Hi everyone,
re-sending this because Daniel was requesting a background why this is useful.
When TTM creates a buffer this object initially should not have any backing store and there no resource object associated with it. The same can happen when a driver requests that the backing store of an object is destroyed without allocating a new one.
This is really useful during initial buffer creation as well as temporary buffers and page tables which content doesn't need to be preserved when they are evicted.
Currently TTM allocates dummy system resources for that because drivers couldn't handle a NULL pointer there. Audit the drivers and then clean up TTM to stop making those dummy allocations.
Please review and comment, Christian.
Use the new interface instead.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/radeon/radeon_object.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 6c4a6802ca96..1d414ff4ab0c 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -133,9 +133,12 @@ int radeon_bo_create(struct radeon_device *rdev, struct dma_resv *resv, struct radeon_bo **bo_ptr) { - struct radeon_bo *bo; - enum ttm_bo_type type; unsigned long page_align = roundup(byte_align, PAGE_SIZE) >> PAGE_SHIFT; + + /* Kernel allocation are uninterruptible */ + struct ttm_operation_ctx ctx = { !kernel, false }; + enum ttm_bo_type type; + struct radeon_bo *bo; int r;
size = ALIGN(size, PAGE_SIZE); @@ -200,11 +203,13 @@ int radeon_bo_create(struct radeon_device *rdev, #endif
radeon_ttm_placement_from_domain(bo, domain); - /* Kernel allocation are uninterruptible */ down_read(&rdev->pm.mclk_lock); - r = ttm_bo_init(&rdev->mman.bdev, &bo->tbo, size, type, - &bo->placement, page_align, !kernel, sg, resv, - &radeon_ttm_bo_destroy); + r = ttm_bo_init_reserved(&rdev->mman.bdev, &bo->tbo, size, type, + &bo->placement, page_align, &ctx, sg, resv, + &radeon_ttm_bo_destroy); + if (!r) + ttm_bo_unreserve(&bo->tbo); + up_read(&rdev->pm.mclk_lock); if (unlikely(r != 0)) { return r;
Use the new interface instead.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/nouveau/nouveau_bo.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 05076e530e7d..858b9382036c 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -302,19 +302,23 @@ nouveau_bo_init(struct nouveau_bo *nvbo, u64 size, int align, u32 domain, struct sg_table *sg, struct dma_resv *robj) { int type = sg ? ttm_bo_type_sg : ttm_bo_type_device; + struct ttm_operation_ctx ctx = { false, false }; int ret;
nouveau_bo_placement_set(nvbo, domain, 0); INIT_LIST_HEAD(&nvbo->io_reserve_lru);
- ret = ttm_bo_init(nvbo->bo.bdev, &nvbo->bo, size, type, - &nvbo->placement, align >> PAGE_SHIFT, false, sg, - robj, nouveau_bo_del_ttm); + ret = ttm_bo_init_reserved(nvbo->bo.bdev, &nvbo->bo, size, type, + &nvbo->placement, align >> PAGE_SHIFT, &ctx, + sg, robj, nouveau_bo_del_ttm); if (ret) { /* ttm will call nouveau_bo_del_ttm if it fails.. */ return ret; }
+ if (!robj) + ttm_bo_unreserve(&nvbo->bo); + return 0; }
Use the new interface instead.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/drm_gem_vram_helper.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c index 123045b58fec..7449cbc2f925 100644 --- a/drivers/gpu/drm/drm_gem_vram_helper.c +++ b/drivers/gpu/drm/drm_gem_vram_helper.c @@ -186,6 +186,7 @@ struct drm_gem_vram_object *drm_gem_vram_create(struct drm_device *dev, size_t size, unsigned long pg_align) { + struct ttm_operation_ctx ctx = { false, false }; struct drm_gem_vram_object *gbo; struct drm_gem_object *gem; struct drm_vram_mm *vmm = dev->vram_mm; @@ -225,12 +226,13 @@ struct drm_gem_vram_object *drm_gem_vram_create(struct drm_device *dev, * A failing ttm_bo_init will call ttm_buffer_object_destroy * to release gbo->bo.base and kfree gbo. */ - ret = ttm_bo_init(bdev, &gbo->bo, size, ttm_bo_type_device, - &gbo->placement, pg_align, false, NULL, NULL, - ttm_buffer_object_destroy); - if (ret) + ret = ttm_bo_init_reserved(bdev, &gbo->bo, size, ttm_bo_type_device, + &gbo->placement, pg_align, &ctx, NULL, NULL, + ttm_buffer_object_destroy); + if (ret) return ERR_PTR(ret);
