]> pilppa.org Git - linux-2.6-omap-h63xx.git/blobdiff - net/ipv4/tcp_output.c
Merge ssh://master.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
[linux-2.6-omap-h63xx.git] / net / ipv4 / tcp_output.c
index 7114031fdc70e3ce8c6adcdd7beda2f6d2b5d049..a7623ead39a8b6a91fe34b81afd5193d9ed6cce3 100644 (file)
@@ -51,8 +51,8 @@ int sysctl_tcp_retrans_collapse = 1;
  */
 int sysctl_tcp_tso_win_divisor = 3;
 
-static inline void update_send_head(struct sock *sk, struct tcp_sock *tp,
-                                   struct sk_buff *skb)
+static void update_send_head(struct sock *sk, struct tcp_sock *tp,
+                            struct sk_buff *skb)
 {
        sk->sk_send_head = skb->next;
        if (sk->sk_send_head == (struct sk_buff *)&sk->sk_write_queue)
@@ -124,8 +124,8 @@ static void tcp_cwnd_restart(struct sock *sk, struct dst_entry *dst)
        tp->snd_cwnd_used = 0;
 }
 
-static inline void tcp_event_data_sent(struct tcp_sock *tp,
-                                      struct sk_buff *skb, struct sock *sk)
+static void tcp_event_data_sent(struct tcp_sock *tp,
+                               struct sk_buff *skb, struct sock *sk)
 {
        struct inet_connection_sock *icsk = inet_csk(sk);
        const u32 now = tcp_time_stamp;
@@ -142,7 +142,7 @@ static inline void tcp_event_data_sent(struct tcp_sock *tp,
                icsk->icsk_ack.pingpong = 1;
 }
 
-static __inline__ void tcp_event_ack_sent(struct sock *sk, unsigned int pkts)
+static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts)
 {
        tcp_dec_quickack_mode(sk, pkts);
        inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK);
@@ -212,7 +212,7 @@ void tcp_select_initial_window(int __space, __u32 mss,
  * value can be stuffed directly into th->window for an outgoing
  * frame.
  */
-static __inline__ u16 tcp_select_window(struct sock *sk)
+static u16 tcp_select_window(struct sock *sk)
 {
        struct tcp_sock *tp = tcp_sk(sk);
        u32 cur_win = tcp_receive_window(tp);
@@ -250,6 +250,75 @@ static __inline__ u16 tcp_select_window(struct sock *sk)
        return new_win;
 }
 
+static void tcp_build_and_update_options(__u32 *ptr, struct tcp_sock *tp,
+                                        __u32 tstamp)
+{
+       if (tp->rx_opt.tstamp_ok) {
+               *ptr++ = __constant_htonl((TCPOPT_NOP << 24) |
+                                         (TCPOPT_NOP << 16) |
+                                         (TCPOPT_TIMESTAMP << 8) |
+                                         TCPOLEN_TIMESTAMP);
+               *ptr++ = htonl(tstamp);
+               *ptr++ = htonl(tp->rx_opt.ts_recent);
+       }
+       if (tp->rx_opt.eff_sacks) {
+               struct tcp_sack_block *sp = tp->rx_opt.dsack ? tp->duplicate_sack : tp->selective_acks;
+               int this_sack;
+
+               *ptr++ = htonl((TCPOPT_NOP  << 24) |
+                              (TCPOPT_NOP  << 16) |
+                              (TCPOPT_SACK <<  8) |
+                              (TCPOLEN_SACK_BASE + (tp->rx_opt.eff_sacks *
+                                                    TCPOLEN_SACK_PERBLOCK)));
+               for(this_sack = 0; this_sack < tp->rx_opt.eff_sacks; this_sack++) {
+                       *ptr++ = htonl(sp[this_sack].start_seq);
+                       *ptr++ = htonl(sp[this_sack].end_seq);
+               }
+               if (tp->rx_opt.dsack) {
+                       tp->rx_opt.dsack = 0;
+                       tp->rx_opt.eff_sacks--;
+               }
+       }
+}
+
+/* Construct a tcp options header for a SYN or SYN_ACK packet.
+ * If this is every changed make sure to change the definition of
+ * MAX_SYN_SIZE to match the new maximum number of options that you
+ * can generate.
+ */
+static void tcp_syn_build_options(__u32 *ptr, int mss, int ts, int sack,
+                                 int offer_wscale, int wscale, __u32 tstamp,
+                                 __u32 ts_recent)
+{
+       /* We always get an MSS option.
+        * The option bytes which will be seen in normal data
+        * packets should timestamps be used, must be in the MSS
+        * advertised.  But we subtract them from tp->mss_cache so
+        * that calculations in tcp_sendmsg are simpler etc.
+        * So account for this fact here if necessary.  If we
+        * don't do this correctly, as a receiver we won't
+        * recognize data packets as being full sized when we
+        * should, and thus we won't abide by the delayed ACK
+        * rules correctly.
+        * SACKs don't matter, we never delay an ACK when we
+        * have any of those going out.
+        */
+       *ptr++ = htonl((TCPOPT_MSS << 24) | (TCPOLEN_MSS << 16) | mss);
+       if (ts) {
+               if(sack)
+                       *ptr++ = __constant_htonl((TCPOPT_SACK_PERM << 24) | (TCPOLEN_SACK_PERM << 16) |
+                                                 (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
+               else
+                       *ptr++ = __constant_htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
+                                                 (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
+               *ptr++ = htonl(tstamp);         /* TSVAL */
+               *ptr++ = htonl(ts_recent);      /* TSECR */
+       } else if(sack)
+               *ptr++ = __constant_htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
+                                         (TCPOPT_SACK_PERM << 8) | TCPOLEN_SACK_PERM);
+       if (offer_wscale)
+               *ptr++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_WINDOW << 16) | (TCPOLEN_WINDOW << 8) | (wscale));
+}
 
 /* This routine actually transmits TCP packets queued in by
  * tcp_do_sendmsg().  This is used by both the initial
@@ -262,122 +331,139 @@ static __inline__ u16 tcp_select_window(struct sock *sk)
  * We are working here with either a clone of the original
  * SKB, or a fresh unique copy made by the retransmit engine.
  */
-static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb)
+static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, gfp_t gfp_mask)
 {
-       if (skb != NULL) {
-               const struct inet_connection_sock *icsk = inet_csk(sk);
-               struct inet_sock *inet = inet_sk(sk);
-               struct tcp_sock *tp = tcp_sk(sk);
-               struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
-               int tcp_header_size = tp->tcp_header_len;
-               struct tcphdr *th;
-               int sysctl_flags;
-               int err;
+       const struct inet_connection_sock *icsk = inet_csk(sk);
+       struct inet_sock *inet;
+       struct tcp_sock *tp;
+       struct tcp_skb_cb *tcb;
+       int tcp_header_size;
+       struct tcphdr *th;
+       int sysctl_flags;
+       int err;
 
-               BUG_ON(!tcp_skb_pcount(skb));
+       BUG_ON(!skb || !tcp_skb_pcount(skb));
+
+       /* If congestion control is doing timestamping, we must
+        * take such a timestamp before we potentially clone/copy.
+        */
+       if (icsk->icsk_ca_ops->rtt_sample)
+               __net_timestamp(skb);
+
+       if (likely(clone_it)) {
+               if (unlikely(skb_cloned(skb)))
+                       skb = pskb_copy(skb, gfp_mask);
+               else
+                       skb = skb_clone(skb, gfp_mask);
+               if (unlikely(!skb))
+                       return -ENOBUFS;
+       }
+
+       inet = inet_sk(sk);
+       tp = tcp_sk(sk);
+       tcb = TCP_SKB_CB(skb);
+       tcp_header_size = tp->tcp_header_len;
 
 #define SYSCTL_FLAG_TSTAMPS    0x1
 #define SYSCTL_FLAG_WSCALE     0x2
 #define SYSCTL_FLAG_SACK       0x4
 
