netdev - [PATCH net-next 2/2] net/sched: act_csum: don't use spinlock in the fast path

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <16125b3f16fb9f8c068ec80ce3f6550d0465521c.1513104506.git.dcaratti@redhat.com>
Date:   Wed, 13 Dec 2017 10:48:38 +0100
From:   Davide Caratti <dcaratti@...hat.com>
To:     Cong Wang <xiyou.wangcong@...il.com>,
        Jiri Pirko <jiri@...lanox.com>,
        "David S . Miller" <davem@...emloft.net>
Cc:     netdev@...r.kernel.org
Subject: [PATCH net-next 2/2] net/sched: act_csum: don't use spinlock in the fast path

there is no need for using spin_{,unlock}_bh() to protect simultaneous
read/write of act_csum configuration: it's sufficient to ensure that RTNL
lock has been taken when tcf_action and update_flags are written, so that
configuration is not racy. Then, in the data path, use READ_ONCE() to
read those values, to avoid lock contention among multiple readers.

Signed-off-by: Davide Caratti <dcaratti@...hat.com>
---
 net/sched/act_csum.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/net/sched/act_csum.c b/net/sched/act_csum.c
index 22a555ba3985..ac8402d53cd5 100644
--- a/net/sched/act_csum.c
+++ b/net/sched/act_csum.c
@@ -80,10 +80,9 @@ static int tcf_csum_init(struct net *net, struct nlattr *nla,
 	}
 
 	p = to_tcf_csum(*a);
-	spin_lock_bh(&p->tcf_lock);
+	ASSERT_RTNL();
 	p->tcf_action = parm->action;
 	p->update_flags = parm->update_flags;
-	spin_unlock_bh(&p->tcf_lock);
 
 	if (ret == ACT_P_CREATED)
 		tcf_idr_insert(tn, *a);
@@ -544,14 +543,12 @@ static int tcf_csum(struct sk_buff *skb, const struct tc_action *a,
 
 	tcf_lastuse_update(&p->tcf_tm);
 	bstats_cpu_update(this_cpu_ptr(p->common.cpu_bstats), skb);
-	spin_lock(&p->tcf_lock);
-	action = p->tcf_action;
-	update_flags = p->update_flags;
-	spin_unlock(&p->tcf_lock);
 
+	action = READ_ONCE(p->tcf_action);
 	if (unlikely(action == TC_ACT_SHOT))
 		goto drop;
 
+	update_flags = READ_ONCE(p->update_flags);
 	switch (tc_skb_protocol(skb)) {
 	case cpu_to_be16(ETH_P_IP):
 		if (!tcf_csum_ipv4(skb, update_flags))
-- 
2.13.6