/*
 * INET         An implementation of the TCP/IP protocol suite for the LINUX
 *              operating system.  INET is implemented using the  BSD Socket
 *              interface as the means of communication with the user level.
 *
 *              Implementation of the Transmission Control Protocol(TCP).
 *
 * Version:     $Id: tcp_input.c,v 1.241.2.1 2002/02/13 05:37:15 davem Exp $
 *
 * Authors:     Ross Biro, <bir7@leland.Stanford.Edu>
 *              Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *              Mark Evans, <evansmp@uhura.aston.ac.uk>
 *              Corey Minyard <wf-rch!minyard@relay.EU.net>
 *              Florian La Roche, <flla@stud.uni-sb.de>
 *              Charles Hedrick, <hedrick@klinzhai.rutgers.edu>
 *              Linus Torvalds, <torvalds@cs.helsinki.fi>
 *              Alan Cox, <gw4pts@gw4pts.ampr.org>
 *              Matthew Dillon, <dillon@apollo.west.oic.com>
 *              Arnt Gulbrandsen, <agulbra@nvg.unit.no>
 *              Jorge Cwik, <jorge@laser.satlink.net>
 */

/*
 * Changes:
 *              Pedro Roque     :       Fast Retransmit/Recovery.
 *                                      Two receive queues.
 *                                      Retransmit queue handled by TCP.
 *                                      Better retransmit timer handling.
 *                                      New congestion avoidance.
 *                                      Header prediction.
 *                                      Variable renaming.
 *
 *              Eric            :       Fast Retransmit.
 *              Randy Scott     :       MSS option defines.
 *              Eric Schenk     :       Fixes to slow start algorithm.
 *              Eric Schenk     :       Yet another double ACK bug.
 *              Eric Schenk     :       Delayed ACK bug fixes.
 *              Eric Schenk     :       Floyd style fast retrans war avoidance.
 *              David S. Miller :       Don't allow zero congestion window.
 *              Eric Schenk     :       Fix retransmitter so that it sends
 *                                      next packet on ack of previous packet.
 *              Andi Kleen      :       Moved open_request checking here
 *                                      and process RSTs for open_requests.
 *              Andi Kleen      :       Better prune_queue, and other fixes.
 *              Andrey Savochkin:       Fix RTT measurements in the presnce of
 *                                      timestamps.
 *              Andrey Savochkin:       Check sequence numbers correctly when
 *                                      removing SACKs due to in sequence incoming
 *                                      data segments.
 *              Andi Kleen:             Make sure we never ack data there is not
 *                                      enough room for. Also make this condition
 *                                      a fatal error if it might still happen.
 *              Andi Kleen:             Add tcp_measure_rcv_mss to make 
 *                                      connections with MSS<min(MTU,ann. MSS)
 *                                      work without delayed acks. 
 *              Andi Kleen:             Process packets with PSH set in the
 *                                      fast path.
 *              J Hadi Salim:           ECN support
 *              Andrei Gurtov,
 *              Pasi Sarolahti,
 *              Panu Kuhlberg:          Experimental audit of TCP (re)transmission
 *                                      engine. Lots of bugs are found.
 *              Pasi Sarolahti:         F-RTO for dealing with spurious RTOs
 *              Angelo Dell'Aera:       TCP Westwood+ support
 */

#include <linux/config.h>
#include <linux/mm.h>
#include <linux/sysctl.h>
#include <net/tcp.h>
#include <net/inet_common.h>
#include <linux/ipsec.h>

int sysctl_tcp_timestamps = 1;
int sysctl_tcp_window_scaling = 1;
int sysctl_tcp_sack = 1;
int sysctl_tcp_fack = 1;
int sysctl_tcp_reordering = TCP_FASTRETRANS_THRESH;
#ifdef CONFIG_INET_ECN
int sysctl_tcp_ecn = 1;
#else
int sysctl_tcp_ecn = 0;
#endif
int sysctl_tcp_dsack = 1;
int sysctl_tcp_app_win = 31;
int sysctl_tcp_adv_win_scale = 2;

int sysctl_tcp_stdurg = 0;
int sysctl_tcp_rfc1337 = 0;
int sysctl_tcp_max_orphans = NR_FILE;
int sysctl_tcp_frto = 0;

int sysctl_tcp_nometrics_save = 0;

int sysctl_tcp_westwood = 0;
int sysctl_tcp_vegas_cong_avoid = 0;

int sysctl_tcp_moderate_rcvbuf = 0;

/* 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
int sysctl_tcp_vegas_alpha = 1<<V_PARAM_SHIFT;
int sysctl_tcp_vegas_beta  = 3<<V_PARAM_SHIFT;
int sysctl_tcp_vegas_gamma = 1<<V_PARAM_SHIFT;
int sysctl_tcp_bic;
int sysctl_tcp_bic_fast_convergence = 1;
int sysctl_tcp_bic_low_window = 14;
int sysctl_tcp_bic_beta = 819;          /* = 819/1024 (BICTCP_BETA_SCALE) */

#define FLAG_DATA               0x01 /* Incoming frame contained data.          */
#define FLAG_WIN_UPDATE         0x02 /* Incoming ACK was a window update.       */
#define FLAG_DATA_ACKED         0x04 /* This ACK acknowledged new data.         */
#define FLAG_RETRANS_DATA_ACKED 0x08 /* "" "" some of which was retransmitted.  */
#define FLAG_SYN_ACKED          0x10 /* This ACK acknowledged SYN.              */
#define FLAG_DATA_SACKED        0x20 /* New SACK.                               */
#define FLAG_ECE                0x40 /* ECE in this ACK                         */
#define FLAG_DATA_LOST          0x80 /* SACK detected data lossage.             */
#define FLAG_SLOWPATH           0x100 /* Do not skip RFC checks for window update.*/

#define FLAG_ACKED              (FLAG_DATA_ACKED|FLAG_SYN_ACKED)
#define FLAG_NOT_DUP            (FLAG_DATA|FLAG_WIN_UPDATE|FLAG_ACKED)
#define FLAG_CA_ALERT           (FLAG_DATA_SACKED|FLAG_ECE)
#define FLAG_FORWARD_PROGRESS   (FLAG_ACKED|FLAG_DATA_SACKED)

#define IsReno(tp) ((tp)->sack_ok == 0)
#define IsFack(tp) ((tp)->sack_ok & 2)
#define IsDSack(tp) ((tp)->sack_ok & 4)

#define TCP_REMNANT (TCP_FLAG_FIN|TCP_FLAG_URG|TCP_FLAG_SYN|TCP_FLAG_PSH)

/* Adapt the MSS value used to make delayed ack decision to the 
 * real world.
 */ 
static __inline__ void tcp_measure_rcv_mss(struct tcp_opt *tp, struct sk_buff *skb)
{
        unsigned int len, lss;

        lss = tp->ack.last_seg_size; 
        tp->ack.last_seg_size = 0; 

        /* skb->len may jitter because of SACKs, even if peer
         * sends good full-sized frames.
         */
        len = skb->len;
        if (len >= tp->ack.rcv_mss) {
                tp->ack.rcv_mss = len;
        } else {
                /* Otherwise, we make more careful check taking into account,
                 * that SACKs block is variable.
                 *
                 * "len" is invariant segment length, including TCP header.
                 */
                len += skb->data - skb->h.raw;
                if (len >= TCP_MIN_RCVMSS + sizeof(struct tcphdr) ||
                    /* If PSH is not set, packet should be
                     * full sized, provided peer TCP is not badly broken.
                     * This observation (if it is correct 8)) allows
                     * to handle super-low mtu links fairly.
                     */
                    (len >= TCP_MIN_MSS + sizeof(struct tcphdr) &&
                     !(tcp_flag_word(skb->h.th)&TCP_REMNANT))) {
                        /* Subtract also invariant (if peer is RFC compliant),
                         * tcp header plus fixed timestamp option length.
                         * Resulting "len" is MSS free of SACK jitter.
                         */
                        len -= tp->tcp_header_len;
                        tp->ack.last_seg_size = len;
                        if (len == lss) {
                                tp->ack.rcv_mss = len;
                                return;
                        }
                }
                tp->ack.pending |= TCP_ACK_PUSHED;
        }
}

static void tcp_incr_quickack(struct tcp_opt *tp)
{
        unsigned quickacks = tp->rcv_wnd/(2*tp->ack.rcv_mss);

        if (quickacks==0)
                quickacks=2;
        if (quickacks > tp->ack.quick)
                tp->ack.quick = min(quickacks, TCP_MAX_QUICKACKS);
}

void tcp_enter_quickack_mode(struct tcp_opt *tp)
{
        tcp_incr_quickack(tp);
        tp->ack.pingpong = 0;
        tp->ack.ato = TCP_ATO_MIN;
}

/* Send ACKs quickly, if "quick" count is not exhausted
 * and the session is not interactive.
 */

static __inline__ int tcp_in_quickack_mode(struct tcp_opt *tp)
{
        return (tp->ack.quick && !tp->ack.pingpong);
}

/* Buffer size and advertised window tuning.
 *
 * 1. Tuning sk->sndbuf, when connection enters established state.
 */

static void tcp_fixup_sndbuf(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        int sndmem = tp->mss_clamp+MAX_TCP_HEADER+16+sizeof(struct sk_buff);

        if (sk->sndbuf < 3*sndmem)
                sk->sndbuf = min(3*sndmem, sysctl_tcp_wmem[2]);
}

/* 2. Tuning advertised window (window_clamp, rcv_ssthresh)
 *
 * All tcp_full_space() is split to two parts: "network" buffer, allocated
 * forward and advertised in receiver window (tp->rcv_wnd) and
 * "application buffer", required to isolate scheduling/application
 * latencies from network.
 * window_clamp is maximal advertised window. It can be less than
 * tcp_full_space(), in this case tcp_full_space() - window_clamp
 * is reserved for "application" buffer. The less window_clamp is
 * the smoother our behaviour from viewpoint of network, but the lower
 * throughput and the higher sensitivity of the connection to losses. 8)
 *
 * rcv_ssthresh is more strict window_clamp used at "slow start"
 * phase to predict further behaviour of this connection.
 * It is used for two goals:
 * - to enforce header prediction at sender, even when application
 *   requires some significant "application buffer". It is check #1.
 * - to prevent pruning of receive queue because of misprediction
 *   of receiver window. Check #2.
 *
 * The scheme does not work when sender sends good segments opening
 * window and then starts to feed us spagetti. But it should work
 * in common situations. Otherwise, we have to rely on queue collapsing.
 */

/* Slow part of check#2. */
static int
__tcp_grow_window(struct sock *sk, struct tcp_opt *tp, struct sk_buff *skb)
{
        /* Optimize this! */
        int truesize = tcp_win_from_space(skb->truesize)/2;
        int window = tcp_full_space(sk)/2;

        while (tp->rcv_ssthresh <= window) {
                if (truesize <= skb->len)
                        return 2*tp->ack.rcv_mss;

                truesize >>= 1;
                window >>= 1;
        }
        return 0;
}

static __inline__ void
tcp_grow_window(struct sock *sk, struct tcp_opt *tp, struct sk_buff *skb)
{
        /* Check #1 */
        if (tp->rcv_ssthresh < tp->window_clamp &&
            (int)tp->rcv_ssthresh < tcp_space(sk) &&
            !tcp_memory_pressure) {
                int incr;

                /* Check #2. Increase window, if skb with such overhead
                 * will fit to rcvbuf in future.
                 */
                if (tcp_win_from_space(skb->truesize) <= skb->len)
                        incr = 2*tp->advmss;
                else
                        incr = __tcp_grow_window(sk, tp, skb);

                if (incr) {
                        tp->rcv_ssthresh = min(tp->rcv_ssthresh + incr, tp->window_clamp);
                        tp->ack.quick |= 1;
                }
        }
}

/* 3. Tuning rcvbuf, when connection enters established state. */

static void tcp_fixup_rcvbuf(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        int rcvmem = tp->advmss+MAX_TCP_HEADER+16+sizeof(struct sk_buff);

        /* Try to select rcvbuf so that 4 mss-sized segments
         * will fit to window and correspoding skbs will fit to our rcvbuf.
         * (was 3; 4 is minimum to allow fast retransmit to work.)
         */
        while (tcp_win_from_space(rcvmem) < tp->advmss)
                rcvmem += 128;
        if (sk->rcvbuf < 4*rcvmem)
                sk->rcvbuf = min(4*rcvmem, sysctl_tcp_rmem[2]);
}

/* 4. Try to fixup all. It is made iimediately after connection enters
 *    established state.
 */
static void tcp_init_buffer_space(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        int maxwin;

        if (!(sk->userlocks&SOCK_RCVBUF_LOCK))
                tcp_fixup_rcvbuf(sk);
        if (!(sk->userlocks&SOCK_SNDBUF_LOCK))
                tcp_fixup_sndbuf(sk);

        tp->rcvq_space.space = tp->rcv_wnd;

        maxwin = tcp_full_space(sk);

        if (tp->window_clamp >= maxwin) {
                tp->window_clamp = maxwin;

                if (sysctl_tcp_app_win && maxwin>4*tp->advmss)
                        tp->window_clamp = max(maxwin-(maxwin>>sysctl_tcp_app_win), 4*tp->advmss);
        }

        /* Force reservation of one segment. */
        if (sysctl_tcp_app_win &&
            tp->window_clamp > 2*tp->advmss &&
            tp->window_clamp + tp->advmss > maxwin)
                tp->window_clamp = max(2*tp->advmss, maxwin-tp->advmss);

        tp->rcv_ssthresh = min(tp->rcv_ssthresh, tp->window_clamp);
        tp->snd_cwnd_stamp = tcp_time_stamp;
}

static void init_bictcp(struct tcp_opt *tp)
{
        tp->bictcp.cnt = 0;

        tp->bictcp.last_max_cwnd = 0;
        tp->bictcp.last_cwnd = 0;
        tp->bictcp.last_stamp = 0;
}

