tcp: fix overflow in __tcp_retransmit_skb()
authorEric Dumazet <edumazet@google.com>
Thu, 15 Sep 2016 15:12:33 +0000 (08:12 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 15 Nov 2016 06:46:35 +0000 (07:46 +0100)
commita35ce624a3ae06894db065727d33c17aff5cdaec
treee63e2bcc147950b49eeb32ee1eabd55f7f83151e
parentbeb996c1c39990c22248d3afdc23d7d2af13272f
tcp: fix overflow in __tcp_retransmit_skb()

[ Upstream commit ffb4d6c8508657824bcef68a36b2a0f9d8c09d10 ]

If a TCP socket gets a large write queue, an overflow can happen
in a test in __tcp_retransmit_skb() preventing all retransmits.

The flow then stalls and resets after timeouts.

Tested:

sysctl -w net.core.wmem_max=1000000000
netperf -H dest -- -s 1000000000

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/ipv4/tcp_output.c