On 19/02/2024 20:24, Bryan O'Donoghue wrote: > On 19/02/2024 5:44 p.m., Dmitry Baryshkov wrote: > >> On Mon, 19 Feb 2024 at 19:29, Konrad Dybcio wrote: >>> >>> On 19.02.2024 18:18, Marc Gonzalez wrote: >>> >>>> On our msm8998-based device, calling venus_sys_set_power_control() >>>> breaks playback. Since the vendor kernel never calls it, we assume >>>> it should not be called for this device/FW combo. >>> >>> FWIW, this is also broken on other SoCs.. 8280/8350 and 6115 >>> to name a couple. >> >> Then let's just disable it until it gets unbroken? > > Its functional on most of our upstream stuff though, why switch if off > unless necessary ? > > Maybe it should be an opt-in instead of an opt-out, TBH my own feeling > is its better to minimize the amount of work and opt as per the proposed > patch. > > Perhaps the qcom vidc team can give insights on 8280xp and 8350 when we > come to tackling new HFI6XX and later SoCs ... I was wondering if the chosen property name might cause issues later... Thinking "qcom,no-low-power" might be a bit too general? Perhaps would need to mention venus somewhere in the name, to limit this to the video decoder? Regards
