From 79320d7e14900c549c3520791a297328f53ff71e Mon Sep 17 00:00:00 2001 From: Aki M Nyrhinen Date: Sun, 11 Jun 2006 21:18:56 -0700 Subject: [TCP]: continued: reno sacked_out count fix From: Aki M Nyrhinen IMHO the current fix to the problem (in_flight underflow in reno) is incorrect. it treats the symptons but ignores the problem. the problem is timing out packets other than the head packet when we don't have sack. i try to explain (sorry if explaining the obvious). with sack, scanning the retransmit queue for timed out packets is fine because we know which packets in our retransmit queue have been acked by the receiver. without sack, we know only how many packets in our retransmit queue the receiver has acknowledged, but no idea which packets. think of a "typical" slow-start overshoot case, where for example every third packet in a window get lost because a router buffer gets full. with sack, we check for timeouts on those every third packet (as the rest have been sacked). the packet counting works out and if there is no reordering, we'll retransmit exactly the packets that were lost. without sack, however, we check for timeout on every packet and end up retransmitting consecutive packets in the retransmit queue. in our slow-start example, 2/3 of those retransmissions are unnecessary. these unnecessary retransmissions eat the congestion window and evetually prevent fast recovery from continuing, if enough packets were lost. Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 4a538bc1683d..b5521a9d3dc1 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1649,7 +1649,7 @@ static void tcp_update_scoreboard(struct sock *sk, struct tcp_sock *tp) * Hence, we can detect timed out packets during fast * retransmit without falling to slow start. */ - if (tcp_head_timedout(sk, tp)) { + if (!IsReno(tp) && tcp_head_timedout(sk, tp)) { struct sk_buff *skb; skb = tp->scoreboard_skb_hint ? tp->scoreboard_skb_hint @@ -1662,8 +1662,6 @@ static void tcp_update_scoreboard(struct sock *sk, struct tcp_sock *tp) if (!(TCP_SKB_CB(skb)->sacked&TCPCB_TAGBITS)) { TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; tp->lost_out += tcp_skb_pcount(skb); - if (IsReno(tp)) - tcp_remove_reno_sacks(sk, tp, tcp_skb_pcount(skb) + 1); /* clear xmit_retrans hint */ if (tp->retransmit_skb_hint && -- cgit v1.2.3 From 1a2449a87bb7606113b1aa1a9d3c3e78ef189a1c Mon Sep 17 00:00:00 2001 From: Chris Leech Date: Tue, 23 May 2006 18:05:53 -0700 Subject: [I/OAT]: TCP recv offload to I/OAT Locks down user pages and sets up for DMA in tcp_recvmsg, then calls dma_async_try_early_copy in tcp_v4_do_rcv Signed-off-by: Chris Leech Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 103 +++++++++++++++++++++++++++++++++++++++++++++------ net/ipv4/tcp_input.c | 74 ++++++++++++++++++++++++++++++++---- net/ipv4/tcp_ipv4.c | 18 ++++++++- net/ipv6/tcp_ipv6.c | 12 +++++- 4 files changed, 185 insertions(+), 22 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 4e067d25a63c..ff6ccda9ff46 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -263,7 +263,7 @@ #include #include #include - +#include #include #include @@ -1110,6 +1110,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, int target; /* Read at least this many bytes */ long timeo; struct task_struct *user_recv = NULL; + int copied_early = 0; lock_sock(sk); @@ -1133,6 +1134,17 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, target = sock_rcvlowat(sk, flags & MSG_WAITALL, len); +#ifdef CONFIG_NET_DMA + tp->ucopy.dma_chan = NULL; + preempt_disable(); + if ((len > sysctl_tcp_dma_copybreak) && !