Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 2d086fde authored by Felix Kuehling's avatar Felix Kuehling Committed by Alex Deucher
Browse files

drm/amdgpu: Simplify eviction fence handling



Temporarily removing eviction fences to avoid triggering them by
accident is no longer necessary due to the fence_owner logic in
amdgpu_sync_resv.

As a result the ef_list usage of amdgpu_amdkfd_remove_eviction_fence
and amdgpu_amdkfd_add_eviction_fence are no longer needed.

Signed-off-by: default avatarFelix Kuehling <Felix.Kuehling@amd.com>
Acked-by: default avatarChristian König <christian.koenig@amd.com>
Reviewed-by: default avatarHarish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
parent 8db588d5
Loading
Loading
Loading
Loading
+11 −123
Original line number Original line Diff line number Diff line
@@ -204,38 +204,25 @@ void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
}
}




/* amdgpu_amdkfd_remove_eviction_fence - Removes eviction fence(s) from BO's
/* amdgpu_amdkfd_remove_eviction_fence - Removes eviction fence from BO's
 *  reservation object.
 *  reservation object.
 *
 *
 * @bo: [IN] Remove eviction fence(s) from this BO
 * @bo: [IN] Remove eviction fence(s) from this BO
 * @ef: [IN] If ef is specified, then this eviction fence is removed if it
 * @ef: [IN] This eviction fence is removed if it
 *  is present in the shared list.
 *  is present in the shared list.
 * @ef_list: [OUT] Returns list of eviction fences. These fences are removed
 *  from BO's reservation object shared list.
 * @ef_count: [OUT] Number of fences in ef_list.
 *
 *
 * NOTE: If called with ef_list, then amdgpu_amdkfd_add_eviction_fence must be
 *  called to restore the eviction fences and to avoid memory leak. This is
 *  useful for shared BOs.
 * NOTE: Must be called with BO reserved i.e. bo->tbo.resv->lock held.
 * NOTE: Must be called with BO reserved i.e. bo->tbo.resv->lock held.
 */
 */
static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
					struct amdgpu_amdkfd_fence *ef,
					struct amdgpu_amdkfd_fence *ef)
					struct amdgpu_amdkfd_fence ***ef_list,
					unsigned int *ef_count)
{
{
	struct reservation_object *resv = bo->tbo.resv;
	struct reservation_object *resv = bo->tbo.resv;
	struct reservation_object_list *old, *new;
	struct reservation_object_list *old, *new;
	unsigned int i, j, k;
	unsigned int i, j, k;


	if (!ef && !ef_list)
	if (!ef)
		return -EINVAL;
		return -EINVAL;


	if (ef_list) {
		*ef_list = NULL;
		*ef_count = 0;
	}

	old = reservation_object_get_list(resv);
	old = reservation_object_get_list(resv);
	if (!old)
	if (!old)
		return 0;
		return 0;
@@ -254,8 +241,7 @@ static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
		f = rcu_dereference_protected(old->shared[i],
		f = rcu_dereference_protected(old->shared[i],
					      reservation_object_held(resv));
					      reservation_object_held(resv));


		if ((ef && f->context == ef->base.context) ||
		if (f->context == ef->base.context)
		    (!ef && to_amdgpu_amdkfd_fence(f)))
			RCU_INIT_POINTER(new->shared[--j], f);
			RCU_INIT_POINTER(new->shared[--j], f);
		else
		else
			RCU_INIT_POINTER(new->shared[k++], f);
			RCU_INIT_POINTER(new->shared[k++], f);
@@ -263,21 +249,6 @@ static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
	new->shared_max = old->shared_max;
	new->shared_max = old->shared_max;
	new->shared_count = k;
	new->shared_count = k;


	if (!ef) {
		unsigned int count = old->shared_count - j;

		/* Alloc memory for count number of eviction fence pointers.
		 * Fill the ef_list array and ef_count
		 */
		*ef_list = kcalloc(count, sizeof(**ef_list), GFP_KERNEL);
		*ef_count = count;

		if (!*ef_list) {
			kfree(new);
			return -ENOMEM;
		}
	}

	/* Install the new fence list, seqcount provides the barriers */
	/* Install the new fence list, seqcount provides the barriers */
	preempt_disable();
	preempt_disable();
	write_seqcount_begin(&resv->seq);
	write_seqcount_begin(&resv->seq);
@@ -291,9 +262,6 @@ static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,


		f = rcu_dereference_protected(new->shared[i],
		f = rcu_dereference_protected(new->shared[i],
					      reservation_object_held(resv));
					      reservation_object_held(resv));
		if (!ef)
			(*ef_list)[k++] = to_amdgpu_amdkfd_fence(f);
		else
		dma_fence_put(f);
		dma_fence_put(f);
	}
	}
	kfree_rcu(old, rcu);
	kfree_rcu(old, rcu);
@@ -301,36 +269,6 @@ static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
	return 0;
	return 0;
}
}


