Skip to contentSkip to conversation

ipv6: Fix tcp_v6_send_response(): it didn't set skb transport header

PATCH Applied Cosmin Ratiu 2 participants 2 messages Sep 3, 2009 – Sep 4, 2009

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipv6: Fix tcp_v6_send_response(): it didn't set skb transport header

Hello,

Here is a patch which fixes an issue observed when using TCP over IPv6 and AH from IPsec.

When a connection gets closed the 4-way method and the last ACK from the server gets dropped, the subsequent FINs from the client do not get ACKed because tcp_v6_send_response does not set the transport header pointer. This causes ah6_output to try to allocate a lot of memory, which typically fails, so the ACKs never make it out of the stack.

I have reproduced the problem on kernel 2.6.7, but after looking at the latest kernel it seems the problem is still there.

Cosmin.

---
 net/ipv6/tcp_ipv6.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
net/ipv6/tcp_ipv6.c +1 -0
index d849dd5..776e911 100644
@@ -1003,6 +1003,7 @@ static void tcp_v6_send_response(struct sk_buff *skb, u32 seq, u32 ack, u32 win,
 	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
 
 	t1 = (struct tcphdr *) skb_push(buff, tot_len);
+	skb_reset_transport_header(skb);
 
 	/* Swap the send and the receive. */
 	memset(t1, 0, sizeof(*t1));
maintainer Applied
#1 lore raw

From: Cosmin Ratiu <[email protected]>
Date: Thu, 3 Sep 2009 19:25:53 +0300

Quoted 11 lines

Hello,

Here is a patch which fixes an issue observed when using TCP over IPv6 and AH from IPsec.

When a connection gets closed the 4-way method and the last ACK from the server gets dropped, the subsequent FINs from the client do not get ACKed because tcp_v6_send_response does not set the transport header pointer. This causes ah6_output to try to allocate a lot of memory, which typically fails, so the ACKs never make it out of the stack.

Quoted 4 lines

I have reproduced the problem on kernel 2.6.7, but after looking at the latest kernel it seems the problem is still there.

Great catch. IPV4 handles this transparently via ip_append_data() which is what it uses to send these kinds of frames. IPV6 does this stuff by hand.

Applied to net-next-2.6, thanks!