-               /* If congestion control is doing timestamping */
-               if (icsk->icsk_ca_ops->rtt_sample)
-                       __net_timestamp(skb);
-
-               sysctl_flags = 0;
-               if (tcb->flags & TCPCB_FLAG_SYN) {
-                       tcp_header_size = sizeof(struct tcphdr) + TCPOLEN_MSS;
-                       if(sysctl_tcp_timestamps) {
-                               tcp_header_size += TCPOLEN_TSTAMP_ALIGNED;
-                               sysctl_flags |= SYSCTL_FLAG_TSTAMPS;
-                       }
-                       if(sysctl_tcp_window_scaling) {
-                               tcp_header_size += TCPOLEN_WSCALE_ALIGNED;
-                               sysctl_flags |= SYSCTL_FLAG_WSCALE;
-                       }
-                       if(sysctl_tcp_sack) {
-                               sysctl_flags |= SYSCTL_FLAG_SACK;
-                               if(!(sysctl_flags & SYSCTL_FLAG_TSTAMPS))
-                                       tcp_header_size += TCPOLEN_SACKPERM_ALIGNED;
-                       }
-               } else if (tp->rx_opt.eff_sacks) {
-                       /* A SACK is 2 pad bytes, a 2 byte header, plus
-                        * 2 32-bit sequence numbers for each SACK block.
-                        */
-                       tcp_header_size += (TCPOLEN_SACK_BASE_ALIGNED +
-                                           (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK));
+       sysctl_flags = 0;
+       if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) {
+               tcp_header_size = sizeof(struct tcphdr) + TCPOLEN_MSS;
+               if(sysctl_tcp_timestamps) {
+                       tcp_header_size += TCPOLEN_TSTAMP_ALIGNED;
+                       sysctl_flags |= SYSCTL_FLAG_TSTAMPS;
                }
-               
-               if (tcp_packets_in_flight(tp) == 0)
-                       tcp_ca_event(sk, CA_EVENT_TX_START);
-
-               th = (struct tcphdr *) skb_push(skb, tcp_header_size);
-               skb->h.th = th;
-               skb_set_owner_w(skb, sk);
-
-               /* Build TCP header and checksum it. */
-               th->source              = inet->sport;
-               th->dest                = inet->dport;
-               th->seq                 = htonl(tcb->seq);
-               th->ack_seq             = htonl(tp->rcv_nxt);
-               *(((__u16 *)th) + 6)    = htons(((tcp_header_size >> 2) << 12) | tcb->flags);
-               if (tcb->flags & TCPCB_FLAG_SYN) {
-                       /* RFC1323: The window in SYN & SYN/ACK segments
-                        * is never scaled.
-                        */
-                       th->window      = htons(tp->rcv_wnd);
-               } else {
-                       th->window      = htons(tcp_select_window(sk));
+               if (sysctl_tcp_window_scaling) {
+                       tcp_header_size += TCPOLEN_WSCALE_ALIGNED;
+                       sysctl_flags |= SYSCTL_FLAG_WSCALE;
                }
-               th->check               = 0;
-               th->urg_ptr             = 0;
-
-               if (tp->urg_mode &&
-                   between(tp->snd_up, tcb->seq+1, tcb->seq+0xFFFF)) {
-                       th->urg_ptr             = htons(tp->snd_up-tcb->seq);
-                       th->urg                 = 1;
+               if (sysctl_tcp_sack) {
+                       sysctl_flags |= SYSCTL_FLAG_SACK;
+                       if (!(sysctl_flags & SYSCTL_FLAG_TSTAMPS))
+                               tcp_header_size += TCPOLEN_SACKPERM_ALIGNED;
                }
+       } else if (unlikely(tp->rx_opt.eff_sacks)) {
+               /* A SACK is 2 pad bytes, a 2 byte header, plus
+                * 2 32-bit sequence numbers for each SACK block.
+                */
+               tcp_header_size += (TCPOLEN_SACK_BASE_ALIGNED +
+                                   (tp->rx_opt.eff_sacks *
+                                    TCPOLEN_SACK_PERBLOCK));
+       }
+               
+       if (tcp_packets_in_flight(tp) == 0)
+               tcp_ca_event(sk, CA_EVENT_TX_START);
+
+       th = (struct tcphdr *) skb_push(skb, tcp_header_size);
+       skb->h.th = th;
+       skb_set_owner_w(skb, sk);
+
+       /* Build TCP header and checksum it. */
+       th->source              = inet->sport;
+       th->dest                = inet->dport;
+       th->seq                 = htonl(tcb->seq);
+       th->ack_seq             = htonl(tp->rcv_nxt);
+       *(((__u16 *)th) + 6)    = htons(((tcp_header_size >> 2) << 12) |
+                                       tcb->flags);
+
+       if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) {
+               /* RFC1323: The window in SYN & SYN/ACK segments
+                * is never scaled.
+                */
+               th->window      = htons(tp->rcv_wnd);
+       } else {
+               th->window      = htons(tcp_select_window(sk));
+       }
+       th->check               = 0;
+       th->urg_ptr             = 0;
 
