[NET] Generalise tcp_listen_opt
This chunks out the accept_queue and tcp_listen_opt code and moves
them to net/core/request_sock.c and include/net/request_sock.h, to
make it useful for other transport protocols, DCCP being the first one
to use it.
Next patches will rename tcp_listen_opt to accept_sock and remove the
inline tcp functions that just call a reqsk_queue_ function.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
diff --git a/net/core/Makefile b/net/core/Makefile
index 81f0324..5e0c56b 100644
--- a/net/core/Makefile
+++ b/net/core/Makefile
@@ -2,7 +2,8 @@
# Makefile for the Linux networking core.
#
-obj-y := sock.o skbuff.o iovec.o datagram.o stream.o scm.o gen_stats.o gen_estimator.o
+obj-y := sock.o request_sock.o skbuff.o iovec.o datagram.o stream.o scm.o \
+ gen_stats.o gen_estimator.o
obj-$(CONFIG_SYSCTL) += sysctl_net_core.o
diff --git a/net/core/request_sock.c b/net/core/request_sock.c
new file mode 100644
index 0000000..1258333
--- /dev/null
+++ b/net/core/request_sock.c
@@ -0,0 +1,48 @@
+/*
+ * NET Generic infrastructure for Network protocols.
+ *
+ * Authors: Arnaldo Carvalho de Melo <[email protected]>
+ *
+ * From code originally in include/net/tcp.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/random.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+
+#include <net/request_sock.h>
+
+int reqsk_queue_alloc(struct request_sock_queue *queue,
+ const int nr_table_entries)
+{
+ const int lopt_size = sizeof(struct tcp_listen_opt) +
+ nr_table_entries * sizeof(struct request_sock *);
+ struct tcp_listen_opt *lopt = kmalloc(lopt_size, GFP_KERNEL);
+
+ if (lopt == NULL)
+ return -ENOMEM;
+
+ memset(lopt, 0, lopt_size);
+
+ for (lopt->max_qlen_log = 6;
+ (1 << lopt->max_qlen_log) < sysctl_max_syn_backlog;
+ lopt->max_qlen_log++);
+
+ 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;
+
+ write_lock_bh(&queue->syn_wait_lock);
+ queue->listen_opt = lopt;
+ write_unlock_bh(&queue->syn_wait_lock);
+
+ return 0;
+}
+
+EXPORT_SYMBOL(reqsk_queue_alloc);
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 1c29feb..b85a46d 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -316,7 +316,7 @@
static __inline__ unsigned int tcp_listen_poll(struct sock *sk,
poll_table *wait)
{
- return tcp_sk(sk)->accept_queue ? (POLLIN | POLLRDNORM) : 0;
+ return !reqsk_queue_empty(&tcp_sk(sk)->accept_queue) ? (POLLIN | POLLRDNORM) : 0;
}
/*
@@ -462,28 +462,15 @@
{
struct inet_sock *inet = inet_sk(sk);
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt;
+ int rc = reqsk_queue_alloc(&tp->accept_queue, TCP_SYNQ_HSIZE);
+
+ if (rc != 0)
+ return rc;
sk->sk_max_ack_backlog = 0;
sk->sk_ack_backlog = 0;
- tp->accept_queue = tp->accept_queue_tail = NULL;
- rwlock_init(&tp->syn_wait_lock);
tcp_delack_init(tp);
- lopt = kmalloc(sizeof(struct tcp_listen_opt), GFP_KERNEL);
- if (!lopt)
- return -ENOMEM;
-
- memset(lopt, 0, sizeof(struct tcp_listen_opt));
- for (lopt->max_qlen_log = 6; ; lopt->max_qlen_log++)
- if ((1 << lopt->max_qlen_log) >= sysctl_max_syn_backlog)
- break;
- get_random_bytes(&lopt->hash_rnd, 4);
-
- write_lock_bh(&tp->syn_wait_lock);
- tp->listen_opt = lopt;
- write_unlock_bh(&tp->syn_wait_lock);
-
/* There is race window here: we announce ourselves listening,
* but this transition is still not validated by get_port().
