diff options
Diffstat (limited to 'net/ipv4')
-rw-r--r-- | net/ipv4/inet_hashtables.c | 15 | ||||
-rw-r--r-- | net/ipv4/inet_timewait_sock.c | 5 | ||||
-rw-r--r-- | net/ipv4/syncookies.c | 2 | ||||
-rw-r--r-- | net/ipv4/tcp.c | 90 | ||||
-rw-r--r-- | net/ipv4/tcp_diag.c | 21 | ||||
-rw-r--r-- | net/ipv4/tcp_input.c | 266 | ||||
-rw-r--r-- | net/ipv4/tcp_ipv4.c | 158 | ||||
-rw-r--r-- | net/ipv4/tcp_minisocks.c | 28 | ||||
-rw-r--r-- | net/ipv4/tcp_output.c | 86 | ||||
-rw-r--r-- | net/ipv4/tcp_timer.c | 165 |
10 files changed, 440 insertions, 396 deletions
diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index d94e962958a..e8d29fe736d 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -19,6 +19,7 @@ #include <linux/slab.h> #include <linux/wait.h> +#include <net/inet_connection_sock.h> #include <net/inet_hashtables.h> /* @@ -56,10 +57,9 @@ void inet_bind_bucket_destroy(kmem_cache_t *cachep, struct inet_bind_bucket *tb) void inet_bind_hash(struct sock *sk, struct inet_bind_bucket *tb, const unsigned short snum) { - struct inet_sock *inet = inet_sk(sk); - inet->num = snum; + inet_sk(sk)->num = snum; sk_add_bind_node(sk, &tb->owners); - inet->bind_hash = tb; + inet_csk(sk)->icsk_bind_hash = tb; } EXPORT_SYMBOL(inet_bind_hash); @@ -69,16 +69,15 @@ EXPORT_SYMBOL(inet_bind_hash); */ static void __inet_put_port(struct inet_hashinfo *hashinfo, struct sock *sk) { - struct inet_sock *inet = inet_sk(sk); - const int bhash = inet_bhashfn(inet->num, hashinfo->bhash_size); + const int bhash = inet_bhashfn(inet_sk(sk)->num, hashinfo->bhash_size); struct inet_bind_hashbucket *head = &hashinfo->bhash[bhash]; struct inet_bind_bucket *tb; spin_lock(&head->lock); - tb = inet->bind_hash; + tb = inet_csk(sk)->icsk_bind_hash; __sk_del_bind_node(sk); - inet->bind_hash = NULL; - inet->num = 0; + inet_csk(sk)->icsk_bind_hash = NULL; + inet_sk(sk)->num = 0; inet_bind_bucket_destroy(hashinfo->bind_bucket_cachep, tb); spin_unlock(&head->lock); } diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index ceb577c7423..5cba59b869f 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -56,6 +56,7 @@ void __inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, struct inet_hashinfo *hashinfo) { const struct inet_sock *inet = inet_sk(sk); + const struct inet_connection_sock *icsk = inet_csk(sk); struct inet_ehash_bucket *ehead = &hashinfo->ehash[sk->sk_hashent]; struct inet_bind_hashbucket *bhead; /* Step 1: Put TW into bind hash. Original socket stays there too. @@ -64,8 +65,8 @@ void __inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, */ bhead = &hashinfo->bhash[inet_bhashfn(inet->num, hashinfo->bhash_size)]; spin_lock(&bhead->lock); - tw->tw_tb = inet->bind_hash; - BUG_TRAP(inet->bind_hash); + tw->tw_tb = icsk->icsk_bind_hash; + BUG_TRAP(icsk->icsk_bind_hash); inet_twsk_add_bind_node(tw, &tw->tw_tb->owners); spin_unlock(&bhead->lock); diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c index 72d01444218..8692cb9d4bd 100644 --- a/net/ipv4/syncookies.c +++ b/net/ipv4/syncookies.c @@ -180,7 +180,7 @@ static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb, child = tp->af_specific->syn_recv_sock(sk, skb, req, dst); if (child) - tcp_acceptq_queue(sk, req, child); + inet_csk_reqsk_queue_add(sk, req, child); else reqsk_free(req); diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index f1a708bf7a9..8177b86570d 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -313,7 +313,7 @@ EXPORT_SYMBOL(tcp_enter_memory_pressure); static __inline__ unsigned int tcp_listen_poll(struct sock *sk, poll_table *wait) { - return !reqsk_queue_empty(&tcp_sk(sk)->accept_queue) ? (POLLIN | POLLRDNORM) : 0; + return !reqsk_queue_empty(&inet_csk(sk)->icsk_accept_queue) ? (POLLIN | POLLRDNORM) : 0; } /* @@ -458,15 +458,15 @@ int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg) int