/* amdgpu_amdkfd_add_eviction_fence - Adds eviction fence(s) back into BO's
 *  reservation object.
 *
 * @bo: [IN] Add eviction fences to this BO
 * @ef_list: [IN] List of eviction fences to be added
 * @ef_count: [IN] Number of fences in ef_list.
 *
 * NOTE: Must call amdgpu_amdkfd_remove_eviction_fence before calling this
 *  function.
 */
static void amdgpu_amdkfd_add_eviction_fence(struct amdgpu_bo *bo,
				struct amdgpu_amdkfd_fence **ef_list,
				unsigned int ef_count)
{
	int i;

	if (!ef_list || !ef_count)
		return;

	for (i = 0; i < ef_count; i++) {
		amdgpu_bo_fence(bo, &ef_list[i]->base, true);
		/* Re-adding the fence takes an additional reference. Drop that
		 * reference.
		 */
		dma_fence_put(&ef_list[i]->base);
	}

	kfree(ef_list);
}

static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain,
static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain,
				     bool wait)
				     bool wait)
{
{
@@ -346,18 +284,8 @@ static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain,
	ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
	ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
	if (ret)
	if (ret)
		goto validate_fail;
		goto validate_fail;
	if (wait) {
	if (wait)
		struct amdgpu_amdkfd_fence **ef_list;
		unsigned int ef_count;

		ret = amdgpu_amdkfd_remove_eviction_fence(bo, NULL, &ef_list,
							  &ef_count);
		if (ret)
			goto validate_fail;

		amdgpu_bo_sync_wait(bo, AMDGPU_FENCE_OWNER_KFD, false);
		amdgpu_bo_sync_wait(bo, AMDGPU_FENCE_OWNER_KFD, false);
		amdgpu_amdkfd_add_eviction_fence(bo, ef_list, ef_count);
	}


validate_fail:
validate_fail:
	return ret;
	return ret;
@@ -444,7 +372,6 @@ static int add_bo_to_vm(struct amdgpu_device *adev, struct kgd_mem *mem,
{
{
	int ret;
	int ret;
	struct kfd_bo_va_list *bo_va_entry;
	struct kfd_bo_va_list *bo_va_entry;
	struct amdgpu_bo *pd = vm->root.base.bo;
	struct amdgpu_bo *bo = mem->bo;
	struct amdgpu_bo *bo = mem->bo;
	uint64_t va = mem->va;
	uint64_t va = mem->va;
	struct list_head *list_bo_va = &mem->bo_va_list;
	struct list_head *list_bo_va = &mem->bo_va_list;
@@ -484,14 +411,8 @@ static int add_bo_to_vm(struct amdgpu_device *adev, struct kgd_mem *mem,
		*p_bo_va_entry = bo_va_entry;
		*p_bo_va_entry = bo_va_entry;


	/* Allocate new page tables if needed and validate
	/* Allocate new page tables if needed and validate
	 * them. Clearing of new page tables and validate need to wait
	 * them.
	 * on move fences. We don't want that to trigger the eviction
	 * fence, so remove it temporarily.
	 */
	 */
	amdgpu_amdkfd_remove_eviction_fence(pd,
					vm->process_info->eviction_fence,
					NULL, NULL);

	ret = amdgpu_vm_alloc_pts(adev, vm, va, amdgpu_bo_size(bo));
	ret = amdgpu_vm_alloc_pts(adev, vm, va, amdgpu_bo_size(bo));
	if (ret) {
	if (ret) {
		pr_err("Failed to allocate pts, err=%d\n", ret);
		pr_err("Failed to allocate pts, err=%d\n", ret);
@@ -504,13 +425,9 @@ static int add_bo_to_vm(struct amdgpu_device *adev, struct kgd_mem *mem,
		goto err_alloc_pts;
		goto err_alloc_pts;
	}
	}


