Re: [PATCH v3 0/3] Add support for the Iris codec on Milos

[Alexander Koskovich]

On Monday, July 6th, 2026 at 10:44 AM, Dmitry Baryshkov <dmitry.baryshkov@xxxxxxxxxxxxxxxx> wrote:

> On Sun, Jul 05, 2026 at 10:10:18AM +0000, Alexander Koskovich wrote:
> > This series adds the bindings, nodes and platform data for the Milos platform
> > for the Iris video codec, allowing Milos to use hardware‑accelerated video
> > encoding and decoding.
> >
> > Ran v4l2-compliance and some fluster tests, though a concerning amount of them
> > failed. Attaching v4l2-compliance output and the full fluster results below.
> >
> > nothing-asteroids:~$ v4l2-compliance -d /dev/video1 -s
> > v4l2-compliance 1.32.0, 64 bits, 64-bit time_t
> >
> >
> > nothing-asteroids:~/fluster$ ./fluster.py --no-emoji run -ts JVT-AVC_V1 -d GStreamer-H.264-V4L2
> 
> Could you please run the tests against the full test suite and submit
> the results to https://github.com/linux-msm/fluster-tests ? There is a
> script to run the tests.

Will do.

> 
> > ****************************************************************************************************
> > Running test suite JVT-AVC_V1 with decoder GStreamer-H.264-V4L2
> > Using 8 parallel job(s)
> 
> -j1 might also be nicer.

Yeah I did run -j1 for VP9, but I kept the other two at defaults since I was
trying to match the testing done here since it's another vpu2 gen2 platform,
and assumed there was a specific reason only VP9 was set to -j1:
https://lore.kernel.org/linux-arm-msm/20260529-kodiak-gen2-support-v4-v6-0-9a81bfa797d9@xxxxxxxxxxxxxxxx

Will make sure to use -j1 for all of the tests going forward (assuming the
script you mentioned does that).

> 
> > ****************************************************************************************************
> >
> 
> --
> With best wishes
> Dmitry
> 
>