hnyman:

duplicate (with x86 evidence) of FS#2414

ynezz:

duplicate (with x86 evidence) of FS#2414

symptoms looks same, but there's no evidence (log of the crash) as in FS#2414 so we don't know if it's duplicate (yet).

If you're able to reproduce this on x86, then it's very likely, that you can simply run the upstream kernel 5.2 or 5.3 and thus can try to reproduce it there. If the issue is present in the upstream kernel as well, then you can ask for the help on the ath10k mailing list.

Hauke:

This branch still contains some more backports version between 4.19 and 5.2, could you please check with which version this problem started:
https://git.openwrt.org/?p=openwrt/staging/hauke.git;a=shortlog;h=refs/heads/mac80211-5.0

hnyman:

A brief summary of my longer test report in the IPQ806X bug report.

Thanks for pointer to those intermediate commits in your staging tree. I checked it out and built minimal versions of the firmware. baseline and three first commits worked ok, but then the chrasing starts :-(

The culprit is the bump to 5.1-rc2-1:

commit ba5b3ce570ea3abb7a05d00fd6d4d4a32804b0e4
mac80211: update to version 5.0.5-1
OK

commit a9e9bdb8ba26ae3cf9df0a230d8eacd92deac821
mac80211: Update to version 5.1-rc2-1
CRASH

https://git.openwrt.org/?p=openwrt/staging/hauke.git;a=commit;h=a9e9bdb8ba26ae3cf9df0a230d8eacd92deac821

Pure guess, but that commit seems to remove some airtime related patches, and like I have said earlier, the crash always happens at:
PC is at ieee80211_sta_register_airtime+0x24/0x148 [mac80211]

Hauke:

I proposed a patch to test in the last comment of FS#2414