]> err.no Git - linux-2.6/blobdiff - net/ipv4/tcp_minisocks.c
Merge refs/heads/ieee80211-wifi from master.kernel.org:/pub/scm/linux/kernel/git...
[linux-2.6] / net / ipv4 / tcp_minisocks.c
index 5b5a49335fbbdc787e562835aa49ce00828b57c6..a88db28b0af7d83db35c516cabcc337169a88aa3 100644 (file)
 #define SYNC_INIT 1
 #endif
 
-int sysctl_tcp_tw_recycle;
-int sysctl_tcp_max_tw_buckets = NR_FILE*2;
-
 int sysctl_tcp_syncookies = SYNC_INIT; 
 int sysctl_tcp_abort_on_overflow;
 
-static void tcp_tw_schedule(struct inet_timewait_sock *tw, int timeo);
+struct inet_timewait_death_row tcp_death_row = {
+       .sysctl_max_tw_buckets = NR_FILE * 2,
+       .period         = TCP_TIMEWAIT_LEN / INET_TWDR_TWKILL_SLOTS,
+       .death_lock     = SPIN_LOCK_UNLOCKED,
+       .hashinfo       = &tcp_hashinfo,
+       .tw_timer       = TIMER_INITIALIZER(inet_twdr_hangman, 0,
+                                           (unsigned long)&tcp_death_row),
+       .twkill_work    = __WORK_INITIALIZER(tcp_death_row.twkill_work,
+                                            inet_twdr_twkill_work,
+                                            &tcp_death_row),
+/* Short-time timewait calendar */
+
+       .twcal_hand     = -1,
+       .twcal_timer    = TIMER_INITIALIZER(inet_twdr_twcal_tick, 0,
+                                           (unsigned long)&tcp_death_row),
+};
+
+EXPORT_SYMBOL_GPL(tcp_death_row);
 
 static __inline__ int tcp_in_window(u32 seq, u32 end_seq, u32 s_win, u32 e_win)
 {
@@ -52,46 +66,6 @@ static __inline__ int tcp_in_window(u32 seq, u32 end_seq, u32 s_win, u32 e_win)
        return (seq == e_win && seq == end_seq);
 }
 
-/* New-style handling of TIME_WAIT sockets. */
-
-int tcp_tw_count;
-
-
-/* Must be called with locally disabled BHs. */
-static void tcp_timewait_kill(struct inet_timewait_sock *tw)
-{
-       struct inet_bind_hashbucket *bhead;
-       struct inet_bind_bucket *tb;
-       /* Unlink from established hashes. */
-       struct inet_ehash_bucket *ehead = &tcp_hashinfo.ehash[tw->tw_hashent];
-
-       write_lock(&ehead->lock);
-       if (hlist_unhashed(&tw->tw_node)) {
-               write_unlock(&ehead->lock);
-               return;
-       }
-       __hlist_del(&tw->tw_node);
-       sk_node_init(&tw->tw_node);
-       write_unlock(&ehead->lock);
-
-       /* Disassociate with bind bucket. */
-       bhead = &tcp_hashinfo.bhash[inet_bhashfn(tw->tw_num, tcp_hashinfo.bhash_size)];
-       spin_lock(&bhead->lock);
-       tb = tw->tw_tb;
-       __hlist_del(&tw->tw_bind_node);
-       tw->tw_tb = NULL;
-       inet_bind_bucket_destroy(tcp_hashinfo.bind_bucket_cachep, tb);
-       spin_unlock(&bhead->lock);
-
-#ifdef SOCK_REFCNT_DEBUG
-       if (atomic_read(&tw->tw_refcnt) != 1) {
-               printk(KERN_DEBUG "%s timewait_sock %p refcnt=%d\n",
-                      tw->tw_prot->name, tw, atomic_read(&tw->tw_refcnt));
-       }
-#endif
-       inet_twsk_put(tw);
-}
-
 /* 
  * * Main purpose of TIME-WAIT state is to close connection gracefully,
  *   when one of ends sits in LAST-ACK or CLOSING retransmitting FIN
@@ -168,7 +142,7 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
                if (!th->fin ||
                    TCP_SKB_CB(skb)->end_seq != tcptw->tw_rcv_nxt + 1) {
 kill_with_rst:
-                       tcp_tw_deschedule(tw);
+                       inet_twsk_deschedule(tw, &tcp_death_row);
                        inet_twsk_put(tw);
                        return TCP_TW_RST;
                }
@@ -187,11 +161,13 @@ kill_with_rst:
                 * do not undertsnad recycling in any case, it not
                 * a big problem in practice. --ANK */
                if (tw->tw_family == AF_INET &&
-                   sysctl_tcp_tw_recycle && tcptw->tw_ts_recent_stamp &&
+                   tcp_death_row.sysctl_tw_recycle && tcptw->tw_ts_recent_stamp &&
                    tcp_v4_tw_remember_stamp(tw))
-                       tcp_tw_schedule(tw, tw->tw_timeout);
+                       inet_twsk_schedule(tw, &tcp_death_row, tw->tw_timeout,
+                                          TCP_TIMEWAIT_LEN);
                else
-                       tcp_tw_schedule(tw, TCP_TIMEWAIT_LEN);
+                       inet_twsk_schedule(tw, &tcp_death_row, TCP_TIMEWAIT_LEN,
+                                          TCP_TIMEWAIT_LEN);
                return TCP_TW_ACK;
        }
 
