09:53 <dlan> apritzel: do you have any suggestion which branch I could work on in uboot? to prepare an EMAC0 patch for A523/T527.. as I think the basic A523 patch has not been accepted into mainline?

10:11 <apritzel> dlan: you wouldn't need any patch, would you? Beauty of DT fallbacks: it just works the moment you give it the right DT

10:12 <apritzel> so just update the DT, from Linux' sunxi/next branch

10:13 <apritzel> once sunxi/next reaches Linus' tree, it will be synced into the git rebasing repo, and I can cherry-pick it from there, into U-Boot - or we could wait until it gets pulled automatically

11:16 <dlan> apritzel: oh, that's sweet! so, probably we could switch to OF_UPSTREAM in future..

11:22 <spawacz> Is it possible to overclock D1 just a tiny bit via the dtb?

11:50 <apritzel> dlan: that was the plan from the beginning: it would be OF_UPSTREAM only. I just need to wait for the DTs to actually appear in upstream, which would probably be in like 3-4 weeks from now

11:53 <apritzel> spawacz: have you tried just changing the frequency in opp_table_cpu, in arch/riscv/boot/dts/allwinner/sun20i-d1s.dtsi?

14:30 <luka177> Hello, ive made a custom board arounf allwinner t113-s4 with ADV7513 rgb to hdmi bridge, i booted linux on it with no issues, and bridge do work, BUT there is an issue, after tcon driver is probbed one of signal (specifically d21 on PD15) is stuck high, resulting in a slight red tint. Anything obvious i might be missing? Pin is clearly used by tcon ```pin 111 (PD15): device 5461000.lcd-controller function lcd0 group PD15```

15:22 <apritzel> luka177: maybe it's the internal pull-up accidentally configured for that pin?

15:24 <luka177> apritzel: ```

15:24 <luka177> ```lcd0_rgb888_pins: lcd0-rgb888 { pins = "PB2", "PB3", "PD0", "PD1", "PD2", "PD3", "PD4", "PD5", "PB4", "PB5", "PD6", "PD7", "PD8", "PD9", "PD10", "PD11", "PB6", "PB7", "PD12", "PD13", "PD14", "PD15", "PD16", "PD17", "PD21", "PD20", "PD18", "PD19"; function = "lcd0"; bias-pull-down; drive-strength = <10>; };``` ive tried enabling even pull down, with no effect

15:24 <apritzel> Linux wouldn't change that setting, unless explicitly told so. You can force that off with "bias-disable;" in the pinctrl DT node

15:27 <luka177> apritzel: Just tried bias-disable, no result either, also it takes +/- what is specified in drive-strength to sink it (tested with 40ma and 10ma) so i'd assume its not a pupd

15:29 <apritzel> luka177: that's all with non-upstream drivers, right? As on a quick glance I don't see any display support in mainline ...

15:32 <luka177> apritzel: That is with upstream kernel with no edits beside DT

15:33 <spawacz> apritzel: it worked \o/

15:33 <apritzel> ah, scratch that, was looking at the wrong DT

15:34 <spawacz> I set it to 1.2 GHz. How far can I push it further?

15:37 <luka177> also here is my dt: https://bpa.st/BTBA

15:37 <apritzel> luka177: have you tried it with the existing rgb666 pins?

15:41 <luka177> apritzel: i just did, its the same

15:43 <luka177> PD15 is still at 3v3 resulting in red tint

16:17 <apritzel> I dimly remember people reporting that sometimes tints are a software/driver issue? Not sure how well tested the display is on this SoC

16:35 <luka177> apritzel: Wekk it is very likely to be software issue but i am not sure what can be the culprit

16:39 <apritzel> just to make sure: I meant a display pipeline issue, not a pinctrl issue. But I have no clue about the internals, I guess jernej might know more here

18:13 <spawacz> I'd like o be able to read temperatures, how can I find out what kernel module is missing to populate /sys/class/thermal on D1? Debian's stock kernel has very little options enabled on riscv.

18:23 <spawacz> It is CONFIG_SUN8I_THERMAL, correct?

21:54 <luka177> Ok so i did try to set invert in SUN4I_TCON0_IO_POL_REG and now its stuck at 0v so i doubt its a hardware issue, but a software iss that leads specifically to D21 not working sounds strange

22:03 <luka177> I followed https://groups.google.com/g/linux-sunxi/c/HxDBpY5HbbQ/m/mX2O2OYlCwAJ?pli=1 and indeed disabling mixer1 helped, but question is, what is the underlying issue

22:23 <apritzel> spawacz: in general you do: "git grep allwinner,sun20i-d1-ths drivers/" to find the driver, then search for the driver filename in the Makefile in that same directory to find the Kconfig symbol

22:24 <apritzel> spawacz: interestingly there is the binding and driver support in mainline, but for some reason the DT node is missing

23:49 <apritzel> ah, I see, there were actually two patches posted end of February, but there were comments, and the v2 that was advertised was never sent

23:49 <apritzel> https://lore.kernel.org/linux-sunxi/20250218020629.1476126-1-alex@studer.dev/T/#u