*
* Implementation of the Transmission Control Protocol(TCP).
*
- * Version: $Id: tcp_ipv4.c,v 1.240 2002/02/01 22:01:04 davem Exp $
- *
* IPv4 specific functions
*
*
int sysctl_tcp_tw_reuse __read_mostly;
int sysctl_tcp_low_latency __read_mostly;
-/* Check TCP sequence numbers in ICMP packets. */
-#define ICMP_MIN_LENGTH 8
-
-/* Socket used for sending RSTs */
-static struct sock *tcp_sock __read_mostly;
-
-void tcp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb);
#ifdef CONFIG_TCP_MD5SIG
static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk,
__be32 addr);
static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
__be32 saddr, __be32 daddr,
- struct tcphdr *th, int protocol,
- unsigned int tcplen);
+ struct tcphdr *th, unsigned int tcplen);
+#else
+static inline
+struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
+{
+ return NULL;
+}
#endif
struct inet_hashinfo __cacheline_aligned tcp_hashinfo = {
key,
ip_hdr(skb)->daddr,
ip_hdr(skb)->saddr,
- &rep.th, IPPROTO_TCP,
- arg.iov[0].iov_len);
+ &rep.th, arg.iov[0].iov_len);
}
#endif
arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
sizeof(struct tcphdr), IPPROTO_TCP, 0);
arg.csumoffset = offsetof(struct tcphdr, check) / 2;
- ip_send_reply(tcp_sock, skb, &arg, arg.iov[0].iov_len);
+ ip_send_reply(dev_net(skb->dst->dev)->ipv4.tcp_sock, skb,
+ &arg, arg.iov[0].iov_len);
TCP_INC_STATS_BH(TCP_MIB_OUTSEGS);
TCP_INC_STATS_BH(TCP_MIB_OUTRSTS);
outside socket context is ugly, certainly. What can I do?
*/
-static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk,
- struct sk_buff *skb, u32 seq, u32 ack,
- u32 win, u32 ts)
+static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
+ u32 win, u32 ts, int oif,
+ struct tcp_md5sig_key *key)
{
struct tcphdr *th = tcp_hdr(skb);
struct {
];
} rep;
struct ip_reply_arg arg;
-#ifdef CONFIG_TCP_MD5SIG
- struct tcp_md5sig_key *key;
- struct tcp_md5sig_key tw_key;
-#endif
memset(&rep.th, 0, sizeof(struct tcphdr));
memset(&arg, 0, sizeof(arg));
rep.th.window = htons(win);
#ifdef CONFIG_TCP_MD5SIG
- /*
- * The SKB holds an imcoming packet, but may not have a valid ->sk
- * pointer. This is especially the case when we're dealing with a
- * TIME_WAIT ack, because the sk structure is long gone, and only
- * the tcp_timewait_sock remains. So the md5 key is stashed in that
- * structure, and we use it in preference. I believe that (twsk ||
- * skb->sk) holds true, but we program defensively.
- */
- if (!twsk && skb->sk) {
- key = tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr);
- } else if (twsk && twsk->tw_md5_keylen) {
- tw_key.key = twsk->tw_md5_key;
- tw_key.keylen = twsk->tw_md5_keylen;
- key = &tw_key;
- } else
- key = NULL;
-
if (key) {
int offset = (ts) ? 3 : 0;
key,
ip_hdr(skb)->daddr,
ip_hdr(skb)->saddr,
- &rep.th, IPPROTO_TCP,
- arg.iov[0].iov_len);
+ &rep.th, arg.iov[0].iov_len);
}
#endif
arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
ip_hdr(skb)->saddr, /* XXX */
arg.iov[0].iov_len, IPPROTO_TCP, 0);
arg.csumoffset = offsetof(struct tcphdr, check) / 2;
- if (twsk)
- arg.bound_dev_if = twsk->tw_sk.tw_bound_dev_if;
+ if (oif)
+ arg.bound_dev_if = oif;
- ip_send_reply(tcp_sock, skb, &arg, arg.iov[0].iov_len);
+ ip_send_reply(dev_net(skb->dev)->ipv4.tcp_sock, skb,
+ &arg, arg.iov[0].iov_len);
TCP_INC_STATS_BH(TCP_MIB_OUTSEGS);
}
struct inet_timewait_sock *tw = inet_twsk(sk);
struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
- tcp_v4_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
+ tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
- tcptw->tw_ts_recent);
+ tcptw->tw_ts_recent,
+ tw->tw_bound_dev_if,
+ tcp_twsk_md5_key(tcptw)
+ );
inet_twsk_put(tw);
}
static void tcp_v4_reqsk_send_ack(struct sk_buff *skb,
struct request_sock *req)
{
- tcp_v4_send_ack(NULL, skb, tcp_rsk(req)->snt_isn + 1,
+ tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1,
tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd,
- req->ts_recent);
+ req->ts_recent,
+ 0,
+ tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr));
}
/*
static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
__be32 saddr, __be32 daddr,
- struct tcphdr *th, int protocol,
+ struct tcphdr *th,
unsigned int tcplen)
{
- struct scatterlist sg[4];
- __u16 data_len;
- int block = 0;
- __sum16 old_checksum;
struct tcp_md5sig_pool *hp;
struct tcp4_pseudohdr *bp;
- struct hash_desc *desc;
int err;
- unsigned int nbytes = 0;
/*
* Okay, so RFC2385 is turned on for this connection,
goto clear_hash_noput;
bp = &hp->md5_blk.ip4;
- desc = &hp->md5_desc;
/*
- * 1. the TCP pseudo-header (in the order: source IP address,
+ * The TCP pseudo-header (in the order: source IP address,
* destination IP address, zero-padded protocol number, and
* segment length)
*/
bp->saddr = saddr;
bp->daddr = daddr;
bp->pad = 0;
- bp->protocol = protocol;
+ bp->protocol = IPPROTO_TCP;
bp->len = htons(tcplen);
- sg_init_table(sg, 4);
-
- sg_set_buf(&sg[block++], bp, sizeof(*bp));
- nbytes += sizeof(*bp);
-
- /* 2. the TCP header, excluding options, and assuming a
- * checksum of zero/
- */
- old_checksum = th->check;
- th->check = 0;
- sg_set_buf(&sg[block++], th, sizeof(struct tcphdr));
- nbytes += sizeof(struct tcphdr);
-
- /* 3. the TCP segment data (if any) */
- data_len = tcplen - (th->doff << 2);
- if (data_len > 0) {
- unsigned char *data = (unsigned char *)th + (th->doff << 2);
- sg_set_buf(&sg[block++], data, data_len);
- nbytes += data_len;
- }
-
- /* 4. an independently-specified key or password, known to both
- * TCPs and presumably connection-specific
- */
- sg_set_buf(&sg[block++], key->key, key->keylen);
- nbytes += key->keylen;
-
- sg_mark_end(&sg[block - 1]);
-
- /* Now store the Hash into the packet */
- err = crypto_hash_init(desc);
- if (err)
- goto clear_hash;
