Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
  [patch 3/3] OCFS2 Configurable timeouts - Protocol changes
  [patch 2/3] OCFS2 Configurable timeouts
  [patch 1/3] OCFS2 - Expose struct o2nm_cluster
  ocfs2: Synchronize feature incompat flags in ocfs2_fs.h
  ocfs2: update mount option documentation
  ocfs2: local mounts

Documentation/filesystems/ocfs2.txt

@@ -54,3 +54,6 @@ errors=panic		Panic and halt the machine if an error occurs.
 intr		(*)	Allow signals to interrupt cluster operations.
 nointr			Do not allow signals to interrupt cluster
 			operations.
+atime_quantum=60(*)	OCFS2 will not update atime unless this number
+			of seconds has passed since the last update.
+			Set to zero to always update atime.

fs/ocfs2/cluster/nodemanager.c

@@ -35,7 +35,7 @@
 /* for now we operate under the assertion that there can be only one
  * cluster active at a time.  Changing this will require trickling
  * cluster references throughout where nodes are looked up */
-static struct o2nm_cluster *o2nm_single_cluster = NULL;
+struct o2nm_cluster *o2nm_single_cluster = NULL;
 
 #define OCFS2_MAX_HB_CTL_PATH 256
 static char ocfs2_hb_ctl_path[OCFS2_MAX_HB_CTL_PATH] = "/sbin/ocfs2_hb_ctl";
@@ -97,17 +97,6 @@ const char *o2nm_get_hb_ctl_path(void)
 }
 EXPORT_SYMBOL_GPL(o2nm_get_hb_ctl_path);
 
-struct o2nm_cluster {
-	struct config_group	cl_group;
-	unsigned		cl_has_local:1;
-	u8			cl_local_node;
-	rwlock_t		cl_nodes_lock;
-	struct o2nm_node  	*cl_nodes[O2NM_MAX_NODES];
-	struct rb_root		cl_node_ip_tree;
-	/* this bitmap is part of a hack for disk bitmap.. will go eventually. - zab */
-	unsigned long	cl_nodes_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
-};
-
 struct o2nm_node *o2nm_get_node_by_num(u8 node_num)
 {
 	struct o2nm_node *node = NULL;
@@ -543,6 +532,179 @@ static struct o2nm_node_group *to_o2nm_node_group(struct config_group *group)
 }
 #endif
 
