Announcement

Collapse
No announcement yet.

[GMA X4500] GPU video decoding (VA-API?)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #41
    Originally posted by darkbasic View Post
    Uh? Does it already work with GMA X4500? O_O
    I think it still requires kernel patches?

    Comment


    • #42
      Interesting, but what now?

      I have Linux 2.6.35-rc2, libdrm git 20100609, and libva git HEAD.

      After building with --enable-i965-driver I get:

      Code:
      $ vainfo 
      libva: libva version 0.31.1
      libva: va_getDriverName() returns 0
      libva: Trying to open /usr/lib/dri/i965_drv_video.so
      libva: va_openDriver() returns 0
      vainfo: VA API version: 0.31
      vainfo: Driver version: i965 Driver 0.1
      vainfo: Supported profile and entrypoints
            VAProfileMPEG2Simple            :	VAEntrypointVLD
            VAProfileMPEG2Main              :	VAEntrypointVLD
            VAProfileH264Baseline           :	VAEntrypointVLD
            VAProfileH264Main               :	VAEntrypointVLD
            VAProfileH264High               :	VAEntrypointVLD
      vainfo: i965_media.c:295: i965_media_terminate: Assertion `media_state->free_private_context' failed.
      Aborted
      And using mplayer-vaapi

      Code:
      $ mplayer -quiet -vo vaapi -va vaapi h264.mp4 
      MPlayer SVN-r31303-4.4.3 (C) 2000-2010 MPlayer Team
      
      Playing h264.mp4.
      libavformat file format detected.
      [lavf] stream 0: video (h264), -vid 0
      VIDEO:  [H264]  320x240  24bpp  12.000 fps  1159.2 kbps (141.5 kbyte/s)
      Clip info:
       major_brand: isom
       minor_version: 512
       compatible_brands: isomiso2avc1mp41
       encoder: Lavf52.32.0
      libva: libva version 0.31.1
      libva: va_getDriverName() returns 0
      libva: Trying to open /usr/lib/dri/i965_drv_video.so
      libva: va_openDriver() returns 0
      ==========================================================================
      Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
      [VD_FFMPEG] VA API accelerated codec.
      Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
      ==========================================================================
      Audio: no sound
      Starting playback...
      Unsupported PixelFormat 61
      [VD_FFMPEG] Trying pixfmt=1.
      Movie-Aspect is 1.33:1 - prescaling to correct movie aspect.
      VO: [vaapi] 320x240 => 320x240 H.264 VA-API Acceleration  [fs]
      [VD_FFMPEG] XVMC-accelerated MPEG-2.
      mplayer: i965_avc_ildb.c:621: i965_avc_ildb_decode_init: Assertion `avc_ildb_context->urb.cs_start + avc_ildb_context->urb.num_cs_entries * avc_ildb_context->urb.size_cs_entry <= ((((&i965->intel)->device_id == 0x0042) || ((&i965->intel)->device_id == 0x0046)) ? 1024 : (((&i965->intel)->device_id == 0x2E02 || (&i965->intel)->device_id == 0x2E12 || (&i965->intel)->device_id == 0x2E22 || (&i965->intel)->device_id == 0x2E32) || ((&i965->intel)->device_id == 0x2A42)) ? 384 : 256)' failed.
      Am I testing too early?

      Comment


      • #43
        I'm still not sure if the kernel side of things have been mainlined yet? And that assertion failure you're getting when just running vainfo seems to indicate badness...

        A status update from the Intel devs, or a short Phoronix article would be very welcome.

        Comment


        • #44
          From the notes in the new libdrm release, it sounds like it should work with 2.6.35. I guess filing a bug is in order, or trying the latest kernel snapshot first.

          Comment


          • #45
            I’m not sure if GM45 chipset is already supported, or if this is only meant to support Ironlake yet.

            Do you have concrete mentions of GM45, G45?

            Comment


            • #46
              From the libdrm release notes:

              1) media ring support on kernel 2.6.35 for doing media decode on G45+
              I guess this includes GM45 too?

              Comment


              • #47
                But there is no mention here:


                I guess some bits are still missing?

                Comment


                • #48
                  Word from the developer:

                  No, GM45 support will schedule for Q3 this year.

                  Thanks
                  Zou Nan hai

                  Comment


                  • #49
                    I was wondering why does it take so long to implement this feature when they have all the specs they need ? Does this represent a lot of lines of codes in the driver ?

                    Comment


                    • #50
                      Originally posted by beecher View Post
                      I was wondering why does it take so long to implement this feature when they have all the specs they need ? Does this represent a lot of lines of codes in the driver ?
                      Yes, the patches that went in recently had quite a long, invasive diffstat.

                      Comment

                      Working...
                      X