+ ttm_bo_unreserve(&gbo->bo); return gbo; } EXPORT_SYMBOL(drm_gem_vram_create);
It's the only driver using this.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/vmwgfx/vmwgfx_bo.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index 85a66014c2b6..c4f376d5e1d0 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -462,6 +462,9 @@ int vmw_bo_create(struct vmw_private *vmw, return -ENOMEM; }
+ if (!bo_free) + bo_free = vmw_bo_default_destroy; + ret = vmw_bo_init(vmw, *p_bo, size, placement, interruptible, pin, bo_free);
Not used any more.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_bo.c | 26 --------------------- include/drm/ttm/ttm_bo_api.h | 44 ------------------------------------ 2 files changed, 70 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 296af2b89951..e652055b5175 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -985,32 +985,6 @@ int ttm_bo_init_reserved(struct ttm_device *bdev, } EXPORT_SYMBOL(ttm_bo_init_reserved);
-int ttm_bo_init(struct ttm_device *bdev, - struct ttm_buffer_object *bo, - size_t size, - enum ttm_bo_type type, - struct ttm_placement *placement, - uint32_t page_alignment, - bool interruptible, - struct sg_table *sg, - struct dma_resv *resv, - void (*destroy) (struct ttm_buffer_object *)) -{ - struct ttm_operation_ctx ctx = { interruptible, false }; - int ret; - - ret = ttm_bo_init_reserved(bdev, bo, size, type, placement, - page_alignment, &ctx, sg, resv, destroy); - if (ret) - return ret; - - if (!resv) - ttm_bo_unreserve(bo); - - return 0; -} -EXPORT_SYMBOL(ttm_bo_init); - /* * buffer object vm functions. */ diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h index 2d524f8b0802..29384e2cb704 100644 --- a/include/drm/ttm/ttm_bo_api.h +++ b/include/drm/ttm/ttm_bo_api.h @@ -361,50 +361,6 @@ int ttm_bo_init_reserved(struct ttm_device *bdev, struct sg_table *sg, struct dma_resv *resv, void (*destroy) (struct ttm_buffer_object *));
-/** - * ttm_bo_init - * - * @bdev: Pointer to a ttm_device struct. - * @bo: Pointer to a ttm_buffer_object to be initialized. - * @size: Requested size of buffer object. - * @type: Requested type of buffer object. - * @placement: Initial placement for buffer object. - * @page_alignment: Data alignment in pages. - * @interruptible: If needing to sleep to wait for GPU resources, - * sleep interruptible. - * pinned in physical memory. If this behaviour is not desired, this member - * holds a pointer to a persistent shmem object. Typically, this would - * point to the shmem object backing a GEM object if TTM is used to back a - * GEM user interface. - * @sg: Scatter-gather table. - * @resv: Pointer to a dma_resv, or NULL to let ttm allocate one. - * @destroy: Destroy function. Use NULL for kfree(). - * - * This function initializes a pre-allocated struct ttm_buffer_object. - * As this object may be part of a larger structure, this function, - * together with the @destroy function, - * enables driver-specific objects derived from a ttm_buffer_object. - * - * On successful return, the caller owns an object kref to @bo. The kref and - * list_kref are usually set to 1, but note that in some situations, other - * tasks may already be holding references to @bo as well. - * - * If a failure occurs, the function will call the @destroy function, or - * kfree() if @destroy is NULL. Thus, after a failure, dereferencing @bo is - * illegal and will likely cause memory corruption. - * - * Returns - * -ENOMEM: Out of memory. - * -EINVAL: Invalid placement flags. - * -ERESTARTSYS: Interrupted by signal while sleeping waiting for resources. - */ -int ttm_bo_init(struct ttm_device *bdev, struct ttm_buffer_object *bo, - size_t size, enum ttm_bo_type type, - struct ttm_placement *placement, - uint32_t page_alignment, bool interrubtible, - struct sg_table *sg, struct dma_resv *resv, - void (*destroy) (struct ttm_buffer_object *)); - /** * ttm_kmap_obj_virtual *
Rename ttm_bo_init_reserved to ttm_bo_init_validate since that better matches what the function is actually doing.
