[PATCH 1/5] dma-buf: add optional invalidate_mappings callback v3

[PATCH 1/5] dma-buf: add optional invalidate_mappings callback v3

[Christian König]

Each importer can now provide an invalidate_mappings callback.

This allows the exporter to provide the mappings without the need to pin
the backing store.

v2: don't try to invalidate mappings when the callback is NULL,
    lock the reservation obj while using the attachments,
    add helper to set the callback
v3: move flag for invalidation support into the DMA-buf,
    use new attach_info structure to set the callback

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/dma-buf/dma-buf.c | 43 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/dma-buf.h   | 28 ++++++++++++++++++++++++++++
 2 files changed, 71 insertions(+)

diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
index d2e8ca0d9427..ffaa2f9a9c2c 100644
--- a/drivers/dma-buf/dma-buf.c
+++ b/drivers/dma-buf/dma-buf.c
@@ -566,6 +566,7 @@ struct dma_buf_attachment *dma_buf_attach(const struct dma_buf_attach_info *info
 	attach->dev = info->dev;
 	attach->dmabuf = dmabuf;
 	attach->priv = info->priv;
+	attach->invalidate = info->invalidate;
 
 	mutex_lock(&dmabuf->lock);
 
@@ -574,7 +575,9 @@ struct dma_buf_attachment *dma_buf_attach(const struct dma_buf_attach_info *info
 		if (ret)
 			goto err_attach;
 	}
+	reservation_object_lock(dmabuf->resv, NULL);
 	list_add(&attach->node, &dmabuf->attachments);
+	reservation_object_unlock(dmabuf->resv);
 
 	mutex_unlock(&dmabuf->lock);
 	return attach;
@@ -600,7 +603,9 @@ void dma_buf_detach(struct dma_buf *dmabuf, struct dma_buf_attachment *attach)
 		return;
 
 	mutex_lock(&dmabuf->lock);
+	reservation_object_lock(dmabuf->resv, NULL);
 	list_del(&attach->node);
+	reservation_object_unlock(dmabuf->resv);
 	if (dmabuf->ops->detach)
 		dmabuf->ops->detach(dmabuf, attach);
 
@@ -634,10 +639,23 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *attach,
 	if (WARN_ON(!attach || !attach->dmabuf))
 		return ERR_PTR(-EINVAL);
 
+	/*
+	 * Mapping a DMA-buf can trigger its invalidation, prevent sending this
+	 * event to the caller by temporary removing this attachment from the
+	 * list.
+	 */
+	if (attach->invalidate) {
+		reservation_object_assert_held(attach->dmabuf->resv);
+		list_del(&attach->node);
+	}
+
 	sg_table = attach->dmabuf->ops->map_dma_buf(attach, direction);
 	if (!sg_table)
 		sg_table = ERR_PTR(-ENOMEM);
 
+	if (attach->invalidate)
+		list_add(&attach->node, &attach->dmabuf->attachments);
+
 	return sg_table;
 }
 EXPORT_SYMBOL_GPL(dma_buf_map_attachment);
@@ -658,6 +676,9 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach,
 {
 	might_sleep();
 
+	if (attach->invalidate)
+		reservation_object_assert_held(attach->dmabuf->resv);
+
 	if (WARN_ON(!attach || !attach->dmabuf || !sg_table))
 		return;
 
@@ -666,6 +687,26 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach,
 }
 EXPORT_SYMBOL_GPL(dma_buf_unmap_attachment);
 
+/**
+ * dma_buf_invalidate_mappings - invalidate all mappings of this dma_buf
+ *
+ * @dmabuf:	[in]	buffer which mappings should be invalidated
+ *
+ * Informs all attachmenst that they need to destroy and recreated all their
+ * mappings.
+ */
+void dma_buf_invalidate_mappings(struct dma_buf *dmabuf)
+{
+	struct dma_buf_attachment *attach;
+
+	reservation_object_assert_held(dmabuf->resv);
+
+	list_for_each_entry(attach, &dmabuf->attachments, node)
+		if (attach->invalidate)
+			attach->invalidate(attach);
+}
+EXPORT_SYMBOL_GPL(dma_buf_invalidate_mappings);
+
 /**
  * DOC: cpu access
  *
@@ -1123,10 +1164,12 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused)
 		seq_puts(s, "\tAttached Devices:\n");
 		attach_count = 0;
 
+		reservation_object_lock(buf_obj->resv, NULL);
 		list_for_each_entry(attach_obj, &buf_obj->attachments, node) {
 			seq_printf(s, "\t%s\n", dev_name(attach_obj->dev));
 			attach_count++;
 		}
+		reservation_object_unlock(buf_obj->resv);
 
 		seq_printf(s, "Total %d devices attached\n\n",
 				attach_count);
diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h
index 2c27568d44af..15dd8598bff1 100644
--- a/include/linux/dma-buf.h
+++ b/include/linux/dma-buf.h
@@ -270,6 +270,8 @@ struct dma_buf_ops {
  * @poll: for userspace poll support
  * @cb_excl: for userspace poll support
  * @cb_shared: for userspace poll support
+ * @invalidation_supported: True when the exporter supports unpinned operation
+ *                          using the reservation lock.
  *
  * This represents a shared buffer, created by calling dma_buf_export(). The
  * userspace representation is a normal file descriptor, which can be created by
@@ -293,6 +295,7 @@ struct dma_buf {
 	struct list_head list_node;
 	void *priv;
 	struct reservation_object *resv;
+	bool invalidation_supported;
 
 	/* poll support */
 	wait_queue_head_t poll;
@@ -326,6 +329,28 @@ struct dma_buf_attachment {
 	struct device *dev;
 	struct list_head node;
 	void *priv;
+
+	/**
+	 * @invalidate:
+	 *
+	 * Optional callback provided by the importer of the dma-buf.
+	 *
+	 * If provided the exporter can avoid pinning the backing store while
+	 * mappings exists.
+	 *
+	 * The function is called with the lock of the reservation object
+	 * associated with the dma_buf held and the mapping function must be
+	 * called with this lock held as well. This makes sure that no mapping
+	 * is created concurrently with an ongoing invalidation.
+	 *
+	 * After the callback all existing mappings are still valid until all
+	 * fences in the dma_bufs reservation object are signaled, but should be
+	 * destroyed by the importer as soon as possible.
+	 *
+	 * New mappings can be created immediately, but can't be used before the
+	 * exclusive fence in the dma_bufs reservation object is signaled.
+	 */
+	void (*invalidate)(struct dma_buf_attachment *attach);
 };
 
 /**
@@ -367,6 +392,7 @@ struct dma_buf_export_info {
  * @dmabuf:	the exported dma_buf
  * @dev:	the device which wants to import the attachment
  * @priv:	private data of importer to this attachment
+ * @invalidate:	callback to use for invalidating mappings
  *
  * This structure holds the information required to attach to a buffer. Used
  * with dma_buf_attach() only.
@@ -375,6 +401,7 @@ struct dma_buf_attach_info {
 	struct dma_buf *dmabuf;
 	struct device *dev;
 	void *priv;
+	void (*invalidate)(struct dma_buf_attachment *attach);
 };
 
 /**
@@ -406,6 +433,7 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *,
 					enum dma_data_direction);
 void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *,
 				enum dma_data_direction);
+void dma_buf_invalidate_mappings(struct dma_buf *dma_buf);
 int dma_buf_begin_cpu_access(struct dma_buf *dma_buf,
 			     enum dma_data_direction dir);
 int dma_buf_end_cpu_access(struct dma_buf *dma_buf,
-- 
2.14.1

[PATCH 1/5] dma-buf: add optional invalidate_mappings callback v3

[Christian König]

Each importer can now provide an invalidate_mappings callback.

This allows the exporter to provide the mappings without the need to pin
the backing store.

v2: don't try to invalidate mappings when the callback is NULL,
    lock the reservation obj while using the attachments,
    add helper to set the callback
v3: move flag for invalidation support into the DMA-buf,
    use new attach_info structure to set the callback

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/dma-buf/dma-buf.c | 43 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/dma-buf.h   | 28 ++++++++++++++++++++++++++++
 2 files changed, 71 insertions(+)

diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
index d2e8ca0d9427..ffaa2f9a9c2c 100644
--- a/drivers/dma-buf/dma-buf.c
+++ b/drivers/dma-buf/dma-buf.c
@@ -566,6 +566,7 @@ struct dma_buf_attachment *dma_buf_attach(const struct dma_buf_attach_info *info
 	attach->dev = info->dev;
 	attach->dmabuf = dmabuf;
 	attach->priv = info->priv;
+	attach->invalidate = info->invalidate;
 
 	mutex_lock(&dmabuf->lock);
 
@@ -574,7 +575,9 @@ struct dma_buf_attachment *dma_buf_attach(const struct dma_buf_attach_info *info
 		if (ret)
 			goto err_attach;
 	}
+	reservation_object_lock(dmabuf->resv, NULL);
 	list_add(&attach->node, &dmabuf->attachments);
+	reservation_object_unlock(dmabuf->resv);
 
 	mutex_unlock(&dmabuf->lock);
 	return attach;
@@ -600,7 +603,9 @@ void dma_buf_detach(struct dma_buf *dmabuf, struct dma_buf_attachment *attach)
 		return;
 
 	mutex_lock(&dmabuf->lock);
+	reservation_object_lock(dmabuf->resv, NULL);
 	list_del(&attach->node);
+	reservation_object_unlock(dmabuf->resv);
 	if (dmabuf->ops->detach)
 		dmabuf->ops->detach(dmabuf, attach);
 
@@ -634,10 +639,23 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *attach,
 	if (WARN_ON(!attach || !attach->dmabuf))
 		return ERR_PTR(-EINVAL);
 
+	/*
+	 * Mapping a DMA-buf can trigger its invalidation, prevent sending this
+	 * event to the caller by temporary removing this attachment from the
+	 * list.
+	 */
+	if (attach->invalidate) {
+		reservation_object_assert_held(attach->dmabuf->resv);
+		list_del(&attach->node);
+	}
+
 	sg_table = attach->dmabuf->ops->map_dma_buf(attach, direction);
 	if (!sg_table)
 		sg_table = ERR_PTR(-ENOMEM);
 
+	if (attach->invalidate)
+		list_add(&attach->node, &attach->dmabuf->attachments);
+
 	return sg_table;
 }
 EXPORT_SYMBOL_GPL(dma_buf_map_attachment);
@@ -658,6 +676,9 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach,
 {
 	might_sleep();
 
+	if (attach->invalidate)
+		reservation_object_assert_held(attach->dmabuf->resv);
+
 	if (WARN_ON(!attach || !attach->dmabuf || !sg_table))
 		return;
 
@@ -666,6 +687,26 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach,
 }
 EXPORT_SYMBOL_GPL(dma_buf_unmap_attachment);
 
+/**
+ * dma_buf_invalidate_mappings - invalidate all mappings of this dma_buf
+ *
+ * @dmabuf:	[in]	buffer which mappings should be invalidated
+ *
+ * Informs all attachmenst that they need to destroy and recreated all their
+ * mappings.
+ */
+void dma_buf_invalidate_mappings(struct dma_buf *dmabuf)
+{
+	struct dma_buf_attachment *attach;
+
+	reservation_object_assert_held(dmabuf->resv);
+
+	list_for_each_entry(attach, &dmabuf->attachments, node)
+		if (attach->invalidate)
+			attach->invalidate(attach);
+}
+EXPORT_SYMBOL_GPL(dma_buf_invalidate_mappings);
+
 /**
  * DOC: cpu access
  *
@@ -1123,10 +1164,12 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused)
 		seq_puts(s, "\tAttached Devices:\n");
 		attach_count = 0;
 
+		reservation_object_lock(buf_obj->resv, NULL);
 		list_for_each_entry(attach_obj, &buf_obj->attachments, node) {
 			seq_printf(s, "\t%s\n", dev_name(attach_obj->dev));
 			attach_count++;
 		}
+		reservation_object_unlock(buf_obj->resv);
 
 		seq_printf(s, "Total %d devices attached\n\n",
 				attach_count);
diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h
index 2c27568d44af..15dd8598bff1 100644
--- a/include/linux/dma-buf.h
+++ b/include/linux/dma-buf.h
@@ -270,6 +270,8 @@ struct dma_buf_ops {
  * @poll: for userspace poll support
  * @cb_excl: for userspace poll support
  * @cb_shared: for userspace poll support
+ * @invalidation_supported: True when the exporter supports unpinned operation
+ *                          using the reservation lock.
  *
  * This represents a shared buffer, created by calling dma_buf_export(). The
  * userspace representation is a normal file descriptor, which can be created by
@@ -293,6 +295,7 @@ struct dma_buf {
 	struct list_head list_node;
 	void *priv;
 	struct reservation_object *resv;
+	bool invalidation_supported;
 
 	/* poll support */
 	wait_queue_head_t poll;
@@ -326,6 +329,28 @@ struct dma_buf_attachment {
 	struct device *dev;
 	struct list_head node;
 	void *priv;
+
+	/**
+	 * @invalidate:
+	 *
+	 * Optional callback provided by the importer of the dma-buf.
+	 *
+	 * If provided the exporter can avoid pinning the backing store while
+	 * mappings exists.
+	 *
+	 * The function is called with the lock of the reservation object
+	 * associated with the dma_buf held and the mapping function must be
+	 * called with this lock held as well. This makes sure that no mapping
+	 * is created concurrently with an ongoing invalidation.
+	 *
+	 * After the callback all existing mappings are still valid until all
+	 * fences in the dma_bufs reservation object are signaled, but should be
+	 * destroyed by the importer as soon as possible.
+	 *
+	 * New mappings can be created immediately, but can't be used before the
+	 * exclusive fence in the dma_bufs reservation object is signaled.
+	 */
+	void (*invalidate)(struct dma_buf_attachment *attach);
 };
 
 /**
@@ -367,6 +392,7 @@ struct dma_buf_export_info {
  * @dmabuf:	the exported dma_buf
  * @dev:	the device which wants to import the attachment
  * @priv:	private data of importer to this attachment
+ * @invalidate:	callback to use for invalidating mappings
  *
  * This structure holds the information required to attach to a buffer. Used
  * with dma_buf_attach() only.
@@ -375,6 +401,7 @@ struct dma_buf_attach_info {
 	struct dma_buf *dmabuf;
 	struct device *dev;
 	void *priv;
+	void (*invalidate)(struct dma_buf_attachment *attach);
 };
 
 /**
@@ -406,6 +433,7 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *,
 					enum dma_data_direction);
 void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *,
 				enum dma_data_direction);
+void dma_buf_invalidate_mappings(struct dma_buf *dma_buf);
 int dma_buf_begin_cpu_access(struct dma_buf *dma_buf,
 			     enum dma_data_direction dir);
 int dma_buf_end_cpu_access(struct dma_buf *dma_buf,
-- 
2.14.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH 2/5] drm/ttm: keep a reference to transfer pipelined BOs

[Christian König]

Make sure the transfered BO is never destroy before the transfer BO.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/ttm_bo_util.c | 50 +++++++++++++++++++++++----------------
 1 file changed, 30 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 1f730b3f18e5..1ee20558ee31 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -39,6 +39,11 @@
 #include <linux/module.h>
 #include <linux/reservation.h>
 
+struct ttm_transfer_obj {
+	struct ttm_buffer_object base;
+	struct ttm_buffer_object *bo;
+};
+
 void ttm_bo_free_old_node(struct ttm_buffer_object *bo)
 {
 	ttm_bo_mem_put(bo, &bo->mem);
@@ -435,7 +440,11 @@ EXPORT_SYMBOL(ttm_bo_move_memcpy);
 
 static void ttm_transfered_destroy(struct ttm_buffer_object *bo)
 {
-	kfree(bo);
+	struct ttm_transfer_obj *fbo;
+
+	fbo = container_of(bo, struct ttm_transfer_obj, base);
+	ttm_bo_unref(&fbo->bo);
+	kfree(fbo);
 }
 
