1 files changed, 93 insertions, 38 deletions

diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 920e9f048bd8..a5163296d9d9 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -384,8 +384,8 @@ static int parse_reply_info_readdir(void **p, void *end,
 	}
 
 done:
-	if (*p != end)
-		goto bad;
+	/* Skip over any unrecognized fields */
+	*p = end;
 	return 0;
 
 bad:
@@ -406,12 +406,10 @@ static int parse_reply_info_filelock(void **p, void *end,
 		goto bad;
 
 	info->filelock_reply = *p;
-	*p += sizeof(*info->filelock_reply);
 
-	if (unlikely(*p != end))
-		goto bad;
+	/* Skip over any unrecognized fields */
+	*p = end;
 	return 0;
-
 bad:
 	return -EIO;
 }
@@ -425,18 +423,21 @@ static int parse_reply_info_create(void **p, void *end,
 {
 	if (features == (u64)-1 ||
 	    (features & CEPH_FEATURE_REPLY_CREATE_INODE)) {
+		/* Malformed reply? */
 		if (*p == end) {
 			info->has_create_ino = false;
 		} else {
 			info->has_create_ino = true;
-			info->ino = ceph_decode_64(p);
+			ceph_decode_64_safe(p, end, info->ino, bad);
 		}
+	} else {
+		if (*p != end)
+			goto bad;
 	}
 
-	if (unlikely(*p != end))
-		goto bad;
+	/* Skip over any unrecognized fields */
+	*p = end;
 	return 0;
-
 bad:
 	return -EIO;
 }
@@ -639,7 +640,6 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
 	s->s_renew_seq = 0;
 	INIT_LIST_HEAD(&s->s_caps);
 	s->s_nr_caps = 0;
-	s->s_trim_caps = 0;
 	refcount_set(&s->s_ref, 1);
 	INIT_LIST_HEAD(&s->s_waiting);
 	INIT_LIST_HEAD(&s->s_unsafe);
@@ -1270,6 +1270,7 @@ static void cleanup_session_requests(struct ceph_mds_client *mdsc,
 {
 	struct ceph_mds_request *req;
 	struct rb_node *p;
+	struct ceph_inode_info *ci;
 
 	dout("cleanup_session_requests mds%d\n", session->s_mds);
 	mutex_lock(&mdsc->mutex);
@@ -1278,6 +1279,16 @@ static void cleanup_session_requests(struct ceph_mds_client *mdsc,
 				       struct ceph_mds_request, r_unsafe_item);
 		pr_warn_ratelimited(" dropping unsafe request %llu\n",
 				    req->r_tid);
+		if (req->r_target_inode) {
+			/* dropping unsafe change of inode's attributes */
+			ci = ceph_inode(req->r_target_inode);
+			errseq_set(&ci->i_meta_err, -EIO);
+		}
+		if (req->r_unsafe_dir) {
+			/* dropping unsafe directory operation */
+			ci = ceph_inode(req->r_unsafe_dir);
+			errseq_set(&ci->i_meta_err, -EIO);
+		}
 		__unregister_request(mdsc, req);
 	}
 	/* zero r_attempts, so kick_requests() will re-send requests */
@@ -1370,7 +1381,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 	struct ceph_fs_client *fsc = (struct ceph_fs_client *)arg;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	LIST_HEAD(to_remove);
-	bool drop = false;
+	bool dirty_dropped = false;
 	bool invalidate = false;
 
 	dout("removing cap %p, ci is %p, inode is %p\n",
@@ -1383,9 +1394,12 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 		struct ceph_cap_flush *cf;
 		struct ceph_mds_client *mdsc = fsc->mdsc;
 
-		if (ci->i_wrbuffer_ref > 0 &&
-		    READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN)
-			invalidate = true;
+		if (READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
+			if (inode->i_data.nrpages > 0)
+				invalidate = true;
+			if (ci->i_wrbuffer_ref > 0)
+				mapping_set_error(&inode->i_data, -EIO);
+		}
 
 		while (!list_empty(&ci->i_cap_flush_list)) {
 			cf = list_first_entry(&ci->i_cap_flush_list,
@@ -1405,7 +1419,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 				inode, ceph_ino(inode));
 			ci->i_dirty_caps = 0;
 			list_del_init(&ci->i_dirty_item);
-			drop = true;
+			dirty_dropped = true;
 		}
 		if (!list_empty(&ci->i_flushing_item)) {
 			pr_warn_ratelimited(
@@ -1415,10 +1429,22 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 			ci->i_flushing_caps = 0;
 			list_del_init(&ci->i_flushing_item);
 			mdsc->num_cap_flushing--;
-			drop = true;
+			dirty_dropped = true;
 		}
 		spin_unlock(&mdsc->cap_dirty_lock);
 
+		if (dirty_dropped) {
+			errseq_set(&ci->i_meta_err, -EIO);
+
+			if (ci->i_wrbuffer_ref_head == 0 &&
+			    ci->i_wr_ref == 0 &&
+			    ci->i_dirty_caps == 0 &&
+			    ci->i_flushing_caps == 0) {
+				ceph_put_snap_context(ci->i_head_snapc);
+				ci->i_head_snapc = NULL;
+			}
+		}
+
 		if (atomic_read(&ci->i_filelock_ref) > 0) {
 			/* make further file lock syscall return -EIO */
 			ci->i_ceph_flags |= CEPH_I_ERROR_FILELOCK;
@@ -1430,15 +1456,6 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 			list_add(&ci->i_prealloc_cap_flush->i_list, &to_remove);
 			ci->i_prealloc_cap_flush = NULL;
 		}
-
-               if (drop &&
-                  ci->i_wrbuffer_ref_head == 0 &&
-                  ci->i_wr_ref == 0 &&
-                  ci->i_dirty_caps == 0 &&
-                  ci->i_flushing_caps == 0) {
-                      ceph_put_snap_context(ci->i_head_snapc);
-                      ci->i_head_snapc = NULL;
-               }
 	}
 	spin_unlock(&ci->i_ceph_lock);
 	while (!list_empty(&to_remove)) {
@@ -1452,7 +1469,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 	wake_up_all(&ci->i_cap_wq);
 	if (invalidate)
 		ceph_queue_invalidate(inode);
-	if (drop)
+	if (dirty_dropped)
 		iput(inode);
 	return 0;
 }
@@ -1705,11 +1722,11 @@ out:
  */
 static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg)
 {
-	struct ceph_mds_session *session = arg;
+	int *remaining = arg;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	int used, wanted, oissued, mine;
 
-	if (session->s_trim_caps <= 0)
+	if (*remaining <= 0)
 		return -1;
 
 	spin_lock(&ci->i_ceph_lock);
@@ -1746,7 +1763,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg)
 	if (oissued) {
 		/* we aren't the only cap.. just remove us */
 		__ceph_remove_cap(cap, true);
-		session->s_trim_caps--;
+		(*remaining)--;
 	} else {
 		struct dentry *dentry;
 		/* try dropping referring dentries */
@@ -1758,7 +1775,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg)
 			d_prune_aliases(inode);
 			count = atomic_read(&inode->i_count);
 			if (count == 1)
-				session->s_trim_caps--;
+				(*remaining)--;
 			dout("trim_caps_cb %p cap %p pruned, count now %d\n",
 			     inode, cap, count);
 		} else {
@@ -1784,12 +1801,12 @@ int ceph_trim_caps(struct ceph_mds_client *mdsc,
 	dout("trim_caps mds%d start: %d / %d, trim %d\n",
 	     session->s_mds, session->s_nr_caps, max_caps, trim_caps);
 	if (trim_caps > 0) {
-		session->s_trim_caps = trim_caps;
-		ceph_iterate_session_caps(session, trim_caps_cb, session);
+		int remaining = trim_caps;
+
+		ceph_iterate_session_caps(session, trim_caps_cb, &remaining);
 		dout("trim_caps mds%d done: %d / %d, trimmed %d\n",
 		     session->s_mds, session->s_nr_caps, max_caps,
-			trim_caps - session->s_trim_caps);
-		session->s_trim_caps = 0;
+			trim_caps - remaining);
 	}
 
 	ceph_flush_cap_releases(mdsc, session);
@@ -3015,18 +3032,23 @@ bad:
 	pr_err("mdsc_handle_forward decode error err=%d\n", err);
 }
 