/* 5. Recalculate window clamp after socket hit its memory bounds. */
static void tcp_clamp_window(struct sock *sk, struct tcp_opt *tp)
{
        struct sk_buff *skb;
        unsigned int app_win = tp->rcv_nxt - tp->copied_seq;
        int ofo_win = 0;

        tp->ack.quick = 0;

        skb_queue_walk(&tp->out_of_order_queue, skb) {
                ofo_win += skb->len;
        }

        /* If overcommit is due to out of order segments,
         * do not clamp window. Try to expand rcvbuf instead.
         */
        if (ofo_win) {
                if (sk->rcvbuf < sysctl_tcp_rmem[2] &&
                    !(sk->userlocks&SOCK_RCVBUF_LOCK) &&
                    !tcp_memory_pressure &&
                    atomic_read(&tcp_memory_allocated) < sysctl_tcp_mem[0])
                        sk->rcvbuf = min(atomic_read(&sk->rmem_alloc), sysctl_tcp_rmem[2]);
        }
        if (atomic_read(&sk->rmem_alloc) > sk->rcvbuf) {
                app_win += ofo_win;
                if (atomic_read(&sk->rmem_alloc) >= 2*sk->rcvbuf)
                        app_win >>= 1;
                if (app_win > tp->ack.rcv_mss)
                        app_win -= tp->ack.rcv_mss;
                app_win = max(app_win, 2U*tp->advmss);

                if (!ofo_win)
                        tp->window_clamp = min(tp->window_clamp, app_win);
                tp->rcv_ssthresh = min(tp->window_clamp, 2U*tp->advmss);
        }
}

/* Receiver "autotuning" code.
 *
 * The algorithm for RTT estimation w/o timestamps is based on
 * Dynamic Right-Sizing (DRS) by Wu Feng and Mike Fisk of LANL.
 * <http://www.lanl.gov/radiant/website/pubs/drs/lacsi2001.ps>
 *
 * More detail on this code can be found at
 * <http://www.psc.edu/~jheffner/senior_thesis.ps>,
 * though this reference is out of date.  A new paper
 * is pending.
 */
static void tcp_rcv_rtt_update(struct tcp_opt *tp, u32 sample, int win_dep)
{
        u32 new_sample = tp->rcv_rtt_est.rtt;
        long m = sample;

        if (m == 0)
                m = 1;

        if (new_sample != 0) {
                /* If we sample in larger samples in the non-timestamp
                 * case, we could grossly overestimate the RTT especially
                 * with chatty applications or bulk transfer apps which
                 * are stalled on filesystem I/O.
                 *
                 * Also, since we are only going for a minimum in the
                 * non-timestamp case, we do not smoothe things out
                 * else with timestamps disabled convergance takes too
                 * long.
                 */
                if (!win_dep) {
                        m -= (new_sample >> 3);
                        new_sample += m;
                } else if (m < new_sample)
                        new_sample = m << 3;
        } else {
                /* No previous mesaure. */
                new_sample = m << 3;
        }

        if (tp->rcv_rtt_est.rtt != new_sample)
                tp->rcv_rtt_est.rtt = new_sample;
}

static inline void tcp_rcv_rtt_measure(struct tcp_opt *tp)
{
        if (tp->rcv_rtt_est.time == 0)
                goto new_measure;
        if (before(tp->rcv_nxt, tp->rcv_rtt_est.seq))
                return;
        tcp_rcv_rtt_update(tp,
                           jiffies - tp->rcv_rtt_est.time,
                           1);

new_measure:
        tp->rcv_rtt_est.seq = tp->rcv_nxt + tp->rcv_wnd;
        tp->rcv_rtt_est.time = tcp_time_stamp;
}

static inline void tcp_rcv_rtt_measure_ts(struct tcp_opt *tp, struct sk_buff *skb)
{
        if (tp->rcv_tsecr &&
            (TCP_SKB_CB(skb)->end_seq -
             TCP_SKB_CB(skb)->seq >= tp->ack.rcv_mss))
                tcp_rcv_rtt_update(tp, tcp_time_stamp - tp->rcv_tsecr, 0);
}

/*
 * This function should be called every time data is copied to user space.
 * It calculates the appropriate TCP receive buffer space.
 */
void tcp_rcv_space_adjust(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        int time;
        int space;
        
        if (tp->rcvq_space.time == 0)
                goto new_measure;
        
        time = tcp_time_stamp - tp->rcvq_space.time;
        if (time < (tp->rcv_rtt_est.rtt >> 3) ||
            tp->rcv_rtt_est.rtt == 0)
                return;
        
        space = 2 * (tp->copied_seq - tp->rcvq_space.seq);

        space = max(tp->rcvq_space.space, space);

        if (tp->rcvq_space.space != space) {
                int rcvmem;

                tp->rcvq_space.space = space;

                if (sysctl_tcp_moderate_rcvbuf) {
                        int new_clamp = space;

                        /* Receive space grows, normalize in order to
                         * take into account packet headers and sk_buff
                         * structure overhead.
                         */
                        space /= tp->advmss;
                        if (!space)
                                space = 1;
                        rcvmem = (tp->advmss + MAX_TCP_HEADER +
                                  16 + sizeof(struct sk_buff));
                        while (tcp_win_from_space(rcvmem) < tp->advmss)
                                rcvmem += 128;
                        space *= rcvmem;
                        space = min(space, sysctl_tcp_rmem[2]);
                        if (space > sk->rcvbuf) {
                                sk->rcvbuf = space;

                                /* Make the window clamp follow along.  */
                                tp->window_clamp = new_clamp;
                        }
                }
        }
        
new_measure:
        tp->rcvq_space.seq = tp->copied_seq;
        tp->rcvq_space.time = tcp_time_stamp;
}

/* There is something which you must keep in mind when you analyze the
 * behavior of the tp->ato delayed ack timeout interval.  When a
 * connection starts up, we want to ack as quickly as possible.  The
 * problem is that "good" TCP's do slow start at the beginning of data
 * transmission.  The means that until we send the first few ACK's the
 * sender will sit on his end and only queue most of his data, because
 * he can only send snd_cwnd unacked packets at any given time.  For
 * each ACK we send, he increments snd_cwnd and transmits more of his
 * queue.  -DaveM
 */
static void tcp_event_data_recv(struct sock *sk, struct tcp_opt *tp, struct sk_buff *skb)
{
        u32 now;

        tcp_schedule_ack(tp);

        tcp_measure_rcv_mss(tp, skb);

        tcp_rcv_rtt_measure(tp);

        now = tcp_time_stamp;

        if (!tp->ack.ato) {
                /* The _first_ data packet received, initialize
                 * delayed ACK engine.
                 */
                tcp_incr_quickack(tp);
                tp->ack.ato = TCP_ATO_MIN;
        } else {
                int m = now - tp->ack.lrcvtime;

                if (m <= TCP_ATO_MIN/2) {
                        /* The fastest case is the first. */
                        tp->ack.ato = (tp->ack.ato>>1) + TCP_ATO_MIN/2;
                } else if (m < tp->ack.ato) {
                        tp->ack.ato = (tp->ack.ato>>1) + m;
                        if (tp->ack.ato > tp->rto)
                                tp->ack.ato = tp->rto;
                } else if (m > tp->rto) {
                        /* Too long gap. Apparently sender falled to
                         * restart window, so that we send ACKs quickly.
                         */
                        tcp_incr_quickack(tp);
                        tcp_mem_reclaim(sk);
                }
        }
        tp->ack.lrcvtime = now;

        TCP_ECN_check_ce(tp, skb);

        if (skb->len >= 128)
                tcp_grow_window(sk, tp, skb);
}

/* When starting a new connection, pin down the current choice of 
 * congestion algorithm.
 */
void tcp_ca_init(struct tcp_opt *tp)
{
        if (sysctl_tcp_westwood) 
                tp->adv_cong = TCP_WESTWOOD;
        else if (sysctl_tcp_bic)
                tp->adv_cong = TCP_BIC;
        else if (sysctl_tcp_vegas_cong_avoid) {
                tp->adv_cong = TCP_VEGAS;
                tp->vegas.baseRTT = 0x7fffffff;
                tcp_vegas_enable(tp);
        } 
}

/* Do RTT sampling needed for Vegas.
 * Basically we:
 *   o min-filter RTT samples from within an RTT to get the current
 *     propagation delay + queuing delay (we are min-filtering to try to
 *     avoid the effects of delayed ACKs)
 *   o min-filter RTT samples from a much longer window (forever for now)
 *     to find the propagation delay (baseRTT)
 */
static inline void vegas_rtt_calc(struct tcp_opt *tp, __u32 rtt)
{
        __u32 vrtt = rtt + 1; /* Never allow zero rtt or baseRTT */

        /* Filter to find propagation delay: */
        if (vrtt < tp->vegas.baseRTT) 
                tp->vegas.baseRTT = vrtt;

        /* Find the min RTT during the last RTT to find
         * the current prop. delay + queuing delay:
         */
        tp->vegas.minRTT = min(tp->vegas.minRTT, vrtt);
        tp->vegas.cntRTT++;
}

/* Called to compute a smoothed rtt estimate. The data fed to this
 * routine either comes from timestamps, or from segments that were
 * known _not_ to have been retransmitted [see Karn/Partridge
 * Proceedings SIGCOMM 87]. The algorithm is from the SIGCOMM 88
 * piece by Van Jacobson.
 * NOTE: the next three routines used to be one big routine.
 * To save cycles in the RFC 1323 implementation it was better to break
 * it up into three procedures. -- erics
 */
static __inline__ void tcp_rtt_estimator(struct tcp_opt *tp, __u32 mrtt)
{
        long m = mrtt; /* RTT */

        if (tcp_vegas_enabled(tp))
                vegas_rtt_calc(tp, mrtt);

        /*      The following amusing code comes from Jacobson's
         *      article in SIGCOMM '88.  Note that rtt and mdev
         *      are scaled versions of rtt and mean deviation.
         *      This is designed to be as fast as possible 
         *      m stands for "measurement".
         *
         *      On a 1990 paper the rto value is changed to:
         *      RTO = rtt + 4 * mdev
         *
         * Funny. This algorithm seems to be very broken.
         * These formulae increase RTO, when it should be decreased, increase
         * too slowly, when it should be incresed fastly, decrease too fastly
         * etc. I guess in BSD RTO takes ONE value, so that it is absolutely
         * does not matter how to _calculate_ it. Seems, it was trap
         * that VJ failed to avoid. 8)
         */
        if(m == 0)
                m = 1;
        if (tp->srtt != 0) {
                m -= (tp->srtt >> 3);   /* m is now error in rtt est */
                tp->srtt += m;          /* rtt = 7/8 rtt + 1/8 new */
                if (m < 0) {
                        m = -m;         /* m is now abs(error) */
                        m -= (tp->mdev >> 2);   /* similar update on mdev */
                        /* This is similar to one of Eifel findings.
                         * Eifel blocks mdev updates when rtt decreases.
                         * This solution is a bit different: we use finer gain
                         * for mdev in this case (alpha*beta).
                         * Like Eifel it also prevents growth of rto,
                         * but also it limits too fast rto decreases,
                         * happening in pure Eifel.
                         */
                        if (m > 0)
                                m >>= 3;
                } else {
                        m -= (tp->mdev >> 2);   /* similar update on mdev */
                }
                tp->mdev += m;          /* mdev = 3/4 mdev + 1/4 new */
                if (tp->mdev > tp->mdev_max) {
                        tp->mdev_max = tp->mdev;
                        if (tp->mdev_max > tp->rttvar)
                                tp->rttvar = tp->mdev_max;
                }
                if (after(tp->snd_una, tp->rtt_seq)) {
                        if (tp->mdev_max < tp->rttvar)
                                tp->rttvar -= (tp->rttvar-tp->mdev_max)>>2;
                        tp->rtt_seq = tp->snd_nxt;
                        tp->mdev_max = TCP_RTO_MIN;
                }
        } else {
                /* no previous measure. */
                tp->srtt = m<<3;        /* take the measured time to be rtt */
                tp->mdev = m<<1;        /* make sure rto = 3*rtt */
                tp->mdev_max = tp->rttvar = max(tp->mdev, TCP_RTO_MIN);
                tp->rtt_seq = tp->snd_nxt;
        }

        tcp_westwood_update_rtt(tp, tp->srtt >> 3);
}

/* Calculate rto without backoff.  This is the second half of Van Jacobson's
 * routine referred to above.
 */
static __inline__ void tcp_set_rto(struct tcp_opt *tp)
{
        /* Old crap is replaced with new one. 8)
         *
         * More seriously:
         * 1. If rtt variance happened to be less 50msec, it is hallucination.
         *    It cannot be less due to utterly erratic ACK generation made
         *    at least by solaris and freebsd. "Erratic ACKs" has _nothing_
         *    to do with delayed acks, because at cwnd>2 true delack timeout
         *    is invisible. Actually, Linux-2.4 also generates erratic
         *    ACKs in some curcumstances.
         */
        tp->rto = (tp->srtt >> 3) + tp->rttvar;

        /* 2. Fixups made earlier cannot be right.
         *    If we do not estimate RTO correctly without them,
         *    all the algo is pure shit and should be replaced
         *    with correct one. It is exaclty, which we pretend to do.
         */
}