(flags & MSG_PEEK) && + !sysctl_tcp_low_latency && __get_cpu_var(softnet_data.net_dma)) { + preempt_enable_no_resched(); + tp->ucopy.pinned_list = dma_pin_iovec_pages(msg->msg_iov, len); + } else + preempt_enable_no_resched(); +#endif + do { struct sk_buff *skb; u32 offset; @@ -1274,6 +1286,10 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, } else sk_wait_data(sk, &timeo); +#ifdef CONFIG_NET_DMA + tp->ucopy.wakeup = 0; +#endif + if (user_recv) { int chunk; @@ -1329,13 +1345,39 @@ do_prequeue: } if (!(flags & MSG_TRUNC)) { - err = skb_copy_datagram_iovec(skb, offset, - msg->msg_iov, used); - if (err) { - /* Exception. Bailout! */ - if (!copied) - copied = -EFAULT; - break; +#ifdef CONFIG_NET_DMA + if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list) + tp->ucopy.dma_chan = get_softnet_dma(); + + if (tp->ucopy.dma_chan) { + tp->ucopy.dma_cookie = dma_skb_copy_datagram_iovec( + tp->ucopy.dma_chan, skb, offset, + msg->msg_iov, used, + tp->ucopy.pinned_list); + + if (tp->ucopy.dma_cookie < 0) { + + printk(KERN_ALERT "dma_cookie < 0\n"); + + /* Exception. Bailout! */ + if (!copied) + copied = -EFAULT; + break; + } + if ((offset + used) == skb->len) + copied_early = 1; + + } else +#endif + { + err = skb_copy_datagram_iovec(skb, offset, + msg->msg_iov, used); + if (err) { + /* Exception. Bailout! */ + if (!copied) + copied = -EFAULT; + break; + } } } @@ -1355,15 +1397,19 @@ skip_copy: if (skb->h.th->fin) goto found_fin_ok; - if (!(flags & MSG_PEEK)) - sk_eat_skb(sk, skb, 0); + if (!(flags & MSG_PEEK)) { + sk_eat_skb(sk, skb, copied_early); + copied_early = 0; + } continue; found_fin_ok: /* Process the FIN. */ ++*seq; - if (!(flags & MSG_PEEK)) - sk_eat_skb(sk, skb, 0); + if (!(flags & MSG_PEEK)) { + sk_eat_skb(sk, skb, copied_early); + copied_early = 0; + } break; } while (len > 0); @@ -1386,6 +1432,36 @@ skip_copy: tp->ucopy.len = 0; } +#ifdef CONFIG_NET_DMA + if (tp->ucopy.dma_chan) { + struct sk_buff *skb; + dma_cookie_t done, used; + + dma_async_memcpy_issue_pending(tp->ucopy.dma_chan); + + while (dma_async_memcpy_complete(tp->ucopy.dma_chan, + tp->ucopy.dma_cookie, &done, + &used) == DMA_IN_PROGRESS) { + /* do partial cleanup of sk_async_wait_queue */ + while ((skb = skb_peek(&sk->sk_async_wait_queue)) && + (dma_async_is_complete(skb->dma_cookie, done, + used) == DMA_SUCCESS)) { + __skb_dequeue(&sk->sk_async_wait_queue); + kfree_skb(skb); + } + } + + /* Safe to free early-copied skbs now */ + __skb_queue_purge(&sk->sk_async_wait_queue); + dma_chan_put(tp->ucopy.dma_chan); + tp->ucopy.dma_chan = NULL; + } + if (tp->ucopy.pinned_list) { + dma_unpin_iovec_pages(tp->ucopy.pinned_list); + tp->ucopy.pinned_list = NULL; + } +#endif + /* According to UNIX98, msg_name/msg_namelen are ignored * on connected socket. I was just happy when found this 8) --ANK */ @@ -1658,6 +1734,9 @@ int tcp_disconnect(struct sock *sk, int flags) __skb_queue_purge(&sk->sk_receive_queue); sk_stream_writequeue_purge(sk); __skb_queue_purge(&tp->out_of_order_queue); +#ifdef CONFIG_NET_DMA + __skb_queue_purge(&sk->sk_async_wait_queue); +#endif inet->dport = 0; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index b5521a9d3dc1..c6d62f0a9966 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -71,6 +71,7 @@ #include #include #include +#include int sysctl_tcp_timestamps = 1; int sysctl_tcp_window_scaling = 1; @@ -3785,6 +3786,50 @@ static inline int tcp_checksum_complete_user(struct sock *sk, struct sk_buff *sk __tcp_checksum_complete_user(sk, skb); } +#ifdef CONFIG_NET_DMA +static int tcp_dma_try_early_copy(struct sock *sk, struct sk_buff *skb, int hlen) +{ + struct tcp_sock *tp = tcp_sk(sk); + int chunk = skb->len - hlen; + int dma_cookie; + int copied_early = 0; + + if (tp->ucopy.wakeup) + return 0; + + if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list) + tp->ucopy.dma_chan = get_softnet_dma(); + + if (tp->ucopy.dma_chan && skb->ip_summed == CHECKSUM_UNNECESSARY) { + + dma_cookie = dma_skb_copy_datagram_iovec(tp->ucopy.dma_chan, + skb, hlen, tp->ucopy.iov, chunk, tp->ucopy.pinned_list); + + if (dma_cookie < 0) + goto out; + + tp->ucopy.dma_cookie = dma_cookie; + copied_early = 1; + + tp->ucopy.len -= chunk; + tp->copied_seq += chunk; + tcp_rcv_space_adjust(sk); + + if ((tp->ucopy.len == 0) || + (tcp_flag_word(skb->h.th) & TCP_FLAG_PSH) || + (atomic_read(&sk->sk_rmem_alloc) > (sk->sk_rcvbuf >> 1))) { + tp->ucopy.wakeup = 1; + sk->sk_data_ready(sk, 0); + } + } else if (chunk > 0) { + tp->ucopy.wakeup = 1; + sk->sk_data_ready(sk, 0); + } +out: + return copied_early; +} +#endif /* CONFIG_NET_DMA */ + /* * TCP receive function for the ESTABLISHED state. * @@ -3901,14 +3946,23 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, } } else { int eaten = 0; + int copied_early = 0; - if (tp->ucopy.task == current && - tp->copied_seq == tp->rcv_nxt && - len - tcp_header_len <= tp->ucopy.len && - sock_owned_by_user(sk)) { - __set_current_state(TASK_RUNNING); + if (tp->copied_seq == tp->rcv_nxt && + len - tcp_header_len <= tp->ucopy.len) { +#ifdef CONFIG_NET_DMA + if (tcp_dma_try_early_copy(sk, skb, tcp_header_len)) { + copied_early = 1; + eaten = 1; + } +#endif + if (tp->ucopy.task == current && sock_owned_by_user(sk) && !copied_early) { + __set_current_state(TASK_RUNNING); - if (!tcp_copy_to_iovec(sk, skb, tcp_header_len)) { + if (!tcp_copy_to_iovec(sk, skb, tcp_header_len)) + eaten = 1; + } + if (eaten) { /* Predicted packet is in window by definition. * seq == rcv_nxt and rcv_wup <= rcv_nxt. * Hence, check seq<=rcv_wup reduces to: @@ -3924,8 +3978,9 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, __skb_pull(skb, tcp_header_len); tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq; NET_INC_STATS_BH(LINUX_MIB_TCPHPHITSTOUSER); - eaten = 1; } + if (copied_early) + tcp_cleanup_rbuf(sk, skb->len); } if (!eaten) { if (tcp_checksum_complete_user(sk, skb)) @@ -3966,6 +4021,11 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, __tcp_ack_snd_check(sk, 0); no_ack: +#ifdef CONFIG_NET_DMA + if (copied_early) + __skb_queue_tail(&sk->sk_async_wait_queue, skb); + else +#endif if (eaten) __kfree_skb(skb); else diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 672950e54c49..25ecc6e2478b 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -71,6 +71,7 @@ #include #include #include +#include #include #include @@ -1091,8 +1092,18 @@ process: bh_lock_sock(sk); ret = 0; if (!sock_owned_by_user(sk)) { - if (!tcp_prequeue(sk, skb)) +#ifdef CONFIG_NET_DMA + struct tcp_sock *tp = tcp_sk(sk); + if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list) + tp->ucopy.dma_chan = get_softnet_dma(); + if (tp->ucopy.dma_chan) ret = tcp_v4_do_rcv(sk, skb); + else +#endif + { + if (!tcp_prequeue(sk, skb)) + ret = tcp_v4_do_rcv(sk, skb); + } } else sk_add_backlog(sk, skb); bh_unlock_sock(sk); @@ -1296,6 +1307,11 @@ int tcp_v4_destroy_sock(struct sock *sk) /* Cleans up our, hopefully empty, out_of_order_queue. */ __skb_queue_purge(&tp->out_of_order_queue); +#ifdef CONFIG_NET_DMA + /* Cleans up our sk_async_wait_queue */ + __skb_queue_purge(&sk->sk_async_wait_queue); +#endif + /* Clean prequeue, it must be empty really */ __skb_queue_purge(&tp->ucopy.prequeue); diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 301eee726b0f..a50eb306e9e2 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1218,8 +1218,16 @@ process: bh_lock_sock(sk); ret = 0; if (!sock_owned_by_user(sk)) { - if (!tcp_prequeue(sk, skb)) - ret = tcp_v6_do_rcv(sk, skb); +#ifdef CONFIG_NET_DMA + struct tcp_sock *tp = tcp_sk(sk); + if (tp->ucopy.dma_chan) + ret = tcp_v6_do_rcv(sk, skb); + else +#endif + { + if (!tcp_prequeue(sk, skb)) + ret = tcp_v6_do_rcv(sk, skb); + } } else sk_add_backlog(sk, skb); bh_unlock_sock(sk); -- cgit v1.2.3 From 15986e1aadbbf40a331cddd0470bb434d156431d Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 25 May 2006 16:11:14 -0700 Subject: [TCP]: tcp_rcv_rtt_measure_ts() call in pure-ACK path is superfluous We only want to take receive RTT mesaurements for data bearing frames, here in the header prediction fast path for a pure-sender, we know that we have a pure-ACK and thus the checks in tcp_rcv_rtt_mesaure_ts() will not pass. Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index c6d62f0a9966..6d167889a4b0 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3931,8 +3931,6 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, tp->rcv_nxt == tp->rcv_wup) tcp_store_ts_recent(tp); - tcp_rcv_rtt_measure_ts(sk, skb); - /* We know that such packets are checksummed * on entry. */ -- cgit v1.2.3 From 72dc5b9225c53310c010b68a70ea97c8c8e24bdf Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Mon, 5 Jun 2006 17:30:08 -0700 Subject: [TCP]: Minimum congestion window consolidation. Many of the TCP congestion methods all just use ssthresh as the minimum congestion window on decrease. Rather than duplicating the code, just have that be the default if that handle in the ops structure is not set. Minor behaviour change to TCP compound. It probably wants to use this (ssthresh) as lower bound, rather than ssthresh/2 because the latter causes undershoot on loss. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- include/net/tcp.h | 4 ++-- net/ipv4/tcp_bic.c | 7 ------- net/ipv4/tcp_compound.c | 1 - net/ipv4/tcp_cong.c | 6 +++--- net/ipv4/tcp_cubic.c | 6 ------ net/ipv4/tcp_htcp.c | 9 --------- net/ipv4/tcp_input.c | 13 +++++++++++-- net/ipv4/tcp_veno.c | 7 ------- net/ipv4/tcp_westwood.c | 18 +++++++----------- 9 files changed, 23 insertions(+), 48 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/include/net/tcp.h b/include/net/tcp.h index f1f472746e6c..de88c5472bfc 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -632,7 +632,7 @@ struct tcp_congestion_ops { /* return slow start threshold (required) */ u32 (*ssthresh)(struct sock *sk); /* lower bound for congestion window (optional) */ - u32 (*min_cwnd)(struct sock *sk); + u32 (*min_cwnd)(const struct sock *sk); /* do new cwnd calculation (required) */ void (*cong_avoid)(struct sock *sk, u32 ack, u32 rtt, u32 in_flight, int good_ack); @@ -667,7 +667,7 @@ extern struct tcp_congestion_ops tcp_init_congestion_ops; extern u32 tcp_reno_ssthresh(struct sock *sk); extern void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 rtt, u32 in_flight, int flag); -extern u32 tcp_reno_min_cwnd(struct sock *sk); +extern u32 tcp_reno_min_cwnd(const struct sock *sk); extern struct tcp_congestion_ops tcp_reno; static inline void tcp_set_ca_state(struct