* It is OK, because this socket enters to hash table only
@@ -500,10 +487,7 @@
}
sk->sk_state = TCP_CLOSE;
- write_lock_bh(&tp->syn_wait_lock);
- tp->listen_opt = NULL;
- write_unlock_bh(&tp->syn_wait_lock);
- kfree(lopt);
+ reqsk_queue_destroy(&tp->accept_queue);
return -EADDRINUSE;
}
@@ -515,18 +499,16 @@
static void tcp_listen_stop (struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
- struct request_sock *acc_req = tp->accept_queue;
+ struct tcp_listen_opt *lopt;
+ struct request_sock *acc_req;
struct request_sock *req;
int i;
tcp_delete_keepalive_timer(sk);
/* make all the listen_opt local to us */
- write_lock_bh(&tp->syn_wait_lock);
- tp->listen_opt = NULL;
- write_unlock_bh(&tp->syn_wait_lock);
- tp->accept_queue = tp->accept_queue_tail = NULL;
+ lopt = reqsk_queue_yank_listen_sk(&tp->accept_queue);
+ acc_req = reqsk_queue_yank_acceptq(&tp->accept_queue);
if (lopt->qlen) {
for (i = 0; i < TCP_SYNQ_HSIZE; i++) {
@@ -1867,11 +1849,11 @@
prepare_to_wait_exclusive(sk->sk_sleep, &wait,
TASK_INTERRUPTIBLE);
release_sock(sk);
- if (!tp->accept_queue)
+ if (reqsk_queue_empty(&tp->accept_queue))
timeo = schedule_timeout(timeo);
lock_sock(sk);
err = 0;
- if (tp->accept_queue)
+ if (!reqsk_queue_empty(&tp->accept_queue))
break;
err = -EINVAL;
if (sk->sk_state != TCP_LISTEN)
@@ -1894,7 +1876,6 @@
struct sock *tcp_accept(struct sock *sk, int flags, int *err)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct request_sock *req;
struct sock *newsk;
int error;
@@ -1905,37 +1886,31 @@
*/
error = -EINVAL;
if (sk->sk_state != TCP_LISTEN)
- goto out;
+ goto out_err;
/* Find already established connection */
- if (!tp->accept_queue) {
+ if (reqsk_queue_empty(&tp->accept_queue)) {
long timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
/* If this is a non blocking socket don't sleep */
error = -EAGAIN;
if (!timeo)
- goto out;
+ goto out_err;
error = wait_for_connect(sk, timeo);
if (error)
- goto out;
+ goto out_err;
}
- req = tp->accept_queue;
- if ((tp->accept_queue = req->dl_next) == NULL)
- tp->accept_queue_tail = NULL;
-
- newsk = req->sk;
- sk_acceptq_removed(sk);
- __reqsk_free(req);
+ newsk = reqsk_queue_get_child(&tp->accept_queue, sk);
BUG_TRAP(newsk->sk_state != TCP_SYN_RECV);
- release_sock(sk);
- return newsk;
-
out:
release_sock(sk);
+ return newsk;
+out_err:
+ newsk = NULL;
*err = error;
- return NULL;
+ goto out;
}
/*
diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c
index 6727780..c3328fa 100644
--- a/net/ipv4/tcp_diag.c
+++ b/net/ipv4/tcp_diag.c
@@ -529,9 +529,9 @@
entry.family = sk->sk_family;
- read_lock_bh(&tp->syn_wait_lock);
+ read_lock_bh(&tp->accept_queue.syn_wait_lock);
- lopt = tp->listen_opt;
+ lopt = tp->accept_queue.listen_opt;
if (!lopt || !lopt->qlen)
goto out;
@@ -588,7 +588,7 @@
}
out:
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
return err;
}
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 95528a7..1745dc8 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -874,7 +874,7 @@
__u16 rport,
__u32 raddr, __u32 laddr)
{
- struct tcp_listen_opt *lopt = tp->listen_opt;
+ struct tcp_listen_opt *lopt = tp->accept_queue.listen_opt;
struct request_sock *req, **prev;
for (prev = &lopt->syn_table[tcp_v4_synq_hash(raddr, rport, lopt->hash_rnd)];
@@ -898,18 +898,10 @@
static void tcp_v4_synq_add(struct sock *sk, struct request_sock *req)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
+ struct tcp_listen_opt *lopt = tp->accept_queue.listen_opt;
u32 h = tcp_v4_synq_hash(inet_rsk(req)->rmt_addr, inet_rsk(req)->rmt_port, lopt->hash_rnd);
- req->expires = jiffies + TCP_TIMEOUT_INIT;
- req->retrans = 0;
- req->sk = NULL;
- req->dl_next = lopt->syn_table[h];
-
- write_lock(&tp->syn_wait_lock);