Remove the unused size parameter, move the function's kerneldoc to the implementation and cleanup the whole error handling.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 2 +- drivers/gpu/drm/drm_gem_vram_helper.c | 2 +- drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 5 +- drivers/gpu/drm/nouveau/nouveau_bo.c | 2 +- drivers/gpu/drm/qxl/qxl_object.c | 2 +- drivers/gpu/drm/radeon/radeon_object.c | 2 +- drivers/gpu/drm/ttm/ttm_bo.c | 92 ++++++++++++++-------- drivers/gpu/drm/vmwgfx/vmwgfx_bo.c | 12 ++- include/drm/ttm/ttm_bo_api.h | 44 +---------- 9 files changed, 75 insertions(+), 88 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 5444515c1476..116c8d31e646 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -590,7 +590,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev, if (!bp->destroy) bp->destroy = &amdgpu_bo_destroy;
- r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, bp->type, + r = ttm_bo_init_validate(&adev->mman.bdev, &bo->tbo, bp->type, &bo->placement, page_align, &ctx, NULL, bp->resv, bp->destroy); if (unlikely(r != 0)) diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c index 7449cbc2f925..73e91baccea0 100644 --- a/drivers/gpu/drm/drm_gem_vram_helper.c +++ b/drivers/gpu/drm/drm_gem_vram_helper.c @@ -226,7 +226,7 @@ struct drm_gem_vram_object *drm_gem_vram_create(struct drm_device *dev, * A failing ttm_bo_init will call ttm_buffer_object_destroy * to release gbo->bo.base and kfree gbo. */ - ret = ttm_bo_init_reserved(bdev, &gbo->bo, size, ttm_bo_type_device, + ret = ttm_bo_init_validate(bdev, &gbo->bo, ttm_bo_type_device, &gbo->placement, pg_align, &ctx, NULL, NULL, ttm_buffer_object_destroy); if (ret) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c index 4c25d9b2f138..253188da91eb 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c @@ -1229,9 +1229,8 @@ int __i915_gem_ttm_object_init(struct intel_memory_region *mem, * Similarly, in delayed_destroy, we can't call ttm_bo_put() * until successful initialization. */ - ret = ttm_bo_init_reserved(&i915->bdev, i915_gem_to_ttm(obj), size, - bo_type, &i915_sys_placement, - page_size >> PAGE_SHIFT, + ret = ttm_bo_init_validate(&i915->bdev, i915_gem_to_ttm(obj), bo_type, + &i915_sys_placement, page_size >> PAGE_SHIFT, &ctx, NULL, NULL, i915_ttm_bo_destroy); if (ret) return i915_ttm_err_to_gem(ret); diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 858b9382036c..666941804297 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -308,7 +308,7 @@ nouveau_bo_init(struct nouveau_bo *nvbo, u64 size, int align, u32 domain, nouveau_bo_placement_set(nvbo, domain, 0); INIT_LIST_HEAD(&nvbo->io_reserve_lru);
- ret = ttm_bo_init_reserved(nvbo->bo.bdev, &nvbo->bo, size, type, + ret = ttm_bo_init_validate(nvbo->bo.bdev, &nvbo->bo, type, &nvbo->placement, align >> PAGE_SHIFT, &ctx, sg, robj, nouveau_bo_del_ttm); if (ret) { diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c index b42a657e4c2f..da3f76f508ea 100644 --- a/drivers/gpu/drm/qxl/qxl_object.c +++ b/drivers/gpu/drm/qxl/qxl_object.c @@ -141,7 +141,7 @@ int qxl_bo_create(struct qxl_device *qdev, unsigned long size, qxl_ttm_placement_from_domain(bo, domain);
bo->tbo.priority = priority; - r = ttm_bo_init_reserved(&qdev->mman.bdev, &bo->tbo, size, type, + r = ttm_bo_init_validate(&qdev->mman.bdev, &bo->tbo, type, &bo->placement, 0, &ctx, NULL, NULL, &qxl_ttm_bo_destroy); if (unlikely(r != 0)) { diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 1d414ff4ab0c..550ca056b3ac 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -204,7 +204,7 @@ int radeon_bo_create(struct radeon_device *rdev,
