From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752718AbZBZHDx (ORCPT ); Thu, 26 Feb 2009 02:03:53 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751948AbZBZHDl (ORCPT ); Thu, 26 Feb 2009 02:03:41 -0500 Received: from cn.fujitsu.com ([222.73.24.84]:54449 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751849AbZBZHDk (ORCPT ); Thu, 26 Feb 2009 02:03:40 -0500 Message-ID: <49A63EE0.3030406@cn.fujitsu.com> Date: Thu, 26 Feb 2009 15:04:00 +0800 From: Li Zefan User-Agent: Thunderbird 2.0.0.9 (X11/20071115) MIME-Version: 1.0 To: Andrew Morton CC: Paul Menage , LKML Subject: [PATCH 3/4] cpuset: avoid changing cpuset's mems when errno returned References: <49A63E77.9060603@cn.fujitsu.com> In-Reply-To: <49A63E77.9060603@cn.fujitsu.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org When writing to cpuset.mems, cpuset has to update its mems_allowed before calling update_tasks_nodemask(), but this function might return -ENOMEM. To avoid this rare case, we allocate the memory before changing mems_allowed, and then pass to update_tasks_nodemask(). Similar to what update_cpumask() does. Signed-off-by: Li Zefan --- kernel/cpuset.c | 25 ++++++++++++++++--------- 1 files changed, 16 insertions(+), 9 deletions(-) diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 01a111f..3202364 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1057,13 +1057,15 @@ static void *cpuset_being_rebound; * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed * @oldmem: old mems_allowed of cpuset cs + * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() * * Called with cgroup_mutex held - * Return 0 if successful, -errno if not. + * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 + * if @heap != NULL. */ -static int update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem) +static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, + struct ptr_heap *heap) { - int retval; struct cgroup_scanner scan; cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ @@ -1071,7 +1073,7 @@ static int update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem) scan.cg = cs->css.cgroup; scan.test_task = NULL; scan.process_task = cpuset_change_nodemask; - scan.heap = NULL; + scan.heap = heap; scan.data = (nodemask_t *)oldmem; /* @@ -1084,12 +1086,10 @@ static int update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem) * It's ok if we rebind the same mm twice; mpol_rebind_mm() * is idempotent. Also migrate pages in each mm to new nodes. */ - retval = cgroup_scan_tasks(&scan); + cgroup_scan_tasks(&scan); /* We're done rebinding vmas to this cpuset's new mems_allowed. */ cpuset_being_rebound = NULL; - - return retval; } /* @@ -1110,6 +1110,7 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, { nodemask_t oldmem; int retval; + struct ptr_heap heap; /* * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY]; @@ -1144,12 +1145,18 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, if (retval < 0) goto done; + retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); + if (retval < 0) + goto done; + mutex_lock(&callback_mutex); cs->mems_allowed = trialcs->mems_allowed; cs->mems_generation = cpuset_mems_generation++; mutex_unlock(&callback_mutex); - retval = update_tasks_nodemask(cs, &oldmem); + update_tasks_nodemask(cs, &oldmem, &heap); + + heap_free(&heap); done: return retval; } @@ -2002,7 +2009,7 @@ static void scan_for_empty_cpusets(struct cpuset *root) remove_tasks_in_empty_cpuset(cp); else { update_tasks_cpumask(cp, NULL); - update_tasks_nodemask(cp, &oldmems); + update_tasks_nodemask(cp, &oldmems, NULL); } } } -- 1.5.4.rc3