Bug#847701: kodi: Movie player unusable with libva error
Bálint Réczey
balint at balintreczey.hu
Fri Jan 13 15:15:59 UTC 2017
Control: reassign -1 libgl1-mesa-dri 13.0.3-1
Control: affects -1 kodi
2017-01-12 17:42 GMT+01:00 Leandro sullo sgorbio <lnoferin at cybervalley.org>:
> Il giorno mer, 11/01/2017 alle 20.54 +0100, Bálint Réczey ha scritto:
>
>> >> I don't think that your GPU is supported by i965-va-driver. So using the VA API
>> >> rendering engine won't work for you.
>> >
>> > Yes, the module is the i915. Also disabling all the acceleration engines
>> > the player does not work.
>>
>> Please disable both VDPAU and VAAPI and try playing a low resolution video
>> file. If it does not work please attach kodi's log and detail how it failed.
>
> I tried again with VDPAU and VAAPI disabled and again, the reproduction
> of a "low resolution video" :-) did not work.
>
> These are the relevant (I hope) lines of log:
>
> 17:29:33.623 T:139943960140672 NOTICE: VideoPlayer:
> Opening: /dati/Film/Serie TV/Star Trek TNG/StarTrek.TNG-s01e03-The Naked
> Now (Contaminazione).avi
> 17:29:33.623 T:139943960140672 WARNING: CDVDMessageQueue(player)::Put
> MSGQ_NOT_INITIALIZED
> 17:29:33.624 T:139940386023168 NOTICE: Creating InputStream
> 17:29:33.631 T:139940386023168 NOTICE: Creating Demuxer
> 17:29:33.845 T:139940386023168 NOTICE: Opening stream: 0 source: 256
> 17:29:33.845 T:139940386023168 NOTICE: Creating video codec with codec
> id: 13
> 17:29:33.845 T:139940386023168 NOTICE: CDVDVideoCodecFFmpeg::Open()
> Using codec: MPEG-4 part 2
> 17:29:33.846 T:139940386023168 NOTICE: Creating video thread
> 17:29:33.846 T:139940386023168 NOTICE: Opening stream: 1 source: 256
> 17:29:33.846 T:139940180850432 NOTICE: running thread: video_thread
> 17:29:33.846 T:139940386023168 NOTICE: Finding audio codec for: 86017
> 17:29:33.847 T:139940386023168 NOTICE: Creating audio thread
> 17:29:33.847 T:139940147287808 NOTICE: running thread:17:29:33.623
> T:139943960140672 NOTICE: VideoPlayer: Opening: /dati/Film/Serie
> TV/Star Trek TNG/StarTrek.TNG-s01e03-The Naked Now (Contaminazione).avi
...
> 17:30:17.611 T:139940180850432 WARNING: Previous line repeats 226 times.
> 17:30:17.611 T:139940180850432 WARNING: CRenderManager::WaitForBuffer -
> timeout waiting for buffer
> 17:30:21.925 T:139943960140672 WARNING: Previous line repeats 2 times.
> 17:30:21.925 T:139943960140672 NOTICE: CVideoPlayer::CloseFile()
> 17:30:21.925 T:139943960140672 NOTICE: VideoPlayer: waiting for threads
> to exit
> 17:30:21.931 T:139940386023168 NOTICE: CVideoPlayer::OnExit()
> 17:30:21.931 T:139940386023168 NOTICE: Closing stream player 1
> 17:30:21.931 T:139940386023168 NOTICE: Waiting for audio thread to exit
> 17:30:22.085 T:139940147287808 NOTICE: thread end:
> CVideoPlayerAudio::OnExit()
...
> The file is:
>
> StarTrek.TNG-s01e03-The Naked Now (Contaminazione).avi: RIFF
> (little-endian) data, AVI, 640 x 480, 25.00 fps, video: DivX 5, audio:
> MPEG-1 Layer 3 (stereo, 48000 Hz)
>
>> > So with this GPU this version of kodi is unusable?
>>
>> Most likely kodi is usable but video decoding will be slow thus you
>> can watch only low resolution videos (max ~720p) or videos which can use fast codecs for
>> decoding.
It looks like the system is really slow for the video. I have
reinstalled a similar system and playing videos was very slow on it,
indeed.
The system:
$ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 28
model name : Intel(R) Atom(TM) CPU N270 @ 1.60GHz
stepping : 2
microcode : 0x208
cpu MHz : 1067.000
cache size : 512 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 1
...
# lspci -v -s 00:02.1
00:02.1 Display controller: Intel Corporation Mobile 945GM/GMS/GME,
943/940GML Express Integrated Graphics Controller (rev 03)
Subsystem: Fujitsu Technology Solutions Mobile 945GM/GMS/GME,
943/940GML Express Integrated Graphics Controller
Flags: bus master, fast devsel, latency 0
Memory at f8180000 (32-bit, non-prefetchable) [size=512K]
Capabilities: [d0] Power Management version 2
I have collected some profiling data by running:
$ perf record -g kodi
Which shows that ~70% of the time is spent in intel's mesa driver:
$ perf report --sort comm,dso | cat
Failed to open /tmp/perf-6001.map, continuing without symbols
# To display the perf.data header info, please use
--header/--header-only options.
#
#
# Total Lost Samples: 0
#
# Samples: 406K of event 'cycles:p'
# Event count (approx.): 4725363215008
#
# Children Self Command Shared Object
# ........ ........ ............... ..................................
#
72.68% 71.68% kodi.bin i915_dri.so
|
|--18.88%--0
| |
| |--3.33%--fetch_vector4
| |
| |--2.31%--fetch_texel_lod
| |
| |--1.75%--_mesa_execute_program
| |
| |--1.70%--lerp_rgba_2d
| |
| |--1.67%--.L184
| |
| |--1.32%--__x86.get_pc_thunk.bx
| |
| |--1.10%--.L180
| |
| |--1.09%--.L200
| |
| --0.99%--.L149
|
|--9.49%--.L149
|
|--6.60%--.L147
|
|--4.75%--store_vector4
|
|--3.77%--_einittext
| |
| |--2.21%--fetch_vector4
| |
| --1.06%--.L117
|
|--2.28%--0x20247f81
| |
| |--1.63%--fetch_vector4
| |
| --0.65%--.L119
|
|--1.90%--0x81ffe5cf
| _swrast_span_render_finish
| |
| --1.90%--_swrast_exec_fragment_program
|
|--1.78%--sample_2d_linear.isra.28
|
|--1.76%--0x380978b
| |
| --0.95%--.L116
|
|--1.43%--linear_texel_locations
|
|--1.15%--_mesa_unpack_rgba_row
|
|--0.97%--0x3
| |
| --0.65%--fetch_texel_lod
|
|--0.93%--fetch_vector4
|
|--0.79%--fetch_texel_2d_A_UNORM8
|
|--0.66%--__x86.get_pc_thunk.bx
|
|--0.65%--0x3d808081
|
|--0.65%--fetch_texel_2d_L_UNORM8
|
--0.63%--_mesa_convert_colors
...
I believe this is a regression in the driver thus I'm reassigning it to mesa.
Dear mesa maintainers, please take a look.
Cheers,
Balint
More information about the pkg-multimedia-maintainers
mailing list