On Mon, Feb 06, 2023 at 03:43:20PM -0800, [email protected] wrote:
> >Synopsis:    Resuming from zzz/ZZZ produces an unstable machine that locks 
> >up after a few seconds
> >Category:    suspend/resume
> >Environment:
>       System      : OpenBSD 7.2
>       Details     : OpenBSD 7.2-current (GENERIC.MP) #1021: Sun Feb  5 
> 09:52:50 MST 2023
>                        
> [email protected]:/usr/src/sys/arch/amd64/compile/GENERIC.MP
> 
>       Architecture: OpenBSD.amd64
>       Machine     : amd64
> >Description:
>       On recent snapshots (within the last week or so), this machine is
>       unstable after zzz/ZZZ (both methods produce the same symptoms).
>       The symptoms are that the machine resumes back into X, and the X
>       mouse cursor works, but the keyboard doesn't, *except* ctrl-alt-f1
>       which takes you back to tty01. Pressing the power button does cause
>       a clean shutdown but it takes several minutes for that to occur.
>       When resumed, sometimes any open xterm windows are missing all their
>       text (but when focus goes into the window, the cursor is re-hilighted
>       in the place it was before zzz). It does not appear to matter what
>       is running on the machine; a cold boot followed immediately by zzz/ZZZ
>       is sufficient to cause the issue.
> 
>       After a few minutes, the machine apparently lands in ddb since the
>       X cursor stops moving and "bo re" (boot/reboot) works to reset the
>       machine. If I zzz/ZZZ *from* tty01, then on resume I still can't type
>       anything and eventually the machine still lands in ddb, but there 
>       is nothing printed to that tty.
> 
>       Another symptom is being able to ctrl-alt-f1 after un-zzz/ZZZ but then
>         you can't type anything. This happens 100% of the time with recent
>       snaps.
> 
>       If nobody has suggestions, I will start bisecting snaps and seeing what
>       date the issue was introduced.
> >How-To-Repeat:
>       zzz then un-zzz, ctrl-alt-f1, then see if you can type. Or just wait,
>       and eventually the machine will enter ddb.
> >Fix:
>       unknown

The root cause of this has been tracked down. Hopefully a fix will land
soon.

Reply via email to