continued messenger-related changes

In ceph_osdc_release_request(), a reference to the r_reply message is dropped. But just after that, that same message is revoked if it was in use to receive an incoming reply. Reorder these so we are sure we hold a reference until we're actually done with the message.

Signed-off-by Alex Elder <[email protected]>

---
 net/ceph/osd_client.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

net/ceph/osd_client.c +2 -2

index e30efbc..d8b6d31 100644
@@ -139,8 +139,6 @@ void ceph_osdc_release_request(struct kref *kref)
 
 	if (req->r_request)
 		ceph_msg_put(req->r_request);
-	if (req->r_reply)
-		ceph_msg_put(req->r_reply);
 	if (req->r_con_filling_msg) {
 		dout("release_request revoking pages %p from con %p\n",
 		     req->r_pages, req->r_con_filling_msg);

@@ -148,6 +146,8 @@ void ceph_osdc_release_request(struct kref *kref)
 				      req->r_reply);
 		ceph_con_put(req->r_con_filling_msg);
 	}
+	if (req->r_reply)
+		ceph_msg_put(req->r_reply);
 	if (req->r_own_pages)
 		ceph_release_page_vector(req->r_pages,
 					 req->r_num_pages);

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

A monitor client has a pointer to a ceph connection structure in it. This is the only one of the three ceph client types that do it this way; the OSD and MDS clients embed the connection into their main structures. There is always exactly one ceph connection for a monitor client, so there is no need to allocate it separate from the monitor client structure.

So switch the ceph_mon_client structure to embed its ceph_connection structure.

Signed-off-by Alex Elder <[email protected]>

---
 include/linux/ceph/mon_client.h |    2 +-
 net/ceph/mon_client.c           |   47
++++++++++++++++----------------------
 2 files changed, 21 insertions(+), 28 deletions(-)

include/linux/ceph/mon_client.h +1 -1

index 545f859..2113e38 100644
@@ -70,7 +70,7 @@ struct ceph_mon_client {
 	bool hunting;
 	int cur_mon;                       /* last monitor i contacted */
 	unsigned long sub_sent, sub_renew_after;
-	struct ceph_connection *con;
+	struct ceph_connection con;
 	bool have_fsid;
 
 	/* pending generic requests */

net/ceph/mon_client.c +2 -2

index 704dc95..ac4d6b1 100644
@@ -106,9 +106,9 @@ static void __send_prepared_auth_request(struct
 ceph_mon_client *monc, int len)
 	monc->pending_auth = 1;
 	monc->m_auth->front.iov_len = len;
 	monc->m_auth->hdr.front_len = cpu_to_le32(len);
-	ceph_con_revoke(monc->con, monc->m_auth);
+	ceph_con_revoke(&monc->con, monc->m_auth);
 	ceph_msg_get(monc->m_auth);  /* keep our ref */
-	ceph_con_send(monc->con, monc->m_auth);
+	ceph_con_send(&monc->con, monc->m_auth);
 }
 

/*

net/ceph/mon_client.c +2 -2

@@ -117,8 +117,8 @@ static void __send_prepared_auth_request(struct
 ceph_mon_client *monc, int len)
 static void __close_session(struct ceph_mon_client *monc)
 {
 	dout("__close_session closing mon%d\n", monc->cur_mon);
-	ceph_con_revoke(monc->con, monc->m_auth);
-	ceph_con_close(monc->con);
+	ceph_con_revoke(&monc->con, monc->m_auth);
+	ceph_con_close(&monc->con);
 	monc->cur_mon = -1;
 	monc->pending_auth = 0;

net/ceph/mon_client.c +5 -5

 	ceph_auth_reset(monc->auth);

@@ -142,9 +142,9 @@ static int __open_session(struct ceph_mon_client *monc)
 		monc->want_next_osdmap = !!monc->want_next_osdmap;
 
 		dout("open_session mon%d opening\n", monc->cur_mon);
-		monc->con->peer_name.type = CEPH_ENTITY_TYPE_MON;
-		monc->con->peer_name.num = cpu_to_le64(monc->cur_mon);
-		ceph_con_open(monc->con,
+		monc->con.peer_name.type = CEPH_ENTITY_TYPE_MON;
+		monc->con.peer_name.num = cpu_to_le64(monc->cur_mon);
+		ceph_con_open(&monc->con,
 			      &monc->monmap->mon_inst[monc->cur_mon].addr);
 
 		/* initiatiate authentication handshake */

@@ -226,8 +226,8 @@ static void __send_subscribe(struct ceph_mon_client
 *monc)
 
 		msg->front.iov_len = p - msg->front.iov_base;
 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
-		ceph_con_revoke(monc->con, msg);
-		ceph_con_send(monc->con, ceph_msg_get(msg));
+		ceph_con_revoke(&monc->con, msg);
+		ceph_con_send(&monc->con, ceph_msg_get(msg));
 
 		monc->sub_sent = jiffies | 1;  /* never 0 */

net/ceph/mon_client.c +1 -1

 	}

@@ -247,7 +247,7 @@ static void handle_subscribe_ack(struct
 ceph_mon_client *monc,
 	if (monc->hunting) {
 		pr_info("mon%d %s session established\n",
 			monc->cur_mon,
-			ceph_pr_addr(&monc->con->peer_addr.in_addr));
+			ceph_pr_addr(&monc->con.peer_addr.in_addr));
 		monc->hunting = false;
 	}

net/ceph/mon_client.c +1 -1

 	dout("handle_subscribe_ack after %d seconds\n", seconds);