@@ -224,12 +200,13 @@ kill_with_rst:
                         */
                        if (sysctl_tcp_rfc1337 == 0) {
 kill:
-                               tcp_tw_deschedule(tw);
+                               inet_twsk_deschedule(tw, &tcp_death_row);
                                inet_twsk_put(tw);
                                return TCP_TW_SUCCESS;
                        }
                }
-               tcp_tw_schedule(tw, TCP_TIMEWAIT_LEN);
+               inet_twsk_schedule(tw, &tcp_death_row, TCP_TIMEWAIT_LEN,
+                                  TCP_TIMEWAIT_LEN);
 
                if (tmp_opt.saw_tstamp) {
                        tcptw->tw_ts_recent       = tmp_opt.rcv_tsval;
@@ -279,7 +256,8 @@ kill:
                 * Do not reschedule in the last case.
                 */
                if (paws_reject || th->ack)
-                       tcp_tw_schedule(tw, TCP_TIMEWAIT_LEN);
+                       inet_twsk_schedule(tw, &tcp_death_row, TCP_TIMEWAIT_LEN,
+                                          TCP_TIMEWAIT_LEN);
 
                /* Send ACK. Note, we do not put the bucket,
                 * it will be released by caller.
@@ -290,40 +268,6 @@ kill:
        return TCP_TW_SUCCESS;
 }
 
-/* Enter the time wait state.  This is called with locally disabled BH.
- * Essentially we whip up a timewait bucket, copy the
- * relevant info into it from the SK, and mess with hash chains
- * and list linkage.
- */
-static void __tcp_tw_hashdance(struct sock *sk, struct inet_timewait_sock *tw)
-{
-       const struct inet_sock *inet = inet_sk(sk);
-       struct inet_ehash_bucket *ehead = &tcp_hashinfo.ehash[sk->sk_hashent];
-       struct inet_bind_hashbucket *bhead;
-       /* Step 1: Put TW into bind hash. Original socket stays there too.
-          Note, that any socket with inet->num != 0 MUST be bound in
-          binding cache, even if it is closed.
-        */
-       bhead = &tcp_hashinfo.bhash[inet_bhashfn(inet->num, tcp_hashinfo.bhash_size)];
-       spin_lock(&bhead->lock);
-       tw->tw_tb = inet->bind_hash;
-       BUG_TRAP(inet->bind_hash);
-       inet_twsk_add_bind_node(tw, &tw->tw_tb->owners);
-       spin_unlock(&bhead->lock);
-
-       write_lock(&ehead->lock);
-
-       /* Step 2: Remove SK from established hash. */
-       if (__sk_del_node_init(sk))
-               sock_prot_dec_use(sk->sk_prot);
-
-       /* Step 3: Hash TW into TIMEWAIT half of established hash table. */
-       inet_twsk_add_node(tw, &(ehead + tcp_hashinfo.ehash_size)->chain);
-       atomic_inc(&tw->tw_refcnt);
-
-       write_unlock(&ehead->lock);
-}
-
 /* 
  * Move a socket to time-wait or dead fin-wait-2 state.
  */ 
@@ -333,41 +277,23 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
        const struct tcp_sock *tp = tcp_sk(sk);
        int recycle_ok = 0;
 
