| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <3b3d36f1-344f-4cf8-a4f2-98a93052141e@oss.qualcomm.com>
Date: Wed, 10 Sep 2025 16:26:59 -0700
From: Deepa Guthyappa Madivalara <deepa.madivalara@....qualcomm.com>
To: Nicolas Dufresne <nicolas@...fresne.ca>,
Mauro Carvalho Chehab <mchehab@...nel.org>,
Vikash Garodia <quic_vgarodia@...cinc.com>,
Dikshita Agarwal <quic_dikshita@...cinc.com>,
Abhinav Kumar <abhinav.kumar@...ux.dev>,
Bryan O'Donoghue <bryan.odonoghue@...aro.org>
Cc: linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH RFC 0/5] Enable support for AV1 stateful decoder
On 9/3/2025 6:07 AM, Nicolas Dufresne wrote:
> Le mardi 02 septembre 2025 à 16:00 -0700, DEEPA GUTHYAPPA MADIVALARA a écrit :
>> Hi all,
>>
>> This patch series adds initial support for the AV1 stateful
>> decoder codecs in iris decoder. Also it adds support for AV1
>> stateful decoder in V4l2. The objective of this work is to
>> extend the Iris decoder's capabilities to handle AV1 format
>> codec streams, including necessary format handling and buffer
>> management. I'm sharing this series as an RFC because conformance
>> testing and gstreamer testing are still in progress. While initial
>> functional tests show positive results, I would appreciate early
>> feedback on the design, implementation, and fixes before moving to
>> a formal submission. I plan to submit a formal patch series after
>> completing all the compliance checks. Meanwhile, any feedback or
>> suggestion to improve this work are very welcome and will be of
>> great help.
>>
>> Gstreamer testing:
>> Gstreamer MR for enabling AV1 stateful decoder:
>> https://gitlab.freedesktop.org/dmadival/gstreamer/-/merge_requests/1
Hi Nicholas,
Thank you so much for a quick response.
Can you please provide feedback for the AV1 stateful decoder plugin - GST MR ?
Thanks to Nicolas Dufresne for proving the MR
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9469
However, Gst testing with parsebin is not resolving to av1parser as
below:
Ex: With the following command parsebin is unable to resolve to
av1parser.
GST_DEBUG=*:2,parsebin:6 gst-launch-1.0 --no-fault
filesrc
location=/media/sd/fluster/fluster/resources/AV1-ARGON-PROFILE0-CORE-ANNEX-B/
argon_coveragetool_av1_base_and_extended_profiles_v2.1/profile0_core/streams/test10220.obu
! parsebin ! v4l2av1dec ! video/x-raw ! videoconvert dither=none !
video/x-raw,format=I420
! filesink location=gst_decoder_output.yuv
> Be aware that Argon integration with fluster/GStreamer still needs some work,
> see issue tracker:
>
> https://github.com/fluendo/fluster/issues/222
>
> It is fine to proceed with testing the other (IVF or MKV based) test suites.
> Feel free to report all wrongly identified files onto that issue, this will be
> addressed in GStreamer (and eventually in FFMPEG too if need be).
Thanks for the above information. I will continue to work on ivf and mkv
based test-suites.
Is it okay to post the AV1 patch series even though Argon test-suite has
failures?
>> 0:00:00.051674896 400 0xffff8c000b90 DEBUG
>> parsebin gstparsebin.c:2439:type_found:<parsebin0> typefind found caps
>> video/x-h263, variant=(string)itu
>>
>> The same test with the av1parse command parses correctly:
>> GST_DEBUG=*:2,av1parse:6 gst-launch-1.0 --no-fault
>> filesrc
>> location=/media/sd/fluster/fluster/resources/AV1-ARGON-PROFILE0-CORE-ANNEX-B/
>> argon_coveragetool_av1_base_and_extended_profiles_v2.1/profile0_core/streams/test10220.obu
>> ! av1parse ! v4l2av1dec ! video/x-raw ! videoconvert dither=none !
>> video/x-raw,format=I420
>> ! filesink location=/tmp/gst_decoder_output.yuv
>>
>> Got EOS from element "pipeline0".
>> Execution ended after 0:00:01.599088176
>> Setting pipeline to NULL ...