radeon_ttm_placement_from_domain(bo, domain); down_read(&rdev->pm.mclk_lock); - r = ttm_bo_init_reserved(&rdev->mman.bdev, &bo->tbo, size, type, + r = ttm_bo_init_validate(&rdev->mman.bdev, &bo->tbo, type, &bo->placement, page_align, &ctx, sg, resv, &radeon_ttm_bo_destroy); if (!r) diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index e652055b5175..2b01cb30694a 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -915,36 +915,61 @@ int ttm_bo_validate(struct ttm_buffer_object *bo, } EXPORT_SYMBOL(ttm_bo_validate);
-int ttm_bo_init_reserved(struct ttm_device *bdev, - struct ttm_buffer_object *bo, - size_t size, - enum ttm_bo_type type, - struct ttm_placement *placement, - uint32_t page_alignment, - struct ttm_operation_ctx *ctx, - struct sg_table *sg, - struct dma_resv *resv, +/** + * ttm_bo_init_validate + * + * @bdev: Pointer to a ttm_device struct. + * @bo: Pointer to a ttm_buffer_object to be initialized. + * @type: Requested type of buffer object. + * @placement: Initial placement for buffer object. + * @alignment: Data alignment in pages. + * @ctx: TTM operation context for memory allocation. + * @sg: Scatter-gather table. + * @resv: Pointer to a dma_resv, or NULL to let ttm allocate one. + * @destroy: Destroy function. Use NULL for kfree(). + * + * This function initializes a pre-allocated struct ttm_buffer_object. + * As this object may be part of a larger structure, this function, + * together with the @destroy function, enables driver-specific objects + * derived from a ttm_buffer_object. + * + * On successful return, the caller owns an object kref to @bo. The kref and + * list_kref are usually set to 1, but note that in some situations, other + * tasks may already be holding references to @bo as well. + * Furthermore, if resv == NULL, the buffer's reservation lock will be held, + * and it is the caller's responsibility to call ttm_bo_unreserve. + * + * If a failure occurs, the function will call the @destroy function. Thus, + * after a failure, dereferencing @bo is illegal and will likely cause memory + * corruption. + * + * Returns + * -ENOMEM: Out of memory. + * -EINVAL: Invalid placement flags. + * -ERESTARTSYS: Interrupted by signal while sleeping waiting for resources. + */ +int ttm_bo_init_validate(struct ttm_device *bdev, struct ttm_buffer_object *bo, + enum ttm_bo_type type, struct ttm_placement *placement, + uint32_t alignment, struct ttm_operation_ctx *ctx, + struct sg_table *sg, struct dma_resv *resv, void (*destroy) (struct ttm_buffer_object *)) { static const struct ttm_place sys_mem = { .mem_type = TTM_PL_SYSTEM }; - bool locked; int ret;
- bo->destroy = destroy; kref_init(&bo->kref); INIT_LIST_HEAD(&bo->ddestroy); bo->bdev = bdev; bo->type = type; - bo->page_alignment = page_alignment; + bo->page_alignment = alignment; + bo->destroy = destroy; bo->pin_count = 0; bo->sg = sg; bo->bulk_move = NULL; - if (resv) { + if (resv) bo->base.resv = resv; - dma_resv_assert_held(bo->base.resv); - } else { + else bo->base.resv = &bo->base._resv; - } atomic_inc(&ttm_glob.bo_count);
ret = ttm_resource_alloc(bo, &sys_mem, &bo->resource); @@ -957,33 +982,36 @@ int ttm_bo_init_reserved(struct ttm_device *bdev, * For ttm_bo_type_device buffers, allocate * address space from the device. */ - if (bo->type == ttm_bo_type_device || - bo->type == ttm_bo_type_sg) + if (bo->type == ttm_bo_type_device || bo->type == ttm_bo_type_sg) { ret = drm_vma_offset_add(bdev->vma_manager, &bo->base.vma_node, - bo->resource->num_pages); + PFN_UP(bo->base.size)); + if (ret) + goto err_put; + }
/* passed reservation objects should already be locked, * since otherwise lockdep will be angered in radeon. */ - if (!resv) { - locked = dma_resv_trylock(bo->base.resv); - WARN_ON(!locked); - } + if (!resv) + WARN_ON(!dma_resv_trylock(bo->base.resv)); + else + dma_resv_assert_held(resv);
