Re: [PATCH 01/18] media: venus: hfi_venus: Set venus_sys_idle_indicator to false on V6

[Konrad Dybcio]

On 28.02.2023 16:38, Bryan O'Donoghue wrote:
> On 28/02/2023 15:37, Konrad Dybcio wrote:
>>
>>
>> On 28.02.2023 16:31, Bryan O'Donoghue wrote:
>>> On 28/02/2023 15:26, Bryan O'Donoghue wrote:
>>>> On 28/02/2023 15:24, Konrad Dybcio wrote:
>>>>> This call does not seem to have been cast on any kernel with support
>>>>> for VPU-1.0 or newer (and by extension, HFI6 and newer).
>>>>
>>>> We tested this on sm8250
>>>>
>>>> Restrict it
>>>>> to V4 only, as it seems to have been enabled by mistake and causes a
>>>>> hang & reboot to EDL on at least one occasion with SM6115 / AR50L
>>>>>
>>>>> Fixes: 7ed9e0b3393c ("media: venus: hfi, vdec: v6 Add IS_V6() to existing IS_V4() if locations")
>>>>> Signed-off-by: Konrad Dybcio<konrad.dybcio@xxxxxxxxxx>
>>>>
>>>> Right. This may indeed fix it for you on SM6115, could you test it on RB5 and verify the above statement ?
>>>>
>>>> ---
>>>> bod
>>>
>>> For example.
>>>
>>> Doesn't your later patch take account of VPU h/w version ? IRIS_1, IRIS_2 etc.
>>>
>>> When we added for V6 here, we meant for current tested V6 hardware at that point - at least sm8250.
>>>
>>> Can you not differentiate sm6115 based on VPU hardware identifier ? We want to retain this logic for 8250 and then assuming your patch is correct, not do this for sm6115.
>> As far as my only source of information (msm-4.19 techpack) goes, this is
>> unnecessary/incorrect on 8250 as well. I doubt downstream would ship Venus
>> with no/broken low-power modes..
> 
> Can you test it and make sure ?
As I mentioned in the cover letter, 8250 still seems to work with this
patchset. I have no idea how one would go about validating the
functionality enabled through this call.

Konrad
> 
> ---
> bod
>