aboutsummaryrefslogtreecommitdiff
path: root/net/ceph/messenger.c
diff options
context:
space:
mode:
Diffstat (limited to 'net/ceph/messenger.c')
-rw-r--r--net/ceph/messenger.c106
1 files changed, 78 insertions, 28 deletions
diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index 1679f47280e2..525f454f7531 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -6,6 +6,7 @@
#include <linux/inet.h>
#include <linux/kthread.h>
#include <linux/net.h>
+#include <linux/nsproxy.h>
#include <linux/slab.h>
#include <linux/socket.h>
#include <linux/string.h>
@@ -162,6 +163,7 @@ static struct kmem_cache *ceph_msg_data_cache;
static char tag_msg = CEPH_MSGR_TAG_MSG;
static char tag_ack = CEPH_MSGR_TAG_ACK;
static char tag_keepalive = CEPH_MSGR_TAG_KEEPALIVE;
+static char tag_keepalive2 = CEPH_MSGR_TAG_KEEPALIVE2;
#ifdef CONFIG_LOCKDEP
static struct lock_class_key socket_class;
@@ -175,7 +177,7 @@ static struct lock_class_key socket_class;
static void queue_con(struct ceph_connection *con);
static void cancel_con(struct ceph_connection *con);
-static void con_work(struct work_struct *);
+static void ceph_con_workfn(struct work_struct *);
static void con_fault(struct ceph_connection *con);
/*
@@ -275,22 +277,22 @@ static void _ceph_msgr_exit(void)
ceph_msgr_wq = NULL;
}
- ceph_msgr_slab_exit();
-
BUG_ON(zero_page == NULL);
page_cache_release(zero_page);
zero_page = NULL;
+
+ ceph_msgr_slab_exit();
}
int ceph_msgr_init(void)
{
+ if (ceph_msgr_slab_init())
+ return -ENOMEM;
+
BUG_ON(zero_page != NULL);
zero_page = ZERO_PAGE(0);
page_cache_get(zero_page);
- if (ceph_msgr_slab_init())
- return -ENOMEM;
-
/*
* The number of active work items is limited by the number of
* connections, so leave @max_active at default.
@@ -479,7 +481,7 @@ static int ceph_tcp_connect(struct ceph_connection *con)
int ret;
BUG_ON(con->sock);
- ret = sock_create_kern(&init_net, con->peer_addr.in_addr.ss_family,
+ ret = sock_create_kern(read_pnet(&con->msgr->net), paddr->ss_family,
SOCK_STREAM, IPPROTO_TCP, &sock);
if (ret)
return ret;
@@ -748,7 +750,7 @@ void ceph_con_init(struct ceph_connection *con, void *private,
mutex_init(&con->mutex);
INIT_LIST_HEAD(&con->out_queue);
INIT_LIST_HEAD(&con->out_sent);
- INIT_DELAYED_WORK(&con->work, con_work);
+ INIT_DELAYED_WORK(&con->work, ceph_con_workfn);
con->state = CON_STATE_CLOSED;
}
@@ -1350,7 +1352,15 @@ static void prepare_write_keepalive(struct ceph_connection *con)
{
dout("prepare_write_keepalive %p\n", con);
con_out_kvec_reset(con);
- con_out_kvec_add(con, sizeof (tag_keepalive), &tag_keepalive);
+ if (con->peer_features & CEPH_FEATURE_MSGR_KEEPALIVE2) {
+ struct timespec ts = CURRENT_TIME;
+ struct ceph_timespec ceph_ts;
+ ceph_encode_timespec(&ceph_ts, &ts);
+ con_out_kvec_add(con, sizeof(tag_keepalive2), &tag_keepalive2);
+ con_out_kvec_add(con, sizeof(ceph_ts), &ceph_ts);
+ } else {
+ con_out_kvec_add(con, sizeof(tag_keepalive), &tag_keepalive);
+ }
con_flag_set(con, CON_FLAG_WRITE_PENDING);
}
@@ -1624,6 +1634,12 @@ static void prepare_read_tag(struct ceph_connection *con)
con->in_tag = CEPH_MSGR_TAG_READY;
}
+static void prepare_read_keepalive_ack(struct ceph_connection *con)
+{
+ dout("prepare_read_keepalive_ack %p\n", con);
+ con->in_base_pos = 0;
+}
+
/*
* Prepare to read a message.
*/
@@ -1731,17 +1747,17 @@ static int verify_hello(struct ceph_connection *con)
static bool addr_is_blank(struct sockaddr_storage *ss)
{
+ struct in_addr *addr = &((struct sockaddr_in *)ss)->sin_addr;
+ struct in6_addr *addr6 = &((struct sockaddr_in6 *)ss)->sin6_addr;
+
switch (ss->ss_family) {
case AF_INET:
- return ((struct sockaddr_in *)ss)->sin_addr.s_addr == 0;
+ return addr->s_addr == htonl(INADDR_ANY);
case AF_INET6:
- return
- ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[0] == 0 &&
- ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[1] == 0 &&
- ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[2] == 0 &&
- ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[3] == 0;
+ return ipv6_addr_any(addr6);
+ default:
+ return true;
}
- return false;
}
static int addr_port(struct sockaddr_storage *ss)
@@ -2321,13 +2337,6 @@ static int read_partial_message(struct ceph_connection *con)
return ret;
BUG_ON(!con->in_msg ^ skip);
- if (con->in_msg && data_len > con->in_msg->data_length) {
- pr_warn("%s skipping long message (%u > %zd)\n",
- __func__, data_len, con->in_msg->data_length);
- ceph_msg_put(con->in_msg);
- con->in_msg = NULL;
- skip = 1;
- }
if (skip) {
/* skip this message */
dout("alloc_msg said skip message\n");
@@ -2456,6 +2465,17 @@ static void process_message(struct ceph_connection *con)
mutex_lock(&con->mutex);
}
+static int read_keepalive_ack(struct ceph_connection *con)
+{
+ struct ceph_timespec ceph_ts;
+ size_t size = sizeof(ceph_ts);
+ int ret = read_partial(con, size, size, &ceph_ts);
+ if (ret <= 0)
+ return ret;
+ ceph_decode_timespec(&con->last_keepalive_ack, &ceph_ts);
+ prepare_read_tag(con);
+ return 1;
+}
/*
* Write something to the socket. Called in a worker thread when the
@@ -2525,6 +2545,10 @@ more_kvec:
do_next:
if (con->state == CON_STATE_OPEN) {
+ if (con_flag_test_and_clear(con, CON_FLAG_KEEPALIVE_PENDING)) {
+ prepare_write_keepalive(con);
+ goto more;
+ }
/* is anything else pending? */
if (!list_empty(&con->out_queue)) {
prepare_write_message(con);
@@ -2534,10 +2558,6 @@ do_next:
prepare_write_ack(con);
goto more;
}
- if (con_flag_test_and_clear(con, CON_FLAG_KEEPALIVE_PENDING)) {
- prepare_write_keepalive(con);
- goto more;
- }
}
/* Nothing to do! */
@@ -2640,6 +2660,9 @@ more:
case CEPH_MSGR_TAG_ACK:
prepare_read_ack(con);
break;
+ case CEPH_MSGR_TAG_KEEPALIVE2_ACK:
+ prepare_read_keepalive_ack(con);
+ break;
case CEPH_MSGR_TAG_CLOSE:
con_close_socket(con);
con->state = CON_STATE_CLOSED;
@@ -2683,6 +2706,12 @@ more:
process_ack(con);
goto more;
}
+ if (con->in_tag == CEPH_MSGR_TAG_KEEPALIVE2_ACK) {
+ ret = read_keepalive_ack(con);
+ if (ret <= 0)
+ goto out;
+ goto more;
+ }
out:
dout("try_read done on %p ret %d\n", con, ret);
@@ -2798,7 +2827,7 @@ static void con_fault_finish(struct ceph_connection *con)
/*
* Do some work on a connection. Drop a connection ref when we're done.
*/
-static void con_work(struct work_struct *work)
+static void ceph_con_workfn(struct work_struct *work)
{
struct ceph_connection *con = container_of(work, struct ceph_connection,
work.work);
@@ -2944,11 +2973,18 @@ void ceph_messenger_init(struct ceph_messenger *msgr,
msgr->tcp_nodelay = tcp_nodelay;
atomic_set(&msgr->stopping, 0);
+ write_pnet(&msgr->net, get_net(current->nsproxy->net_ns));
dout("%s %p\n", __func__, msgr);
}
EXPORT_SYMBOL(ceph_messenger_init);
+void ceph_messenger_fini(struct ceph_messenger *msgr)
+{
+ put_net(read_pnet(&msgr->net));
+}
+EXPORT_SYMBOL(ceph_messenger_fini);
+
static void clear_standby(struct ceph_connection *con)
{
/* come back from STANDBY? */
@@ -3093,6 +3129,20 @@ void ceph_con_keepalive(struct ceph_connection *con)
}
EXPORT_SYMBOL(ceph_con_keepalive);
+bool ceph_con_keepalive_expired(struct ceph_connection *con,
+ unsigned long interval)
+{
+ if (interval > 0 &&
+ (con->peer_features & CEPH_FEATURE_MSGR_KEEPALIVE2)) {
+ struct timespec now = CURRENT_TIME;
+ struct timespec ts;
+ jiffies_to_timespec(interval, &ts);
+ ts = timespec_add(con->last_keepalive_ack, ts);
+ return timespec_compare(&now, &ts) >= 0;
+ }
+ return false;
+}
+
static struct ceph_msg_data *ceph_msg_data_create(enum ceph_msg_data_type type)
{
struct ceph_msg_data *data;
Powered by cgit, Gruvboxed by Blaster4385.