- err = crypto_hash_update(desc, sg, nbytes);
- if (err)
- goto clear_hash;
- err = crypto_hash_final(desc, md5_hash);
+ err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp),
+ th, tcplen, hp);
if (err)
goto clear_hash;
- /* Reset header, and free up the crypto */
+ /* Free up the crypto pool */
tcp_put_md5sig_pool();
- th->check = old_checksum;
-
out:
return 0;
clear_hash:
struct sock *sk,
struct dst_entry *dst,
struct request_sock *req,
- struct tcphdr *th, int protocol,
+ struct tcphdr *th,
unsigned int tcplen)
{
__be32 saddr, daddr;
}
return tcp_v4_do_calc_md5_hash(md5_hash, key,
saddr, daddr,
- th, protocol, tcplen);
+ th, tcplen);
}
EXPORT_SYMBOL(tcp_v4_calc_md5_hash);
struct tcp_md5sig_key *hash_expected;
const struct iphdr *iph = ip_hdr(skb);
struct tcphdr *th = tcp_hdr(skb);
- int length = (th->doff << 2) - sizeof(struct tcphdr);
int genhash;
- unsigned char *ptr;
unsigned char newhash[16];
hash_expected = tcp_v4_md5_do_lookup(sk, iph->saddr);
+ hash_location = tcp_parse_md5sig_option(th);
- /*
- * If the TCP option length is less than the TCP_MD5SIG
- * option length, then we can shortcut
- */
- if (length < TCPOLEN_MD5SIG) {
- if (hash_expected)
- return 1;
- else
- return 0;
- }
-
- /* Okay, we can't shortcut - we have to grub through the options */
- ptr = (unsigned char *)(th + 1);
- while (length > 0) {
- int opcode = *ptr++;
- int opsize;
-
- switch (opcode) {
- case TCPOPT_EOL:
- goto done_opts;
- case TCPOPT_NOP:
- length--;
- continue;
- default:
- opsize = *ptr++;
- if (opsize < 2)
- goto done_opts;
- if (opsize > length)
- goto done_opts;
-
- if (opcode == TCPOPT_MD5SIG) {
- hash_location = ptr;
- goto done_opts;
- }
- }
- ptr += opsize-2;
- length -= opsize;
- }
-done_opts:
/* We've parsed the options - do we have a hash? */
if (!hash_expected && !hash_location)
return 0;
genhash = tcp_v4_do_calc_md5_hash(newhash,
hash_expected,
iph->saddr, iph->daddr,
- th, sk->sk_protocol,
- skb->len);
+ th, skb->len);
if (genhash || memcmp(hash_location, newhash, 16) != 0) {
if (net_ratelimit()) {
if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
goto drop;
- req = reqsk_alloc(&tcp_request_sock_ops);
+ req = inet_reqsk_alloc(&tcp_request_sock_ops);
if (!req)
goto drop;
tcp_parse_options(skb, &tmp_opt, 0);
- if (want_cookie) {
+ if (want_cookie && !tmp_opt.saw_tstamp)
tcp_clear_options(&tmp_opt);
- tmp_opt.saw_tstamp = 0;
- }
if (tmp_opt.saw_tstamp && !tmp_opt.rcv_tsval) {
/* Some OSes (unknown ones, but I see them on web server, which
if (want_cookie) {
#ifdef CONFIG_SYN_COOKIES
syn_flood_warning(skb);
+ req->cookie_ts = tmp_opt.tstamp_ok;
#endif
isn = cookie_v4_init_sequence(sk, skb, &req->mss);
} else if (!isn) {
* to the moment of synflood.
*/
LIMIT_NETDEBUG(KERN_DEBUG "TCP: drop open "
- "request from %u.%u.%u.%u/%u\n",
+ "request from " NIPQUAD_FMT "/%u\n",
NIPQUAD(saddr),
ntohs(tcp_hdr(skb)->source));
goto drop_and_release;
return 0;
}
-int tcp_v4_destroy_sock(struct sock *sk)
+void tcp_v4_destroy_sock(struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
sk->sk_sndmsg_page = NULL;
}
- if (tp->defer_tcp_accept.request) {
- reqsk_free(tp->defer_tcp_accept.request);
- sock_put(tp->defer_tcp_accept.listen_sk);
- sock_put(sk);
- tp->defer_tcp_accept.listen_sk = NULL;
- tp->defer_tcp_accept.request = NULL;
- }
-
atomic_dec(&tcp_sockets_allocated);
-
- return 0;
}
EXPORT_SYMBOL(tcp_v4_destroy_sock);
struct hlist_node *node;
struct sock *sk = cur;
struct tcp_iter_state* st = seq->private;
- struct net *net = st->net;
+ struct net *net = seq_file_net(seq);
if (!sk) {
st->bucket = 0;
static void *established_get_first(struct seq_file *seq)
{
struct tcp_iter_state* st = seq->private;
- struct net *net = st->net;
+ struct net *net = seq_file_net(seq);
void *rc = NULL;
for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
struct inet_timewait_sock *tw;
struct hlist_node *node;
struct tcp_iter_state* st = seq->private;
- struct net *net = st->net;
+ struct net *net = seq_file_net(seq);
++st->num;
static int tcp_seq_open(struct inode *inode, struct file *file)
{
struct tcp_seq_afinfo *afinfo = PDE(inode)->data;
- struct seq_file *seq;
struct tcp_iter_state *s;
- struct net *net;
- int rc;
-
- if (unlikely(afinfo == NULL))
- return -EINVAL;
-
- s = kzalloc(sizeof(*s), GFP_KERNEL);
- if (!s)
- return -ENOMEM;
+ int err;
- rc = -ENXIO;
- net = get_proc_net(inode);
- if (!net)
- goto out_kfree;
+ err = seq_open_net(inode, file, &afinfo->seq_ops,
+ sizeof(struct tcp_iter_state));
+ if (err < 0)
+ return err;
+ s = ((struct seq_file *)file->private_data)->private;
s->family = afinfo->family;
- s->seq_ops.start = tcp_seq_start;
- s->seq_ops.next = tcp_seq_next;
- s->seq_ops.show = afinfo->seq_show;
- s->seq_ops.stop = tcp_seq_stop;
- s->net = net;
-
- rc = seq_open(file, &s->seq_ops);
- if (rc)
- goto out_put_net;
- seq = file->private_data;
- seq->private = s;
-out:
- return rc;
-out_put_net:
- put_net(net);
-out_kfree:
- kfree(s);
- goto out;
-}
-
-static int tcp_seq_release(struct inode *inode, struct file *file)
-{
- struct seq_file *seq = file->private_data;
- struct tcp_iter_state *s = seq->private;
-
- put_net(s->net);
- seq_release_private(inode, file);
return 0;
}
int rc = 0;
struct proc_dir_entry *p;
- if (!afinfo)
- return -EINVAL;
- afinfo->seq_fops->owner = afinfo->owner;
- afinfo->seq_fops->open = tcp_seq_open;
- afinfo->seq_fops->read = seq_read;
- afinfo->seq_fops->llseek = seq_lseek;
- afinfo->seq_fops->release = tcp_seq_release;
-
- p = proc_net_fops_create(net, afinfo->name, S_IRUGO, afinfo->seq_fops);
- if (p)
- p->data = afinfo;
- else
+ afinfo->seq_fops.open = tcp_seq_open;
+ afinfo->seq_fops.read = seq_read;
+ afinfo->seq_fops.llseek = seq_lseek;
+ afinfo->seq_fops.release = seq_release_net;
+
+ afinfo->seq_ops.start = tcp_seq_start;
+ afinfo->seq_ops.next = tcp_seq_next;
+ afinfo->seq_ops.stop = tcp_seq_stop;
+
+ p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
+ &afinfo->seq_fops, afinfo);
+ if (!p)
rc = -ENOMEM;
return rc;
}
void tcp_proc_unregister(struct net *net, struct tcp_seq_afinfo *afinfo)
{
- if (!afinfo)
- return;
proc_net_remove(net, afinfo->name);
- memset(afinfo->seq_fops, 0, sizeof(*afinfo->seq_fops));
}
static void get_openreq4(struct sock *sk, struct request_sock *req,
- char *tmpbuf, int i, int uid)
+ struct seq_file *f, int i, int uid, int *len)
{
const struct inet_request_sock *ireq = inet_rsk(req);
int ttd = req->expires - jiffies;
- sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
- " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p",
+ seq_printf(f, "%4d: %08X:%04X %08X:%04X"
+ " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p%n",
i,
ireq->loc_addr,
ntohs(inet_sk(sk)->sport),