- if (likely(!ret)) - ret = ttm_bo_validate(bo, placement, ctx); + ret = ttm_bo_validate(bo, placement, ctx); + if (unlikely(ret)) + goto err_unlock;
- if (unlikely(ret)) { - if (!resv) - ttm_bo_unreserve(bo); + return 0;
- ttm_bo_put(bo); - return ret; - } +err_unlock: + if (!resv) + dma_resv_unlock(bo->base.resv);
+err_put: + ttm_bo_put(bo); return ret; } -EXPORT_SYMBOL(ttm_bo_init_reserved); +EXPORT_SYMBOL(ttm_bo_init_validate);
/* * buffer object vm functions. diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index c4f376d5e1d0..2bda298e51f0 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -429,9 +429,9 @@ int vmw_bo_create_kernel(struct vmw_private *dev_priv, unsigned long size,
drm_gem_private_object_init(vdev, &bo->base, size);
- ret = ttm_bo_init_reserved(&dev_priv->bdev, bo, size, - ttm_bo_type_kernel, placement, 0, - &ctx, NULL, NULL, vmw_bo_default_destroy); + ret = ttm_bo_init_validate(&dev_priv->bdev, bo, ttm_bo_type_kernel, + placement, 0, &ctx, NULL, NULL, + vmw_bo_default_destroy); if (unlikely(ret)) goto error_free;
@@ -515,10 +515,8 @@ int vmw_bo_init(struct vmw_private *dev_priv, size = ALIGN(size, PAGE_SIZE); drm_gem_private_object_init(vdev, &vmw_bo->base.base, size);
- ret = ttm_bo_init_reserved(bdev, &vmw_bo->base, size, - ttm_bo_type_device, - placement, - 0, &ctx, NULL, NULL, bo_free); + ret = ttm_bo_init_validate(bdev, &vmw_bo->base, ttm_bo_type_device, + placement, 0, &ctx, NULL, NULL, bo_free); if (unlikely(ret)) { return ret; } diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h index 29384e2cb704..d3821a130d74 100644 --- a/include/drm/ttm/ttm_bo_api.h +++ b/include/drm/ttm/ttm_bo_api.h @@ -317,47 +317,9 @@ void ttm_bo_unlock_delayed_workqueue(struct ttm_device *bdev, int resched); bool ttm_bo_eviction_valuable(struct ttm_buffer_object *bo, const struct ttm_place *place);
-/** - * ttm_bo_init_reserved - * - * @bdev: Pointer to a ttm_device struct. - * @bo: Pointer to a ttm_buffer_object to be initialized. - * @size: Requested size of buffer object. - * @type: Requested type of buffer object. - * @placement: Initial placement for buffer object. - * @page_alignment: Data alignment in pages. - * @ctx: TTM operation context for memory allocation. - * @sg: Scatter-gather table. - * @resv: Pointer to a dma_resv, or NULL to let ttm allocate one. - * @destroy: Destroy function. Use NULL for kfree(). - * - * This function initializes a pre-allocated struct ttm_buffer_object. - * As this object may be part of a larger structure, this function, - * together with the @destroy function, - * enables driver-specific objects derived from a ttm_buffer_object. - * - * On successful return, the caller owns an object kref to @bo. The kref and - * list_kref are usually set to 1, but note that in some situations, other - * tasks may already be holding references to @bo as well. - * Furthermore, if resv == NULL, the buffer's reservation lock will be held, - * and it is the caller's responsibility to call ttm_bo_unreserve. - * - * If a failure occurs, the function will call the @destroy function, or - * kfree() if @destroy is NULL. Thus, after a failure, dereferencing @bo is - * illegal and will likely cause memory corruption. - * - * Returns - * -ENOMEM: Out of memory. - * -EINVAL: Invalid placement flags. - * -ERESTARTSYS: Interrupted by signal while sleeping waiting for resources. - */ - -int ttm_bo_init_reserved(struct ttm_device *bdev, - struct ttm_buffer_object *bo, - size_t size, enum ttm_bo_type type, - struct ttm_placement *placement, - uint32_t page_alignment, - struct ttm_operation_ctx *ctx, +int ttm_bo_init_validate(struct ttm_device *bdev, struct ttm_buffer_object *bo, + enum ttm_bo_type type, struct ttm_placement *placement, + uint32_t alignment, struct ttm_operation_ctx *ctx, struct sg_table *sg, struct dma_resv *resv, void (*destroy) (struct ttm_buffer_object *));
Make sure we can at least move and release BOs without backing store.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 3 ++- 2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 116c8d31e646..5cf3a58bc925 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -1302,7 +1302,7 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo) if (bo->base.resv == &bo->base._resv) amdgpu_amdkfd_remove_fence_on_pt_pd_bos(abo);
- if (bo->resource->mem_type != TTM_PL_VRAM || + if (!bo->resource || bo->resource->mem_type != TTM_PL_VRAM || !(abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE) || adev->in_suspend || adev->shutdown) return; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index ec26edd4f4d8..b79c93812342 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -471,7 +471,8 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
adev = amdgpu_ttm_adev(bo->bdev);
- if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) { + if (!old_mem || (old_mem->mem_type == TTM_PL_SYSTEM && + bo->ttm == NULL)) { ttm_bo_move_null(bo, new_mem); goto out; }
Make sure we can at least move and release BOs without backing store.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/nouveau/nouveau_bo.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 666941804297..fb903c62d322 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -1010,7 +1010,8 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, }
/* Fake bo copy. */ - if (old_reg->mem_type == TTM_PL_SYSTEM && !bo->ttm) { + if (!old_reg || (old_reg->mem_type == TTM_PL_SYSTEM && + !bo->ttm)) { ttm_bo_move_null(bo, new_reg); goto out; }
Allow BOs to exist without backing store.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_bo.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 2b01cb30694a..a55564c8b57c 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -117,12 +117,13 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo, struct ttm_operation_ctx *ctx, struct ttm_place *hop) { - struct ttm_resource_manager *old_man, *new_man; struct ttm_device *bdev = bo->bdev; + bool old_use_tt, new_use_tt; int ret;
- old_man = ttm_manager_type(bdev, bo->resource->mem_type); - new_man = ttm_manager_type(bdev, mem->mem_type); + old_use_tt = bo->resource && + ttm_manager_type(bdev, bo->resource->mem_type)->use_tt; + new_use_tt = ttm_manager_type(bdev, mem->mem_type)->use_tt;
ttm_bo_unmap_virtual(bo);
@@ -130,11 +131,11 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo, * Create and bind a ttm if required. */
- if (new_man->use_tt) { + if (new_use_tt) { /* Zero init the new TTM structure if the old location should * have used one as well. */ - ret = ttm_tt_create(bo, old_man->use_tt); + ret = ttm_tt_create(bo, old_use_tt); if (ret) goto out_err;
@@ -160,8 +161,7 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo, return 0;
out_err: - new_man = ttm_manager_type(bdev, bo->resource->mem_type); - if (!new_man->use_tt) + if (!old_use_tt) ttm_bo_tt_destroy(bo);
return ret; @@ -898,7 +898,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo, /* * Check whether we need to move buffer. */ - if (!ttm_resource_compat(bo->resource, placement)) { + if (!bo->resource || !ttm_resource_compat(bo->resource, placement)) { ret = ttm_bo_move_buffer(bo, placement, ctx); if (ret) return ret;
