drm/ttm: rework bulk move handling v5

Instead of providing the bulk move structure for each LRU update set
this as property of the BO. This should avoid costly bulk move rebuilds
with some games under RADV.

v2: some name polishing, add a few more kerneldoc words.
v3: add some lockdep
v4: fix bugs, handle pin/unpin as well
v5: improve kerneldoc

Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220321132601.2161-5-christian.koenig@amd.com

3 files changed, 10 insertions, 66 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 5859ed0552a4..57ac118fc266 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1498,7 +1498,6 @@ static struct ttm_device_funcs amdgpu_bo_driver = {
 	.io_mem_reserve = &amdgpu_ttm_io_mem_reserve,
 	.io_mem_pfn = amdgpu_ttm_io_mem_pfn,
 	.access_memory = &amdgpu_ttm_access_memory,
-	.del_from_lru_notify = &amdgpu_vm_del_from_lru_notify
 };
 
 /*
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index f2ce5a0defd9..28f5e8b21a99 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -375,7 +375,7 @@ static void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base,
 	if (bo->tbo.base.resv != vm->root.bo->tbo.base.resv)
 		return;
 
-	vm->bulk_moveable = false;
+	ttm_bo_set_bulk_move(&bo->tbo, &vm->lru_bulk_move);
 	if (bo->tbo.type == ttm_bo_type_kernel && bo->parent)
 		amdgpu_vm_bo_relocated(base);
 	else
@@ -638,36 +638,6 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 }
 
 /**
- * amdgpu_vm_del_from_lru_notify - update bulk_moveable flag
- *
- * @bo: BO which was removed from the LRU
- *
- * Make sure the bulk_moveable flag is updated when a BO is removed from the
- * LRU.
- */
-void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo)
-{
-	struct amdgpu_bo *abo;
-	struct amdgpu_vm_bo_base *bo_base;
-
-	if (!amdgpu_bo_is_amdgpu_bo(bo))
-		return;
-
-	if (bo->pin_count)
-		return;
-
-	abo = ttm_to_amdgpu_bo(bo);
-	if (!abo->parent)
-		return;
-	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.bo->tbo.base.resv)
-			vm->bulk_moveable = false;
-	}
-
-}
-/**
  * amdgpu_vm_move_to_lru_tail - move all BOs to the end of LRU
  *
  * @adev: amdgpu device pointer
@@ -679,33 +649,9 @@ void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo)
 void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm)
 {
-	struct amdgpu_vm_bo_base *bo_base;
-
-	if (vm->bulk_moveable) {
-		spin_lock(&adev->mman.bdev.lru_lock);
-		ttm_lru_bulk_move_tail(&vm->lru_bulk_move);
-		spin_unlock(&adev->mman.bdev.lru_lock);
-		return;
-	}
-
-	ttm_lru_bulk_move_init(&vm->lru_bulk_move);
-
 	spin_lock(&adev->mman.bdev.lru_lock);
-	list_for_each_entry(bo_base, &vm->idle, vm_status) {
-		struct amdgpu_bo *bo = bo_base->bo;
-		struct amdgpu_bo *shadow = amdgpu_bo_shadowed(bo);
-
-		if (!bo->parent)
-			continue;
-
-		ttm_bo_move_to_lru_tail(&bo->tbo, &vm->lru_bulk_move);
-		if (shadow)
-			ttm_bo_move_to_lru_tail(&shadow->tbo,
-						&vm->lru_bulk_move);
-	}
+	ttm_lru_bulk_move_tail(&vm->lru_bulk_move);
 	spin_unlock(&adev->mman.bdev.lru_lock);
-
-	vm->bulk_moveable = true;
 }
 
 /**
@@ -728,8 +674,6 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	struct amdgpu_vm_bo_base *bo_base, *tmp;
 	int r;
 
-	vm->bulk_moveable &= list_empty(&vm->evicted);
-
 	list_for_each_entry_safe(bo_base, tmp, &vm->evicted, vm_status) {
 		struct amdgpu_bo *bo = bo_base->bo;
 		struct amdgpu_bo *shadow = amdgpu_bo_shadowed(bo);
@@ -1047,10 +991,16 @@ static void amdgpu_vm_free_table(struct amdgpu_vm_bo_base *entry)
 
 	if (!entry->bo)
 		return;
+
 	shadow = amdgpu_bo_shadowed(entry->bo);
+	if (shadow) {
+		ttm_bo_set_bulk_move(&shadow->tbo, NULL);
+		amdgpu_bo_unref(&shadow);
+	}
+
+	ttm_bo_set_bulk_move(&entry->bo->tbo, NULL);
 	entry->bo->vm_bo = NULL;
 	list_del(&entry->vm_status);
-	amdgpu_bo_unref(&shadow);
 	amdgpu_bo_unref(&entry->bo);
 }
 
@@ -1070,8 +1020,6 @@ static void amdgpu_vm_free_pts(struct amdgpu_device *adev,
 	struct amdgpu_vm_pt_cursor cursor;
 	struct amdgpu_vm_bo_base *entry;
 
-	vm->bulk_moveable = false;
-
 	for_each_amdgpu_vm_pt_dfs_safe(adev, vm, start, cursor, entry)
 		amdgpu_vm_free_table(entry);
 
@@ -2651,7 +2599,7 @@ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
 
 	if (bo) {
 		if (bo->tbo.base.resv == vm->root.bo->tbo.base.resv)
-			vm->bulk_moveable = false;
+			ttm_bo_set_bulk_move(&bo->tbo, NULL);
 
 		for (base = &bo_va->base.bo->vm_bo; *base;
 		     base = &(*base)->next) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 85fcfb8c5efd..4d236682a118 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -317,8 +317,6 @@ struct amdgpu_vm {
 
 	/* Store positions of group of BOs */
 	struct ttm_lru_bulk_move lru_bulk_move;
-	/* mark whether can do the bulk move */
-	bool			bulk_moveable;
 	/* Flag to indicate if VM is used for compute */
 	bool			is_compute_context;
 };
@@ -454,7 +452,6 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm);
 
 void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
-void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo);
 void amdgpu_vm_get_memory(struct amdgpu_vm *vm, uint64_t *vram_mem,
 				uint64_t *gtt_mem, uint64_t *cpu_mem);