On 07/08/25(Thu) 17:17, Kirill A. Korinsky wrote:
> On Wed, 06 Aug 2025 14:41:51 +0200,
> Kirill A. Korinsky <[email protected]> wrote:
> > 
> > On Wed, 06 Aug 2025 11:34:45 +0200,
> > Kirill A. Korinsky <[email protected]> wrote:
> > > 
> > > >How-To-Repeat:
> > >   Not sure, can guess that the next build will crash as well.
> > 
> > Indeed it does. Serveral attempts lead to the same state.

What is /etc/fstab for this machine?  Do you have a swap partition on a
local disk?

> > 
> > Each attempt to build it, leads to the crash or frozen device.
> > 
> > If I mount the swap via /etc/fstab:
> > 
> > 172.31.2.23:/volume1/octeon/swap none swap sw,nfsmntpt=/swap
> > 
> > it frozes the device. When I use the following mount
> > 
> > 172.31.2.23:/volume1/octeon /mnt nfs rw,nodev,soft,intr,tcp,-x=2 0 0
> > 
> > and add swap via swapon /mnt/swap, it crashes on ISSET.

Could you reproduce this crash on an amd64?  octeon's ddb is too limited
and we don't even have a usable stack trace.

> Next attempt with swap on the local eMMC leads to somehow frozen device.

This might be a completely different bug.

> I can't connect to it via SSH, but it replies to a ping. Nor input to serial
> console has any affect, but it had continius messages which repeats each few
> seconds (like 5-10): pagedaemon: wait_pla deadlock detected!

That means the pagedaemon needs memory to write to swap.  Could you turn
this printf into a panic and report:

show panic
trace
ps
show uvmexp
show bcstats


Reply via email to