/* NOTE: clamping at TCP_RTO_MIN is not required, current algo
 * guarantees that rto is higher.
 */
static __inline__ void tcp_bound_rto(struct tcp_opt *tp)
{
        if (tp->rto > TCP_RTO_MAX)
                tp->rto = TCP_RTO_MAX;
}

/* Save metrics learned by this TCP session.
   This function is called only, when TCP finishes successfully
   i.e. when it enters TIME-WAIT or goes from LAST-ACK to CLOSE.
 */
void tcp_update_metrics(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        struct dst_entry *dst = __sk_dst_get(sk);

        if (sysctl_tcp_nometrics_save)
                return;

        dst_confirm(dst);

        if (dst && (dst->flags&DST_HOST)) {
                int m;

                if (tp->backoff || !tp->srtt) {
                        /* This session failed to estimate rtt. Why?
                         * Probably, no packets returned in time.
                         * Reset our results.
                         */
                        if (!(dst->mxlock&(1<<RTAX_RTT)))
                                dst->rtt = 0;
                        return;
                }

                m = dst->rtt - tp->srtt;

                /* If newly calculated rtt larger than stored one,
                 * store new one. Otherwise, use EWMA. Remember,
                 * rtt overestimation is always better than underestimation.
                 */
                if (!(dst->mxlock&(1<<RTAX_RTT))) {
                        if (m <= 0)
                                dst->rtt = tp->srtt;
                        else
                                dst->rtt -= (m>>3);
                }

                if (!(dst->mxlock&(1<<RTAX_RTTVAR))) {
                        if (m < 0)
                                m = -m;

                        /* Scale deviation to rttvar fixed point */
                        m >>= 1;
                        if (m < tp->mdev)
                                m = tp->mdev;

                        if (m >= dst->rttvar)
                                dst->rttvar = m;
                        else
                                dst->rttvar -= (dst->rttvar - m)>>2;
                }

                if (tp->snd_ssthresh >= 0xFFFF) {
                        /* Slow start still did not finish. */
                        if (dst->ssthresh &&
                            !(dst->mxlock&(1<<RTAX_SSTHRESH)) &&
                            (tp->snd_cwnd>>1) > dst->ssthresh)
                                dst->ssthresh = (tp->snd_cwnd>>1);
                        if (!(dst->mxlock&(1<<RTAX_CWND)) &&
                            tp->snd_cwnd > dst->cwnd)
                                dst->cwnd = tp->snd_cwnd;
                } else if (tp->snd_cwnd > tp->snd_ssthresh &&
                           tp->ca_state == TCP_CA_Open) {
                        /* Cong. avoidance phase, cwnd is reliable. */
                        if (!(dst->mxlock&(1<<RTAX_SSTHRESH)))
                                dst->ssthresh = max(tp->snd_cwnd>>1, tp->snd_ssthresh);
                        if (!(dst->mxlock&(1<<RTAX_CWND)))
                                dst->cwnd = (dst->cwnd + tp->snd_cwnd)>>1;
                } else {
                        /* Else slow start did not finish, cwnd is non-sense,
                           ssthresh may be also invalid.
                         */
                        if (!(dst->mxlock&(1<<RTAX_CWND)))
                                dst->cwnd = (dst->cwnd + tp->snd_ssthresh)>>1;
                        if (dst->ssthresh &&
                            !(dst->mxlock&(1<<RTAX_SSTHRESH)) &&
                            tp->snd_ssthresh > dst->ssthresh)
                                dst->ssthresh = tp->snd_ssthresh;
                }

                if (!(dst->mxlock&(1<<RTAX_REORDERING))) {
                        if (dst->reordering < tp->reordering &&
                            tp->reordering != sysctl_tcp_reordering)
                                dst->reordering = tp->reordering;
                }
        }
}

/* Increase initial CWND conservatively: if estimated
 * RTT is low enough (<20msec) or if we have some preset ssthresh.
 *
 * Numbers are taken from RFC2414.
 */
__u32 tcp_init_cwnd(struct tcp_opt *tp)
{
        __u32 cwnd;

        if (tp->mss_cache > 1460)
                return 2;

        cwnd = (tp->mss_cache > 1095) ? 3 : 4;

        if (!tp->srtt || (tp->snd_ssthresh >= 0xFFFF && tp->srtt > ((HZ/50)<<3)))
                cwnd = 2;
        else if (cwnd > tp->snd_ssthresh)
                cwnd = tp->snd_ssthresh;

        return min_t(__u32, cwnd, tp->snd_cwnd_clamp);
}

/* Initialize metrics on socket. */

static void tcp_init_metrics(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        struct dst_entry *dst = __sk_dst_get(sk);

        if (dst == NULL)
                goto reset;

        dst_confirm(dst);

        if (dst->mxlock&(1<<RTAX_CWND))
                tp->snd_cwnd_clamp = dst->cwnd;
        if (dst->ssthresh) {
                tp->snd_ssthresh = dst->ssthresh;
                if (tp->snd_ssthresh > tp->snd_cwnd_clamp)
                        tp->snd_ssthresh = tp->snd_cwnd_clamp;
        }
        if (dst->reordering && tp->reordering != dst->reordering) {
                tp->sack_ok &= ~2;
                tp->reordering = dst->reordering;
        }

        if (dst->rtt == 0)
                goto reset;

        if (!tp->srtt && dst->rtt < (TCP_TIMEOUT_INIT<<3))
                goto reset;

        /* Initial rtt is determined from SYN,SYN-ACK.
         * The segment is small and rtt may appear much
         * less than real one. Use per-dst memory
         * to make it more realistic.
         *
         * A bit of theory. RTT is time passed after "normal" sized packet
         * is sent until it is ACKed. In normal curcumstances sending small
         * packets force peer to delay ACKs and calculation is correct too.
         * The algorithm is adaptive and, provided we follow specs, it
         * NEVER underestimate RTT. BUT! If peer tries to make some clever
         * tricks sort of "quick acks" for time long enough to decrease RTT
         * to low value, and then abruptly stops to do it and starts to delay
         * ACKs, wait for troubles.
         */
        if (dst->rtt > tp->srtt) {
                tp->srtt = dst->rtt;
                tp->rtt_seq = tp->snd_nxt;
        }
        if (dst->rttvar > tp->mdev) {
                tp->mdev = dst->rttvar;
                tp->mdev_max = tp->rttvar = max(tp->mdev, TCP_RTO_MIN);
        }
        tcp_set_rto(tp);
        tcp_bound_rto(tp);
        if (tp->rto < TCP_TIMEOUT_INIT && !tp->saw_tstamp)
                goto reset;
        tp->snd_cwnd = tcp_init_cwnd(tp);
        tp->snd_cwnd_stamp = tcp_time_stamp;
        return;

reset:
        /* Play conservative. If timestamps are not
         * supported, TCP will fail to recalculate correct
         * rtt, if initial rto is too small. FORGET ALL AND RESET!
         */
        if (!tp->saw_tstamp && tp->srtt) {
                tp->srtt = 0;
                tp->mdev = tp->mdev_max = tp->rttvar = TCP_TIMEOUT_INIT;
                tp->rto = TCP_TIMEOUT_INIT;
        }
}

static void tcp_update_reordering(struct tcp_opt *tp, int metric, int ts)
{
        if (metric > tp->reordering) {
                tp->reordering = min(TCP_MAX_REORDERING, metric);

                /* This exciting event is worth to be remembered. 8) */
                if (ts)
                        NET_INC_STATS_BH(TCPTSReorder);
                else if (IsReno(tp))
                        NET_INC_STATS_BH(TCPRenoReorder);
                else if (IsFack(tp))
                        NET_INC_STATS_BH(TCPFACKReorder);
                else
                        NET_INC_STATS_BH(TCPSACKReorder);
#if FASTRETRANS_DEBUG > 1
                printk(KERN_DEBUG "Disorder%d %d %u f%u s%u rr%d\n",
                       tp->sack_ok, tp->ca_state,
                       tp->reordering, tp->fackets_out, tp->sacked_out,
                       tp->undo_marker ? tp->undo_retrans : 0);
#endif
                /* Disable FACK yet. */
                tp->sack_ok &= ~2;
        }
}

/* This procedure tags the retransmission queue when SACKs arrive.
 *
 * We have three tag bits: SACKED(S), RETRANS(R) and LOST(L).
 * Packets in queue with these bits set are counted in variables
 * sacked_out, retrans_out and lost_out, correspondingly.
 *
 * Valid combinations are:
 * Tag  InFlight        Description
 * 0    1               - orig segment is in flight.
 * S    0               - nothing flies, orig reached receiver.
 * L    0               - nothing flies, orig lost by net.
 * R    2               - both orig and retransmit are in flight.
 * L|R  1               - orig is lost, retransmit is in flight.
 * S|R  1               - orig reached receiver, retrans is still in flight.
 * (L|S|R is logically valid, it could occur when L|R is sacked,
 *  but it is equivalent to plain S and code short-curcuits it to S.
 *  L|S is logically invalid, it would mean -1 packet in flight 8))
 *
 * These 6 states form finite state machine, controlled by the following events:
 * 1. New ACK (+SACK) arrives. (tcp_sacktag_write_queue())
 * 2. Retransmission. (tcp_retransmit_skb(), tcp_xmit_retransmit_queue())
 * 3. Loss detection event of one of three flavors:
 *      A. Scoreboard estimator decided the packet is lost.
 *         A'. Reno "three dupacks" marks head of queue lost.
 *         A''. Its FACK modfication, head until snd.fack is lost.
 *      B. SACK arrives sacking data transmitted after never retransmitted
 *         hole was sent out.
 *      C. SACK arrives sacking SND.NXT at the moment, when the
 *         segment was retransmitted.
 * 4. D-SACK added new rule: D-SACK changes any tag to S.
 *
 * It is pleasant to note, that state diagram turns out to be commutative,
 * so that we are allowed not to be bothered by order of our actions,
 * when multiple events arrive simultaneously. (see the function below).
 *
 * Reordering detection.
 * --------------------
 * Reordering metric is maximal distance, which a packet can be displaced
 * in packet stream. With SACKs we can estimate it:
 *
 * 1. SACK fills old hole and the corresponding segment was not
 *    ever retransmitted -> reordering. Alas, we cannot use it
 *    when segment was retransmitted.
 * 2. The last flaw is solved with D-SACK. D-SACK arrives
 *    for retransmitted and already SACKed segment -> reordering..
 * Both of these heuristics are not used in Loss state, when we cannot
 * account for retransmits accurately.
 */
static int
tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_una)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        unsigned char *ptr = ack_skb->h.raw + TCP_SKB_CB(ack_skb)->sacked;
        struct tcp_sack_block *sp = (struct tcp_sack_block *)(ptr+2);
        int num_sacks = (ptr[1] - TCPOLEN_SACK_BASE)>>3;
        int reord = tp->packets_out;
        int prior_fackets;
        u32 lost_retrans = 0;
        int flag = 0;
        int i;

        if (!tp->sacked_out)
                tp->fackets_out = 0;
        prior_fackets = tp->fackets_out;

        for (i=0; i<num_sacks; i++, sp++) {
                struct sk_buff *skb;
                __u32 start_seq = ntohl(sp->start_seq);
                __u32 end_seq = ntohl(sp->end_seq);
                int fack_count = 0;
                int dup_sack = 0;

                /* Check for D-SACK. */
                if (i == 0) {
                        u32 ack = TCP_SKB_CB(ack_skb)->ack_seq;

                        if (before(start_seq, ack)) {
                                dup_sack = 1;
                                tp->sack_ok |= 4;
                                NET_INC_STATS_BH(TCPDSACKRecv);
                        } else if (num_sacks > 1 &&
                                   !after(end_seq, ntohl(sp[1].end_seq)) &&

                                   !before(start_seq, ntohl(sp[1].start_seq))) {
                                dup_sack = 1;
                                tp->sack_ok |= 4;
                                NET_INC_STATS_BH(TCPDSACKOfoRecv);
                        }

                        /* D-SACK for already forgotten data...
                         * Do dumb counting. */
                        if (dup_sack &&
                            !after(end_seq, prior_snd_una) &&
                            after(end_seq, tp->undo_marker))
                                tp->undo_retrans--;

                        /* Eliminate too old ACKs, but take into
                         * account more or less fresh ones, they can
                         * contain valid SACK info.
                         */
                        if (before(ack, prior_snd_una-tp->max_window))
                                return 0;
                }

                /* Event "B" in the comment above. */
                if (after(end_seq, tp->high_seq))
                        flag |= FLAG_DATA_LOST;

                for_retrans_queue(skb, sk, tp) {
                        u8 sacked = TCP_SKB_CB(skb)->sacked;
                        int in_sack;

                        /* The retransmission queue is always in order, so
                         * we can short-circuit the walk early.
                         */
                        if(!before(TCP_SKB_CB(skb)->seq, end_seq))
                                break;

                        fack_count++;

                        in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq) &&
                                !before(end_seq, TCP_SKB_CB(skb)->end_seq);

                        /* Account D-SACK for retransmitted packet. */
                        if ((dup_sack && in_sack) &&
                            (sacked & TCPCB_RETRANS) &&
                            after(TCP_SKB_CB(skb)->end_seq, tp->undo_marker))
                                tp->undo_retrans--;

                        /* The frame is ACKed. */
                        if (!after(TCP_SKB_CB(skb)->end_seq, tp->snd_una)) {
                                if (sacked&TCPCB_RETRANS) {
                                        if ((dup_sack && in_sack) &&
                                            (sacked&TCPCB_SACKED_ACKED))
                                                reord = min(fack_count, reord);
                                } else {
                                        /* If it was in a hole, we detected reordering. */
                                        if (fack_count < prior_fackets &&
                                            !(sacked&TCPCB_SACKED_ACKED))
                                                reord = min(fack_count, reord);
                                }

                                /* Nothing to do; acked frame is about to be dropped. */
                                continue;
                        }

                        if ((sacked&TCPCB_SACKED_RETRANS) &&
                            after(end_seq, TCP_SKB_CB(skb)->ack_seq) &&
                            (!lost_retrans || after(end_seq, lost_retrans)))
                                lost_retrans = end_seq;

                        if (!in_sack)
                                continue;

                        if (!(sacked&TCPCB_SACKED_ACKED)) {
                                if (sacked & TCPCB_SACKED_RETRANS) {
                                        /* If the segment is not tagged as lost,
                                         * we do not clear RETRANS, believing
                                         * that retransmission is still in flight.
                                         */
                                        if (sacked & TCPCB_LOST) {
                                                TCP_SKB_CB(skb)->sacked &= ~(TCPCB_LOST|TCPCB_SACKED_RETRANS);
                                                tp->lost_out--;
                                                tp->retrans_out--;
                                        }
                                } else {
                                        /* New sack for not retransmitted frame,
                                         * which was in hole. It is reordering.
                                         */
                                        if (!(sacked & TCPCB_RETRANS) &&
                                            fack_count < prior_fackets)
                                                reord = min(fack_count, reord);

                                        if (sacked & TCPCB_LOST) {
                                                TCP_SKB_CB(skb)->sacked &= ~TCPCB_LOST;
                                                tp->lost_out--;
                                        }
                                }

                                TCP_SKB_CB(skb)->sacked |= TCPCB_SACKED_ACKED;
                                flag |= FLAG_DATA_SACKED;
                                tp->sacked_out++;

                                if (fack_count > tp->fackets_out)
                                        tp->fackets_out = fack_count;
                        } else {
                                if (dup_sack && (sacked&TCPCB_RETRANS))
                                        reord = min(fack_count, reord);
                        }

                        /* D-SACK. We can detect redundant retransmission
                         * in S|R and plain R frames and clear it.
                         * undo_retrans is decreased above, L|R frames
                         * are accounted above as well.
                         */
                        if (dup_sack &&
                            (TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_RETRANS)) {
                                TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
                                tp->retrans_out--;
                        }
                }
        }

        /* Check for lost retransmit. This superb idea is
         * borrowed from "ratehalving". Event "C".
         * Later note: FACK people cheated me again 8),
         * we have to account for reordering! Ugly,
         * but should help.
         */
        if (lost_retrans && tp->ca_state == TCP_CA_Recovery) {
                struct sk_buff *skb;

                for_retrans_queue(skb, sk, tp) {
                        if (after(TCP_SKB_CB(skb)->seq, lost_retrans))
                                break;
                        if (!after(TCP_SKB_CB(skb)->end_seq, tp->snd_una))
                                continue;
                        if ((TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_RETRANS) &&
                            after(lost_retrans, TCP_SKB_CB(skb)->ack_seq) &&
                            (IsFack(tp) ||
                             !before(lost_retrans, TCP_SKB_CB(skb)->ack_seq+tp->reordering*tp->mss_cache))) {
                                TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
                                tp->retrans_out--;

                                if (!(TCP_SKB_CB(skb)->sacked&(TCPCB_LOST|TCPCB_SACKED_ACKED))) {
                                        tp->lost_out++;
                                        TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                                        flag |= FLAG_DATA_SACKED;
                                        NET_INC_STATS_BH(TCPLostRetransmit);
                                }
                        }
                }
        }

        tp->left_out = tp->sacked_out + tp->lost_out;

        if (reord < tp->fackets_out && tp->ca_state != TCP_CA_Loss)
                tcp_update_reordering(tp, (tp->fackets_out+1)-reord, 0);

#if FASTRETRANS_DEBUG > 0
        BUG_TRAP((int)tp->sacked_out >= 0);
        BUG_TRAP((int)tp->lost_out >= 0);
        BUG_TRAP((int)tp->retrans_out >= 0);
        BUG_TRAP((int)tcp_packets_in_flight(tp) >= 0);
#endif
        return flag;
}

/* RTO occurred, but do not yet enter loss state. Instead, transmit two new
 * segments to see from the next ACKs whether any data was really missing.
 * If the RTO was spurious, new ACKs should arrive.
 */
void tcp_enter_frto(struct sock *sk)
{
        struct tcp_opt *tp = &sk->tp_pinfo.af_tcp;
        struct sk_buff *skb;

        tp->frto_counter = 1;

        if (tp->ca_state <= TCP_CA_Disorder ||
            tp->snd_una == tp->high_seq ||
            (tp->ca_state == TCP_CA_Loss && !tp->retransmits)) {
                tp->prior_ssthresh = tcp_current_ssthresh(tp);
                tp->snd_ssthresh = tcp_recalc_ssthresh(tp);
        }

        /* Have to clear retransmission markers here to keep the bookkeeping
         * in shape, even though we are not yet in Loss state.
         * If something was really lost, it is eventually caught up
         * in tcp_enter_frto_loss.
         */
        tp->retrans_out = 0;
        tp->undo_marker = tp->snd_una;
        tp->undo_retrans = 0;

        for_retrans_queue(skb, sk, tp) {
                TCP_SKB_CB(skb)->sacked &= ~TCPCB_RETRANS;
        }
        tcp_sync_left_out(tp);

        tcp_set_ca_state(tp, TCP_CA_Open);
        tp->frto_highmark = tp->snd_nxt;
}

/* Enter Loss state after F-RTO was applied. Dupack arrived after RTO,
 * which indicates that we should follow the traditional RTO recovery,
 * i.e. mark everything lost and do go-back-N retransmission.
 */
void tcp_enter_frto_loss(struct sock *sk)
{
        struct tcp_opt *tp = &sk->tp_pinfo.af_tcp;
        struct sk_buff *skb;
        int cnt = 0;

        tp->sacked_out = 0;
        tp->lost_out = 0;
        tp->fackets_out = 0;

        for_retrans_queue(skb, sk, tp) {
                cnt++;
                TCP_SKB_CB(skb)->sacked &= ~TCPCB_LOST;
                if (!(TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_ACKED)) {

                        /* Do not mark those segments lost that were
                         * forward transmitted after RTO
                         */
                        if(!after(TCP_SKB_CB(skb)->end_seq,
                                   tp->frto_highmark)) {
                                TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                                tp->lost_out++;
                        }
                } else {
                        tp->sacked_out++;
                        tp->fackets_out = cnt;
                }
        }
        tcp_sync_left_out(tp);

        tp->snd_cwnd = tp->frto_counter + tcp_packets_in_flight(tp)+1;
        tp->snd_cwnd_cnt = 0;
        tp->snd_cwnd_stamp = tcp_time_stamp;
        tp->undo_marker = 0;
        tp->frto_counter = 0;

        tp->reordering = min_t(unsigned int, tp->reordering,
                                             sysctl_tcp_reordering);
        tcp_set_ca_state(tp, TCP_CA_Loss);
        tp->high_seq = tp->frto_highmark;
        TCP_ECN_queue_cwr(tp);

        init_bictcp(tp);
}

void tcp_clear_retrans(struct tcp_opt *tp)
{
        tp->left_out = 0;
        tp->retrans_out = 0;

        tp->fackets_out = 0;
        tp->sacked_out = 0;
        tp->lost_out = 0;

        tp->undo_marker = 0;
        tp->undo_retrans = 0;
}

/* Enter Loss state. If "how" is not zero, forget all SACK information
 * and reset tags completely, otherwise preserve SACKs. If receiver
 * dropped its ofo queue, we will know this due to reneging detection.
 */
void tcp_enter_loss(struct sock *sk, int how)
{
        struct tcp_opt *tp = &sk->tp_pinfo.af_tcp;
        struct sk_buff *skb;
        int cnt = 0;

        /* Reduce ssthresh if it has not yet been made inside this window. */
        if (tp->ca_state <= TCP_CA_Disorder ||
            tp->snd_una == tp->high_seq ||
            (tp->ca_state == TCP_CA_Loss && !tp->retransmits)) {
                tp->prior_ssthresh = tcp_current_ssthresh(tp);

                if (!(tcp_westwood_ssthresh(tp)))
                        tp->snd_ssthresh = tcp_recalc_ssthresh(tp);
        }
        tp->snd_cwnd = 1;
        tp->snd_cwnd_cnt = 0;
        tp->snd_cwnd_stamp = tcp_time_stamp;

        tcp_clear_retrans(tp);

        /* Push undo marker, if it was plain RTO and nothing
         * was retransmitted. */
        if (!how)
                tp->undo_marker = tp->snd_una;

        for_retrans_queue(skb, sk, tp) {
                cnt++;
                if (TCP_SKB_CB(skb)->sacked&TCPCB_RETRANS)
                        tp->undo_marker = 0;
                TCP_SKB_CB(skb)->sacked &= (~TCPCB_TAGBITS)|TCPCB_SACKED_ACKED;
                if (!(TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_ACKED) || how) {
                        TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_ACKED;
                        TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                        tp->lost_out++;
                } else {
                        tp->sacked_out++;
                        tp->fackets_out = cnt;
                }
        }
        tcp_sync_left_out(tp);

        tp->reordering = min_t(unsigned int, tp->reordering, sysctl_tcp_reordering);
        tcp_set_ca_state(tp, TCP_CA_Loss);
        tp->high_seq = tp->snd_nxt;
        TCP_ECN_queue_cwr(tp);
}

static int tcp_check_sack_reneging(struct sock *sk, struct tcp_opt *tp)
{
        struct sk_buff *skb;

        /* If ACK arrived pointing to a remembered SACK,
         * it means that our remembered SACKs do not reflect
         * real state of receiver i.e.
         * receiver _host_ is heavily congested (or buggy).
         * Do processing similar to RTO timeout.
         */
        if ((skb = skb_peek(&sk->write_queue)) != NULL &&
            (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) {
                NET_INC_STATS_BH(TCPSACKReneging);

                tcp_enter_loss(sk, 1);
                tp->retransmits++;
                tcp_retransmit_skb(sk, skb_peek(&sk->write_queue));
                tcp_reset_xmit_timer(sk, TCP_TIME_RETRANS, tp->rto);
                return 1;
        }
        return 0;
}

static inline int tcp_fackets_out(struct tcp_opt *tp)
{
        return IsReno(tp) ? tp->sacked_out+1 : tp->fackets_out;
}

static inline int tcp_skb_timedout(struct tcp_opt *tp, struct sk_buff *skb)
{
        return (tcp_time_stamp - TCP_SKB_CB(skb)->when > tp->rto);
}

static inline int tcp_head_timedout(struct sock *sk, struct tcp_opt *tp)
{
        return tp->packets_out && tcp_skb_timedout(tp, skb_peek(&sk->write_queue));
}

/* Linux NewReno/SACK/FACK/ECN state machine.
 * --------------------------------------
 *
 * "Open"       Normal state, no dubious events, fast path.
 * "Disorder"   In all the respects it is "Open",
 *              but requires a bit more attention. It is entered when
 *              we see some SACKs or dupacks. It is split of "Open"
 *              mainly to move some processing from fast path to slow one.
 * "CWR"        CWND was reduced due to some Congestion Notification event.
 *              It can be ECN, ICMP source quench, local device congestion.
 * "Recovery"   CWND was reduced, we are fast-retransmitting.
 * "Loss"       CWND was reduced due to RTO timeout or SACK reneging.
 *
 * tcp_fastretrans_alert() is entered:
 * - each incoming ACK, if state is not "Open"
 * - when arrived ACK is unusual, namely:
 *      * SACK
 *      * Duplicate ACK.
 *      * ECN ECE.
 *
 * Counting packets in flight is pretty simple.
 *
 *      in_flight = packets_out - left_out + retrans_out
 *
 *      packets_out is SND.NXT-SND.UNA counted in packets.
 *
 *      retrans_out is number of retransmitted segments.
 *
 *      left_out is number of segments left network, but not ACKed yet.
 *
 *              left_out = sacked_out + lost_out
 *
 *     sacked_out: Packets, which arrived to receiver out of order
 *                 and hence not ACKed. With SACKs this number is simply
 *                 amount of SACKed data. Even without SACKs
 *                 it is easy to give pretty reliable estimate of this number,
 *                 counting duplicate ACKs.
 *
 *       lost_out: Packets lost by network. TCP has no explicit
 *                 "loss notification" feedback from network (for now).
 *                 It means that this number can be only _guessed_.
 *                 Actually, it is the heuristics to predict lossage that
 *                 distinguishes different algorithms.
 *
 *      F.e. after RTO, when all the queue is considered as lost,
 *      lost_out = packets_out and in_flight = retrans_out.
 *
 *              Essentially, we have now two algorithms counting
 *              lost packets.
 *
 *              FACK: It is the simplest heuristics. As soon as we decided
 *              that something is lost, we decide that _all_ not SACKed
 *              packets until the most forward SACK are lost. I.e.
 *              lost_out = fackets_out - sacked_out and left_out = fackets_out.
 *              It is absolutely correct estimate, if network does not reorder
 *              packets. And it loses any connection to reality when reordering
 *              takes place. We use FACK by default until reordering
 *              is suspected on the path to this destination.
 *
 *              NewReno: when Recovery is entered, we assume that one segment
 *              is lost (classic Reno). While we are in Recovery and
 *              a partial ACK arrives, we assume that one more packet
 *              is lost (NewReno). This heuristics are the same in NewReno
 *              and SACK.
 *
 *  Imagine, that's all! Forget about all this shamanism about CWND inflation
 *  deflation etc. CWND is real congestion window, never inflated, changes
 *  only according to classic VJ rules.
 *
 * Really tricky (and requiring careful tuning) part of algorithm
 * is hidden in functions tcp_time_to_recover() and tcp_xmit_retransmit_queue().
 * The first determines the moment _when_ we should reduce CWND and,
 * hence, slow down forward transmission. In fact, it determines the moment
 * when we decide that hole is caused by loss, rather than by a reorder.
 *
 * tcp_xmit_retransmit_queue() decides, _what_ we should retransmit to fill
 * holes, caused by lost packets.
 *
 * And the most logically complicated part of algorithm is undo
 * heuristics. We detect false retransmits due to both too early
 * fast retransmit (reordering) and underestimated RTO, analyzing
 * timestamps and D-SACKs. When we detect that some segments were
 * retransmitted by mistake and CWND reduction was wrong, we undo
 * window reduction and abort recovery phase. This logic is hidden
 * inside several functions named tcp_try_undo_<something>.
 */