-static int __decode_and_drop_session_metadata(void **p, void *end)
+static int __decode_session_metadata(void **p, void *end,
+				     bool *blacklisted)
 {
 	/* map<string,string> */
 	u32 n;
+	bool err_str;
 	ceph_decode_32_safe(p, end, n, bad);
 	while (n-- > 0) {
 		u32 len;
 		ceph_decode_32_safe(p, end, len, bad);
 		ceph_decode_need(p, end, len, bad);
+		err_str = !strncmp(*p, "error_string", len);
 		*p += len;
 		ceph_decode_32_safe(p, end, len, bad);
 		ceph_decode_need(p, end, len, bad);
+		if (err_str && strnstr(*p, "blacklisted", len))
+			*blacklisted = true;
 		*p += len;
 	}
 	return 0;
@@ -3050,6 +3072,7 @@ static void handle_session(struct ceph_mds_session *session,
 	u64 seq;
 	unsigned long features = 0;
 	int wake = 0;
+	bool blacklisted = false;
 
 	/* decode */
 	ceph_decode_need(&p, end, sizeof(*h), bad);
@@ -3062,7 +3085,7 @@ static void handle_session(struct ceph_mds_session *session,
 	if (msg_version >= 3) {
 		u32 len;
 		/* version >= 2, metadata */
-		if (__decode_and_drop_session_metadata(&p, end) < 0)
+		if (__decode_session_metadata(&p, end, &blacklisted) < 0)
 			goto bad;
 		/* version >= 3, feature bits */
 		ceph_decode_32_safe(&p, end, len, bad);
@@ -3149,6 +3172,8 @@ static void handle_session(struct ceph_mds_session *session,
 		session->s_state = CEPH_MDS_SESSION_REJECTED;
 		cleanup_session_requests(mdsc, session);
 		remove_session_caps(session);
+		if (blacklisted)
+			mdsc->fsc->blacklisted = true;
 		wake = 2; /* for good measure */
 		break;
 
@@ -3998,7 +4023,27 @@ static void lock_unlock_sessions(struct ceph_mds_client *mdsc)
 	mutex_unlock(&mdsc->mutex);
 }
 
+static void maybe_recover_session(struct ceph_mds_client *mdsc)
+{
+	struct ceph_fs_client *fsc = mdsc->fsc;
+
+	if (!ceph_test_mount_opt(fsc, CLEANRECOVER))
+		return;
+
+	if (READ_ONCE(fsc->mount_state) != CEPH_MOUNT_MOUNTED)
+		return;
+
+	if (!READ_ONCE(fsc->blacklisted))
+		return;
 
+	if (fsc->last_auto_reconnect &&
+	    time_before(jiffies, fsc->last_auto_reconnect + HZ * 60 * 30))
+		return;
+
+	pr_info("auto reconnect after blacklisted\n");
+	fsc->last_auto_reconnect = jiffies;
+	ceph_force_reconnect(fsc->sb);
+}
 
 /*
  * delayed work -- periodically trim expired leases, renew caps with mds
@@ -4044,7 +4089,9 @@ static void delayed_work(struct work_struct *work)
 				pr_info("mds%d hung\n", s->s_mds);
 			}
 		}
-		if (s->s_state < CEPH_MDS_SESSION_OPEN) {
+		if (s->s_state == CEPH_MDS_SESSION_NEW ||
+		    s->s_state == CEPH_MDS_SESSION_RESTARTING ||
+		    s->s_state == CEPH_MDS_SESSION_REJECTED) {
 			/* this mds is failed or recovering, just wait */
 			ceph_put_mds_session(s);
 			continue;
@@ -4072,6 +4119,8 @@ static void delayed_work(struct work_struct *work)
 
 	ceph_trim_snapid_map(mdsc);
 
+	maybe_recover_session(mdsc);
+
 	schedule_delayed(mdsc);
 }
 
@@ -4355,7 +4404,12 @@ void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc)
 		session = __ceph_lookup_mds_session(mdsc, mds);
 		if (!session)
 			continue;
+
+		if (session->s_state == CEPH_MDS_SESSION_REJECTED)
+			__unregister_session(mdsc, session);
+		__wake_requests(mdsc, &session->s_waiting);
 		mutex_unlock(&mdsc->mutex);
+
 		mutex_lock(&session->s_mutex);
 		__close_session(mdsc, session);
 		if (session->s_state == CEPH_MDS_SESSION_CLOSING) {
@@ -4364,6 +4418,7 @@ void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc)
 		}
 		mutex_unlock(&session->s_mutex);
 		ceph_put_mds_session(session);
+
 		mutex_lock(&mdsc->mutex);
 		kick_requests(mdsc, mds);
 	}