Allow per-route window scale limiting

There are black box devices out there, routers and firewalls and
whatnot, that simply cannot grok the TCP window scaling option
correctly.

People should and do bark at the site running the device causing
the problems, but in the mean time folks do want a way to deal
with the problem.  We don't want them to turn off window scaling
completely as that hurts performance of connections that would run
just fine with window scaling enabled.

So give a way to do this on a per-route basis by limiting the
window scaling by the per-connection window clamp.  Stephen's
changelog message explains how to do this using a route metric.

[TCP]: Limit window scaling if window is clamped.

This small change allows for easy per-route workarounds for broken hosts or
middleboxes that are not compliant with TCP standards for window scaling.
Rather than having to turn off window scaling globally. This patch allows
reducing or disabling window scaling if window clamp is present.

Example: Mark Lord reported a problem with 2.6.17 kernel being unable to
access http://www.everymac.com

# ip route add 216.145.246.23/32 via 10.8.0.1 window 65535

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

1 files changed, 1 insertions, 0 deletions

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index f33c9dddaa12..e0657b96019c 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -197,6 +197,7 @@ void tcp_select_initial_window(int __space, __u32 mss,
 		 * See RFC1323 for an explanation of the limit to 14 
 		 */
 		space = max_t(u32, sysctl_tcp_rmem[2], sysctl_rmem_max);
+		space = min_t(u32, space, *window_clamp);
 		while (space > 65535 && (*rcv_wscale) < 14) {
 			space >>= 1;
 			(*rcv_wscale)++;