-               if (tcb->flags & TCPCB_FLAG_SYN) {
-                       tcp_syn_build_options((__u32 *)(th + 1),
-                                             tcp_advertise_mss(sk),
-                                             (sysctl_flags & SYSCTL_FLAG_TSTAMPS),
-                                             (sysctl_flags & SYSCTL_FLAG_SACK),
-                                             (sysctl_flags & SYSCTL_FLAG_WSCALE),
-                                             tp->rx_opt.rcv_wscale,
-                                             tcb->when,
-                                             tp->rx_opt.ts_recent);
-               } else {
-                       tcp_build_and_update_options((__u32 *)(th + 1),
-                                                    tp, tcb->when);
+       if (unlikely(tp->urg_mode &&
+                    between(tp->snd_up, tcb->seq+1, tcb->seq+0xFFFF))) {
+               th->urg_ptr             = htons(tp->snd_up-tcb->seq);
+               th->urg                 = 1;
+       }
 
-                       TCP_ECN_send(sk, tp, skb, tcp_header_size);
-               }
-               tp->af_specific->send_check(sk, th, skb->len, skb);
+       if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) {
+               tcp_syn_build_options((__u32 *)(th + 1),
+                                     tcp_advertise_mss(sk),
+                                     (sysctl_flags & SYSCTL_FLAG_TSTAMPS),
+                                     (sysctl_flags & SYSCTL_FLAG_SACK),
+                                     (sysctl_flags & SYSCTL_FLAG_WSCALE),
+                                     tp->rx_opt.rcv_wscale,
+                                     tcb->when,
+                                     tp->rx_opt.ts_recent);
+       } else {
+               tcp_build_and_update_options((__u32 *)(th + 1),
+                                            tp, tcb->when);
+               TCP_ECN_send(sk, tp, skb, tcp_header_size);
+       }
 
-               if (tcb->flags & TCPCB_FLAG_ACK)
-                       tcp_event_ack_sent(sk, tcp_skb_pcount(skb));
+       icsk->icsk_af_ops->send_check(sk, skb->len, skb);
 
-               if (skb->len != tcp_header_size)
-                       tcp_event_data_sent(tp, skb, sk);
+       if (likely(tcb->flags & TCPCB_FLAG_ACK))
+               tcp_event_ack_sent(sk, tcp_skb_pcount(skb));
 
-               TCP_INC_STATS(TCP_MIB_OUTSEGS);
+       if (skb->len != tcp_header_size)
+               tcp_event_data_sent(tp, skb, sk);
 
-               err = tp->af_specific->queue_xmit(skb, 0);
-               if (err <= 0)
-                       return err;
+       TCP_INC_STATS(TCP_MIB_OUTSEGS);
 