-       if (sysctl_tcp_tw_recycle && tp->rx_opt.ts_recent_stamp)
+       if (tcp_death_row.sysctl_tw_recycle && tp->rx_opt.ts_recent_stamp)
                recycle_ok = tp->af_specific->remember_stamp(sk);
 
-       if (tcp_tw_count < sysctl_tcp_max_tw_buckets)
-               tw = kmem_cache_alloc(sk->sk_prot_creator->twsk_slab, SLAB_ATOMIC);
+       if (tcp_death_row.tw_count < tcp_death_row.sysctl_max_tw_buckets)
+               tw = inet_twsk_alloc(sk, state);
 
        if (tw != NULL) {
                struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);
-               const struct inet_sock *inet = inet_sk(sk);
-               const int rto = (tp->rto << 2) - (tp->rto >> 1);
-
-               /* Remember our protocol */
-               tw->tw_prot             = sk->sk_prot_creator;
-
-               /* Give us an identity. */
-               tw->tw_daddr            = inet->daddr;
-               tw->tw_rcv_saddr        = inet->rcv_saddr;
-               tw->tw_bound_dev_if     = sk->sk_bound_dev_if;
-               tw->tw_num              = inet->num;
-               tw->tw_state            = TCP_TIME_WAIT;
-               tw->tw_substate         = state;
-               tw->tw_sport            = inet->sport;
-               tw->tw_dport            = inet->dport;
-               tw->tw_family           = sk->sk_family;
-               tw->tw_reuse            = sk->sk_reuse;
-               tw->tw_rcv_wscale       = tp->rx_opt.rcv_wscale;
-               atomic_set(&tw->tw_refcnt, 1);
+               const struct inet_connection_sock *icsk = inet_csk(sk);
+               const int rto = (icsk->icsk_rto << 2) - (icsk->icsk_rto >> 1);
 
-               tw->tw_hashent          = sk->sk_hashent;
+               tw->tw_rcv_wscale       = tp->rx_opt.rcv_wscale;
                tcptw->tw_rcv_nxt       = tp->rcv_nxt;
                tcptw->tw_snd_nxt       = tp->snd_nxt;
                tcptw->tw_rcv_wnd       = tcp_receive_window(tp);
                tcptw->tw_ts_recent     = tp->rx_opt.ts_recent;
                tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
-               inet_twsk_dead_node_init(tw);
 
 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
                if (tw->tw_family == PF_INET6) {
@@ -377,11 +303,10 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
                        ipv6_addr_copy(&tcp6tw->tw_v6_daddr, &np->daddr);
                        ipv6_addr_copy(&tcp6tw->tw_v6_rcv_saddr, &np->rcv_saddr);
                        tw->tw_ipv6only = np->ipv6only;
-               } else
-                       tw->tw_ipv6only = 0;
+               }
 #endif
                /* Linkage updates. */
-               __tcp_tw_hashdance(sk, tw);
+               __inet_twsk_hashdance(tw, sk, &tcp_hashinfo);
 
                /* Get the TIME_WAIT timeout firing. */
                if (timeo < rto)
@@ -395,7 +320,8 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
                                timeo = TCP_TIMEWAIT_LEN;
                }
 
