From: Ian Campbell Subject: Re: pv 2.6.31 (kernel.org) and save/migrate fails, domU BUG Date: Wed, 25 Nov 2009 20:03:05 +0000 Message-ID: <1259179385.30589.0.camel@localhost.localdomain> References: <20091108154153.GM1434@reaktio.net> <693ea516-aa5b-4f82-ad48-1bd51cfa3480@default> <20091108172747.GR1434@reaktio.net> <20091110100806.GE16033@reaktio.net> <4AFC9BFF.9030707@goop.org> <1259072824.7590.389.camel@zakaz.uk.xensource.com> <1259158328.7590.539.camel@zakaz.uk.xensource.com> <4B0D8554.3010103@goop.org> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <4B0D8554.3010103@goop.org> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jeremy Fitzhardinge Cc: "Rafael J. Wysocki" , Dan Magenheimer , "Xen-Devel (E-mail)" , Gleixner , Thomas List-Id: xen-devel@lists.xenproject.org On Wed, 2009-11-25 at 19:28 +0000, Jeremy Fitzhardinge wrote: > On 11/25/09 06:12, Ian Campbell wrote: > > tick_resume() is never called on secondary processors. Presumably this > > is because they are offlined for suspend on native and so this is > > normally taken care of in the CPU onlining path. Under Xen we keep all > > CPUs online over a suspend. > > > > This patch papers over the issue for me but I will investigate a more > > generic, less hacky, way of doing to the same. > > > > tick_suspend is also only called on the boot CPU which I presume should > > be fixed too. > > > > Yep. I wonder how it ever worked? There's been a fair amount of change > in the PM code, so that could have changed things. I don't know if > there's a deep reason for not calling tick_resume() on all processors. > > Rafael, tglx: suspend/resume under Xen doesn't need to hot unplug all > the CPUs, so we don't; the hypervisor can manage the context > save/restore for all CPUs. Is there a deep reason why > timekeeping_resume() can't call the CLOCK_EVT_NOTIFY_RESUME notifier on > all online CPUs? Interrupts are disabled at that point where it currently calls the notifier, so none of the SMP function call primitives work. > > void xen_arch_resume(void) > > { > > - /* nothing */ > > + smp_call_function_many(cpu_online_mask, xen_vcpu_notify_restore, > > + (void *)CLOCK_EVT_NOTIFY_RESUME, 1); > > } > > > > This is equivalent to smp_call_function(). Oh yeah. Ian.