00:15 <Lightsword> apritzel, yeah essentially, I just stuck it in a fat partition for now on the SD card

00:21 <apritzel> ah, that would have been my suggestion as well. I guess a FIT image is just not working great if you are short of memory, since U-Boot might end up copying things around, so you need more space

00:22 <apritzel> is the initrd your eventual rootfs, or do you just use it to mount a ubifs as a rootfs? So do you really need the initrd?

00:26 <Lightsword> apritzel, initrd is used as rootfs as well

00:27 <Lightsword> would mount a small ubifs config partition is plan

00:28 <Lightsword> keeping as much as possible in ramdisk helps with reliability IMO

00:36 <apritzel> sure, just asking because of your atomic update requirement. I wonder if you could eventually use a ubifs volume for holding kernel and initrd, avoiding the FIT?

00:45 <Lightsword> apritzel, I mean sure...that's possible, but that's kind of adding another layer and point of failure(ubifs) without a good reason IMO, and I would then have to do something funky like swap the entire ubifs volumes as well

00:45 <Lightsword> I like using FIT images since everything is nicely bundled together.

00:48 <Lightsword> apritzel, been trying to get this working with new kernel API, pwm was broken https://gist.github.com/jameshilliard/21f454cbe4a3abc629fe18a6e48419b4

00:49 <Lightsword> still not quite working though, not sure why

08:00 <Lightsword> any idea what might be causing this error? "i2c i2c-0: mv64xxx: I2C bus locked, block: 1, time_left: 0" seems it's preventing ethernet from getting set up

08:16 <Lightsword> dts: https://gist.github.com/jameshilliard/4e93931a53472aaaafeacc67fe275447 dmesg: https://gist.github.com/jameshilliard/eb46a257b45bce9ed97dcad1cfb1b05b kernel source: https://github.com/jameshilliard/linux-h616/commits/ac300/

08:17 <tokyovigilante> I've just realised the device tree enablement patch didn't get picked up along with the rest of the headphone detection code for the RG35XX for 6.15, should I resend just that patch, the whole series, or reply to myself?

16:59 <jernej> apritzel: what happened to your H616 PD patch? Do you need to fix anything?

17:06 <apritzel> jernej: Ulf had two comments, but should be easy to fix, just wasn't high priority due to it not working properly. With loki666's fixes it should now work, I think, so will look into sending a v2

17:06 <jernej> that would be nice, thanks!

17:07 <jernej> because I'm wondering, if panfrost patches stalled a bit due to missing PD driver

17:09 <apritzel> chicken/egg you mean? ;-)

17:10 <apritzel> there was still the question of how or whether the panfrost changes affect the other SoCs

17:15 <jernej> something like that :)

17:17 <jernej> I think PD handling in panfrost just worked because most of platforms have only one PD. That's handled already by common framework

18:59 <montjoie> hello, dwmac_sun8i do now work on 6.14.2, anyone saw the same ?

19:59 <montjoie> mmmh everything normal in fact

22:16 <Lightsword> smaeul, btw we can add params being passed to mkimage like this https://patchwork.ozlabs.org/project/uboot/patch/20250411011025.1434660-1-james.hilliard1@gmail.com/