tcp md5 fraglist...

tcp md5 fraglist...

[David Miller]

Eric, would you please formally submit that tcp md5 fraglist
fix to net/ipv4/tcp.c with proper signoffs?

Thanks!

Re: tcp md5 fraglist...

[Eric Dumazet]

Le lundi 17 mai 2010 à 23:14 -0700, David Miller a écrit :
> Eric, would you please formally submit that tcp md5 fraglist
> fix to net/ipv4/tcp.c with proper signoffs?
> 
> Thanks!

Sure, will do, but I doubt it is a real problem, because GRO or RSC
should not coalesce two segments with MD5 signature (they are guaranteed
to differ for each segment ?)

Re: tcp md5 fraglist...

[David Miller]

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 18 May 2010 08:30:24 +0200

> Le lundi 17 mai 2010 à 23:14 -0700, David Miller a écrit :
>> Eric, would you please formally submit that tcp md5 fraglist
>> fix to net/ipv4/tcp.c with proper signoffs?
>> 
>> Thanks!
> 
> Sure, will do, but I doubt it is a real problem, because GRO or RSC
> should not coalesce two segments with MD5 signature (they are guaranteed
> to differ for each segment ?)

Right, and I just checked that GRO does in fact check if the TCP
options are different in any way.

[PATCH] tcp: tcp_md5_hash_skb_data() frag_list handling

[Eric Dumazet]

Le lundi 17 mai 2010 à 23:14 -0700, David Miller a écrit :
> Eric, would you please formally submit that tcp md5 fraglist
> fix to net/ipv4/tcp.c with proper signoffs?

Here it is, thanks !


[PATCH] tcp: tcp_md5_hash_skb_data() frag_list handling

tcp_md5_hash_skb_data() should handle skb->frag_list, and eventually
recurse.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
 net/ipv4/tcp.c |    5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 6596b4f..49d0d2b 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2999,6 +2999,7 @@ int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp,
 	const unsigned head_data_len = skb_headlen(skb) > header_len ?
 				       skb_headlen(skb) - header_len : 0;
 	const struct skb_shared_info *shi = skb_shinfo(skb);
+	struct sk_buff *frag_iter;
 
 	sg_init_table(&sg, 1);
 
@@ -3013,6 +3014,10 @@ int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp,
 			return 1;
 	}
 
+	skb_walk_frags(skb, frag_iter)
+		if (tcp_md5_hash_skb_data(hp, frag_iter, 0))
+			return 1;
+
 	return 0;
 }
 

Re: [PATCH] tcp: tcp_md5_hash_skb_data() frag_list handling

[David Miller]

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 18 May 2010 08:40:51 +0200

> Le lundi 17 mai 2010 à 23:14 -0700, David Miller a écrit :
>> Eric, would you please formally submit that tcp md5 fraglist
>> fix to net/ipv4/tcp.c with proper signoffs?
> 
> Here it is, thanks !
> 
> 
> [PATCH] tcp: tcp_md5_hash_skb_data() frag_list handling
> 
> tcp_md5_hash_skb_data() should handle skb->frag_list, and eventually
> recurse.
> 
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>

Applied.