Patchwork [2/2] tcp: miscounts due to tcp_fragment pcount reset

login
register
mail settings
Submitter Ilpo Järvinen
Date April 2, 2009, 9:18 a.m.
Message ID <Pine.LNX.4.64.0904021216170.30150@wrl-59.cs.helsinki.fi>
Download mbox | patch
Permalink /patch/25524/
State Accepted
Delegated to: David Miller
Headers show

Comments

Ilpo Järvinen - April 2, 2009, 9:18 a.m.
It seems that trivial reset of pcount to one was not sufficient
in tcp_retransmit_skb. Multiple counters experience a positive
miscount when skb's pcount gets lowered without the necessary
adjustments (depending on skb's sacked bits which exactly), at
worst a packets_out miscount can crash at RTO if the write queue
is empty!

Triggering this requires mss change, so bidir tcp or mtu probe or
like.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Tested-by: Uwe Bugla <uwe.bugla@gmx.de>
---

Markus probably tested this as well though I didn't hear any reports back 
yet.

 net/ipv4/tcp_output.c |    7 ++++++-
 1 files changed, 6 insertions(+), 1 deletions(-)
David Miller - April 2, 2009, 11:32 p.m.
From: "Ilpo Järvinen" <ilpo.jarvinen@helsinki.fi>
Date: Thu, 2 Apr 2009 12:18:20 +0300 (EEST)

> 
> It seems that trivial reset of pcount to one was not sufficient
> in tcp_retransmit_skb. Multiple counters experience a positive
> miscount when skb's pcount gets lowered without the necessary
> adjustments (depending on skb's sacked bits which exactly), at
> worst a packets_out miscount can crash at RTO if the write queue
> is empty!
> 
> Triggering this requires mss change, so bidir tcp or mtu probe or
> like.
> 
> Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
> Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
> Tested-by: Uwe Bugla <uwe.bugla@gmx.de>

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

Patch

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index f1db89b..53300fa 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1893,7 +1893,12 @@  int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
 		if (tcp_fragment(sk, skb, cur_mss, cur_mss))
 			return -ENOMEM; /* We'll try again later. */
 	} else {
-		tcp_init_tso_segs(sk, skb, cur_mss);
+		int oldpcount = tcp_skb_pcount(skb);
+
+		if (unlikely(oldpcount > 1)) {
+			tcp_init_tso_segs(sk, skb, cur_mss);
+			tcp_adjust_pcount(sk, skb, oldpcount - tcp_skb_pcount(skb));
+		}
 	}
 
 	tcp_retrans_try_collapse(sk, skb, cur_mss);