[TGL] gst-msdk decode vc1 cmd runtime hang on TGL

[wenqingx]

System information

• CPU information(cat /proc/cpuinfo | grep "model name" | uniq):11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz
• GPU information(lspci -nn | grep -E 'VGA|isplay):0000:00:02.0 VGA compatible controller [0300]: Intel Corporation TigerLake-LP GT2 [Iris Xe Graphics] [8086:9a49] (rev 01)
• Display server if rendering to display(X or wayland):

Issue behavior

Describe the current behavior

gst-launch-1.0 -vf filesrc location=input.vc1 ! 'video/x-wmv,profile=(string)advanced',width=176,height=144,framerate=14/1 ! msdkvc1dec ! videoconvert chroma-mode=none dither=0 ! video/x-raw,format=I420 ! checksumsink2 qos=false file-checksum=false frame-checksum=false plane-checksum=false eos-after=15 dump-output=true dump-location=output.yuv

Describe the expected behavior

pass

Debug information

reproduce commit: libva : ef1df02f3ad4 gmmlib : 4552654ed52f intel-media-driver : c7132d869bed onevpl : 250d806b5867 onevpl-gpu : ab051c80094c cartwheel-gstreamer:36e60d64d1a6

[wenqingx]

[wenqingx]

debug from @mengker33

[mengker33]

Actually, this hang issue occurred on all platforms which support vc1 decode, i.e. ICL, CFL, ADL, etc.

[Jexu]

Could you check if it is same issue mentioned in https://github.com/oneapi-src/oneVPL-intel-gpu/issues/230?

[mengker33]

Could you check if it is same issue mentioned in #230?

It might not be the same issue. The hang mentioned in #230 happened when we pass MFX_IOPATTERN_OUT_SYSTEM_MEMORY type output mfxFrameSurface1, and I didn't see the same gdb debug log as this one shown in the pic above. The hang occurred in this ticket occurred using our cartwheel-gstreamer repo, in which we pass MFX_IOPATTERN_OUT_VIDEO_MEMORY type for output mfxFrameSurface1.

[mengker33]

@Jexu, can anyone help with this issue? Thanks

[Jexu]

@mengker33 Yeah, the issue is on track and i will let you know if having any finding.

[wenqingx]

Are there any updates? thanks

[wenqingx]

skip test decode vc1, so close this ticket.