Commit 377b5b39 authored by Dave Airlie's avatar Dave Airlie
Browse files

Merge tag 'amd-drm-next-6.10-2024-04-19' of... 

Merge tag 'amd-drm-next-6.10-2024-04-19' of https://gitlab.freedesktop.org/agd5f/linux

 into drm-next

amd-drm-next-6.10-2024-04-19:

amdgpu:
- DC resource allocation logic updates
- DC IPS fixes
- DC YUV fixes
- DMCUB fixes
- DML2 fixes
- Devcoredump updates
- USB-C DSC fix
- Misc display code cleanups
- PSR fixes
- MES timeout fix
- RAS updates
- UAF fix in VA IOCTL
- Fix visible VRAM handling during faults
- Fix IP discovery handling during PCI rescans
- Misc code cleanups
- PSP 14 updates
- More runtime PM code rework
- SMU 14.0.2 support
- GPUVM page fault redirection to secondary IH rings for IH 6.x
- Suspend/resume fixes
- SR-IOV fixes

amdkfd:
- Fix eviction fence handling
- Fix leak in GPU memory allocation failure case
- DMABuf import handling fix

radeon:
- Silence UBSAN warnings related to flexible arrays

Signed-off-by: default avatarDave Airlie <airlied@redhat.com>

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240419224332.2938259-1-alexander.deucher@amd.com
parents fad3dad8 81bf1451

drivers/gpu/drm/amd/amdgpu/amdgpu.h

+1 −0
Original line number Diff line number Diff line
@@ -1409,6 +1409,7 @@ bool amdgpu_device_supports_px(struct drm_device *dev); 
bool amdgpu_device_supports_boco(struct drm_device *dev);

bool amdgpu_device_supports_smart_shift(struct drm_device *dev);

int amdgpu_device_supports_baco(struct drm_device *dev);

void amdgpu_device_detect_runtime_pm_mode(struct amdgpu_device *adev);

bool amdgpu_device_is_peer_accessible(struct amdgpu_device *adev,

				      struct amdgpu_device *peer_adev);

int amdgpu_device_baco_enter(struct drm_device *dev);

drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c

+8 −15
Original line number Diff line number Diff line
@@ -753,23 +753,13 @@ int aca_bank_info_decode(struct aca_bank *bank, struct aca_bank_info *info) 


static int aca_bank_get_error_code(struct amdgpu_device *adev, struct aca_bank *bank)

{

	int error_code;



	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {

	case IP_VERSION(13, 0, 6):

		if (!(adev->flags & AMD_IS_APU) && adev->pm.fw_version >= 0x00555600) {

			error_code = ACA_REG__SYND__ERRORINFORMATION(bank->regs[ACA_REG_IDX_SYND]);

			return error_code & 0xff;

		}

		break;

	default:

		break;

	}

	struct amdgpu_aca *aca = &adev->aca;

	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;



	/* NOTE: the true error code is encoded in status.errorcode[0:7] */

	error_code = ACA_REG__STATUS__ERRORCODE(bank->regs[ACA_REG_IDX_STATUS]);

	if (!smu_funcs || !smu_funcs->parse_error_code)

		return -EOPNOTSUPP;



	return error_code & 0xff;

	return smu_funcs->parse_error_code(adev, bank);

}



int aca_bank_check_error_codes(struct amdgpu_device *adev, struct aca_bank *bank, int *err_codes, int size)
@@ -780,6 +770,9 @@ int aca_bank_check_error_codes(struct amdgpu_device *adev, struct aca_bank *bank 
		return -EINVAL;



	error_code = aca_bank_get_error_code(adev, bank);

	if (error_code < 0)

		return error_code;



	for (i = 0; i < size; i++) {

		if (err_codes[i] == error_code)

			return 0;

drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h

+1 −0
Original line number Diff line number Diff line
@@ -173,6 +173,7 @@ struct aca_smu_funcs { 
	int (*set_debug_mode)(struct amdgpu_device *adev, bool enable);

	int (*get_valid_aca_count)(struct amdgpu_device *adev, enum aca_smu_type type, u32 *count);

	int (*get_valid_aca_bank)(struct amdgpu_device *adev, enum aca_smu_type type, int idx, struct aca_bank *bank);

	int (*parse_error_code)(struct amdgpu_device *adev, struct aca_bank *bank);

};



struct amdgpu_aca {

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c

+21 −14
Original line number Diff line number Diff line
@@ -1854,6 +1854,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( 
err_bo_create:

	amdgpu_amdkfd_unreserve_mem_limit(adev, aligned_size, flags, xcp_id);

err_reserve_limit:

	amdgpu_sync_free(&(*mem)->sync);

	mutex_destroy(&(*mem)->lock);

	if (gobj)

		drm_gem_object_put(gobj);
@@ -2900,13 +2901,12 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu * 


	amdgpu_sync_create(&sync_obj);



	/* Validate BOs and map them to GPUVM (update VM page tables). */

	/* Validate BOs managed by KFD */

	list_for_each_entry(mem, &process_info->kfd_bo_list,

			    validate_list) {



		struct amdgpu_bo *bo = mem->bo;

		uint32_t domain = mem->domain;

		struct kfd_mem_attachment *attachment;

		struct dma_resv_iter cursor;

		struct dma_fence *fence;
@@ -2931,6 +2931,25 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu * 
				goto validate_map_fail;

			}

		}

	}



	if (failed_size)

		pr_debug("0x%lx/0x%lx in system\n", failed_size, total_size);



	/* Validate PDs, PTs and evicted DMABuf imports last. Otherwise BO

	 * validations above would invalidate DMABuf imports again.

	 */

	ret = process_validate_vms(process_info, &exec.ticket);

	if (ret) {

		pr_debug("Validating VMs failed, ret: %d\n", ret);

		goto validate_map_fail;

	}



	/* Update mappings managed by KFD. */

	list_for_each_entry(mem, &process_info->kfd_bo_list,

			    validate_list) {

		struct kfd_mem_attachment *attachment;



		list_for_each_entry(attachment, &mem->attachments, list) {

			if (!attachment->is_mapped)

				continue;
@@ -2947,18 +2966,6 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu * 
		}

	}



	if (failed_size)

		pr_debug("0x%lx/0x%lx in system\n", failed_size, total_size);



	/* Validate PDs, PTs and evicted DMABuf imports last. Otherwise BO

	 * validations above would invalidate DMABuf imports again.

	 */

	ret = process_validate_vms(process_info, &exec.ticket);

	if (ret) {

		pr_debug("Validating VMs failed, ret: %d\n", ret);

		goto validate_map_fail;

	}



	/* Update mappings not managed by KFD */

	list_for_each_entry(peer_vm, &process_info->vm_list_head,

			vm_list_node) {

drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

+1 −1
Original line number Diff line number Diff line
@@ -819,7 +819,7 @@ static int amdgpu_cs_bo_validate(void *param, struct amdgpu_bo *bo) 


	p->bytes_moved += ctx.bytes_moved;

	if (!amdgpu_gmc_vram_full_visible(&adev->gmc) &&

	    amdgpu_bo_in_cpu_visible_vram(bo))

	    amdgpu_res_cpu_visible(adev, bo->tbo.resource))

		p->bytes_moved_vis += ctx.bytes_moved;



	if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {