Bugzilla – Bug 1222715
segmentation fault with vdr-plugin-xineliboutput and libvdpau
Last modified: 2024-06-05 07:55:32 UTC
Hello, With a fresh new installation of Leap-15.5 I get a segmentation fault when trying to run vdr with xineliboutput and vdpau video driver. Steps to reproduce: - new installation of Leap-15.5 (basic graphical interface) - zypper ar https://download.opensuse.org/repositories/vdr/15.5/vdr.repo - zypper in vdr-plugin-xineliboutput - vdr -P"xineliboutput -Vvdpau" Last lines on the terminal: video_decoder: first_frame_flag = 2. load_plugins: plugin dxr3-mpeg2 failed to instantiate itself. video_out_vdpau: video surface #9 (1920 x 1080) aligned to (1920 x 1088). video_out: vo_open (0x7f414c541640) load_plugins: plugin vdpau_mpeg12 will be used for video streamtype 00. video_out_vdpau: video surface #10 (704 x 576). Segmentation fault Last line of journal: vdr[4629]: [4639] [xine..put] cXinelibLocal::Action - fe config OK vdr[4629]: [4646] [demux_vdr] video fifo flush and decoder reset after still image vdr[4629]: [4646] [demux_vdr] H.264 scanner: Possible MPEG2 start code (0xb3) kernel: Local decoder/d[4643]: segfault at 48 ip 00007f4157688e25 sp 00007f413e9fe460 error 4 in libvdpau_r600.so.1.0.0[7f4157200000+b71000] likely on CPU 1 (core 1, socket 0) kernel: Code: 45 00 01 f0 83 2e 01 0f 84 c0 01 00 00 0f 1f 84 00 00 00 00 00 48 8b 83 d8 00 00 00 48 89 ab e8 00 00 00 66 0f ef c0 49 89 e4 <48> 8b 68 48 0f b7 45 44 0f b7 55 40 0f 29 04 24 48 89 ee c6 43 3f Kind regards, Peter
Sorry, I can only test vdr-plugin-xineliboutput in remote mode (via vdr-sxfe) and that one works. I also have no radeon card to reproduce the crash (a quick test locally on intel graphics did not segfault, but I had no input device to test with). So unless you have a patch that I could apply, there is unfortunately not much I can do. Does it work without VDPAU?
(In reply to Stefan Seyfried from comment #1) > Does it work without VDPAU? Yes.
Then please check if it is a generic xine problem, with something like xine -V vdpau file.mp4 If this also segfaults, it is a xine problem. But even if this is not a xine but a vdr-xineliboutput problem, unless there is an upstream patch there is not much I can do to resolve this.
(In reply to Stefan Seyfried from comment #3) > Then please check if it is a generic xine problem, with something like > > xine -V vdpau file.mp4 No segfault here...
Too bad ;-) Still, nothing I can do here unless upstream has an patch for the issue. I'd suggest to just not use VDPAU then.
Hm, maybe it is specific to mpeg2 playback? so maybe also try playing a mpeg2 (SDTV vdr recording) with xine..., maybe this will trigger the segfault there, too?
there is actually a new release 2.3.0 on sourceforge But there is no mentioning of vdpau fixes in the changelog I submitted it to the vdr project anyway after a quick smoke test, maybe this helps your issue
(In reply to Stefan Seyfried from comment #6) > so maybe also try playing a mpeg2 (SDTV vdr recording) with xine Then I have another error: This is xine (X11 gui) - a free video player v0.99.14. (c) 2000-2019 The xine Team. xiTK received SIGSEGV signal, RIP. Aborted (core dumped)
I’ve tried with VLC now. Output on terminal: VLC media player 3.0.20 Vetinari (revision 3.0.20-0-g6f0d0ab126b) [0000559cd56ff0d0] vlcpulse audio output error: PulseAudio server connection failure: Connection refused [0000559cd56674c0] main libvlc: Running vlc with the default interface. Use 'cvlc' to use vlc without interface. [00007f22f0007da0] gl gl: Initialized libplacebo v4.208.0 (API v208) libva info: VA-API version 1.16.0 libva info: Trying to open /usr/lib64/dri/r600_drv_video.so libva info: Found init function __vaDriverInit_1_16 libva info: va_openDriver() returns 0 [00007f2300c07480] avcodec decoder: Using Mesa Gallium driver 22.3.5 for AMD RV730 (DRM 2.50.0 / 6.8.8-lp155.9.1-default, LLVM 15.0.7) for hardware decoding Segmentation fault (core dumped) Syslog: May 29 15:29:44 media kernel: vlc[10743]: segfault at 48 ip 00007f233119d045 sp 00007f235c7fe4d0 error 4 in r600_dri.so[7f2330600000+17d6000] likely on CPU 1 (core 1, socket 0) May 29 15:29:44 media kernel: Code: 45 00 01 f0 83 2e 01 0f 84 c0 01 00 00 0f 1f 84 00 00 00 00 00 48 8b 83 d8 00 00 00 48 89 ab e8 00 00 00 66 0f ef c0 49 89 e4 <48> 8b 68 48 0f b7 45 44 0f b7 55 40 0f 29 04 24 48 89 ee c6 43 3f
so it looks like this is not related to vdr-xineliboutput but to something VDPAU side :-)
> "RESOLVED WORKSFORME" Does this mean, that you cannot reproduce it? And that it won’t be fixed?
This means it is not a bug in vdr-plugin-xineliboutput AFAICT and thus there is not much I could do, even if I would be able to reproduce it (which I am not as I don't have radeon hardware). You need to report this to the maintainers of the VDPAU stuff.
I’ve just seen, that the error is with "r600_dri.so" instead of "libvdpau_r600.so.1.0.0", when triggered by VLC. What do you think: 2 different issues, or the same one?
I have no idea. I don't know anything about dri, vdpau and the likes.