[GIT PULL] cgroup fixes for v3.5-rc3

[Tejun Heo]

Hello, Linus.

Please pull from the following branch to receive two patches fixing a
refcnt race bug during css_put().  Decing and checking the value
weren't atomic and two tasks could think that they both pushed the
counter to zero.

  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git for-3.5-fixes

Salman Qazi (2):
      cgroup: make sure that decisions in __css_put are atomic
      cgroups: Account for CSS_DEACT_BIAS in __css_put

Thanks.
---
 kernel/cgroup.c |   13 ++++++++++---
 1 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 72fcd30..2097684 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -255,12 +255,17 @@ int cgroup_lock_is_held(void)
 
 EXPORT_SYMBOL_GPL(cgroup_lock_is_held);
 
+static int css_unbias_refcnt(int refcnt)
+{
+	return refcnt >= 0 ? refcnt : refcnt - CSS_DEACT_BIAS;
+}
+
 /* the current nr of refs, always >= 0 whether @css is deactivated or not */
 static int css_refcnt(struct cgroup_subsys_state *css)
 {
 	int v = atomic_read(&css->refcnt);
 
-	return v >= 0 ? v : v - CSS_DEACT_BIAS;
+	return css_unbias_refcnt(v);
 }
 
 /* convenient tests for these bits */
@@ -4982,10 +4987,12 @@ EXPORT_SYMBOL_GPL(__css_tryget);
 void __css_put(struct cgroup_subsys_state *css)
 {
 	struct cgroup *cgrp = css->cgroup;
+	int v;
 
 	rcu_read_lock();
-	atomic_dec(&css->refcnt);
-	switch (css_refcnt(css)) {
+	v = css_unbias_refcnt(atomic_dec_return(&css->refcnt));
+
+	switch (v) {
 	case 1:
 		if (notify_on_release(cgrp)) {
 			set_bit(CGRP_RELEASABLE, &cgrp->flags);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/