07:35 <naoki> hmm...... DT for sound card is too complicated for me... ;)

07:35 <dsimic> it can get tricky :)

07:37 <naoki> audio handling on userland is difficult too. I can understand only alsa :D

07:38 <dsimic> yup, the audio subsystem is quite complex

15:26 <qschulz> wens: regarding the discussion we had on pinctrl+gpio with having .strict = True + pinctrl-0 + *-gpios in DT

15:27 <qschulz> wens: we absolutely need what you were suggesting in the pinctrl subsystem (I assume) to allow pinctrl-0 + *-gpios in DT (if the pinctrl-0 is muxing in GPIO function of course!) because we need to be able to provide the pinconf there

15:28 <qschulz> (PU/PD/HIGH-Z/DS, etc...)

18:10 <helene> hello hello, i've been trying for a few days to make Linux 6.11 & U-Boot 2024.10-rc5 run on my RK3588-based CM3588 (since the DTs for that board were merged in for that kernel release! thank you for the work, Kwiboo!), but i've been encountering issues on getting bootup, which probably are caused by me misconfiguring something in my kconfig but i can't quite figure out what

18:12 <CounterPillow> Show serial output of bootup, and your kconfig, in a paste

18:12 <helene> i've been hit (again) with the "feb20000.spi: deferred probe pending: (reason unknown)"; which last time was caused by me missing the fact that SCMI clock drivers were required, as the registers for PMIC (which held regulators); but this time i'm not sure what's up

18:12 <helene> sure thing

18:15 <helene> https://brpaste.xyz/SPcWAw/raw here's the dmesg and bootup output (from serial)

18:17 <CounterPillow> >Linux version 6.10.9

18:17 <CounterPillow> Well for starters that's not 6.11

18:17 <helene> oops, yeah sorry

18:17 <helene> it did use to work on 6.10 mainline, however; but yes, that's silly of me since drivers could've changed

18:18 <CounterPillow> But this does look like no pmic or no spi for the pmic to work

18:18 <helene> it did use to work on 6.9, still (just with the Armbian DTBs, which don't work anymore either now)

18:19 <CounterPillow> Do keep in mind that things may be turned off when moving between kernel versions and doing make oldconfig, as sometimes kconfig reorganising isn't thought out in the best ways

18:20 <helene> yeah, so i made sure to check /proc/config.gz; but maybe i missed something

18:23 <CounterPillow> So are you gonna show the config?

18:26 <helene> https://brpaste.xyz/e9U9xA/raw here's the config, sorry for the delay

18:30 <robmur01> "CONFIG_REGULATOR_FIXED_VOLTAGE=m" - gonna guess that module's not loading

18:32 <helene> oops... that seems very obvious in hindsight now, i've probably missed quite a few (i didn't know about that driver)

18:32 <helene> thank you!

18:33 <helene> i'm surprised that the kconfig let me do that, though

18:33 <robmur01> thankfully the missing pcie supplies were the tell in this case, having no dependency chain

18:34 <CounterPillow> The i2c controller driver is a module too, don't know if it's included in the initramfs though but i2c might be hooked up to PMIC stuff as well (though I think with all rk3588 stuff it's spi)

18:35 <helene> i tried figuring out myself from reading the DTs and looking for which driver matched which device, and i couldn't figure it out; so thank you! i lack a lot of the linux "internals" knowledge wrt/ drivers myself, i couldn't have possibly figured it out without bisecting from a full config that may possibly have worked

18:36 <helene> CounterPillow: devicetree seems to say it's SPI (if i'm reading it right), but if I'm encountering any issues I'll enable it entirely too, thank you!

18:40 <robmur01> scripts/dtc/dt_to_config tries to give a reasonable starting point

18:45 <helene> oh, i didn't know that was a thing! thank you, i'll write that down somewhere for future endeavours!

22:13 <wens> qschulz: AFAIK you can specify pin config options without claiming the pins (just don't specify the mux function)