-               tcp_enter_cwr(sk);
+       err = icsk->icsk_af_ops->queue_xmit(skb, 0);
+       if (unlikely(err <= 0))
+               return err;
+
+       tcp_enter_cwr(sk);
+
+       /* NET_XMIT_CN is special. It does not guarantee,
+        * that this packet is lost. It tells that device
+        * is about to start to drop packets or already
+        * drops some packets of the same priority and
+        * invokes us to send less aggressively.
+        */
+       return err == NET_XMIT_CN ? 0 : err;
 
-               /* NET_XMIT_CN is special. It does not guarantee,
-                * that this packet is lost. It tells that device
-                * is about to start to drop packets or already
-                * drops some packets of the same priority and
-                * invokes us to send less aggressively.
-                */
-               return err == NET_XMIT_CN ? 0 : err;
-       }
-       return -ENOBUFS;
 #undef SYSCTL_FLAG_TSTAMPS
 #undef SYSCTL_FLAG_WSCALE
 #undef SYSCTL_FLAG_SACK
@@ -435,17 +521,9 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss
        int nsize, old_factor;
        u16 flags;
 
-       if (unlikely(len >= skb->len)) {
-               if (net_ratelimit()) {
-                       printk(KERN_DEBUG "TCP: seg_size=%u, mss=%u, seq=%u, "
-                              "end_seq=%u, skb->len=%u.\n", len, mss_now,
-                              TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq,
-                              skb->len);
-                       WARN_ON(1);
-               }
-               return 0;
-       }
+       BUG_ON(len > skb->len);
 
+       clear_all_retrans_hints(tp);
        nsize = skb_headlen(skb) - len;
        if (nsize < 0)
                nsize = 0;
@@ -609,10 +687,10 @@ int tcp_trim_head(struct sock *sk, struct sk_buff *skb, u32 len)
    for TCP options, but includes only bare TCP header.
 
    tp->rx_opt.mss_clamp is mss negotiated at connection setup.
-   It is minumum of user_mss and mss received with SYN.
+   It is minimum of user_mss and mss received with SYN.
    It also does not include TCP options.
 
-   tp->pmtu_cookie is last pmtu, seen by this function.
+   inet_csk(sk)->icsk_pmtu_cookie is last pmtu, seen by this function.
 
    tp->mss_cache is current effective sending mss, including
    all tcp options except for SACKs. It is evaluated,
@@ -622,26 +700,26 @@ int tcp_trim_head(struct sock *sk, struct sk_buff *skb, u32 len)
    NOTE1. rfc1122 clearly states that advertised MSS
    DOES NOT include either tcp or ip options.
 
-   NOTE2. tp->pmtu_cookie and tp->mss_cache are READ ONLY outside
-   this function.                      --ANK (980731)
+   NOTE2. inet_csk(sk)->icsk_pmtu_cookie and tp->mss_cache
+   are READ ONLY outside this function.                --ANK (980731)
  */
 
 unsigned int tcp_sync_mss(struct sock *sk, u32 pmtu)
 {
        struct tcp_sock *tp = tcp_sk(sk);
-       int mss_now;
-
+       struct inet_connection_sock *icsk = inet_csk(sk);
        /* Calculate base mss without TCP options:
           It is MMS_S - sizeof(tcphdr) of rfc1122
         */
-       mss_now = pmtu - tp->af_specific->net_header_len - sizeof(struct tcphdr);
+       int mss_now = (pmtu - icsk->icsk_af_ops->net_header_len -
+                      sizeof(struct tcphdr));
 
        /* Clamp it (mss_clamp does not include tcp options) */
        if (mss_now > tp->rx_opt.mss_clamp)
                mss_now = tp->rx_opt.mss_clamp;
 
        /* Now subtract optional transport overhead */
-       mss_now -= tp->ext_header_len;
+       mss_now -= icsk->icsk_ext_hdr_len;
 
        /* Then reserve room for full set of TCP options and 8 bytes of data */
        if (mss_now < 48)
@@ -655,7 +733,7 @@ unsigned int tcp_sync_mss(struct sock *sk, u32 pmtu)
                mss_now = max((tp->max_window>>1), 68U - tp->tcp_header_len);
 