@@ -461,7 +461,7 @@ static int do_generic_request(struct ceph_mon_client
 *monc,
 	req->request->hdr.tid = cpu_to_le64(req->tid);
 	__insert_generic_request(monc, req);
 	monc->num_generic_requests++;
-	ceph_con_send(monc->con, ceph_msg_get(req->request));
+	ceph_con_send(&monc->con, ceph_msg_get(req->request));
 	mutex_unlock(&monc->mutex);
 

net/ceph/mon_client.c +7 -10

 	err = wait_for_completion_interruptible(&req->completion);

@@ -684,8 +684,8 @@ static void __resend_generic_request(struct
 ceph_mon_client *monc)
 
 	for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) {
 		req = rb_entry(p, struct ceph_mon_generic_request, node);
-		ceph_con_revoke(monc->con, req->request);
-		ceph_con_send(monc->con, ceph_msg_get(req->request));
+		ceph_con_revoke(&monc->con, req->request);
+		ceph_con_send(&monc->con, ceph_msg_get(req->request));
 	}
 }

@@ -705,7 +705,7 @@ static void delayed_work(struct work_struct *work)
 		__close_session(monc);
 		__open_session(monc);  /* continue hunting */
 	} else {
-		ceph_con_keepalive(monc->con);
+		ceph_con_keepalive(&monc->con);
 
 		__validate_auth(monc);
 

@@ -760,19 +760,16 @@ int ceph_monc_init(struct ceph_mon_client *monc,
 struct ceph_client *cl)
 		goto out;
 
 	/* connection */
-	monc->con = kmalloc(sizeof(*monc->con), GFP_KERNEL);
-	if (!monc->con)
-		goto out_monmap;
-	ceph_con_init(&monc->client->msgr, monc->con);
-	monc->con->private = monc;
-	monc->con->ops = &mon_con_ops;
+	ceph_con_init(&monc->client->msgr, &monc->con);
+	monc->con.private = monc;
+	monc->con.ops = &mon_con_ops;
 
 	/* authentication */
 	monc->auth = ceph_auth_init(cl->options->name,
 				    cl->options->key);
 	if (IS_ERR(monc->auth)) {
 		err = PTR_ERR(monc->auth);
-		goto out_con;
+		goto out_monmap;
 	}
 	monc->auth->want_keys =

net/ceph/mon_client.c +2 -6

 		CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON |

@@ -824,8 +821,6 @@ out_subscribe_ack:
 	ceph_msg_put(monc->m_subscribe_ack);
 out_auth:
 	ceph_auth_destroy(monc->auth);
-out_con:
-	monc->con->ops->put(monc->con);
 out_monmap:
 	kfree(monc->monmap);
 out:

@@ -841,9 +836,7 @@ void ceph_monc_stop(struct ceph_mon_client *monc)
 	mutex_lock(&monc->mutex);
 	__close_session(monc);
 
-	monc->con->private = NULL;
-	monc->con->ops->put(monc->con);
-	monc->con = NULL;
+	monc->con.private = NULL;
 
 	mutex_unlock(&monc->mutex);
 

@@ -1021,7 +1014,7 @@ static void mon_fault(struct ceph_connection *con)
 	if (!monc->hunting)
 		pr_info("mon%d %s session lost, "
 			"hunting for new mon\n", monc->cur_mon,
-			ceph_pr_addr(&monc->con->peer_addr.in_addr));
+			ceph_pr_addr(&monc->con.peer_addr.in_addr));
 
 	__close_session(monc);
 	if (!monc->hunting) {

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

A monitor client has a pointer to a ceph connection structure in it. This is the only one of the three ceph client types that do it this way; the OSD and MDS clients embed the connection into their main structures. There is always exactly one ceph connection for a monitor client, so there is no need to allocate it separate from the monitor client structure.

So switch the ceph_mon_client structure to embed its ceph_connection structure.

Signed-off-by Alex Elder <[email protected]>

---
 include/linux/ceph/mon_client.h |    2 +-
 net/ceph/mon_client.c           |   47
++++++++++++++++----------------------
 2 files changed, 21 insertions(+), 28 deletions(-)

include/linux/ceph/mon_client.h +1 -1

index 545f859..2113e38 100644
@@ -70,7 +70,7 @@ struct ceph_mon_client {
 	bool hunting;
 	int cur_mon;                       /* last monitor i contacted */
 	unsigned long sub_sent, sub_renew_after;
-	struct ceph_connection *con;
+	struct ceph_connection con;
 	bool have_fsid;
 
 	/* pending generic requests */

net/ceph/mon_client.c +2 -2

index 704dc95..ac4d6b1 100644
@@ -106,9 +106,9 @@ static void __send_prepared_auth_request(struct
 ceph_mon_client *monc, int len)
 	monc->pending_auth = 1;
 	monc->m_auth->front.iov_len = len;
 	monc->m_auth->hdr.front_len = cpu_to_le32(len);
-	ceph_con_revoke(monc->con, monc->m_auth);
+	ceph_con_revoke(&monc->con, monc->m_auth);
 	ceph_msg_get(monc->m_auth);  /* keep our ref */
-	ceph_con_send(monc->con, monc->m_auth);
+	ceph_con_send(&monc->con, monc->m_auth);
 }
 

/*

net/ceph/mon_client.c +2 -2

@@ -117,8 +117,8 @@ static void __send_prepared_auth_request(struct
 ceph_mon_client *monc, int len)
 static void __close_session(struct ceph_mon_client *monc)
 {
 	dout("__close_session closing mon%d\n", monc->cur_mon);
-	ceph_con_revoke(monc->con, monc->m_auth);
-	ceph_con_close(monc->con);
+	ceph_con_revoke(&monc->con, monc->m_auth);
+	ceph_con_close(&monc->con);
 	monc->cur_mon = -1;
 	monc->pending_auth = 0;

net/ceph/mon_client.c +5 -5

 	ceph_auth_reset(monc->auth);

