[Bug 112133] AMD Radeon 5700 / Navi: VAAPI encoding generates garbled output

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Fri Oct 25 10:37:58 UTC 2019


https://bugs.freedesktop.org/show_bug.cgi?id=112133

            Bug ID: 112133
           Summary: AMD Radeon 5700 / Navi: VAAPI encoding generates
                    garbled output
           Product: DRI
           Version: DRI git
          Hardware: x86-64 (AMD64)
                OS: Linux (All)
            Status: NEW
          Severity: normal
          Priority: not set
         Component: DRM/AMDgpu
          Assignee: dri-devel at lists.freedesktop.org
          Reporter: toscar at gmx.net

Using ffmpeg to encode video to h264/hevc with constant or variable bitrate
results in completely garbled output video on navi/radeon 5700 gpu.

ffmpeg command used:
ffmpeg -init_hw_device vaapi=foo:/dev/dri/renderD128 -hwaccel vaapi
-hwaccel_output_format vaapi -hwaccel_device foo \
-i <input file> -filter_hw_device foo -vf 'format=nv12|vaapi,hwupload' -c:v
hevc_vaapi -b:v 5000k -maxrate 5000k -f matroska <output file>
-This issue also applies to h264 encoding (replace "hevc_vaapi" with
"h264_vaapi" above).

HW used:
AMD Radeon 5700 XT

SW used:
linux-amd-staging-drm-next-git-5.4.858569.1c2677f991d3-1-x86_64
mesa-git 1:19.3.0_devel.116641.f11ea226664
libva-utils 2.5.0
ffmpeg 1:4.2.1

This issue sounds a bit like BUG 109538, however #109538 speaks of "garbled
line at the bottom" where I get completely garbled output, not just at the
bottom.

In additon, ffmpeg stalls/crashes sometimes as described in #109538 with:
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring vcn_enc0 timeout, signaled
seq=150025, emitted seq=150027
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process ffmpeg
pid 228611 thread ffmpeg:cs0 pid 228612

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/dri-devel/attachments/20191025/211b9439/attachment.html>


More information about the dri-devel mailing list