 /**
@@ -456,14 +465,15 @@ static void ttm_transfered_destroy(struct ttm_buffer_object *bo)
 static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 				      struct ttm_buffer_object **new_obj)
 {
-	struct ttm_buffer_object *fbo;
+	struct ttm_transfer_obj *fbo;
 	int ret;
 
 	fbo = kmalloc(sizeof(*fbo), GFP_KERNEL);
 	if (!fbo)
 		return -ENOMEM;
 
-	*fbo = *bo;
+	fbo->base = *bo;
+	fbo->bo = ttm_bo_reference(bo);
 
 	/**
 	 * Fix up members that we shouldn't copy directly:
@@ -471,25 +481,25 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	 */
 
 	atomic_inc(&bo->bdev->glob->bo_count);
-	INIT_LIST_HEAD(&fbo->ddestroy);
-	INIT_LIST_HEAD(&fbo->lru);
-	INIT_LIST_HEAD(&fbo->swap);
-	INIT_LIST_HEAD(&fbo->io_reserve_lru);
-	mutex_init(&fbo->wu_mutex);
-	fbo->moving = NULL;
-	drm_vma_node_reset(&fbo->vma_node);
-	atomic_set(&fbo->cpu_writers, 0);
-
-	kref_init(&fbo->list_kref);
-	kref_init(&fbo->kref);
-	fbo->destroy = &ttm_transfered_destroy;
-	fbo->acc_size = 0;
-	fbo->resv = &fbo->ttm_resv;
-	reservation_object_init(fbo->resv);
-	ret = reservation_object_trylock(fbo->resv);
+	INIT_LIST_HEAD(&fbo->base.ddestroy);
+	INIT_LIST_HEAD(&fbo->base.lru);
+	INIT_LIST_HEAD(&fbo->base.swap);
+	INIT_LIST_HEAD(&fbo->base.io_reserve_lru);
+	mutex_init(&fbo->base.wu_mutex);
+	fbo->base.moving = NULL;
+	drm_vma_node_reset(&fbo->base.vma_node);
+	atomic_set(&fbo->base.cpu_writers, 0);
+
+	kref_init(&fbo->base.list_kref);
+	kref_init(&fbo->base.kref);
+	fbo->base.destroy = &ttm_transfered_destroy;
+	fbo->base.acc_size = 0;
+	fbo->base.resv = &fbo->base.ttm_resv;
+	reservation_object_init(fbo->base.resv);
+	ret = reservation_object_trylock(fbo->base.resv);
 	WARN_ON(!ret);
 
-	*new_obj = fbo;
+	*new_obj = &fbo->base;
 	return 0;
 }
 
-- 
2.14.1

[PATCH 2/5] drm/ttm: keep a reference to transfer pipelined BOs

[Christian König]

Make sure the transfered BO is never destroy before the transfer BO.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/ttm_bo_util.c | 50 +++++++++++++++++++++++----------------
 1 file changed, 30 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 1f730b3f18e5..1ee20558ee31 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -39,6 +39,11 @@
 #include <linux/module.h>
 #include <linux/reservation.h>
 
+struct ttm_transfer_obj {
+	struct ttm_buffer_object base;
+	struct ttm_buffer_object *bo;
+};
+
 void ttm_bo_free_old_node(struct ttm_buffer_object *bo)
 {
 	ttm_bo_mem_put(bo, &bo->mem);
@@ -435,7 +440,11 @@ EXPORT_SYMBOL(ttm_bo_move_memcpy);
 
 static void ttm_transfered_destroy(struct ttm_buffer_object *bo)
 {
-	kfree(bo);
+	struct ttm_transfer_obj *fbo;
+
+	fbo = container_of(bo, struct ttm_transfer_obj, base);
+	ttm_bo_unref(&fbo->bo);
+	kfree(fbo);
 }
 
 /**
@@ -456,14 +465,15 @@ static void ttm_transfered_destroy(struct ttm_buffer_object *bo)
 static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 				      struct ttm_buffer_object **new_obj)
 {
-	struct ttm_buffer_object *fbo;
+	struct ttm_transfer_obj *fbo;
 	int ret;
 
 	fbo = kmalloc(sizeof(*fbo), GFP_KERNEL);
 	if (!fbo)
 		return -ENOMEM;
 
-	*fbo = *bo;
+	fbo->base = *bo;
+	fbo->bo = ttm_bo_reference(bo);
 
 	/**
 	 * Fix up members that we shouldn't copy directly:
@@ -471,25 +481,25 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	 */
 
 	atomic_inc(&bo->bdev->glob->bo_count);
-	INIT_LIST_HEAD(&fbo->ddestroy);
-	INIT_LIST_HEAD(&fbo->lru);
-	INIT_LIST_HEAD(&fbo->swap);
-	INIT_LIST_HEAD(&fbo->io_reserve_lru);
-	mutex_init(&fbo->wu_mutex);
-	fbo->moving = NULL;
-	drm_vma_node_reset(&fbo->vma_node);
-	atomic_set(&fbo->cpu_writers, 0);
-
-	kref_init(&fbo->list_kref);
-	kref_init(&fbo->kref);
-	fbo->destroy = &ttm_transfered_destroy;
-	fbo->acc_size = 0;
-	fbo->resv = &fbo->ttm_resv;
-	reservation_object_init(fbo->resv);
-	ret = reservation_object_trylock(fbo->resv);
+	INIT_LIST_HEAD(&fbo->base.ddestroy);
+	INIT_LIST_HEAD(&fbo->base.lru);
+	INIT_LIST_HEAD(&fbo->base.swap);
+	INIT_LIST_HEAD(&fbo->base.io_reserve_lru);
+	mutex_init(&fbo->base.wu_mutex);
+	fbo->base.moving = NULL;
+	drm_vma_node_reset(&fbo->base.vma_node);
+	atomic_set(&fbo->base.cpu_writers, 0);
+
+	kref_init(&fbo->base.list_kref);
+	kref_init(&fbo->base.kref);
+	fbo->base.destroy = &ttm_transfered_destroy;
+	fbo->base.acc_size = 0;
+	fbo->base.resv = &fbo->base.ttm_resv;
+	reservation_object_init(fbo->base.resv);
+	ret = reservation_object_trylock(fbo->base.resv);
 	WARN_ON(!ret);
 
-	*new_obj = fbo;
+	*new_obj = &fbo->base;
 	return 0;
 }
 
-- 
2.14.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH 3/5] drm/ttm: remove the backing store if no placement is given

[Christian König]

Pipeline removal of the BOs backing store when the placement is given
during validation.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/ttm_bo.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 98e06f8bf23b..17e821f01d0a 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1078,6 +1078,18 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	uint32_t new_flags;
 
 	reservation_object_assert_held(bo->resv);
+
+	/*
+	 * Remove the backing store if no placement is given.
+	 */
+	if (!placement->num_placement && !placement->num_busy_placement) {
+		ret = ttm_bo_pipeline_gutting(bo);
+		if (ret)
+			return ret;
+
+		return ttm_tt_create(bo, false);
+	}
+
 	/*
 	 * Check whether we need to move buffer.
 	 */
-- 
2.14.1

[PATCH 3/5] drm/ttm: remove the backing store if no placement is given

[Christian König]

Pipeline removal of the BOs backing store when the placement is given
during validation.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/ttm_bo.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 98e06f8bf23b..17e821f01d0a 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1078,6 +1078,18 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	uint32_t new_flags;
 
 	reservation_object_assert_held(bo->resv);
+
+	/*
+	 * Remove the backing store if no placement is given.
+	 */
+	if (!placement->num_placement && !placement->num_busy_placement) {
+		ret = ttm_bo_pipeline_gutting(bo);
+		if (ret)
+			return ret;
+
+		return ttm_tt_create(bo, false);
+	}
+
 	/*
 	 * Check whether we need to move buffer.
 	 */
-- 
2.14.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 4/5] drm/amdgpu: add independent DMA-buf export v2

[Christian König]

Instead of relying on the DRM functions just implement our own export
functions. This adds support for taking care of unpinned DMA-buf.