@@ -142,9 +142,9 @@ static int __open_session(struct ceph_mon_client *monc)
 		monc->want_next_osdmap = !!monc->want_next_osdmap;
 
 		dout("open_session mon%d opening\n", monc->cur_mon);
-		monc->con->peer_name.type = CEPH_ENTITY_TYPE_MON;
-		monc->con->peer_name.num = cpu_to_le64(monc->cur_mon);
-		ceph_con_open(monc->con,
+		monc->con.peer_name.type = CEPH_ENTITY_TYPE_MON;
+		monc->con.peer_name.num = cpu_to_le64(monc->cur_mon);
+		ceph_con_open(&monc->con,
 			      &monc->monmap->mon_inst[monc->cur_mon].addr);
 
 		/* initiatiate authentication handshake */

@@ -226,8 +226,8 @@ static void __send_subscribe(struct ceph_mon_client
 *monc)
 
 		msg->front.iov_len = p - msg->front.iov_base;
 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
-		ceph_con_revoke(monc->con, msg);
-		ceph_con_send(monc->con, ceph_msg_get(msg));
+		ceph_con_revoke(&monc->con, msg);
+		ceph_con_send(&monc->con, ceph_msg_get(msg));
 
 		monc->sub_sent = jiffies | 1;  /* never 0 */

net/ceph/mon_client.c +1 -1

 	}

@@ -247,7 +247,7 @@ static void handle_subscribe_ack(struct
 ceph_mon_client *monc,
 	if (monc->hunting) {
 		pr_info("mon%d %s session established\n",
 			monc->cur_mon,
-			ceph_pr_addr(&monc->con->peer_addr.in_addr));
+			ceph_pr_addr(&monc->con.peer_addr.in_addr));
 		monc->hunting = false;
 	}

net/ceph/mon_client.c +1 -1

 	dout("handle_subscribe_ack after %d seconds\n", seconds);

@@ -461,7 +461,7 @@ static int do_generic_request(struct ceph_mon_client
 *monc,
 	req->request->hdr.tid = cpu_to_le64(req->tid);
 	__insert_generic_request(monc, req);
 	monc->num_generic_requests++;
-	ceph_con_send(monc->con, ceph_msg_get(req->request));
+	ceph_con_send(&monc->con, ceph_msg_get(req->request));
 	mutex_unlock(&monc->mutex);
 

net/ceph/mon_client.c +7 -10

 	err = wait_for_completion_interruptible(&req->completion);

@@ -684,8 +684,8 @@ static void __resend_generic_request(struct
 ceph_mon_client *monc)
 
 	for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) {
 		req = rb_entry(p, struct ceph_mon_generic_request, node);
-		ceph_con_revoke(monc->con, req->request);
-		ceph_con_send(monc->con, ceph_msg_get(req->request));
+		ceph_con_revoke(&monc->con, req->request);
+		ceph_con_send(&monc->con, ceph_msg_get(req->request));
 	}
 }

@@ -705,7 +705,7 @@ static void delayed_work(struct work_struct *work)
 		__close_session(monc);
 		__open_session(monc);  /* continue hunting */
 	} else {
-		ceph_con_keepalive(monc->con);
+		ceph_con_keepalive(&monc->con);
 
 		__validate_auth(monc);
 

@@ -760,19 +760,16 @@ int ceph_monc_init(struct ceph_mon_client *monc,
 struct ceph_client *cl)
 		goto out;
 
 	/* connection */
-	monc->con = kmalloc(sizeof(*monc->con), GFP_KERNEL);
-	if (!monc->con)
-		goto out_monmap;
-	ceph_con_init(&monc->client->msgr, monc->con);
-	monc->con->private = monc;
-	monc->con->ops = &mon_con_ops;
+	ceph_con_init(&monc->client->msgr, &monc->con);
+	monc->con.private = monc;
+	monc->con.ops = &mon_con_ops;
 
 	/* authentication */
 	monc->auth = ceph_auth_init(cl->options->name,
 				    cl->options->key);
 	if (IS_ERR(monc->auth)) {
 		err = PTR_ERR(monc->auth);
-		goto out_con;
+		goto out_monmap;
 	}
 	monc->auth->want_keys =

net/ceph/mon_client.c +2 -6

 		CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON |

@@ -824,8 +821,6 @@ out_subscribe_ack:
 	ceph_msg_put(monc->m_subscribe_ack);
 out_auth:
 	ceph_auth_destroy(monc->auth);
-out_con:
-	monc->con->ops->put(monc->con);
 out_monmap:
 	kfree(monc->monmap);
 out:

@@ -841,9 +836,7 @@ void ceph_monc_stop(struct ceph_mon_client *monc)
 	mutex_lock(&monc->mutex);
 	__close_session(monc);
 
-	monc->con->private = NULL;
-	monc->con->ops->put(monc->con);
-	monc->con = NULL;
+	monc->con.private = NULL;
 
 	mutex_unlock(&monc->mutex);
 

@@ -1021,7 +1014,7 @@ static void mon_fault(struct ceph_connection *con)
 	if (!monc->hunting)
 		pr_info("mon%d %s session lost, "
 			"hunting for new mon\n", monc->cur_mon,
-			ceph_pr_addr(&monc->con->peer_addr.in_addr));
+			ceph_pr_addr(&monc->con.peer_addr.in_addr));
 
 	__close_session(monc);
 	if (!monc->hunting) {

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

The function ceph_alloc_msg() is only used to allocate a message that will be assigned to a connection's in_msg pointer. Rename the function so this implied usage is more clear.

In addition, make that assignment inside the function (again, since that's precisely what it's intended to be used for). This allows us to return what is now provided via the passed-in address of a "skip" variable. The return type is now Boolean to be explicit that there are only two possible outcomes.

Signed-off-by Alex Elder <[email protected]>

---
 net/ceph/messenger.c |   58
++++++++++++++++++++++++++-----------------------
 1 files changed, 31 insertions(+), 27 deletions(-)

net/ceph/messenger.c +2 -3

index 3b65f6e..f7c9061 100644
@@ -1655,9 +1655,8 @@ static int read_partial_message_section(struct
 ceph_connection *con,
 	return 1;
 }
 
-static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
-				struct ceph_msg_header *hdr,
-				int *skip);
+static bool ceph_con_in_msg_alloc(struct ceph_connection *con,
+				struct ceph_msg_header *hdr);
 
 

static int read_partial_message_pages(struct ceph_connection *con,

net/ceph/messenger.c +1 -4

@@ -1740,7 +1739,6 @@ static int read_partial_message(struct
 ceph_connection *con)
 	int ret;
 	unsigned front_len, middle_len, data_len;
 	bool do_datacrc = !con->msgr->nocrc;
-	int skip;
 	u64 seq;
 	u32 crc;

@@ -1793,9 +1791,7 @@ static int read_partial_message(struct
 ceph_connection *con)
 	if (!con->in_msg) {
 		dout("got hdr type %d front %d data %d\n", con->in_hdr.type,
 		     con->in_hdr.front_len, con->in_hdr.data_len);
-		skip = 0;
-		con->in_msg = ceph_alloc_msg(con, &con->in_hdr, &skip);
-		if (skip) {
+		if (ceph_con_in_msg_alloc(con, &con->in_hdr)) {
 			/* skip this message */
 			dout("alloc_msg said skip message\n");

net/ceph/messenger.c +0 -0

 			BUG_ON(con->in_msg);

net/ceph/messenger.c +4 -4

  */
-static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
-				struct ceph_msg_header *hdr,
-				int *skip)
+static bool ceph_con_in_msg_alloc(struct ceph_connection *con,
+				struct ceph_msg_header *hdr)
 {
 	int type = le16_to_cpu(hdr->type);
 	int front_len = le32_to_cpu(hdr->front_len);
 	int middle_len = le32_to_cpu(hdr->middle_len);
-	struct ceph_msg *msg = NULL;
 	int ret;
 
+	BUG_ON(con->in_msg != NULL);
+
 	if (con->ops->alloc_msg) {

net/ceph/messenger.c +11 -12

 	}
-	if (!msg) {
-		*skip = 0;
-		msg = ceph_msg_new(type, front_len, GFP_NOFS, false);
-		if (!msg) {
+	if (!con->in_msg) {
+		con->in_msg = ceph_msg_new(type, front_len, GFP_NOFS, false);
+		if (!con->in_msg) {
 			pr_err("unable to allocate msg type %d len %d\n",
 			       type, front_len);
-			return NULL;
+			return false;
 		}
-		msg->page_alignment = le16_to_cpu(hdr->data_off);
+		con->in_msg->page_alignment = le16_to_cpu(hdr->data_off);
 	}
-	memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr));
+	memcpy(&con->in_msg->hdr, &con->in_hdr, sizeof(con->in_hdr));
 
-	if (middle_len && !msg->middle) {
-		ret = ceph_alloc_middle(con, msg);
+	if (middle_len && !con->in_msg->middle) {
+		ret = ceph_alloc_middle(con, con->in_msg);
 		if (ret < 0) {
-			ceph_msg_put(msg);
-			return NULL;
+			ceph_msg_put(con->in_msg);
+			con->in_msg = NULL;
 		}
 	}
 
-	return msg;
+	return false;
 }
 

Signature · 7 lines

-- 
1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

When a ceph message is queued for sending it is placed on a list of pending messages (ceph_connection->out_queue). When they are actually sent over the wire, they are moved from that list to another (ceph_connection->out_sent). When acknowledgement for the message is received, it is removed from the sent messages list.

During that entire time the message is "in the possession" of a single ceph connection. Keep track of that connection in the message. This will be used in the next patch (and is a helpful bit of information for debugging anyway).

Signed-off-by Alex Elder <[email protected]>

---
 include/linux/ceph/messenger.h |    3 +++
 net/ceph/messenger.c           |   27 +++++++++++++++++++++++++--
 2 files changed, 28 insertions(+), 2 deletions(-)

include/linux/ceph/messenger.h +3 -0

index dd27837..6df837f 100644
@@ -77,7 +77,10 @@ struct ceph_msg {
 	unsigned nr_pages;              /* size of page array */
 	unsigned page_alignment;        /* io offset in first page */
 	struct ceph_pagelist *pagelist; /* instead of pages */
+
+	struct ceph_connection *con;
 	struct list_head list_head;
+
 	struct kref kref;
 	struct bio  *bio;		/* instead of pages/pagelist */
 	struct bio  *bio_iter;		/* bio iterator */

net/ceph/messenger.c +5 -0

index f7c9061..59fa5fb 100644
@@ -414,6 +414,9 @@ static int con_close_socket(struct ceph_connection *con)
 static void ceph_msg_remove(struct ceph_msg *msg)
 {
 	list_del_init(&msg->list_head);
+	BUG_ON(msg->con == NULL);
+	msg->con = NULL;
+
 	ceph_msg_put(msg);
 }
 static void ceph_msg_remove_list(struct list_head *head)

@@ -433,6 +436,8 @@ static void reset_connection(struct ceph_connection
 *con)
 	ceph_msg_remove_list(&con->out_sent);
 
 	if (con->in_msg) {
+		BUG_ON(con->in_msg->con != con);
+		con->in_msg->con = NULL;
 		ceph_msg_put(con->in_msg);
 		con->in_msg = NULL;

net/ceph/messenger.c +2 -0

 	}

@@ -625,8 +630,10 @@ static void prepare_write_message(struct
 ceph_connection *con)
 			&con->out_temp_ack);
 	}
 