That should not be necessary any more when drivers should at least be able to handle the move without a resource.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_bo.c | 7 ------- 1 file changed, 7 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index a55564c8b57c..31aa4b040d1e 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -954,7 +954,6 @@ int ttm_bo_init_validate(struct ttm_device *bdev, struct ttm_buffer_object *bo, struct sg_table *sg, struct dma_resv *resv, void (*destroy) (struct ttm_buffer_object *)) { - static const struct ttm_place sys_mem = { .mem_type = TTM_PL_SYSTEM }; int ret;
kref_init(&bo->kref); @@ -972,12 +971,6 @@ int ttm_bo_init_validate(struct ttm_device *bdev, struct ttm_buffer_object *bo, bo->base.resv = &bo->base._resv; atomic_inc(&ttm_glob.bo_count);
- ret = ttm_resource_alloc(bo, &sys_mem, &bo->resource); - if (unlikely(ret)) { - ttm_bo_put(bo); - return ret; - } - /* * For ttm_bo_type_device buffers, allocate * address space from the device.
That should not be necessary any more when drivers should at least be able to handle a move without a resource.
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_bo_util.c | 15 ++------------- 1 file changed, 2 insertions(+), 13 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c index 1cbfb00c1d65..585fc529cc75 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_util.c +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c @@ -600,16 +600,10 @@ EXPORT_SYMBOL(ttm_bo_move_sync_cleanup); */ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo) { - static const struct ttm_place sys_mem = { .mem_type = TTM_PL_SYSTEM }; struct ttm_buffer_object *ghost; - struct ttm_resource *sys_res; struct ttm_tt *ttm; int ret;
- ret = ttm_resource_alloc(bo, &sys_mem, &sys_res); - if (ret) - return ret; - /* If already idle, no need for ghost object dance. */ ret = ttm_bo_wait(bo, false, true); if (ret != -EBUSY) { @@ -617,14 +611,13 @@ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo) /* See comment below about clearing. */ ret = ttm_tt_create(bo, true); if (ret) - goto error_free_sys_mem; + return ret; } else { ttm_tt_unpopulate(bo->bdev, bo->ttm); if (bo->type == ttm_bo_type_device) ttm_tt_mark_for_clear(bo->ttm); } ttm_resource_free(bo, &bo->resource); - ttm_bo_assign_mem(bo, sys_res); return 0; }
@@ -641,7 +634,7 @@ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo) ret = ttm_tt_create(bo, true); swap(bo->ttm, ttm); if (ret) - goto error_free_sys_mem; + return ret;
ret = ttm_buffer_object_transfer(bo, &ghost); if (ret) @@ -655,13 +648,9 @@ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo) dma_resv_unlock(&ghost->base._resv); ttm_bo_put(ghost); bo->ttm = ttm; - ttm_bo_assign_mem(bo, sys_res); return 0;
error_destroy_tt: ttm_tt_destroy(bo->bdev, ttm); - -error_free_sys_mem: - ttm_resource_free(bo, &sys_res); return ret; }
On Mon, May 09, 2022 at 03:09:40PM +0200, Christian König wrote:
Hi everyone,
re-sending this because Daniel was requesting a background why this is useful.
Thanks a lot for typing this up. To make sure this isn't lost, could you add a patch to at least add this as a comment to ttm_buffer_object.resource, ideally rolling out the kerneldoc scaffolding for that file while at it and making it kerneldoc?
I think documenting that kind of why&how for key links like ttm_bo->ttm_resource is really important.
When TTM creates a buffer this object initially should not have any backing store and there no resource object associated with it. The same can happen when a driver requests that the backing store of an object is destroyed without allocating a new one.
This is really useful during initial buffer creation as well as temporary buffers and page tables which content doesn't need to be preserved when they are evicted.
