10 files changed, 67 insertions, 37 deletions

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 800930fac388..620096840744 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -270,7 +270,7 @@ struct tcp_sock {
         __u8    frto_counter;   /* Number of new acks after RTO */
 
         __u8    nonagle;        /* Disable Nagle algorithm?             */
-        __u8    defer_accept;   /* User waits for some data after accept() */
+        /* ONE BYTE HOLE, TRY TO PACK */
 
 /* RTT measurement */
         __u32   srtt;           /* smoothed round trip time << 3        */
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index a50f4a4b7b4b..692825fc8135 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -239,4 +239,6 @@ static inline void inet_csk_reqsk_queue_drop(struct sock *sk,
         reqsk_free(req);
 }
 
+extern void inet_csk_listen_stop(struct sock *sk);
+
 #endif /* _INET_CONNECTION_SOCK_H */
diff --git a/include/net/request_sock.h b/include/net/request_sock.h
index b7c7eecbe64d..447d287a38fd 100644
--- a/include/net/request_sock.h
+++ b/include/net/request_sock.h
@@ -97,6 +97,7 @@ struct listen_sock {
  *
  * @rskq_accept_head - FIFO head of established children
  * @rskq_accept_tail - FIFO tail of established children
+ * @rskq_defer_accept - User waits for some data after accept()
  * @syn_wait_lock - serializer
  *
  * %syn_wait_lock is necessary only to avoid proc interface having to grab the main
@@ -112,6 +113,8 @@ struct request_sock_queue {
         struct request_sock     *rskq_accept_head;
         struct request_sock     *rskq_accept_tail;
         rwlock_t                syn_wait_lock;
+        u8                      rskq_defer_accept;
+        /* 3 bytes hole, try to pack */
         struct listen_sock      *listen_opt;
 };
 
@@ -255,4 +258,8 @@ static inline void reqsk_queue_hash_req(struct request_sock_queue *queue,
         write_unlock(&queue->syn_wait_lock);
 }
 
+extern void reqsk_queue_prune(struct request_sock_queue *queue, struct sock *parent,
+                              const unsigned long interval, const unsigned long timeout,
+                              const unsigned long max_rto, int max_retries);
+
 #endif /* _REQUEST_SOCK_H */
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 68f1ec1c583a..2423f059b62b 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -423,7 +423,8 @@ extern int			tcp_recvmsg(struct kiocb *iocb, struct sock *sk,
                                             size_t len, int nonblock, 
                                             int flags, int *addr_len);
 
-extern int                      tcp_listen_start(struct sock *sk);
+extern int                      inet_csk_listen_start(struct sock *sk,
+                                                      const int nr_table_entries);
 
 extern void                     tcp_parse_options(struct sk_buff *skb,
                                                   struct tcp_options_received *opt_rx,
diff --git a/net/core/request_sock.c b/net/core/request_sock.c
index 98f0fc923f91..b8203de5ff07 100644
--- a/net/core/request_sock.c
+++ b/net/core/request_sock.c
@@ -52,6 +52,7 @@ int reqsk_queue_alloc(struct request_sock_queue *queue,
         get_random_bytes(&lopt->hash_rnd, sizeof(lopt->hash_rnd));
         rwlock_init(&queue->syn_wait_lock);
         queue->rskq_accept_head = queue->rskq_accept_head = NULL;
+        queue->rskq_defer_accept = 0;
         lopt->nr_table_entries = nr_table_entries;
 
         write_lock_bh(&queue->syn_wait_lock);
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index f691058cf599..52f5ecc58c46 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -99,6 +99,7 @@
 #include <net/arp.h>
 #include <net/route.h>
 #include <net/ip_fib.h>
+#include <net/inet_connection_sock.h>
 #include <net/tcp.h>
 #include <net/udp.h>
 #include <linux/skbuff.h>
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index a1f812159ced..a4e9eec44895 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -495,7 +495,7 @@ EXPORT_SYMBOL_GPL(inet_csk_listen_start);
  *      This routine closes sockets which have been at least partially
  *      opened, but not yet accepted.
  */
-static void inet_csk_listen_stop(struct sock *sk)
+void inet_csk_listen_stop(struct sock *sk)
 {
         struct inet_connection_sock *icsk = inet_csk(sk);
         struct request_sock *acc_req;
@@ -1947,15 +1947,15 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
                 break;
 
         case TCP_DEFER_ACCEPT:
-                tp->defer_accept = 0;
+                icsk->icsk_accept_queue.rskq_defer_accept = 0;
                 if (val > 0) {
                         /* Translate value in seconds to number of
                          * retransmits */
-                        while (tp->defer_accept < 32 &&
+                        while (icsk->icsk_accept_queue.rskq_defer_accept < 32 &&
                                val > ((TCP_TIMEOUT_INIT / HZ) <<
-                                       tp->defer_accept))
-                                tp->defer_accept++;
-                        tp->defer_accept++;
+                                       icsk->icsk_accept_queue.rskq_defer_accept))
+                                icsk->icsk_accept_queue.rskq_defer_accept++;
+                        icsk->icsk_accept_queue.rskq_defer_accept++;
                 }
                 break;
 
