From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759318Ab0JHRuH (ORCPT ); Fri, 8 Oct 2010 13:50:07 -0400 Received: from e36.co.us.ibm.com ([32.97.110.154]:42204 "EHLO e36.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758269Ab0JHRuF (ORCPT ); Fri, 8 Oct 2010 13:50:05 -0400 Date: Fri, 8 Oct 2010 23:19:58 +0530 From: Balbir Singh To: Andrew Morton , KAMEZAWA Hiroyuki Cc: "linux-mm@kvack.org" , "linux-kernel@vger.kernel.org" Subject: [BUGFIX] memcg CPU hotplug lockdep warning fix Message-ID: <20101008174958.GI5327@balbir.in.ibm.com> Reply-To: balbir@linux.vnet.ibm.com MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org memcg has lockdep warnings (sleep inside rcu lock) From: Balbir Singh Recent move to get_online_cpus() ends up calling get_online_cpus() from mem_cgroup_read_stat(). However mem_cgroup_read_stat() is called under rcu lock. get_online_cpus() can sleep. The dirty limit patches expose this BUG more readily due to their usage of mem_cgroup_page_stat() This patch address this issue as identified by lockdep and moves the hotplug protection to a higher layer. This might increase the time required to hotplug, but not by much. Warning messages BUG: sleeping function called from invalid context at kernel/cpu.c:62 in_atomic(): 0, irqs_disabled(): 0, pid: 6325, name: pagetest 2 locks held by pagetest/6325: #0: (&mm->mmap_sem){......}, at: [] do_page_fault+0x27d/0x4a0 #1: (rcu_read_lock){......}, at: [] mem_cgroup_page_stat+0x0/0x23f Pid: 6325, comm: pagetest Not tainted 2.6.36-rc5-mm1+ #201 Call Trace: [] __might_sleep+0x12d/0x131 [] get_online_cpus+0x1c/0x51 [] mem_cgroup_read_stat+0x27/0xa3 [] mem_cgroup_page_stat+0x131/0x23f [] ? mem_cgroup_page_stat+0x0/0x23f [] global_dirty_limits+0x42/0xf8 [] throttle_vm_writeout+0x3a/0xb4 [] shrink_zone+0x3e6/0x3f8 [] ? ktime_get_ts+0xb2/0xbf [] do_try_to_free_pages+0x106/0x478 [] try_to_free_mem_cgroup_pages+0xe5/0x14c [] mem_cgroup_hierarchical_reclaim+0x314/0x3a2 [] __mem_cgroup_try_charge+0x29b/0x593 [] ? __mem_cgroup_try_charge+0xb4/0x593 [] ? local_clock+0x40/0x59 [] ? sched_clock+0x9/0xd [] ? sched_clock_local+0x1c/0x82 [] mem_cgroup_charge_common+0x4b/0x76 [] ? bio_add_page+0x36/0x38 [] mem_cgroup_cache_charge+0x1f4/0x214 [] add_to_page_cache_locked+0x4a/0x148 .... Signed-off-by: Balbir Singh --- mm/memcontrol.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 116fecd..f4c5665 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -578,7 +578,6 @@ static s64 mem_cgroup_read_stat(struct mem_cgroup *mem, int cpu; s64 val = 0; - get_online_cpus(); for_each_online_cpu(cpu) val += per_cpu(mem->stat->count[idx], cpu); #ifdef CONFIG_HOTPLUG_CPU @@ -586,7 +585,6 @@ static s64 mem_cgroup_read_stat(struct mem_cgroup *mem, val += mem->nocpu_base.count[idx]; spin_unlock(&mem->pcp_counter_lock); #endif - put_online_cpus(); return val; } @@ -1284,6 +1282,7 @@ s64 mem_cgroup_page_stat(enum mem_cgroup_read_page_stat_item item) struct mem_cgroup *iter; s64 value; + get_online_cpus(); rcu_read_lock(); mem = mem_cgroup_from_task(current); if (mem && !mem_cgroup_is_root(mem)) { @@ -1305,6 +1304,7 @@ s64 mem_cgroup_page_stat(enum mem_cgroup_read_page_stat_item item) } else value = -EINVAL; rcu_read_unlock(); + put_online_cpus(); return value; } -- Three Cheers, Balbir