Currently TTM allocates dummy system resources for that because drivers couldn't handle a NULL pointer there. Audit the drivers and then clean up TTM to stop making those dummy allocations.
Also I guess this is good prep work for going the other direction, i.e. allowing more than one ttm_resource per ttm_bo? Or is that all mostly orthogonal issues that bo:resource 1:N for N=0 has vs 1:N for N > 1?
Please review and comment,
I'll ... try. But maybe call in some favours from other i915 ttm folks because I'm just really bad at doing review timeline these days :-/ -Daniel
Patch set reviewed. Good stuff.
Acked-by: Luben Tuikov luben.tuikov@amd.com
Regards, Luben
On 2022-05-09 09:09, Christian König wrote:
Hi everyone,
re-sending this because Daniel was requesting a background why this is useful.
When TTM creates a buffer this object initially should not have any backing store and there no resource object associated with it. The same can happen when a driver requests that the backing store of an object is destroyed without allocating a new one.
This is really useful during initial buffer creation as well as temporary buffers and page tables which content doesn't need to be preserved when they are evicted.
Currently TTM allocates dummy system resources for that because drivers couldn't handle a NULL pointer there. Audit the drivers and then clean up TTM to stop making those dummy allocations.
Please review and comment, Christian.
On Mon, 9 May 2022 at 14:09, Christian König ckoenig.leichtzumerken@gmail.com wrote:
Hi everyone,
re-sending this because Daniel was requesting a background why this is useful.
When TTM creates a buffer this object initially should not have any backing store and there no resource object associated with it. The same can happen when a driver requests that the backing store of an object is destroyed without allocating a new one.
This is really useful during initial buffer creation as well as temporary buffers and page tables which content doesn't need to be preserved when they are evicted.
Currently TTM allocates dummy system resources for that because drivers couldn't handle a NULL pointer there. Audit the drivers and then clean up TTM to stop making those dummy allocations.
Please review and comment,
Any chance we can throw this at intel-ci? Series seems to apply to drm-tip.
Christian.
Am 13.05.22 um 11:21 schrieb Matthew Auld:
On Mon, 9 May 2022 at 14:09, Christian König ckoenig.leichtzumerken@gmail.com wrote:
Hi everyone,
re-sending this because Daniel was requesting a background why this is useful.
When TTM creates a buffer this object initially should not have any backing store and there no resource object associated with it. The same can happen when a driver requests that the backing store of an object is destroyed without allocating a new one.
This is really useful during initial buffer creation as well as temporary buffers and page tables which content doesn't need to be preserved when they are evicted.
Currently TTM allocates dummy system resources for that because drivers couldn't handle a NULL pointer there. Audit the drivers and then clean up TTM to stop making those dummy allocations.
Please review and comment,
Any chance we can throw this at intel-ci? Series seems to apply to drm-tip.
It should be enough to CC intel-gfx@lists.freedesktop.org when sending the patch set out, right?
Christian.
Christian.
On Fri, 13 May 2022 at 14:03, Christian König ckoenig.leichtzumerken@gmail.com wrote:
Am 13.05.22 um 11:21 schrieb Matthew Auld:
On Mon, 9 May 2022 at 14:09, Christian König ckoenig.leichtzumerken@gmail.com wrote:
Hi everyone,
re-sending this because Daniel was requesting a background why this is useful.
When TTM creates a buffer this object initially should not have any backing store and there no resource object associated with it. The same can happen when a driver requests that the backing store of an object is destroyed without allocating a new one.
This is really useful during initial buffer creation as well as temporary buffers and page tables which content doesn't need to be preserved when they are evicted.
Currently TTM allocates dummy system resources for that because drivers couldn't handle a NULL pointer there. Audit the drivers and then clean up TTM to stop making those dummy allocations.
Please review and comment,
Any chance we can throw this at intel-ci? Series seems to apply to drm-tip.
It should be enough to CC intel-gfx@lists.freedesktop.org when sending the patch set out, right?
Yes, that should be enough.
Christian.
Christian.
dri-devel@lists.freedesktop.org
-
Christian König -
Daniel Vetter -
Luben Tuikov -
Matthew Auld