libceph: fix alloc_msg_with_page_vector() memory leaks

Make it so that CEPH_MSG_DATA_PAGES data item can own pages,
fixing a bunch of memory leaks for a page vector allocated in
alloc_msg_with_page_vector().  Currently, only watch-notify
messages trigger this allocation, and normally the page vector
is freed either in handle_watch_notify() or by the caller of
ceph_osdc_notify().  But if the message is freed before that
(e.g. if the session faults while reading in the message or
if the notify is stale), we leak the page vector.

This was supposed to be fixed by switching to a message-owned
pagelist, but that never happened.

Fixes: 1907920324f1 ("libceph: support for sending notifies")
Reported-by: Roman Penyaev <rpenyaev@suse.de>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Roman Penyaev <rpenyaev@suse.de>

2 files changed, 10 insertions, 13 deletions

diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index 5b4bd8261002..f8ca5edc5f2c 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -3248,12 +3248,16 @@ static struct ceph_msg_data *ceph_msg_data_add(struct ceph_msg *msg)
 
 static void ceph_msg_data_destroy(struct ceph_msg_data *data)
 {
-	if (data->type == CEPH_MSG_DATA_PAGELIST)
+	if (data->type == CEPH_MSG_DATA_PAGES && data->own_pages) {
+		int num_pages = calc_pages_for(data->alignment, data->length);
+		ceph_release_page_vector(data->pages, num_pages);
+	} else if (data->type == CEPH_MSG_DATA_PAGELIST) {
 		ceph_pagelist_release(data->pagelist);
+	}
 }
 
 void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
-		size_t length, size_t alignment)
+			     size_t length, size_t alignment, bool own_pages)
 {
 	struct ceph_msg_data *data;
 
@@ -3265,6 +3269,7 @@ void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
 	data->pages = pages;
 	data->length = length;
 	data->alignment = alignment & ~PAGE_MASK;
+	data->own_pages = own_pages;
 
 	msg->data_length += length;
 }
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index b68b376d8c2f..af868d3923b9 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -962,7 +962,7 @@ static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
 		BUG_ON(length > (u64) SIZE_MAX);
 		if (length)
 			ceph_msg_data_add_pages(msg, osd_data->pages,
-					length, osd_data->alignment);
+					length, osd_data->alignment, false);
 	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
 		BUG_ON(!length);
 		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
@@ -4436,9 +4436,7 @@ static void handle_watch_notify(struct ceph_osd_client *osdc,
 							CEPH_MSG_DATA_PAGES);
 					*lreq->preply_pages = data->pages;
 					*lreq->preply_len = data->length;
-				} else {
-					ceph_release_page_vector(data->pages,
-					       calc_pages_for(0, data->length));
+					data->own_pages = false;
 				}
 			}
 			lreq->notify_finish_error = return_code;
@@ -5506,9 +5504,6 @@ out_unlock_osdc:
 	return m;
 }
 
-/*
- * TODO: switch to a msg-owned pagelist
- */
 static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 {
 	struct ceph_msg *m;
@@ -5522,7 +5517,6 @@ static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 
 	if (data_len) {
 		struct page **pages;
-		struct ceph_osd_data osd_data;
 
 		pages = ceph_alloc_page_vector(calc_pages_for(0, data_len),
 					       GFP_NOIO);
@@ -5531,9 +5525,7 @@ static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 			return NULL;
 		}
 
-		ceph_osd_data_pages_init(&osd_data, pages, data_len, 0, false,
-					 false);
-		ceph_osdc_msg_data_add(m, &osd_data);
+		ceph_msg_data_add_pages(m, pages, data_len, 0, true);
 	}
 
 	return m;