From: Tomas Szepe <szepe@pinerecords.com>
To: davem@redhat.com
Cc: linux-kernel@vger.kernel.org, netdev@oss.sgi.com, grof@dragon.cz
Subject: possible bug in tcp_input.c
Date: Fri, 24 Oct 2003 18:29:59 +0200 [thread overview]
Message-ID: <20031024162959.GB11154@louise.pinerecords.com> (raw)
Hi David,
We came up with the attached patch during a hectic oops tracing session
that got started by our sysadmin writing down an oops using the pen & paper
method, no less. The crashing machine has been a firewall running a very
unusual NAT + QoS configuration, however we believe that we might have
discovered a real bug in 2.4's tcp_input.c. Since our insight into the
internals of the tcp/ip stack is far from even basic, we are seeking your
opinion on whether we are correct.
The static inline function skb_peek() as defined in include/linux/skbuff.h
returns a pointer to a sk_buff, or NULL when its argument is an empty list
or a pointer to the head element. Since this is documented behavior, it is
not surprising that all segments of code within tcp_input.c dealing with
a return from skb_peek() take care not to dereference the returned pointer
if it happens to be NULL. There is an exception, though:
/* tcp_input.c, line 1138 */
static inline int tcp_head_timedout(struct sock *sk, struct tcp_opt *tp)
{
return tp->packets_out && tcp_skb_timedout(tp, skb_peek(&sk->write_queue));
}
The passed NULL (and yes, this is where we are getting one) is dereferenced
immediately in:
/* tcp_input.c, line 1133 */
static inline int tcp_skb_timedout(struct tcp_opt *tp, struct sk_buff *skb)
{
return (tcp_time_stamp - TCP_SKB_CB(skb)->when > tp->rto);
}
with TCP_SKB_CB that is defined as
/* tcp.h, line 1034 */
#define TCP_SKB_CB(__skb) ((struct tcp_skb_cb *)&((__skb)->cb[0]))
We are proposing to cure the problem by adding a simple check in
tcp_head_timedout(), but are not sure whether this is the right
thing to do, because as a friend put it, we seem to be fixing
a leaking faucet in a god damn power plant.
Thanks for any help,
--
Tomas Szepe <szepe@pinerecords.com>
diff -urN a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
--- a/net/ipv4/tcp_input.c 2003-06-13 16:51:39 +0200
+++ b/net/ipv4/tcp_input.c 2003-10-24 17:41:19 +0200
@@ -1138,7 +1138,11 @@
static inline int tcp_head_timedout(struct sock *sk, struct tcp_opt *tp)
{
- return tp->packets_out && tcp_skb_timedout(tp, skb_peek(&sk->write_queue));
+ struct sk_buff *skb = skb_peek(&sk->write_queue);
+ if (skb == NULL)
+ return 1;
+
+ return tp->packets_out && tcp_skb_timedout(tp, skb);
}
/* Linux NewReno/SACK/FACK/ECN state machine.
next reply other threads:[~2003-10-24 16:29 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-10-24 16:29 Tomas Szepe [this message]
2003-10-25 2:30 ` possible bug in tcp_input.c David S. Miller
2003-10-25 9:12 ` Tomas Szepe
2003-10-26 6:55 ` Tomas Szepe
2003-10-27 6:33 ` David S. Miller
[not found] <20031024162959.GB11154@louise.pinerecords.com.suse.lists.linux.kernel>
2003-10-24 17:57 ` Andi Kleen
2003-11-18 13:58 ` Tomas Szepe
2003-11-18 14:01 ` Andi Kleen
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=20031024162959.GB11154@louise.pinerecords.com \
--to=szepe@pinerecords.com \
--cc=davem@redhat.com \
--cc=grof@dragon.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@oss.sgi.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;
as well as URLs for NNTP newsgroup(s).