tcp_listen_start(struct sock *sk) { struct inet_sock *inet = inet_sk(sk); - struct tcp_sock *tp = tcp_sk(sk); - int rc = reqsk_queue_alloc(&tp->accept_queue, TCP_SYNQ_HSIZE); + struct inet_connection_sock *icsk = inet_csk(sk); + int rc = reqsk_queue_alloc(&icsk->icsk_accept_queue, TCP_SYNQ_HSIZE); if (rc != 0) return rc; sk->sk_max_ack_backlog = 0; sk->sk_ack_backlog = 0; - tcp_delack_init(tp); + inet_csk_delack_init(sk); /* There is race window here: we announce ourselves listening, * but this transition is still not validated by get_port(). @@ -484,7 +484,7 @@ int tcp_listen_start(struct sock *sk) } sk->sk_state = TCP_CLOSE; - __reqsk_queue_destroy(&tp->accept_queue); + __reqsk_queue_destroy(&icsk->icsk_accept_queue); return -EADDRINUSE; } @@ -495,14 +495,14 @@ int tcp_listen_start(struct sock *sk) static void tcp_listen_stop (struct sock *sk) { - struct tcp_sock *tp = tcp_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); struct request_sock *acc_req; struct request_sock *req; - tcp_delete_keepalive_timer(sk); + inet_csk_delete_keepalive_timer(sk); /* make all the listen_opt local to us */ - acc_req = reqsk_queue_yank_acceptq(&tp->accept_queue); + acc_req = reqsk_queue_yank_acceptq(&icsk->icsk_accept_queue); /* Following specs, it would be better either to send FIN * (and enter FIN-WAIT-1, it is normal close) @@ -512,7 +512,7 @@ static void tcp_listen_stop (struct sock *sk) * To be honest, we are not able to make either * of the variants now. --ANK */ - reqsk_queue_destroy(&tp->accept_queue); + reqsk_queue_destroy(&icsk->icsk_accept_queue); while ((req = acc_req) != NULL) { struct sock *child = req->sk; @@ -1039,20 +1039,21 @@ static void cleanup_rbuf(struct sock *sk, int copied) BUG_TRAP(!skb || before(tp->copied_seq, TCP_SKB_CB(skb)->end_seq)); #endif - if (tcp_ack_scheduled(tp)) { + if (inet_csk_ack_scheduled(sk)) { + const struct inet_connection_sock *icsk = inet_csk(sk); /* Delayed ACKs frequently hit locked sockets during bulk * receive. */ - if (tp->ack.blocked || + if (icsk->icsk_ack.blocked || /* Once-per-two-segments ACK was not sent by tcp_input.c */ - tp->rcv_nxt - tp->rcv_wup > tp->ack.rcv_mss || + tp->rcv_nxt - tp->rcv_wup > icsk->icsk_ack.rcv_mss || /* * If this read emptied read buffer, we send ACK, if * connection is not bidirectional, user drained * receive buffer and there was a small segment * in queue. */ - (copied > 0 && (tp->ack.pending & TCP_ACK_PUSHED) && - !tp->ack.pingpong && !atomic_read(&sk->sk_rmem_alloc))) + (copied > 0 && (icsk->icsk_ack.pending & ICSK_ACK_PUSHED) && + !icsk->icsk_ack.pingpong && !atomic_read(&sk->sk_rmem_alloc))) time_to_ack = 1; } @@ -1569,7 +1570,7 @@ void tcp_destroy_sock(struct sock *sk) BUG_TRAP(sk_unhashed(sk)); /* If it has not 0 inet_sk(sk)->num, it must be bound */ - BUG_TRAP(!inet_sk(sk)->num || inet_sk(sk)->bind_hash); + BUG_TRAP(!inet_sk(sk)->num || inet_csk(sk)->icsk_bind_hash); sk->sk_prot->destroy(sk); @@ -1698,10 +1699,10 @@ adjudge_to_death: tcp_send_active_reset(sk, GFP_ATOMIC); NET_INC_STATS_BH(LINUX_MIB_TCPABORTONLINGER); } else { - int tmo = tcp_fin_time(tp); + const int tmo = tcp_fin_time(sk); if (tmo > TCP_TIMEWAIT_LEN) { - tcp_reset_keepalive_timer(sk, tcp_fin_time(tp)); + inet_csk_reset_keepalive_timer(sk, tcp_fin_time(sk)); } else { atomic_inc(&tcp_orphan_count); tcp_time_wait(sk, TCP_FIN_WAIT2, tmo); @@ -1746,6 +1747,7 @@ static inline int tcp_need_reset(int state) int tcp_disconnect(struct sock *sk, int flags) { struct inet_sock *inet = inet_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); int err = 0; int old_state = sk->sk_state; @@ -1782,7 +1784,7 @@ int tcp_disconnect(struct sock *sk, int flags) tp->srtt = 0; if ((tp->write_seq += tp->max_window + 2) == 0) tp->write_seq = 1; - tp->backoff = 0; + icsk->icsk_backoff = 