-               tcp_tw_schedule(tw, timeo);
+               inet_twsk_schedule(tw, &tcp_death_row, timeo,
+                                  TCP_TIMEWAIT_LEN);
                inet_twsk_put(tw);
        } else {
                /* Sorry, if we're out of memory, just CLOSE this
@@ -410,277 +336,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
        tcp_done(sk);
 }
 
-/* Kill off TIME_WAIT sockets once their lifetime has expired. */
-static int tcp_tw_death_row_slot;
-
-static void tcp_twkill(unsigned long);
-
-/* TIME_WAIT reaping mechanism. */
-#define TCP_TWKILL_SLOTS       8       /* Please keep this a power of 2. */
-#define TCP_TWKILL_PERIOD      (TCP_TIMEWAIT_LEN/TCP_TWKILL_SLOTS)
-
-#define TCP_TWKILL_QUOTA       100
-
-static struct hlist_head tcp_tw_death_row[TCP_TWKILL_SLOTS];
-static DEFINE_SPINLOCK(tw_death_lock);
-static struct timer_list tcp_tw_timer = TIMER_INITIALIZER(tcp_twkill, 0, 0);
-static void twkill_work(void *);
-static DECLARE_WORK(tcp_twkill_work, twkill_work, NULL);
-static u32 twkill_thread_slots;
-
-/* Returns non-zero if quota exceeded.  */
-static int tcp_do_twkill_work(int slot, unsigned int quota)
-{
-       struct inet_timewait_sock *tw;
-       struct hlist_node *node;
-       unsigned int killed;
-       int ret;
-
-       /* NOTE: compare this to previous version where lock
-        * was released after detaching chain. It was racy,
-        * because tw buckets are scheduled in not serialized context
-        * in 2.3 (with netfilter), and with softnet it is common, because
-        * soft irqs are not sequenced.
-        */
-       killed = 0;
-       ret = 0;
-rescan:
-       inet_twsk_for_each_inmate(tw, node, &tcp_tw_death_row[slot]) {
-               __inet_twsk_del_dead_node(tw);
-               spin_unlock(&tw_death_lock);
-               tcp_timewait_kill(tw);
-               inet_twsk_put(tw);
-               killed++;
-               spin_lock(&tw_death_lock);
-               if (killed > quota) {
-                       ret = 1;
-                       break;
-               }
-
-               /* While we dropped tw_death_lock, another cpu may have
-                * killed off the next TW bucket in the list, therefore
-                * do a fresh re-read of the hlist head node with the
-                * lock reacquired.  We still use the hlist traversal
-                * macro in order to get the prefetches.
-                */
-               goto rescan;
-       }
-
-       tcp_tw_count -= killed;
-       NET_ADD_STATS_BH(LINUX_MIB_TIMEWAITED, killed);
-
-       return ret;
-}
-
-static void tcp_twkill(unsigned long dummy)
-{
-       int need_timer, ret;
-
-       spin_lock(&tw_death_lock);
-
-       if (tcp_tw_count == 0)
-               goto out;
-
-       need_timer = 0;
-       ret = tcp_do_twkill_work(tcp_tw_death_row_slot, TCP_TWKILL_QUOTA);
-       if (ret) {
-               twkill_thread_slots |= (1 << tcp_tw_death_row_slot);
-               mb();
-               schedule_work(&tcp_twkill_work);
-               need_timer = 1;
-       } else {
-               /* We purged the entire slot, anything left?  */
-               if (tcp_tw_count)
-                       need_timer = 1;
-       }
-       tcp_tw_death_row_slot =
-               ((tcp_tw_death_row_slot + 1) & (TCP_TWKILL_SLOTS - 1));
-       if (need_timer)
-               mod_timer(&tcp_tw_timer, jiffies + TCP_TWKILL_PERIOD);
-out:
-       spin_unlock(&tw_death_lock);
-}
-
-extern void twkill_slots_invalid(void);
-
-static void twkill_work(void *dummy)
-{
-       int i;
-
-       if ((TCP_TWKILL_SLOTS - 1) > (sizeof(twkill_thread_slots) * 8))
-               twkill_slots_invalid();
-
-       while (twkill_thread_slots) {
-               spin_lock_bh(&tw_death_lock);
-               for (i = 0; i < TCP_TWKILL_SLOTS; i++) {
-                       if (!(twkill_thread_slots & (1 << i)))
-                               continue;
-
-                       while (tcp_do_twkill_work(i, TCP_TWKILL_QUOTA) != 0) {
-                               if (need_resched()) {
-                                       spin_unlock_bh(&tw_death_lock);
-                                       schedule();
-                                       spin_lock_bh(&tw_death_lock);
-                               }
-                       }
-
-                       twkill_thread_slots &= ~(1 << i);
-               }
-               spin_unlock_bh(&tw_death_lock);
-       }
-}
-
-/* These are always called from BH context.  See callers in
- * tcp_input.c to verify this.
- */
-
-/* This is for handling early-kills of TIME_WAIT sockets. */
-void tcp_tw_deschedule(struct inet_timewait_sock *tw)
-{
-       spin_lock(&tw_death_lock);
-       if (inet_twsk_del_dead_node(tw)) {
-               inet_twsk_put(tw);
-               if (--tcp_tw_count == 0)
-                       del_timer(&tcp_tw_timer);
-       }
-       spin_unlock(&tw_death_lock);
-       tcp_timewait_kill(tw);
-}
-
-/* Short-time timewait calendar */
-
-static int tcp_twcal_hand = -1;
-static int tcp_twcal_jiffie;
-static void tcp_twcal_tick(unsigned long);
-static struct timer_list tcp_twcal_timer =
-               TIMER_INITIALIZER(tcp_twcal_tick, 0, 0);
-static struct hlist_head tcp_twcal_row[TCP_TW_RECYCLE_SLOTS];
-
-static void tcp_tw_schedule(struct inet_timewait_sock *tw, const int timeo)
-{
-       struct hlist_head *list;
-       int slot;
-
-       /* timeout := RTO * 3.5
-        *
-        * 3.5 = 1+2+0.5 to wait for two retransmits.
-        *
-        * RATIONALE: if FIN arrived and we entered TIME-WAIT state,
-        * our ACK acking that FIN can be lost. If N subsequent retransmitted
-        * FINs (or previous seqments) are lost (probability of such event
-        * is p^(N+1), where p is probability to lose single packet and
-        * time to detect the loss is about RTO*(2^N - 1) with exponential
-        * backoff). Normal timewait length is calculated so, that we
-        * waited at least for one retransmitted FIN (maximal RTO is 120sec).
-        * [ BTW Linux. following BSD, violates this requirement waiting
-        *   only for 60sec, we should wait at least for 240 secs.
-        *   Well, 240 consumes too much of resources 8)
-        * ]
-        * This interval is not reduced to catch old duplicate and
-        * responces to our wandering segments living for two MSLs.
-        * However, if we use PAWS to detect
-        * old duplicates, we can reduce the interval to bounds required
-        * by RTO, rather than MSL. So, if peer understands PAWS, we
-        * kill tw bucket after 3.5*RTO (it is important that this number
-        * is greater than TS tick!) and detect old duplicates with help
-        * of PAWS.
-        */
-       slot = (timeo + (1<<TCP_TW_RECYCLE_TICK) - 1) >> TCP_TW_RECYCLE_TICK;
-
-       spin_lock(&tw_death_lock);
-
-       /* Unlink it, if it was scheduled */
-       if (inet_twsk_del_dead_node(tw))
-               tcp_tw_count--;
-       else
-               atomic_inc(&tw->tw_refcnt);
-
-       if (slot >= TCP_TW_RECYCLE_SLOTS) {
-               /* Schedule to slow timer */
-               if (timeo >= TCP_TIMEWAIT_LEN) {
-                       slot = TCP_TWKILL_SLOTS-1;
-               } else {
-                       slot = (timeo + TCP_TWKILL_PERIOD-1) / TCP_TWKILL_PERIOD;
