From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755163Ab3AEIbi (ORCPT ); Sat, 5 Jan 2013 03:31:38 -0500 Received: from e28smtp02.in.ibm.com ([122.248.162.2]:49131 "EHLO e28smtp02.in.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754773Ab3AEIbg (ORCPT ); Sat, 5 Jan 2013 03:31:36 -0500 Message-ID: <1357374665.716.32.camel@ThinkPad-T5421.cn.ibm.com> Subject: Re: [PATCH] Remove WARN_ON_ONCE in __smp_call_function_single and smp_call_function_single From: Li Zhong To: Hui Zhu Cc: Peter Zijlstra , Thomas Gleixner , Andrew Morton , Michal Nazarewicz , "Srivatsa S. Bhat" , linux-kernel@vger.kernel.org, Zhu Yanhai , yan@linux.vnet.ibm.com, "Paul E. McKenney" Date: Sat, 05 Jan 2013 16:31:05 +0800 In-Reply-To: References: Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.2.3-0ubuntu6 Content-Transfer-Encoding: 7bit Mime-Version: 1.0 X-Content-Scanned: Fidelis XPS MAILER x-cbid: 13010508-5816-0000-0000-0000061E71E9 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 2013-01-05 at 12:26 +0800, Hui Zhu wrote: > The comments of these WARN_ON_ONCE said "Can deadlock when called with > interrupts disabled". > But I didn't find anything that will block this function. It seems that if two cpus try to IPI each other with wait equals true, they will deadlock if interrupts are disabled, because neither of them can take the interrupt. If wait equals false, it seems deadlock is still possible, for example, if smp_call_function_single() is called from interrupt context, then one cpu might -->call smp_call_function_single() -->csd_lock() -->be interrupted before sending the ipi out -->in interrupt context, call smp_call_function_single() again -->csd_lock_wait() forever in csd_lock() I'm not sure whether it is possible that irqs be enabled in interrupt context? If so, maybe we also need check that if wait equals false? Thanks, Zhong > So I think maybe this WARN_ON_ONCE is not need now. So I post a patch for that. > > Signed-off-by: Hui Zhu > Signed-off-by: Zhu Yanhai > > --- a/kernel/smp.c > +++ b/kernel/smp.c > @@ -303,25 +303,13 @@ int smp_call_function_single(int cpu, sm > .flags = 0, > }; > unsigned long flags; > - int this_cpu; > int err = 0; > > /* > * prevent preemption and reschedule on another processor, > * as well as CPU removal > */ > - this_cpu = get_cpu(); > - > - /* > - * Can deadlock when called with interrupts disabled. > - * We allow cpu's that are not yet online though, as no one else can > - * send smp call function interrupt to this cpu and as such deadlocks > - * can't happen. > - */ > - WARN_ON_ONCE(cpu_online(this_cpu) && irqs_disabled() > - && !oops_in_progress); > - > - if (cpu == this_cpu) { > + if (cpu == get_cpu()) { > local_irq_save(flags); > func(info); > local_irq_restore(flags); > @@ -409,17 +397,7 @@ void __smp_call_function_single(int cpu, > unsigned int this_cpu; > unsigned long flags; > > - this_cpu = get_cpu(); > - /* > - * Can deadlock when called with interrupts disabled. > - * We allow cpu's that are not yet online though, as no one else can > - * send smp call function interrupt to this cpu and as such deadlocks > - * can't happen. > - */ > - WARN_ON_ONCE(cpu_online(smp_processor_id()) && wait && irqs_disabled() > - && !oops_in_progress); > - > - if (cpu == this_cpu) { > + if (cpu == get_cpu()) { > local_irq_save(flags); > data->func(data->info); > local_irq_restore(flags); > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ >