Skip to contentSkip to conversation

Fix locking in ip_conntrack_change_expect()

PATCH Applied Harald Welte 2 participants 5 messages Oct 14, 2002

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Fix locking in ip_conntrack_change_expect()

Hi again, Dave!

Below is an important bugfix for netfilter (by Martin Josefsson).

In ip_conntrack_change_expect() we were at least need to grab a readlock before comparing the tuple/mask of an ip_conntrack_expect.

Pleae note that the bug was introduced in early 2.4.20-pre time, as newnat was merged. So there's no need to worry about old (stable) kernel releases.

Please apply the following fix, Thanks.

- Harald

net/ipv4/netfilter/ip_conntrack_core.c renamed from net/ipv4/netfilter/ip_conntrack_core.c.mjufs +9 -7
@@ -1060,7 +1060,10 @@
 int ip_conntrack_change_expect(struct ip_conntrack_expect *expect,
 			       struct ip_conntrack_tuple *newtuple)
 {
+	int ret;
+
 	MUST_BE_READ_LOCKED(&ip_conntrack_lock);
+	WRITE_LOCK(&ip_conntrack_expect_tuple_lock);
 
 	DEBUGP("change_expect:\n");
 	DEBUGP("exp tuple: "); DUMP_TUPLE(&expect->tuple);
@@ -1073,26 +1076,25 @@
 		    && LIST_FIND(&ip_conntrack_expect_list, expect_clash,
 			         struct ip_conntrack_expect *, newtuple, &expect->mask)) {
 			/* Force NAT to find an unused tuple */
-			return -1;
+			ret = -1;
 		} else {
-			WRITE_LOCK(&ip_conntrack_expect_tuple_lock);
 			memcpy(&expect->ct_tuple, &expect->tuple, sizeof(expect->tuple));
 			memcpy(&expect->tuple, newtuple, sizeof(expect->tuple));
-			WRITE_UNLOCK(&ip_conntrack_expect_tuple_lock);
-			return 0;
+			ret = 0;
 		}
 	} else {
 		/* Resent packet */
 		DEBUGP("change expect: resent packet\n");
 		if (ip_ct_tuple_equal(&expect->tuple, newtuple)) {
-			return 0;
+			ret = 0;
 		} else {
 			/* Force NAT to choose again the same port */
-			return -1;
+			ret = -1;
 		}
 	}
+	WRITE_UNLOCK(&ip_conntrack_expect_tuple_lock);
 	
-	return -1;
+	return ret;
 }
 
 /* Alter reply tuple (maybe alter helper).  If it's already taken,
Signature · 5 lines
Live long and prosper
- Harald Welte / [email protected]               http://www.gnumonks.org/
============================================================================
"If this were a dictatorship, it'd be a heck of a lot easier, just so long
 as I'm the dictator."  --  George W. Bush Dec 18, 2000
#1 lore raw

From: Harald Welte <[email protected]>
Date: Mon, 14 Oct 2002 13:40:47 +0200

Pleae note that the bug was introduced in early 2.4.20-pre time, as newnat was merged. So there's no need to worry about old (stable) kernel releases.

Please apply the following fix, Thanks.

Applied.

I guess this means I really should push this stuff in for 2.4.20-preX

#2 lore raw
Quoted 4 lines · David Miller

Applied.

I guess this means I really should push this stuff in for 2.4.20-preX

yes, this is the only one which should be submitted before 2.4.20-final.

Thanks.

Signature · 5 lines
Live long and prosper
- Harald Welte / [email protected]               http://www.gnumonks.org/
============================================================================
"If this were a dictatorship, it'd be a heck of a lot easier, just so long
 as I'm the dictator."  --  George W. Bush Dec 18, 2000
#3 lore raw

From: Harald Welte <[email protected]>
Date: Mon, 14 Oct 2002 16:03:09 +0200

Quoted 2 lines · David Miller

I guess this means I really should push this stuff in for 2.4.20-preX

yes, this is the only one which should be submitted before 2.4.20-final.

There were two locking fixes Harald :-)

What I want to push to Marcelo is:

1) Avoid nesting readlocks in conntrack
2) Fix ip_conntrack_change_expect locking
3) Display ipv6 SIT encapsulation properly

Ok?

#4 lore raw
Quoted 11 lines · David Miller

From: Harald Welte <[email protected]>
Date: Mon, 14 Oct 2002 16:03:09 +0200

Quoted 2 lines · David Miller

I guess this means I really should push this stuff in for 2.4.20-preX

yes, this is the only one which should be submitted before 2.4.20-final.

There were two locking fixes Harald :-)

Well, according to my definition of a locking fix (for a locking bug) there only was one. Having a nested readlock is not a locking bug. It was just the debug statements.

Anyway, it's not worth discussing this further.

Quoted 7 lines

What I want to push to Marcelo is:

1) Avoid nesting readlocks in conntrack
2) Fix ip_conntrack_change_expect locking
3) Display ipv6 SIT encapsulation properly

Ok?

Perfect, thanks. The only harmful bug is number 2.

Signature · 5 lines
Live long and prosper
- Harald Welte / [email protected]               http://www.gnumonks.org/
============================================================================
"If this were a dictatorship, it'd be a heck of a lot easier, just so long
 as I'm the dictator."  --  George W. Bush Dec 18, 2000