>> 0:00:03.580831249 1075 0x3354f960 DEBUG
>> av1parse gstav1parse.c:435:gst_av1_parse_stop:<av1parse0> stop
>>
>> Fluster testing:
>> As fluster.py is using parsebin for gstreamer, seeing the same issue as
>> described above for the following testsuites.
>> AV1-ARGON-PROFILE0-CORE-ANNEX-B
>> AV1-ARGON-PROFILE0-NON-ANNEX-B
>> AV1-ARGON-PROFILE0-NON-ANNEX-B
>>
>> Test suite: AV1-TEST-VECTORS
>> The result of fluster test on SM8550:
>> 134/242 testcases passed while testing AV1-TEST-VECTORS with
>> GStreamer-AV1-V4L2-Gst1.0
>> unsupported content, bit depth: a000a (66 tests)
>> Iris hardware decoder supports only 8bit NV12
>> av1-1-b10-00-quantizer-*
> That is interesting, I believe there was no profile without 10bit, which would
> mean this is non-compliant hardware. Not a blocker, just a remark, and me being
> surprised 10bit isn't supported on modern codec in 2025.
Sorry, hardware does support 10bit. As we continue to add more features,
we plan to
include support for additional color formats as well.
>> Unsupported resolution (36 tests).
>> Iris hardware decoder supports min resolution of 96x96
>> av1-1-b8-01-size-*
>>
>> Unsupported colorformat (1 test)
>> av1-1-b8-24-monochrome
> Fair enough.
>
>> Crc mismatch: debug in progress (5tests)
>> av1-1-b8-03-sizeup
>> av1-1-b8-03-sizedown
> These should be basic dynamic resolution changes cases, let me know what you
> find.
I have a corruption after second sequence change, which propagates further
leading to hardware hang. I am working on fixing this.
>> av1-1-b8-16-intra_only-intrabc-extreme-dv
> No idea about this one.
This is passing actually. It is just an all intra bitstream.
>> av1-1-b8-22-svc-L2T1
>> av1-1-b8-22-svc-L2T2
> These two are spatial SVC. That means that some decode only frames are going to
> be smaller in dimensions. Only the last frame of a TU is displayed. Both do have
> the optional sequence header announcing the maximum resolution though. On
> stateless side, these usually works by simply decoding the hidden frames into
> the much larger buffers. Dealing with firmware can be harder, since firmware may
> silently drop the decode only frames, which leads to wrong timestamp matching,
> which could have side effect in frameworks.
>
> On stateful decoding, SVC have never been done, some documentation will likely
> be needed. Strictly unsupported is the case the sequence header is not there,
> which results in resolution change on non-keyframe. We now have the ability to
> allocate frames of different dimension at run-time, and free old unused frames,
> but in stateful, decoders streamoff the capture queue, which currently results
> in lost of references.
>
> Let us know your progress, inter-frame resolution change is not a strict
> requirement, it is quite rare in real life.
>
> looking forward, there is nothing particularly alarming in this report. Thanks
> for the transparency.
>
> Nicolas
I did not have a chance into debug this yet. But these streams do have a
resolution change
on non-keyframe.
>> Testsuite: CHROMIUM-8bit-AV1-TEST-VECTORS
>> 12/13 testcases passed while testing CHROMIUM-8bit-AV1-TEST-VECTORS with
>> GStreamer-AV1-V4L2-Gst1.0
>> Crc mismatch: debug in progress
>> av1-1-b8-03-sizeup.ivf
Same corruption after second sequence change, which propagates further
leading to hardware hang.
