drm/amdgpu: add independent DMA-buf export v8

Add an DMA-buf export implementation independent of the DRM helpers.

This not only avoids the caching of DMA-buf mappings, but also
allows us to use the new dynamic locking approach.

This is also a prerequisite of unpinned DMA-buf handling.

v2: fix unintended recursion, remove debugging leftovers
v3: split out from unpinned DMA-buf work
v4: rebase on top of new no_sgt_cache flag
v5: fix some warnings by including amdgpu_dma_buf.h
v6: fix locking for non amdgpu exports
v7: rebased on new DMA-buf locking patch
v8: drop extra include
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/337949/

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

@@ -34,26 +34,11 @@
 #include "amdgpu.h"
 #include "amdgpu_display.h"
 #include "amdgpu_gem.h"
+#include "amdgpu_dma_buf.h"
 #include <drm/amdgpu_drm.h>
 #include <linux/dma-buf.h>
 #include <linux/dma-fence-array.h>
 
-/**
- * amdgpu_gem_prime_get_sg_table - &drm_driver.gem_prime_get_sg_table
- * implementation
- * @obj: GEM buffer object (BO)
- *
- * Returns:
- * A scatter/gather table for the pinned pages of the BO's memory.
- */
-struct sg_table *amdgpu_gem_prime_get_sg_table(struct drm_gem_object *obj)
-{
-	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
-	int npages = bo->tbo.num_pages;
-
-	return drm_prime_pages_to_sg(bo->tbo.ttm->pages, npages);
-}
-
 /**
  * amdgpu_gem_prime_vmap - &dma_buf_ops.vmap implementation
  * @obj: GEM BO
@@ -179,35 +164,28 @@ __dma_resv_make_exclusive(struct dma_resv *obj)
 }
 
 /**
- * amdgpu_dma_buf_map_attach - &dma_buf_ops.attach implementation
- * @dma_buf: Shared DMA buffer
- * @attach: DMA-buf attachment
+ * amdgpu_dma_buf_attach - &dma_buf_ops.attach implementation
  *
- * Makes sure that the shared DMA buffer can be accessed by the target device.
- * For now, simply pins it to the GTT domain, where it should be accessible by
- * all DMA devices.
+ * @dmabuf: DMA-buf where we attach to
+ * @attach: attachment to add
  *
- * Returns:
- * 0 on success or a negative error code on failure.
+ * Add the attachment as user to the exported DMA-buf.
  */
-static int amdgpu_dma_buf_map_attach(struct dma_buf *dma_buf,
+static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 				 struct dma_buf_attachment *attach)
 {
-	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_gem_object *obj = dmabuf->priv;
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	long r;
+	int r;
 
-	r = drm_gem_map_attach(dma_buf, attach);
-	if (r)
-		return r;
+	if (attach->dev->driver == adev->dev->driver)
+		return 0;
 
 	r = amdgpu_bo_reserve(bo, false);
 	if (unlikely(r != 0))
-		goto error_detach;
-
+		return r;
 
-	if (attach->dev->driver != adev->dev->driver) {
 	/*
 	 * We only create shared fences for internal use, but importers
 	 * of the dmabuf rely on exclusive fences for implicitly
@@ -218,53 +196,94 @@ static int amdgpu_dma_buf_map_attach(struct dma_buf *dma_buf,
 	 */
 	r = __dma_resv_make_exclusive(bo->tbo.base.resv);
 	if (r)
-			goto error_unreserve;
-	}
+		return r;
+
+	bo->prime_shared_count++;
+	amdgpu_bo_unreserve(bo);
+	return 0;
+}
+
+/**
+ * amdgpu_dma_buf_detach - &dma_buf_ops.detach implementation
+ *
+ * @dmabuf: DMA-buf where we remove the attachment from
+ * @attach: the attachment to remove
+ *
+ * Called when an attachment is removed from the DMA-buf.
+ */
+static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf,
+				  struct dma_buf_attachment *attach)
+{
+	struct drm_gem_object *obj = dmabuf->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+
+	if (attach->dev->driver != adev->dev->driver && bo->prime_shared_count)
+		bo->prime_shared_count--;
+}
+
+/**
+ * amdgpu_dma_buf_map - &dma_buf_ops.map_dma_buf implementation
+ * @attach: DMA-buf attachment
+ * @dir: DMA direction
+ *
+ * Makes sure that the shared DMA buffer can be accessed by the target device.
+ * For now, simply pins it to the GTT domain, where it should be accessible by
+ * all DMA devices.
+ *
+ * Returns:
+ * sg_table filled with the DMA addresses to use or ERR_PRT with negative error
+ * code.
+ */
+static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
+					   enum dma_data_direction dir)
+{
+	struct dma_buf *dma_buf = attach->dmabuf;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct sg_table *sgt;
+	long r;
 
