From: Xiubo Li <xiubli@redhat.com>
Date: Tue, 13 Jun 2023 04:49:59 +0000 (+0800)
Subject: ceph: issue a cap release immediately if no cap exists
X-Git-Tag: v6.6.7~2395^2~1
X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=ce72d4e0f179340cece90d5b826eb63bbf9fefc0;p=platform%2Fkernel%2Flinux-starfive.git

ceph: issue a cap release immediately if no cap exists

In case:

           mds                             client
                                - Releases cap and put Inode
  - Increase cap->seq and sends
    revokes req to the client
  - Receives release req and    - Receives & drops the revoke req
    skip removing the cap and
    then eval the CInode and
    issue or revoke caps again.
                                - Receives & drops the caps update
                                  or revoke req
  - Health warning for client
    isn't responding to
    mclientcaps(revoke)

All the IMPORT/REVOKE/GRANT cap ops will increase the session seq
in MDS side and then the client need to issue a cap release to
unblock MDS to remove the corresponding cap to unblock possible
waiters.

Link: https://tracker.ceph.com/issues/61332
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Milind Changire <mchangir@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
---

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 6fb73e3..cef91dd 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -4092,6 +4092,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
 	struct cap_extra_info extra_info = {};
 	bool queue_trunc;
 	bool close_sessions = false;
+	bool do_cap_release = false;
 
 	dout("handle_caps from mds%d\n", session->s_mds);
 
@@ -4198,17 +4199,14 @@ void ceph_handle_caps(struct ceph_mds_session *session,
 	if (!inode) {
 		dout(" i don't have ino %llx\n", vino.ino);
 
-		if (op == CEPH_CAP_OP_IMPORT) {
-			cap = ceph_get_cap(mdsc, NULL);
-			cap->cap_ino = vino.ino;
-			cap->queue_release = 1;
-			cap->cap_id = le64_to_cpu(h->cap_id);
-			cap->mseq = mseq;
-			cap->seq = seq;
-			cap->issue_seq = seq;
-			spin_lock(&session->s_cap_lock);
-			__ceph_queue_cap_release(session, cap);
-			spin_unlock(&session->s_cap_lock);
+		switch (op) {
+		case CEPH_CAP_OP_IMPORT:
+		case CEPH_CAP_OP_REVOKE:
+		case CEPH_CAP_OP_GRANT:
+			do_cap_release = true;
+			break;
+		default:
+			break;
 		}
 		goto flush_cap_releases;
 	}
@@ -4258,6 +4256,14 @@ void ceph_handle_caps(struct ceph_mds_session *session,
 		     inode, ceph_ino(inode), ceph_snap(inode),
 		     session->s_mds);
 		spin_unlock(&ci->i_ceph_lock);
+		switch (op) {
+		case CEPH_CAP_OP_REVOKE:
+		case CEPH_CAP_OP_GRANT:
+			do_cap_release = true;
+			break;
+		default:
+			break;
+		}
 		goto flush_cap_releases;
 	}
 
@@ -4308,6 +4314,18 @@ flush_cap_releases:
 	 * along for the mds (who clearly thinks we still have this
 	 * cap).
 	 */
+	if (do_cap_release) {
+		cap = ceph_get_cap(mdsc, NULL);
+		cap->cap_ino = vino.ino;
+		cap->queue_release = 1;
+		cap->cap_id = le64_to_cpu(h->cap_id);
+		cap->mseq = mseq;
+		cap->seq = seq;
+		cap->issue_seq = seq;
+		spin_lock(&session->s_cap_lock);
+		__ceph_queue_cap_release(session, cap);
+		spin_unlock(&session->s_cap_lock);
+	}
 	ceph_flush_cap_releases(mdsc, session);
 	goto done;