Blob Blame History Raw
From: Ilya Dryomov <idryomov@gmail.com>
Date: Wed, 20 Mar 2019 09:46:58 +0100
Subject: libceph: wait for latest osdmap in ceph_monc_blacklist_add()
Git-commit: bb229bbb3bf63d23128e851a1f3b85c083178fa1
Patch-mainline: v5.1-rc2
References: bsc#1130412

Because map updates are distributed lazily, an OSD may not know about
the new blacklist for quite some time after "osd blacklist add" command
is completed.  This makes it possible for a blacklisted but still alive
client to overwrite a post-blacklist update, resulting in data
corruption.

Waiting for latest osdmap in ceph_monc_blacklist_add() and thus using
the post-blacklist epoch for all post-blacklist requests ensures that
all such requests "wait" for the blacklist to come into force on their
respective OSDs.

Cc: stable@vger.kernel.org
Fixes: 6305a3b41515 ("libceph: support for blacklisting clients")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Acked-by: Luis Henriques <lhenriques@suse.com>
---
 include/linux/ceph/libceph.h |    2 ++
 net/ceph/ceph_common.c       |   18 +++++++++++++++++-
 net/ceph/mon_client.c        |    9 +++++++++
 3 files changed, 28 insertions(+), 1 deletion(-)

--- a/include/linux/ceph/libceph.h
+++ b/include/linux/ceph/libceph.h
@@ -291,6 +291,8 @@ extern void ceph_destroy_client(struct c
 extern int __ceph_open_session(struct ceph_client *client,
 			       unsigned long started);
 extern int ceph_open_session(struct ceph_client *client);
+int ceph_wait_for_latest_osdmap(struct ceph_client *client,
+				unsigned long timeout);
 
 /* pagevec.c */
 extern void ceph_release_page_vector(struct page **pages, int num_pages);
--- a/net/ceph/ceph_common.c
+++ b/net/ceph/ceph_common.c
@@ -729,7 +729,6 @@ int __ceph_open_session(struct ceph_clie
 }
 EXPORT_SYMBOL(__ceph_open_session);
 
-
 int ceph_open_session(struct ceph_client *client)
 {
 	int ret;
@@ -745,6 +744,23 @@ int ceph_open_session(struct ceph_client
 }
 EXPORT_SYMBOL(ceph_open_session);
 
+int ceph_wait_for_latest_osdmap(struct ceph_client *client,
+				unsigned long timeout)
+{
+	u64 newest_epoch;
+	int ret;
+
+	ret = ceph_monc_get_version(&client->monc, "osdmap", &newest_epoch);
+	if (ret)
+		return ret;
+
+	if (client->osdc.osdmap->epoch >= newest_epoch)
+		return 0;
+
+	ceph_osdc_maybe_request_map(&client->osdc);
+	return ceph_monc_wait_osdmap(&client->monc, newest_epoch, timeout);
+}
+EXPORT_SYMBOL(ceph_wait_for_latest_osdmap);
 
 static int __init init_ceph_lib(void)
 {
--- a/net/ceph/mon_client.c
+++ b/net/ceph/mon_client.c
@@ -921,6 +921,15 @@ int ceph_monc_blacklist_add(struct ceph_
 	mutex_unlock(&monc->mutex);
 
 	ret = wait_generic_request(req);
+	if (!ret)
+		/*
+		 * Make sure we have the osdmap that includes the blacklist
+		 * entry.  This is needed to ensure that the OSDs pick up the
+		 * new blacklist before processing any future requests from
+		 * this client.
+		 */
+		ret = ceph_wait_for_latest_osdmap(monc->client, 0);
+
 out:
 	put_generic_request(req);
 	return ret;