        /* And store cached results */
-       tp->pmtu_cookie = pmtu;
+       icsk->icsk_pmtu_cookie = pmtu;
        tp->mss_cache = mss_now;
 
        return mss_now;
@@ -685,7 +763,7 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed)
 
        if (dst) {
                u32 mtu = dst_mtu(dst);
-               if (mtu != tp->pmtu_cookie)
+               if (mtu != inet_csk(sk)->icsk_pmtu_cookie)
                        mss_now = tcp_sync_mss(sk, mtu);
        }
 
@@ -696,9 +774,10 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed)
        xmit_size_goal = mss_now;
 
        if (doing_tso) {
-               xmit_size_goal = 65535 -
-                       tp->af_specific->net_header_len -
-                       tp->ext_header_len - tp->tcp_header_len;
+               xmit_size_goal = (65535 -
+                                 inet_csk(sk)->icsk_af_ops->net_header_len -
+                                 inet_csk(sk)->icsk_ext_hdr_len -
+                                 tp->tcp_header_len);
 
                if (tp->max_window &&
                    (xmit_size_goal > (tp->max_window >> 1)))
@@ -714,7 +793,7 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed)
 
 /* Congestion window validation. (RFC2861) */
 
-static inline void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp)
+static void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp)
 {
        __u32 packets_out = tp->packets_out;
 
@@ -763,7 +842,7 @@ static inline unsigned int tcp_cwnd_test(struct tcp_sock *tp, struct sk_buff *sk
 /* This must be invoked the first time we consider transmitting
  * SKB onto the wire.
  */
-static inline int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now)
+static int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now)
 {
        int tso_segs = tcp_skb_pcount(skb);
 
@@ -1044,7 +1123,7 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle)
 
                TCP_SKB_CB(skb)->when = tcp_time_stamp;
 
-               if (unlikely(tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC))))
+               if (unlikely(tcp_transmit_skb(sk, skb, 1, GFP_ATOMIC)))
                        break;
 
                /* Advance the send_head.  This one is sent out.
@@ -1117,7 +1196,7 @@ void tcp_push_one(struct sock *sk, unsigned int mss_now)
                /* Send it out now. */
                TCP_SKB_CB(skb)->when = tcp_time_stamp;
 
-               if (likely(!tcp_transmit_skb(sk, skb_clone(skb, sk->sk_allocation)))) {
+               if (likely(!tcp_transmit_skb(sk, skb, 1, sk->sk_allocation))) {
                        update_send_head(sk, tp, skb);
                        tcp_cwnd_validate(sk, tp);
                        return;
@@ -1181,7 +1260,7 @@ u32 __tcp_select_window(struct sock *sk)
 {
        struct inet_connection_sock *icsk = inet_csk(sk);
        struct tcp_sock *tp = tcp_sk(sk);
-       /* MSS for the peer's data.  Previous verions used mss_clamp
+       /* MSS for the peer's data.  Previous versions used mss_clamp
         * here.  I don't know if the value based on our guesses
         * of peer's MSS is better for the performance.  It's more correct
         * but may be worse for the performance because of rcv_mss
@@ -1270,7 +1349,10 @@ static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *skb, int m
                BUG_ON(tcp_skb_pcount(skb) != 1 ||
                       tcp_skb_pcount(next_skb) != 1);
 
-               /* Ok.  We will be able to collapse the packet. */
+               /* changing transmit queue under us so clear hints */
+               clear_all_retrans_hints(tp);
+
+               /* Ok.  We will be able to collapse the packet. */
                __skb_unlink(next_skb, &sk->sk_write_queue);
 
                memcpy(skb_put(skb, next_skb_size), next_skb->data, next_skb_size);
@@ -1340,6 +1422,8 @@ void tcp_simple_retransmit(struct sock *sk)
                }
        }
 
+       clear_all_retrans_hints(tp);
+
        if (!lost)
                return;
 
@@ -1371,7 +1455,7 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
        int err;
 