v2: fix unintended recursion, remove debugging leftovers

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        |   1 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    |   1 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c |   5 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c  | 108 ++++++++++++++++++-----------
 4 files changed, 72 insertions(+), 43 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index a7e22298ce5c..190bada6f487 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -371,7 +371,6 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 void amdgpu_gem_object_close(struct drm_gem_object *obj,
 				struct drm_file *file_priv);
 unsigned long amdgpu_gem_timeout(uint64_t timeout_ns);
-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,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 1bfce79bc074..33893f0311c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -892,7 +892,6 @@ static struct drm_driver kms_driver = {
 	.gem_prime_export = amdgpu_gem_prime_export,
 	.gem_prime_import = amdgpu_gem_prime_import,
 	.gem_prime_res_obj = amdgpu_gem_prime_res_obj,
-	.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,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index ac1fa3394a9e..11d2ac175b30 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -31,6 +31,7 @@
  */
 #include <linux/list.h>
 #include <linux/slab.h>
+#include <linux/dma-buf.h>
 #include <drm/drmP.h>
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_cache.h>
@@ -935,6 +936,10 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 
 	amdgpu_bo_kunmap(abo);
 
+	if (abo->gem_base.dma_buf && !abo->gem_base.import_attach &&
+	    bo->mem.mem_type != TTM_PL_SYSTEM)
+		dma_buf_invalidate_mappings(abo->gem_base.dma_buf);
+
 	/* remember the eviction */
 	if (evict)
 		atomic64_inc(&adev->num_evictions);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 4b584cb75bf4..7fef95f0fed1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -32,14 +32,6 @@
 
 static const struct dma_buf_ops amdgpu_dmabuf_ops;
 
-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);
-}
-
 void *amdgpu_gem_prime_vmap(struct drm_gem_object *obj)
 {
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
@@ -126,23 +118,22 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 	return ERR_PTR(ret);
 }
 
-static int amdgpu_gem_map_attach(struct dma_buf *dma_buf,
-				 struct device *target_dev,
-				 struct dma_buf_attachment *attach)
+static struct sg_table *
+amdgpu_gem_map_dma_buf(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 amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	struct sg_table *sgt;
 	long r;
 
-	r = drm_gem_map_attach(dma_buf, target_dev, attach);
-	if (r)
-		return r;
-
-	r = amdgpu_bo_reserve(bo, false);
-	if (unlikely(r != 0))
-		goto error_detach;
-
+	if (!attach->invalidate) {
+		r = amdgpu_bo_reserve(bo, false);
+		if (unlikely(r != 0))
+			return ERR_PTR(r);
+	}
 
 	if (attach->dev->driver != adev->dev->driver) {
 		/*
@@ -158,42 +149,77 @@ static int amdgpu_gem_map_attach(struct dma_buf *dma_buf,
 		}
 	}
 
-	/* pin buffer into GTT */
-	r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT, NULL);
-	if (r)
+	if (!attach->invalidate) {
+		/* pin buffer into GTT */
+		r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT, NULL);
+		if (r)
+			goto error_unreserve;
+
+	} else {
+		/* move buffer into GTT */
+		struct ttm_operation_ctx ctx = { false, false };
+
+		amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+		if (r)
+			goto error_unreserve;
+	}
+
+	sgt = drm_prime_pages_to_sg(bo->tbo.ttm->pages, bo->tbo.num_pages);
+	if (IS_ERR(sgt)) {
+		r = PTR_ERR(sgt);
 		goto error_unreserve;
+	}
+
+	if (!dma_map_sg_attrs(attach->dev, sgt->sgl, sgt->nents, dir,
+			      DMA_ATTR_SKIP_CPU_SYNC))
+		goto error_free;
 
 	if (attach->dev->driver != adev->dev->driver)
 		bo->prime_shared_count++;
 
+	if (!attach->invalidate)
+		amdgpu_bo_unreserve(bo);
+
+	return sgt;
+
+error_free:
+	sg_free_table(sgt);
+	kfree(sgt);
+	r = -ENOMEM;
+
 error_unreserve:
-	amdgpu_bo_unreserve(bo);
+	if (!attach->invalidate)
+		amdgpu_bo_unreserve(bo);
 
-error_detach:
-	if (r)
-		drm_gem_map_detach(dma_buf, attach);
-	return r;
+	return ERR_PTR(r);
 }
 
-static void amdgpu_gem_map_detach(struct dma_buf *dma_buf,
-				  struct dma_buf_attachment *attach)
+static void amdgpu_gem_unmap_dma_buf(struct dma_buf_attachment *attach,
+				     struct sg_table *sgt,
+				     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 amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	int ret = 0;
 
-	ret = amdgpu_bo_reserve(bo, true);
-	if (unlikely(ret != 0))
-		goto error;
+	if (!attach->invalidate) {
+		amdgpu_bo_reserve(bo, true);
+		amdgpu_bo_unpin(bo);
+	}
 
-	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);
+	if (!attach->invalidate)
+		amdgpu_bo_unreserve(bo);
+
+	if (sgt) {
+		dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+		sg_free_table(sgt);
+		kfree(sgt);
+	}
 }
 
 struct reservation_object *amdgpu_gem_prime_res_obj(struct drm_gem_object *obj)
@@ -232,10 +258,8 @@ static int amdgpu_gem_begin_cpu_access(struct dma_buf *dma_buf,
 }
 
 static const struct dma_buf_ops amdgpu_dmabuf_ops = {
-	.attach = amdgpu_gem_map_attach,
-	.detach = amdgpu_gem_map_detach,
-	.map_dma_buf = drm_gem_map_dma_buf,
-	.unmap_dma_buf = drm_gem_unmap_dma_buf,
+	.map_dma_buf = amdgpu_gem_map_dma_buf,
+	.unmap_dma_buf = amdgpu_gem_unmap_dma_buf,
 	.release = drm_gem_dmabuf_release,
 	.begin_cpu_access = amdgpu_gem_begin_cpu_access,
 	.map = drm_gem_dmabuf_kmap,
@@ -260,6 +284,8 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 
 	buf = drm_gem_prime_export(dev, gobj, flags);
 	if (!IS_ERR(buf)) {
+		buf->invalidation_supported =
+			!!(bo->preferred_domains & AMDGPU_GEM_DOMAIN_GTT);
 		buf->file->f_mapping = dev->anon_inode->i_mapping;
 		buf->ops = &amdgpu_dmabuf_ops;
 	}
-- 
2.14.1

[PATCH 4/5] drm/amdgpu: add independent DMA-buf export v2

[Christian König]

Instead of relying on the DRM functions just implement our own export
functions. This adds support for taking care of unpinned DMA-buf.

v2: fix unintended recursion, remove debugging leftovers

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        |   1 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    |   1 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c |   5 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c  | 108 ++++++++++++++++++-----------
 4 files changed, 72 insertions(+), 43 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index a7e22298ce5c..190bada6f487 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -371,7 +371,6 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 void amdgpu_gem_object_close(struct drm_gem_object *obj,
 				struct drm_file *file_priv);
 unsigned long amdgpu_gem_timeout(uint64_t timeout_ns);
-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,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 1bfce79bc074..33893f0311c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -892,7 +892,6 @@ static struct drm_driver kms_driver = {
 	.gem_prime_export = amdgpu_gem_prime_export,
 	.gem_prime_import = amdgpu_gem_prime_import,
 	.gem_prime_res_obj = amdgpu_gem_prime_res_obj,
-	.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,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index ac1fa3394a9e..11d2ac175b30 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -31,6 +31,7 @@
  */
 #include <linux/list.h>
 #include <linux/slab.h>
+#include <linux/dma-buf.h>
 #include <drm/drmP.h>
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_cache.h>
@@ -935,6 +936,10 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 
 	amdgpu_bo_kunmap(abo);
 
+	if (abo->gem_base.dma_buf && !abo->gem_base.import_attach &&
+	    bo->mem.mem_type != TTM_PL_SYSTEM)
+		dma_buf_invalidate_mappings(abo->gem_base.dma_buf);
+
 	/* remember the eviction */
 	if (evict)
 		atomic64_inc(&adev->num_evictions);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 4b584cb75bf4..7fef95f0fed1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -32,14 +32,6 @@
 
 static const struct dma_buf_ops amdgpu_dmabuf_ops;
 
-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);
-}
-
 void *amdgpu_gem_prime_vmap(struct drm_gem_object *obj)
 {
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
@@ -126,23 +118,22 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 	return ERR_PTR(ret);
 }
 
-static int amdgpu_gem_map_attach(struct dma_buf *dma_buf,
-				 struct device *target_dev,
-				 struct dma_buf_attachment *attach)
+static struct sg_table *
+amdgpu_gem_map_dma_buf(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 amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	struct sg_table *sgt;
 	long r;
 
-	r = drm_gem_map_attach(dma_buf, target_dev, attach);
-	if (r)
-		return r;
-
-	r = amdgpu_bo_reserve(bo, false);
-	if (unlikely(r != 0))
-		goto error_detach;
-
+	if (!attach->invalidate) {
+		r = amdgpu_bo_reserve(bo, false);
+		if (unlikely(r != 0))
+			return ERR_PTR(r);
+	}
 
 	if (attach->dev->driver != adev->dev->driver) {
 		/*
@@ -158,42 +149,77 @@ static int amdgpu_gem_map_attach(struct dma_buf *dma_buf,
 		}
 	}
 
-	/* pin buffer into GTT */
-	r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT, NULL);
-	if (r)
+	if (!attach->invalidate) {
+		/* pin buffer into GTT */
+		r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT, NULL);
+		if (r)
+			goto error_unreserve;
+
+	} else {
+		/* move buffer into GTT */
+		struct ttm_operation_ctx ctx = { false, false };
+
+		amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+		if (r)
+			goto error_unreserve;
+	}
+
+	sgt = drm_prime_pages_to_sg(bo->tbo.ttm->pages, bo->tbo.num_pages);
+	if (IS_ERR(sgt)) {
+		r = PTR_ERR(sgt);
 		goto error_unreserve;
+	}
+
+	if (!dma_map_sg_attrs(attach->dev, sgt->sgl, sgt->nents, dir,
+			      DMA_ATTR_SKIP_CPU_SYNC))
+		goto error_free;
 
 	if (attach->dev->driver != adev->dev->driver)
 		bo->prime_shared_count++;
 
+	if (!attach->invalidate)
+		amdgpu_bo_unreserve(bo);
+
+	return sgt;
+
+error_free:
+	sg_free_table(sgt);
+	kfree(sgt);
+	r = -ENOMEM;
+
 error_unreserve:
-	amdgpu_bo_unreserve(bo);
+	if (!attach->invalidate)
+		amdgpu_bo_unreserve(bo);
 
-error_detach:
-	if (r)
-		drm_gem_map_detach(dma_buf, attach);
-	return r;
+	return ERR_PTR(r);
 }
 
-static void amdgpu_gem_map_detach(struct dma_buf *dma_buf,
-				  struct dma_buf_attachment *attach)
+static void amdgpu_gem_unmap_dma_buf(struct dma_buf_attachment *attach,
+				     struct sg_table *sgt,
+				     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 amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	int ret = 0;
 
-	ret = amdgpu_bo_reserve(bo, true);
-	if (unlikely(ret != 0))
-		goto error;
+	if (!attach->invalidate) {
+		amdgpu_bo_reserve(bo, true);
+		amdgpu_bo_unpin(bo);
+	}
 
-	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);
+	if (!attach->invalidate)
+		amdgpu_bo_unreserve(bo);
+
+	if (sgt) {
+		dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+		sg_free_table(sgt);
+		kfree(sgt);
+	}
 }
 
 struct reservation_object *amdgpu_gem_prime_res_obj(struct drm_gem_object *obj)
@@ -232,10 +258,8 @@ static int amdgpu_gem_begin_cpu_access(struct dma_buf *dma_buf,
 }
 
 static const struct dma_buf_ops amdgpu_dmabuf_ops = {
-	.attach = amdgpu_gem_map_attach,
-	.detach = amdgpu_gem_map_detach,
-	.map_dma_buf = drm_gem_map_dma_buf,
-	.unmap_dma_buf = drm_gem_unmap_dma_buf,
+	.map_dma_buf = amdgpu_gem_map_dma_buf,
+	.unmap_dma_buf = amdgpu_gem_unmap_dma_buf,
 	.release = drm_gem_dmabuf_release,
 	.begin_cpu_access = amdgpu_gem_begin_cpu_access,
 	.map = drm_gem_dmabuf_kmap,
@@ -260,6 +284,8 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 
 	buf = drm_gem_prime_export(dev, gobj, flags);
 	if (!IS_ERR(buf)) {
+		buf->invalidation_supported =
+			!!(bo->preferred_domains & AMDGPU_GEM_DOMAIN_GTT);
 		buf->file->f_mapping = dev->anon_inode->i_mapping;
 		buf->ops = &amdgpu_dmabuf_ops;
 	}
-- 
2.14.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 5/5] drm/amdgpu: add independent DMA-buf import v3

[Christian König]

Instead of relying on the DRM functions just implement our own import
functions. This adds support for taking care of unpinned DMA-buf.

v2: enable for all exporters, not just amdgpu, fix invalidation
    handling, lock reservation object while setting callback
v3: change to new dma_buf attach interface

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c | 72 ++++++++++++++++++++++++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c   | 32 +++++++++++---
 2 files changed, 89 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 7fef95f0fed1..fb43faf88eb0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -86,28 +86,24 @@ int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma
 	return ret;
 }
 
-struct drm_gem_object *
-amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
-				 struct dma_buf_attachment *attach,
-				 struct sg_table *sg)
+static struct drm_gem_object *
+amdgpu_gem_prime_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 {
-	struct reservation_object *resv = attach->dmabuf->resv;
+	struct reservation_object *resv = dma_buf->resv;
 	struct amdgpu_device *adev = dev->dev_private;
 	struct amdgpu_bo *bo;
 	int ret;
 
 	ww_mutex_lock(&resv->lock, NULL);
-	ret = amdgpu_bo_create(adev, attach->dmabuf->size, PAGE_SIZE,
+	ret = amdgpu_bo_create(adev, dma_buf->size, PAGE_SIZE,
 			       AMDGPU_GEM_DOMAIN_CPU, 0, ttm_bo_type_sg,
 			       resv, &bo);
 	if (ret)
 		goto error;
 
-	bo->tbo.sg = sg;
-	bo->tbo.ttm->sg = sg;
 	bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
 	bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
-	if (attach->dmabuf->ops != &amdgpu_dmabuf_ops)
+	if (dma_buf->ops != &amdgpu_dmabuf_ops)
 		bo->prime_shared_count = 1;
 
 	ww_mutex_unlock(&resv->lock);
@@ -118,6 +114,26 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 	return ERR_PTR(ret);
 }
 
+struct drm_gem_object *
+amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
+				 struct dma_buf_attachment *attach,
+				 struct sg_table *sg)
+{
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *bo;
+
+	obj = amdgpu_gem_prime_create_obj(dev, attach->dmabuf);
+	if (IS_ERR(obj))
+		return obj;
+
+	bo = gem_to_amdgpu_bo(obj);
+	amdgpu_bo_reserve(bo, true);
+	bo->tbo.sg = sg;
+	bo->tbo.ttm->sg = sg;
+	amdgpu_bo_unreserve(bo);
+	return obj;
+}
+
 static struct sg_table *
 amdgpu_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		       enum dma_data_direction dir)
@@ -293,9 +309,29 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 	return buf;
 }
 
+static void
+amdgpu_gem_prime_invalidate_mappings(struct dma_buf_attachment *attach)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct drm_gem_object *obj = attach->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct ttm_placement placement = {};
+	int r;
+
+	r = ttm_bo_validate(&bo->tbo, &placement, &ctx);
+	if (r)
+		DRM_ERROR("Failed to unmap DMA-buf import (%d))\n", r);
+}
+
 struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 					    struct dma_buf *dma_buf)
 {
+	struct dma_buf_attach_info attach_info = {
+		.dev = dev->dev,
+		.dmabuf = dma_buf,
+		.invalidate = amdgpu_gem_prime_invalidate_mappings
+	};
+	struct dma_buf_attachment *attach;
 	struct drm_gem_object *obj;
 
 	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
@@ -310,5 +346,21 @@ struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 		}
 	}
 
