This is a note to let you know that I've just added the patch titled ceph: check session state after bumping session->s_seq to the 5.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: ceph-check-session-state-after-bumping-session-s_seq.patch and it can be found in the queue-5.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. commit 137b1125848af6a44ce33ca52c9c684998b3d99b Author: Jeff Layton <jlayton@xxxxxxxxxx> Date: Mon Oct 12 09:39:06 2020 -0400 ceph: check session state after bumping session->s_seq [ Upstream commit 62575e270f661aba64778cbc5f354511cf9abb21 ] Some messages sent by the MDS entail a session sequence number increment, and the MDS will drop certain types of requests on the floor when the sequence numbers don't match. In particular, a REQUEST_CLOSE message can cross with one of the sequence morphing messages from the MDS which can cause the client to stall, waiting for a response that will never come. Originally, this meant an up to 5s delay before the recurring workqueue job kicked in and resent the request, but a recent change made it so that the client would never resend, causing a 60s stall unmounting and sometimes a blockisting event. Add a new helper for incrementing the session sequence and then testing to see whether a REQUEST_CLOSE needs to be resent, and move the handling of CEPH_MDS_SESSION_CLOSING into that function. Change all of the bare sequence counter increments to use the new helper. Reorganize check_session_state with a switch statement. It should no longer be called when the session is CLOSING, so throw a warning if it ever is (but still handle that case sanely). [ idryomov: whitespace, pr_err() call fixup ] URL: https://tracker.ceph.com/issues/47563 Fixes: fa9967734227 ("ceph: fix potential mdsc use-after-free crash") Reported-by: Patrick Donnelly <pdonnell@xxxxxxxxxx> Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx> Reviewed-by: Ilya Dryomov <idryomov@xxxxxxxxx> Reviewed-by: Xiubo Li <xiubli@xxxxxxxxxx> Signed-off-by: Ilya Dryomov <idryomov@xxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 034b3f4fdd3a7..64a64a29f5c79 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -4064,7 +4064,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, vino.snap, inode); mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); dout(" mds%d seq %lld cap seq %u\n", session->s_mds, session->s_seq, (unsigned)seq); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 76d8d9495d1d4..b2214679baf4e 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -4227,7 +4227,7 @@ static void handle_lease(struct ceph_mds_client *mdsc, dname.len, dname.name); mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); if (!inode) { dout("handle_lease no inode %llx\n", vino.ino); @@ -4381,28 +4381,48 @@ static void maybe_recover_session(struct ceph_mds_client *mdsc) bool check_session_state(struct ceph_mds_session *s) { - if (s->s_state == CEPH_MDS_SESSION_CLOSING) { - dout("resending session close request for mds%d\n", - s->s_mds); - request_close_session(s); - return false; - } - if (s->s_ttl && time_after(jiffies, s->s_ttl)) { - if (s->s_state == CEPH_MDS_SESSION_OPEN) { + switch (s->s_state) { + case CEPH_MDS_SESSION_OPEN: + if (s->s_ttl && time_after(jiffies, s->s_ttl)) { s->s_state = CEPH_MDS_SESSION_HUNG; pr_info("mds%d hung\n", s->s_mds); } - } - if (s->s_state == CEPH_MDS_SESSION_NEW || - s->s_state == CEPH_MDS_SESSION_RESTARTING || - s->s_state == CEPH_MDS_SESSION_CLOSED || - s->s_state == CEPH_MDS_SESSION_REJECTED) - /* this mds is failed or recovering, just wait */ + break; + case CEPH_MDS_SESSION_CLOSING: + /* Should never reach this when we're unmounting */ + WARN_ON_ONCE(true); + fallthrough; + case CEPH_MDS_SESSION_NEW: + case CEPH_MDS_SESSION_RESTARTING: + case CEPH_MDS_SESSION_CLOSED: + case CEPH_MDS_SESSION_REJECTED: return false; + } return true; } +/* + * If the sequence is incremented while we're waiting on a REQUEST_CLOSE reply, + * then we need to retransmit that request. + */ +void inc_session_sequence(struct ceph_mds_session *s) +{ + lockdep_assert_held(&s->s_mutex); + + s->s_seq++; + + if (s->s_state == CEPH_MDS_SESSION_CLOSING) { + int ret; + + dout("resending session close request for mds%d\n", s->s_mds); + ret = request_close_session(s); + if (ret < 0) + pr_err("unable to close session to mds%d: %d\n", + s->s_mds, ret); + } +} + /* * delayed work -- periodically trim expired leases, renew caps with mds */ diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 658800605bfb4..11f20a4d36bc5 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -480,6 +480,7 @@ struct ceph_mds_client { extern const char *ceph_mds_op_name(int op); extern bool check_session_state(struct ceph_mds_session *s); +void inc_session_sequence(struct ceph_mds_session *s); extern struct ceph_mds_session * __ceph_lookup_mds_session(struct ceph_mds_client *, int mds); diff --git a/fs/ceph/quota.c b/fs/ceph/quota.c index cc2c4d40b0222..2b213f864c564 100644 --- a/fs/ceph/quota.c +++ b/fs/ceph/quota.c @@ -53,7 +53,7 @@ void ceph_handle_quota(struct ceph_mds_client *mdsc, /* increment msg sequence number */ mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); mutex_unlock(&session->s_mutex); /* lookup inode */ diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 923be9399b21c..cc9a9bfc790a3 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -873,7 +873,7 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, ceph_snap_op_name(op), split, trace_len); mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); mutex_unlock(&session->s_mutex); down_write(&mdsc->snap_rwsem);