02:57 <macromorgan> diederik: I think if you rig the regulators to the GPU power domain like they did with the rk3588 it fixes the DRM issues, at least it did for me. I was going to upstream it once I knew for sure that was the fix.

07:43 <diederik> macromorgan: My first test will be to try the regulator-boot/always-on 'thing', then set the SCMI clocks on all CPUs. I assume that will already bring back the display output. If not, then the thing you're referring to will be the next test

07:45 <diederik> And now that I (finally) have a rk3588 device, I'll likely do a `git bisect` to see what change caused it

14:03 <detlevc> warpme: yes, that's wip... gstreamer has support here: https://gitlab.freedesktop.org/detlev.c/gstreamer/-/tree/add-st-ref-pic-sets-v4l2codecs (for the verison before the commit a66dfa7f09c4eba9dabe5cf380e238325830f1f7 you mentioned)

14:04 <warpme> detlevc thx! Have you plans also for ffmpeg?

14:18 <detlevc> warpme not really familiar with ffmpeg codebase, but let's have a stable API first

14:20 <warpme> ah ok. Is your work: an extension to current uapi or rather adding missing functionality in current uapi?

18:59 <CounterPillow> mmind00: hmmm, I think this patch has slipped between the cracks again, don't see it in next-20250424 :( https://lore.kernel.org/linux-arm-kernel/2364081.ElGaqSPkdT@diego/

21:59 <mmind00> CounterPillow: I don't even know where the otp stuff actually is right now ... from what I gathered there was a mixup between the nvmem-tree and greg's tree and the driver side actually never made it into -rc1

21:59 <CounterPillow> :D I only remembered because I sent out tsadc again

22:00 <mmind00> CounterPillow: last checked somewhere before easter, so not sure if things worked out for 6.15 afterall

22:00 <mmind00> CounterPillow: ah, it made it into Torvald's tree afterall now

22:09 <mmind00> CounterPillow: done now

22:16 <CounterPillow> Thanks!