v4l2 should not need ENUMSTD

[Clément Bœsch]

Problem: lavd/v4l2.c now requires the device to support VIDIOC_ENUMSTD.

This seems to be a regression according to ​https://github.com/umlaeute/v4l2loopback/issues/38

Steps to reproduce:

• install v4l2loopback and load the module

  git clone git://github.com/umlaeute/v4l2loopback.git
  cd v4l2loopback
  make && sudo make install
  modprobe v4l2loopback
  

• push some video to the device with for instance:

  gst-launch-0.10 videotestsrc ! v4l2sink device=/dev/video0 
  

• try to play the device with FFmpeg

  % ./ffplay -v verbose -f v4l2 /dev/video0
  ffplay version N-51026-g3313b9c Copyright (c) 2003-2013 the FFmpeg developers
    built on Mar 17 2013 04:38:52 with gcc 4.7.2 (GCC)
    configuration: --enable-gpl --enable-version3 --enable-fontconfig --enable-libfreetype --enable-libmp3lame --enable-libvorbis --enable-libxvid --enable-libx264 --enable-libvpx --enable-libtheora --enable-x11grab --enable-libopenjpeg --enable-libass --enable-libmodplug --enable-libv4l2 --cc=colorgcc --samples=/home/ubitux/fate-samples --prefix=/tmp/ffinstall --disable-runtime-cpudetect --enable-libcelt --enable-libopencv --enable-frei0r --enable-libcaca --enable-libiec61883 --enable-libopencore-amrwb --enable-libopencore-amrnb --enable-libopus --enable-libpulse --enable-libspeex --enable-openal --enable-libopencv
    libavutil      52. 19.101 / 52. 19.101
    libavcodec     55.  0.100 / 55.  0.100
    libavformat    55.  0.100 / 55.  0.100
    libavdevice    55.  0.100 / 55.  0.100
    libavfilter     3. 47.102 /  3. 47.102
    libswscale      2.  2.100 /  2.  2.100
    libswresample   0. 17.102 /  0. 17.102
    libpostproc    52.  2.100 / 52.  2.100
  [video4linux2,v4l2 @ 0x7f457c000940] fd:7 capabilities:50000030/0   
  [video4linux2,v4l2 @ 0x7f457c000940] Querying the device for the current frame size
  [video4linux2,v4l2 @ 0x7f457c000940] Setting frame size to 320x240
  [video4linux2,v4l2 @ 0x7f457c000940] ioctl(VIDIOC_ENUMSTD): Invalid argument
  /dev/video0: Invalid argument
  

The device plays fine with for example mplayer, and with older ffplay (1.1.3 is working).

[Carl Eugen Hoyos]

Which change in FFmpeg introduced the regression?

[Clément Bœsch]

514216d8a928cf80980ca8dc5daea46e9d41cad9 is the first bad commit.

[Clément Bœsch]

Note that while supporting the latest version of v4l2loopback is also important, it seems some workaround recently got pushed in it to handle that issue (see github ticket). To reproduce properly the problem, you need to checkout 5568514931b063eda0aa83976fd9a68cacc53e52 in v4l2loopback repository.

[holden]

The problem has been fixed on v4l2loopback git/master, so now ffmpeg should work again.
I will try to see if we can apply a workaround to allow ffmpeg to work with older versions of v4l2loopback.

[Clément Bœsch]

Any news from this or it should be ignored?

[mrlika]

This regression effects DVP-7620HE device driver. Latest ffmpeg throws the following error:

ffmpeg -f video4linux2 -i /dev/video1 -an -vcodec copy -s cif  test.h264 -loglevel verbose
ffmpeg version git-2013-07-12-0da0caa Copyright (c) 2000-2013 the FFmpeg developers
  built on Jul 12 2013 16:55:30 with gcc 4.7 (Ubuntu/Linaro 4.7.3-1ubuntu1)
  configuration: --enable-libfaac --enable-libmp3lame --enable-nonfree --disable-vaapi --enable-shared --enable-libx264 --enable-version3 --enable-gpl
  libavutil      52. 39.100 / 52. 39.100
  libavcodec     55. 18.102 / 55. 18.102
  libavformat    55. 12.102 / 55. 12.102
  libavdevice    55.  3.100 / 55.  3.100
  libavfilter     3. 80.100 /  3. 80.100
  libswscale      2.  3.100 /  2.  3.100
  libswresample   0. 17.102 /  0. 17.102
  libpostproc    52.  3.100 / 52.  3.100
[video4linux2,v4l2 @ 0x1467900] fd:3 capabilities:4000001
[video4linux2,v4l2 @ 0x1467900] Querying the device for the current frame size
[video4linux2,v4l2 @ 0x1467900] Setting frame size to 1920x1080
[video4linux2,v4l2 @ 0x1467900] ioctl(VIDIOC_ENUMSTD): Invalid argument
/dev/video1: Invalid argument

Older FFmpeg versions work correct with this device:

ffmpeg -f video4linux2 -i /dev/video1 -an -vcodec copy -s cif  test.h264 -loglevel verbose
ffmpeg version 0.8.6-6:0.8.6-1ubuntu2, Copyright (c) 2000-2013 the Libav developers
  built on Mar 30 2013 22:20:06 with gcc 4.7.2
  configuration: --arch=amd64 --enable-pthreads --enable-runtime-cpudetect --extra-version='6:0.8.6-1ubuntu2' --libdir=/usr/lib/x86_64-linux-gnu --prefix=/usr --enable-bzlib --enable-libdc1394 --enable-libfreetype --enable-gnutls --enable-libgsm --enable-libpulse --enable-libschroedinger --enable-libspeex --enable-libtheora --enable-vaapi --enable-vdpau --enable-libvorbis --enable-libvpx --enable-zlib --enable-gpl --enable-postproc --enable-swscale --enable-libcdio --enable-x11grab --shlibdir=/usr/lib/x86_64-linux-gnu --enable-shared --disable-static
  libavutil    51. 22. 1 / 51. 22. 1
  libavcodec   53. 35. 0 / 53. 35. 0
  libavformat  53. 21. 1 / 53. 21. 1
  libavdevice  53.  2. 0 / 53.  2. 0
  libavfilter   2. 15. 0 /  2. 15. 0
  libswscale    2.  1. 0 /  2.  1. 0
  libpostproc  52.  0. 0 / 52.  0. 0
*** THIS PROGRAM IS DEPRECATED ***
This program is only provided for compatibility and will be removed in a future release. Please use avconv instead.                                                                                     
[video4linux2 @ 0x1b6a1e0] [4]Capabilities: 4000001
[video4linux2 @ 0x1b6a1e0] Querying the device for the current frame size
[video4linux2 @ 0x1b6a1e0] Setting frame size to 1920x1080
[video4linux2 @ 0x1b6a1e0] Estimating duration from bitrate, this may be inaccurate
Input #0, video4linux2, from '/dev/video1':
  Duration: N/A, start: 1378115441.148443, bitrate: 829440 kb/s
    Stream #0.0: Video: rawvideo, yuyv422, 1920x1080, 829440 kb/s, 25 tbr, 1000k tbn, 25 tbc
File 'test.h264' already exists. Overwrite ? [y/N] y
Output #0, h264, to 'test.h264':
  Metadata:
    encoder         : Lavf53.21.1
    Stream #0.0: Video: rawvideo, yuyv422, 1920x1080, q=2-31, 829440 kb/s, 90k tbn, 25 tbc
Stream mapping:
  Stream #0.0 -> #0.0
Press ctrl-c to stop encoding

[Carl Eugen Hoyos]

Replying to mrlika:

Older FFmpeg versions work correct with this device:

ffmpeg -f video4linux2 -i /dev/video1 -an -vcodec copy -s cif  test.h264 -loglevel verbose
ffmpeg version 0.8.6-6:0.8.6-1ubuntu2

This unfortunately is not FFmpeg, testing makes no sense (this intentionally broken version has several hundred known regressions, some of them security relevant).
Please test 402ea62

[mrlika]

Patch that fixes the issue:

diff --git a/libavdevice/v4l2.c b/libavdevice/v4l2.c
index c7cd519..7ada54f 100644
--- a/libavdevice/v4l2.c
+++ b/libavdevice/v4l2.c
@@ -687,6 +687,10 @@ static int v4l2_set_parameters(AVFormatContext *s1)
             standard.index = i;
             if (v4l2_ioctl(s->fd, VIDIOC_ENUMSTD, &standard) < 0) {
                 ret = AVERROR(errno);
+                if (ret == AVERROR(EINVAL)) {
+                    tpf = &streamparm.parm.capture.timeperframe;
+                    break;
+                }
                 av_log(s1, AV_LOG_ERROR, "ioctl(VIDIOC_ENUMSTD): %s\n", av_err2str(ret));
                 return ret;
             }

The patch sent to ffmpeg-devel for review: ​http://ffmpeg.org/pipermail/ffmpeg-devel/2013-September/147793.html

[holden]

Replying to mrlika:

Patch that fixes the issue:

Hi,

thanks for the patch. Could you please apply the attached patch to the latest ffmpeg git and provide the output of the following commands (change you video device if you need to):

ffmpeg -loglevel debug -f video4linux2 -list_formats all -i /dev/video0
ffmpeg -loglevel debug -f video4linux2 -list_standards all -i /dev/video0
ffmpeg -loglevel debug -f video4linux2 -i /dev/video0 out.avi

Thanks.

[mrlika]

Output with 0002-lavd-v4l2-make-sure-VIDIOC_G_STD-does-not-return-std.patch

[mrlika]

Not patched output

[mrlika]

Output with my patch

[mrlika]

Hi Holden,

patched.txt - output of the commands from latest FFmpeg with your patch
not_patched.txt - output of the commands from not patched latest FFmpeg
patched_my.txt - output from FFmpeg with my patch

Only with my patch third command starts encoding.

Regards.

[holden]

Replying to mrlika:

Hi Holden,

patched.txt - output of the commands from latest FFmpeg with your patch
not_patched.txt - output of the commands from not patched latest FFmpeg
patched_my.txt - output from FFmpeg with my patch

Only with my patch third command starts encoding.

Regards.

Thanks for the tests, I changed the commit message on your original patch and sent it to the mailing list. Once it gets applied I think we can close this ticket.

[mrlika]

Thank you

[Carl Eugen Hoyos]

Fixed by mrlika in ed72542

[Carl Eugen Hoyos]

Should the patch be backported to release/2.0 ?

[holden]

Replying to cehoyos:

Should the patch be backported to release/2.0 ?

Sorry for the late replay, but if we're still in time I'd say sure, it can be backported.

[Carl Eugen Hoyos]

Merged into release/2.0 and release/1.2.