+	BUG_ON(list_empty(&con->out_queue));
 	m = list_first_entry(&con->out_queue, struct ceph_msg, list_head);
 	con->out_msg = m;
+	BUG_ON(m->con != con);
 
 	/* put message on sent list */

net/ceph/messenger.c +2 -0

 	ceph_msg_get(m);

@@ -1806,6 +1813,8 @@ static int read_partial_message(struct
 ceph_connection *con)
 				"error allocating memory for incoming message";
 			return -ENOMEM;
 		}
+
+		BUG_ON(con->in_msg->con != con);
 		m = con->in_msg;
 		m->front.iov_len = 0;    /* haven't read it yet */

net/ceph/messenger.c +6 -0

 		if (m->middle)

@@ -1901,6 +1910,8 @@ static void process_message(struct ceph_connection
 *con)
 {
 	struct ceph_msg *msg;
 
+	BUG_ON(con->in_msg->con != con);
+	con->in_msg->con = NULL;
 	msg = con->in_msg;
 	con->in_msg = NULL;

@@ -2260,6 +2271,8 @@ static void ceph_fault(struct ceph_connection *con)
 	con_close_socket(con);
 
 	if (con->in_msg) {
+		BUG_ON(con->in_msg->con != con);
+		con->in_msg->con = NULL;
 		ceph_msg_put(con->in_msg);
 		con->in_msg = NULL;
 	}

@@ -2378,6 +2391,8 @@ void ceph_con_send(struct ceph_connection *con,
 struct ceph_msg *msg)
 
 	/* queue */
 	mutex_lock(&con->mutex);
+	BUG_ON(msg->con != NULL);
+	msg->con = con;
 	BUG_ON(!list_empty(&msg->list_head));
 	list_add_tail(&msg->list_head, &con->out_queue);

net/ceph/messenger.c +5 -2

 	dout("----- %p to %s%lld %d=%s len %d+%d+%d -----\n", msg,

@@ -2403,13 +2418,16 @@ void ceph_con_revoke(struct ceph_connection
 *con, struct ceph_msg *msg)
 {
 	mutex_lock(&con->mutex);
 	if (!list_empty(&msg->list_head)) {
-		dout("con_revoke %p msg %p - was on queue\n", con, msg);
+		dout("%s %p msg %p - was on queue\n", __func__, con, msg);
 		list_del_init(&msg->list_head);
+		BUG_ON(msg->con == NULL);
+		msg->con = NULL;
+
 		ceph_msg_put(msg);
 		msg->hdr.seq = 0;
 	}
 	if (con->out_msg == msg) {
-		dout("con_revoke %p msg %p - was sending\n", con, msg);
+		dout("%s %p msg %p - was sending\n", __func__, con, msg);
 		con->out_msg = NULL;
 		if (con->out_kvec_is_msg) {

net/ceph/messenger.c +2 -0

 			con->out_skip = con->out_kvec_bytes;

@@ -2478,6 +2496,8 @@ struct ceph_msg *ceph_msg_new(int type, int
 front_len, gfp_t flags,
 	if (m == NULL)
 		goto out;
 	kref_init(&m->kref);
+
+	m->con = NULL;
 	INIT_LIST_HEAD(&m->list_head);
 

net/ceph/messenger.c +3 -0

 	m->hdr.tid = 0;

@@ -2595,6 +2615,8 @@ static bool ceph_con_in_msg_alloc(struct
 ceph_connection *con,
 		mutex_unlock(&con->mutex);
 		con->in_msg = con->ops->alloc_msg(con, hdr, &skip);
 		mutex_lock(&con->mutex);
+		if (con->in_msg)
+			con->in_msg->con = con;
 		if (skip)
 			con->in_msg = NULL;

@@ -2608,6 +2630,7 @@ static bool ceph_con_in_msg_alloc(struct
 ceph_connection *con,
 			       type, front_len);
 			return false;
 		}
+		con->in_msg->con = con;
 		con->in_msg->page_alignment = le16_to_cpu(hdr->data_off);
 	}

memcpy(&con->in_msg->hdr, &con->in_hdr, sizeof(con->in_hdr));

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

There are essentially two types of ceph messages: incoming and outgoing. Outgoing messages are always allocated via ceph_msg_new(), and at the time of their allocation they are not associated with any particular connection. Incoming messages are always allocated via ceph_con_in_msg_alloc(), and they are initially associated with the connection from which incoming data will be placed into the message.

When an outgoing message gets sent, it becomes associated with a connection and remains that way until the message is successfully sent. The association of an incoming message goes away at the point it is sent to an upper layer via a con->ops->dispatch method.

This patch implements reference counting for all ceph messages, such that every message holds a reference (and a pointer) to a connection if and only if it is associated with that connection (as described above).

For background, here is an explanation of the ceph message lifecycle, emphasizing when an association exists between a message and a connection.

Outgoing Messages
An outgoing message is "owned" by its allocator, from the time it is allocated in ceph_msg_new() up to the point it gets queued for sending in ceph_con_send(). Prior to that point the message's msg->con pointer is null; at the point it is queued for sending its message pointer is assigned to refer to the connection. At that time the message is inserted into a connection's out_queue list.

When a message on the out_queue list has been sent to the socket layer to be put on the wire, it is transferred out of that list and into the connection's out_sent list. At that point it is still owned by the connection, and will remain so until an acknowledgement is received from the recipient that indicates the message was successfully transferred. When such an acknowledgement is received (in process_ack()), the message is removed from its list (in ceph_msg_remove()), at which point it is no longer associated with the connection.