-                       if (slot >= TCP_TWKILL_SLOTS)
-                               slot = TCP_TWKILL_SLOTS-1;
-               }
-               tw->tw_ttd = jiffies + timeo;
-               slot = (tcp_tw_death_row_slot + slot) & (TCP_TWKILL_SLOTS - 1);
-               list = &tcp_tw_death_row[slot];
-       } else {
-               tw->tw_ttd = jiffies + (slot << TCP_TW_RECYCLE_TICK);
-
-               if (tcp_twcal_hand < 0) {
-                       tcp_twcal_hand = 0;
-                       tcp_twcal_jiffie = jiffies;
-                       tcp_twcal_timer.expires = tcp_twcal_jiffie + (slot<<TCP_TW_RECYCLE_TICK);
-                       add_timer(&tcp_twcal_timer);
-               } else {
-                       if (time_after(tcp_twcal_timer.expires, jiffies + (slot<<TCP_TW_RECYCLE_TICK)))
-                               mod_timer(&tcp_twcal_timer, jiffies + (slot<<TCP_TW_RECYCLE_TICK));
-                       slot = (tcp_twcal_hand + slot)&(TCP_TW_RECYCLE_SLOTS-1);
-               }
-               list = &tcp_twcal_row[slot];
-       }
-
-       hlist_add_head(&tw->tw_death_node, list);
-
-       if (tcp_tw_count++ == 0)
-               mod_timer(&tcp_tw_timer, jiffies+TCP_TWKILL_PERIOD);
-       spin_unlock(&tw_death_lock);
-}
-
-void tcp_twcal_tick(unsigned long dummy)
-{
-       int n, slot;
-       unsigned long j;
-       unsigned long now = jiffies;
-       int killed = 0;
-       int adv = 0;
-
-       spin_lock(&tw_death_lock);
-       if (tcp_twcal_hand < 0)
-               goto out;
-
-       slot = tcp_twcal_hand;
-       j = tcp_twcal_jiffie;
-
-       for (n=0; n<TCP_TW_RECYCLE_SLOTS; n++) {
-               if (time_before_eq(j, now)) {
-                       struct hlist_node *node, *safe;
-                       struct inet_timewait_sock *tw;
-
-                       inet_twsk_for_each_inmate_safe(tw, node, safe,
-                                                      &tcp_twcal_row[slot]) {
-                               __inet_twsk_del_dead_node(tw);
-                               tcp_timewait_kill(tw);
-                               inet_twsk_put(tw);
-                               killed++;
-                       }
-               } else {
-                       if (!adv) {
-                               adv = 1;
-                               tcp_twcal_jiffie = j;
-                               tcp_twcal_hand = slot;
-                       }
-
-                       if (!hlist_empty(&tcp_twcal_row[slot])) {
-                               mod_timer(&tcp_twcal_timer, j);
-                               goto out;
-                       }
-               }
-               j += (1<<TCP_TW_RECYCLE_TICK);
-               slot = (slot+1)&(TCP_TW_RECYCLE_SLOTS-1);
-       }
-       tcp_twcal_hand = -1;
-
-out:
-       if ((tcp_tw_count -= killed) == 0)
-               del_timer(&tcp_tw_timer);
-       NET_ADD_STATS_BH(LINUX_MIB_TIMEWAITKILLED, killed);
-       spin_unlock(&tw_death_lock);
-}
-
 /* This is not only more efficient than what we used to do, it eliminates
  * a lot of code duplication between IPv4/IPv6 SYN recv processing. -DaveM
  *
@@ -689,77 +344,27 @@ out:
  */
 struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req, struct sk_buff *skb)
 {
-       /* allocate the newsk from the same slab of the master sock,
-        * if not, at sk_free time we'll try to free it from the wrong
-        * slabcache (i.e. is it TCPv4 or v6?), this is handled thru sk->sk_prot -acme */
-       struct sock *newsk = sk_alloc(PF_INET, GFP_ATOMIC, sk->sk_prot, 0);
+       struct sock *newsk = inet_csk_clone(sk, req, GFP_ATOMIC);
 
-       if(newsk != NULL) {
-               struct inet_request_sock *ireq = inet_rsk(req);
+       if (newsk != NULL) {
+               const struct inet_request_sock *ireq = inet_rsk(req);
                struct tcp_request_sock *treq = tcp_rsk(req);
-               struct inet_sock *newinet = inet_sk(newsk);
+               struct inet_connection_sock *newicsk = inet_csk(sk);
                struct tcp_sock *newtp;
-               struct sk_filter *filter;
-
-               memcpy(newsk, sk, sizeof(struct tcp_sock));
-               newsk->sk_state = TCP_SYN_RECV;
-
-               /* SANITY */
-               sk_node_init(&newsk->sk_node);
-               newinet->bind_hash = NULL;
-
-               /* Clone the TCP header template */
-               newinet->dport = ireq->rmt_port;
-
-               sock_lock_init(newsk);
-               bh_lock_sock(newsk);
-
-               rwlock_init(&newsk->sk_dst_lock);
-               newsk->sk_dst_cache = NULL;
-               atomic_set(&newsk->sk_rmem_alloc, 0);
-               skb_queue_head_init(&newsk->sk_receive_queue);
-               atomic_set(&newsk->sk_wmem_alloc, 0);
-               skb_queue_head_init(&newsk->sk_write_queue);
-               atomic_set(&newsk->sk_omem_alloc, 0);
-               newsk->sk_wmem_queued = 0;
-               newsk->sk_forward_alloc = 0;
-
-               sock_reset_flag(newsk, SOCK_DONE);
-               newsk->sk_userlocks = sk->sk_userlocks & ~SOCK_BINDPORT_LOCK;
-               newsk->sk_backlog.head = newsk->sk_backlog.tail = NULL;
-               newsk->sk_send_head = NULL;
-               rwlock_init(&newsk->sk_callback_lock);
-               skb_queue_head_init(&newsk->sk_error_queue);
-               newsk->sk_write_space = sk_stream_write_space;
-
-               if ((filter = newsk->sk_filter) != NULL)
-                       sk_filter_charge(newsk, filter);
-
-               if (unlikely(xfrm_sk_clone_policy(newsk))) {
-                       /* It is still raw copy of parent, so invalidate
-                        * destructor and make plain sk_free() */
-                       newsk->sk_destruct = NULL;
-                       sk_free(newsk);
-                       return NULL;
-               }
 