@@ -2058,6 +2058,7 @@ EXPORT_SYMBOL_GPL(tcp_get_info);
 int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
                    int __user *optlen)
 {
+        struct inet_connection_sock *icsk = inet_csk(sk);
         struct tcp_sock *tp = tcp_sk(sk);
         int val, len;
 
@@ -2095,7 +2096,7 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
                 val = tp->keepalive_probes ? : sysctl_tcp_keepalive_probes;
                 break;
         case TCP_SYNCNT:
-                val = inet_csk(sk)->icsk_syn_retries ? : sysctl_tcp_syn_retries;
+                val = icsk->icsk_syn_retries ? : sysctl_tcp_syn_retries;
                 break;
         case TCP_LINGER2:
                 val = tp->linger2;
@@ -2103,8 +2104,8 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
                         val = (val ? : sysctl_tcp_fin_timeout) / HZ;
                 break;
         case TCP_DEFER_ACCEPT:
-                val = !tp->defer_accept ? 0 : ((TCP_TIMEOUT_INIT / HZ) <<
-                                               (tp->defer_accept - 1));
+                val = !icsk->icsk_accept_queue.rskq_defer_accept ? 0 :
+                        ((TCP_TIMEOUT_INIT / HZ) << (icsk->icsk_accept_queue.rskq_defer_accept - 1));
                 break;
         case TCP_WINDOW_CLAMP:
                 val = tp->window_clamp;
@@ -2125,7 +2126,7 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
                 return 0;
         }
         case TCP_QUICKACK:
-                val = !inet_csk(sk)->icsk_ack.pingpong;
+                val = !icsk->icsk_ack.pingpong;
                 break;
 
         case TCP_CONGESTION:
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index b35badf53aa5..71d456148de7 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3831,6 +3831,7 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
         tcp_parse_options(skb, &tp->rx_opt, 0);
 
         if (th->ack) {
+                struct inet_connection_sock *icsk;
                 /* rfc793:
                  * "If the state is SYN-SENT then
                  *    first check the ACK bit
@@ -3956,7 +3957,11 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
                         sk_wake_async(sk, 0, POLL_OUT);
                 }
 
-                if (sk->sk_write_pending || tp->defer_accept || inet_csk(sk)->icsk_ack.pingpong) {
+                icsk = inet_csk(sk);
+
+                if (sk->sk_write_pending ||
+                    icsk->icsk_accept_queue.rskq_defer_accept ||
+                    icsk->icsk_ack.pingpong) {
                         /* Save one ACK. Data will be ready after
                          * several ticks, if write_pending is set.
                          *
@@ -3965,8 +3970,8 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
                          * to stand against the temptation 8)     --ANK
                          */
                         inet_csk_schedule_ack(sk);
-                        inet_csk(sk)->icsk_ack.lrcvtime = tcp_time_stamp;
-                        inet_csk(sk)->icsk_ack.ato       = TCP_ATO_MIN;
+                        icsk->icsk_ack.lrcvtime = tcp_time_stamp;
+                        icsk->icsk_ack.ato       = TCP_ATO_MIN;
                         tcp_incr_quickack(sk);
                         tcp_enter_quickack_mode(sk);
                         inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK,
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 4cfbe1d1c920..2d95afe5b393 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -787,9 +787,10 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
            does sequence test, SYN is truncated, and thus we consider
            it a bare ACK.
 