So basically, any time a message is on one of a connection's lists, it is associated with that connection. Reference counting outgoing messages can thus be done at the points a message is added to the out_queue (in ceph_con_send()) and the point it is removed from either its two lists (in ceph_msg_remove())--at which point its connection pointer becomes null.

Incoming Messages
When an incoming message on a connection is getting read (in read_partial_message()) and there is no message in con->in_msg, a new one is allocated using ceph_con_in_msg_alloc(). At that point the message is associated with the connection. Once that message has been completely and successfully read, it is passed to upper layer code using the connection's con->ops->dispatch method. At that point the association between the message and the connection no longer exists.

Reference counting of connections for incoming messages can be done by taking a reference to the connection when the message gets allocated, and releasing that reference when it gets handed off using the dispatch method.

We should never fail to get a connection reference for a message--the since the caller should already hold one.

Signed-off-by Alex Elder <[email protected]>

---
 net/ceph/messenger.c |   24 ++++++++++++++++++------
 1 files changed, 18 insertions(+), 6 deletions(-)

net/ceph/messenger.c +2 -0

index 59fa5fb..336707b 100644
@@ -415,6 +415,7 @@ static void ceph_msg_remove(struct ceph_msg *msg)
 {
 	list_del_init(&msg->list_head);
 	BUG_ON(msg->con == NULL);
+	ceph_con_put(msg->con);
 	msg->con = NULL;
 
 	ceph_msg_put(msg);

@@ -440,6 +441,7 @@ static void reset_connection(struct ceph_connection
 *con)
 		con->in_msg->con = NULL;
 		ceph_msg_put(con->in_msg);
 		con->in_msg = NULL;
+		ceph_con_put(con->in_msg->con);
 	}
 

net/ceph/messenger.c +1 -0

 	con->connect_seq = 0;

@@ -1914,6 +1916,7 @@ static void process_message(struct ceph_connection
 *con)
 	con->in_msg->con = NULL;
 	msg = con->in_msg;
 	con->in_msg = NULL;
+	ceph_con_put(con);
 
 	/* if first message, set peer_name */

net/ceph/messenger.c +5 -1

 	if (con->peer_name.type == 0)

@@ -2275,6 +2278,7 @@ static void ceph_fault(struct ceph_connection *con)
 		con->in_msg->con = NULL;
 		ceph_msg_put(con->in_msg);
 		con->in_msg = NULL;
+		ceph_con_put(con);
 	}
 
 	/* Requeue anything that hasn't been acked */

@@ -2391,8 +2395,11 @@ void ceph_con_send(struct ceph_connection *con,
 struct ceph_msg *msg)
 
 	/* queue */
 	mutex_lock(&con->mutex);
+
 	BUG_ON(msg->con != NULL);
-	msg->con = con;
+	msg->con = ceph_con_get(con);
+	BUG_ON(msg->con == NULL);
+
 	BUG_ON(!list_empty(&msg->list_head));
 	list_add_tail(&msg->list_head, &con->out_queue);

net/ceph/messenger.c +2 -1

 	dout("----- %p to %s%lld %d=%s len %d+%d+%d -----\n", msg,

@@ -2421,10 +2428,11 @@ void ceph_con_revoke(struct ceph_connection
 *con, struct ceph_msg *msg)
 		dout("%s %p msg %p - was on queue\n", __func__, con, msg);
 		list_del_init(&msg->list_head);
 		BUG_ON(msg->con == NULL);
+		ceph_con_put(msg->con);
 		msg->con = NULL;
+		msg->hdr.seq = 0;
 
 		ceph_msg_put(msg);
-		msg->hdr.seq = 0;
 	}
 	if (con->out_msg == msg) {

net/ceph/messenger.c +2 -1

 		dout("%s %p msg %p - was sending\n", __func__, con, msg);

@@ -2433,8 +2441,9 @@ void ceph_con_revoke(struct ceph_connection *con,
 struct ceph_msg *msg)
 			con->out_skip = con->out_kvec_bytes;
 			con->out_kvec_is_msg = false;
 		}
-		ceph_msg_put(msg);
 		msg->hdr.seq = 0;
+
+		ceph_msg_put(msg);
 	}
 	mutex_unlock(&con->mutex);

}

net/ceph/messenger.c +6 -3

@@ -2615,8 +2624,10 @@ static bool ceph_con_in_msg_alloc(struct
 ceph_connection *con,
 		mutex_unlock(&con->mutex);
 		con->in_msg = con->ops->alloc_msg(con, hdr, &skip);
 		mutex_lock(&con->mutex);
-		if (con->in_msg)
-			con->in_msg->con = con;
+		if (con->in_msg) {
+			con->in_msg->con = ceph_con_get(con);
+			BUG_ON(con->in_msg->con == NULL);
+		}
 		if (skip)
 			con->in_msg = NULL;

@@ -2630,7 +2641,8 @@ static bool ceph_con_in_msg_alloc(struct
 ceph_connection *con,
 			       type, front_len);
 			return false;
 		}
-		con->in_msg->con = con;
+		con->in_msg->con = ceph_con_get(con);
+		BUG_ON(con->in_msg->con == NULL);
 		con->in_msg->page_alignment = le16_to_cpu(hdr->data_off);
 	}

memcpy(&con->in_msg->hdr, &con->in_hdr, sizeof(con->in_hdr));

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

I have some reservations about this one. Not because it's wrong, but because I'm not sure what it will fix (now or later). Right now the con has references to the messages. If we add refs the other way around, we need to make sure those refs go away or else the con will leak. And in order to do that, the upper layer has to properly clean up. Currently, if it doesn't clean up, the messages would point to an invalid/released con, but nothing would reference them.. they'd leak. With this change, we would also leak the con and it's containing structure. I'm not sure that's any better or worse.