0, /* non standard timer */
0, /* open_requests have no inode */
atomic_read(&sk->sk_refcnt),
- req);
+ req,
+ len);
}
-static void get_tcp4_sock(struct sock *sk, char *tmpbuf, int i)
+static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i, int *len)
{
int timer_active;
unsigned long timer_expires;
timer_expires = jiffies;
}
- sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
- "%08X %5d %8d %lu %d %p %u %u %u %u %d",
+ seq_printf(f, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
+ "%08X %5d %8d %lu %d %p %lu %lu %u %u %d%n",
i, src, srcp, dest, destp, sk->sk_state,
tp->write_seq - tp->snd_una,
sk->sk_state == TCP_LISTEN ? sk->sk_ack_backlog :
icsk->icsk_probes_out,
sock_i_ino(sk),
atomic_read(&sk->sk_refcnt), sk,
- icsk->icsk_rto,
- icsk->icsk_ack.ato,
+ jiffies_to_clock_t(icsk->icsk_rto),
+ jiffies_to_clock_t(icsk->icsk_ack.ato),
(icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
tp->snd_cwnd,
- tp->snd_ssthresh >= 0xFFFF ? -1 : tp->snd_ssthresh);
+ tp->snd_ssthresh >= 0xFFFF ? -1 : tp->snd_ssthresh,
+ len);
}
static void get_timewait4_sock(struct inet_timewait_sock *tw,
- char *tmpbuf, int i)
+ struct seq_file *f, int i, int *len)
{
__be32 dest, src;
__u16 destp, srcp;
destp = ntohs(tw->tw_dport);
srcp = ntohs(tw->tw_sport);
- sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
- " %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p",
+ seq_printf(f, "%4d: %08X:%04X %08X:%04X"
+ " %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p%n",
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);
+ atomic_read(&tw->tw_refcnt), tw, len);
}
#define TMPSZ 150
static int tcp4_seq_show(struct seq_file *seq, void *v)
{
struct tcp_iter_state* st;
- char tmpbuf[TMPSZ + 1];
+ int len;
if (v == SEQ_START_TOKEN) {
seq_printf(seq, "%-*s\n", TMPSZ - 1,
switch (st->state) {
case TCP_SEQ_STATE_LISTENING:
case TCP_SEQ_STATE_ESTABLISHED:
- get_tcp4_sock(v, tmpbuf, st->num);
+ get_tcp4_sock(v, seq, st->num, &len);
break;
case TCP_SEQ_STATE_OPENREQ:
- get_openreq4(st->syn_wait_sk, v, tmpbuf, st->num, st->uid);
+ get_openreq4(st->syn_wait_sk, v, seq, st->num, st->uid, &len);
break;
case TCP_SEQ_STATE_TIME_WAIT:
- get_timewait4_sock(v, tmpbuf, st->num);
+ get_timewait4_sock(v, seq, st->num, &len);
break;
}
- seq_printf(seq, "%-*s\n", TMPSZ - 1, tmpbuf);
+ seq_printf(seq, "%*s\n", TMPSZ - 1 - len, "");
out:
return 0;
}
-static struct file_operations tcp4_seq_fops;
static struct tcp_seq_afinfo tcp4_seq_afinfo = {
- .owner = THIS_MODULE,
.name = "tcp",
.family = AF_INET,
- .seq_show = tcp4_seq_show,
- .seq_fops = &tcp4_seq_fops,
+ .seq_fops = {
+ .owner = THIS_MODULE,
+ },
+ .seq_ops = {
+ .show = tcp4_seq_show,
+ },
};
static int tcp4_proc_init_net(struct net *net)
#endif
};
+
+static int __net_init tcp_sk_init(struct net *net)
+{
+ return inet_ctl_sock_create(&net->ipv4.tcp_sock,
+ PF_INET, SOCK_RAW, IPPROTO_TCP, net);
+}
+
+static void __net_exit tcp_sk_exit(struct net *net)
+{
+ inet_ctl_sock_destroy(net->ipv4.tcp_sock);
+}
+
+static struct pernet_operations __net_initdata tcp_sk_ops = {
+ .init = tcp_sk_init,
+ .exit = tcp_sk_exit,
+};
+
void __init tcp_v4_init(void)
{
- if (inet_ctl_sock_create(&tcp_sock, PF_INET, SOCK_RAW,
- IPPROTO_TCP, &init_net) < 0)
+ if (register_pernet_device(&tcp_sk_ops))
panic("Failed to create the TCP control socket.\n");
}