Ubuntu
linux package

  1. Bug #1690085
  2. Comment #711

Comment 711 for bug 1690085

Revision history for this message
In , zheilbron (zheilbron-linux-kernel-bugs) wrote :

(In reply to zheilbron from comment #624)
> As @C0rn3j mentioned, I'm in a similar boat. With a Ryzen 3700x, I continue
> to get the following errors:
>
> ```
> rcu: INFO: rcu_sched self-detected stall on CPU
> ...
> watchdog: BUG: soft lockup - CPU#1 stuck for 23s! [kworker/1:1:94]
> ```
>
> I've tried just about every workaround that I've read about:
> * Disable C5/C6 states (from BIOS, from zenstates.py)
> * Set "Typical Current Idle" in BIOS
> * Add a combination of kernel parameters (idle=nomwait, noapci, acpi=off,
> rcu_nocbs=0-15, processor.max_cstate=1)
>
> Nothing works.
>
> System specs:
> Linux ... 4.19.0-5-amd64 #1 SMP Debian 4.19.37-5+deb10u2 (2019-08-08) x86_64
> MSI B450 Gaming Pro Carbon AC (BIOS version 7B85v18)
> Ryzen 7 3700X

It seems that leaving the machine idle does not produce the issue. However, connecting over Wireguard + SSH (which is how I had been accessing the machine) seems to cause the issue to manifest. After following the advice here (https://bbs.archlinux.org/viewtopic.php?id=247900), I've added the `nordrand` boot flag. This seems to be keeping the issue at bay so far. Perhaps I was chasing the wrong issue!

Time will tell... I'll report back with any new findings or if it remains stable for some time.