The net effect here is that we're doing a lot of additional atomic operations on the containing object's ref count. Not sure if that cost is important.

What do you think?

ceph_con_revoke() is passed both a message and a ceph connection. Now that any message associated with a connection holds a pointer to that connection, there's no need to provide the connection when revoking a message.

This has the added benefit of precluding the possibility of the providing the wrong connection pointer. If the message's connection pointer is null, it is not being tracked by any connection, so revoking it is a no-op. This is supported as a convenience for upper layers, so they can revoke a message that is not actually "in flight."

Rename the function ceph_msg_revoke() to reflect that it is really an operation on a message, not a connection.

Signed-off-by Alex Elder <[email protected]>

---
 include/linux/ceph/messenger.h |    3 ++-
 net/ceph/messenger.c           |    7 ++++++-
 net/ceph/mon_client.c          |    8 ++++----
 net/ceph/osd_client.c          |    4 ++--
 4 files changed, 14 insertions(+), 8 deletions(-)

include/linux/ceph/messenger.h +2 -1

index 6df837f..9008f81 100644
@@ -239,7 +239,8 @@ extern void ceph_con_open(struct ceph_connection *con,
 extern bool ceph_con_opened(struct ceph_connection *con);
 extern void ceph_con_close(struct ceph_connection *con);
 extern void ceph_con_send(struct ceph_connection *con, struct ceph_msg
 *msg);
-extern void ceph_con_revoke(struct ceph_connection *con, struct
 ceph_msg *msg);
+
+extern void ceph_msg_revoke(struct ceph_msg *msg);
 extern void ceph_con_revoke_message(struct ceph_connection *con,

include/linux/ceph/messenger.h +0 -0

 				  struct ceph_msg *msg);
 extern void ceph_con_keepalive(struct ceph_connection *con);

net/ceph/messenger.c +6 -1

index 336707b..79f3032 100644
@@ -2421,8 +2421,13 @@ EXPORT_SYMBOL(ceph_con_send);
 /*
  * Revoke a message that was previously queued for send
  */
-void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg)
+void ceph_msg_revoke(struct ceph_msg *msg)
 {
+	struct ceph_connection *con = msg->con;
+
+	if (!con)
+		return;		/* Message not in our possession */
+
 	mutex_lock(&con->mutex);
 	if (!list_empty(&msg->list_head)) {
 		dout("%s %p msg %p - was on queue\n", __func__, con, msg);

net/ceph/mon_client.c +1 -1

index ab6b24a..cf82e04 100644
@@ -106,7 +106,7 @@ static void __send_prepared_auth_request(struct
 ceph_mon_client *monc, int len)
 	monc->pending_auth = 1;
 	monc->m_auth->front.iov_len = len;
 	monc->m_auth->hdr.front_len = cpu_to_le32(len);
-	ceph_con_revoke(&monc->con, monc->m_auth);
+	ceph_msg_revoke(monc->m_auth);
 	ceph_msg_get(monc->m_auth);  /* keep our ref */
 	ceph_con_send(&monc->con, monc->m_auth);

}

net/ceph/mon_client.c +1 -1

@@ -117,7 +117,7 @@ static void __send_prepared_auth_request(struct
 ceph_mon_client *monc, int len)
 static void __close_session(struct ceph_mon_client *monc)
 {
 	dout("__close_session closing mon%d\n", monc->cur_mon);
-	ceph_con_revoke(&monc->con, monc->m_auth);
+	ceph_msg_revoke(monc->m_auth);
 	ceph_con_close(&monc->con);
 	monc->con.private = NULL;

net/ceph/mon_client.c +1 -1

 	monc->cur_mon = -1;

@@ -229,7 +229,7 @@ static void __send_subscribe(struct ceph_mon_client
 *monc)
 
 		msg->front.iov_len = p - msg->front.iov_base;
 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
-		ceph_con_revoke(&monc->con, msg);
+		ceph_msg_revoke(msg);
 		ceph_con_send(&monc->con, ceph_msg_get(msg));
 

net/ceph/mon_client.c +1 -1

 		monc->sub_sent = jiffies | 1;  /* never 0 */

@@ -687,7 +687,7 @@ static void __resend_generic_request(struct
 ceph_mon_client *monc)
 
 	for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) {
 		req = rb_entry(p, struct ceph_mon_generic_request, node);
-		ceph_con_revoke(&monc->con, req->request);
+		ceph_msg_revoke(req->request);
 		ceph_con_send(&monc->con, ceph_msg_get(req->request));
 	}

}

net/ceph/osd_client.c +1 -1

index 448c9da..403fefb 100644
@@ -852,7 +852,7 @@ static void __unregister_request(struct
 ceph_osd_client *osdc,
 
 	if (req->r_osd) {
 		/* make sure the original request isn't in flight. */
-		ceph_con_revoke(&req->r_osd->o_con, req->r_request);
+		ceph_msg_revoke(req->r_request);
 
 		list_del_init(&req->r_osd_item);

net/ceph/osd_client.c +1 -1

 		if (list_empty(&req->r_osd->o_requests) &&

@@ -879,7 +879,7 @@ static void __unregister_request(struct
 ceph_osd_client *osdc,
 static void __cancel_request(struct ceph_osd_request *req)
 {
 	if (req->r_sent && req->r_osd) {
-		ceph_con_revoke(&req->r_osd->o_con, req->r_request);
+		ceph_msg_revoke(req->r_request);
 		req->r_sent = 0;
 	}

}

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

I'm mildly concerned that if the msg isn't on the con the upper layer thinks it is it is a bug, and we might be better off keeping this arg and asserting that con == msg->con...

ceph_con_revoke_message() is passed both a message and a ceph connection. A ceph_msg allocated for incoming messages on a connection always has a pointer to that connection, so there's no need to provide the connection when revoking such a message.