- lopt->syn_table[h] = req;
- write_unlock(&tp->syn_wait_lock);
-
+ reqsk_queue_hash_req(&tp->accept_queue, h, req, TCP_TIMEOUT_INIT);
tcp_synq_added(sk);
}
@@ -2167,17 +2159,17 @@
if (++st->sbucket >= TCP_SYNQ_HSIZE)
break;
get_req:
- req = tp->listen_opt->syn_table[st->sbucket];
+ req = tp->accept_queue.listen_opt->syn_table[st->sbucket];
}
sk = sk_next(st->syn_wait_sk);
st->state = TCP_SEQ_STATE_LISTENING;
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
} else {
tp = tcp_sk(sk);
- read_lock_bh(&tp->syn_wait_lock);
- if (tp->listen_opt && tp->listen_opt->qlen)
+ read_lock_bh(&tp->accept_queue.syn_wait_lock);
+ if (reqsk_queue_len(&tp->accept_queue))
goto start_req;
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
sk = sk_next(sk);
}
get_sk:
@@ -2187,8 +2179,8 @@
goto out;
}
tp = tcp_sk(sk);
- read_lock_bh(&tp->syn_wait_lock);
- if (tp->listen_opt && tp->listen_opt->qlen) {
+ read_lock_bh(&tp->accept_queue.syn_wait_lock);
+ if (reqsk_queue_len(&tp->accept_queue)) {
start_req:
st->uid = sock_i_uid(sk);
st->syn_wait_sk = sk;
@@ -2196,7 +2188,7 @@
st->sbucket = 0;
goto get_req;
}
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
}
if (++st->bucket < TCP_LHTABLE_SIZE) {
sk = sk_head(&tcp_listening_hash[st->bucket]);
@@ -2383,7 +2375,7 @@
case TCP_SEQ_STATE_OPENREQ:
if (v) {
struct tcp_sock *tp = tcp_sk(st->syn_wait_sk);
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
}
case TCP_SEQ_STATE_LISTENING:
if (v != SEQ_START_TOKEN)
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 0e6d525..b3943e7 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -790,10 +790,8 @@
newtp->probes_out = 0;
newtp->rx_opt.num_sacks = 0;
newtp->urg_data = 0;
- newtp->listen_opt = NULL;
- newtp->accept_queue = newtp->accept_queue_tail = NULL;
- /* Deinitialize syn_wait_lock to trap illegal accesses. */
- memset(&newtp->syn_wait_lock, 0, sizeof(newtp->syn_wait_lock));
+ /* 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;
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index f03efe5..d97d191 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -464,7 +464,7 @@
static void tcp_synack_timer(struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
+ struct tcp_listen_opt *lopt = tp->accept_queue.listen_opt;
int max_retries = tp->syn_retries ? : sysctl_tcp_synack_retries;
int thresh = max_retries;
unsigned long now = jiffies;
@@ -527,12 +527,8 @@
}
/* Drop this request */
- write_lock(&tp->syn_wait_lock);
- *reqp = req->dl_next;
- write_unlock(&tp->syn_wait_lock);
- lopt->qlen--;
- if (req->retrans == 0)
- lopt->qlen_young--;
+ tcp_synq_unlink(tp, req, reqp);
+ reqsk_queue_removed(&tp->accept_queue, req);
reqsk_free(req);
continue;
}
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 068cd4a..84091da 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -401,7 +401,7 @@
struct in6_addr *laddr,
int iif)
{
- struct tcp_listen_opt *lopt = tp->listen_opt;
+ struct tcp_listen_opt *lopt = tp->accept_queue.listen_opt;
struct request_sock *req, **prev;
for (prev = &lopt->syn_table[tcp_v6_synq_hash(raddr, rport, lopt->hash_rnd)];
@@ -1267,18 +1267,10 @@
static void tcp_v6_synq_add(struct sock *sk, struct request_sock *req)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
+ struct tcp_listen_opt *lopt = tp->accept_queue.listen_opt;
u32 h = tcp_v6_synq_hash(&tcp6_rsk(req)->rmt_addr, inet_rsk(req)->rmt_port, lopt->hash_rnd);
- req->sk = NULL;
- req->expires = jiffies + TCP_TIMEOUT_INIT;
- req->retrans = 0;
- req->dl_next = lopt->syn_table[h];
-
- write_lock(&tp->syn_wait_lock);
- lopt->syn_table[h] = req;
- write_unlock(&tp->syn_wait_lock);
-
+ reqsk_queue_hash_req(&tp->accept_queue, h, req, TCP_TIMEOUT_INIT);
tcp_synq_added(sk);
}
