referenced anymore. */ - r = amdgpu_bo_reserve(fpriv->vm.root.base.bo, true); + r = amdgpu_bo_reserve(fpriv->vm.root.bo, true); if (r) return r; @@ -3333,7 +3319,7 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) if (r < 0) return r; - amdgpu_bo_unreserve(fpriv->vm.root.base.bo); + amdgpu_bo_unreserve(fpriv->vm.root.bo); amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB_0); break; default: @@ -3399,19 +3385,20 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, { bool is_compute_context = false; struct amdgpu_bo *root; + unsigned long irqflags; uint64_t value, flags; struct amdgpu_vm *vm; int r; - spin_lock(&adev->vm_manager.pasid_lock); + spin_lock_irqsave(&adev->vm_manager.pasid_lock, irqflags); vm = idr_find(&adev->vm_manager.pasid_idr, pasid); if (vm) { - root = amdgpu_bo_ref(vm->root.base.bo); + root = amdgpu_bo_ref(vm->root.bo); is_compute_context = vm->is_compute_context; } else { root = NULL; } - spin_unlock(&adev->vm_manager.pasid_lock); + spin_unlock_irqrestore(&adev->vm_manager.pasid_lock, irqflags); if (!root) return false; @@ -3429,11 +3416,11 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, goto error_unref; /* Double check that the VM still exists */ - spin_lock(&adev->vm_manager.pasid_lock); + spin_lock_irqsave(&adev->vm_manager.pasid_lock, irqflags); vm = idr_find(&adev->vm_manager.pasid_idr, pasid); - if (vm && vm->root.base.bo != root) + if (vm && vm->root.bo != root) vm = NULL; - spin_unlock(&adev->vm_manager.pasid_lock); + spin_unlock_irqrestore(&adev->vm_manager.pasid_lock, irqflags); if (!vm) goto error_unlock; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index 1f089da1e615..ddb85a85cbba 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -152,13 +152,6 @@ struct amdgpu_vm_bo_base { bool moved; }; -struct amdgpu_vm_pt { - struct amdgpu_vm_bo_base base; - - /* array of page tables, one for each directory entry */ - struct amdgpu_vm_pt *entries; -}; - /* provided by hw blocks that can write ptes, e.g., sdma */ struct amdgpu_vm_pte_funcs { /* number of dw to reserve per operation */ @@ -284,7 +277,7 @@ struct amdgpu_vm { struct list_head done; /* contains the page directory */ - struct amdgpu_vm_pt root; + struct amdgpu_vm_bo_base root; struct dma_fence *last_update; /* Scheduler entities for page table updates */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c index 422958152c2b..dbb551762805 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c @@ -112,7 +112,7 @@ static int amdgpu_vm_sdma_commit(struct amdgpu_vm_update_params *p, swap(p->vm->last_unlocked, f); dma_fence_put(tmp); } else { - amdgpu_bo_fence(p->vm->root.base.bo, f, true); + amdgpu_bo_fence(p->vm->root.bo, f, true); } if (fence && !p->immediate) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c index c0352dcc89be..1769c4cba2ad 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c @@ -782,11 +782,6 @@ void gfx_v9_4_2_set_power_brake_sequence(struct amdgpu_device *adev) tmp = REG_SET_FIELD(tmp, GC_THROTTLE_CTRL1, PWRBRK_STALL_EN, 1); WREG32_SOC15(GC, 0, regGC_THROTTLE_CTRL1, tmp); - WREG32_SOC15(GC, 0, regDIDT_IND_INDEX, ixDIDT_SQ_THROTTLE_CTRL); - tmp = 0; - tmp = REG_SET_FIELD(tmp, DIDT_SQ_THROTTLE_CTRL, PWRBRK_STALL_EN, 1); - WREG32_SOC15(GC, 0, regDIDT_IND_DATA, tmp); - WREG32_SOC15(GC, 0, regGC_CAC_IND_INDEX, ixPWRBRK_STALL_PATTERN_CTRL); tmp = 0; tmp = REG_SET_FIELD(tmp, PWRBRK_STALL_PATTERN_CTRL, PWRBRK_END_STEP, 0x12); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h index f6d3180febc4..dd0dce254901 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h +++ b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h @@ -332,11 +332,16 @@ struct psp_gfx_uresp_fwar_db_info uint32_t fwar_db_addr_hi; }; +/* Command-specific response for boot config. */ +struct psp_gfx_uresp_bootcfg { + uint32_t boot_cfg; /* boot config data */ +}; + /* Union of command-specific responses for GPCOM ring. */ -union psp_gfx_uresp -{ - struct psp_gfx_uresp_reserved reserved; - struct psp_gfx_uresp_fwar_db_info fwar_db_info; +union psp_gfx_uresp { + struct psp_gfx_uresp_reserved reserved; + struct psp_gfx_uresp_bootcfg boot_cfg; + struct psp_gfx_uresp_fwar_db_info fwar_db_info; }; /* Structure of GFX Response buffer. diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index fc400d95b83b..bc133db2d538 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -461,6 +461,7 @@ static int psp_v11_0_ring_create(struct psp_context *psp, struct amdgpu_device *adev = psp->adev; if (amdgpu_sriov_vf(adev)) { + ring->ring_wptr = 0; ret = psp_v11_0_ring_stop(psp, ring_type); if (ret) { DRM_ERROR("psp_v11_0_ring_stop_sriov failed!