>> Unsupported test suites:
>> Iris Hardware Decoder supports only
>> PROFILE0/V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN
>> and 8 bit, 420 only
>> AV1-ARGON-PROFILE1-CORE-ANNEX-B
>> AV1-ARGON-PROFILE1-NON-ANNEX-B
>> AV1-ARGON-PROFILE1-STRESS-ANNEX-B
>> AV1-ARGON-PROFILE2-CORE-ANNEX-B
>> AV1-ARGON-PROFILE2-NON-ANNEX-B
>> AV1-ARGON-PROFILE2-STRESS-ANNEX-B
>> CHROMIUM-10bit-AV1-TEST-VECTORS
>>
>> Compliance test for iris_driver device /dev/video0:
>>
>> Driver Info:
>> Driver name : iris_driver
>> Card type : Iris Decoder
>> Bus info : platform:aa00000.video-codec
>> Driver version : 6.16.0
>> Capabilities : 0x84204000
>> Video Memory-to-Memory Multiplanar
>> Streaming
>> Extended Pix Format
>> Device Capabilities
>> Device Caps : 0x04204000
>> Video Memory-to-Memory Multiplanar
>> Streaming
>> Extended Pix Format
>> Detected Stateful Decoder
>>
>> Required ioctls:
>> test VIDIOC_QUERYCAP: OK
>> test invalid ioctls: OK
>>
>> Allow for multiple opens:
>> test second /dev/video0 open: OK
>> test VIDIOC_QUERYCAP: OK
>> test VIDIOC_G/S_PRIORITY: OK
>> test for unlimited opens: OK
>>
>> Debug ioctls:
>> test VIDIOC_DBG_G/S_REGISTER: OK (Not Supported)
>> test VIDIOC_LOG_STATUS: OK (Not Supported)
>>
>> Input ioctls:
>> test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
>> test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
>> test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
>> test VIDIOC_ENUMAUDIO: OK (Not Supported)
>> test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
>> test VIDIOC_G/S_AUDIO: OK (Not Supported)
>> Inputs: 0 Audio Inputs: 0 Tuners: 0
>>
>> Output ioctls:
>> test VIDIOC_G/S_MODULATOR: OK (Not Supported)
>> test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
>> test VIDIOC_ENUMAUDOUT: OK (Not Supported)
>> test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
>> test VIDIOC_G/S_AUDOUT: OK (Not Supported)
>> Outputs: 0 Audio Outputs: 0 Modulators: 0
>>
>> Input/Output configuration ioctls:
>> test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
>> test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
>> test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
>> test VIDIOC_G/S_EDID: OK (Not Supported)
>>
>> Control ioctls:
>> test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK
>> test VIDIOC_QUERYCTRL: OK
>> test VIDIOC_G/S_CTRL: OK
>> test VIDIOC_G/S/TRY_EXT_CTRLS: OK
>> test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK
>> test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
>> Standard Controls: 12 Private Controls: 0
>>
>> Format ioctls:
>> test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK
>> test VIDIOC_G/S_PARM: OK (Not Supported)
>> test VIDIOC_G_FBUF: OK (Not Supported)
>> test VIDIOC_G_FMT: OK
>> test VIDIOC_TRY_FMT: OK
>> test VIDIOC_S_FMT: OK
>> test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
>> test Cropping: OK
>> test Composing: OK
>> test Scaling: OK (Not Supported)
>>
>> Codec ioctls:
>> test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
>> test VIDIOC_G_ENC_INDEX: OK (Not Supported)
>> test VIDIOC_(TRY_)DECODER_CMD: OK
>>
>> Buffer ioctls:
>> test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
>> test CREATE_BUFS maximum buffers: OK
>> test VIDIOC_REMOVE_BUFS: OK
>> test VIDIOC_EXPBUF: OK
>> test Requests: OK (Not Supported)
>> [65391.311675] qcom-iris aa00000.video-codec: invalid plane
>> [65395.340586] qcom-iris aa00000.video-codec: invalid plane
>> test blocking wait: OK
>>
>> Total for iris_driver device /dev/video0: 48, Succeeded: 48, Failed: 0,
>> Warnings: 0
>>
>> V4l2-ctl Test verified for 2 streams as well.
