ceph: update the auth cap when the async create req is forwarded

For async create we will always try to choose the auth MDS of frag
the dentry belonged to of the parent directory to send the request
and ususally this works fine, but if the MDS migrated the directory
to another MDS before it could be handled the request will be
forwarded. And then the auth cap will be changed.

We need to update the auth cap in this case before the request is
forwarded.

Link: https://tracker.ceph.com/issues/55857
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>

1 files changed, 58 insertions, 0 deletions

diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 9cfa7b775fdb..80f8b9ec1a31 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -2960,6 +2960,64 @@ static void __do_request(struct ceph_mds_client *mdsc,
 	if (req->r_request_started == 0)   /* note request start time */
 		req->r_request_started = jiffies;
 
+	/*
+	 * For async create we will choose the auth MDS of frag in parent
+	 * directory to send the request and ususally this works fine, but
+	 * if the migrated the dirtory to another MDS before it could handle
+	 * it the request will be forwarded.
+	 *
+	 * And then the auth cap will be changed.
+	 */
+	if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags) && req->r_num_fwd) {
+		struct ceph_dentry_info *di = ceph_dentry(req->r_dentry);
+		struct ceph_inode_info *ci;
+		struct ceph_cap *cap;
+
+		/*
+		 * The request maybe handled very fast and the new inode
+		 * hasn't been linked to the dentry yet. We need to wait
+		 * for the ceph_finish_async_create(), which shouldn't be
+		 * stuck too long or fail in thoery, to finish when forwarding
+		 * the request.
+		 */
+		if (!d_inode(req->r_dentry)) {
+			err = wait_on_bit(&di->flags, CEPH_DENTRY_ASYNC_CREATE_BIT,
+					  TASK_KILLABLE);
+			if (err) {
+				mutex_lock(&req->r_fill_mutex);
+				set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags);
+				mutex_unlock(&req->r_fill_mutex);
+				goto out_session;
+			}
+		}
+
+		ci = ceph_inode(d_inode(req->r_dentry));
+
+		spin_lock(&ci->i_ceph_lock);
+		cap = ci->i_auth_cap;
+		if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE && mds != cap->mds) {
+			dout("do_request session changed for auth cap %d -> %d\n",
+			     cap->session->s_mds, session->s_mds);
+
+			/* Remove the auth cap from old session */
+			spin_lock(&cap->session->s_cap_lock);
+			cap->session->s_nr_caps--;
+			list_del_init(&cap->session_caps);
+			spin_unlock(&cap->session->s_cap_lock);
+
+			/* Add the auth cap to the new session */
+			cap->mds = mds;
+			cap->session = session;
+			spin_lock(&session->s_cap_lock);
+			session->s_nr_caps++;
+			list_add_tail(&cap->session_caps, &session->s_caps);
+			spin_unlock(&session->s_cap_lock);
+
+			change_auth_cap_ses(ci, session);
+		}
+		spin_unlock(&ci->i_ceph_lock);
+	}
+
 	err = __send_request(session, req, false);
 
 out_session: