diff mbox

[FFmpeg-devel,2/2] avcodec/v4l2: fix segmentation fault on codec exit

Message ID 1507276322-19752-1-git-send-email-jorge.ramirez-ortiz@linaro.org
State New
Headers show

Commit Message

Jorge Ramirez-Ortiz Oct. 6, 2017, 7:52 a.m. UTC 
It occurs when the codec is closed while buffer references still
exist. This is a regression from the original patchset where support
for this use-case was implemented.

The regression occurred while cleaning the code for the last patchset
(decoding was tested only with ffplay which disposes of the buffer
straightaway hence the feature went in broken/untested)
---
 libavcodec/v4l2_m2m.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

Comments

Jorge Ramirez-Ortiz Oct. 6, 2017, 6:59 p.m. UTC | #1 
On 10/06/2017 09:52 AM, Jorge Ramirez-Ortiz wrote:
> It occurs when the codec is closed while buffer references still
> exist. This is a regression from the original patchset where support
> for this use-case was implemented.
>
> The regression occurred while cleaning the code for the last patchset
> (decoding was tested only with ffplay which disposes of the buffer
> straightaway hence the feature went in broken/untested)
> ---
>   libavcodec/v4l2_m2m.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)

bug fix. needed in 3.4

>
> diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c
> index bd96a6d..5e85bcb 100644
> --- a/libavcodec/v4l2_m2m.c
> +++ b/libavcodec/v4l2_m2m.c
> @@ -331,8 +331,10 @@ int ff_v4l2_m2m_codec_end(AVCodecContext *avctx)
>   
>       ff_v4l2_context_release(&s->output);
>   
> -    if (atomic_load(&s->refcount))
> -        av_log(avctx, AV_LOG_ERROR, "ff_v4l2m2m_codec_end leaving pending buffers\n");
> +    if (atomic_load(&s->refcount)) {
> +        av_log(avctx, AV_LOG_DEBUG, "ff_v4l2m2m_codec_end leaving pending buffers\n");
> +        return 0;
> +    }
>   
>       ff_v4l2_context_release(&s->capture);
>       sem_destroy(&s->refsync);
Mark Thompson Oct. 6, 2017, 7:53 p.m. UTC | #2 
On 06/10/17 08:52, Jorge Ramirez-Ortiz wrote:
> It occurs when the codec is closed while buffer references still
> exist. This is a regression from the original patchset where support
> for this use-case was implemented.
> 
> The regression occurred while cleaning the code for the last patchset
> (decoding was tested only with ffplay which disposes of the buffer
> straightaway hence the feature went in broken/untested)
> ---
>  libavcodec/v4l2_m2m.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c
> index bd96a6d..5e85bcb 100644
> --- a/libavcodec/v4l2_m2m.c
> +++ b/libavcodec/v4l2_m2m.c
> @@ -331,8 +331,10 @@ int ff_v4l2_m2m_codec_end(AVCodecContext *avctx)
>  
>      ff_v4l2_context_release(&s->output);
>  
> -    if (atomic_load(&s->refcount))
> -        av_log(avctx, AV_LOG_ERROR, "ff_v4l2m2m_codec_end leaving pending buffers\n");
> +    if (atomic_load(&s->refcount)) {
> +        av_log(avctx, AV_LOG_DEBUG, "ff_v4l2m2m_codec_end leaving pending buffers\n");
> +        return 0;
> +    }
>  
>      ff_v4l2_context_release(&s->capture);
>      sem_destroy(&s->refsync);
> 

Trying to test this, I segfault when decoding finishes with this patch applied (exynos s5p-mfc):

(gdb) r
Starting program: /home/mrt/ffmpeg/v4l2/build/ffmpeg_g -threads 1 -v 55 -y -c:v h264_v4l2m2m -i /home/mrt/bbb_1080_264.mp4 -an -frames:v 100 -f null -
...
Output file #0 (pipe:):
  Output stream #0:0 (video): 100 frames encoded; 100 packets muxed (40000 bytes); 
  Total: 100 packets (40000 bytes) muxed
[h264_v4l2m2m @ 0x1403720] ff_v4l2m2m_codec_end leaving pending buffers
100 frames successfully decoded, 0 decoding errors
...
Thread 1 "ffmpeg_g" received signal SIGSEGV, Segmentation fault.
ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at src/libavcodec/v4l2_context.c:489
489         int type = ctx->type;
(gdb) bt
#0  ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at src/libavcodec/v4l2_context.c:489
#1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at src/libavcodec/v4l2_m2m.c:319
#2  0x00835aba in buffer_replace (src=0x0, dst=<optimized out>) at src/libavutil/buffer.c:120
#3  av_buffer_unref (buf=buf@entry=0x14d8b2c) at src/libavutil/buffer.c:130
#4  0x008417ac in av_frame_unref (frame=frame@entry=0x14d8a30) at src/libavutil/frame.c:515
#5  0x00841d58 in av_frame_unref (frame=<optimized out>) at src/libavutil/frame.c:147
#6  av_frame_free (frame=frame@entry=0x1403d9c) at src/libavutil/frame.c:168
#7  0x00082bce in ffmpeg_cleanup (ret=0) at src/fftools/ffmpeg.c:541
#8  0x00079c42 in exit_program (ret=0) at src/fftools/cmdutils.c:138
#9  0x0006efba in main (argc=<optimized out>, argv=<optimized out>) at src/fftools/ffmpeg.c:4793
(gdb) p ctx
$1 = (V4L2Context *) 0x10fc
(gdb) f 1
#1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at src/libavcodec/v4l2_m2m.c:319
319         ret = ff_v4l2_context_set_status(&s->output, VIDIOC_STREAMOFF);
(gdb) p s
$2 = (V4L2m2mContext *) 0x0
(gdb)
Mark Thompson Oct. 6, 2017, 8:01 p.m. UTC | #3 
On 06/10/17 20:53, Mark Thompson wrote:
> On 06/10/17 08:52, Jorge Ramirez-Ortiz wrote:
>> It occurs when the codec is closed while buffer references still
>> exist. This is a regression from the original patchset where support
>> for this use-case was implemented.
>>
>> The regression occurred while cleaning the code for the last patchset
>> (decoding was tested only with ffplay which disposes of the buffer
>> straightaway hence the feature went in broken/untested)
>> ---
>>  libavcodec/v4l2_m2m.c | 6 ++++--
>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c
>> index bd96a6d..5e85bcb 100644
>> --- a/libavcodec/v4l2_m2m.c
