Comment # 26 on bug 1202138 from Marius Kittler

I know. That's also why I'm currently unsure what's best to try next. Btw,
these are the kernel versions we've tried so far:

* 5.3.18 (default) from Leap 15.3: the last kernel version that works reliably
* 5.8.15 (default) from home:tiwai:kernel:5.8: QEMU gets frequently stuck
(leading to openQA tests exceeding the timeout, so basically not able to handle
the workload)
* 5.14.21 (default) from Leap 15.4: frequent crashes (no recorded kernel panic,
system is just stuck and needs power reset via IPMI), can otherwise handle the
workload
* 5.19.12 (default) from Kernel project: same as 5.14
* 6.0.0 (vanilla) from Kernel project: same as 5.14

So the problem we see as of 5.14 must have been introduced after 5.3 with
possible other regressions (that then apparently have been fixed before 5.14
but potentially mask the issue we see as of 5.14).