-	return drm_gem_prime_import(dev, dma_buf);
+	if (!dma_buf->invalidation_supported)
+		return drm_gem_prime_import(dev, dma_buf);
+
+	obj = amdgpu_gem_prime_create_obj(dev, dma_buf);
+	if (IS_ERR(obj))
+		return obj;
+
+	attach_info.priv = obj;
+	attach = dma_buf_attach(&attach_info);
+	if (IS_ERR(attach)) {
+		drm_gem_object_put(obj);
+		return ERR_CAST(attach);
+	}
+
+	get_dma_buf(dma_buf);
+	obj->import_attach = attach;
+	return obj;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index d2ab40494a4c..ad93f201e7b6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -43,6 +43,7 @@
 #include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include <linux/iommu.h>
+#include <linux/dma-buf.h>
 #include "amdgpu.h"
 #include "amdgpu_object.h"
 #include "amdgpu_trace.h"
@@ -685,6 +686,7 @@ struct amdgpu_ttm_gup_task_list {
 
 struct amdgpu_ttm_tt {
 	struct ttm_dma_tt	ttm;
+	struct amdgpu_bo	*bo;
 	u64			offset;
 	uint64_t		userptr;
 	struct mm_struct	*usermm;
@@ -993,6 +995,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 		return NULL;
 	}
 	gtt->ttm.ttm.func = &amdgpu_backend_func;
+	gtt->bo = ttm_to_amdgpu_bo(bo);
 	if (ttm_sg_tt_init(&gtt->ttm, bo, page_flags)) {
 		kfree(gtt);
 		return NULL;
@@ -1005,7 +1008,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
 	if (gtt && gtt->userptr) {
 		ttm->sg = kzalloc(sizeof(struct sg_table), GFP_KERNEL);
@@ -1017,7 +1019,19 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 		return 0;
 	}
 
-	if (slave && ttm->sg) {
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG) {
+		if (!ttm->sg) {
+			struct dma_buf_attachment *attach;
+			struct sg_table *sgt;
+
+			attach = gtt->bo->gem_base.import_attach;
+			sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+			if (IS_ERR(sgt))
+				return PTR_ERR(sgt);
+
+			ttm->sg = sgt;
+		}
+
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address,
 						 ttm->num_pages);
@@ -1036,9 +1050,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 
 static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev;
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
+	struct amdgpu_device *adev;
 
 	if (gtt && gtt->userptr) {
 		amdgpu_ttm_tt_set_user_pages(ttm, NULL);
@@ -1047,7 +1060,16 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 		return;
 	}
 
-	if (slave)
+	if (ttm->sg && !gtt->bo->tbo.sg) {
+		struct dma_buf_attachment *attach;
+
+		attach = gtt->bo->gem_base.import_attach;
+		dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL);
+		ttm->sg = NULL;
+		return;
+	}
+
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG)
 		return;
 
 	adev = amdgpu_ttm_adev(ttm->bdev);
-- 
2.14.1

[PATCH 5/5] drm/amdgpu: add independent DMA-buf import v3

[Christian König]

Instead of relying on the DRM functions just implement our own import
functions. This adds support for taking care of unpinned DMA-buf.

v2: enable for all exporters, not just amdgpu, fix invalidation
    handling, lock reservation object while setting callback
v3: change to new dma_buf attach interface

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c | 72 ++++++++++++++++++++++++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c   | 32 +++++++++++---
 2 files changed, 89 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 7fef95f0fed1..fb43faf88eb0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -86,28 +86,24 @@ int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma
 	return ret;
 }
 
-struct drm_gem_object *
-amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
-				 struct dma_buf_attachment *attach,
-				 struct sg_table *sg)
+static struct drm_gem_object *
+amdgpu_gem_prime_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 {
-	struct reservation_object *resv = attach->dmabuf->resv;
+	struct reservation_object *resv = dma_buf->resv;
 	struct amdgpu_device *adev = dev->dev_private;
 	struct amdgpu_bo *bo;
 	int ret;
 
 	ww_mutex_lock(&resv->lock, NULL);
-	ret = amdgpu_bo_create(adev, attach->dmabuf->size, PAGE_SIZE,
+	ret = amdgpu_bo_create(adev, dma_buf->size, PAGE_SIZE,
 			       AMDGPU_GEM_DOMAIN_CPU, 0, ttm_bo_type_sg,
 			       resv, &bo);
 	if (ret)
 		goto error;
 
-	bo->tbo.sg = sg;
-	bo->tbo.ttm->sg = sg;
 	bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
 	bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
-	if (attach->dmabuf->ops != &amdgpu_dmabuf_ops)
+	if (dma_buf->ops != &amdgpu_dmabuf_ops)
 		bo->prime_shared_count = 1;
 
 	ww_mutex_unlock(&resv->lock);
@@ -118,6 +114,26 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 	return ERR_PTR(ret);
 }
 
+struct drm_gem_object *
+amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
+				 struct dma_buf_attachment *attach,
+				 struct sg_table *sg)
+{
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *bo;
+
+	obj = amdgpu_gem_prime_create_obj(dev, attach->dmabuf);
+	if (IS_ERR(obj))
+		return obj;
+
+	bo = gem_to_amdgpu_bo(obj);
+	amdgpu_bo_reserve(bo, true);
+	bo->tbo.sg = sg;
+	bo->tbo.ttm->sg = sg;
+	amdgpu_bo_unreserve(bo);
+	return obj;
+}
+
 static struct sg_table *
 amdgpu_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		       enum dma_data_direction dir)
@@ -293,9 +309,29 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 	return buf;
 }
 
+static void
+amdgpu_gem_prime_invalidate_mappings(struct dma_buf_attachment *attach)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct drm_gem_object *obj = attach->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct ttm_placement placement = {};
+	int r;
+
+	r = ttm_bo_validate(&bo->tbo, &placement, &ctx);
+	if (r)
+		DRM_ERROR("Failed to unmap DMA-buf import (%d))\n", r);
+}
+
 struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 					    struct dma_buf *dma_buf)
 {
+	struct dma_buf_attach_info attach_info = {
+		.dev = dev->dev,
+		.dmabuf = dma_buf,
+		.invalidate = amdgpu_gem_prime_invalidate_mappings
+	};
+	struct dma_buf_attachment *attach;
 	struct drm_gem_object *obj;
 
 	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
@@ -310,5 +346,21 @@ struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 		}
 	}
 
-	return drm_gem_prime_import(dev, dma_buf);
+	if (!dma_buf->invalidation_supported)
+		return drm_gem_prime_import(dev, dma_buf);
+
+	obj = amdgpu_gem_prime_create_obj(dev, dma_buf);
+	if (IS_ERR(obj))
+		return obj;
+
+	attach_info.priv = obj;
+	attach = dma_buf_attach(&attach_info);
+	if (IS_ERR(attach)) {
+		drm_gem_object_put(obj);
+		return ERR_CAST(attach);
+	}
+
+	get_dma_buf(dma_buf);
+	obj->import_attach = attach;
+	return obj;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index d2ab40494a4c..ad93f201e7b6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -43,6 +43,7 @@
 #include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include <linux/iommu.h>
+#include <linux/dma-buf.h>
 #include "amdgpu.h"
 #include "amdgpu_object.h"
 #include "amdgpu_trace.h"
@@ -685,6 +686,7 @@ struct amdgpu_ttm_gup_task_list {
 
 struct amdgpu_ttm_tt {
 	struct ttm_dma_tt	ttm;
+	struct amdgpu_bo	*bo;
 	u64			offset;
 	uint64_t		userptr;
 	struct mm_struct	*usermm;
@@ -993,6 +995,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 		return NULL;
 	}
 	gtt->ttm.ttm.func = &amdgpu_backend_func;
+	gtt->bo = ttm_to_amdgpu_bo(bo);
 	if (ttm_sg_tt_init(&gtt->ttm, bo, page_flags)) {
 		kfree(gtt);
 		return NULL;
@@ -1005,7 +1008,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
 	if (gtt && gtt->userptr) {
 		ttm->sg = kzalloc(sizeof(struct sg_table), GFP_KERNEL);
@@ -1017,7 +1019,19 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 		return 0;
 	}
 
