00:09 <apritzel> macromorgan: what are you after anyway? That error about not setting up the CPU frequency?

00:10 <apritzel> this seems to happen because Linux doesn't reset the PMIC to I2C mode, so the SPL fails to talk to the PMIC

00:13 <apritzel> so it's not U-Boot or TF-A at fault here, I think

00:22 <macromorgan> `regulator list` in U-Boot doesn't show anything unless I enable the rsb driver, which requires setting SUNXI_RSB_BASE (in cpu_sun50i_h6.h) as well as bypassing prcm_apb0_enable() with an ifdef in sun8i_rsb.c

00:23 <macromorgan> not that most users need to tinker with regulators in U-Boot, but if I want to do things like enable USB host or the 2nd mmc slot I do (plus it helps for debugging)

00:23 <macromorgan> well, actually neither of those need the axp regulators, but still

07:08 <ReallyDarkGray> Hello! I cannot boot FreeBSD 13.0-RELEASE on BPI M1+. It hangs after Kernel args: (null). Given that this OS has been known to boot on this board, I assume that the culprit is the (more recent) u-boot loader. I'm going to compile an older version of u-boot and see if I can boot. Am I on the right path?

12:17 <apritzel> macromorgan: for the regulators: you seem to be looking at the old U-Boot way. In U-Boot proper, the AXPs should work in a fully clean DT driven way, without any hacks

12:17 <apritzel> there are some symbols you need to enable for that, though, but then the xxx-supply properties should be handled

12:18 <apritzel> so far there are not too many users, but Ethernet is one of them

12:18 <apritzel> for USB there is a pending series, but there is some kind of bug in there

15:58 <macromorgan> okay... I'll defer on that for now and look at the pwms next, then circle back to some more PMIC stuff in Linux proper.

15:59 <macromorgan> I've dumped some vital data from the RG28XX which is another H700 based device here: https://gist.github.com/macromorgan/6bf2fcdc8093b88765c48343ad63a995

16:00 <macromorgan> I'm still using a different device, but eventually we can support that RG28XX as well. The question (and I have an open question to Anbernic on this topic) is if there's any way to runtime differentiate the device from U-Boot. That way we can probably just mainline the RG35XX-2024 device and use that bootloader for all of them.

16:03 <macromorgan> so far the differences are (relative to the most basic 35XX-2024): 28XX has a unique screen; 35XXH has wifi, ADC joysticks, and a 2nd USB port; and the 35XXPlus has wifi

17:27 <macromorgan> is there a datasheet on the pwm controller of the R329? I think (one of the pwms anyway) that it looks similar to the H616

20:24 <macromorgan> I found it. Looks like it's extremely similar but with slightly different registers (like I believe you said previously)

21:04 <apritzel> macromorgan: yes, but we figured that already, that's why we wanted to adapt the R329 driver

21:04 <macromorgan> right, I'm looking at that right now. Looks like it doesn't have clock gating like the 329 though

21:05 <macromorgan> otherwise looks like a copy-pasta from the sun20i but with different register values and no clock gating

21:05 <apritzel> are you looking at U-Boot or Linux?

21:05 <macromorgan> Linux

21:51 <tokyovigilante> jernej: sorry I'm basing the H616 audio codec driver on the out-of-tree OrangePi/BigTree one, but it's near-identical to the existing sun4i codec driver so just adding the H616 bits to that. There is also a driver for the audio hub (which seems to not be required for basic line-out and more complex routing for SPDIF and HDMI, as well as the DMICs for Karaoke mode ;)) and then an HDMI audio

21:51 <tokyovigilante> driver, but that code is pretty wild so will start with just the codec

22:31 <apritzel> macromorgan: regarding auto-detection: now that you have U-Boot running, can you try what I suggested: pick one button that isn't connected on a device, and try to read its input status with the pin pulled up and pulled down?

22:37 <apritzel> for instance try "gpio status pe8", on the -2024 device, then try changing bits 16 & 17 in register 0x300b0a4

22:37 <apritzel> and repeat that on the -Plus, to see if it differs