        /* Do not sent more than we queued. 1/4 is reserved for possible
-        * copying overhead: frgagmentation, tunneling, mangling etc.
+        * copying overhead: fragmentation, tunneling, mangling etc.
         */
        if (atomic_read(&sk->sk_wmem_alloc) >
            min(sk->sk_wmem_queued + (sk->sk_wmem_queued >> 2), sk->sk_sndbuf))
@@ -1408,7 +1492,7 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
           (sysctl_tcp_retrans_collapse != 0))
                tcp_retrans_try_collapse(sk, skb, cur_mss);
 
-       if(tp->af_specific->rebuild_header(sk))
+       if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk))
                return -EHOSTUNREACH; /* Routing failure or similar. */
 
        /* Some Solaris stacks overoptimize and ignore the FIN on a
@@ -1432,9 +1516,7 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
         */
        TCP_SKB_CB(skb)->when = tcp_time_stamp;
 
-       err = tcp_transmit_skb(sk, (skb_cloned(skb) ?
-                                   pskb_copy(skb, GFP_ATOMIC):
-                                   skb_clone(skb, GFP_ATOMIC)));
+       err = tcp_transmit_skb(sk, skb, 1, GFP_ATOMIC);
 
        if (err == 0) {
                /* Update global TCP statistics. */
@@ -1478,13 +1560,25 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
        const struct inet_connection_sock *icsk = inet_csk(sk);
        struct tcp_sock *tp = tcp_sk(sk);
        struct sk_buff *skb;
-       int packet_cnt = tp->lost_out;
+       int packet_cnt;
+
+       if (tp->retransmit_skb_hint) {
+               skb = tp->retransmit_skb_hint;
+               packet_cnt = tp->retransmit_cnt_hint;
+       }else{
+               skb = sk->sk_write_queue.next;
+               packet_cnt = 0;
+       }
 
        /* First pass: retransmit lost packets. */
-       if (packet_cnt) {
-               sk_stream_for_retrans_queue(skb, sk) {
+       if (tp->lost_out) {
+               sk_stream_for_retrans_queue_from(skb, sk) {
                        __u8 sacked = TCP_SKB_CB(skb)->sacked;
 
+                       /* we could do better than to assign each time */
+                       tp->retransmit_skb_hint = skb;
+                       tp->retransmit_cnt_hint = packet_cnt;
+
                        /* Assume this retransmit will generate
                         * only one packet for congestion window
                         * calculation purposes.  This works because
@@ -1495,10 +1589,12 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
                        if (tcp_packets_in_flight(tp) >= tp->snd_cwnd)
                                return;
 
-                       if (sacked&TCPCB_LOST) {
+                       if (sacked & TCPCB_LOST) {
                                if (!(sacked&(TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))) {
-                                       if (tcp_retransmit_skb(sk, skb))
+                                       if (tcp_retransmit_skb(sk, skb)) {
+                                               tp->retransmit_skb_hint = NULL;
                                                return;
+                                       }
                                        if (icsk->icsk_ca_state != TCP_CA_Loss)
                                                NET_INC_STATS_BH(LINUX_MIB_TCPFASTRETRANS);
                                        else
@@ -1511,8 +1607,8 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
                                                                          TCP_RTO_MAX);
                                }
 
-                               packet_cnt -= tcp_skb_pcount(skb);
-                               if (packet_cnt <= 0)
+                               packet_cnt += tcp_skb_pcount(skb);
+                               if (packet_cnt >= tp->lost_out)
                                        break;
                        }
                }
@@ -1538,9 +1634,18 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
        if (tcp_may_send_now(sk, tp))
                return;
 
-       packet_cnt = 0;
+       if (tp->forward_skb_hint) {
+               skb = tp->forward_skb_hint;
+               packet_cnt = tp->forward_cnt_hint;
+       } else{
+               skb = sk->sk_write_queue.next;
+               packet_cnt = 0;
+       }
+
+       sk_stream_for_retrans_queue_from(skb, sk) {
+               tp->forward_cnt_hint = packet_cnt;
+               tp->forward_skb_hint = skb;
 