/* This function decides, when we should leave Disordered state
 * and enter Recovery phase, reducing congestion window.
 *
 * Main question: may we further continue forward transmission
 * with the same cwnd?
 */
static int
tcp_time_to_recover(struct sock *sk, struct tcp_opt *tp)
{
        /* Trick#1: The loss is proven. */
        if (tp->lost_out)
                return 1;

        /* Not-A-Trick#2 : Classic rule... */
        if (tcp_fackets_out(tp) > tp->reordering)
                return 1;

        /* Trick#3 : when we use RFC2988 timer restart, fast
         * retransmit can be triggered by timeout of queue head.
         */
        if (tcp_head_timedout(sk, tp))
                return 1;

        /* Trick#4: It is still not OK... But will it be useful to delay
         * recovery more?
         */
        if (tp->packets_out <= tp->reordering &&
            tp->sacked_out >= max_t(__u32, tp->packets_out/2, sysctl_tcp_reordering) &&
            !tcp_may_send_now(sk, tp)) {
                /* We have nothing to send. This connection is limited
                 * either by receiver window or by application.
                 */
                return 1;
        }

        return 0;
}

/* If we receive more dupacks than we expected counting segments
 * in assumption of absent reordering, interpret this as reordering.
 * The only another reason could be bug in receiver TCP.
 */
static void tcp_check_reno_reordering(struct tcp_opt *tp, int addend)
{
        u32 holes;

        holes = max(tp->lost_out, 1U);
        holes = min(holes, tp->packets_out);

        if (tp->sacked_out + holes > tp->packets_out) {
                tp->sacked_out = tp->packets_out - holes;
                tcp_update_reordering(tp, tp->packets_out+addend, 0);
        }
}

/* Emulate SACKs for SACKless connection: account for a new dupack. */

static void tcp_add_reno_sack(struct tcp_opt *tp)
{
        ++tp->sacked_out;
        tcp_check_reno_reordering(tp, 0);
        tcp_sync_left_out(tp);
}

/* Account for ACK, ACKing some data in Reno Recovery phase. */

static void tcp_remove_reno_sacks(struct sock *sk, struct tcp_opt *tp, int acked)
{
        if (acked > 0) {
                /* One ACK acked hole. The rest eat duplicate ACKs. */
                if (acked-1 >= tp->sacked_out)
                        tp->sacked_out = 0;
                else
                        tp->sacked_out -= acked-1;
        }
        tcp_check_reno_reordering(tp, acked);
        tcp_sync_left_out(tp);
}

static inline void tcp_reset_reno_sack(struct tcp_opt *tp)
{
        tp->sacked_out = 0;
        tp->left_out = tp->lost_out;
}

/* Mark head of queue up as lost. */
static void
tcp_mark_head_lost(struct sock *sk, struct tcp_opt *tp, int packets, u32 high_seq)
{
        struct sk_buff *skb;
        int cnt = packets;

        BUG_TRAP(cnt <= tp->packets_out);

        for_retrans_queue(skb, sk, tp) {
                if (--cnt < 0 || after(TCP_SKB_CB(skb)->end_seq, high_seq))
                        break;
                if (!(TCP_SKB_CB(skb)->sacked&TCPCB_TAGBITS)) {
                        TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                        tp->lost_out++;
                }
        }
        tcp_sync_left_out(tp);
}

/* Account newly detected lost packet(s) */

static void tcp_update_scoreboard(struct sock *sk, struct tcp_opt *tp)
{
        if (IsFack(tp)) {
                int lost = tp->fackets_out - tp->reordering;
                if (lost <= 0)
                        lost = 1;
                tcp_mark_head_lost(sk, tp, lost, tp->high_seq);
        } else {
                tcp_mark_head_lost(sk, tp, 1, tp->high_seq);
        }

        /* New heuristics: it is possible only after we switched
         * to restart timer each time when something is ACKed.
         * Hence, we can detect timed out packets during fast
         * retransmit without falling to slow start.
         */
        if (tcp_head_timedout(sk, tp)) {
                struct sk_buff *skb;

                for_retrans_queue(skb, sk, tp) {
                        if (tcp_skb_timedout(tp, skb) &&
                            !(TCP_SKB_CB(skb)->sacked&TCPCB_TAGBITS)) {
                                TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                                tp->lost_out++;
                        }
                }
                tcp_sync_left_out(tp);
        }
}

/* CWND moderation, preventing bursts due to too big ACKs
 * in dubious situations.
 */
static __inline__ void tcp_moderate_cwnd(struct tcp_opt *tp)
{
        tp->snd_cwnd = min(tp->snd_cwnd,
                           tcp_packets_in_flight(tp)+tcp_max_burst(tp));
        tp->snd_cwnd_stamp = tcp_time_stamp;
}

/* Decrease cwnd each second ack. */

static void tcp_cwnd_down(struct tcp_opt *tp)
{
        int decr = tp->snd_cwnd_cnt + 1;
        __u32 limit;

        /*
         * TCP Westwood
         * Here limit is evaluated as BWestimation*RTTmin (for obtaining it
         * in packets we use mss_cache). If sysctl_tcp_westwood is off
         * tcp_westwood_bw_rttmin() returns 0. In such case snd_ssthresh is
         * still used as usual. It prevents other strange cases in which
         * BWE*RTTmin could assume value 0. It should not happen but...
         */ 

        if (!(limit = tcp_westwood_bw_rttmin(tp)))
                limit = tp->snd_ssthresh/2;

        tp->snd_cwnd_cnt = decr&1;
        decr >>= 1;

        if (decr && tp->snd_cwnd > limit)
                tp->snd_cwnd -= decr;

        tp->snd_cwnd = min(tp->snd_cwnd, tcp_packets_in_flight(tp)+1);
        tp->snd_cwnd_stamp = tcp_time_stamp;
}

/* Nothing was retransmitted or returned timestamp is less
 * than timestamp of the first retransmission.
 */
static __inline__ int tcp_packet_delayed(struct tcp_opt *tp)
{
        return !tp->retrans_stamp ||
                (tp->saw_tstamp && tp->rcv_tsecr &&
                 (__s32)(tp->rcv_tsecr - tp->retrans_stamp) < 0);
}

/* Undo procedures. */

#if FASTRETRANS_DEBUG > 1
static void DBGUNDO(struct sock *sk, struct tcp_opt *tp, const char *msg)
{
        printk(KERN_DEBUG "Undo %s %u.%u.%u.%u/%u c%u l%u ss%u/%u p%u\n",
               msg,
               NIPQUAD(sk->daddr), ntohs(sk->dport),
               tp->snd_cwnd, tp->left_out,
               tp->snd_ssthresh, tp->prior_ssthresh, tp->packets_out);
}
#else
#define DBGUNDO(x...) do { } while (0)
#endif

static void tcp_undo_cwr(struct tcp_opt *tp, int undo)
{
        if (tp->prior_ssthresh) {
                if (tcp_is_bic(tp))
                        tp->snd_cwnd = max(tp->snd_cwnd, tp->bictcp.last_max_cwnd);
                else
                        tp->snd_cwnd = max(tp->snd_cwnd, tp->snd_ssthresh<<1);

                if (undo && tp->prior_ssthresh > tp->snd_ssthresh) {
                        tp->snd_ssthresh = tp->prior_ssthresh;
                        TCP_ECN_withdraw_cwr(tp);
                }
        } else {
                tp->snd_cwnd = max(tp->snd_cwnd, tp->snd_ssthresh);
        }
        tcp_moderate_cwnd(tp);
        tp->snd_cwnd_stamp = tcp_time_stamp;
}

static inline int tcp_may_undo(struct tcp_opt *tp)
{
        return tp->undo_marker &&
                (!tp->undo_retrans || tcp_packet_delayed(tp));
}

/* People celebrate: "We love our President!" */
static int tcp_try_undo_recovery(struct sock *sk, struct tcp_opt *tp)
{
        if (tcp_may_undo(tp)) {
                /* Happy end! We did not retransmit anything
                 * or our original transmission succeeded.
                 */
                DBGUNDO(sk, tp, tp->ca_state == TCP_CA_Loss ? "loss" : "retrans");
                tcp_undo_cwr(tp, 1);
                if (tp->ca_state == TCP_CA_Loss)
                        NET_INC_STATS_BH(TCPLossUndo);
                else
                        NET_INC_STATS_BH(TCPFullUndo);
                tp->undo_marker = 0;
        }
        if (tp->snd_una == tp->high_seq && IsReno(tp)) {
                /* Hold old state until something *above* high_seq
                 * is ACKed. For Reno it is MUST to prevent false
                 * fast retransmits (RFC2582). SACK TCP is safe. */
                tcp_moderate_cwnd(tp);
                return 1;
        }
        tcp_set_ca_state(tp, TCP_CA_Open);
        return 0;
}

/* Try to undo cwnd reduction, because D-SACKs acked all retransmitted data */
static void tcp_try_undo_dsack(struct sock *sk, struct tcp_opt *tp)
{
        if (tp->undo_marker && !tp->undo_retrans) {
                DBGUNDO(sk, tp, "D-SACK");
                tcp_undo_cwr(tp, 1);
                tp->undo_marker = 0;
                NET_INC_STATS_BH(TCPDSACKUndo);
        }
}

/* Undo during fast recovery after partial ACK. */

static int tcp_try_undo_partial(struct sock *sk, struct tcp_opt *tp, int acked)
{
        /* Partial ACK arrived. Force Hoe's retransmit. */
        int failed = IsReno(tp) || tp->fackets_out>tp->reordering;

        if (tcp_may_undo(tp)) {
                /* Plain luck! Hole if filled with delayed
                 * packet, rather than with a retransmit.
                 */
                if (tp->retrans_out == 0)
                        tp->retrans_stamp = 0;

                tcp_update_reordering(tp, tcp_fackets_out(tp)+acked, 1);

                DBGUNDO(sk, tp, "Hoe");
                tcp_undo_cwr(tp, 0);
                NET_INC_STATS_BH(TCPPartialUndo);

                /* So... Do not make Hoe's retransmit yet.
                 * If the first packet was delayed, the rest
                 * ones are most probably delayed as well.
                 */
                failed = 0;
        }
        return failed;
}

/* Undo during loss recovery after partial ACK. */
static int tcp_try_undo_loss(struct sock *sk, struct tcp_opt *tp)
{
        if (tcp_may_undo(tp)) {
                struct sk_buff *skb;
                for_retrans_queue(skb, sk, tp) {
                        TCP_SKB_CB(skb)->sacked &= ~TCPCB_LOST;
                }
                DBGUNDO(sk, tp, "partial loss");
                tp->lost_out = 0;
                tp->left_out = tp->sacked_out;
                tcp_undo_cwr(tp, 1);
                NET_INC_STATS_BH(TCPLossUndo);
                tp->retransmits = 0;
                tp->undo_marker = 0;
                if (!IsReno(tp))
                        tcp_set_ca_state(tp, TCP_CA_Open);
                return 1;
        }
        return 0;
}

static __inline__ void tcp_complete_cwr(struct tcp_opt *tp)
{
        if (!(tcp_westwood_complete_cwr(tp)))
                tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_ssthresh);
        tp->snd_cwnd_stamp = tcp_time_stamp;
}

static void tcp_try_to_open(struct sock *sk, struct tcp_opt *tp, int flag)
{
        tp->left_out = tp->sacked_out;

        if (tp->retrans_out == 0)
                tp->retrans_stamp = 0;

        if (flag&FLAG_ECE)
                tcp_enter_cwr(tp);

        if (tp->ca_state != TCP_CA_CWR) {
                int state = TCP_CA_Open;

                if (tp->left_out ||
                    tp->retrans_out ||
                    tp->undo_marker)
                        state = TCP_CA_Disorder;

                if (tp->ca_state != state) {
                        tcp_set_ca_state(tp, state);
                        tp->high_seq = tp->snd_nxt;
                }
                tcp_moderate_cwnd(tp);
        } else {
                tcp_cwnd_down(tp);
        }
}

/* Process an event, which can update packets-in-flight not trivially.
 * Main goal of this function is to calculate new estimate for left_out,
 * taking into account both packets sitting in receiver's buffer and
 * packets lost by network.
 *
 * Besides that it does CWND reduction, when packet loss is detected
 * and changes state of machine.
 *
 * It does _not_ decide what to send, it is made in function
 * tcp_xmit_retransmit_queue().
 */
static void
tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una,
                      int prior_packets, int flag)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        int is_dupack = (tp->snd_una == prior_snd_una && !(flag&FLAG_NOT_DUP));

        /* Some technical things:
         * 1. Reno does not count dupacks (sacked_out) automatically. */
        if (!tp->packets_out)
                tp->sacked_out = 0;
        /* 2. SACK counts snd_fack in packets inaccurately. */
        if (tp->sacked_out == 0)
                tp->fackets_out = 0;

        /* Now state machine starts.
         * A. ECE, hence prohibit cwnd undoing, the reduction is required. */
        if (flag&FLAG_ECE)
                tp->prior_ssthresh = 0;

        /* B. In all the states check for reneging SACKs. */
        if (tp->sacked_out && tcp_check_sack_reneging(sk, tp))
                return;

        /* C. Process data loss notification, provided it is valid. */
        if ((flag&FLAG_DATA_LOST) &&
            before(tp->snd_una, tp->high_seq) &&
            tp->ca_state != TCP_CA_Open &&
            tp->fackets_out > tp->reordering) {
                tcp_mark_head_lost(sk, tp, tp->fackets_out-tp->reordering, tp->high_seq);
                NET_INC_STATS_BH(TCPLoss);
        }

        /* D. Synchronize left_out to current state. */
        tcp_sync_left_out(tp);

        /* E. Check state exit conditions. State can be terminated
         *    when high_seq is ACKed. */
        if (tp->ca_state == TCP_CA_Open) {
                if (!sysctl_tcp_frto)
                        BUG_TRAP(tp->retrans_out == 0);
                tp->retrans_stamp = 0;
        } else if (!before(tp->snd_una, tp->high_seq)) {
                switch (tp->ca_state) {
                case TCP_CA_Loss:
                        tp->retransmits = 0;
                        if (tcp_try_undo_recovery(sk, tp))
                                return;
                        break;

                case TCP_CA_CWR:
                        /* CWR is to be held something *above* high_seq
                         * is ACKed for CWR bit to reach receiver. */
                        if (tp->snd_una != tp->high_seq) {
                                tcp_complete_cwr(tp);
                                tcp_set_ca_state(tp, TCP_CA_Open);
                        }
                        break;

                case TCP_CA_Disorder:
                        tcp_try_undo_dsack(sk, tp);
                        if (!tp->undo_marker ||
                            /* For SACK case do not Open to allow to undo
                             * catching for all duplicate ACKs. */
                            IsReno(tp) || tp->snd_una != tp->high_seq) {
                                tp->undo_marker = 0;
                                tcp_set_ca_state(tp, TCP_CA_Open);
                        }
                        break;

                case TCP_CA_Recovery:
                        if (IsReno(tp))
                                tcp_reset_reno_sack(tp);
                        if (tcp_try_undo_recovery(sk, tp))
                                return;
                        tcp_complete_cwr(tp);
                        break;
                }
        }

        /* F. Process state. */
        switch (tp->ca_state) {
        case TCP_CA_Recovery:
                if (prior_snd_una == tp->snd_una) {
                        if (IsReno(tp) && is_dupack)
                                tcp_add_reno_sack(tp);
                } else {
                        int acked = prior_packets - tp->packets_out;
                        if (IsReno(tp))
                                tcp_remove_reno_sacks(sk, tp, acked);
                        is_dupack = tcp_try_undo_partial(sk, tp, acked);
                }
                break;
        case TCP_CA_Loss:
                if (flag&FLAG_DATA_ACKED)
                        tp->retransmits = 0;
                if (!tcp_try_undo_loss(sk, tp)) {
                        tcp_moderate_cwnd(tp);
                        tcp_xmit_retransmit_queue(sk);
                        return;
                }
                if (tp->ca_state != TCP_CA_Open)
                        return;
                /* Loss is undone; fall through to processing in Open state. */
        default:
                if (IsReno(tp)) {
                        if (tp->snd_una != prior_snd_una)
                                tcp_reset_reno_sack(tp);
                        if (is_dupack)
                                tcp_add_reno_sack(tp);
                }

                if (tp->ca_state == TCP_CA_Disorder)
                        tcp_try_undo_dsack(sk, tp);

                if (!tcp_time_to_recover(sk, tp)) {
                        tcp_try_to_open(sk, tp, flag);
                        return;
                }

                /* Otherwise enter Recovery state */

                if (IsReno(tp))
                        NET_INC_STATS_BH(TCPRenoRecovery);
                else
                        NET_INC_STATS_BH(TCPSackRecovery);

                tp->high_seq = tp->snd_nxt;
                tp->prior_ssthresh = 0;
                tp->undo_marker = tp->snd_una;
                tp->undo_retrans = tp->retrans_out;

                if (tp->ca_state < TCP_CA_CWR) {
                        if (!(flag&FLAG_ECE))
                                tp->prior_ssthresh = tcp_current_ssthresh(tp);
                        tp->snd_ssthresh = tcp_recalc_ssthresh(tp);
                        TCP_ECN_queue_cwr(tp);
                }

                tp->snd_cwnd_cnt = 0;
                tcp_set_ca_state(tp, TCP_CA_Recovery);
        }

        if (is_dupack || tcp_head_timedout(sk, tp))
                tcp_update_scoreboard(sk, tp);
        tcp_cwnd_down(tp);
        tcp_xmit_retransmit_queue(sk);
}

/* Read draft-ietf-tcplw-high-performance before mucking
 * with this code. (Superceeds RFC1323)
 */
static void tcp_ack_saw_tstamp(struct tcp_opt *tp, int flag)
{
        __u32 seq_rtt;

        /* RTTM Rule: A TSecr value received in a segment is used to
         * update the averaged RTT measurement only if the segment
         * acknowledges some new data, i.e., only if it advances the
         * left edge of the send window.
         *
         * See draft-ietf-tcplw-high-performance-00, section 3.3.
         * 1998/04/10 Andrey V. Savochkin <saw@msu.ru>
         *
         * Changed: reset backoff as soon as we see the first valid sample.
         * If we do not, we get strongly overstimated rto. With timestamps
         * samples are accepted even from very old segments: f.e., when rtt=1
         * increases to 8, we retransmit 5 times and after 8 seconds delayed
         * answer arrives rto becomes 120 seconds! If at least one of segments
         * in window is lost... Voila.                          --ANK (010210)
         */
        seq_rtt = tcp_time_stamp - tp->rcv_tsecr;
        tcp_rtt_estimator(tp, seq_rtt);
        tcp_set_rto(tp);
        tp->backoff = 0;
        tcp_bound_rto(tp);
}

static void tcp_ack_no_tstamp(struct tcp_opt *tp, u32 seq_rtt, int flag)
{
        /* We don't have a timestamp. Can only use
         * packets that are not retransmitted to determine
         * rtt estimates. Also, we must not reset the
         * backoff for rto until we get a non-retransmitted
         * packet. This allows us to deal with a situation
         * where the network delay has increased suddenly.
         * I.e. Karn's algorithm. (SIGCOMM '87, p5.)
         */

        if (flag & FLAG_RETRANS_DATA_ACKED)
                return;

        tcp_rtt_estimator(tp, seq_rtt);
        tcp_set_rto(tp);
        tp->backoff = 0;
        tcp_bound_rto(tp);
}

static __inline__ void
tcp_ack_update_rtt(struct tcp_opt *tp, int flag, s32 seq_rtt)

{
        /* Note that peer MAY send zero echo. In this case it is ignored. (rfc1323) */
        if (tp->saw_tstamp && tp->rcv_tsecr)
                tcp_ack_saw_tstamp(tp, flag);
        else if (seq_rtt >= 0)
                tcp_ack_no_tstamp(tp, seq_rtt, flag);
}

/*
 * Compute congestion window to use.
 *
 * This is from the implementation of BICTCP in
 * Lison-Xu, Kahaled Harfoush, and Injog Rhee.
 *  "Binary Increase Congestion Control for Fast, Long Distance
 *  Networks" in InfoComm 2004
 * Available from:
 *  http://www.csc.ncsu.edu/faculty/rhee/export/bitcp.pdf
 *
 * Unless BIC is enabled and congestion window is large
 * this behaves the same as the original Reno.
 */
static inline __u32 bictcp_cwnd(struct tcp_opt *tp)
{
        /* orignal Reno behaviour */
        if (!tcp_is_bic(tp))
                return tp->snd_cwnd;

        if (tp->bictcp.last_cwnd == tp->snd_cwnd &&
           (s32)(tcp_time_stamp - tp->bictcp.last_stamp) <= (HZ>>5))
                return tp->bictcp.cnt;

        tp->bictcp.last_cwnd = tp->snd_cwnd;
        tp->bictcp.last_stamp = tcp_time_stamp;
      
        /* start off normal */
        if (tp->snd_cwnd <= sysctl_tcp_bic_low_window)
                tp->bictcp.cnt = tp->snd_cwnd;

        /* binary increase */
        else if (tp->snd_cwnd < tp->bictcp.last_max_cwnd) {
                __u32   dist = (tp->bictcp.last_max_cwnd - tp->snd_cwnd)
                        / BICTCP_B;

                if (dist > BICTCP_MAX_INCREMENT)
                        /* linear increase */
                        tp->bictcp.cnt = tp->snd_cwnd / BICTCP_MAX_INCREMENT;
                else if (dist <= 1U)
                        /* binary search increase */
                        tp->bictcp.cnt = tp->snd_cwnd * BICTCP_FUNC_OF_MIN_INCR
                                / BICTCP_B;
                else
                        /* binary search increase */
                        tp->bictcp.cnt = tp->snd_cwnd / dist;
        } else {
                /* slow start amd linear increase */
                if (tp->snd_cwnd < tp->bictcp.last_max_cwnd + BICTCP_B)
                        /* slow start */
                        tp->bictcp.cnt = tp->snd_cwnd * BICTCP_FUNC_OF_MIN_INCR
                                / BICTCP_B;
                else if (tp->snd_cwnd < tp->bictcp.last_max_cwnd
                                        + BICTCP_MAX_INCREMENT*(BICTCP_B-1))
                        /* slow start */
                        tp->bictcp.cnt = tp->snd_cwnd * (BICTCP_B-1)
                                / (tp->snd_cwnd-tp->bictcp.last_max_cwnd);
                else
                        /* linear increase */
                        tp->bictcp.cnt = tp->snd_cwnd / BICTCP_MAX_INCREMENT;
        }
        return tp->bictcp.cnt;
}

/* This is Jacobson's slow start and congestion avoidance. 
 * SIGCOMM '88, p. 328.
 */
static __inline__ void reno_cong_avoid(struct tcp_opt *tp)
{
        if (tp->snd_cwnd <= tp->snd_ssthresh) {
                /* In "safe" area, increase. */
                if (tp->snd_cwnd < tp->snd_cwnd_clamp)
                        tp->snd_cwnd++;
        } else {
                /* In dangerous area, increase slowly.
                 * In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd
                 */
                if (tp->snd_cwnd_cnt >= bictcp_cwnd(tp)) {
                        if (tp->snd_cwnd < tp->snd_cwnd_clamp)
                                tp->snd_cwnd++;
                        tp->snd_cwnd_cnt=0;
                } else
                        tp->snd_cwnd_cnt++;
        }
        tp->snd_cwnd_stamp = tcp_time_stamp;
}

/* This is based on the congestion detection/avoidance scheme described in
 *    Lawrence S. Brakmo and Larry L. Peterson.
 *    "TCP Vegas: End to end congestion avoidance on a global internet."
 *    IEEE Journal on Selected Areas in Communication, 13(8):1465--1480,
 *    October 1995. Available from:
 *      ftp://ftp.cs.arizona.edu/xkernel/Papers/jsac.ps
 *
 * See http://www.cs.arizona.edu/xkernel/ for their implementation.
 * The main aspects that distinguish this implementation from the
 * Arizona Vegas implementation are:
 *   o We do not change the loss detection or recovery mechanisms of
 *     Linux in any way. Linux already recovers from losses quite well,
 *     using fine-grained timers, NewReno, and FACK.
 *   o To avoid the performance penalty imposed by increasing cwnd
 *     only every-other RTT during slow start, we increase during
 *     every RTT during slow start, just like Reno.
 *   o Largely to allow continuous cwnd growth during slow start,
 *     we use the rate at which ACKs come back as the "actual"
 *     rate, rather than the rate at which data is sent.
 *   o To speed convergence to the right rate, we set the cwnd
 *     to achieve the right ("actual") rate when we exit slow start.
 *   o To filter out the noise caused by delayed ACKs, we use the
 *     minimum RTT sample observed during the last RTT to calculate
 *     the actual rate.
 *   o When the sender re-starts from idle, it waits until it has
 *     received ACKs for an entire flight of new data before making
 *     a cwnd adjustment decision. The original Vegas implementation
 *     assumed senders never went idle.
 */
static void vegas_cong_avoid(struct tcp_opt *tp, u32 ack, u32 seq_rtt)
{
        /* The key players are v_beg_snd_una and v_beg_snd_nxt.
         *
         * These are so named because they represent the approximate values
         * of snd_una and snd_nxt at the beginning of the current RTT. More
         * precisely, they represent the amount of data sent during the RTT.
         * At the end of the RTT, when we receive an ACK for v_beg_snd_nxt,
         * we will calculate that (v_beg_snd_nxt - v_beg_snd_una) outstanding
         * bytes of data have been ACKed during the course of the RTT, giving
         * an "actual" rate of:
         *
         *     (v_beg_snd_nxt - v_beg_snd_una) / (rtt duration)
         *
         * Unfortunately, v_beg_snd_una is not exactly equal to snd_una,
         * because delayed ACKs can cover more than one segment, so they
         * don't line up nicely with the boundaries of RTTs.
         *
         * Another unfortunate fact of life is that delayed ACKs delay the
         * advance of the left edge of our send window, so that the number
         * of bytes we send in an RTT is often less than our cwnd will allow.
         * So we keep track of our cwnd separately, in v_beg_snd_cwnd.
         */

        if (after(ack, tp->vegas.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 = (tp->vegas.beg_snd_nxt - tp->vegas.beg_snd_una) /
                        tp->mss_cache;
                old_snd_cwnd = tp->vegas.beg_snd_cwnd;