\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c index ce7377d2368f..b86dcbabb635 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c @@ -227,6 +227,7 @@ static int psp_v3_1_ring_create(struct psp_context *psp, psp_v3_1_reroute_ih(psp); if (amdgpu_sriov_vf(adev)) { + ring->ring_wptr = 0; ret = psp_v3_1_ring_stop(psp, ring_type); if (ret) { DRM_ERROR("psp_v3_1_ring_stop_sriov failed!\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c index 89d20adfa001..af59a35788e3 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c @@ -234,7 +234,7 @@ static void umc_v8_7_query_error_address(struct amdgpu_device *adev, err_addr &= ~((0x1ULL << lsb) - 1); /* translate umc channel address to soc pa, 3 parts are included */ - retired_page = ADDR_OF_8KB_BLOCK(err_addr) | + retired_page = ADDR_OF_4KB_BLOCK(err_addr) | ADDR_OF_256B_BLOCK(channel_index) | OFFSET_IN_256B_BLOCK(err_addr); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 4c36fc5c9738..47d4f04cbd69 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -87,21 +87,18 @@ static void vcn_v3_0_enc_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v3_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i; if (amdgpu_sriov_vf(adev)) { - adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID; + for (i = 0; i < VCN_INSTANCES_SIENNA_CICHLID; i++) + if (amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, i)) + adev->vcn.num_vcn_inst++; adev->vcn.harvest_config = 0; adev->vcn.num_enc_rings = 1; - if (adev->asic_type == CHIP_BEIGE_GOBY) { - adev->vcn.num_vcn_inst = 1; - adev->vcn.num_enc_rings = 0; - } - } else { if (adev->asic_type == CHIP_SIENNA_CICHLID) { u32 harvest; - int i; adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID; for (i = 0; i < adev->vcn.num_vcn_inst; i++) { @@ -156,7 +153,8 @@ static int vcn_v3_0_sw_init(void *handle) adev->firmware.fw_size += ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - if (adev->vcn.num_vcn_inst == VCN_INSTANCES_SIENNA_CICHLID) { + if ((adev->vcn.num_vcn_inst == VCN_INSTANCES_SIENNA_CICHLID) || + (amdgpu_sriov_vf(adev) && adev->asic_type == CHIP_SIENNA_CICHLID)) { adev->firmware.ucode[AMDGPU_UCODE_ID_VCN1].ucode_id = AMDGPU_UCODE_ID_VCN1; adev->firmware.ucode[AMDGPU_UCODE_ID_VCN1].fw = adev->vcn.fw; adev->firmware.fw_size += @@ -330,19 +328,17 @@ static int vcn_v3_0_hw_init(void *handle) continue; ring = &adev->vcn.inst[i].ring_dec; - if (ring->sched.ready) { - ring->wptr = 0; - ring->wptr_old = 0; - vcn_v3_0_dec_ring_set_wptr(ring); - } + ring->wptr = 0; + ring->wptr_old = 0; + vcn_v3_0_dec_ring_set_wptr(ring); + ring->sched.ready = true; for (j = 0; j < adev->vcn.num_enc_rings; ++j) { ring = &adev->vcn.inst[i].ring_enc[j]; - if (ring->sched.ready) { - ring->wptr = 0; - ring->wptr_old = 0; - vcn_v3_0_enc_ring_set_wptr(ring); - } + ring->wptr = 0; + ring->wptr_old = 0; + vcn_v3_0_enc_ring_set_wptr(ring); + ring->sched.ready = true; } } } else { @@ -1309,8 +1305,6 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) uint32_t table_size; uint32_t size, size_dw; - bool is_vcn_ready; - struct mmsch_v3_0_cmd_direct_write direct_wt = { {0} }; struct mmsch_v3_0_cmd_direct_read_modify_write @@ -1502,30 +1496,6 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) } } - /* 6, check each VCN's init_status - * if it remains as 0, then this VCN is not assigned to current VF - * do not start ring for this VCN - */ - size = sizeof(struct mmsch_v3_0_init_header); - table_loc = (uint32_t *)table->cpu_addr; - memcpy(&header, (void *)table_loc, size); - - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - - is_vcn_ready = (header.inst[i].init_status == 1); - if (!is_vcn_ready) - DRM_INFO("VCN(%d) engine is disabled by hypervisor\n", i); - - ring = &adev->vcn.inst[i].ring_dec; - ring->sched.ready = is_vcn_ready; - for (j = 0; j < adev->vcn.num_enc_rings; ++j) { - ring = &adev->vcn.inst[i].ring_enc[j]; - ring->sched.ready = is_vcn_ready; - } - } - return 0; } |