Re: media: iris: enable SM8350 and SC8280XP support

[Dmitry Baryshkov]

On Sat, 4 Apr 2026 at 03:00, Bryan O'Donoghue
<bryan.odonoghue@xxxxxxxxxx> wrote:
>
> On 03/04/2026 22:59, Dmitry Baryshkov wrote:
> > On Fri, Apr 03, 2026 at 04:41:39PM -0500, Steev Klimaszewski wrote:
> >> Hi Vikash,
> >>
> >>>> This is not just with totem, showtime shows the same issue.  If I let the video
> >>>> play, it plays just fine, however, if I attempt to skip forward, back, or even
> >>>> play*after* the video has played, then I see the smmu fault
> >>>>
> >>>>    arm-smmu 15000000.iommu: Unhandled context fault: fsr=0x402, iova=0xd51e3e00, fsynr=0x600002, cbfrsynra=0x2a00, cb=6
> >>> please log the info about which "iova=0xd51e3e00" buffer is faulting in
> >>> this case
> >> I will include the diff that I applied against my kernel, in case this isn't the
> >> correct information, but I get the following output in EL2:
> >
> > But these logs don't capture, which buffers are causing the IOMMU error.
>
> I'd be interested to see if we can run this test on venus - instead of iris.

At least when I tried testing it, venus wouldn't even boot the
firmware on SC8280XP, citing UC_REGION error:

2026-04-03T01:51:05.376407+03:00 rohan kernel: qcom-venus
aa00000.video-codec: non legacy binding
2026-04-03T01:51:05.437232+03:00 rohan kernel: qcom-venus
aa00000.video-codec: invalid setting for UC_REGION
2026-04-03T01:51:05.437838+03:00 rohan kernel: qcom-venus
aa00000.video-codec: failed to reset venus core
2026-04-03T01:51:05.437850+03:00 rohan kernel: qcom-venus
aa00000.video-codec: probe with driver qcom-venus failed with error
-22


>
> Konrad's patches are still floating around somewhere.

I think it was captured by Steev:

> This is much better than the previous venus patchset where accessing the
> hardware decoding would cause the machine to hard reset at least!

-- 
With best wishes
Dmitry