From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755512AbYFJPkm (ORCPT ); Tue, 10 Jun 2008 11:40:42 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752241AbYFJPke (ORCPT ); Tue, 10 Jun 2008 11:40:34 -0400 Received: from E23SMTP03.au.ibm.com ([202.81.18.172]:49077 "EHLO e23smtp03.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751309AbYFJPkd (ORCPT ); Tue, 10 Jun 2008 11:40:33 -0400 Date: Tue, 10 Jun 2008 08:33:01 -0700 From: "Paul E. McKenney" To: Peter Zijlstra Cc: linux-kernel@vger.kernel.org, Ingo Molnar , Thomas Gleixner , Steven Rostedt , Clark Williams , Gregory Haskins , Gautham R Shenoy , Pekka Enberg , Arnaldo Carvalho de Melo Subject: Re: [PATCH -rt 5/5] cpu-hotplug: cpu_down vs preempt-rt Message-ID: <20080610153301.GD15481@linux.vnet.ibm.com> Reply-To: paulmck@linux.vnet.ibm.com References: <20080610111259.766940257@chello.nl> <20080610111832.969119014@chello.nl> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20080610111832.969119014@chello.nl> User-Agent: Mutt/1.5.13 (2006-08-11) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Jun 10, 2008 at 01:13:04PM +0200, Peter Zijlstra wrote: > idle_task_exit() calls mmdrop() from the idle thread, but in PREEMPT_RT all the > allocator locks are sleeping locks - for obvious reasons scheduling away the > idle thread gives some curious problems. > > Solve this by pushing the mmdrop() into an RCU callback, however we can't use > RCU because the CPU is already down and all the local RCU state has been > destroyed. > > Therefore create a new call_rcu() variant that enqueues the callback on an > online cpu. I am a bit nervous about the non-determinism, but on the other hand CPU online/offline events can only happen so often due to the locking. So... Reviewed-by: Paul E. McKenney > Signed-off-by: Peter Zijlstra > --- > include/linux/mm_types.h | 5 +++++ > include/linux/rcupreempt.h | 2 ++ > kernel/rcupreempt.c | 29 +++++++++++++++++++++++++++++ > kernel/sched.c | 13 +++++++++++++ > 4 files changed, 49 insertions(+) > > Index: linux-2.6.24.7.noarch/include/linux/mm_types.h > =================================================================== > --- linux-2.6.24.7.noarch.orig/include/linux/mm_types.h > +++ linux-2.6.24.7.noarch/include/linux/mm_types.h > @@ -10,6 +10,7 @@ > #include > #include > #include > +#include > #include > #include > > @@ -222,6 +223,10 @@ struct mm_struct { > /* aio bits */ > rwlock_t ioctx_list_lock; > struct kioctx *ioctx_list; > + > +#ifdef CONFIG_PREEMPT_RT > + struct rcu_head rcu_head; > +#endif > }; > > #endif /* _LINUX_MM_TYPES_H */ > Index: linux-2.6.24.7.noarch/include/linux/rcupreempt.h > =================================================================== > --- linux-2.6.24.7.noarch.orig/include/linux/rcupreempt.h > +++ linux-2.6.24.7.noarch/include/linux/rcupreempt.h > @@ -83,6 +83,8 @@ extern void FASTCALL(call_rcu_classic(st > void (*func)(struct rcu_head *head))); > extern void FASTCALL(call_rcu_preempt(struct rcu_head *head, > void (*func)(struct rcu_head *head))); > +extern void FASTCALL(call_rcu_preempt_online(struct rcu_head *head, > + void (*func)(struct rcu_head *head))); > extern void __rcu_read_lock(void); > extern void __rcu_read_unlock(void); > extern void __synchronize_sched(void); > Index: linux-2.6.24.7.noarch/kernel/rcupreempt.c > =================================================================== > --- linux-2.6.24.7.noarch.orig/kernel/rcupreempt.c > +++ linux-2.6.24.7.noarch/kernel/rcupreempt.c > @@ -916,6 +916,35 @@ void fastcall call_rcu_preempt(struct rc > } > EXPORT_SYMBOL_GPL(call_rcu_preempt); > > +void fastcall call_rcu_preempt_online(struct rcu_head *head, > + void (*func)(struct rcu_head *rcu)) > +{ > + struct rcu_data *rdp; > + unsigned long flags; > + int cpu; > + > + head->func = func; > + head->next = NULL; > +again: > + cpu = first_cpu(cpu_online_map); > + rdp = RCU_DATA_CPU(cpu); > + > + spin_lock_irqsave(&rdp->lock, flags); > + if (unlikely(!cpu_online(cpu))) { > + /* > + * cpu is removed from the online map before rcu_offline_cpu > + * is called. > + */ > + spin_unlock_irqrestore(&rdp->lock, flags); > + goto again; > + } > + > + *rdp->nexttail = head; > + rdp->nexttail = &head->next; > + spin_unlock_irqrestore(&rdp->lock, flags); > + > +} > + > /* > * Check to see if any future RCU-related work will need to be done > * by the current CPU, even if none need be done immediately, returning > Index: linux-2.6.24.7.noarch/kernel/sched.c > =================================================================== > --- linux-2.6.24.7.noarch.orig/kernel/sched.c > +++ linux-2.6.24.7.noarch/kernel/sched.c > @@ -5888,6 +5888,15 @@ void sched_idle_next(void) > spin_unlock_irqrestore(&rq->lock, flags); > } > > +#ifdef CONFIG_PREEMPT_RT > +void mmdrop_rcu(struct rcu_head *head) > +{ > + struct mm_struct *mm = container_of(head, struct mm_struct, rcu_head); > + > + mmdrop(mm); > +} > +#endif > + > /* > * Ensures that the idle task is using init_mm right before its cpu goes > * offline. > @@ -5900,7 +5909,11 @@ void idle_task_exit(void) > > if (mm != &init_mm) > switch_mm(mm, &init_mm, current); > +#ifdef CONFIG_PREEMPT_RT > + call_rcu_preempt_online(&mm->rcu_head, mmdrop_rcu); > +#else > mmdrop(mm); > +#endif > } > > /* called under rq->lock with disabled interrupts */ > > -- >