                /* Save the extent of the current window so we can use this
                 * at the end of the next RTT.
                 */
                tp->vegas.beg_snd_una  = tp->vegas.beg_snd_nxt;
                tp->vegas.beg_snd_nxt  = tp->snd_nxt;
                tp->vegas.beg_snd_cwnd = tp->snd_cwnd;

                /* Take into account the current RTT sample too, to
                 * decrease the impact of delayed acks. This double counts
                 * this sample since we count it for the next window as well,
                 * but that's not too awful, since we're taking the min,
                 * rather than averaging.
                 */
                vegas_rtt_calc(tp, seq_rtt);

                /* We do the Vegas calculations only if we got enough RTT
                 * samples that we can be reasonably sure that we got
                 * at least one RTT sample that wasn't from a delayed ACK.
                 * If we only had 2 samples total,
                 * then that means we're getting only 1 ACK per RTT, which
                 * means they're almost certainly delayed ACKs.
                 * If  we have 3 samples, we should be OK.
                 */

                if (tp->vegas.cntRTT <= 2) {
                        /* We don't have enough RTT samples to do the Vegas
                         * calculation, so we'll behave like Reno.
                         */
                        if (tp->snd_cwnd > tp->snd_ssthresh)
                                tp->snd_cwnd++;
                } else {
                        u32 rtt, target_cwnd, diff;

                        /* We have enough RTT samples, so, using the Vegas
                         * algorithm, we determine if we should increase or
                         * decrease cwnd, and by how much.
                         */

                        /* Pluck out the RTT we are using for the Vegas
                         * calculations. This is the min RTT seen during the
                         * last RTT. Taking the min filters out the effects
                         * of delayed ACKs, at the cost of noticing congestion
                         * a bit later.
                         */
                        rtt = tp->vegas.minRTT;

                        /* Calculate the cwnd we should have, if we weren't
                         * going too fast.
                         *
                         * 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 = ((old_wnd * tp->vegas.baseRTT)
                                       << V_PARAM_SHIFT) / 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;

                        if (tp->snd_cwnd < tp->snd_ssthresh) {
                                /* Slow start.  */
                                if (diff > sysctl_tcp_vegas_gamma) {
                                        /* Going too fast. Time to slow down
                                         * and switch to congestion avoidance.
                                         */
                                        tp->snd_ssthresh = 2;

                                        /* Set cwnd to match the actual rate
                                         * exactly:
                                         *   cwnd = (actual rate) * baseRTT
                                         * Then we add 1 because the integer
                                         * truncation robs us of full link
                                         * utilization.
                                         */
                                        tp->snd_cwnd = min(tp->snd_cwnd,
                                                           (target_cwnd >>
                                                            V_PARAM_SHIFT)+1);

                                }
                        } else {
                                /* Congestion avoidance. */
                                u32 next_snd_cwnd;

                                /* Figure out where we would like cwnd
                                 * to be.
                                 */
                                if (diff > sysctl_tcp_vegas_beta) {
                                        /* The old window was too fast, so
                                         * we slow down.
                                         */
                                        next_snd_cwnd = old_snd_cwnd - 1;
                                } else if (diff < sysctl_tcp_vegas_alpha) {
                                        /* We don't have enough extra packets
                                         * in the network, so speed up.
                                         */
                                        next_snd_cwnd = old_snd_cwnd + 1;
                                } 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--;
                        }
                }

                /* Wipe the slate clean for the next RTT. */
                tp->vegas.cntRTT = 0;
                tp->vegas.minRTT = 0x7fffffff;
        }

        /* The following code is executed for every ack we receive,
         * except for conditions checked in should_advance_cwnd()
         * before the call to tcp_cong_avoid(). Mainly this means that
         * we only execute this code if the ack actually acked some
         * data.
         */

        /* If we are in slow start, increase our cwnd in response to this ACK.
         * (If we are not in slow start then we are in congestion avoidance,
         * and adjust our congestion window only once per RTT. See the code
         * above.)
         */
        if (tp->snd_cwnd <= tp->snd_ssthresh) 
                tp->snd_cwnd++;

        /* to keep cwnd from growing without bound */
        tp->snd_cwnd = min_t(u32, tp->snd_cwnd, tp->snd_cwnd_clamp);

        /* Make sure that we are never so timid as to reduce our cwnd below
         * 2 MSS.
         *
         * Going below 2 MSS would risk huge delayed ACKs from our receiver.
         */
        tp->snd_cwnd = max(tp->snd_cwnd, 2U);

        tp->snd_cwnd_stamp = tcp_time_stamp;
}

static inline void tcp_cong_avoid(struct tcp_opt *tp, u32 ack, u32 seq_rtt)
{
        if (tcp_vegas_enabled(tp))
                vegas_cong_avoid(tp, ack, seq_rtt);
        else
                reno_cong_avoid(tp);
}

/* Restart timer after forward progress on connection.
 * RFC2988 recommends to restart timer to now+rto.
 */

static __inline__ void tcp_ack_packets_out(struct sock *sk, struct tcp_opt *tp)
{
        if (tp->packets_out==0) {
                tcp_clear_xmit_timer(sk, TCP_TIME_RETRANS);
        } else {
                tcp_reset_xmit_timer(sk, TCP_TIME_RETRANS, tp->rto);
        }
}

/* Remove acknowledged frames from the retransmission queue. */
static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        struct sk_buff *skb;
        __u32 now = tcp_time_stamp;
        int acked = 0;
        __s32 seq_rtt = -1;

        while((skb=skb_peek(&sk->write_queue)) && (skb != tp->send_head)) {
                struct tcp_skb_cb *scb = TCP_SKB_CB(skb); 
                __u8 sacked = scb->sacked;

                /* If our packet is before the ack sequence we can
                 * discard it as it's confirmed to have arrived at
                 * the other end.
                 */
                if (after(scb->end_seq, tp->snd_una))
                        break;

                /* Initial outgoing SYN's get put onto the write_queue
                 * just like anything else we transmit.  It is not
                 * true data, and if we misinform our callers that
                 * this ACK acks real data, we will erroneously exit
                 * connection startup slow start one packet too
                 * quickly.  This is severely frowned upon behavior.
                 */
                if(!(scb->flags & TCPCB_FLAG_SYN)) {
                        acked |= FLAG_DATA_ACKED;
                } else {
                        acked |= FLAG_SYN_ACKED;
                        tp->retrans_stamp = 0;
                }

                if (sacked) {
                        if(sacked & TCPCB_RETRANS) {
                                if(sacked & TCPCB_SACKED_RETRANS)
                                        tp->retrans_out--;
                                acked |= FLAG_RETRANS_DATA_ACKED;
                                seq_rtt = -1;
                        } else if (seq_rtt < 0)
                                seq_rtt = now - scb->when;
                        if(sacked & TCPCB_SACKED_ACKED)
                                tp->sacked_out--;
                        if(sacked & TCPCB_LOST)
                                tp->lost_out--;
                        if(sacked & TCPCB_URG) {
                                if (tp->urg_mode &&
                                    !before(scb->end_seq, tp->snd_up))
                                        tp->urg_mode = 0;
                        }
                } else if (seq_rtt < 0)
                        seq_rtt = now - scb->when;
                if(tp->fackets_out)
                        tp->fackets_out--;
                tp->packets_out--;
                __skb_unlink(skb, skb->list);
                tcp_free_skb(sk, skb);
        }

        if (acked&FLAG_ACKED) {
                tcp_ack_update_rtt(tp, acked, seq_rtt);
                tcp_ack_packets_out(sk, tp);
        }

#if FASTRETRANS_DEBUG > 0
        BUG_TRAP((int)tp->sacked_out >= 0);
        BUG_TRAP((int)tp->lost_out >= 0);
        BUG_TRAP((int)tp->retrans_out >= 0);
        if (tp->packets_out==0 && tp->sack_ok) {
                if (tp->lost_out) {
                        printk(KERN_DEBUG "Leak l=%u %d\n", tp->lost_out, tp->ca_state);
                        tp->lost_out = 0;
                }
                if (tp->sacked_out) {
                        printk(KERN_DEBUG "Leak s=%u %d\n", tp->sacked_out, tp->ca_state);
                        tp->sacked_out = 0;
                }
                if (tp->retrans_out) {
                        printk(KERN_DEBUG "Leak r=%u %d\n", tp->retrans_out, tp->ca_state);
                        tp->retrans_out = 0;
                }
        }
#endif
        *seq_rtt_p = seq_rtt;
        return acked;
}

static void tcp_ack_probe(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        /* Was it a usable window open? */

        if (!after(TCP_SKB_CB(tp->send_head)->end_seq, tp->snd_una + tp->snd_wnd)) {
                tp->backoff = 0;
                tcp_clear_xmit_timer(sk, TCP_TIME_PROBE0);
                /* Socket must be waked up by subsequent tcp_data_snd_check().
                 * This function is not for random using!
                 */
        } else {
                tcp_reset_xmit_timer(sk, TCP_TIME_PROBE0,
                                     min(tp->rto << tp->backoff, TCP_RTO_MAX));
        }
}

static __inline__ int tcp_ack_is_dubious(struct tcp_opt *tp, int flag)
{
        return (!(flag & FLAG_NOT_DUP) || (flag & FLAG_CA_ALERT) ||
                tp->ca_state != TCP_CA_Open);
}

static __inline__ int tcp_may_raise_cwnd(struct tcp_opt *tp, int flag)
{
        return (!(flag & FLAG_ECE) || tp->snd_cwnd < tp->snd_ssthresh) &&
                !((1<<tp->ca_state)&(TCPF_CA_Recovery|TCPF_CA_CWR));
}

/* Check that window update is acceptable.
 * The function assumes that snd_una<=ack<=snd_next.
 */
static __inline__ int
tcp_may_update_window(struct tcp_opt *tp, u32 ack, u32 ack_seq, u32 nwin)
{
        return (after(ack, tp->snd_una) ||
                after(ack_seq, tp->snd_wl1) ||
                (ack_seq == tp->snd_wl1 && nwin > tp->snd_wnd));
}

/* Update our send window.
 *
 * Window update algorithm, described in RFC793/RFC1122 (used in linux-2.2
 * and in FreeBSD. NetBSD's one is even worse.) is wrong.
 */
static int tcp_ack_update_window(struct sock *sk, struct tcp_opt *tp,
                                 struct sk_buff *skb, u32 ack, u32 ack_seq)
{
        int flag = 0;
        u32 nwin = ntohs(skb->h.th->window);

        if (likely(!skb->h.th->syn))
                nwin <<= tp->snd_wscale;

        if (tcp_may_update_window(tp, ack, ack_seq, nwin)) {
                flag |= FLAG_WIN_UPDATE;
                tcp_update_wl(tp, ack, ack_seq);

                if (tp->snd_wnd != nwin) {
                        tp->snd_wnd = nwin;

                        /* Note, it is the only place, where
                         * fast path is recovered for sending TCP.
                         */
                        tcp_fast_path_check(sk, tp);

                        if (nwin > tp->max_window) {
                                tp->max_window = nwin;
                                tcp_sync_mss(sk, tp->pmtu_cookie);
                        }
                }
        }

        tp->snd_una = ack;

        return flag;
}

static void tcp_process_frto(struct sock *sk, u32 prior_snd_una)
{
        struct tcp_opt *tp = &sk->tp_pinfo.af_tcp;
        
        tcp_sync_left_out(tp);
        
        if (tp->snd_una == prior_snd_una ||
            !before(tp->snd_una, tp->frto_highmark)) {
                /* RTO was caused by loss, start retransmitting in
                 * go-back-N slow start
                 */
                tcp_enter_frto_loss(sk);
                return;
        }

        if (tp->frto_counter == 1) {
                /* First ACK after RTO advances the window: allow two new
                 * segments out.
                 */
                tp->snd_cwnd = tcp_packets_in_flight(tp) + 2;
        } else {
                /* Also the second ACK after RTO advances the window.
                 * The RTO was likely spurious. Reduce cwnd and continue
                 * in congestion avoidance
                 */
                tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_ssthresh);
                tcp_moderate_cwnd(tp);
        }

        /* F-RTO affects on two new ACKs following RTO.
         * At latest on third ACK the TCP behavor is back to normal.
         */
        tp->frto_counter = (tp->frto_counter + 1) % 3;
}

/*
 * TCP Westwood+
 */

/*
 * @westwood_do_filter
 * Low-pass filter. Implemented using constant coeffients.
 */

static inline __u32 westwood_do_filter(__u32 a, __u32 b)
{
        return (((7 * a) + b) >> 3);
}

static void westwood_filter(struct sock *sk, __u32 delta)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        tp->westwood.bw_ns_est =
                westwood_do_filter(tp->westwood.bw_ns_est, 
                                   tp->westwood.bk / delta);
        tp->westwood.bw_est =
                westwood_do_filter(tp->westwood.bw_est,
                                   tp->westwood.bw_ns_est);
}

/* @westwood_update_rttmin
 * It is used to update RTTmin. In this case we MUST NOT use
 * WESTWOOD_RTT_MIN minimum bound since we could be on a LAN!
 */

static inline __u32 westwood_update_rttmin(const struct sock *sk)
{
        const struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        __u32 rttmin = tp->westwood.rtt_min;

        if (tp->westwood.rtt != 0 && 
            (tp->westwood.rtt < tp->westwood.rtt_min || !rttmin))
                rttmin = tp->westwood.rtt;

        return rttmin;
}

/*
 * @westwood_acked
 * Evaluate increases for dk. 
 */

static __u32 westwood_acked(const struct sock *sk)
{
        const struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        return ((tp->snd_una) - (tp->westwood.snd_una));
}

