[PATCH v3] media: vimc: streamer: if kthread_stop fails, ignore the error

[PATCH v3] media: vimc: streamer: if kthread_stop fails, ignore the error

[Dafna Hirschfeld]

Ignore errors returned from kthread_stop since the
vimc subdevices should still be notified that
streaming stopped so they can release the memory for
the streaming, and also kthread should be set to NULL.
kthread_stop can return -EINTR in case the thread
did not yet run. This can happen if userspace calls
streamon and streamoff right after.

Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
---
changes since v2: add a comment explaning why we ignore the error

 drivers/media/platform/vimc/vimc-streamer.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/media/platform/vimc/vimc-streamer.c b/drivers/media/platform/vimc/vimc-streamer.c
index cd6b55433c9e..6a6fa6c21138 100644
--- a/drivers/media/platform/vimc/vimc-streamer.c
+++ b/drivers/media/platform/vimc/vimc-streamer.c
@@ -215,9 +215,13 @@ int vimc_streamer_s_stream(struct vimc_stream *stream,
 			return 0;
 
 		ret = kthread_stop(stream->kthread);
-		if (ret)
-			return ret;
 
+		/*
+		 * ignore errors from kthread_stop since we have to set kthread
+		 * to NULL and notify the subdevs that the stream stopped anyway
+		 */
+		if (ret)
+			dev_warn(ved->dev, "kthread_stop returned '%d'\n", ret);
 		stream->kthread = NULL;
 
 		vimc_streamer_pipeline_terminate(stream);
-- 
2.17.1

Re: [PATCH v3] media: vimc: streamer: if kthread_stop fails, ignore the error

[Helen Koike]

On 1/17/20 7:50 AM, Dafna Hirschfeld wrote:
> Ignore errors returned from kthread_stop since the
> vimc subdevices should still be notified that
> streaming stopped so they can release the memory for
> the streaming, and also kthread should be set to NULL.
> kthread_stop can return -EINTR in case the thread
> did not yet run. This can happen if userspace calls
> streamon and streamoff right after.
> 
> Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
> ---
> changes since v2: add a comment explaning why we ignore the error
> 
>  drivers/media/platform/vimc/vimc-streamer.c | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/platform/vimc/vimc-streamer.c b/drivers/media/platform/vimc/vimc-streamer.c
> index cd6b55433c9e..6a6fa6c21138 100644
> --- a/drivers/media/platform/vimc/vimc-streamer.c
> +++ b/drivers/media/platform/vimc/vimc-streamer.c
> @@ -215,9 +215,13 @@ int vimc_streamer_s_stream(struct vimc_stream *stream,
>  			return 0;
>  
>  		ret = kthread_stop(stream->kthread);
> -		if (ret)
> -			return ret;
>  

this new line is not required

> +		/*
> +		 * ignore errors from kthread_stop since we have to set kthread
> +		 * to NULL and notify the subdevs that the stream stopped anyway
> +		 */

hmm, I think what we could make it clear is in which situation kthread_stop would return
an error.

How about:

"kthread_stop returns error in cases when streamon was immediately followed by
streamoff, and the thread didn't had a chance to run.
Ignore errors to stop the stream in the pipeline."

Helen

> +		if (ret)
> +			dev_warn(ved->dev, "kthread_stop returned '%d'\n", ret);
>  		stream->kthread = NULL;
>  
>  		vimc_streamer_pipeline_terminate(stream);
>