>>
>> Thanks,
>> Deepa
>>
>> Signed-off-by: Deepa Guthyappa Madivalara <deepa.madivalara@....qualcomm.com>
>> ---
>> DEEPA GUTHYAPPA MADIVALARA (5):
>> media: uapi: videodev2: Add support for AV1 stateful decoder
>> media: v4l2: Add description for V4L2_PIX_FMT_AV1 in v4l_fill_fmtdesc()
>> media: iris: Add support for AV1 format in iris decoder
>> media: iris: Add internal buffer calculation for AV1 decoder
>> media: iris: Define AV1-specific platform capabilities and properties
>>
>> drivers/media/platform/qcom/iris/iris_buffer.h | 2 +
>> drivers/media/platform/qcom/iris/iris_ctrls.c | 8 +
>> drivers/media/platform/qcom/iris/iris_hfi_common.h | 3 +
>> .../platform/qcom/iris/iris_hfi_gen2_command.c | 109 ++++++++-
>> .../platform/qcom/iris/iris_hfi_gen2_defines.h | 10 +
>> .../platform/qcom/iris/iris_hfi_gen2_response.c | 22 ++
>> drivers/media/platform/qcom/iris/iris_instance.h | 1 +
>> .../platform/qcom/iris/iris_platform_common.h | 15 ++
>> .../media/platform/qcom/iris/iris_platform_gen2.c | 156 ++++++++++++-
>> .../platform/qcom/iris/iris_platform_sm8250.c | 17 ++
>> drivers/media/platform/qcom/iris/iris_vdec.c | 25 +-
>> drivers/media/platform/qcom/iris/iris_vidc.c | 1 +
>> drivers/media/platform/qcom/iris/iris_vpu_buffer.c | 255 ++++++++++++++++++++-
>> drivers/media/platform/qcom/iris/iris_vpu_buffer.h | 105 +++++++++
>> drivers/media/v4l2-core/v4l2-ioctl.c | 1 +
>> include/uapi/linux/videodev2.h | 1 +
>> 16 files changed, 703 insertions(+), 28 deletions(-)
>> ---
>> base-commit: 88a6b4187eacb700a678296afb0c610eb3781e2f
>> change-id: 20250821-rfc_split-c3ff834bb2c9
>> prerequisite-change-id: 20250704-iris-video-encoder-b193350b487a:v3
>> prerequisite-patch-id: 8a566690da276da34430c10dbc2fe64c1d623a9c
>> prerequisite-patch-id: 1430a33603b425d0b142aab98befcda771fb885e
>> prerequisite-patch-id: 32024cd49d2445ff396e31f40739b32597be59a4
>> prerequisite-patch-id: 65b569952650647174e8221dc7adde9b000a7ae3
>> prerequisite-patch-id: da128980fab8538bf668f19016c5121fb03759c2
>> prerequisite-patch-id: 079823dffbe8b89990797bf7f7640b754382d8ce
>> prerequisite-patch-id: 6ce10e03d7b3b96b2391e26cda703b650bde7cd0
>> prerequisite-patch-id: b5950670ac5068a0c5b26651ebd433f7d3bbe6ca
>> prerequisite-patch-id: 4de7a934f6bdfe28c84e461f70495925aa98365e
>> prerequisite-patch-id: 07682a6d2530b5796122bf8763f94b5bc92949ec
>> prerequisite-patch-id: 72b7eba20f1a222908d41323f28be3ba84106759
>> prerequisite-patch-id: fd9e2e1b157112c39c69486799493ee99e6033a7
>> prerequisite-patch-id: ae0ad8a04a04dd3434a092d4c2bb3f493417c6e1
>> prerequisite-patch-id: 52631eec348735d1dc5f5804b573e3cf942550a0
>> prerequisite-patch-id: 4109c59edb1b757162db46297914c8f7c14408dc
>> prerequisite-patch-id: fc0b713eb4822047e8172d11fd4cd5a097ef23a5
>> prerequisite-patch-id: 20ac8e7307f1f852b2a43268b2474178fbc0b94c
>> prerequisite-patch-id: e8419d716573beb64ad89968f0074d6bddfa86d3
>> prerequisite-patch-id: bdc72f5876ceb2e981d594c86a45cb21a6264af3
>> prerequisite-patch-id: fc5d26d01cab94d229a00eab819ae80196f3f5d5
>> prerequisite-patch-id: 32a9fe1371fffc9abd9a862b2814050a144d1968
>> prerequisite-patch-id: 25184583b5de886f78ee0444a4a59d5f3c271ce5
>> prerequisite-patch-id: df3376b9de27b23ae81a4c7b7a8fe4b429c32423
>> prerequisite-patch-id: e68fbf7c82567d2e9f3fdd0fdf2e2911329d5ccd
>> prerequisite-patch-id: 580fa40de01a81a8685e56420f562d299bfc60fa
>> prerequisite-patch-id: 89548da6690681854ee1de992a491bed73202b83
>>
>> Best regards,
Powered by blists - more mailing lists