	/* Add the eviction fence back */
	amdgpu_bo_fence(pd, &vm->process_info->eviction_fence->base, true);

	return 0;
	return 0;


err_alloc_pts:
err_alloc_pts:
	amdgpu_bo_fence(pd, &vm->process_info->eviction_fence->base, true);
	amdgpu_vm_bo_rmv(adev, bo_va_entry->bo_va);
	amdgpu_vm_bo_rmv(adev, bo_va_entry->bo_va);
	list_del(&bo_va_entry->bo_list);
	list_del(&bo_va_entry->bo_list);
err_vmadd:
err_vmadd:
@@ -809,24 +726,11 @@ static int unmap_bo_from_gpuvm(struct amdgpu_device *adev,
{
{
	struct amdgpu_bo_va *bo_va = entry->bo_va;
	struct amdgpu_bo_va *bo_va = entry->bo_va;
	struct amdgpu_vm *vm = bo_va->base.vm;
	struct amdgpu_vm *vm = bo_va->base.vm;
	struct amdgpu_bo *pd = vm->root.base.bo;


	/* Remove eviction fence from PD (and thereby from PTs too as
	 * they share the resv. object). Otherwise during PT update
	 * job (see amdgpu_vm_bo_update_mapping), eviction fence would
	 * get added to job->sync object and job execution would
	 * trigger the eviction fence.
	 */
	amdgpu_amdkfd_remove_eviction_fence(pd,
					    vm->process_info->eviction_fence,
					    NULL, NULL);
	amdgpu_vm_bo_unmap(adev, bo_va, entry->va);
	amdgpu_vm_bo_unmap(adev, bo_va, entry->va);


	amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
	amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);


	/* Add the eviction fence back */
	amdgpu_bo_fence(pd, &vm->process_info->eviction_fence->base, true);

	amdgpu_sync_fence(NULL, sync, bo_va->last_pt_update, false);
	amdgpu_sync_fence(NULL, sync, bo_va->last_pt_update, false);


	return 0;
	return 0;
@@ -1389,8 +1293,7 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
	 * attached
	 * attached
	 */
	 */
	amdgpu_amdkfd_remove_eviction_fence(mem->bo,
	amdgpu_amdkfd_remove_eviction_fence(mem->bo,
					process_info->eviction_fence,
					process_info->eviction_fence);
					NULL, NULL);
	pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va,
	pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va,
		mem->va + bo_size * (1 + mem->aql_queue));
		mem->va + bo_size * (1 + mem->aql_queue));


@@ -1617,8 +1520,7 @@ int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
	if (mem->mapped_to_gpu_memory == 0 &&
	if (mem->mapped_to_gpu_memory == 0 &&
	    !amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm) && !mem->bo->pin_count)
	    !amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm) && !mem->bo->pin_count)
		amdgpu_amdkfd_remove_eviction_fence(mem->bo,
		amdgpu_amdkfd_remove_eviction_fence(mem->bo,
						process_info->eviction_fence,
						process_info->eviction_fence);
						    NULL, NULL);


unreserve_out:
unreserve_out:
	unreserve_bo_and_vms(&ctx, false, false);
	unreserve_bo_and_vms(&ctx, false, false);
@@ -1679,7 +1581,7 @@ int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
	}
	}


	amdgpu_amdkfd_remove_eviction_fence(
	amdgpu_amdkfd_remove_eviction_fence(
		bo, mem->process_info->eviction_fence, NULL, NULL);
		bo, mem->process_info->eviction_fence);
	list_del_init(&mem->validate_list.head);
	list_del_init(&mem->validate_list.head);


	if (size)
	if (size)
@@ -1945,16 +1847,6 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)


	amdgpu_sync_create(&sync);
	amdgpu_sync_create(&sync);


	/* Avoid triggering eviction fences when unmapping invalid
	 * userptr BOs (waits for all fences, doesn't use
	 * FENCE_OWNER_VM)
	 */
	list_for_each_entry(peer_vm, &process_info->vm_list_head,
			    vm_list_node)
		amdgpu_amdkfd_remove_eviction_fence(peer_vm->root.base.bo,
						process_info->eviction_fence,
						NULL, NULL);

	ret = process_validate_vms(process_info);
	ret = process_validate_vms(process_info);
	if (ret)
	if (ret)
		goto unreserve_out;
		goto unreserve_out;
@@ -2015,10 +1907,6 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
	ret = process_update_pds(process_info, &sync);
	ret = process_update_pds(process_info, &sync);


unreserve_out:
unreserve_out:
	list_for_each_entry(peer_vm, &process_info->vm_list_head,
			    vm_list_node)
		amdgpu_bo_fence(peer_vm->root.base.bo,
				&process_info->eviction_fence->base, true);
	ttm_eu_backoff_reservation(&ticket, &resv_list);
	ttm_eu_backoff_reservation(&ticket, &resv_list);
	amdgpu_sync_wait(&sync, false);
	amdgpu_sync_wait(&sync, false);
	amdgpu_sync_free(&sync);
	amdgpu_sync_free(&sync);