From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Campbell Subject: [PATCH 4/4] xen: do not suspend IPI IRQs. Date: Wed, 28 Jul 2010 11:54:27 +0100 Message-ID: <1280314467-5637-4-git-send-email-ian.campbell@citrix.com> References: <1280314445.24292.90.camel@zakaz.uk.xensource.com> Return-path: In-Reply-To: <1280314445.24292.90.camel@zakaz.uk.xensource.com> Sender: linux-kernel-owner@vger.kernel.org To: linux-kernel@vger.kernel.org Cc: Ian Campbell , Thomas Gleixner , Jeremy Fitzhardinge , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org In general the semantics of IPIs are that they are are expected to continue functioning after dpm_suspend_noirq(). Specifically I have seen a deadlock between the callfunc IPI and the stop machine used by xen's do_suspend() routine. If one CPU has already called dpm_suspend_noirq() then there is a window where it can be sent a callfunc IPI before all the other CPUs have entered stop_cpu(). If this happens then the first CPU ends up spinning in stop_cpu() waiting for the other to rendezvous in state STOPMACHINE_PREPARE while the other is spinning in csd_lock_wait(). Signed-off-by: Ian Campbell Cc: Thomas Gleixner Cc: Jeremy Fitzhardinge Cc: xen-devel@lists.xensource.com Cc: # .32.x+: 517d6934: irq: rename IRQF_TIMER to IRQF_NO_SUSPEND Cc: # .32.x+ --- drivers/xen/events.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/drivers/xen/events.c b/drivers/xen/events.c index db8f506..28f133a 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -536,6 +536,7 @@ int bind_ipi_to_irqhandler(enum ipi_vector ipi, if (irq < 0) return irq; + irqflags |= IRQF_NO_SUSPEND; retval = request_irq(irq, handler, irqflags, devname, dev_id); if (retval != 0) { unbind_from_irq(irq); -- 1.5.6.5