>> +++ b/libavcodec/v4l2_m2m.c
>> @@ -331,8 +331,10 @@ int ff_v4l2_m2m_codec_end(AVCodecContext *avctx)
>>  
>>      ff_v4l2_context_release(&s->output);
>>  
>> -    if (atomic_load(&s->refcount))
>> -        av_log(avctx, AV_LOG_ERROR, "ff_v4l2m2m_codec_end leaving pending buffers\n");
>> +    if (atomic_load(&s->refcount)) {
>> +        av_log(avctx, AV_LOG_DEBUG, "ff_v4l2m2m_codec_end leaving pending buffers\n");
>> +        return 0;
>> +    }
>>  
>>      ff_v4l2_context_release(&s->capture);
>>      sem_destroy(&s->refsync);
>>
> 
> Trying to test this, I segfault when decoding finishes with this patch applied (exynos s5p-mfc):
> 
> (gdb) r
> Starting program: /home/mrt/ffmpeg/v4l2/build/ffmpeg_g -threads 1 -v 55 -y -c:v h264_v4l2m2m -i /home/mrt/bbb_1080_264.mp4 -an -frames:v 100 -f null -
> ...
> Output file #0 (pipe:):
>   Output stream #0:0 (video): 100 frames encoded; 100 packets muxed (40000 bytes); 
>   Total: 100 packets (40000 bytes) muxed
> [h264_v4l2m2m @ 0x1403720] ff_v4l2m2m_codec_end leaving pending buffers
> 100 frames successfully decoded, 0 decoding errors
> ...
> Thread 1 "ffmpeg_g" received signal SIGSEGV, Segmentation fault.
> ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at src/libavcodec/v4l2_context.c:489
> 489         int type = ctx->type;
> (gdb) bt
> #0  ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at src/libavcodec/v4l2_context.c:489
> #1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at src/libavcodec/v4l2_m2m.c:319
> #2  0x00835aba in buffer_replace (src=0x0, dst=<optimized out>) at src/libavutil/buffer.c:120
> #3  av_buffer_unref (buf=buf@entry=0x14d8b2c) at src/libavutil/buffer.c:130
> #4  0x008417ac in av_frame_unref (frame=frame@entry=0x14d8a30) at src/libavutil/frame.c:515
> #5  0x00841d58 in av_frame_unref (frame=<optimized out>) at src/libavutil/frame.c:147
> #6  av_frame_free (frame=frame@entry=0x1403d9c) at src/libavutil/frame.c:168
> #7  0x00082bce in ffmpeg_cleanup (ret=0) at src/fftools/ffmpeg.c:541
> #8  0x00079c42 in exit_program (ret=0) at src/fftools/cmdutils.c:138
> #9  0x0006efba in main (argc=<optimized out>, argv=<optimized out>) at src/fftools/ffmpeg.c:4793
> (gdb) p ctx
> $1 = (V4L2Context *) 0x10fc
> (gdb) f 1
> #1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at src/libavcodec/v4l2_m2m.c:319
> 319         ret = ff_v4l2_context_set_status(&s->output, VIDIOC_STREAMOFF);
> (gdb) p s
> $2 = (V4L2m2mContext *) 0x0
> (gdb) 
> 

