Hi, On Mon, 2020-11-16 at 15:06, Oliver Neukum wrote: > Am Samstag, den 14.11.2020, 13:55 +0100 schrieb Milan P. Stanić: > Hi, > > > Driver loads and work some time but at unpredictable time it stops and > > starts to flood 'dmesg' output until shutdown (eth) interface. > > > > Machine where I test this is arm64 ACER R13 chromebook. > > Are you sure this started in v5.10 or did you not try earlier kernels? I never used much mainline kernels on this machine because previous kernels had issues and quirks making them practically unusable. Earlier I used old chromeOS kernel (3.18 series) and from July this summer I switched to kernels from https://gitlab.collabora.com/eballetbo/linux/-/tree/topic/chromeos/somewhat-stable-5.7 https://gitlab.collabora.com/eballetbo/linux/-/tree/topic/chromeos/somewhat-stable-5.8 https://gitlab.collabora.com/eballetbo/linux/-/tree/topic/chromeos/somewhat-stable-5.9 They had also some issues but on somewhat-stable-5.8 and somewhat-stable-5.9 r8152 worked, or I missed to notice this problem. First mainline kernel which works fine is 5.10-rc1 and up (well, still some quirks and issues) and it is usable for everyday work. And on this kernel and up to 5.10-rc3 problem exist, right now I'm testing 5.10-rc4 to see will that problem appear again. Sometimes it need full 24 hours or more to manifest and I can't conclude how to 'trigger' it. -- Kind regards > > 'dmesg' shows a lot of such messages: > > > > [305186.211284] r8152 2-1.2.4:1.0 eth0: Couldn't submit rx[000000005d575699], ret = -11 > > EAGAIN is an unusual error return. > > Regards > Oliver > >
