Re: HVM guests hanging in hpet_rtc_interrupt

[Olaf Hering <olaf@aepfle.de>]

On Mon, Feb 04, Ian Campbell wrote:

> On Mon, 2013-02-04 at 15:27 +0000, Olaf Hering wrote:
> > On Mon, Feb 04, Olaf Hering wrote:
> > 
> > > Now I'm going to boot into sles11 dom0 and see if the hpet1 messsages
> > > are really gone.
> > 
> > With sles11 dom0+domU there are no more hpet1 messages when started from
> > xend.
> > 
> > 
> > With xl however I now get a hanging guest, which I did not see before this
> > patch. But this happens only if it is booted with quiet, with 'debug
> > ignore_loglevel' booting works fine.
> > Also xl list shows the domain in paused state?!
> 
> [...]
> >   [<ffffffff812a9543>] acpi_os_write_port+0xe  <--
> 
> I expect this will have trapped to qemu, which would explain why the
> vcpu is paused, but not why qemu appears to not be responding. Maybe
> something in the qemu logs will shed some light?

I was running xenctx in another window, and it leads to this sort of
hang. Without the xenctx monitoring booting works fine.

watch -n 0,1 '/usr/lib64/xen/bin/xenctx -C -s /boot/System.map-3.0.51-0.7.9-default \
        `xl domid sles11sp2_full_minimal_nfsroot_bug694863 2>&1` 2>&1'

Olaf