0; tp->snd_cwnd = 2; tp->probes_out = 0; tp->packets_out = 0; @@ -1790,13 +1792,13 @@ int tcp_disconnect(struct sock *sk, int flags) tp->snd_cwnd_cnt = 0; tcp_set_ca_state(tp, TCP_CA_Open); tcp_clear_retrans(tp); - tcp_delack_init(tp); + inet_csk_delack_init(sk); sk->sk_send_head = NULL; tp->rx_opt.saw_tstamp = 0; tcp_sack_reset(&tp->rx_opt); __sk_dst_reset(sk); - BUG_TRAP(!inet->num || inet->bind_hash); + BUG_TRAP(!inet->num || icsk->icsk_bind_hash); sk->sk_error_report(sk); return err; @@ -1808,7 +1810,7 @@ int tcp_disconnect(struct sock *sk, int flags) */ static int wait_for_connect(struct sock *sk, long timeo) { - struct tcp_sock *tp = tcp_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); DEFINE_WAIT(wait); int err; @@ -1830,11 +1832,11 @@ static int wait_for_connect(struct sock *sk, long timeo) prepare_to_wait_exclusive(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE); release_sock(sk); - if (reqsk_queue_empty(&tp->accept_queue)) + if (reqsk_queue_empty(&icsk->icsk_accept_queue)) timeo = schedule_timeout(timeo); lock_sock(sk); err = 0; - if (!reqsk_queue_empty(&tp->accept_queue)) + if (!reqsk_queue_empty(&icsk->icsk_accept_queue)) break; err = -EINVAL; if (sk->sk_state != TCP_LISTEN) @@ -1854,9 +1856,9 @@ static int wait_for_connect(struct sock *sk, long timeo) * This will accept the next outstanding connection. */ -struct sock *tcp_accept(struct sock *sk, int flags, int *err) +struct sock *inet_csk_accept(struct sock *sk, int flags, int *err) { - struct tcp_sock *tp = tcp_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); struct sock *newsk; int error; @@ -1870,7 +1872,7 @@ struct sock *tcp_accept(struct sock *sk, int flags, int *err) goto out_err; /* Find already established connection */ - if (reqsk_queue_empty(&tp->accept_queue)) { + if (reqsk_queue_empty(&icsk->icsk_accept_queue)) { long timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK); /* If this is a non blocking socket don't sleep */ @@ -1883,7 +1885,7 @@ struct sock *tcp_accept(struct sock *sk, int flags, int *err) goto out_err; } - newsk = reqsk_queue_get_child(&tp->accept_queue, sk); + newsk = reqsk_queue_get_child(&icsk->icsk_accept_queue, sk); BUG_TRAP(newsk->sk_state != TCP_SYN_RECV); out: release_sock(sk); @@ -1901,6 +1903,7 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval, int optlen) { struct tcp_sock *tp = tcp_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); int val; int err = 0; @@ -1999,7 +2002,7 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval, elapsed = tp->keepalive_time - elapsed; else elapsed = 0; - tcp_reset_keepalive_timer(sk, elapsed); + inet_csk_reset_keepalive_timer(sk, elapsed); } } break; @@ -2019,7 +2022,7 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval, if (val < 1 || val > MAX_TCP_SYNCNT) err = -EINVAL; else - tp->syn_retries = val; + icsk->icsk_syn_retries = val; break; case TCP_LINGER2: @@ -2058,16 +2061,16 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval, case TCP_QUICKACK: if (!val) { - tp->ack.pingpong = 1; + icsk->icsk_ack.pingpong = 1; } else { - tp->ack.pingpong = 0; + icsk->icsk_ack.pingpong = 0; if ((1 << sk->sk_state) & (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) && - tcp_ack_scheduled(tp)) { - tp->ack.pending |= TCP_ACK_PUSHED; + inet_csk_ack_scheduled(sk)) { + icsk->icsk_ack.pending |= ICSK_ACK_PUSHED; cleanup_rbuf(sk, 1); if (!(val & 1)) - tp->ack.pingpong = 1; + icsk->icsk_ack.pingpong = 1; } } break; @@ -2084,15 +2087,16 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval, void tcp_get_info(struct sock *sk, struct tcp_info *info) { struct tcp_sock *tp = tcp_sk(sk); + const struct inet_connection_sock *icsk = inet_csk(sk); u32 now = tcp_time_stamp; memset(info, 0, sizeof(*info)); info->tcpi_state = sk->sk_state; info->tcpi_ca_state = tp->ca_state; - info->tcpi_retransmits = tp->retransmits; + info->tcpi_retransmits = icsk->icsk_retransmits; info->tcpi_probes = tp->probes_out; - info->tcpi_backoff = tp->backoff; + info->tcpi_backoff = icsk->icsk_backoff; if (tp->rx_opt.tstamp_ok) info->tcpi_options |= TCPI_OPT_TIMESTAMPS; @@ -2107,10 +2111,10 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) if (tp->ecn_flags&TCP_ECN_OK) info->tcpi_options |= TCPI_OPT_ECN; - info->tcpi_rto = jiffies_to_usecs(tp->rto); - info->tcpi_ato = jiffies_to_usecs(tp->ack.ato); + info->tcpi_rto = jiffies_to_usecs(icsk->icsk_rto); + info->tcpi_ato = jiffies_to_usecs(icsk->icsk_ack.ato); info->tcpi_snd_mss = tp->mss_cache; - info->tcpi_rcv_mss = tp->ack.rcv_mss; + info->tcpi_rcv_mss = icsk->icsk_ack.rcv_mss; info->tcpi_unacked = tp->packets_out; info->tcpi_sacked = tp->sacked_out; @@ -2119,7 +2123,7 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) info->tcpi_fackets = tp->fackets_out; info->tcpi_last_data_sent = jiffies_to_msecs(now - tp->lsndtime); - info->tcpi_last_data_recv = jiffies_to_msecs(now - tp->ack.lrcvtime); + info->tcpi_last_data_recv = jiffies_to_msecs(now - icsk->icsk_ack.lrcvtime); info->tcpi_last_ack_recv = jiffies_to_msecs(now - tp->rcv_tstamp); info->tcpi_pmtu = tp->pmtu_cookie; @@ -2179,7 +2183,7 @@ 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 = tp->syn_retries ? : sysctl_tcp_syn_retries; + val = inet_csk(sk)->icsk_syn_retries ? : sysctl_tcp_syn_retries; break; case TCP_LINGER2: val = tp->linger2; @@ -2209,7 +2213,7 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval, return 0; } case TCP_QUICKACK: - val = !tp->ack.pingpong; + val = !inet_csk(sk)->icsk_ack.pingpong; break; case TCP_CONGESTION: @@ -2340,7 +2344,7 @@ void __init tcp_init(void) tcp_register_congestion_control(&tcp_reno); } -EXPORT_SYMBOL(tcp_accept); +EXPORT_SYMBOL(inet_csk_accept); EXPORT_SYMBOL(tcp_close); EXPORT_SYMBOL(tcp_destroy_sock); EXPORT_SYMBOL(tcp_disconnect); diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c index 60c6a797cc5..5f4c74f45e8 100644 --- a/net/ipv4/tcp_diag.c +++ b/net/ipv4/tcp_diag.c @@ -48,8 +48,9 @@ static struct sock *tcpnl; static int tcpdiag_fill(struct sk_buff *skb, struct sock *sk, int ext, u32 pid, u32 seq, u16 nlmsg_flags) { - struct inet_sock *inet = inet_sk(sk); + const struct inet_sock *inet = inet_sk(sk); struct tcp_sock *tp = tcp_sk(sk); + const struct inet_connection_sock *icsk = inet_csk(sk); struct tcpdiagmsg *r; struct nlmsghdr *nlh; struct tcp_info *info = NULL; @@ -129,14 +130,14 @@ static int tcpdiag_fill(struct sk_buff *skb, struct sock *sk, #define EXPIRES_IN_MS(tmo) ((tmo-jiffies)*1000+HZ-1)/HZ - if (tp->pending == TCP_TIME_RETRANS) { + if (icsk->icsk_pending == ICSK_TIME_RETRANS) { r->tcpdiag_timer = 1; - r->tcpdiag_retrans = tp->retransmits; - r->tcpdiag_expires = EXPIRES_IN_MS(tp->timeout); - } else if (tp->pending == TCP_TIME_PROBE0) { + r->tcpdiag_retrans = icsk->icsk_retransmits; + r->tcpdiag_expires = EXPIRES_IN_MS(icsk->icsk_timeout); + } else if (icsk->icsk_pending == ICSK_TIME_PROBE0) { r->tcpdiag_timer = 4; r->tcpdiag_retrans = tp->probes_out; - r->tcpdiag_expires = EXPIRES_IN_MS(tp->timeout); + r->tcpdiag_expires = EXPIRES_IN_MS(icsk->icsk_timeout); } else if (timer_pending(&sk->sk_timer)) { r->tcpdiag_timer = 2; r->tcpdiag_retrans = tp->probes_out; @@ -497,7 +498,7 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk, { struct tcpdiag_entry entry; struct tcpdiagreq *r = NLMSG_DATA(cb->nlh); - struct tcp_sock *tp = tcp_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); struct listen_sock *lopt; struct rtattr *bc = NULL; struct inet_sock *inet = inet_sk(sk); @@ -513,9 +514,9 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk, entry.family = sk->sk_family; - read_lock_bh(&tp->accept_queue.syn_wait_lock); + read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock); - lopt = tp->accept_queue.listen_opt; + lopt = icsk->icsk_accept_queue.listen_opt; if (!lopt || !lopt->qlen) goto out; @@ -572,7 +573,7 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk, } out: - read_unlock_bh(&tp->accept_queue.syn_wait_lock); + read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock); return err; } diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index ffa24025cd0..8a8c5c2d90c 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -114,20 +114,21 @@ int sysctl_tcp_moderate_rcvbuf = 1; /* Adapt the MSS value used to make delayed ack decision to the * real world. */ -static inline void tcp_measure_rcv_mss(struct tcp_sock *tp, - struct sk_buff *skb) +static inline void tcp_measure_rcv_mss(struct sock *sk, + const struct sk_buff *skb) { - unsigned int len, lss; + struct inet_connection_sock *icsk = inet_csk(sk); + const unsigned int lss = icsk->icsk_ack.last_seg_size; + unsigned int len; - lss = tp->ack.last_seg_size; - tp->ack.last_seg_size = 0; + icsk->icsk_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; + if (len >= icsk->icsk_ack.rcv_mss) { + icsk->icsk_ack.rcv_mss = len; } else { /* Otherwise, we make more careful check taking into account, * that SACKs block is variable. @@ -147,41 +148,44 @@ static inline void tcp_measure_rcv_mss(struct tcp_sock *tp, * 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; + len -= tcp_sk(sk)->tcp_header_len; + icsk->icsk_ack.last_seg_size = len; if (len == lss) { - tp->ack.rcv_mss = len; + icsk->icsk_ack.rcv_mss = len; return; } } - tp->ack.pending |= TCP_ACK_PUSHED; + icsk->icsk_ack.pending |= ICSK_ACK_PUSHED; } } -static void tcp_incr_quickack(struct tcp_sock *tp) +static void tcp_incr_quickack(struct sock *sk) { - unsigned quickacks = tp->rcv_wnd/(2*tp->ack.rcv_mss); + struct inet_connection_sock *icsk = inet_csk(sk); + unsigned quickacks = tcp_sk(sk)->rcv_wnd / (2 * icsk->icsk_ack.rcv_mss); if (quickacks==0) quickacks=2; - if (quickacks > tp->ack.quick) - tp->ack.quick = min(quickacks, TCP_MAX_QUICKACKS); + if (quickacks > icsk->icsk_ack.quick) + icsk->icsk_ack.quick = min(quickacks, TCP_MAX_QUICKACKS); } -void tcp_enter_quickack_mode(struct tcp_sock *tp) +void tcp_enter_quickack_mode(struct sock *sk) { - tcp_incr_quickack(tp); - tp->ack.pingpong = 0; - tp->ack.ato = TCP_ATO_MIN; + struct inet_connection_sock *icsk = inet_csk(sk); + tcp_incr_quickack(sk); + icsk->icsk_ack.pingpong = 0; + icsk->icsk_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_sock *tp) +static inline int tcp_in_quickack_mode(const struct sock *sk) { - return (tp->ack.quick && !tp->ack.pingpong); + const struct inet_connection_sock *icsk = inet_csk(sk); + return icsk->icsk_ack.quick && !icsk->icsk_ack.pingpong; } /* Buffer size and advertised window tuning. @@ -224,8 +228,8 @@ static void tcp_fixup_sndbuf(struct sock *sk) */ /* Slow part of check#2. */ -static int __tcp_grow_window(struct sock *sk, struct tcp_sock *tp, - struct sk_buff *skb) +static int __tcp_grow_window(const struct sock *sk, struct tcp_sock *tp, + const struct sk_buff *skb) { /* Optimize this! */ int truesize = tcp_win_from_space(skb->truesize)/2; @@ -233,7 +237,7 @@ static int __tcp_grow_window(struct sock *sk, struct tcp_sock *tp, while (tp->rcv_ssthresh <= window) { if (truesize <= skb->len) - return 2*tp->ack.rcv_mss; + return 2 * inet_csk(sk)->icsk_ack.rcv_mss; truesize >>= 1; window >>= 1; @@ -260,7 +264,7 @@ static inline void tcp_grow_window(struct sock *sk, struct tcp_sock *tp, if (incr) { tp->rcv_ssthresh = min(tp->rcv_ssthresh + incr, tp->window_clamp); - tp->ack.quick |= 1; + inet_csk(sk)->icsk_ack.quick |= 1; } } } @@ -325,7 +329,7 @@ static void tcp_clamp_window(struct sock *sk, struct tcp_sock *tp) unsigned int app_win = tp->rcv_nxt - tp->copied_seq; int