To clarify, it segfaults in the same way without the patch as well.

- Mark
Jorge Ramirez-Ortiz Oct. 6, 2017, 9:38 p.m. UTC | #4 
On 10/06/2017 10:01 PM, Mark Thompson wrote:
> On 06/10/17 20:53, Mark Thompson wrote:
>> On 06/10/17 08:52, Jorge Ramirez-Ortiz wrote:
>>> It occurs when the codec is closed while buffer references still
>>> exist. This is a regression from the original patchset where support
>>> for this use-case was implemented.
>>>
>>> The regression occurred while cleaning the code for the last patchset
>>> (decoding was tested only with ffplay which disposes of the buffer
>>> straightaway hence the feature went in broken/untested)
>>> ---
>>>   libavcodec/v4l2_m2m.c | 6 ++++--
>>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c
>>> index bd96a6d..5e85bcb 100644
>>> --- a/libavcodec/v4l2_m2m.c
>>> +++ b/libavcodec/v4l2_m2m.c
>>> @@ -331,8 +331,10 @@ int ff_v4l2_m2m_codec_end(AVCodecContext *avctx)
>>>   
>>>       ff_v4l2_context_release(&s->output);
>>>   
>>> -    if (atomic_load(&s->refcount))
>>> -        av_log(avctx, AV_LOG_ERROR, "ff_v4l2m2m_codec_end leaving pending buffers\n");
>>> +    if (atomic_load(&s->refcount)) {
>>> +        av_log(avctx, AV_LOG_DEBUG, "ff_v4l2m2m_codec_end leaving pending buffers\n");
>>> +        return 0;
>>> +    }
>>>   
>>>       ff_v4l2_context_release(&s->capture);
>>>       sem_destroy(&s->refsync);
>>>
>> Trying to test this, I segfault when decoding finishes with this patch applied (exynos s5p-mfc):
>>
>> (gdb) r
>> Starting program: /home/mrt/ffmpeg/v4l2/build/ffmpeg_g -threads 1 -v 55 -y -c:v h264_v4l2m2m -i /home/mrt/bbb_1080_264.mp4 -an -frames:v 100 -f null -
>> ...
>> Output file #0 (pipe:):
>>    Output stream #0:0 (video): 100 frames encoded; 100 packets muxed (40000 bytes);
>>    Total: 100 packets (40000 bytes) muxed
>> [h264_v4l2m2m @ 0x1403720] ff_v4l2m2m_codec_end leaving pending buffers
>> 100 frames successfully decoded, 0 decoding errors
>> ...
>> Thread 1 "ffmpeg_g" received signal SIGSEGV, Segmentation fault.
>> ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at src/libavcodec/v4l2_context.c:489
>> 489         int type = ctx->type;
>> (gdb) bt
>> #0  ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at src/libavcodec/v4l2_context.c:489
>> #1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at src/libavcodec/v4l2_m2m.c:319
>> #2  0x00835aba in buffer_replace (src=0x0, dst=<optimized out>) at src/libavutil/buffer.c:120
>> #3  av_buffer_unref (buf=buf@entry=0x14d8b2c) at src/libavutil/buffer.c:130
>> #4  0x008417ac in av_frame_unref (frame=frame@entry=0x14d8a30) at src/libavutil/frame.c:515
>> #5  0x00841d58 in av_frame_unref (frame=<optimized out>) at src/libavutil/frame.c:147
>> #6  av_frame_free (frame=frame@entry=0x1403d9c) at src/libavutil/frame.c:168
>> #7  0x00082bce in ffmpeg_cleanup (ret=0) at src/fftools/ffmpeg.c:541
>> #8  0x00079c42 in exit_program (ret=0) at src/fftools/cmdutils.c:138
>> #9  0x0006efba in main (argc=<optimized out>, argv=<optimized out>) at src/fftools/ffmpeg.c:4793
>> (gdb) p ctx
>> $1 = (V4L2Context *) 0x10fc
>> (gdb) f 1
>> #1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at src/libavcodec/v4l2_m2m.c:319
>> 319         ret = ff_v4l2_context_set_status(&s->output, VIDIOC_STREAMOFF);
>> (gdb) p s
>> $2 = (V4L2m2mContext *) 0x0
>> (gdb)
>>
> To clarify, it segfaults in the same way without the patch as well.

vow, very strange.

would you mind doing a clean build with the patch applied?
I am asking because that test works fine on my db410c and 820 (and hardware 
shouldnt make a difference in this test)

also the team at Kodi also tested the fix on their imx6



>
> - Mark
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Jorge Ramirez-Ortiz Oct. 6, 2017, 9:45 p.m. UTC | #5 
On 10/06/2017 11:38 PM, Jorge Ramirez-Ortiz wrote:
> On 10/06/2017 10:01 PM, Mark Thompson wrote:
>> On 06/10/17 20:53, Mark Thompson wrote:
>>> On 06/10/17 08:52, Jorge Ramirez-Ortiz wrote:
>>>> It occurs when the codec is closed while buffer references still
>>>> exist. This is a regression from the original patchset where support
>>>> for this use-case was implemented.
>>>>
>>>> The regression occurred while cleaning the code for the last patchset
>>>> (decoding was tested only with ffplay which disposes of the buffer
>>>> straightaway hence the feature went in broken/untested)
>>>> ---
>>>>   libavcodec/v4l2_m2m.c | 6 ++++--
>>>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c
>>>> index bd96a6d..5e85bcb 100644
>>>> --- a/libavcodec/v4l2_m2m.c
>>>> +++ b/libavcodec/v4l2_m2m.c
>>>> @@ -331,8 +331,10 @@ int ff_v4l2_m2m_codec_end(AVCodecContext *avctx)
>>>>         ff_v4l2_context_release(&s->output);
>>>>   -    if (atomic_load(&s->refcount))
>>>> -        av_log(avctx, AV_LOG_ERROR, "ff_v4l2m2m_codec_end leaving pending 
>>>> buffers\n");
>>>> +    if (atomic_load(&s->refcount)) {
>>>> +        av_log(avctx, AV_LOG_DEBUG, "ff_v4l2m2m_codec_end leaving pending 
>>>> buffers\n");
>>>> +        return 0;
>>>> +    }
>>>>         ff_v4l2_context_release(&s->capture);
>>>>       sem_destroy(&s->refsync);
>>>>
>>> Trying to test this, I segfault when decoding finishes with this patch 
>>> applied (exynos s5p-mfc):
>>>
>>> (gdb) r
>>> Starting program: /home/mrt/ffmpeg/v4l2/build/ffmpeg_g -threads 1 -v 55 -y 
>>> -c:v h264_v4l2m2m -i /home/mrt/bbb_1080_264.mp4 -an -frames:v 100 -f null -
>>> ...
>>> Output file #0 (pipe:):
>>>    Output stream #0:0 (video): 100 frames encoded; 100 packets muxed (40000 
>>> bytes);
>>>    Total: 100 packets (40000 bytes) muxed
>>> [h264_v4l2m2m @ 0x1403720] ff_v4l2m2m_codec_end leaving pending buffers
>>> 100 frames successfully decoded, 0 decoding errors
>>> ...
>>> Thread 1 "ffmpeg_g" received signal SIGSEGV, Segmentation fault.
>>> ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at 
>>> src/libavcodec/v4l2_context.c:489
>>> 489         int type = ctx->type;
>>> (gdb) bt
>>> #0  ff_v4l2_context_set_status (ctx=ctx@entry=0x10fc, cmd=1074026003) at 
>>> src/libavcodec/v4l2_context.c:489
>>> #1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at 
>>> src/libavcodec/v4l2_m2m.c:319
>>> #2  0x00835aba in buffer_replace (src=0x0, dst=<optimized out>) at 
>>> src/libavutil/buffer.c:120
>>> #3  av_buffer_unref (buf=buf@entry=0x14d8b2c) at src/libavutil/buffer.c:130
>>> #4  0x008417ac in av_frame_unref (frame=frame@entry=0x14d8a30) at 
>>> src/libavutil/frame.c:515
>>> #5  0x00841d58 in av_frame_unref (frame=<optimized out>) at 
>>> src/libavutil/frame.c:147
>>> #6  av_frame_free (frame=frame@entry=0x1403d9c) at src/libavutil/frame.c:168
>>> #7  0x00082bce in ffmpeg_cleanup (ret=0) at src/fftools/ffmpeg.c:541
>>> #8  0x00079c42 in exit_program (ret=0) at src/fftools/cmdutils.c:138
>>> #9  0x0006efba in main (argc=<optimized out>, argv=<optimized out>) at 
>>> src/fftools/ffmpeg.c:4793
>>> (gdb) p ctx
>>> $1 = (V4L2Context *) 0x10fc
>>> (gdb) f 1
>>> #1  0x00798c24 in ff_v4l2_m2m_codec_end (avctx=0x1403720) at 
>>> src/libavcodec/v4l2_m2m.c:319
>>> 319         ret = ff_v4l2_context_set_status(&s->output, VIDIOC_STREAMOFF);
>>> (gdb) p s
>>> $2 = (V4L2m2mContext *) 0x0
>>> (gdb)
>>>
>> To clarify, it segfaults in the same way without the patch as well.
>
> vow, very strange.
>
> would you mind doing a clean build with the patch applied?
> I am asking because that test works fine on my db410c and 820 (and hardware 
> shouldnt make a difference in this test)
>
> also the team at Kodi also tested the fix on their imx6
>
>

this is what I see when I run your command

linaro@db820 git.zoltan.ffmpeg (ffmpeg/master $)]$ ./ffmpeg_g -threads 1 -v 55 
-y -c:v h264_v4l2m2m -i ~/Videos/h264.FVDO_Freeway_720p.264 -an -frames:v 100 -f 
null -

[...]