-	if (slave && ttm->sg) {
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG) {
+		if (!ttm->sg) {
+			struct dma_buf_attachment *attach;
+			struct sg_table *sgt;
+
+			attach = gtt->bo->gem_base.import_attach;
+			sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+			if (IS_ERR(sgt))
+				return PTR_ERR(sgt);
+
+			ttm->sg = sgt;
+		}
+
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address,
 						 ttm->num_pages);
@@ -1036,9 +1050,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 
 static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev;
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
+	struct amdgpu_device *adev;
 
 	if (gtt && gtt->userptr) {
 		amdgpu_ttm_tt_set_user_pages(ttm, NULL);
@@ -1047,7 +1060,16 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 		return;
 	}
 
-	if (slave)
+	if (ttm->sg && !gtt->bo->tbo.sg) {
+		struct dma_buf_attachment *attach;
+
+		attach = gtt->bo->gem_base.import_attach;
+		dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL);
+		ttm->sg = NULL;
+		return;
+	}
+
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG)
 		return;
 
 	adev = amdgpu_ttm_adev(ttm->bdev);
-- 
2.14.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH 3/5] drm/ttm: remove the backing store if no placement is given

[He, Roger]

Acked-by: Roger He <Hongbo.He@amd.com>

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Christian K?nig
Sent: Sunday, March 25, 2018 6:58 PM
To: linaro-mm-sig@lists.linaro.org; linux-media@vger.kernel.org; dri-devel@lists.freedesktop.org; amd-gfx@lists.freedesktop.org; sumit.semwal@linaro.org
Subject: [PATCH 3/5] drm/ttm: remove the backing store if no placement is given

Pipeline removal of the BOs backing store when the placement is given during validation.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/ttm_bo.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 98e06f8bf23b..17e821f01d0a 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1078,6 +1078,18 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	uint32_t new_flags;
 
 	reservation_object_assert_held(bo->resv);
+
+	/*
+	 * Remove the backing store if no placement is given.
+	 */
+	if (!placement->num_placement && !placement->num_busy_placement) {
+		ret = ttm_bo_pipeline_gutting(bo);
+		if (ret)
+			return ret;
+
+		return ttm_tt_create(bo, false);
+	}
+
 	/*
 	 * Check whether we need to move buffer.
 	 */
--
2.14.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH 3/5] drm/ttm: remove the backing store if no placement is given

[He, Roger]

Acked-by: Roger He <Hongbo.He@amd.com>

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Christian K?nig
Sent: Sunday, March 25, 2018 6:58 PM
To: linaro-mm-sig@lists.linaro.org; linux-media@vger.kernel.org; dri-devel@lists.freedesktop.org; amd-gfx@lists.freedesktop.org; sumit.semwal@linaro.org
Subject: [PATCH 3/5] drm/ttm: remove the backing store if no placement is given

Pipeline removal of the BOs backing store when the placement is given during validation.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/ttm_bo.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 98e06f8bf23b..17e821f01d0a 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1078,6 +1078,18 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	uint32_t new_flags;
 
 	reservation_object_assert_held(bo->resv);
+
+	/*
+	 * Remove the backing store if no placement is given.
+	 */
+	if (!placement->num_placement && !placement->num_busy_placement) {
+		ret = ttm_bo_pipeline_gutting(bo);
+		if (ret)
+			return ret;
+
+		return ttm_tt_create(bo, false);
+	}
+
 	/*
 	 * Check whether we need to move buffer.
 	 */
--
2.14.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH 5/5] drm/amdgpu: add independent DMA-buf import v3

[Christian König]

Instead of relying on the DRM functions just implement our own import
functions. This adds support for taking care of unpinned DMA-buf.

v2: enable for all exporters, not just amdgpu, fix invalidation
    handling, lock reservation object while setting callback
v3: change to new dma_buf attach interface

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c | 72 ++++++++++++++++++++++++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c   | 32 +++++++++++---
 2 files changed, 89 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 7fef95f0fed1..58dcfba0225a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -86,28 +86,24 @@ int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma
 	return ret;
 }
 
-struct drm_gem_object *
-amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
-				 struct dma_buf_attachment *attach,
-				 struct sg_table *sg)
+static struct drm_gem_object *
+amdgpu_gem_prime_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 {
-	struct reservation_object *resv = attach->dmabuf->resv;
+	struct reservation_object *resv = dma_buf->resv;
 	struct amdgpu_device *adev = dev->dev_private;
 	struct amdgpu_bo *bo;
 	int ret;
 
 	ww_mutex_lock(&resv->lock, NULL);
-	ret = amdgpu_bo_create(adev, attach->dmabuf->size, PAGE_SIZE,
+	ret = amdgpu_bo_create(adev, dma_buf->size, PAGE_SIZE,
 			       AMDGPU_GEM_DOMAIN_CPU, 0, ttm_bo_type_sg,
 			       resv, &bo);
 	if (ret)
 		goto error;
 
-	bo->tbo.sg = sg;
-	bo->tbo.ttm->sg = sg;
 	bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
 	bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
-	if (attach->dmabuf->ops != &amdgpu_dmabuf_ops)
+	if (dma_buf->ops != &amdgpu_dmabuf_ops)
 		bo->prime_shared_count = 1;
 
 	ww_mutex_unlock(&resv->lock);
@@ -118,6 +114,26 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 	return ERR_PTR(ret);
 }
 
+struct drm_gem_object *
+amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
+				 struct dma_buf_attachment *attach,
+				 struct sg_table *sg)
+{
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *bo;
+
+	obj = amdgpu_gem_prime_create_obj(dev, attach->dmabuf);
+	if (IS_ERR(obj))
+		return obj;
+
+	bo = gem_to_amdgpu_bo(obj);
+	amdgpu_bo_reserve(bo, true);
+	bo->tbo.sg = sg;
+	bo->tbo.ttm->sg = sg;
+	amdgpu_bo_unreserve(bo);
+	return obj;
+}
+
 static struct sg_table *
 amdgpu_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		       enum dma_data_direction dir)
@@ -293,9 +309,29 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 	return buf;
 }
 
+static void
+amdgpu_gem_prime_invalidate_mappings(struct dma_buf_attachment *attach)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct drm_gem_object *obj = attach->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct ttm_placement placement = {};
+	int r;
+
+	r = ttm_bo_validate(&bo->tbo, &placement, &ctx);
+	if (r)
+		DRM_ERROR("Failed to unmap DMA-buf import (%d))\n", r);
+}
+
 struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 					    struct dma_buf *dma_buf)
 {
+	struct dma_buf_attach_info attach_info = {
+		.dev = dev->dev,
+		.dmabuf = dma_buf,
+		.invalidate = amdgpu_gem_prime_invalidate_mappings
+	};
+	struct dma_buf_attachment *attach;
 	struct drm_gem_object *obj;
 
 	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
@@ -310,5 +346,21 @@ struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 		}
 	}
 
-	return drm_gem_prime_import(dev, dma_buf);
+	if (!dma_buf->invalidation_supported)
+		return drm_gem_prime_import(dev, dma_buf);
+
+	obj = amdgpu_gem_prime_create_obj(dev, dma_buf);
+	if (IS_ERR(obj))
+		return obj;
+
+	attach_info.importer_priv = obj;
+	attach = dma_buf_attach(&attach_info);
+	if (IS_ERR(attach)) {
+		drm_gem_object_put(obj);
+		return ERR_CAST(attach);
+	}
+
+	get_dma_buf(dma_buf);
+	obj->import_attach = attach;
+	return obj;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index ab73300e6c7f..2a8f328918cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -43,6 +43,7 @@
 #include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include <linux/iommu.h>
+#include <linux/dma-buf.h>
 #include "amdgpu.h"
 #include "amdgpu_object.h"
 #include "amdgpu_trace.h"
@@ -680,6 +681,7 @@ struct amdgpu_ttm_gup_task_list {
 
 struct amdgpu_ttm_tt {
 	struct ttm_dma_tt	ttm;
+	struct amdgpu_bo	*bo;
 	u64			offset;
 	uint64_t		userptr;
 	struct mm_struct	*usermm;
@@ -988,6 +990,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 		return NULL;
 	}
 	gtt->ttm.ttm.func = &amdgpu_backend_func;
+	gtt->bo = ttm_to_amdgpu_bo(bo);
 	if (ttm_sg_tt_init(&gtt->ttm, bo, page_flags)) {
 		kfree(gtt);
 		return NULL;
@@ -1000,7 +1003,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
 	if (gtt && gtt->userptr) {
 		ttm->sg = kzalloc(sizeof(struct sg_table), GFP_KERNEL);
@@ -1012,7 +1014,19 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 		return 0;
 	}
 
