bugzilla-daemon at freedesktop.org
2019-May-15  05:51 UTC
[Nouveau] [Bug 110679] New: va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679
            Bug ID: 110679
           Summary: va-api scaling with nv50-class cards incomplete?
           Product: Mesa
           Version: git
          Hardware: x86 (IA32)
                OS: Linux (All)
            Status: NEW
          Severity: normal
          Priority: medium
         Component: Drivers/DRI/nouveau
          Assignee: nouveau at lists.freedesktop.org
          Reporter: randrik at mail.ru
        QA Contact: nouveau at lists.freedesktop.org
Created attachment 144278
  --> https://bugs.freedesktop.org/attachment.cgi?id=144278&action=edit
ffmpeg's output and command line
Hello!
I was trying to find answer to 'good ffmpeg commandline showing how to use
scale_vaapi filter' [1], but unfortunately none of my attempts worked  for
me
:/
Mesa/va state tarcker:
glxinfo | grep OpenGL
OpenGL vendor string: nouveau
OpenGL renderer string: NV92
OpenGL core profile version string: 3.3 (Core Profile) Mesa 19.2.0-devel
(git-4fd8161773)
OpenGL core profile shading language version string: 3.30
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile
OpenGL core profile extensions:
OpenGL version string: 3.3 (Compatibility Profile) Mesa 19.2.0-devel
(git-4fd8161773)
OpenGL shading language version string: 3.30
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile
OpenGL extensions:
OpenGL ES profile version string: OpenGL ES 3.0 Mesa 19.2.0-devel
(git-4fd8161773)
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.00
OpenGL ES profile extensions:
vainfo:
vainfo
libva info: VA-API version 0.40.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/X11R7/lib/dri/nouveau_drv_video.so
libva info: Found init function __vaDriverInit_0_40
libva info: va_openDriver() returns 0
vainfo: VA-API version: 0.40 (libva )
vainfo: Driver version: Mesa Gallium driver 19.2.0-devel for NV92
vainfo: Supported profile and entrypoints
      VAProfileMPEG2Simple            : VAEntrypointVLD
      VAProfileMPEG2Simple            : VAEntrypointEncSlice
      VAProfileMPEG2Main              : VAEntrypointVLD
      VAProfileMPEG2Main              : VAEntrypointEncSlice
      VAProfileNone                   : VAEntrypointVideoProc
First, I think EncSlice lines are invalid (or mesa can magically (shaders?)
_encode_ mpeg2 videos?)
Second .. ffmpeg command failed for me :/
ffmpeg -vaapi_device /dev/dri/renderD128 -benchmark -c:v libdav1d   -i
/mnt/sdb1/Stream2_AV1_720p_3.2mbps.webm  -vf
format=nv12,hwupload,scale_vaapi=w=240:h=120 -map 0:v:0 -v trace -f null -
ffmpeg version N-93862-gf49cec2ba8 Copyright (c) 2000-2019 the FFmpeg
developers
  built with gcc 5.5.0 (GCC)
[skip]
Applying option vaapi_device (set VAAPI hardware device (DRM path or X11
display name)) with argument /dev/dri/renderD128.
[AVHWDeviceContext @ 0xaef7d00] Opened VA display via DRM device
/dev/dri/renderD128.
libva info: VA-API version 0.40.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/X11R7/lib/dri/nouveau_drv_video.so
libva info: Found init function __vaDriverInit_0_40
libva info: va_openDriver() returns 0
[AVHWDeviceContext @ 0xaef7d00] Initialised VAAPI connection: version 0.40
[AVHWDeviceContext @ 0xaef7d00] Format 0x3231564e -> nv12.
[AVHWDeviceContext @ 0xaef7d00] Format 0x30313050 -> p010le.
[AVHWDeviceContext @ 0xaef7d00] Format 0x36313050 -> unknown.
[AVHWDeviceContext @ 0xaef7d00] Format 0x32315659 -> yuv420p.
[AVHWDeviceContext @ 0xaef7d00] Format 0x56595559 -> unknown.
[AVHWDeviceContext @ 0xaef7d00] Format 0x59565955 -> uyvy422.
[AVHWDeviceContext @ 0xaef7d00] Format 0x41524742 -> bgra.
[AVHWDeviceContext @ 0xaef7d00] Format 0x41424752 -> rgba.
[AVHWDeviceContext @ 0xaef7d00] Format 0x58524742 -> bgr0.
[AVHWDeviceContext @ 0xaef7d00] Format 0x58424752 -> rgb0.
