]> pilppa.com Git - linux-2.6-omap-h63xx.git/commitdiff
cpuset: avoid changing cpuset's cpus when -errno returned
authorLi Zefan <lizf@cn.fujitsu.com>
Sat, 13 Sep 2008 09:33:08 +0000 (02:33 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sat, 13 Sep 2008 21:41:50 +0000 (14:41 -0700)
After the patch:

commit 0b2f630a28d53b5a2082a5275bc3334b10373508
Author: Miao Xie <miaox@cn.fujitsu.com>
Date:   Fri Jul 25 01:47:21 2008 -0700

    cpusets: restructure the function update_cpumask() and update_nodemask()

It might happen that 'echo 0 > /cpuset/sub/cpus' returned failure but 'cpus'
has been changed, because cpus was changed before calling heap_init() which
may return -ENOMEM.

This patch restores the orginal behavior.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/cpuset.c

index f227bc1726906a52a803907299230f62b39edd84..827cd9adccb272f3d54a0ea8a98ba8c028cde211 100644 (file)
@@ -843,37 +843,25 @@ static void cpuset_change_cpumask(struct task_struct *tsk,
 /**
  * update_tasks_cpumask - Update the cpumasks of tasks in the cpuset.
  * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed
+ * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()
  *
  * Called with cgroup_mutex held
  *
  * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,
  * calling callback functions for each.
  *
- * 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_cpumask(struct cpuset *cs)
+static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap)
 {
        struct cgroup_scanner scan;
-       struct ptr_heap heap;
-       int retval;
-
-       /*
-        * cgroup_scan_tasks() will initialize heap->gt for us.
-        * heap_init() is still needed here for we should not change
-        * cs->cpus_allowed when heap_init() fails.
-        */
-       retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL);
-       if (retval)
-               return retval;
 
        scan.cg = cs->css.cgroup;
        scan.test_task = cpuset_test_cpumask;
        scan.process_task = cpuset_change_cpumask;
-       scan.heap = &heap;
-       retval = cgroup_scan_tasks(&scan);
-
-       heap_free(&heap);
-       return retval;
+       scan.heap = heap;
+       cgroup_scan_tasks(&scan);
 }
 
 /**
@@ -883,6 +871,7 @@ static int update_tasks_cpumask(struct cpuset *cs)
  */
 static int update_cpumask(struct cpuset *cs, const char *buf)
 {
+       struct ptr_heap heap;
        struct cpuset trialcs;
        int retval;
        int is_load_balanced;
@@ -917,6 +906,10 @@ static int update_cpumask(struct cpuset *cs, const char *buf)
        if (cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed))
                return 0;
 
+       retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL);
+       if (retval)
+               return retval;
+
        is_load_balanced = is_sched_load_balance(&trialcs);
 
        mutex_lock(&callback_mutex);
@@ -927,9 +920,9 @@ static int update_cpumask(struct cpuset *cs, const char *buf)
         * Scan tasks in the cpuset, and update the cpumasks of any
         * that need an update.
         */
-       retval = update_tasks_cpumask(cs);
-       if (retval < 0)
-               return retval;
+       update_tasks_cpumask(cs, &heap);
+
+       heap_free(&heap);
 
        if (is_load_balanced)
                async_rebuild_sched_domains();
@@ -1965,7 +1958,7 @@ static void scan_for_empty_cpusets(const struct cpuset *root)
                     nodes_empty(cp->mems_allowed))
                        remove_tasks_in_empty_cpuset(cp);
                else {
-                       update_tasks_cpumask(cp);
+                       update_tasks_cpumask(cp, NULL);
                        update_tasks_nodemask(cp, &oldmems);
                }
        }