Skip to contentSkip to conversation

cgroup: remove rcu_read_lock() from css_get_next()

PATCH In review Bharata B Rao 2 participants 2 messages Mar 4, 2009

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cgroup: remove rcu_read_lock() from css_get_next()

cgroup: remove rcu_read_lock() from css_get_next()

Callers of css_get_next() are already calling css_get_next() under rcu_read_lock() and hence there there is no need for css_get_next() to re-acquire rcu_read_lock(). As per the comments in css_get_next(), let the callers accquire the rcu_read_lock().

---
 kernel/cgroup.c |    3 ---
 1 file changed, 3 deletions(-)
kernel/cgroup.c +0 -3
@@ -3553,7 +3553,6 @@ css_get_next(struct cgroup_subsys *ss, i
 		return NULL;
 
 	BUG_ON(!ss->use_id);
-	rcu_read_lock();
 	/* fill start point for scan */
 	tmpid = id;
 	while (1) {
@@ -3577,8 +3576,6 @@ css_get_next(struct cgroup_subsys *ss, i
 		/* continue to scan from next id */
 		tmpid = tmpid + 1;
 	}
-
-	rcu_read_unlock();
 	return ret;
 }
 

Regards,
Bharata.

Acked-by on v1 1/1
#1 lore raw
Quoted 9 lines · On Wed, 4 Mar 2009 09:59:20 +0530 Bharata B Rao <[email protected]>…

cgroup: remove rcu_read_lock() from css_get_next()

Callers of css_get_next() are already calling css_get_next() under rcu_read_lock() and hence there there is no need for css_get_next() to re-acquire rcu_read_lock(). As per the comments in css_get_next(), let the callers accquire the rcu_read_lock().

Yes, thank you for pointing out. This should be done by caller.

Quoted 26 lines
---
 kernel/cgroup.c |    3 ---
 1 file changed, 3 deletions(-)
kernel/cgroup.c +0 -3
@@ -3553,7 +3553,6 @@ css_get_next(struct cgroup_subsys *ss, i
 		return NULL;
 
 	BUG_ON(!ss->use_id);
-	rcu_read_lock();
 	/* fill start point for scan */
 	tmpid = id;
 	while (1) {
@@ -3577,8 +3576,6 @@ css_get_next(struct cgroup_subsys *ss, i
 		/* continue to scan from next id */
 		tmpid = tmpid + 1;
 	}
-
-	rcu_read_unlock();
 	return ret;
 }
 

Regards,
Bharata.