Re: kernel crash in tcp_subr.c:2386

In message <20240212193044.E089D185@slippy.cwsent.com>, Cy Schubert writes:
> In message <625E0EA4-9413-45AD-B05C-500833A1D527@freebsd.org>, 
> tuexen@freebsd.o
> rg writes:
> > > On Feb 12, 2024, at 10:36, Alexander Leidinger =
> > <alexleidingerde@gmail.com> wrote:
> > >=20
> > > Hi,
> > >=20
> > > I got a coredump with sources from 2024-02-10-144617 (GMT+0100):
> > Hi Alexander,
> >
> > we are aware of this problem, but haven't found a way to reproduce it.
> > Do you know how to reproduce this?
>
> I've reproduced this by rebooting any one of my machines in my basement. 
> The other machines will panic as below.
>
> I've reverted the three tcp timer commits, expecting one of them to be the 
> cause.

Another data point:

I build on a build machine and NFS mount /usr/obj on my other machines. 
Another symptom of this problem is that the NFS share will appear 
corrupted. And df -htnfs will sometimes not display the mounted NFS share. 
If not a kernel page fault, random kernel memory can be overwritten 
resulting in bizarre behaviour prior.


-- 
Cheers,
Cy Schubert <Cy.Schubert@cschubert.com>
FreeBSD UNIX:  <cy@FreeBSD.org>   Web:  https://FreeBSD.org
NTP:           <cy@nwtime.org>    Web:  https://nwtime.org

			e^(i*pi)+1=0