2022-07-07 04:13:31

by jie1zhan

Subject: [PATCH v1] drm/syncobj: Fix sync syncobj issue

enable signaling after flatten dma_fence_chains on transfer

Signed-off-by: jie1zhan <[email protected]>

Reviewed-by: Christian König <[email protected]>

Reviewed-by: Nirmoy Das <[email protected]>
---
drivers/gpu/drm/drm_syncobj.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c
index 7e48dcd1bee4..0d9d3577325f 100644
--- a/drivers/gpu/drm/drm_syncobj.c
+++ b/drivers/gpu/drm/drm_syncobj.c
@@ -920,6 +920,7 @@ static int drm_syncobj_transfer_to_timeline(struct drm_file *file_private,
if (ret)
goto err_free_fence;

+ dma_fence_enable_sw_signaling(fence);
chain = dma_fence_chain_alloc();
if (!chain) {
ret = -ENOMEM;
--
2.25.1

2022-07-07 08:38:22

by Christian König

Subject: Re: [PATCH v1] drm/syncobj: Fix sync syncobj issue

Am 07.07.22 um 05:08 schrieb jie1zhan:
> enable signaling after flatten dma_fence_chains on transfer
>
> Signed-off-by: jie1zhan <[email protected]>
>
> Reviewed-by: Christian König <[email protected]>
>
> Reviewed-by: Nirmoy Das <[email protected]>

Well please drop those Reviewed-by lines immediately! Neither Nirmoy nor
me gave you the permission to add an rb to this patch. It looks like you
have a major misunderstanding when to add that.

> ---
> drivers/gpu/drm/drm_syncobj.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c
> index 7e48dcd1bee4..0d9d3577325f 100644
> --- a/drivers/gpu/drm/drm_syncobj.c
> +++ b/drivers/gpu/drm/drm_syncobj.c
> @@ -920,6 +920,7 @@ static int drm_syncobj_transfer_to_timeline(struct drm_file *file_private,
> if (ret)
> goto err_free_fence;
>
> + dma_fence_enable_sw_signaling(fence);

That looks better, but ideally we would enable signaling when the fence
is retrieved from the drm_syncobj and queried for it's signaled state.

Regards,
Christian.

> chain = dma_fence_chain_alloc();
> if (!chain) {
> ret = -ENOMEM;