Hi Laurent, Thankyou for the speedy review. On 04/08/17 17:03, Laurent Pinchart wrote: > Hi Kieran, > > Thank you for the patch. > > On Friday 04 Aug 2017 16:57:05 Kieran Bingham wrote: >> Presently any received buffers are only released back to vb2 if >> vsp1_video_stop_streaming() is called. If vsp1_video_start_streaming() >> encounters an error, we will be warned by the vb2 handlers that buffers >> have not been returned. >> >> Move the buffer cleanup code to it's own function to prevent duplication > > s/it's/its/ Ah yes - I'm always terrible with my its'y bits. > >> and call from both vsp1_video_stop_streaming() and the error path in >> vsp1_video_start_streaming() > > s/$/./ :D > >> >> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@xxxxxxxxxxxxxxxx> >> --- >> drivers/media/platform/vsp1/vsp1_video.c | 22 +++++++++++++++------- >> 1 file changed, 15 insertions(+), 7 deletions(-) >> >> diff --git a/drivers/media/platform/vsp1/vsp1_video.c >> b/drivers/media/platform/vsp1/vsp1_video.c index 5af3486afe07..a24033429cd7 >> 100644 >> --- a/drivers/media/platform/vsp1/vsp1_video.c >> +++ b/drivers/media/platform/vsp1/vsp1_video.c >> @@ -822,6 +822,19 @@ static int vsp1_video_setup_pipeline(struct >> vsp1_pipeline *pipe) return 0; >> } >> >> +static void vsp1_video_cleanup_pipeline(struct vsp1_video *video) > > Should this function take a pipe pointer instead of a video pointer for > symmetry with vsp1_video_setup_pipeline() ? I passed this way because the cleanup needed a *video. Is it possible to get from a *pipe to a *video? >> +{ >> + struct vsp1_vb2_buffer *buffer; >> + unsigned long flags; >> + >> + /* Remove all buffers from the IRQ queue. */ >> + spin_lock_irqsave(&video->irqlock, flags); >> + list_for_each_entry(buffer, &video->irqqueue, queue) >> + vb2_buffer_done(&buffer->buf.vb2_buf, VB2_BUF_STATE_ERROR); >> + INIT_LIST_HEAD(&video->irqqueue); >> + spin_unlock_irqrestore(&video->irqlock, flags); >> +} >> + >> static int vsp1_video_start_streaming(struct vb2_queue *vq, unsigned int >> count) { >> struct vsp1_video *video = vb2_get_drv_priv(vq); >> @@ -835,6 +848,7 @@ static int vsp1_video_start_streaming(struct vb2_queue >> *vq, unsigned int count) ret = vsp1_video_setup_pipeline(pipe); >> if (ret < 0) { >> mutex_unlock(&pipe->lock); >> + vsp1_video_cleanup_pipeline(video); >> return ret; >> } >> >> @@ -866,7 +880,6 @@ static void vsp1_video_stop_streaming(struct vb2_queue >> *vq) { >> struct vsp1_video *video = vb2_get_drv_priv(vq); >> struct vsp1_pipeline *pipe = video->rwpf->pipe; >> - struct vsp1_vb2_buffer *buffer; >> unsigned long flags; >> int ret; >> >> @@ -893,12 +906,7 @@ static void vsp1_video_stop_streaming(struct vb2_queue >> *vq) media_pipeline_stop(&video->video.entity); >> vsp1_video_pipeline_put(pipe); >> >> - /* Remove all buffers from the IRQ queue. */ >> - spin_lock_irqsave(&video->irqlock, flags); >> - list_for_each_entry(buffer, &video->irqqueue, queue) >> - vb2_buffer_done(&buffer->buf.vb2_buf, VB2_BUF_STATE_ERROR); >> - INIT_LIST_HEAD(&video->irqqueue); >> - spin_unlock_irqrestore(&video->irqlock, flags); >> + vsp1_video_cleanup_pipeline(video); > > The vsp1_video_cleanup_pipeline() call should go before > vsp1_video_pipeline_put(), as you've noticed in patch 7/7. I chose to do the move in 3/7 so that this patch did not change the existing functionality. There is no (explicit) need for the cleanup to happen before the pipeline_put() until the cleanup function references the pipe... > > With all that fixed, > > Reviewed-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> > >> } >> >> static const struct vb2_ops vsp1_video_queue_qops = { >