-       sk_stream_for_retrans_queue(skb, sk) {
                /* Similar to the retransmit loop above we
                 * can pretend that the retransmitted SKB
                 * we send out here will be composed of one
@@ -1557,8 +1662,10 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
                        continue;
 
                /* Ok, retransmit it. */
-               if (tcp_retransmit_skb(sk, skb))
+               if (tcp_retransmit_skb(sk, skb)) {
+                       tp->forward_skb_hint = NULL;
                        break;
+               }
 
                if (skb == skb_peek(&sk->sk_write_queue))
                        inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
@@ -1643,7 +1750,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority)
        TCP_SKB_CB(skb)->seq = tcp_acceptable_seq(sk, tp);
        TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq;
        TCP_SKB_CB(skb)->when = tcp_time_stamp;
-       if (tcp_transmit_skb(sk, skb))
+       if (tcp_transmit_skb(sk, skb, 0, priority))
                NET_INC_STATS(LINUX_MIB_TCPABORTFAILED);
 }
 
@@ -1678,7 +1785,7 @@ int tcp_send_synack(struct sock *sk)
                TCP_ECN_send_synack(tcp_sk(sk), skb);
        }
        TCP_SKB_CB(skb)->when = tcp_time_stamp;
-       return tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC));
+       return tcp_transmit_skb(sk, skb, 1, GFP_ATOMIC);
 }
 
 /*
@@ -1756,7 +1863,7 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 /* 
  * Do all connect socket setups that can be done AF independent.
  */ 
-static inline void tcp_connect_init(struct sock *sk)
+static void tcp_connect_init(struct sock *sk)
 {
        struct dst_entry *dst = __sk_dst_get(sk);
        struct tcp_sock *tp = tcp_sk(sk);
@@ -1839,7 +1946,7 @@ int tcp_connect(struct sock *sk)
        __skb_queue_tail(&sk->sk_write_queue, buff);
        sk_charge_skb(sk, buff);
        tp->packets_out += tcp_skb_pcount(buff);
-       tcp_transmit_skb(sk, skb_clone(buff, GFP_KERNEL));
+       tcp_transmit_skb(sk, buff, 1, GFP_KERNEL);
        TCP_INC_STATS(TCP_MIB_ACTIVEOPENS);
 
        /* Timer for repeating the SYN until an answer. */
@@ -1935,7 +2042,7 @@ void tcp_send_ack(struct sock *sk)
                /* Send it off, this clears delayed acks for us. */
                TCP_SKB_CB(buff)->seq = TCP_SKB_CB(buff)->end_seq = tcp_acceptable_seq(sk, tp);
                TCP_SKB_CB(buff)->when = tcp_time_stamp;
-               tcp_transmit_skb(sk, buff);
+               tcp_transmit_skb(sk, buff, 0, GFP_ATOMIC);
        }
 }
 
@@ -1975,7 +2082,7 @@ static int tcp_xmit_probe_skb(struct sock *sk, int urgent)
        TCP_SKB_CB(skb)->seq = urgent ? tp->snd_una : tp->snd_una - 1;
        TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq;
        TCP_SKB_CB(skb)->when = tcp_time_stamp;
-       return tcp_transmit_skb(sk, skb);
+       return tcp_transmit_skb(sk, skb, 0, GFP_ATOMIC);
 }
 
 int tcp_write_wakeup(struct sock *sk)
@@ -2008,7 +2115,7 @@ int tcp_write_wakeup(struct sock *sk)
 
                        TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
                        TCP_SKB_CB(skb)->when = tcp_time_stamp;
-                       err = tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC));
+                       err = tcp_transmit_skb(sk, skb, 1, GFP_ATOMIC);
                        if (!err) {
                                update_send_head(sk, tp, skb);
                        }
@@ -2068,3 +2175,4 @@ EXPORT_SYMBOL(tcp_connect);
 EXPORT_SYMBOL(tcp_make_synack);
 EXPORT_SYMBOL(tcp_simple_retransmit);
 EXPORT_SYMBOL(tcp_sync_mss);
+EXPORT_SYMBOL(sysctl_tcp_tso_win_divisor);