[AVHWDeviceContext @ 0xaef7d00] VAAPI driver: Mesa Gallium driver 19.2.0-devel
for NV92.
[AVHWDeviceContext @ 0xaef7d00] Driver not found in known nonstandard list,
using standard behaviour.
[skip]
[Parsed_format_0 @ 0xaf4dcc0] Setting 'pix_fmts' to value 'nv12'
[Parsed_scale_vaapi_2 @ 0xb005a40] Setting 'w' to value '240'
[Parsed_scale_vaapi_2 @ 0xb005a40] Setting 'h' to value '120'
[graph 0 input from stream 0:0 @ 0xb006300] Setting 'video_size' to
value
'1280x720'
[graph 0 input from stream 0:0 @ 0xb006300] Setting 'pix_fmt' to value
'0'
[graph 0 input from stream 0:0 @ 0xb006300] Setting 'time_base' to value
'1/1000'
[graph 0 input from stream 0:0 @ 0xb006300] Setting 'pixel_aspect' to
value
'1/1'
[graph 0 input from stream 0:0 @ 0xb006300] Setting 'sws_param' to value
'flags=2'
[graph 0 input from stream 0:0 @ 0xb006300] Setting 'frame_rate' to
value
'25/1'
[graph 0 input from stream 0:0 @ 0xb006300] w:1280 h:720 pixfmt:yuv420p
tb:1/1000 fr:25/1 sar:1/1 sws_param:flags=2
[auto_scaler_0 @ 0xb007b40] Setting 'flags' to value 'bicubic'
[auto_scaler_0 @ 0xb007b40] w:iw h:ih flags:'bicubic' interl:0
[Parsed_format_0 @ 0xaf4dcc0] auto-inserting filter 'auto_scaler_0'
between the
filter 'graph 0 input from stream 0:0' and the filter
'Parsed_format_0'
[AVFilterGraph @ 0xaf3ed00] query_formats: 5 queried, 3 merged, 1 already done,
0 delayed
[auto_scaler_0 @ 0xb007b40] w:1280 h:720 fmt:yuv420p sar:1/1 -> w:1280 h:720
fmt:nv12 sar:1/1 flags:0x4
[hwupload @ 0xaf4e200] Surface format is nv12.
[AVHWFramesContext @ 0xaf4e4c0] Created surface 0x1.
[AVHWFramesContext @ 0xaf4e4c0] Direct mapping disabled: deriving image does
not work: 1 (operation failed).
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x3.
[AVHWFramesContext @ 0xb08d2c0] Direct mapping disabled: deriving image does
not work: 1 (operation failed).
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x4.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x5.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x6.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x7.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x8.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0x9.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0xa.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0xb.
[AVHWFramesContext @ 0xb08d2c0] Created surface 0xc.
[AVHWFramesContext @ 0xaf4e4c0] Map surface 0x1.
[AVHWFramesContext @ 0xaf4e4c0] Unmap surface 0x1.
[Parsed_scale_vaapi_2 @ 0xb005a40] Filter input: vaapi_vld, 1280x720 (0).
[Parsed_scale_vaapi_2 @ 0xb005a40] Using surface 0x1 for scale input.
[Parsed_scale_vaapi_2 @ 0xb005a40] Using surface 0xc for scale output.
[Parsed_scale_vaapi_2 @ 0xb005a40] Pipeline parameter buffer is 0xe.
[Parsed_scale_vaapi_2 @ 0xb005a40] Failed to render parameter buffer: 2
(resource allocation failed).
Error while filtering: Input/output error
Failed to inject frame into filter network: Input/output error
Error while processing the decoded data for stream #0:0
bench: maxrss=106420kB
[AVIOContext @ 0xaf40640] Statistics: 759514 bytes read, 0 seeks
Conversion failed!
Ow :( [full log attached]
Note, i use libva1:
libva-1.8.3
Can libva2 be installed in parallel with libva1 ?
[1] - https://lists.cinelerra-gg.org/pipermail/cin/2019-May/000665.html
-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://lists.freedesktop.org/archives/nouveau/attachments/20190515/8b9c5f19/attachment.html>
bugzilla-daemon at freedesktop.org
2019-Jun-05  14:49 UTC
[Nouveau] [Bug 110679] va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679 --- Comment #1 from Andrew Randrianasulu <randrik at mail.ru> --- After adding XVMC_VL=1 variable ffmpeg command above started to work ?! Apparently, 'normal' nouveau video allocator dislikes two differently-sized buffers (command also worked when I set "scale_vaapi=w=1280:h=720" , in other words to strictly same dimensions as input ....) -- You are receiving this mail because: You are the QA Contact for the bug. You are the assignee for the bug. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://lists.freedesktop.org/archives/nouveau/attachments/20190605/efb4c958/attachment.html>
bugzilla-daemon at freedesktop.org
2019-Jun-05  15:01 UTC
[Nouveau] [Bug 110679] va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679 --- Comment #2 from Ilia Mirkin <imirkin at alum.mit.edu> --- (In reply to Andrew Randrianasulu from comment #1)> After adding XVMC_VL=1 variable ffmpeg command above started to work ?!I believe XVMC_VL uses shaders to do the mpeg decoding. It's not the built-in vp2 (or vpe) processor doing the decoding. I *have* seen some weird stuff with vdpau where it looks like surfaces are sized weird and everything breaks (like something is off by 8px in width or height). I have never tried to really track it down. -- You are receiving this mail because: You are the QA Contact for the bug. You are the assignee for the bug. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://lists.freedesktop.org/archives/nouveau/attachments/20190605/389aacf7/attachment.html>
bugzilla-daemon at freedesktop.org
2019-Jun-05  16:38 UTC
[Nouveau] [Bug 110679] va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679 --- Comment #3 from Andrew Randrianasulu <randrik at mail.ru> --- using those two command lines XVMC_VL=1 LIBVA_MESSAGING_LEVEL=2 LIBVA_TRACE=/dev/shm/log_file ./ffmpeg -vaapi_device /dev/dri/renderD128 -benchmark -c:v libdav1d -i /mnt/sdb1/Stream2_AV1_720p_3.2mbps.webm -vf format=nv12,hwupload,scale_vaapi=w=3840:h=2160:mode=hq,hwdownload,format=yuv420p -map 0:v:0 /dev/shm/1.avi ---------working-------- LIBVA_MESSAGING_LEVEL=2 LIBVA_TRACE=/dev/shm/log_file_no_vl ./ffmpeg -vaapi_device /dev/dri/renderD128 -benchmark -c:v libdav1d -i /mnt/sdb1/Stream2_AV1_720p_3.2mbps.webm -vf format=nv12,hwupload,scale_vaapi=w=3840:h=2160:mode=hq,hwdownload,format=yuv420p -map 0:v:0 /dev/shm/1.avi --------failing---------- I got two trace files, will attach them. -- You are receiving this mail because: You are the assignee for the bug. You are the QA Contact for the bug. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://lists.freedesktop.org/archives/nouveau/attachments/20190605/69ff06f4/attachment.html>
bugzilla-daemon at freedesktop.org
2019-Jun-05  16:39 UTC
[Nouveau] [Bug 110679] va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679 --- Comment #4 from Andrew Randrianasulu <randrik at mail.ru> --- Created attachment 144462 --> https://bugs.freedesktop.org/attachment.cgi?id=144462&action=edit libva log file (with VL activated) -- You are receiving this mail because: You are the QA Contact for the bug. You are the assignee for the bug. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://lists.freedesktop.org/archives/nouveau/attachments/20190605/1955e52f/attachment.html>
bugzilla-daemon at freedesktop.org
2019-Jun-05  16:39 UTC
[Nouveau] [Bug 110679] va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679 --- Comment #5 from Andrew Randrianasulu <randrik at mail.ru> --- Created attachment 144463 --> https://bugs.freedesktop.org/attachment.cgi?id=144463&action=edit libva log without VL -- You are receiving this mail because: You are the assignee for the bug. You are the QA Contact for the bug. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://lists.freedesktop.org/archives/nouveau/attachments/20190605/d94fdc67/attachment.html>
bugzilla-daemon at freedesktop.org
2019-Sep-18  20:47 UTC
[Nouveau] [Bug 110679] va-api scaling with nv50-class cards incomplete?
https://bugs.freedesktop.org/show_bug.cgi?id=110679
GitLab Migration User <gitlab-migration at fdo.invalid> changed:
           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |MOVED
             Status|NEW                         |RESOLVED
--- Comment #6 from GitLab Migration User <gitlab-migration at
fdo.invalid> ---
-- GitLab Migration Automatic Message --
This bug has been migrated to freedesktop.org's GitLab instance and has been
closed from further activity.
You can subscribe and participate further through the new bug through this link
to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/1179.
-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://lists.freedesktop.org/archives/nouveau/attachments/20190918/529b8bc3/attachment.html>