ofo_win = 0; - tp->ack.quick = 0; + inet_csk(sk)->icsk_ack.quick = 0; skb_queue_walk(&tp->out_of_order_queue, skb) { ofo_win += skb->len; @@ -346,8 +350,8 @@ static void tcp_clamp_window(struct sock *sk, struct tcp_sock *tp) app_win += ofo_win; if (atomic_read(&sk->sk_rmem_alloc) >= 2 * sk->sk_rcvbuf) app_win >>= 1; - if (app_win > tp->ack.rcv_mss) - app_win -= tp->ack.rcv_mss; + if (app_win > inet_csk(sk)->icsk_ack.rcv_mss) + app_win -= inet_csk(sk)->icsk_ack.rcv_mss; app_win = max(app_win, 2U*tp->advmss); if (!ofo_win) @@ -415,11 +419,12 @@ new_measure: tp->rcv_rtt_est.time = tcp_time_stamp; } -static inline void tcp_rcv_rtt_measure_ts(struct tcp_sock *tp, struct sk_buff *skb) +static inline void tcp_rcv_rtt_measure_ts(struct sock *sk, const struct sk_buff *skb) { + struct tcp_sock *tp = tcp_sk(sk); if (tp->rx_opt.rcv_tsecr && (TCP_SKB_CB(skb)->end_seq - - TCP_SKB_CB(skb)->seq >= tp->ack.rcv_mss)) + TCP_SKB_CB(skb)->seq >= inet_csk(sk)->icsk_ack.rcv_mss)) tcp_rcv_rtt_update(tp, tcp_time_stamp - tp->rx_opt.rcv_tsecr, 0); } @@ -492,41 +497,42 @@ new_measure: */ static void tcp_event_data_recv(struct sock *sk, struct tcp_sock *tp, struct sk_buff *skb) { + struct inet_connection_sock *icsk = inet_csk(sk); u32 now; - tcp_schedule_ack(tp); + inet_csk_schedule_ack(sk); - tcp_measure_rcv_mss(tp, skb); + tcp_measure_rcv_mss(sk, skb); tcp_rcv_rtt_measure(tp); now = tcp_time_stamp; - if (!tp->ack.ato) { + if (!icsk->icsk_ack.ato) { /* The _first_ data packet received, initialize * delayed ACK engine. */ - tcp_incr_quickack(tp); - tp->ack.ato = TCP_ATO_MIN; + tcp_incr_quickack(sk); + icsk->icsk_ack.ato = TCP_ATO_MIN; } else { - int m = now - tp->ack.lrcvtime; + int m = now - icsk->icsk_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) { + icsk->icsk_ack.ato = (icsk->icsk_ack.ato >> 1) + TCP_ATO_MIN / 2; + } else if (m < icsk->icsk_ack.ato) { + icsk->icsk_ack.ato = (icsk->icsk_ack.ato >> 1) + m; + if (icsk->icsk_ack.ato > icsk->icsk_rto) + icsk->icsk_ack.ato = icsk->icsk_rto; + } else if (m > icsk->icsk_rto) { /* Too long gap. Apparently sender falled to * restart window, so that we send ACKs quickly. */ - tcp_incr_quickack(tp); + tcp_incr_quickack(sk); sk_stream_mem_reclaim(sk); } } - tp->ack.lrcvtime = now; + icsk->icsk_ack.lrcvtime = now; TCP_ECN_check_ce(tp, skb); @@ -611,8 +617,9 @@ static void tcp_rtt_estimator(struct tcp_sock *tp, __u32 mrtt, u32 *usrtt) /* 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_sock *tp) +static inline void tcp_set_rto(struct sock *sk) { + const struct tcp_sock *tp = tcp_sk(sk); /* Old crap is replaced with new one. 8) * * More seriously: @@ -623,7 +630,7 @@ static inline void tcp_set_rto(struct tcp_sock *tp) * is invisible. Actually, Linux-2.4 also generates erratic * ACKs in some curcumstances. */ - tp->rto = (tp->srtt >> 3) + tp->rttvar; + inet_csk(sk)->icsk_rto = (tp->srtt >> 3) + tp->rttvar; /* 2. Fixups made earlier cannot be right. * If we do not estimate RTO correctly without them, @@ -635,10 +642,10 @@ static inline void tcp_set_rto(struct tcp_sock *tp) /* NOTE: clamping at TCP_RTO_MIN is not required, current algo * guarantees that rto is higher. */ -static inline void tcp_bound_rto(struct tcp_sock *tp) +static inline void tcp_bound_rto(struct sock *sk) { - if (tp->rto > TCP_RTO_MAX) - tp->rto = TCP_RTO_MAX; + if (inet_csk(sk)->icsk_rto > TCP_RTO_MAX) + inet_csk(sk)->icsk_rto = TCP_RTO_MAX; } /* Save metrics learned by this TCP session. @@ -658,7 +665,7 @@ void tcp_update_metrics(struct sock *sk) if (dst && (dst->flags&DST_HOST)) { int m; - if (tp->backoff || !tp->srtt) { + if (inet_csk(sk)->icsk_backoff || !tp->srtt) { /* This session failed to estimate rtt. Why? * Probably, no packets returned in time. * Reset our results. @@ -801,9 +808,9 @@ static void tcp_init_metrics(struct sock *sk) tp->mdev = dst_metric(dst, RTAX_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->rx_opt.saw_tstamp) + tcp_set_rto(sk); + tcp_bound_rto(sk); + if (inet_csk(sk)->icsk_rto < TCP_TIMEOUT_INIT && !tp->rx_opt.saw_tstamp) goto reset; tp->snd_cwnd = tcp_init_cwnd(tp, dst); tp->snd_cwnd_stamp = tcp_time_stamp; @@ -817,7 +824,7 @@ reset: if (!tp->rx_opt.saw_tstamp && tp->srtt) { tp->srtt = 0; tp->mdev = tp->mdev_max = tp->rttvar = TCP_TIMEOUT_INIT; - tp->rto = TCP_TIMEOUT_INIT; + inet_csk(sk)->icsk_rto = TCP_TIMEOUT_INIT; } } @@ -1118,7 +1125,7 @@ void tcp_enter_frto(struct sock *sk) if (tp->ca_state <= TCP_CA_Disorder || tp->snd_una == tp->high_seq || - (tp->ca_state == TCP_CA_Loss && !tp->retransmits)) { + (tp->ca_state == TCP_CA_Loss && !inet_csk(sk)->icsk_retransmits)) { tp->prior_ssthresh = tcp_current_ssthresh(tp); tp->snd_ssthresh = tp->ca_ops->ssthresh(tp); tcp_ca_event(tp, CA_EVENT_FRTO); @@ -1214,7 +1221,7 @@ void tcp_enter_loss(struct sock *sk, int how) /* 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->ca_state == TCP_CA_Loss && !inet_csk(sk)->icsk_retransmits)) { tp->prior_ssthresh = tcp_current_ssthresh(tp); tp->snd_ssthresh = tp->ca_ops->ssthresh(tp); tcp_ca_event(tp, CA_EVENT_LOSS); @@ -1253,7 +1260,7 @@ void tcp_enter_loss(struct sock *sk, int how) TCP_ECN_queue_cwr(tp); } -static int tcp_check_sack_reneging(struct sock *sk, struct tcp_sock *tp) +static int tcp_check_sack_reneging(struct sock *sk) { struct sk_buff *skb; @@ -1268,9 +1275,10 @@ static int tcp_check_sack_reneging(struct sock *sk, struct tcp_sock *tp) NET_INC_STATS_BH(LINUX_MIB_TCPSACKRENEGING); tcp_enter_loss(sk, 1); - tp->retransmits++; + inet_csk(sk)->icsk_retransmits++; tcp_retransmit_skb(sk, skb_peek(&sk->sk_write_queue)); - tcp_reset_xmit_timer(sk, TCP_TIME_RETRANS, tp->rto); + inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, + inet_csk(sk)->icsk_rto); return 1; } return 0; @@ -1281,15 +1289,15 @@ static inline int tcp_fackets_out(struct tcp_sock *tp) return IsReno(tp) ? tp->sacked_out+1 : tp->fackets_out; } -static inline int tcp_skb_timedout(struct tcp_sock *tp, struct sk_buff *skb) +static inline int tcp_skb_timedout(struct sock *sk, struct sk_buff *skb) { - return (tcp_time_stamp - TCP_SKB_CB(skb)->when > tp->rto); + return (tcp_time_stamp - TCP_SKB_CB(skb)->when > inet_csk(sk)->icsk_rto); } static inline int tcp_head_timedout(struct sock *sk, struct tcp_sock *tp) { return tp->packets_out && - tcp_skb_timedout(tp, skb_peek(&sk->sk_write_queue)); + tcp_skb_timedout(sk, skb_peek(&sk->sk_write_queue)); } /* Linux NewReno/SACK/FACK/ECN state machine. @@ -1509,7 +1517,7 @@ static void tcp_update_scoreboard(struct sock *sk, struct tcp_sock *tp) struct sk_buff *skb; sk_stream_for_retrans_queue(skb, sk) { - if (tcp_skb_timedout(tp, skb) && + if (tcp_skb_timedout(sk, skb) && !(TCP_SKB_CB(skb)->sacked&TCPCB_TAGBITS)) { TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; tp->lost_out += tcp_skb_pcount(skb); @@ -1676,7 +1684,7 @@ static int tcp_try_undo_loss(struct sock *sk, struct tcp_sock *tp) tp->left_out = tp->sacked_out; tcp_undo_cwr(tp, 1); NET_INC_STATS_BH(LINUX_MIB_TCPLOSSUNDO); - tp->retransmits = 0; + inet_csk(sk)->icsk_retransmits = 0; tp->undo_marker = 0; if (!IsReno(tp)) tcp_set_ca_state(tp, TCP_CA_Open); @@ -1750,7 +1758,7 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una, tp->prior_ssthresh = 0; /* B. In all the states check for reneging SACKs. */ - if (tp->sacked_out && tcp_check_sack_reneging(sk, tp)) + if (tp->sacked_out && tcp_check_sack_reneging(sk)) return; /* C. Process data loss notification, provided it is valid. */ @@ -1774,7 +1782,7 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una, } else if (!before(tp->snd_una, tp->high_seq)) { switch (tp->ca_state) { case TCP_CA_Loss: - tp->retransmits = 0; + inet_csk(sk)->icsk_retransmits = 0; if (tcp_try_undo_recovery(sk, tp)) return; break; @@ -1824,7 +1832,7 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una, break; case TCP_CA_Loss: if (flag&FLAG_DATA_ACKED) - tp->retransmits = 0; + inet_csk(sk)->icsk_retransmits = 0; if (!tcp_try_undo_loss(sk, tp)) { tcp_moderate_cwnd(tp); tcp_xmit_retransmit_queue(sk); @@ -1881,10 +1889,8 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una, /* Read draft-ietf-tcplw-high-performance before mucking * with this code. (Superceeds RFC1323) */ -static void tcp_ack_saw_tstamp(struct tcp_sock *tp, u32 *usrtt, int flag) +static void tcp_ack_saw_tstamp(struct sock *sk, u32 *usrtt, 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 @@ -1900,14 +1906,15 @@ static void tcp_ack_saw_tstamp(struct tcp_sock *tp, u32 *usrtt, int flag) * 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->rx_opt.rcv_tsecr; + struct tcp_sock *tp = tcp_sk(sk); + const __u32 seq_rtt = tcp_time_stamp - tp->rx_opt.rcv_tsecr; tcp_rtt_estimator(tp, seq_rtt, usrtt); - tcp_set_rto(tp); - tp->backoff = 0; - tcp_bound_rto(tp); + tcp_set_rto(sk); + inet_csk(sk)->icsk_backoff = 0; + tcp_bound_rto(sk); } -static void tcp_ack_no_tstamp(struct tcp_sock *tp, u32 seq_rtt, u32 *usrtt, int flag) +static void tcp_ack_no_tstamp(struct sock *sk, u32 seq_rtt, u32 *usrtt, int flag) { /* We don't have a timestamp. Can only use * packets that are not retransmitted to determine @@ -1921,20 +1928,21 @@ static void tcp_ack_no_tstamp(struct tcp_sock *tp, u32 seq_rtt, u32 *usrtt, int if (flag & FLAG_RETRANS_DATA_ACKED) return; - tcp_rtt_estimator(tp, seq_rtt, usrtt); - tcp_set_rto(tp); - tp->backoff = 0; - tcp_bound_rto(tp); + tcp_rtt_estimator(tcp_sk(sk), seq_rtt, usrtt); + tcp_set_rto(sk); + inet_csk(sk)->icsk_backoff = 0; + tcp_bound_rto(sk); } -static inline void tcp_ack_update_rtt(struct tcp_sock *tp, - int flag, s32 seq_rtt, u32 *usrtt) +static inline void tcp_ack_update_rtt(struct sock *sk, const int flag, + const s32 seq_rtt, u32 *usrtt) { + const struct tcp_sock *tp = tcp_sk(sk); /* Note that peer MAY send zero echo. In this case it is ignored. (rfc1323) */ if (tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr) - tcp_ack_saw_tstamp(tp, usrtt, flag); + tcp_ack_saw_tstamp(sk, usrtt, flag); else if (seq_rtt >= 0) - tcp_ack_no_tstamp(tp, seq_rtt, usrtt, flag); + tcp_ack_no_tstamp(sk, seq_rtt, usrtt, flag); } static inline void tcp_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt, @@ -1951,9 +1959,9 @@ static inline void tcp_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt, static inline void tcp_ack_packets_out(struct sock *sk, struct tcp_sock *tp) { if (!tp->packets_out) { - tcp_clear_xmit_timer(sk, TCP_TIME_RETRANS); + inet_csk_clear_xmit_timer(sk, ICSK_TIME_RETRANS); } else { - tcp_reset_xmit_timer(sk, TCP_TIME_RETRANS, tp->rto); + inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, inet_csk(sk)->icsk_rto); } } @@ -2090,7 +2098,7 @@ static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p, s32 *seq_usrtt } if (acked&FLAG_ACKED) { - tcp_ack_update_rtt(tp, acked, seq_rtt, seq_usrtt); + tcp_ack_update_rtt(sk, acked, seq_rtt, seq_usrtt); tcp_ack_packets_out(sk, tp); if (tp->ca_ops->pkts_acked) @@ -2125,20 +2133,21 @@ static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p, s32 *seq_usrtt static void tcp_ack_probe(struct sock *sk) { - struct tcp_sock *tp = tcp_sk(sk); + const struct tcp_sock *tp = tcp_sk(sk); + struct inet_connection_sock *icsk = inet_csk(sk); /* Was it a usable window open? */ if (!after(TCP_SKB_CB(sk->sk_send_head)->end_seq, tp->snd_una + tp->snd_wnd)) { - tp->backoff = 0; - tcp_clear_xmit_timer(sk, TCP_TIME_PROBE0); + icsk->icsk_backoff = 0; + inet_csk_clear_xmit_timer(sk, ICSK_TIME_PROBE0); /* Socket must be waked up by subsequent tcp_data_s |