From: Eric Dumazet Date: Wed, 28 Jan 2015 13:47:11 +0000 (-0800) Subject: tcp: ipv4: initialize unicast_sock sk_pacing_rate X-Git-Url: http://git.mmlx.us/?a=commitdiff_plain;h=23990c29a7bc207bfcb3026e286e804c7cdee933;p=linux-edison.git tcp: ipv4: initialize unicast_sock sk_pacing_rate [ Upstream commit 811230cd853d62f09ed0addd0ce9a1b9b0e13fb5 ] When I added sk_pacing_rate field, I forgot to initialize its value in the per cpu unicast_sock used in ip_send_unicast_reply() This means that for sch_fq users, RST packets, or ACK packets sent on behalf of TIME_WAIT sockets might be sent to slowly or even dropped once we reach the per flow limit. Signed-off-by: Eric Dumazet Fixes: 95bd09eb2750 ("tcp: TSO packets automatic sizing") Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman --- diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index 22fa05e041e..ee104cbcefd 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -1465,6 +1465,7 @@ static DEFINE_PER_CPU(struct inet_sock, unicast_sock) = { .sk_wmem_alloc = ATOMIC_INIT(1), .sk_allocation = GFP_ATOMIC, .sk_flags = (1UL << SOCK_USE_WRITE_QUEUE), + .sk_pacing_rate = ~0U, }, .pmtudisc = IP_PMTUDISC_WANT, .uc_ttl = -1,