/*
 * @westwood_new_window
 * It evaluates if we are receiving data inside the same RTT window as
 * when we started.
 * Return value:
 * It returns 0 if we are still evaluating samples in the same RTT
 * window, 1 if the sample has to be considered in the next window.
 */

static int westwood_new_window(const struct sock *sk)
{
        const struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        __u32 left_bound;
        __u32 rtt;
        int ret = 0;

        left_bound = tp->westwood.rtt_win_sx;
        rtt = max(tp->westwood.rtt, (__u32)TCP_WESTWOOD_RTT_MIN);

        /*
         * A RTT-window has passed. Be careful since if RTT is less than
         * 50ms we don't filter but we continue 'building the sample'.
         * This minimum limit was choosen since an estimation on small
         * time intervals is better to avoid...
         * Obvioulsy on a LAN we reasonably will always have
         * right_bound = left_bound + WESTWOOD_RTT_MIN
         */

        if ((left_bound + rtt) < tcp_time_stamp)
                ret = 1;

        return ret;
}

/*
 * @westwood_update_window
 * It updates RTT evaluation window if it is the right moment to do
 * it. If so it calls filter for evaluating bandwidth. 
 */
 
static void __westwood_update_window(struct sock *sk, __u32 now)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        __u32 delta = now - tp->westwood.rtt_win_sx;

        if (delta) {
                if (tp->westwood.rtt)
                        westwood_filter(sk, delta);

                tp->westwood.bk = 0;
                tp->westwood.rtt_win_sx = tcp_time_stamp;
        }
}

static void westwood_update_window(struct sock *sk, __u32 now)
{
        if (westwood_new_window(sk))
                __westwood_update_window(sk, now);
}

/*
 * @__tcp_westwood_fast_bw
 * It is called when we are in fast path. In particular it is called when
 * header prediction is successfull. In such case infact update is
 * straight forward and doesn't need any particular care.
 */

void __tcp_westwood_fast_bw(struct sock *sk, struct sk_buff *skb)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        westwood_update_window(sk, tcp_time_stamp);

        tp->westwood.bk += westwood_acked(sk);
        tp->westwood.snd_una = tp->snd_una;
        tp->westwood.rtt_min = westwood_update_rttmin(sk);
}

/*
 * @westwood_mss
 * This function was inserted just to have the possibility to evaluate
 * which value of MSS is better. Infact we can use neither mss_cache or
 * mss_cache. Just testing we will know it!
 */

static inline __u32 westwood_mss(struct tcp_opt *tp)
{
        return ((__u32)(tp->mss_cache));
}

/*
 * @tcp_westwood_dupack_update
 * It updates accounted and cumul_ack when receiving a dupack.
 */

static void westwood_dupack_update(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        tp->westwood.accounted += westwood_mss(tp);
        tp->westwood.cumul_ack = westwood_mss(tp);
}

static inline int westwood_may_change_cumul(struct tcp_opt *tp)
{
        return (tp->westwood.cumul_ack > westwood_mss(tp));
}

static inline void westwood_partial_update(struct tcp_opt *tp)
{
        tp->westwood.accounted -= tp->westwood.cumul_ack;
        tp->westwood.cumul_ack = westwood_mss(tp);
}

static inline void westwood_complete_update(struct tcp_opt *tp)
{
        tp->westwood.cumul_ack -= tp->westwood.accounted;
        tp->westwood.accounted = 0;
}

/*
 * @westwood_acked_count
 * This function evaluates cumul_ack for evaluating dk in case of
 * delayed or partial acks.
 */

static inline __u32 westwood_acked_count(struct sock *sk)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        tp->westwood.cumul_ack = westwood_acked(sk);

        /* If cumul_ack is 0 this is a dupack since it's not moving
         * tp->snd_una.
         */
        if (!(tp->westwood.cumul_ack))
                westwood_dupack_update(sk);

        if (westwood_may_change_cumul(tp)) {
                /* Partial or delayed ack */
                if (tp->westwood.accounted >= tp->westwood.cumul_ack)
                        westwood_partial_update(tp);
                else 
                        westwood_complete_update(tp);
        }

        tp->westwood.snd_una = tp->snd_una;

        return tp->westwood.cumul_ack;
}

/*
 * @__tcp_westwood_slow_bw
 * It is called when something is going wrong..even if there could
 * be no problems! Infact a simple delayed packet may trigger a
 * dupack. But we need to be careful in such case.
 */

void __tcp_westwood_slow_bw(struct sock *sk, struct sk_buff *skb)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

        westwood_update_window(sk, tcp_time_stamp);

        tp->westwood.bk += westwood_acked_count(sk);
        tp->westwood.rtt_min = westwood_update_rttmin(sk);
}

/* TCP Westwood+ routines end here */

/* This routine deals with incoming acks, but not outgoing ones. */
static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
{
        struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
        u32 prior_snd_una = tp->snd_una;
        u32 ack_seq = TCP_SKB_CB(skb)->seq;
        u32 ack = TCP_SKB_CB(skb)->ack_seq;
        u32 prior_in_flight;
        s32 seq_rtt;
        int prior_packets;

        /* If the ack is newer than sent or older than previous acks
         * then we can probably ignore it.
         */
        if (after(ack, tp->snd_nxt))
                goto uninteresting_ack;

        if (before(ack, prior_snd_una))
                goto old_ack;

        if (!(flag&FLAG_SLOWPATH) && after(ack, prior_snd_una)) {
                /* Window is constant, pure forward advance.
                 * No more checks are required.
                 * Note, we use the fact that SND.UNA>=SND.WL2.
                 */
                tcp_update_wl(tp, ack, ack_seq);
                tp->snd_una = ack;
                tcp_westwood_fast_bw(sk, skb);
                flag |= FLAG_WIN_UPDATE;

                NET_INC_STATS_BH(TCPHPAcks);
        } else {
                if (ack_seq != TCP_SKB_CB(skb)->end_seq)
                        flag |= FLAG_DATA;
                else
                        NET_INC_STATS_BH(TCPPureAcks);

                flag |= tcp_ack_update_window(sk, tp, skb, ack, ack_seq);

                if (TCP_SKB_CB(skb)->sacked)
                        flag |= tcp_sacktag_write_queue(sk, skb, prior_snd_una);

                if (TCP_ECN_rcv_ecn_echo(tp, skb->h.th))
                        flag |= FLAG_ECE;

                tcp_westwood_slow_bw(sk, skb);
        }

        /* We passed data and got it acked, remove any soft error
         * log. Something worked...
         */
        sk->err_soft = 0;
        tp->rcv_tstamp = tcp_time_stamp;
        if ((prior_packets = tp->packets_out) == 0)
                goto no_queue;

        prior_in_flight = tcp_packets_in_flight(tp);

        /* See if we can take anything off of the retransmit queue. */
        flag |= tcp_clean_rtx_queue(sk, &seq_rtt);

        if (tp->frto_counter)
                tcp_process_frto(sk, prior_snd_una);

        if (tcp_ack_is_dubious(tp, flag)) {
                /* Advanve CWND, if state allows this. */
                if ((flag&FLAG_DATA_ACKED) &&
                    (tcp_vegas_enabled(tp) || prior_in_flight >= tp->snd_cwnd) &&
                    tcp_may_raise_cwnd(tp, flag))
                        tcp_cong_avoid(tp, ack, seq_rtt);
                tcp_fastretrans_alert(sk, prior_snd_una, prior_packets, flag);
        } else {
                if ((flag & FLAG_DATA_ACKED) && 
                    (tcp_vegas_enabled(tp) || prior_in_flight >= tp->snd_cwnd))
                        tcp_cong_avoid(tp, ack, seq_rtt);
        }

        if ((flag & FLAG_FORWARD_PROGRESS) || !(flag&FLAG_NOT_DUP))
                dst_confirm(sk->dst_cache);

        return 1;

no_queue:
        tp->probes_out = 0;

        /* If this ack opens up a zero window, clear backoff.  It was
         * being used to time the probes, and is probably far higher than
         * it needs to be for normal retransmission.
         */
        if (tp->send_head)
                tcp_ack_probe(sk);
        return 1;

old_ack:
        if (TCP_SKB_CB(skb)->sacked)
                tcp_sacktag_write_queue(sk, skb, prior_snd_una);

uninteresting_ack:
        SOCK_DEBUG(sk, "Ack %u out of %u:%u\n", ack, tp->snd_una, tp->snd_nxt);
        return 0;
}


/* Look for tcp options. Normally only called on SYN and SYNACK packets.
 * But, this can also be called on packets in the established flow when
 * the fast version below fails.
 */
void tcp_parse_options(struct sk_buff *skb, struct tcp_opt *tp, int estab)
{
        unsigned char *ptr;
        struct tcphdr *th = skb->h.th;
        int length=(th->doff*4)-sizeof(struct tcphdr);

        ptr = (unsigned char *)(th + 1);
        tp->saw_tstamp = 0;

        while(length>0) {
                int opcode=*ptr++;
                int opsize;

                switch (opcode) {
                        case TCPOPT_EOL:
                                return;
                        case TCPOPT_NOP:        /* Ref: RFC 793 section 3.1 */
                                length--;
                                continue;
                        default:
                                opsize=*ptr++;
                                if (opsize < 2) /* "silly options" */
                                        return;
                                if (opsize > length)
                                        return; /* don't parse partial options */
                                switch(opcode) {
                                case TCPOPT_MSS:
                                        if(opsize==TCPOLEN_MSS && th->syn && !estab) {
                                                u16 in_mss = ntohs(*(__u16 *)ptr);
                                                if (in_mss) {
                                                        if (tp->user_mss && tp->user_mss < in_mss)
                                                                in_mss = tp->user_mss;
                                                        tp->mss_clamp = in_mss;
                                                }
                                        }
                                        break;
                                case TCPOPT_WINDOW:
                                        if(opsize==TCPOLEN_WINDOW && th->syn && !estab)
                                                if (sysctl_tcp_window_scaling) {
                                                        tp->wscale_ok = 1;
                                                        tp->snd_wscale = *(__u8 *)ptr;
                                                        if(tp->snd_wscale > 14) {
                                                                if(net_ratelimit())
                                                                        printk(KERN_INFO "tcp_parse_options: Illegal window "
                                                                               "scaling value %d >14 received.\n",
                                                                               tp->snd_wscale);
                                                                tp->snd_wscale = 14;
                                                        }
                                                }
                                        break;
                                case TCPOPT_TIMESTAMP:
                                        if(opsize==TCPOLEN_TIMESTAMP) {
                                                if ((estab && tp->tstamp_ok) ||
                                                    (!estab && sysctl_tcp_timestamps)) {
                                                        tp->saw_tstamp = 1;
                                                        tp->rcv_tsval = ntohl(*(__u32 *)ptr);
                                                        tp->rcv_tsecr = ntohl(*(__u32 *)(ptr+4));
                                                }
                                        }
                                        break;
                                case TCPOPT_SACK_PERM:
                                        if(opsize==TCPOLEN_SACK_PERM && th->syn && !estab) {
                                                if (sysctl_tcp_sack) {
                                                        tp->sack_ok = 1;
                                                        tcp_sack_reset(tp);
                                                }
                                        }
                                        break;

                                case TCPOPT_SACK:
                                        if((opsize >= (TCPOLEN_SACK_BASE + TCPOLEN_SACK_PERBLOCK)) &&
                                           !((opsize - TCPOLEN_SACK_BASE) % TCPOLEN_SACK_PERBLOCK) &&
                                           tp->sack_ok) {
                                                TCP_SKB_CB(skb)->sacked = (ptr - 2) - (unsigned char *)th;
                                        }
                                };
                                ptr+=opsize-2;
                                length-=opsize;
                };
        }
}

/* Fast parse options. This hopes to only see timestamps.
 * If it is wrong it falls back on tcp_parse_options().
 */
static __inline__ int tcp_fast_parse_options(struct sk_buff *skb, struct tcphdr *th, struct tcp_opt *tp)
{
        if (th->doff == sizeof(struct tcphdr)>>2) {
                tp->saw_tstamp = 0;
                return 0;
        } else if (tp->tstamp_ok &&
                   th->doff == (sizeof(struct tcphdr)>>2)+(TCPOLEN_TSTAMP_ALIGNED>>2)) {
                __u32 *ptr = (__u32 *)(th + 1);
                if (*ptr == ntohl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16)
                                  | (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP)) {
                        tp->saw_tstamp = 1;
                        ++ptr;
                        tp->rcv_tsval = ntohl(*ptr);
                        ++ptr;
                        tp->rcv_tsecr = ntohl(*ptr);
                        return 1;
                }
        }
        tcp_parse_options(skb, tp, 1);
        return 1;
}

extern __inline__ void
tcp_store_ts_recent(struct tcp_opt *tp)
{
        tp->ts_recent = tp->rcv_tsval;
        tp->ts_recent_stamp = xtime.tv_sec;
}

extern __inline__ void
tcp_replace_ts_recent(struct tcp_opt *tp, u32 seq)
{
        if (tp->saw_tstamp && !after(seq, tp->rcv_wup)) {
                /* PAWS bug workaround wrt. ACK frames, the PAWS discard
                 * extra check below makes sure this can only happen
                 * for pure ACK frames.  -DaveM
                 *
                 * Not only, also it occurs for expired timestamps.
                 */

                if((s32)(tp->rcv_tsval - tp->ts_recent) >= 0 ||
                   xtime.tv_sec >= tp->ts_recent_stamp + TCP_PAWS_24DAYS)
                        tcp_store_ts_recent(tp);
        }