Skip to contentSkip to conversation

Missing device binding relating to tcp_v4_send_reset?

PATCH Ben Greear 2 participants 3 messages Aug 9, 2010 – Aug 10, 2010

Subsystems (3)
Networking Network drivers TCP
lore
Share & tools
lore

Ben Greear

author
#0 lore raw

This snippet is from some patches Patrick did for me some time back. I think the rest of his work has been merged upstream, but this patch was not. I'm honestly not sure if it's needed or not, but we've been running with it for at least a year or so and it's been working fine for us.

net/ipv4/tcp_ipv4.c +1 -0
index 0207662..1af47db 100644
@@ -641,6 +641,7 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
 				      arg.iov[0].iov_len, IPPROTO_TCP, 0);
 	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
 	arg.flags = (sk && inet_sk(sk)->transparent) ? IP_REPLY_ARG_NOSRCCHECK : 0;
+	arg.bound_dev_if = skb_rtable(skb)->fl.iif;
 
 	net = dev_net(skb_dst(skb)->dev);
 	ip_send_reply(net->ipv4.tcp_sock, skb,

Thanks,
Ben

Signature · 2 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com
maintainer
#1 lore raw

From: Ben Greear <[email protected]>
Date: Mon, 09 Aug 2010 10:46:26 -0700

Quoted 5 lines

This snippet is from some patches Patrick did for me some time back. I think the rest of his work has been merged upstream, but this patch was not. I'm honestly not sure if it's needed or not, but we've been running with it for at least a year or so and it's been working fine for us.

There is no reason at all to force a reply to a packet to go out of the same interface as the packet which triggered that reply arrived upon.

author
#2 lore raw
Quoted 9 lines · David Miller

From: Ben Greear<[email protected]>
Date: Mon, 09 Aug 2010 10:46:26 -0700

Quoted 5 lines

This snippet is from some patches Patrick did for me some time back. I think the rest of his work has been merged upstream, but this patch was not. I'm honestly not sure if it's needed or not, but we've been running with it for at least a year or so and it's been working fine for us.

Quoted 3 lines

There is no reason at all to force a reply to a packet to go out of the same interface as the packet which triggered that reply arrived upon.

Ok, in my case I *did* want this behaviour..it makes send-to-self, BIND_TO_DEVICE and virtual routers and such work properly if I recall correctly.

Would you consider a sysctl to enable this feature, with it disabled by default? Please note there are similar sysctls for ICMP already...

Thanks,
Ben

Signature · 2 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com