11:16 <jernej> smaeul: I'm looking into OrangePi 3 LTS support and I noticed that AXP driver doesn't work under Linux. I suspect this is due to overclocked RSB bus: "sunxi-rsb 7083000.rsb: RSB running at 4000000 Hz"

11:16 <jernej> even though RSB node has "clock-frequency = <3000000>;"

11:16 <jernej> any idea what's going on?

14:39 <smaeul> jernej: crust sets the r-apb2 parent to iosc, so your closest options are 2 MHz and 4 MHz

14:40 <jernej> smaeul: I did a bit of testing and even 1 MHz didn't work

14:40 <jernej> what worked though, is 100 kHz

14:41 <jernej> it seems that routing of the traces is bad, at least for high frequency signals

14:41 <jernej> since bsp kernel always uses i2c, so I guess this was never considered

14:41 <smaeul> yeah, I think Icenowy reported another H6 board that couldn't do normal RSB speeds.

14:42 <jernej> PMIC placement between LTS and non-LTS is rotated for 180 deg. and few components are moved

14:42 <jernej> in any case, I'll leave 100 kHz, it works that way

14:44 <smaeul> I don't see any problem with that. RSB at 100 kHz is still nicer than I2C in terms of CPU-side overhead

14:45 <jernej> I hoped that there wouldn't be need for different DTS file, but it's needed anyway, due to different ethernet PHY delays and different wifi

17:52 <junari> jernej: Have you made new changes for the H616 DRAM driver lately? Except h616-dram-imp branch

17:53 <junari> I will try to add support for LPDDR3 on the weekend

17:54 <jernej> junari: I did, but I didn't push the branch

17:54 <jernej> give me a moment

17:55 <junari> jernej: do not hurry, I will start tomorrow

17:55 <jernej> junari: done

17:55 <junari> jernej: nice, thanks

17:56 <jernej> junari, apritzel: however, I worry that changes that I send for H616 driver, are dangerously close to be copyright enforcable...

17:58 <jernej> code isn't copied, of course, but using same configuration values makes it more similar nevertheless

18:03 <junari> My TV box is the best example of cost reduction. Combined RAM and eMMC with ES (Engineering Sample) (!!!) mark, without any ESD protection, without transformer in ethernet port, only a few SMD components around the power chip.

18:10 <junari> transpeed t98 https://drive.google.com/file/d/1D7zsjnDYPDUUaeYlxCJ0IM8vnPhp2lqO/view?usp=sharing

18:30 <gamiee> absolutely offtopic, but does anyone have SDWire for sale here?

18:43 <BroderTuck> smaeul: I saw that you had posted a few patches that aimed to get the mainline NAND stuff to no longer be intentionally unusable as rescue system for Allwinner/Android-formatted NAND chips, for A33. How much work would it be to extend this to e.g. the H3-based https://linux-sunxi.org/HYH-TBH3 that I got ....2015?

18:57 <smaeul> BroderTuck: H3 support wouldn't be hard. The NAND controller has barely changed. But it will only ever be able to do a backup/restore. You shouldn't expect to be able to mount vendor NAND partitions.

18:59 <smaeul> the vendor driver uses different page/block layout from mainline, and includes a totally custom wear leveling layer

18:59 <BroderTuck> I'm only after being able to read the FEXbin, to help with a proper DTS.

18:59 <smaeul> ah, well you would be able to dump U-Boot

19:00 <smaeul> were you unable to dump the FEX from RAM when booted to the vendor U-Boot/Android?

19:01 <BroderTuck> Yep, only got some sort of access violantion error when I tried the tool.

19:02 <BroderTuck> The uboot on there had timeout:0 and I never got the serial adapter to be able to interrupt that

19:03 <smaeul> did you try holding the 's' key?

19:04 <BroderTuck> Don't remember, been quite some time since I tried anything with it

19:05 <BroderTuck> Most guides I remember sayd to hold '2'

19:18 <BroderTuck> I did place an android boot log and a few other things here: https://palvencia.se/H3/

19:19 <BroderTuck> NB, that page is from a few years ago