diff mbox series

[3/6] dma-buf: Warn about dma_fence_chain container rules v2

Message ID 20220204100429.2049-4-christian.koenig@amd.com (mailing list archive)
State New, archived
Headers show
Series [1/6] dma-buf: consolidate dma_fence subclass checking | expand

Commit Message

Christian König Feb. 4, 2022, 10:04 a.m. UTC
Chaining of dma_fence_chain objects is only allowed through the prev
fence and not through the contained fence.

Warn about that when we create a dma_fence_chain.

v2: fix comment style

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/dma-buf/dma-fence-chain.c | 9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Thomas Hellström Feb. 9, 2022, 2:02 p.m. UTC | #1
On Fri, 2022-02-04 at 11:04 +0100, Christian König wrote:
> Chaining of dma_fence_chain objects is only allowed through the prev
> fence and not through the contained fence.
> 
> Warn about that when we create a dma_fence_chain.
> 
> v2: fix comment style
> 
> Signed-off-by: Christian König <christian.koenig@amd.com>
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>

It looks like this blows up in generic drm code...

https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_22201/shard-skl10/igt@syncobj_timeline@transfer-timeline-point.html

/Thomas
Christian König Feb. 9, 2022, 2:42 p.m. UTC | #2
Am 09.02.22 um 15:02 schrieb Thomas Hellström:
> On Fri, 2022-02-04 at 11:04 +0100, Christian König wrote:
>> Chaining of dma_fence_chain objects is only allowed through the prev
>> fence and not through the contained fence.
>>
>> Warn about that when we create a dma_fence_chain.
>>
>> v2: fix comment style
>>
>> Signed-off-by: Christian König <christian.koenig@amd.com>
>> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> It looks like this blows up in generic drm code...
>
> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_22201/shard-skl10/igt@syncobj_timeline@transfer-timeline-point.html

Thanks for the notice.  Going to take a look.

I'm wondering why the last CI report I've got didn't showed that.

Christian.

>
> /Thomas
>
>
diff mbox series

Patch

diff --git a/drivers/dma-buf/dma-fence-chain.c b/drivers/dma-buf/dma-fence-chain.c
index 1b4cb3e5cec9..084c6927b735 100644
--- a/drivers/dma-buf/dma-fence-chain.c
+++ b/drivers/dma-buf/dma-fence-chain.c
@@ -254,5 +254,14 @@  void dma_fence_chain_init(struct dma_fence_chain *chain,
 
 	dma_fence_init(&chain->base, &dma_fence_chain_ops,
 		       &chain->lock, context, seqno);
+
+	/*
+	 * Chaining dma_fence_chain container together is only allowed through
+	 * the prev fence and not through the contained fence.
+	 *
+	 * The correct way of handling this is to flatten out the fence
+	 * structure into a dma_fence_array by the caller instead.
+	 */
+	WARN_ON(dma_fence_is_chain(fence));
 }
 EXPORT_SYMBOL(dma_fence_chain_init);