[h264_v4l2m2m @ 0xaaaadd3bbb50] capture: NV12 20 buffers initialized: 1280x0736, 
sizeimage 01425408, bytesperline 00001280
cur_dts is invalid (this is harmless if it occurs once at the start per stream)
     Last message repeated 9 times
detected 4 logical cores
[graph 0 input from stream 0:0 @ 0xaaaadd419680] Setting 'video_size' to value 
'1280x720'
[graph 0 input from stream 0:0 @ 0xaaaadd419680] Setting 'pix_fmt' to value '25'
[graph 0 input from stream 0:0 @ 0xaaaadd419680] Setting 'time_base' to value 
'1/1200000'
[graph 0 input from stream 0:0 @ 0xaaaadd419680] Setting 'pixel_aspect' to value 
'0/1'
[graph 0 input from stream 0:0 @ 0xaaaadd419680] Setting 'sws_param' to value 
'flags=2'
[graph 0 input from stream 0:0 @ 0xaaaadd419680] Setting 'frame_rate' to value 
'25/1'
[graph 0 input from stream 0:0 @ 0xaaaadd419680] w:1280 h:720 pixfmt:nv12 
tb:1/1200000 fr:25/1 sar:0/1 sws_param:flags=2
[AVFilterGraph @ 0xaaaadd438530] query_formats: 3 queried, 2 merged, 0 already 
done, 0 delayed
Output #0, null, to 'pipe:':
   Metadata:
     encoder         : Lavf57.82.102
     Stream #0:0, 0, 1/25: Video: wrapped_avframe, 1 reference frame, 
nv12(left), 1280x720, 0/1, q=2-31, 200 kb/s, 25 fps, 25 tbn, 25 tbc
     Metadata:
       encoder         : Lavc57.106.104 wrapped_avframe
No more output streams to write to, finishing.:00.04 bitrate=N/A speed=0.0398x
frame=  100 fps= 91 q=-0.0 Lsize=N/A time=00:00:00.04 bitrate=N/A speed=0.0363x
video:52kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing 
overhead: unknown
Input file #0 (/home/linaro/Videos/h264.FVDO_Freeway_720p.264):
   Input stream #0:0 (video): 209 packets read (4070226 bytes); 101 frames decoded;
   Total: 209 packets (4070226 bytes) demuxed
Output file #0 (pipe:):
   Output stream #0:0 (video): 100 frames encoded; 100 packets muxed (52800 bytes);
   Total: 100 packets (52800 bytes) muxed
[h264_v4l2m2m @ 0xaaaadd3bbb50] ff_v4l2m2m_codec_end leaving pending buffers   
<---------------------------------------
101 frames successfully decoded, 0 decoding errors
[AVIOContext @ 0xaaaadd366d70] Statistics: 4096000 bytes read, 0 seeks


>
>>
>> - Mark
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel@ffmpeg.org
>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
>
diff mbox

Patch

diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c
index bd96a6d..5e85bcb 100644
--- a/libavcodec/v4l2_m2m.c
+++ b/libavcodec/v4l2_m2m.c
@@ -331,8 +331,10 @@  int ff_v4l2_m2m_codec_end(AVCodecContext *avctx)
 
     ff_v4l2_context_release(&s->output);
 
-    if (atomic_load(&s->refcount))
-        av_log(avctx, AV_LOG_ERROR, "ff_v4l2m2m_codec_end leaving pending buffers\n");
+    if (atomic_load(&s->refcount)) {
+        av_log(avctx, AV_LOG_DEBUG, "ff_v4l2m2m_codec_end leaving pending buffers\n");
+        return 0;
+    }
 
     ff_v4l2_context_release(&s->capture);
     sem_destroy(&s->refsync);