diff mbox

[net] act_pedit: check binding before calling tcf_hash_release()

Message ID 1438301541-26192-2-git-send-email-xiyou.wangcong@gmail.com
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Cong Wang July 31, 2015, 12:12 a.m. UTC
When we share an action within a filter, the bind refcnt
should increase, therefore we should not call tcf_hash_release().

Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Cong Wang <cwang@twopensource.com>
---
 net/sched/act_pedit.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

Comments

Daniel Borkmann July 31, 2015, 11:10 a.m. UTC | #1
On 07/31/2015 02:12 AM, Cong Wang wrote:
> When we share an action within a filter, the bind refcnt
> should increase, therefore we should not call tcf_hash_release().
>
> Cc: Jamal Hadi Salim <jhs@mojatatu.com>
> Cc: Daniel Borkmann <daniel@iogearbox.net>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
> Signed-off-by: Cong Wang <cwang@twopensource.com>

Seems like this slipped in via commit 1a29321ed045. The ugly thing is that
this leads to a use-after-free.

# tc actions add action pedit munge offset 2 u16 at 0 0f0000000 22 set 11500
# tc actions show action pedit
   action order 0:  pedit action pass keys 1
   index 1 ref 1 bind 0
   key #0  at 0: val 00002cec mask ffff0000
# tc filter add dev foo parent 1: bpf bytecode "$FOO" flowid 1:1 action pedit index 1
# tc filter show dev foo
   filter parent 1: protocol all pref 49152 bpf
   filter parent 1: protocol all pref 49152 bpf handle 0x1 flowid 1:1 bytecode '1,6 0 0 4294967295'
   action order 1: pedit action pass keys 1
   index 1 ref 1 bind 0
   key #0  at 0: val 00002cec mask ffff0000
# tc actions del action pedit index 1

.... and now you can wait for the next egress packet. ;)

Thanks for the fix Cong!

Fixes: 1a29321ed045 ("net_sched: act: Dont increment refcnt on replace")
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller July 31, 2015, 10:23 p.m. UTC | #2
From: Cong Wang <xiyou.wangcong@gmail.com>
Date: Thu, 30 Jul 2015 17:12:21 -0700

> When we share an action within a filter, the bind refcnt
> should increase, therefore we should not call tcf_hash_release().
> 
> Cc: Jamal Hadi Salim <jhs@mojatatu.com>
> Cc: Daniel Borkmann <daniel@iogearbox.net>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
> Signed-off-by: Cong Wang <cwang@twopensource.com>

Applied.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/net/sched/act_pedit.c b/net/sched/act_pedit.c
index 17e6d66..ff8b466 100644
--- a/net/sched/act_pedit.c
+++ b/net/sched/act_pedit.c
@@ -68,13 +68,12 @@  static int tcf_pedit_init(struct net *net, struct nlattr *nla,
 		}
 		ret = ACT_P_CREATED;
 	} else {
-		p = to_pedit(a);
-		tcf_hash_release(a, bind);
 		if (bind)
 			return 0;
+		tcf_hash_release(a, bind);
 		if (!ovr)
 			return -EEXIST;
-
+		p = to_pedit(a);
 		if (p->tcfp_nkeys && p->tcfp_nkeys != parm->nkeys) {
 			keys = kmalloc(ksize, GFP_KERNEL);
 			if (keys == NULL)