Note that the existing logic does not preclude the message supplied being a null/bogus message pointer. The only user of this interface is the OSD client, and the only value an osd client passes is a request's r_reply field. That is always non-null (except briefly in an error path in ceph_osdc_alloc_request(), and that drops the only reference so the request won't ever have a reply to revoke). So we can safely assume the passed-in message is non-null, but add a BUG_ON() to make it very obvious we are imposing this restriction.

Rename the function ceph_msg_revoke_incoming() to reflect that it is really an operation on an incoming message.

Signed-off-by Alex Elder <[email protected]>

---
 include/linux/ceph/messenger.h |    4 ++--
 net/ceph/messenger.c           |   19 +++++++++++++------
 net/ceph/osd_client.c          |    9 ++++-----
 3 files changed, 19 insertions(+), 13 deletions(-)

include/linux/ceph/messenger.h +2 -2

index 9008f81..a334dbd 100644
@@ -241,8 +241,8 @@ extern void ceph_con_close(struct ceph_connection *con);
 extern void ceph_con_send(struct ceph_connection *con, struct ceph_msg
 *msg);
 
 extern void ceph_msg_revoke(struct ceph_msg *msg);
-extern void ceph_con_revoke_message(struct ceph_connection *con,
-				  struct ceph_msg *msg);
+extern void ceph_msg_revoke_incoming(struct ceph_msg *msg);
+
 extern void ceph_con_keepalive(struct ceph_connection *con);
 extern struct ceph_connection *ceph_con_get(struct ceph_connection *con);

extern void ceph_con_put(struct ceph_connection *con);

net/ceph/messenger.c +0 -0

index 79f3032..a036141 100644
@@ -2456,17 +2456,24 @@ void ceph_msg_revoke(struct ceph_msg *msg)
 /*
  * Revoke a message that we may be reading data into

net/ceph/messenger.c +5 -3

+
+	con = msg->con;
 	mutex_lock(&con->mutex);
-	if (con->in_msg && con->in_msg == msg) {
+	if (con->in_msg == msg) {
 		unsigned front_len = le32_to_cpu(con->in_hdr.front_len);
 		unsigned middle_len = le32_to_cpu(con->in_hdr.middle_len);
 		unsigned data_len = le32_to_cpu(con->in_hdr.data_len);
 
 		/* skip rest of message */
-		dout("con_revoke_pages %p msg %p revoked\n", con, msg);
-			con->in_base_pos = con->in_base_pos -
+		dout("%s %p msg %p revoked\n", __func__, con, msg);
+		con->in_base_pos = con->in_base_pos -
 				sizeof(struct ceph_msg_header) -
 				front_len -

net/ceph/messenger.c +2 -2

 				middle_len -

@@ -2477,8 +2484,8 @@ void ceph_con_revoke_message(struct
 ceph_connection *con, struct ceph_msg *msg)
 		con->in_tag = CEPH_MSGR_TAG_READY;
 		con->in_seq++;
 	} else {
-		dout("con_revoke_pages %p msg %p pages %p no-op\n",
-		     con, con->in_msg, msg);
+		dout("%s %p in_msg %p msg %p no-op\n",
+		     __func__, con, con->in_msg, msg);
 	}
 	mutex_unlock(&con->mutex);

}

net/ceph/osd_client.c +4 -5

index 403fefb..9f87edd 100644
@@ -140,10 +140,9 @@ void ceph_osdc_release_request(struct kref *kref)
 	if (req->r_request)
 		ceph_msg_put(req->r_request);
 	if (req->r_con_filling_msg) {
-		dout("release_request revoking pages %p from con %p\n",
+		dout("%s revoking pages %p from con %p\n", __func__,
 		     req->r_pages, req->r_con_filling_msg);
-		ceph_con_revoke_message(req->r_con_filling_msg,
-				      req->r_reply);
+		ceph_msg_revoke_incoming(req->r_reply);
 		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
 	}
 	if (req->r_reply)

@@ -2022,9 +2021,9 @@ static struct ceph_msg *get_reply(struct
 ceph_connection *con,
 	}
 
 	if (req->r_con_filling_msg) {
-		dout("get_reply revoking msg %p from old con %p\n",
+		dout("%s revoking msg %p from old con %p\n", __func__,
 		     req->r_reply, req->r_con_filling_msg);
-		ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply);
+		ceph_msg_revoke_incoming(req->r_reply);
 		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
 		req->r_con_filling_msg = NULL;

}

Signature · 6 lines

1.7.5.4

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Quoted 9 lines

net/ceph/messenger.c +0 -0

Quoted 8 lines

index 79f3032..a036141 100644

net/ceph/messenger.c +0 -1

@@ -2456,17 +2456,24 @@ void ceph_msg_revoke(struct ceph_msg *msg)
 /*
  * Revoke a message that we may be reading data into
  */
-void ceph_con_revoke_message(struct ceph_connection *con, struct

Quoted 8 lines

net/ceph/messenger.c void ceph_msg_revoke(struct ceph_msg *msg) · line 2466

Quoted 8 lines

net/ceph/messenger.c +6 -0

 ceph_msg *msg)
+void ceph_msg_revoke_incoming(struct ceph_msg *msg)
 {
+	struct ceph_connection *con;
+
+	BUG_ON(msg == NULL);
+	if (!msg->con)
+		return;		/* Message not in our posession */

This case weirds me out. I think it can happen, but maybe we should at least get a debug msg, like the no-op one below.

Reviewed-by Sage Weil <[email protected]>