From 8d3a564da34e5844aca4f991b73f8ca512246b23 Mon Sep 17 00:00:00 2001 From: Doug Leith Date: Tue, 9 Dec 2008 00:13:04 -0800 Subject: tcp: tcp_vegas cong avoid fix This patch addresses a book-keeping issue in tcp_vegas.c. At present tcp_vegas does separate book-keeping of cwnd based on packet sequence numbers. A mismatch can develop between this book-keeping and tp->snd_cwnd due, for example, to delayed acks acking multiple packets. When vegas transitions to reno operation (e.g. following loss), then this mismatch leads to incorrect behaviour (akin to a cwnd backoff). This seems mostly to affect operation at low cwnds where delayed acking can lead to a significant fraction of cwnd being covered by a single ack, leading to the book-keeping mismatch. This patch modifies the congestion avoidance update to avoid the need for separate book-keeping while leaving vegas congestion avoidance functionally unchanged. A secondary advantage of this modification is that the use of fixed-point (via V_PARAM_SHIFT) and 64 bit arithmetic is no longer necessary, simplifying the code. Some example test measurements with the patched code (confirming no functional change in the congestion avoidance algorithm) can be seen at: http://www.hamilton.ie/doug/vegaspatch/ Signed-off-by: Doug Leith Signed-off-by: David S. Miller --- net/ipv4/tcp_vegas.c | 80 +++++++--------------------------------------------- 1 file changed, 10 insertions(+), 70 deletions(-) (limited to 'net/ipv4') diff --git a/net/ipv4/tcp_vegas.c b/net/ipv4/tcp_vegas.c index 7cd22262de3a..a453aac91bd3 100644 --- a/net/ipv4/tcp_vegas.c +++ b/net/ipv4/tcp_vegas.c @@ -40,18 +40,14 @@ #include "tcp_vegas.h" -/* Default values of the Vegas variables, in fixed-point representation - * with V_PARAM_SHIFT bits to the right of the binary point. - */ -#define V_PARAM_SHIFT 1 -static int alpha = 2<beg_snd_nxt)) { /* Do the Vegas once-per-RTT cwnd adjustment. */ - u32 old_wnd, old_snd_cwnd; - - - /* Here old_wnd is essentially the window of data that was - * sent during the previous RTT, and has all - * been acknowledged in the course of the RTT that ended - * with the ACK we just received. Likewise, old_snd_cwnd - * is the cwnd during the previous RTT. - */ - old_wnd = (vegas->beg_snd_nxt - vegas->beg_snd_una) / - tp->mss_cache; - old_snd_cwnd = vegas->beg_snd_cwnd; /* Save the extent of the current window so we can use this * at the end of the next RTT. */ - vegas->beg_snd_una = vegas->beg_snd_nxt; vegas->beg_snd_nxt = tp->snd_nxt; - vegas->beg_snd_cwnd = tp->snd_cwnd; /* We do the Vegas calculations only if we got enough RTT * samples that we can be reasonably sure that we got @@ -252,22 +212,14 @@ static void tcp_vegas_cong_avoid(struct sock *sk, u32 ack, u32 in_flight) * * This is: * (actual rate in segments) * baseRTT - * We keep it as a fixed point number with - * V_PARAM_SHIFT bits to the right of the binary point. */ - target_cwnd = ((u64)old_wnd * vegas->baseRTT); - target_cwnd <<= V_PARAM_SHIFT; - do_div(target_cwnd, rtt); + target_cwnd = tp->snd_cwnd * vegas->baseRTT / rtt; /* Calculate the difference between the window we had, * and the window we would like to have. This quantity * is the "Diff" from the Arizona Vegas papers. - * - * Again, this is a fixed point number with - * V_PARAM_SHIFT bits to the right of the binary - * point. */ - diff = (old_wnd << V_PARAM_SHIFT) - target_cwnd; + diff = tp->snd_cwnd * (rtt-vegas->baseRTT) / vegas->baseRTT; if (diff > gamma && tp->snd_ssthresh > 2 ) { /* Going too fast. Time to slow down @@ -282,16 +234,13 @@ static void tcp_vegas_cong_avoid(struct sock *sk, u32 ack, u32 in_flight) * truncation robs us of full link * utilization. */ - tp->snd_cwnd = min(tp->snd_cwnd, - ((u32)target_cwnd >> - V_PARAM_SHIFT)+1); + tp->snd_cwnd = min(tp->snd_cwnd, (u32)target_cwnd+1); } else if (tp->snd_cwnd <= tp->snd_ssthresh) { /* Slow start. */ tcp_slow_start(tp); } else { /* Congestion avoidance. */ - u32 next_snd_cwnd; /* Figure out where we would like cwnd * to be. @@ -300,26 +249,17 @@ static void tcp_vegas_cong_avoid(struct sock *sk, u32 ack, u32 in_flight) /* The old window was too fast, so * we slow down. */ - next_snd_cwnd = old_snd_cwnd - 1; + tp->snd_cwnd--; } else if (diff < alpha) { /* We don't have enough extra packets * in the network, so speed up. */ - next_snd_cwnd = old_snd_cwnd + 1; + tp->snd_cwnd++; } else { /* Sending just as fast as we * should be. */ - next_snd_cwnd = old_snd_cwnd; } - - /* Adjust cwnd upward or downward, toward the - * desired value. - */ - if (next_snd_cwnd > tp->snd_cwnd) - tp->snd_cwnd++; - else if (next_snd_cwnd < tp->snd_cwnd) - tp->snd_cwnd--; } if (tp->snd_cwnd < 2) -- cgit v1.2.1 From be70ed189bc0d16e1609a1c6c04ec9418b4dd11a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 15 Dec 2008 00:19:14 -0800 Subject: netfilter: update rwlock initialization for nat_table The commit e099a173573ce1ba171092aee7bb3c72ea686e59 (netfilter: netns nat: per-netns NAT table) renamed the nat_table from __nat_table to nat_table without updating the __RW_LOCK_UNLOCKED(__nat_table.lock). Signed-off-by: Steven Rostedt Signed-off-by: David S. Miller --- net/ipv4/netfilter/nf_nat_rule.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv4') diff --git a/net/ipv4/netfilter/nf_nat_rule.c b/net/ipv4/netfilter/nf_nat_rule.c index bea54a685109..8d489e746b21 100644 --- a/net/ipv4/netfilter/nf_nat_rule.c +++ b/net/ipv4/netfilter/nf_nat_rule.c @@ -61,7 +61,7 @@ static struct static struct xt_table nat_table = { .name = "nat", .valid_hooks = NAT_VALID_HOOKS, - .lock = __RW_LOCK_UNLOCKED(__nat_table.lock), + .lock = __RW_LOCK_UNLOCKED(nat_table.lock), .me = THIS_MODULE, .af = AF_INET, }; -- cgit v1.2.1