From: Eric Dumazet <eric.dumazet@gmail.com>
To: Changli Gao <xiaosuo@gmail.com>, David Miller <davem@davemloft.net>
Cc: netdev <netdev@vger.kernel.org>,
Stephen Hemminger <shemminger@vyatta.com>,
Jarek Poplawski <jarkao2@gmail.com>,
Patrick McHardy <kaber@trash.net>
Subject: [PATCH net-2.6] pkt_sched: gen_estimator: add a new lock
Date: Mon, 07 Jun 2010 19:18:27 +0200 [thread overview]
Message-ID: <1275931108.2545.168.camel@edumazet-laptop> (raw)
In-Reply-To: <1275929761.2545.159.camel@edumazet-laptop>
Le lundi 07 juin 2010 à 18:56 +0200, Eric Dumazet a écrit :
> >
> > For your information, bug is already there before my patch.
> >
> > So this est_lock is a wrong protection, in the sense its so convoluted
> > that nobody but you and me even noticed it was buggy in the first place.
> >
> > (see commit 5d944c640b4 for a first patch)
> >
> >
>
> Here is v2 of the patch.
>
> Even if its a bug correction, I cooked it for net-next-2.6 since bug
> probably never occured, and patch is too large to be sent to
> net-2.6/linux-2.6 before testing.
>
> Another bug comes from net/netfilter/xt_RATEEST.c : It apparently
> calls gen_kill_estimator() / gen_new_estimator() without holding RTNL ?
>
> So we should add another lock to protect things (est_root, elist[], ...)
>
> David, I can send a net-2.6 patch for this one, since it should be small
> enough. If yes, I'll respin this patch of course ;)
[PATCH net-2.6] pkt_sched: gen_estimator: add a new lock
gen_kill_estimator() / gen_new_estimator() is not always called with
RTNL held.
net/netfilter/xt_RATEEST.c is one user of these API that do not hold
RTNL, so random corruptions can occur between "tc" and "iptables"
Add a new fine grained lock instead of trying to use RTNL in xt_RATEEST
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
net/core/gen_estimator.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/net/core/gen_estimator.c b/net/core/gen_estimator.c
index cf8e703..3d11203 100644
--- a/net/core/gen_estimator.c
+++ b/net/core/gen_estimator.c
@@ -107,6 +107,7 @@ static DEFINE_RWLOCK(est_lock);
/* Protects against soft lockup during large deletion */
static struct rb_root est_root = RB_ROOT;
+static DEFINE_SPINLOCK(est_tree_lock);
static void est_timer(unsigned long arg)
{
@@ -201,7 +202,6 @@ struct gen_estimator *gen_find_node(const struct gnet_stats_basic_packed *bstats
*
* Returns 0 on success or a negative error code.
*
- * NOTE: Called under rtnl_mutex
*/
int gen_new_estimator(struct gnet_stats_basic_packed *bstats,
struct gnet_stats_rate_est *rate_est,
@@ -222,6 +222,7 @@ int gen_new_estimator(struct gnet_stats_basic_packed *bstats,
if (est == NULL)
return -ENOBUFS;
+ spin_lock(&est_tree_lock);
idx = parm->interval + 2;
est->bstats = bstats;
est->rate_est = rate_est;
@@ -242,6 +243,7 @@ int gen_new_estimator(struct gnet_stats_basic_packed *bstats,
list_add_rcu(&est->list, &elist[idx].list);
gen_add_node(est);
+ spin_unlock(&est_tree_lock);
return 0;
}
@@ -261,13 +263,13 @@ static void __gen_kill_estimator(struct rcu_head *head)
*
* Removes the rate estimator specified by &bstats and &rate_est.
*
- * NOTE: Called under rtnl_mutex
*/
void gen_kill_estimator(struct gnet_stats_basic_packed *bstats,
struct gnet_stats_rate_est *rate_est)
{
struct gen_estimator *e;
+ spin_lock(&est_tree_lock);
while ((e = gen_find_node(bstats, rate_est))) {
rb_erase(&e->node, &est_root);
@@ -278,6 +280,7 @@ void gen_kill_estimator(struct gnet_stats_basic_packed *bstats,
list_del_rcu(&e->list);
call_rcu(&e->e_rcu, __gen_kill_estimator);
}
+ spin_unlock(&est_tree_lock);
}
EXPORT_SYMBOL(gen_kill_estimator);
next prev parent reply other threads:[~2010-06-07 17:18 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-06-07 14:32 [PATCH net-next-2.6] pkt_sched: gen_estimator: kill est_lock rwlock Eric Dumazet
2010-06-07 14:53 ` Changli Gao
2010-06-07 15:30 ` Eric Dumazet
2010-06-07 15:55 ` Eric Dumazet
2010-06-07 16:56 ` [PATCH net-next-2.6 v2] " Eric Dumazet
2010-06-07 17:18 ` Eric Dumazet [this message]
2010-06-08 1:00 ` [PATCH net-2.6] pkt_sched: gen_estimator: add a new lock Changli Gao
2010-06-08 4:30 ` Eric Dumazet
2010-06-08 4:57 ` Changli Gao
2010-06-08 4:58 ` Eric Dumazet
2010-06-08 5:20 ` Changli Gao
2010-06-08 5:39 ` Eric Dumazet
2010-06-09 9:39 ` [PATCH net-2.6 v2] " Eric Dumazet
2010-06-09 11:33 ` Jarek Poplawski
2010-06-09 11:55 ` Eric Dumazet
2010-06-11 5:54 ` David Miller
2010-06-08 12:15 ` [PATCH net-next-2.6 v2] pkt_sched: gen_estimator: kill est_lock rwlock Jarek Poplawski
2010-06-08 12:27 ` Eric Dumazet
2010-06-08 12:40 ` Jarek Poplawski
2010-06-08 19:29 ` Jarek Poplawski
2010-06-08 19:45 ` Eric Dumazet
2010-06-08 20:24 ` Jarek Poplawski
2010-06-08 20:52 ` Eric Dumazet
2010-06-08 21:18 ` Jarek Poplawski
2010-06-09 6:13 ` pkt_sched: gen_estimator: more fuel for Jarek and Changli Eric Dumazet
2010-06-09 6:51 ` Jarek Poplawski
2010-06-09 7:36 ` Eric Dumazet
2010-06-09 8:14 ` Jarek Poplawski
2010-06-09 9:40 ` [PATCH] pkt_sched: gen_kill_estimator() rcu fixes Eric Dumazet
2010-06-09 9:56 ` Eric Dumazet
2010-06-09 10:41 ` Jarek Poplawski
2010-06-09 12:09 ` Eric Dumazet
2010-06-09 12:50 ` Jarek Poplawski
2010-06-09 13:05 ` Eric Dumazet
2010-06-12 1:39 ` David Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1275931108.2545.168.camel@edumazet-laptop \
--to=eric.dumazet@gmail.com \
--cc=davem@davemloft.net \
--cc=jarkao2@gmail.com \
--cc=kaber@trash.net \
--cc=netdev@vger.kernel.org \
--cc=shemminger@vyatta.com \
--cc=xiaosuo@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox