09:48 <hexdump0815> xdarklight: i think the m8s box is simply broken - memtester shows reproducably errors and i then compiled a static memtester binary and ran it on android where it gave errors too - linux kernel cmdline memtest=64 did not help either - so bye bye m8s most probably ...

09:49 <hexdump0815> xdarklight: did you really push your mxiii usb fix yesterday? i could not find it

09:54 <xdarklight> hexdump0815: https://github.com/xdarklight/linux/commits/meson-mx-integration-5.8-20200501/drivers/phy/amlogic/phy-meson8b-usb2.c - it does not need any mxiii .dts change ;)

09:55 <xdarklight> hexdump0815: sad to see hardware go, I only had this once (for my GXBB board) but it also threw memory errors (so many that sometimes even u-boot wouldn't initialize)

10:00 <hexdump0815> xdarklight: i looked through the commit history of your latest branch, but was not able to see that commit and still cannot see it there ...

10:01 <xdarklight> hexdump0815: for me it's on the bottom of page 2

10:03 <hexdump0815> xdarklight: ah - ok i did not look that far - i think i'll never understand the order of the github commit history

10:05 <hexdump0815> xdarklight: the m8s problem looks like the soldering of the memory chips or the soc getting bad - the warmer the box gets the more unstable it gets - looks like with memtest=64 it does not even boot up properly anymore ... maybe i should put some ice cubes on top of it :)

10:08 <xdarklight> hexdump0815: heh - or use liquid nitrogen ;)

10:38 <hexdump0815> xdarklight: the usb phy fix does not seem to work - this is the dmesg: https://pastebin.com/raw/vCUcRU1c and this is the patch i used (slightly adapted to fit the 5.6.8 kernel i use): https://github.com/hexdump0815/linux-mainline-and-mali-amlogic-kernel/blob/master/misc.m8x/mxiii-usb-phy-fix.patch

10:39 <hexdump0815> xdarklight: does it actually apply to the mxiii, which as we found out is meson8 and not meson8m2?

10:40 <xdarklight> hexdump0815: can you please pick all four patches which I committed yesterday to https://github.com/xdarklight/linux/commits/meson-mx-integration-5.8-20200501/drivers/phy/amlogic/phy-meson8b-usb2.c ? these should all apply cleanly, even to 5.6

10:41 <hexdump0815> xdarklight: ah - i see - will do ...

11:23 <hexdump0815> xdarklight: good news - it works - my mxiii has usb now - thanks a lot!

11:27 <xdarklight> hexdump0815: nice, so I'll be sending patches later :)

11:32 <hexdump0815> xdarklight: please put me on cc, so that i can add my tested-by, as i'm not on the list myself

11:43 <hexdump0815> xdarklight: do you think it would make sense to bring in the dts files i'm using on my boxes (in cleaned up form) too? i'm now using meson8-mxii.dts* and meson8m2-m8s-dvfs.dts from https://github.com/hexdump0815/linux-mainline-and-mali-amlogic-kernel/tree/master/misc.m8x/dtb

11:44 <xdarklight> hexdump0815: I think upstreaming them makes sense, because whenever I do larger changes (like fixing RGMII Ethernet for example) I'm always searching in all existing .dts and update them accordingly