08:35 <tokyovigilante> have tested what I think is the DCDC3 line on the RG35XX+ before and after runnning the u-boot SPL, and getting -0.01v with power on but before running SPL, and a very reproducible 1.109v after, so it looks like it is being set ok

08:35 <tokyovigilante> However the mystery of why it is then not loading the TF module deepens

08:37 <apritzel> tokyovigilante: ah, good, thanks for checking that

08:38 <apritzel> I am on the run, but can give you some instructions later how to move BSS to SRAM, which should make the SPL return to FEL

08:38 <apritzel> then you can use "sunxi-fel readl" and friends to probe memory and dump registers

08:39 <tokyovigilante> sounds good, thanks

08:39 <tokyovigilante> at least we have a more-or-less working AXP717 driver

08:46 <Jookia> okay, i've decided before i finish my audio patches i'm going to try tying in another DIN channel so i can read two codecs at once :)

08:46 <Jookia> It will be interesting to see how far I can push the chip with handling 16 input channels. Maybe I can get to 24

09:04 <apritzel> tokyovigilante: speaking of AXP717: do you or macromorgan have anything for Linux already? I have the regulator and minimum MFD bits, which should make it boot, but cannot test it (yet)

09:05 <apritzel> the AXP717 does much more though: GPIOs/IRQs, power key, battery charging, power supply, plus the entirely new USB-C part, but I think we can add those later

09:17 <tokyovigilante> not at all sorry, I think macromorgan has more experience with the other Anbernic devices so hoping he can help more with the linux part. Yeah the charging aspect is obv. quite complex

09:19 <apritzel> well, it's technically complex, but mostly done in hardware, and I think since we support battery charging on other AXP chips already, that's mostly a copy&paste exercise, with some glance into the datasheet

09:19 <apritzel> the USB-C part is completely new to AXPs, so that's more work

09:20 <apritzel> but from what I saw it only covers role switch and analogue audio, plus notifications. I didn't spot anything PD related, really

09:21 <apritzel> so I guess it's 5V only, or handled entirely internally, so shouldn't interfere with the whole charging aspect, really

09:21 <apritzel> at least that's my hope from looking at the three USB type-C related registers in the datasheet

09:23 <apritzel> the regulator part is absolutely crucial for even getting booted to a prompt, so we will need that ASAP

09:23 <apritzel> I will clean up the patches and put them on my github later

09:32 <apritzel> tokyovigilante: after "make ..._defconfig", try: "scripts/config --set-val CONFIG_SPL_BSS_START_ADDR 0x30000 --set-val CONFIG_SPL_BSS_MAX_SIZE 0x8000"

09:34 <apritzel> That should move the BSS into SRAM. If you then do "sunxi-fel spl spl/sunxi-spl.bin", it should ideally return from FEL after the DRAM init, so you can issue further FEL commands

09:37 <apritzel> at least that seems to work here on an OrangePi Zero3

09:44 <apritzel> tokyovigilante: if you see sunxi-fel return, you should try to write to DRAM and read that back, for instance with sunxi-fel's write and read commands

09:45 <apritzel> it that hangs, it's definitely a DRAM issue. Other thing I can imagine is that some bits are either stuck or redirected, which you might be able to spot in the read-back data

19:36 <libv> so an employee of incircuit gmbh removed my rant about their crappy treatment of the open source work they depend on

19:45 <gamiee> libv : huh? what happened

19:48 <libv> https://linux-sunxi.org/Special:RecentChanges

19:51 <gamiee> loool after that long time

20:00 <libv> locking down that page.

20:12 <libv> it took them 8 years to add upstream support.

23:42 <apritzel> libv: did you trip over a wire after editing the wiki? ;-) Seems to be down ...