-           If tp->defer_accept, we silently drop this bare ACK.  Otherwise,
-           we create an established connection.  Both ends (listening sockets)
-           accept the new incoming connection and try to talk to each other. 8-)
+           If icsk->icsk_accept_queue.rskq_defer_accept, we silently drop this
+           bare ACK.  Otherwise, we create an established connection.  Both
+           ends (listening sockets) accept the new incoming connection and try
+           to talk to each other. 8-)
 
            Note: This case is both harmless, and rare.  Possibility is about the
            same as us discovering intelligent life on another plant tomorrow.
@@ -856,7 +857,8 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
                         return NULL;
 
                 /* If TCP_DEFER_ACCEPT is set, drop bare ACK. */
-                if (tp->defer_accept && TCP_SKB_CB(skb)->end_seq == tcp_rsk(req)->rcv_isn + 1) {
+                if (inet_csk(sk)->icsk_accept_queue.rskq_defer_accept &&
+                    TCP_SKB_CB(skb)->end_seq == tcp_rsk(req)->rcv_isn + 1) {
                         inet_rsk(req)->acked = 1;
                         return NULL;
                 }
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index c03930c48f42..b614ad4d30c9 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -424,16 +424,12 @@ out_unlock:
         sock_put(sk);
 }
 
-/*
- *      Timer for listening sockets
- */
-
-static void tcp_synack_timer(struct sock *sk)
+void reqsk_queue_prune(struct request_sock_queue *queue, struct sock *parent,
+                       const unsigned long interval, const unsigned long timeout,
+                       const unsigned long max_rto, int max_retries)
 {
-        struct tcp_sock *tp = tcp_sk(sk);
-        struct inet_connection_sock *icsk = inet_csk(sk);
-        struct listen_sock *lopt = icsk->icsk_accept_queue.listen_opt;
-        int max_retries = icsk->icsk_syn_retries ? : sysctl_tcp_synack_retries;
+        struct inet_connection_sock *icsk = inet_csk(parent);
+        struct listen_sock *lopt = queue->listen_opt;
         int thresh = max_retries;
         unsigned long now = jiffies;
         struct request_sock **reqp, *req;
@@ -470,10 +466,10 @@ static void tcp_synack_timer(struct sock *sk)
                 }
         }
 
-        if (tp->defer_accept)
-                max_retries = tp->defer_accept;
+        if (queue->rskq_defer_accept)
+                max_retries = queue->rskq_defer_accept;
 
-        budget = 2*(TCP_SYNQ_HSIZE/(TCP_TIMEOUT_INIT/TCP_SYNQ_INTERVAL));
+        budget = 2 * (lopt->nr_table_entries / (timeout / interval));
         i = lopt->clock_hand;
 
         do {
@@ -482,20 +478,19 @@ static void tcp_synack_timer(struct sock *sk)
                         if (time_after_eq(now, req->expires)) {
                                 if ((req->retrans < thresh ||
                                      (inet_rsk(req)->acked && req->retrans < max_retries))
-                                    && !req->rsk_ops->rtx_syn_ack(sk, req, NULL)) {
+                                    && !req->rsk_ops->rtx_syn_ack(parent, req, NULL)) {
                                         unsigned long timeo;
 
                                         if (req->retrans++ == 0)
                                                 lopt->qlen_young--;
-                                        timeo = min((TCP_TIMEOUT_INIT << req->retrans),
-                                                    TCP_RTO_MAX);
+                                        timeo = min((timeout << req->retrans), max_rto);
                                         req->expires = now + timeo;
                                         reqp = &req->dl_next;
                                         continue;
                                 }
 
                                 /* Drop this request */
-                                inet_csk_reqsk_queue_unlink(sk, req, reqp);
+                                inet_csk_reqsk_queue_unlink(parent, req, reqp);
                                 reqsk_queue_removed(&icsk->icsk_accept_queue, req);
                                 reqsk_free(req);
                                 continue;
@@ -503,14 +498,29 @@ static void tcp_synack_timer(struct sock *sk)
                         reqp = &req->dl_next;
                 }
 
-                i = (i+1)&(TCP_SYNQ_HSIZE-1);
+                i = (i + 1) & (lopt->nr_table_entries - 1);
 
         } while (--budget > 0);
 
         lopt->clock_hand = i;
 
         if (lopt->qlen)
-                inet_csk_reset_keepalive_timer(sk, TCP_SYNQ_INTERVAL);
+                inet_csk_reset_keepalive_timer(parent, interval);
+}
+
+EXPORT_SYMBOL_GPL(reqsk_queue_prune);
+
+/*
+ *      Timer for listening sockets
+ */
+
+static void tcp_synack_timer(struct sock *sk)
+{
+        struct inet_connection_sock *icsk = inet_csk(sk);
+        const int max_retries = icsk->icsk_syn_retries ? : sysctl_tcp_synack_retries;
+
+        reqsk_queue_prune(&icsk->icsk_accept_queue, sk, TCP_SYNQ_INTERVAL,
+                          TCP_TIMEOUT_INIT, TCP_RTO_MAX, max_retries);
 }
 
 void tcp_set_keepalive(struct sock *sk, int val)