-	if (slave && ttm->sg) {
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG) {
+		if (!ttm->sg) {
+			struct dma_buf_attachment *attach;
+			struct sg_table *sgt;
+
+			attach = gtt->bo->gem_base.import_attach;
+			sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+			if (IS_ERR(sgt))
+				return PTR_ERR(sgt);
+
+			ttm->sg = sgt;
+		}
+
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address,
 						 ttm->num_pages);
@@ -1031,9 +1045,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 
 static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev;
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
+	struct amdgpu_device *adev;
 
 	if (gtt && gtt->userptr) {
 		amdgpu_ttm_tt_set_user_pages(ttm, NULL);
@@ -1042,7 +1055,16 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 		return;
 	}
 
-	if (slave)
+	if (ttm->sg && !gtt->bo->tbo.sg) {
+		struct dma_buf_attachment *attach;
+
+		attach = gtt->bo->gem_base.import_attach;
+		dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL);
+		ttm->sg = NULL;
+		return;
+	}
+
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG)
 		return;
 
 	adev = amdgpu_ttm_adev(ttm->bdev);
-- 
2.14.1

[PATCH 5/5] drm/amdgpu: add independent DMA-buf import v3

[Christian König]

Instead of relying on the DRM functions just implement our own import
functions. This adds support for taking care of unpinned DMA-buf.

v2: enable for all exporters, not just amdgpu, fix invalidation
    handling, lock reservation object while setting callback
v3: change to new dma_buf attach interface

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c | 72 ++++++++++++++++++++++++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c   | 32 +++++++++++---
 2 files changed, 89 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 7fef95f0fed1..58dcfba0225a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -86,28 +86,24 @@ int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma
 	return ret;
 }
 
-struct drm_gem_object *
-amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
-				 struct dma_buf_attachment *attach,
-				 struct sg_table *sg)
+static struct drm_gem_object *
+amdgpu_gem_prime_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 {
-	struct reservation_object *resv = attach->dmabuf->resv;
+	struct reservation_object *resv = dma_buf->resv;
 	struct amdgpu_device *adev = dev->dev_private;
 	struct amdgpu_bo *bo;
 	int ret;
 
 	ww_mutex_lock(&resv->lock, NULL);
-	ret = amdgpu_bo_create(adev, attach->dmabuf->size, PAGE_SIZE,
+	ret = amdgpu_bo_create(adev, dma_buf->size, PAGE_SIZE,
 			       AMDGPU_GEM_DOMAIN_CPU, 0, ttm_bo_type_sg,
 			       resv, &bo);
 	if (ret)
 		goto error;
 
-	bo->tbo.sg = sg;
-	bo->tbo.ttm->sg = sg;
 	bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
 	bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
-	if (attach->dmabuf->ops != &amdgpu_dmabuf_ops)
+	if (dma_buf->ops != &amdgpu_dmabuf_ops)
 		bo->prime_shared_count = 1;
 
 	ww_mutex_unlock(&resv->lock);
@@ -118,6 +114,26 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 	return ERR_PTR(ret);
 }
 
+struct drm_gem_object *
+amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
+				 struct dma_buf_attachment *attach,
+				 struct sg_table *sg)
+{
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *bo;
+
+	obj = amdgpu_gem_prime_create_obj(dev, attach->dmabuf);
+	if (IS_ERR(obj))
+		return obj;
+
+	bo = gem_to_amdgpu_bo(obj);
+	amdgpu_bo_reserve(bo, true);
+	bo->tbo.sg = sg;
+	bo->tbo.ttm->sg = sg;
+	amdgpu_bo_unreserve(bo);
+	return obj;
+}
+
 static struct sg_table *
 amdgpu_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		       enum dma_data_direction dir)
@@ -293,9 +309,29 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 	return buf;
 }
 
+static void
+amdgpu_gem_prime_invalidate_mappings(struct dma_buf_attachment *attach)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct drm_gem_object *obj = attach->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct ttm_placement placement = {};
+	int r;
+
+	r = ttm_bo_validate(&bo->tbo, &placement, &ctx);
+	if (r)
+		DRM_ERROR("Failed to unmap DMA-buf import (%d))\n", r);
+}
+
 struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 					    struct dma_buf *dma_buf)
 {
+	struct dma_buf_attach_info attach_info = {
+		.dev = dev->dev,
+		.dmabuf = dma_buf,
+		.invalidate = amdgpu_gem_prime_invalidate_mappings
+	};
+	struct dma_buf_attachment *attach;
 	struct drm_gem_object *obj;
 
 	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
@@ -310,5 +346,21 @@ struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev,
 		}
 	}
 
-	return drm_gem_prime_import(dev, dma_buf);
+	if (!dma_buf->invalidation_supported)
+		return drm_gem_prime_import(dev, dma_buf);
+
+	obj = amdgpu_gem_prime_create_obj(dev, dma_buf);
+	if (IS_ERR(obj))
+		return obj;
+
+	attach_info.importer_priv = obj;
+	attach = dma_buf_attach(&attach_info);
+	if (IS_ERR(attach)) {
+		drm_gem_object_put(obj);
+		return ERR_CAST(attach);
+	}
+
+	get_dma_buf(dma_buf);
+	obj->import_attach = attach;
+	return obj;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index ab73300e6c7f..2a8f328918cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -43,6 +43,7 @@
 #include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include <linux/iommu.h>
+#include <linux/dma-buf.h>
 #include "amdgpu.h"
 #include "amdgpu_object.h"
 #include "amdgpu_trace.h"
@@ -680,6 +681,7 @@ struct amdgpu_ttm_gup_task_list {
 
 struct amdgpu_ttm_tt {
 	struct ttm_dma_tt	ttm;
+	struct amdgpu_bo	*bo;
 	u64			offset;
 	uint64_t		userptr;
 	struct mm_struct	*usermm;
@@ -988,6 +990,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 		return NULL;
 	}
 	gtt->ttm.ttm.func = &amdgpu_backend_func;
+	gtt->bo = ttm_to_amdgpu_bo(bo);
 	if (ttm_sg_tt_init(&gtt->ttm, bo, page_flags)) {
 		kfree(gtt);
 		return NULL;
@@ -1000,7 +1003,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
 	if (gtt && gtt->userptr) {
 		ttm->sg = kzalloc(sizeof(struct sg_table), GFP_KERNEL);
@@ -1012,7 +1014,19 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 		return 0;
 	}
 
-	if (slave && ttm->sg) {
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG) {
+		if (!ttm->sg) {
+			struct dma_buf_attachment *attach;
+			struct sg_table *sgt;
+
+			attach = gtt->bo->gem_base.import_attach;
+			sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+			if (IS_ERR(sgt))
+				return PTR_ERR(sgt);
+
+			ttm->sg = sgt;
+		}
+
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address,
 						 ttm->num_pages);
@@ -1031,9 +1045,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 
 static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev;
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
+	struct amdgpu_device *adev;
 
 	if (gtt && gtt->userptr) {
 		amdgpu_ttm_tt_set_user_pages(ttm, NULL);
@@ -1042,7 +1055,16 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 		return;
 	}
 
-	if (slave)
+	if (ttm->sg && !gtt->bo->tbo.sg) {
+		struct dma_buf_attachment *attach;
+
+		attach = gtt->bo->gem_base.import_attach;
+		dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL);
+		ttm->sg = NULL;
+		return;
+	}
+
+	if (ttm->page_flags & TTM_PAGE_FLAG_SG)
 		return;
 
 	adev = amdgpu_ttm_adev(ttm->bdev);
-- 
2.14.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx