Bug#644604:

[Moritz Mühlenhoff <jmm@inutil.org>]

On Fri, Oct 14, 2011 at 02:22:37PM +0100, Ian Campbell wrote:
> On Fri, 2011-10-14 at 14:11 +0200, Mauro wrote:
> > Hello, sorry for bother.
> > I've exactly the same problem explained in the bug but with xen 4.0.1
> > distributed with squeeze.
> > I can't do live migrations too.
> > Sorry for question but when do you think you can solve the bug?
> 
> I sent out a fix on Monday[0] but haven't heard from the upstream
> maintainers yet. I'll chase them up.

We fixed that at work by reverting the following patch from 2.6.32.42,
which is sufficient to fix the regression:

commit 652c98bac315a2253628885f05cfd5f30b553ae5
Author: Thomas Gleixner <tglx@linutronix.de>
Date:   Sat Feb 5 20:08:59 2011 +0000

    xen: Use IRQF_FORCE_RESUME

    commit 676dc3cf5bc36a9e129a3ad8fe3bd7b2ebf20f5d upstream.

    Mark the IRQF_NO_SUSPEND interrupts IRQF_FORCE_RESUME and remove the extra
    walk through the interrupt descriptors.

    Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
    Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
    Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
    Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Cheers,
        Moritz