-	/* pin buffer into GTT */
 	r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
 	if (r)
-		goto error_unreserve;
+		return ERR_PTR(r);
 
-	if (attach->dev->driver != adev->dev->driver)
-		bo->prime_shared_count++;
+	sgt = drm_prime_pages_to_sg(bo->tbo.ttm->pages, bo->tbo.num_pages);
+	if (IS_ERR(sgt))
+		return sgt;
 
-error_unreserve:
-	amdgpu_bo_unreserve(bo);
+	if (!dma_map_sg_attrs(attach->dev, sgt->sgl, sgt->nents, dir,
+			      DMA_ATTR_SKIP_CPU_SYNC))
+		goto error_free;
 
-error_detach:
-	if (r)
-		drm_gem_map_detach(dma_buf, attach);
-	return r;
+	return sgt;
+
+error_free:
+	sg_free_table(sgt);
+	kfree(sgt);
+	return ERR_PTR(-ENOMEM);
 }
 
 /**
- * amdgpu_dma_buf_map_detach - &dma_buf_ops.detach implementation
- * @dma_buf: Shared DMA buffer
+ * amdgpu_dma_buf_unmap - &dma_buf_ops.unmap_dma_buf implementation
  * @attach: DMA-buf attachment
+ * @sgt: sg_table to unmap
+ * @dir: DMA direction
  *
  * This is called when a shared DMA buffer no longer needs to be accessible by
  * another device. For now, simply unpins the buffer from GTT.
  */
-static void amdgpu_dma_buf_map_detach(struct dma_buf *dma_buf,
-				      struct dma_buf_attachment *attach)
+static void amdgpu_dma_buf_unmap(struct dma_buf_attachment *attach,
+				 struct sg_table *sgt,
+				 enum dma_data_direction dir)
 {
-	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_gem_object *obj = attach->dmabuf->priv;
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	int ret = 0;
-
-	ret = amdgpu_bo_reserve(bo, true);
-	if (unlikely(ret != 0))
-		goto error;
 
+	dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+	sg_free_table(sgt);
+	kfree(sgt);
 	amdgpu_bo_unpin(bo);
-	if (attach->dev->driver != adev->dev->driver && bo->prime_shared_count)
-		bo->prime_shared_count--;
-	amdgpu_bo_unreserve(bo);
-
-error:
-	drm_gem_map_detach(dma_buf, attach);
 }
 
 /**
@@ -308,10 +327,11 @@ static int amdgpu_dma_buf_begin_cpu_access(struct dma_buf *dma_buf,
 }
 
 const struct dma_buf_ops amdgpu_dmabuf_ops = {
-	.attach = amdgpu_dma_buf_map_attach,
-	.detach = amdgpu_dma_buf_map_detach,
-	.map_dma_buf = drm_gem_map_dma_buf,
-	.unmap_dma_buf = drm_gem_unmap_dma_buf,
+	.dynamic_mapping = true,
+	.attach = amdgpu_dma_buf_attach,
+	.detach = amdgpu_dma_buf_detach,
+	.map_dma_buf = amdgpu_dma_buf_map,
+	.unmap_dma_buf = amdgpu_dma_buf_unmap,
 	.release = drm_gem_dmabuf_release,
 	.begin_cpu_access = amdgpu_dma_buf_begin_cpu_access,
 	.mmap = drm_gem_dmabuf_mmap,

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

@@ -25,7 +25,6 @@
 
 #include <drm/drm_gem.h>
 
-struct sg_table *amdgpu_gem_prime_get_sg_table(struct drm_gem_object *obj);
 struct drm_gem_object *
 amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 				 struct dma_buf_attachment *attach,

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

@@ -1445,7 +1445,6 @@ static struct drm_driver kms_driver = {
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
 	.gem_prime_export = amdgpu_gem_prime_export,
 	.gem_prime_import = amdgpu_gem_prime_import,
-	.gem_prime_get_sg_table = amdgpu_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = amdgpu_gem_prime_import_sg_table,
 	.gem_prime_vmap = amdgpu_gem_prime_vmap,
 	.gem_prime_vunmap = amdgpu_gem_prime_vunmap,