                /* Now setup tcp_sock */
                newtp = tcp_sk(newsk);
                newtp->pred_flags = 0;
                newtp->rcv_nxt = treq->rcv_isn + 1;
-               newtp->snd_nxt = treq->snt_isn + 1;
-               newtp->snd_una = treq->snt_isn + 1;
-               newtp->snd_sml = treq->snt_isn + 1;
+               newtp->snd_nxt = newtp->snd_una = newtp->snd_sml = treq->snt_isn + 1;
 
                tcp_prequeue_init(newtp);
 
                tcp_init_wl(newtp, treq->snt_isn, treq->rcv_isn);
 
-               newtp->retransmits = 0;
-               newtp->backoff = 0;
                newtp->srtt = 0;
                newtp->mdev = TCP_TIMEOUT_INIT;
-               newtp->rto = TCP_TIMEOUT_INIT;
+               newicsk->icsk_rto = TCP_TIMEOUT_INIT;
 
                newtp->packets_out = 0;
                newtp->left_out = 0;
@@ -779,9 +384,9 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
                newtp->frto_counter = 0;
                newtp->frto_highmark = 0;
 
-               newtp->ca_ops = &tcp_reno;
+               newicsk->icsk_ca_ops = &tcp_reno;
 
-               tcp_set_ca_state(newtp, TCP_CA_Open);
+               tcp_set_ca_state(newsk, TCP_CA_Open);
                tcp_init_xmit_timers(newsk);
                skb_queue_head_init(&newtp->out_of_order_queue);
                newtp->rcv_wup = treq->rcv_isn + 1;
@@ -794,38 +399,12 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
                newtp->rx_opt.dsack = 0;
                newtp->rx_opt.eff_sacks = 0;
 
-               newtp->probes_out = 0;
                newtp->rx_opt.num_sacks = 0;
                newtp->urg_data = 0;
-               /* Deinitialize accept_queue to trap illegal accesses. */
-               memset(&newtp->accept_queue, 0, sizeof(newtp->accept_queue));
-
-               /* Back to base struct sock members. */
-               newsk->sk_err = 0;
-               newsk->sk_priority = 0;
-               atomic_set(&newsk->sk_refcnt, 2);
-
-               /*
-                * Increment the counter in the same struct proto as the master
-                * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that
-                * is the same as sk->sk_prot->socks, as this field was copied
-                * with memcpy), same rationale as the first comment in this
-                * function.
-                *
-                * This _changes_ the previous behaviour, where
-                * tcp_create_openreq_child always was incrementing the
-                * equivalent to tcp_prot->socks (inet_sock_nr), so this have
-                * to be taken into account in all callers. -acme
-                */
-               sk_refcnt_debug_inc(newsk);
-
-               atomic_inc(&tcp_sockets_allocated);
 
                if (sock_flag(newsk, SOCK_KEEPOPEN))
-                       tcp_reset_keepalive_timer(newsk,
-                                                 keepalive_time_when(newtp));
-               newsk->sk_socket = NULL;
-               newsk->sk_sleep = NULL;
+                       inet_csk_reset_keepalive_timer(newsk,
+                                                      keepalive_time_when(newtp));
 
                newtp->rx_opt.tstamp_ok = ireq->tstamp_ok;
                if((newtp->rx_opt.sack_ok = ireq->sack_ok) != 0) {
@@ -855,7 +434,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
                        newtp->tcp_header_len = sizeof(struct tcphdr);
                }
                if (skb->len >= TCP_MIN_RCVMSS+newtp->tcp_header_len)
-                       newtp->ack.last_seg_size = skb->len-newtp->tcp_header_len;
+                       newicsk->icsk_ack.last_seg_size = skb->len - newtp->tcp_header_len;
                newtp->rx_opt.mss_clamp = req->mss;
                TCP_ECN_openreq_child(newtp, req);
                if (newtp->ecn_flags&TCP_ECN_OK)
@@ -951,9 +530,10 @@ 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.
@@ -1020,7 +600,8 @@ 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;
                }
@@ -1035,10 +616,10 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
                if (child == NULL)
                        goto listen_overflow;
 
-               tcp_synq_unlink(tp, req, prev);
-               tcp_synq_removed(sk, req);
+               inet_csk_reqsk_queue_unlink(sk, req, prev);
+               inet_csk_reqsk_queue_removed(sk, req);
 
-               tcp_acceptq_queue(sk, req, child);
+               inet_csk_reqsk_queue_add(sk, req, child);
                return child;
 
        listen_overflow:
@@ -1052,7 +633,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
                if (!(flg & TCP_FLAG_RST))
                        req->rsk_ops->send_reset(skb);
 
-               tcp_synq_drop(sk, req, prev);
+               inet_csk_reqsk_queue_drop(sk, req, prev);
                return NULL;
 }
 
@@ -1091,4 +672,3 @@ EXPORT_SYMBOL(tcp_check_req);
 EXPORT_SYMBOL(tcp_child_process);
 EXPORT_SYMBOL(tcp_create_openreq_child);
 EXPORT_SYMBOL(tcp_timewait_state_process);
-EXPORT_SYMBOL(tcp_tw_deschedule);