X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=net%2Fipv4%2Ftcp_ipv4.c;h=68860426e0377c2b2b3dfa16a18825e8df7d0808;hb=c7b5ebbddf7bcd3651947760f423e3783bbe6573;hp=ac92115041faaf7c1fede03c2f9594137872a13f;hpb=5273a3df6485dc2ad6aa7ddd441b9a21970f003b;p=linux-2.6.git diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index ac9211504..68860426e 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -74,6 +74,7 @@ #include #include #include +#include extern int sysctl_ip_dynaddr; int sysctl_tcp_tw_reuse; @@ -181,7 +182,6 @@ void tcp_bind_hash(struct sock *sk, struct tcp_bind_bucket *tb, static inline int tcp_bind_conflict(struct sock *sk, struct tcp_bind_bucket *tb) { - const u32 sk_rcv_saddr = tcp_v4_rcv_saddr(sk); struct sock *sk2; struct hlist_node *node; int reuse = sk->sk_reuse; @@ -194,9 +194,8 @@ static inline int tcp_bind_conflict(struct sock *sk, struct tcp_bind_bucket *tb) sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) { if (!reuse || !sk2->sk_reuse || sk2->sk_state == TCP_LISTEN) { - const u32 sk2_rcv_saddr = tcp_v4_rcv_saddr(sk2); - if (!sk2_rcv_saddr || !sk_rcv_saddr || - sk2_rcv_saddr == sk_rcv_saddr) + if (nx_addr_conflict(sk->sk_nx_info, + tcp_v4_rcv_saddr(sk), sk2)) break; } } @@ -405,6 +404,26 @@ void tcp_unhash(struct sock *sk) wake_up(&tcp_lhash_wait); } + +/* + * Check if a given address matches for a tcp socket + * + * nxi: the socket's nx_info if any + * addr: to be verified address + * saddr: socket addresses + */ +static inline int tcp_addr_match ( + struct nx_info *nxi, + uint32_t addr, + uint32_t saddr) +{ + if (addr && (saddr == addr)) + return 1; + if (!saddr) + return addr_in_nx_info(nxi, addr); + return 0; +} + /* Don't inline this cruft. Here are some nice properties to * exploit here. The BSD API does not allow a listening TCP * to specify the remote port nor the remote address for the @@ -426,11 +445,10 @@ static struct sock *__tcp_v4_lookup_listener(struct hlist_head *head, u32 daddr, __u32 rcv_saddr = inet->rcv_saddr; score = (sk->sk_family == PF_INET ? 1 : 0); - if (rcv_saddr) { - if (rcv_saddr != daddr) - continue; + if (tcp_addr_match(sk->sk_nx_info, daddr, rcv_saddr)) score+=2; - } + else + continue; if (sk->sk_bound_dev_if) { if (sk->sk_bound_dev_if != dif) continue; @@ -460,8 +478,8 @@ inline struct sock *tcp_v4_lookup_listener(u32 daddr, unsigned short hnum, struct inet_opt *inet = inet_sk((sk = __sk_head(head))); if (inet->num == hnum && !sk->sk_node.next && - (!inet->rcv_saddr || inet->rcv_saddr == daddr) && (sk->sk_family == PF_INET || !ipv6_only_sock(sk)) && + tcp_addr_match(sk->sk_nx_info, daddr, inet->rcv_saddr) && !sk->sk_bound_dev_if) goto sherry_cache; sk = __tcp_v4_lookup_listener(head, daddr, hnum, dif); @@ -618,11 +636,11 @@ unique: if (twp) { *twp = tw; - NET_INC_STATS_BH(TimeWaitRecycled); + NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED); } else if (tw) { /* Silly. Should hash-dance instead... */ tcp_tw_deschedule(tw); - NET_INC_STATS_BH(TimeWaitRecycled); + NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED); tcp_tw_put(tw); } @@ -998,14 +1016,14 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) int err; if (skb->len < (iph->ihl << 2) + 8) { - ICMP_INC_STATS_BH(IcmpInErrors); + ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); return; } sk = tcp_v4_lookup(iph->daddr, th->dest, iph->saddr, th->source, tcp_v4_iif(skb)); if (!sk) { - ICMP_INC_STATS_BH(IcmpInErrors); + ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); return; } if (sk->sk_state == TCP_TIME_WAIT) { @@ -1018,7 +1036,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) * servers this needs to be solved differently. */ if (sock_owned_by_user(sk)) - NET_INC_STATS_BH(LockDroppedIcmps); + NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS); if (sk->sk_state == TCP_CLOSE) goto out; @@ -1027,17 +1045,13 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) seq = ntohl(th->seq); if (sk->sk_state != TCP_LISTEN && !between(seq, tp->snd_una, tp->snd_nxt)) { - NET_INC_STATS(OutOfWindowIcmps); + NET_INC_STATS(LINUX_MIB_OUTOFWINDOWICMPS); goto out; } switch (type) { case ICMP_SOURCE_QUENCH: - /* This is deprecated, but if someone generated it, - * we have no reasons to ignore it. - */ - if (!sock_owned_by_user(sk)) - tcp_enter_cwr(tp); + /* Just silently ignore these. */ goto out; case ICMP_PARAMETERPROB: err = EPROTO; @@ -1078,7 +1092,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) BUG_TRAP(!req->sk); if (seq != req->snt_isn) { - NET_INC_STATS_BH(OutOfWindowIcmps); + NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -1096,7 +1110,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) It can f.e. if SYNs crossed. */ if (!sock_owned_by_user(sk)) { - TCP_INC_STATS_BH(TcpAttemptFails); + TCP_INC_STATS_BH(TCP_MIB_ATTEMPTFAILS); sk->sk_err = err; sk->sk_error_report(sk); @@ -1205,8 +1219,8 @@ static void tcp_v4_send_reset(struct sk_buff *skb) ip_send_reply(tcp_socket->sk, skb, &arg, sizeof rth); - TCP_INC_STATS_BH(TcpOutSegs); - TCP_INC_STATS_BH(TcpOutRsts); + TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); + TCP_INC_STATS_BH(TCP_MIB_OUTRSTS); } /* The code following below sending ACKs in SYN-RECV and TIME-WAIT states @@ -1253,7 +1267,7 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, ip_send_reply(tcp_socket->sk, skb, &arg, arg.iov[0].iov_len); - TCP_INC_STATS_BH(TcpOutSegs); + TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); } static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) @@ -1290,12 +1304,12 @@ static struct dst_entry* tcp_v4_route_req(struct sock *sk, .dport = req->rmt_port } } }; if (ip_route_output_flow(&rt, &fl, sk, 0)) { - IP_INC_STATS_BH(IpOutNoRoutes); + IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); return NULL; } if (opt && opt->is_strictroute && rt->rt_dst != rt->rt_gateway) { ip_rt_put(rt); - IP_INC_STATS_BH(IpOutNoRoutes); + IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); return NULL; } return &rt->u.dst; @@ -1442,7 +1456,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb) * clogging syn queue with openreqs with exponentially increasing * timeout. */ - if (tcp_acceptq_is_full(sk) && tcp_synq_young(sk) > 1) + if (sk_acceptq_is_full(sk) && tcp_synq_young(sk) > 1) goto drop; req = tcp_openreq_alloc(); @@ -1505,7 +1519,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb) if (xtime.tv_sec < peer->tcp_ts_stamp + TCP_PAWS_MSL && (s32)(peer->tcp_ts - req->ts_recent) > TCP_PAWS_WINDOW) { - NET_INC_STATS_BH(PAWSPassiveRejected); + NET_INC_STATS_BH(LINUX_MIB_PAWSPASSIVEREJECTED); dst_release(dst); goto drop_and_free; } @@ -1550,7 +1564,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb) drop_and_free: tcp_openreq_free(req); drop: - TCP_INC_STATS_BH(TcpAttemptFails); + TCP_INC_STATS_BH(TCP_MIB_ATTEMPTFAILS); return 0; } @@ -1567,7 +1581,7 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb, struct tcp_opt *newtp; struct sock *newsk; - if (tcp_acceptq_is_full(sk)) + if (sk_acceptq_is_full(sk)) goto exit_overflow; if (!dst && (dst = tcp_v4_route_req(sk, req)) == NULL) @@ -1605,9 +1619,9 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb, return newsk; exit_overflow: - NET_INC_STATS_BH(ListenOverflows); + NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS); exit: - NET_INC_STATS_BH(ListenDrops); + NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS); dst_release(dst); return NULL; } @@ -1725,7 +1739,7 @@ discard: return 0; csum_err: - TCP_INC_STATS_BH(TcpInErrs); + TCP_INC_STATS_BH(TCP_MIB_INERRS); goto discard; } @@ -1743,7 +1757,7 @@ int tcp_v4_rcv(struct sk_buff *skb) goto discard_it; /* Count it even if it's bad */ - TCP_INC_STATS_BH(TcpInSegs); + TCP_INC_STATS_BH(TCP_MIB_INSEGS); if (!pskb_may_pull(skb, sizeof(struct tcphdr))) goto discard_it; @@ -1810,7 +1824,7 @@ no_tcp_socket: if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) { bad_packet: - TCP_INC_STATS_BH(TcpInErrs); + TCP_INC_STATS_BH(TCP_MIB_INERRS); } else { tcp_v4_send_reset(skb); } @@ -1831,7 +1845,7 @@ do_time_wait: } if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) { - TCP_INC_STATS_BH(TcpInErrs); + TCP_INC_STATS_BH(TCP_MIB_INERRS); tcp_tw_put((struct tcp_tw_bucket *) sk); goto discard_it; } @@ -2075,13 +2089,13 @@ static int tcp_v4_init_sock(struct sock *sk) */ tp->snd_ssthresh = 0x7fffffff; /* Infinity */ tp->snd_cwnd_clamp = ~0; - tp->mss_cache = 536; + tp->mss_cache_std = tp->mss_cache = 536; tp->reordering = sysctl_tcp_reordering; sk->sk_state = TCP_CLOSE; - sk->sk_write_space = tcp_write_space; + sk->sk_write_space = sk_stream_write_space; sk->sk_use_write_queue = 1; tp->af_specific = &ipv4_specific; @@ -2094,14 +2108,14 @@ static int tcp_v4_init_sock(struct sock *sk) return 0; } -static int tcp_v4_destroy_sock(struct sock *sk) +int tcp_v4_destroy_sock(struct sock *sk) { struct tcp_opt *tp = tcp_sk(sk); tcp_clear_xmit_timers(sk); /* Cleanup up the write buffer. */ - tcp_writequeue_purge(sk); + sk_stream_writequeue_purge(sk); /* Cleans up our, hopefully empty, out_of_order_queue. */ __skb_queue_purge(&tp->out_of_order_queue); @@ -2113,15 +2127,21 @@ static int tcp_v4_destroy_sock(struct sock *sk) if (tp->bind_hash) tcp_put_port(sk); - /* If sendmsg cached page exists, toss it. */ - if (inet_sk(sk)->sndmsg_page) - __free_page(inet_sk(sk)->sndmsg_page); + /* + * If sendmsg cached page exists, toss it. + */ + if (sk->sk_sndmsg_page) { + __free_page(sk->sk_sndmsg_page); + sk->sk_sndmsg_page = NULL; + } atomic_dec(&tcp_sockets_allocated); return 0; } +EXPORT_SYMBOL(tcp_v4_destroy_sock); + #ifdef CONFIG_PROC_FS /* Proc filesystem TCP sock list dumping. */ @@ -2159,6 +2179,12 @@ static void *listening_get_next(struct seq_file *seq, void *cur) req = req->dl_next; while (1) { while (req) { + vxdprintk(VXD_CBIT(net, 6), + "sk,req: %p [#%d] (from %d)", req->sk, + (req->sk)?req->sk->sk_xid:0, current->xid); + if (req->sk && + !vx_check(req->sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (req->class->family == st->family) { cur = req; goto out; @@ -2173,10 +2199,20 @@ get_req: sk = sk_next(st->syn_wait_sk); st->state = TCP_SEQ_STATE_LISTENING; read_unlock_bh(&tp->syn_wait_lock); - } else + } else { + tp = tcp_sk(sk); + read_lock_bh(&tp->syn_wait_lock); + if (tp->listen_opt && tp->listen_opt->qlen) + goto start_req; + read_unlock_bh(&tp->syn_wait_lock); sk = sk_next(sk); + } get_sk: sk_for_each_from(sk, node) { + vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)", + sk, sk->sk_xid, current->xid); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (sk->sk_family == st->family) { cur = sk; goto out; @@ -2184,6 +2220,7 @@ get_sk: tp = tcp_sk(sk); read_lock_bh(&tp->syn_wait_lock); if (tp->listen_opt && tp->listen_opt->qlen) { +start_req: st->uid = sock_i_uid(sk); st->syn_wait_sk = sk; st->state = TCP_SEQ_STATE_OPENREQ; @@ -2224,18 +2261,26 @@ static void *established_get_first(struct seq_file *seq) read_lock(&tcp_ehash[st->bucket].lock); sk_for_each(sk, node, &tcp_ehash[st->bucket].chain) { - if (sk->sk_family != st->family) { + vxdprintk(VXD_CBIT(net, 6), + "sk,egf: %p [#%d] (from %d)", + sk, sk->sk_xid, current->xid); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; + if (sk->sk_family != st->family) continue; - } rc = sk; goto out; } st->state = TCP_SEQ_STATE_TIME_WAIT; tw_for_each(tw, node, &tcp_ehash[st->bucket + tcp_ehash_size].chain) { - if (tw->tw_family != st->family) { + vxdprintk(VXD_CBIT(net, 6), + "tw: %p [#%d] (from %d)", + tw, tw->tw_xid, current->xid); + if (!vx_check(tw->tw_xid, VX_IDENT|VX_WATCH)) + continue; + if (tw->tw_family != st->family) continue; - } rc = tw; goto out; } @@ -2259,7 +2304,8 @@ static void *established_get_next(struct seq_file *seq, void *cur) tw = cur; tw = tw_next(tw); get_tw: - while (tw && tw->tw_family != st->family) { + while (tw && (tw->tw_family != st->family || + !vx_check(tw->tw_xid, VX_IDENT|VX_WATCH))) { tw = tw_next(tw); } if (tw) { @@ -2279,6 +2325,11 @@ get_tw: sk = sk_next(sk); sk_for_each_from(sk, node) { + vxdprintk(VXD_CBIT(net, 6), + "sk,egn: %p [#%d] (from %d)", + sk, sk->sk_xid, current->xid); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (sk->sk_family == st->family) goto found; } @@ -2452,7 +2503,7 @@ static void get_openreq4(struct sock *sk, struct open_request *req, int ttd = req->expires - jiffies; sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X" - " %02X %08X:%08X %02X:%08X %08X %5d %8d %u %d %p", + " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p", i, req->af.v4_req.loc_addr, ntohs(inet_sk(sk)->sport), @@ -2526,7 +2577,7 @@ static void get_timewait4_sock(struct tcp_tw_bucket *tw, char *tmpbuf, int i) srcp = ntohs(tw->tw_sport); sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X" - " %02X %08X:%08X %02X:%08X %08X %5d %8d %d %d %p", + " %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p", i, src, srcp, dest, destp, tw->tw_substate, 0, 0, 3, jiffies_to_clock_t(ttd), 0, 0, 0, 0, atomic_read(&tw->tw_refcnt), tw); @@ -2586,23 +2637,32 @@ void tcp4_proc_exit(void) #endif /* CONFIG_PROC_FS */ struct proto tcp_prot = { - .name = "TCP", - .close = tcp_close, - .connect = tcp_v4_connect, - .disconnect = tcp_disconnect, - .accept = tcp_accept, - .ioctl = tcp_ioctl, - .init = tcp_v4_init_sock, - .destroy = tcp_v4_destroy_sock, - .shutdown = tcp_shutdown, - .setsockopt = tcp_setsockopt, - .getsockopt = tcp_getsockopt, - .sendmsg = tcp_sendmsg, - .recvmsg = tcp_recvmsg, - .backlog_rcv = tcp_v4_do_rcv, - .hash = tcp_v4_hash, - .unhash = tcp_unhash, - .get_port = tcp_v4_get_port, + .name = "TCP", + .close = tcp_close, + .connect = tcp_v4_connect, + .disconnect = tcp_disconnect, + .accept = tcp_accept, + .ioctl = tcp_ioctl, + .init = tcp_v4_init_sock, + .destroy = tcp_v4_destroy_sock, + .shutdown = tcp_shutdown, + .setsockopt = tcp_setsockopt, + .getsockopt = tcp_getsockopt, + .sendmsg = tcp_sendmsg, + .recvmsg = tcp_recvmsg, + .backlog_rcv = tcp_v4_do_rcv, + .hash = tcp_v4_hash, + .unhash = tcp_unhash, + .get_port = tcp_v4_get_port, + .enter_memory_pressure = tcp_enter_memory_pressure, + .sockets_allocated = &tcp_sockets_allocated, + .memory_allocated = &tcp_memory_allocated, + .memory_pressure = &tcp_memory_pressure, + .sysctl_mem = sysctl_tcp_mem, + .sysctl_wmem = sysctl_tcp_wmem, + .sysctl_rmem = sysctl_tcp_rmem, + .max_header = MAX_TCP_HEADER, + .slab_obj_size = sizeof(struct tcp_sock), };