Re: [PATCH v2 01/12] dt-bindings: media: qcom,venus: Add context bank subnodes to common schema

[Bryan O'Donoghue <bryan.odonoghue@linaro.org>]

On 06/08/2026 10:40, Bryan O'Donoghue wrote:
> On 06/08/2026 08:57, Vikash Garodia wrote:
>> On 8/6/2026 1:24 PM, Bryan O'Donoghue wrote:
>>> On 06/08/2026 06:21, Vikash Garodia wrote:
>>>> On 8/1/2026 3:36 PM, Bryan O'Donoghue wrote:
>>>>> On 31/07/2026 19:22, Vikash Garodia wrote:
>>>>>> Fixes: 41661853ae8e ("arm64: dts: qcom: sm8550: add iris DT node")
>>>>>> Cc:stable@vger.kernel.org
>>>>>> Tested-by: Daniel J Blueman<daniel@quora.org>
>>>>>> Signed-off-by: Vikash Garodia<vikash.garodia@oss.qualcomm.com>
>>>>> What about venus ?
>>>> Plan is to get the reviews for IRIS3 family SOCs, once we have a go
>>>> ahead from binding/DT maintainers, i can either bring in few series
>>>> for remaining venus and iris SOC family OR do all of them in one
>>>> single series. My preference would be to add in multiple series.
>>> I still don't see why we can't just set
>>>
>>> minItems: 1 for the memory nodes - implement a fix for stable and then
>>> iterate around this new way at leisure.
>> you can follow the guidance from binding maintainer [1] and continue to
>> discuss in the same thread if you have any question.
>>
>> [1]
>> https://lore.kernel.org/linux-media/2b0ea2a2-c3ef-4147-801b- 
>> af41733fba7b@kernel.org/
> 
> LoL seriously - system resetting bug.
> 
> How long exactly is this going to take - do you have buy in for sub- 
> nodes and when will that get merged.
> 
> For reference I do not care what happens my concern is an unspecified 
> window where stuff remains broken and we take our leisurely time about 
> doing something about it.
> 
> If we can't target something for -stable then this driver and venus 
> should be marked as BROKEN.
> 
> ---
> bod

I'd just like to be clear.

As far as I understand it a bug that can reset the system has been known 
about for months on the qcom side with little action to remediate in 
-stable.

That's entirely unacceptable. Litigating DT bindings is _not_ addressing 
the fact of life, that a system resetting bug has been sat on for 
months, unaddressed.

Daniel proposed a fix that will stop that reset which is being rejected? 
Arguments are being made to not mark this stuff as broken. And the 
debate about how a DT change should land is still ongoing in this thread.

I tend to think - absent an obvious an immediately applicable fix these 
two drivers must be marked BROKEN, the alternative is to apply Daniel's 
fix and then live with the results for the bindings.

Since the DT schema change is still under debate - what alternative is 
there ?

I'm not far off sending that patch now myself because I'm not seeing 
urgency on fixing this and TBH that's pissing me off.

As a user of a system, I expect resonsible and honest behaviour wrt to 
bugs, not architecture astronauty while my system can still reset 
despite perfectly functional fixes being available.

Great you want to do sub-nodes in DT. Not great, that's not agreed yet, 
even more not great - the venus stuff isn't even on the list.

Please indicate what should be done for -stable this week else both of 
these drivers need to be marked BROKEN.

---
bod