4.11 SMP issues on Intel SE7501CW2

[Jacques Fourie]

On 12/1/05, Jacques Fourie <jacques.fourie at gmail.com> wrote:
> Hi John,
>
> I booted a 6.0-RELEASE CD and the same thing (panic that freezes the
> machine) happens. Can you think of any way in which to reliably reboot
> the machine if this situation occurs?
>
> regards,
> jacques
>
> On 12/1/05, John Baldwin <jhb at freebsd.org> wrote:
> > On Thursday 01 December 2005 08:20 am, Jacques Fourie wrote:
> > > Hi,
> > >
> > > With reference to the following thread :
> > > http://groups.google.com/group/mailing.freebsd.smp/browse_thread/thread/bd4
> > >5afab721e1a85/f66c8476272952af?lnk=st&q=%2Bfreebsd+%2B%22failed!%22+%2Bpanic
> > >&rnum=80#f66c8476272952af
> > >
> > > I am seeing the same issue on an Intel SE7501CW2 dual Xeon machine. 6.0 as
> > > well as -current exhibits the same behaviour. Various postings to the
> > > above thread suggests that this may be due to the APIC ID that the BIOS
> > > claims is assigned to the CPU not being the actual APIC ID assigned to the
> > > CPU. Does anyone have any new information on this issue? If the subsequent
> > > panic succeeded in rebooting the machine this would not be a big issue for
> > > me but unfortunately the machine hangs after pressing 'y' to the "panic
> > > [y/n]" prompt. Is there a way in which to initiate a hard reset in
> > > software?
> >
> > No, there hasn't been any recent info on this and I haven't had any recent
> > reports of these problems, at least not on 5.x or 6.x.  Can you try booting
> > up a 5.4 or 6.0 CD to see if they boot up ok?
> >
> > --
> > John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
> > "Power Users Use the Power to Serve"  =  http://www.FreeBSD.org
> > _______________________________________________
> > freebsd-hackers at freebsd.org mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
> > To unsubscribe, send any mail to "freebsd-hackers-unsubscribe at freebsd.org"
> >
>

Hi John,

In the end a workaround that "solved" the issue for me (on 4.11) was
to call cpu_reset() instead of panic() when failing to start an AP.
This causes the box to reboot reliably instead of freezing and after
the reboot all AP's also start without any issues.  On FreeBSD 6.0
(and -current) the panic() call successfully reboots the box so
although the original problem of failing to start the AP is present on
these platforms the problem is not that severe.

In case anyone is interested in how to reproduce the problem (on 4.11,
6.0 or -current) - just cycle through a few soft reboot cycles (I
placed a /sbin/reboot line in /etc/rc.local).

regards,
jacques