sock *sk, const u8 ca_state) diff --git a/net/ipv4/tcp_bic.c b/net/ipv4/tcp_bic.c index 035f2092d73a..b2d9021ad22b 100644 --- a/net/ipv4/tcp_bic.c +++ b/net/ipv4/tcp_bic.c @@ -198,12 +198,6 @@ static u32 bictcp_undo_cwnd(struct sock *sk) return max(tp->snd_cwnd, ca->last_max_cwnd); } -static u32 bictcp_min_cwnd(struct sock *sk) -{ - const struct tcp_sock *tp = tcp_sk(sk); - return tp->snd_ssthresh; -} - static void bictcp_state(struct sock *sk, u8 new_state) { if (new_state == TCP_CA_Loss) @@ -231,7 +225,6 @@ static struct tcp_congestion_ops bictcp = { .cong_avoid = bictcp_cong_avoid, .set_state = bictcp_state, .undo_cwnd = bictcp_undo_cwnd, - .min_cwnd = bictcp_min_cwnd, .pkts_acked = bictcp_acked, .owner = THIS_MODULE, .name = "bic", diff --git a/net/ipv4/tcp_compound.c b/net/ipv4/tcp_compound.c index ec68cb8081c1..bc54f7e9aea9 100644 --- a/net/ipv4/tcp_compound.c +++ b/net/ipv4/tcp_compound.c @@ -419,7 +419,6 @@ static struct tcp_congestion_ops tcp_compound = { .init = tcp_compound_init, .ssthresh = tcp_reno_ssthresh, .cong_avoid = tcp_compound_cong_avoid, - .min_cwnd = tcp_reno_min_cwnd, .rtt_sample = tcp_compound_rtt_calc, .set_state = tcp_compound_state, .cwnd_event = tcp_compound_cwnd_event, diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c index 91c2f41c7f58..857eefc52aab 100644 --- a/net/ipv4/tcp_cong.c +++ b/net/ipv4/tcp_cong.c @@ -38,7 +38,7 @@ int tcp_register_congestion_control(struct tcp_congestion_ops *ca) int ret = 0; /* all algorithms must implement ssthresh and cong_avoid ops */ - if (!ca->ssthresh || !ca->cong_avoid || !ca->min_cwnd) { + if (!ca->ssthresh || !ca->cong_avoid) { printk(KERN_ERR "TCP %s does not implement required ops\n", ca->name); return -EINVAL; @@ -251,8 +251,8 @@ u32 tcp_reno_ssthresh(struct sock *sk) } EXPORT_SYMBOL_GPL(tcp_reno_ssthresh); -/* Lower bound on congestion window. */ -u32 tcp_reno_min_cwnd(struct sock *sk) +/* Lower bound on congestion window with halving. */ +u32 tcp_reno_min_cwnd(const struct sock *sk) { const struct tcp_sock *tp = tcp_sk(sk); return tp->snd_ssthresh/2; diff --git a/net/ipv4/tcp_cubic.c b/net/ipv4/tcp_cubic.c index 31a4986dfbf7..78b7a6b9e4de 100644 --- a/net/ipv4/tcp_cubic.c +++ b/net/ipv4/tcp_cubic.c @@ -325,11 +325,6 @@ static u32 bictcp_undo_cwnd(struct sock *sk) return max(tcp_sk(sk)->snd_cwnd, ca->last_max_cwnd); } -static u32 bictcp_min_cwnd(struct sock *sk) -{ - return tcp_sk(sk)->snd_ssthresh; -} - static void bictcp_state(struct sock *sk, u8 new_state) { if (new_state == TCP_CA_Loss) @@ -357,7 +352,6 @@ static struct tcp_congestion_ops cubictcp = { .cong_avoid = bictcp_cong_avoid, .set_state = bictcp_state, .undo_cwnd = bictcp_undo_cwnd, - .min_cwnd = bictcp_min_cwnd, .pkts_acked = bictcp_acked, .owner = THIS_MODULE, .name = "cubic", diff --git a/net/ipv4/tcp_htcp.c b/net/ipv4/tcp_htcp.c index 1b2ff53f98ed..3d92c1859267 100644 --- a/net/ipv4/tcp_htcp.c +++ b/net/ipv4/tcp_htcp.c @@ -246,14 +246,6 @@ static void htcp_cong_avoid(struct sock *sk, u32 ack, u32 rtt, } } -/* Lower bound on congestion window. */ -static u32 htcp_min_cwnd(struct sock *sk) -{ - const struct tcp_sock *tp = tcp_sk(sk); - return tp->snd_ssthresh; -} - - static void htcp_init(struct sock *sk) { struct htcp *ca = inet_csk_ca(sk); @@ -285,7 +277,6 @@ static void htcp_state(struct sock *sk, u8 new_state) static struct tcp_congestion_ops htcp = { .init = htcp_init, .ssthresh = htcp_recalc_ssthresh, - .min_cwnd = htcp_min_cwnd, .cong_avoid = htcp_cong_avoid, .set_state = htcp_state, .undo_cwnd = htcp_cwnd_undo, diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 6d167889a4b0..e08245bdda3a 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1689,17 +1689,26 @@ static inline void tcp_moderate_cwnd(struct tcp_sock *tp) tp->snd_cwnd_stamp = tcp_time_stamp; } +/* Lower bound on congestion window is slow start threshold + * unless congestion avoidance choice decides to overide it. + */ +static inline u32 tcp_cwnd_min(const struct sock *sk) +{ + const struct tcp_congestion_ops *ca_ops = inet_csk(sk)->icsk_ca_ops; + + return ca_ops->min_cwnd ? ca_ops->min_cwnd(sk) : tcp_sk(sk)->snd_ssthresh; +} + /* Decrease cwnd each second ack. */ static void tcp_cwnd_down(struct sock *sk) { - const struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); int decr = tp->snd_cwnd_cnt + 1; tp->snd_cwnd_cnt = decr&1; decr >>= 1; - if (decr && tp->snd_cwnd > icsk->icsk_ca_ops->min_cwnd(sk)) + if (decr && tp->snd_cwnd > tcp_cwnd_min(sk)) tp->snd_cwnd -= decr; tp->snd_cwnd = min(tp->snd_cwnd, tcp_packets_in_flight(tp)+1); diff --git a/net/ipv4/tcp_veno.c b/net/ipv4/tcp_veno.c index 1091671751c4..11b42a7135c1 100644 --- a/net/ipv4/tcp_veno.c +++ b/net/ipv4/tcp_veno.c @@ -199,17 +199,10 @@ static u32 tcp_veno_ssthresh(struct sock *sk) return max(tp->snd_cwnd >> 1U, 2U); } -static u32 tcp_veno_min_cwnd(struct sock * sk) -{ - const struct tcp_sock *tp = tcp_sk(sk); - return tp->snd_ssthresh; -} - static struct tcp_congestion_ops tcp_veno = { .init = tcp_veno_init, .ssthresh = tcp_veno_ssthresh, .cong_avoid = tcp_veno_cong_avoid, - .min_cwnd = tcp_veno_min_cwnd, .rtt_sample = tcp_veno_rtt_calc, .set_state = tcp_veno_state, .cwnd_event = tcp_veno_cwnd_event, diff --git a/net/ipv4/tcp_westwood.c b/net/ipv4/tcp_westwood.c index 0c340c3756c2..29eb258b6d82 100644 --- a/net/ipv4/tcp_westwood.c +++ b/net/ipv4/tcp_westwood.c @@ -162,12 +162,6 @@ static inline u32 westwood_acked_count(struct sock *sk) return w->cumul_ack; } -static inline u32 westwood_bw_rttmin(const struct sock *sk) -{ - const struct tcp_sock *tp = tcp_sk(sk); - const struct westwood *w = inet_csk_ca(sk); - return max_t(u32, (w->bw_est * w->rtt_min) / tp->mss_cache, 2); -} /* * TCP Westwood @@ -175,9 +169,11 @@ static inline u32 westwood_bw_rttmin(const struct sock *sk) * in packets we use mss_cache). Rttmin is guaranteed to be >= 2 * so avoids ever returning 0. */ -static u32 tcp_westwood_cwnd_min(struct sock *sk) +static u32 tcp_westwood_bw_rttmin(const struct sock *sk) { - return westwood_bw_rttmin(sk); + const struct tcp_sock *tp = tcp_sk(sk); + const struct westwood *w = inet_csk_ca(sk); + return max_t(u32, (w->bw_est * w->rtt_min) / tp->mss_cache, 2); } static void tcp_westwood_event(struct sock *sk, enum tcp_ca_event event) @@ -191,11 +187,11 @@ static void tcp_westwood_event(struct sock *sk, enum tcp_ca_event event) break; case CA_EVENT_COMPLETE_CWR: - tp->snd_cwnd = tp->snd_ssthresh = westwood_bw_rttmin(sk); + tp->snd_cwnd = tp->snd_ssthresh = tcp_westwood_bw_rttmin(sk); break; case CA_EVENT_FRTO: - tp->snd_ssthresh = westwood_bw_rttmin(sk); + tp->snd_ssthresh = tcp_westwood_bw_rttmin(sk); break; case CA_EVENT_SLOW_ACK: @@ -235,7 +231,7 @@ static struct tcp_congestion_ops tcp_westwood = { .init = tcp_westwood_init, .ssthresh = tcp_reno_ssthresh, .cong_avoid = tcp_reno_cong_avoid, - .min_cwnd = tcp_westwood_cwnd_min, + .min_cwnd = tcp_westwood_bw_rttmin, .cwnd_event = tcp_westwood_event, .get_info = tcp_westwood_info, .pkts_acked = tcp_westwood_pkts_acked, -- cgit v1.2.3