Netdev List
 help / color / mirror / Atom feed
* [PATCH net-next-2.6] gen_estimator: deadlock fix
@ 2010-03-31 17:06 Eric Dumazet
  2010-04-02  1:44 ` David Miller
  0 siblings, 1 reply; 2+ messages in thread
From: Eric Dumazet @ 2010-03-31 17:06 UTC (permalink / raw)
  To: David Miller; +Cc: netdev

David, feel free to apply following to net-next, as I believe
probability to get this deadlock is close to 0.

Deadlock occurred once with a test machine running latest net-next-2.6,
I was not able to reproduce it a second time, so cannot confirm this fix
is correct.

Thanks

[PATCH] gen_estimator: deadlock fix

One of my test machine got a deadlock during "tc" sessions,
adding/deleting classes & filters, using traffic estimators.

After some analysis, I believe we have a potential use after free case
in est_timer() :

spin_lock(e->stats_lock); << HERE >>
read_lock(&est_lock);
if (e->bstats == NULL)   << TEST >>
	goto skip;

Test is done a bit late, because after estimator is killed, and before
rcu grace period elapsed, we might already have freed/reuse memory where
e->stats_locks points to (some qdisc->q.lock)

A possible fix is to respect a rcu grace period at Qdisc dismantle time.

On 64bit, sizeof(struct Qdisc) is exactly 192 bytes. Adding 16 bytes to
it (for struct rcu_head) is a problem because it might change
performance, given QDISC_ALIGNTO is 32 bytes. 

This is why I also change QDISC_ALIGNTO to 64 bytes, to satisfy most
current alignment requirements.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
 include/net/pkt_sched.h   |    2 +-
 include/net/sch_generic.h |    1 +
 net/sched/sch_generic.c   |   15 +++++++++++++--
 3 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/include/net/pkt_sched.h b/include/net/pkt_sched.h
index b6cdc33..9d4d87c 100644
--- a/include/net/pkt_sched.h
+++ b/include/net/pkt_sched.h
@@ -12,7 +12,7 @@ struct qdisc_walker {
 	int	(*fn)(struct Qdisc *, unsigned long cl, struct qdisc_walker *);
 };
 
-#define QDISC_ALIGNTO		32
+#define QDISC_ALIGNTO		64
 #define QDISC_ALIGN(len)	(((len) + QDISC_ALIGNTO-1) & ~(QDISC_ALIGNTO-1))
 
 static inline void *qdisc_priv(struct Qdisc *q)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index 67dc08e..03ca5d8 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -73,6 +73,7 @@ struct Qdisc {
 	struct sk_buff_head	q;
 	struct gnet_stats_basic_packed bstats;
 	struct gnet_stats_queue	qstats;
+	struct rcu_head     rcu_head;
 };
 
 struct Qdisc_class_ops {
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 5173c1e..1751325 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -528,7 +528,7 @@ struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
 	unsigned int size;
 	int err = -ENOBUFS;
 
-	/* ensure that the Qdisc and the private data are 32-byte aligned */
+	/* ensure that the Qdisc and the private data are 64-byte aligned */
 	size = QDISC_ALIGN(sizeof(*sch));
 	size += ops->priv_size + (QDISC_ALIGNTO - 1);
 
@@ -590,6 +590,13 @@ void qdisc_reset(struct Qdisc *qdisc)
 }
 EXPORT_SYMBOL(qdisc_reset);
 
+static void qdisc_rcu_free(struct rcu_head *head)
+{
+	struct Qdisc *qdisc = container_of(head, struct Qdisc, rcu_head);
+
+	kfree((char *) qdisc - qdisc->padded);
+}
+
 void qdisc_destroy(struct Qdisc *qdisc)
 {
 	const struct Qdisc_ops  *ops = qdisc->ops;
@@ -613,7 +620,11 @@ void qdisc_destroy(struct Qdisc *qdisc)
 	dev_put(qdisc_dev(qdisc));
 
 	kfree_skb(qdisc->gso_skb);
-	kfree((char *) qdisc - qdisc->padded);
+	/*
+	 * gen_estimator est_timer() might access qdisc->q.lock,
+	 * wait a RCU grace period before freeing qdisc.
+	 */
+	call_rcu(&qdisc->rcu_head, qdisc_rcu_free);
 }
 EXPORT_SYMBOL(qdisc_destroy);
 



^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2010-04-02  1:44 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-03-31 17:06 [PATCH net-next-2.6] gen_estimator: deadlock fix Eric Dumazet
2010-04-02  1:44 ` David Miller

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox