summary refs log tree commit diff
diff options
context:
space:
mode:
authorXiubo Li <xiubli@redhat.com>2023-07-25 12:03:59 +0800
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>2023-08-11 12:08:19 +0200
commita062da58ed97d1f16a564619726e77ba528c5191 (patch)
tree0f0921c203c7b25d0f5be37fe7515f69509dfc91
parentad82aac732c22e043307933e13fe112682bfff0c (diff)
downloadlinux-a062da58ed97d1f16a564619726e77ba528c5191.tar.gz
ceph: defer stopping mdsc delayed_work
commit e7e607bd00481745550389a29ecabe33e13d67cf upstream.

Flushing the dirty buffer may take a long time if the cluster is
overloaded or if there is network issue. So we should ping the
MDSs periodically to keep alive, else the MDS will blocklist
the kclient.

Cc: stable@vger.kernel.org
Link: https://tracker.ceph.com/issues/61843
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Milind Changire <mchangir@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-rw-r--r--fs/ceph/mds_client.c4
-rw-r--r--fs/ceph/mds_client.h5
-rw-r--r--fs/ceph/super.c10
3 files changed, 17 insertions, 2 deletions
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 1989c8deea55..dcabe2783edf 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -4758,7 +4758,7 @@ static void delayed_work(struct work_struct *work)
 
 	dout("mdsc delayed_work\n");
 
-	if (mdsc->stopping)
+	if (mdsc->stopping >= CEPH_MDSC_STOPPING_FLUSHED)
 		return;
 
 	mutex_lock(&mdsc->mutex);
@@ -4937,7 +4937,7 @@ void send_flush_mdlog(struct ceph_mds_session *s)
 void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc)
 {
 	dout("pre_umount\n");
-	mdsc->stopping = 1;
+	mdsc->stopping = CEPH_MDSC_STOPPING_BEGIN;
 
 	ceph_mdsc_iterate_sessions(mdsc, send_flush_mdlog, true);
 	ceph_mdsc_iterate_sessions(mdsc, lock_unlock_session, false);
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index 18b026b1ac63..9a80658f4167 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -380,6 +380,11 @@ struct cap_wait {
 	int			want;
 };
 
+enum {
+       CEPH_MDSC_STOPPING_BEGIN = 1,
+       CEPH_MDSC_STOPPING_FLUSHED = 2,
+};
+
 /*
  * mds client state
  */
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 3fc48b43cab0..a5f52013314d 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -1374,6 +1374,16 @@ static void ceph_kill_sb(struct super_block *s)
 	ceph_mdsc_pre_umount(fsc->mdsc);
 	flush_fs_workqueues(fsc);
 
+	/*
+	 * Though the kill_anon_super() will finally trigger the
+	 * sync_filesystem() anyway, we still need to do it here
+	 * and then bump the stage of shutdown to stop the work
+	 * queue as earlier as possible.
+	 */
+	sync_filesystem(s);
+
+	fsc->mdsc->stopping = CEPH_MDSC_STOPPING_FLUSHED;
+
 	kill_anon_super(s);
 
 	fsc->client->extra_mon_dispatch = NULL;