+struct o2nm_cluster_attribute {
+	struct configfs_attribute attr;
+	ssize_t (*show)(struct o2nm_cluster *, char *);
+	ssize_t (*store)(struct o2nm_cluster *, const char *, size_t);
+};
+
+static ssize_t o2nm_cluster_attr_write(const char *page, ssize_t count,
+                                       unsigned int *val)
+{
+	unsigned long tmp;
+	char *p = (char *)page;
+
+	tmp = simple_strtoul(p, &p, 0);
+	if (!p || (*p && (*p != '\n')))
+		return -EINVAL;
+
+	if (tmp == 0)
+		return -EINVAL;
+	if (tmp >= (u32)-1)
+		return -ERANGE;
+
+	*val = tmp;
+
+	return count;
+}
+
+static ssize_t o2nm_cluster_attr_idle_timeout_ms_read(
+	struct o2nm_cluster *cluster, char *page)
+{
+	return sprintf(page, "%u\n", cluster->cl_idle_timeout_ms);
+}
+
+static ssize_t o2nm_cluster_attr_idle_timeout_ms_write(
+	struct o2nm_cluster *cluster, const char *page, size_t count)
+{
+	ssize_t ret;
+	unsigned int val;
+
+	ret =  o2nm_cluster_attr_write(page, count, &val);
+
+	if (ret > 0) {
+		if (cluster->cl_idle_timeout_ms != val
+			&& o2net_num_connected_peers()) {
+			mlog(ML_NOTICE,
+			     "o2net: cannot change idle timeout after "
+			     "the first peer has agreed to it."
+			     "  %d connected peers\n",
+			     o2net_num_connected_peers());
+			ret = -EINVAL;
+		} else if (val <= cluster->cl_keepalive_delay_ms) {
+			mlog(ML_NOTICE, "o2net: idle timeout must be larger "
+			     "than keepalive delay\n");
+			ret = -EINVAL;
+		} else {
+			cluster->cl_idle_timeout_ms = val;
+		}
+	}
+
+	return ret;
+}
+
+static ssize_t o2nm_cluster_attr_keepalive_delay_ms_read(
+	struct o2nm_cluster *cluster, char *page)
+{
+	return sprintf(page, "%u\n", cluster->cl_keepalive_delay_ms);
+}
+
+static ssize_t o2nm_cluster_attr_keepalive_delay_ms_write(
+	struct o2nm_cluster *cluster, const char *page, size_t count)
+{
+	ssize_t ret;
+	unsigned int val;
+
+	ret =  o2nm_cluster_attr_write(page, count, &val);
+
+	if (ret > 0) {
+		if (cluster->cl_keepalive_delay_ms != val
+		    && o2net_num_connected_peers()) {
+			mlog(ML_NOTICE,
+			     "o2net: cannot change keepalive delay after"
+			     " the first peer has agreed to it."
+			     "  %d connected peers\n",
+			     o2net_num_connected_peers());
+			ret = -EINVAL;
+		} else if (val >= cluster->cl_idle_timeout_ms) {
+			mlog(ML_NOTICE, "o2net: keepalive delay must be "
+			     "smaller than idle timeout\n");
+			ret = -EINVAL;
+		} else {
+			cluster->cl_keepalive_delay_ms = val;
+		}
+	}
+
+	return ret;
+}
+
+static ssize_t o2nm_cluster_attr_reconnect_delay_ms_read(
+	struct o2nm_cluster *cluster, char *page)
+{
+	return sprintf(page, "%u\n", cluster->cl_reconnect_delay_ms);
+}
+
+static ssize_t o2nm_cluster_attr_reconnect_delay_ms_write(
+	struct o2nm_cluster *cluster, const char *page, size_t count)
+{
+	return o2nm_cluster_attr_write(page, count,
+	                               &cluster->cl_reconnect_delay_ms);
+}
+static struct o2nm_cluster_attribute o2nm_cluster_attr_idle_timeout_ms = {
+	.attr	= { .ca_owner = THIS_MODULE,
+		    .ca_name = "idle_timeout_ms",
+		    .ca_mode = S_IRUGO | S_IWUSR },
+	.show	= o2nm_cluster_attr_idle_timeout_ms_read,
+	.store	= o2nm_cluster_attr_idle_timeout_ms_write,
+};
+
+static struct o2nm_cluster_attribute o2nm_cluster_attr_keepalive_delay_ms = {
+	.attr	= { .ca_owner = THIS_MODULE,
+		    .ca_name = "keepalive_delay_ms",
+		    .ca_mode = S_IRUGO | S_IWUSR },
+	.show	= o2nm_cluster_attr_keepalive_delay_ms_read,
+	.store	= o2nm_cluster_attr_keepalive_delay_ms_write,
+};
+
+static struct o2nm_cluster_attribute o2nm_cluster_attr_reconnect_delay_ms = {
+	.attr	= { .ca_owner = THIS_MODULE,
+		    .ca_name = "reconnect_delay_ms",
+		    .ca_mode = S_IRUGO | S_IWUSR },
+	.show	= o2nm_cluster_attr_reconnect_delay_ms_read,
+	.store	= o2nm_cluster_attr_reconnect_delay_ms_write,
+};
+
+static struct configfs_attribute *o2nm_cluster_attrs[] = {
+	&o2nm_cluster_attr_idle_timeout_ms.attr,
+	&o2nm_cluster_attr_keepalive_delay_ms.attr,
+	&o2nm_cluster_attr_reconnect_delay_ms.attr,
+	NULL,
+};
+static ssize_t o2nm_cluster_show(struct config_item *item,
+                                 struct configfs_attribute *attr,
+                                 char *page)
+{
+	struct o2nm_cluster *cluster = to_o2nm_cluster(item);
+	struct o2nm_cluster_attribute *o2nm_cluster_attr =
+		container_of(attr, struct o2nm_cluster_attribute, attr);
+	ssize_t ret = 0;
+
+	if (o2nm_cluster_attr->show)
+		ret = o2nm_cluster_attr->show(cluster, page);
+	return ret;
+}
+
+static ssize_t o2nm_cluster_store(struct config_item *item,
+                                  struct configfs_attribute *attr,
+                                  const char *page, size_t count)
+{
+	struct o2nm_cluster *cluster = to_o2nm_cluster(item);
+	struct o2nm_cluster_attribute *o2nm_cluster_attr =
+		container_of(attr, struct o2nm_cluster_attribute, attr);
+	ssize_t ret;
+
+	if (o2nm_cluster_attr->store == NULL) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	ret = o2nm_cluster_attr->store(cluster, page, count);
+	if (ret < count)
+		goto out;
+out:
+	return ret;
+}
+
 static struct config_item *o2nm_node_group_make_item(struct config_group *group,
 						     const char *name)
 {
@@ -624,10 +786,13 @@ static void o2nm_cluster_release(struct config_item *item)
 
 static struct configfs_item_operations o2nm_cluster_item_ops = {
 	.release	= o2nm_cluster_release,
+	.show_attribute		= o2nm_cluster_show,
+	.store_attribute	= o2nm_cluster_store,
 };
 
 static struct config_item_type o2nm_cluster_type = {
 	.ct_item_ops	= &o2nm_cluster_item_ops,
+	.ct_attrs	= o2nm_cluster_attrs,
 	.ct_owner	= THIS_MODULE,
 };
 
@@ -678,6 +843,9 @@ static struct config_group *o2nm_cluster_group_make_group(struct config_group *g
 	cluster->cl_group.default_groups[2] = NULL;
 	rwlock_init(&cluster->cl_nodes_lock);
 	cluster->cl_node_ip_tree = RB_ROOT;
+	cluster->cl_reconnect_delay_ms = O2NET_RECONNECT_DELAY_MS_DEFAULT;
+	cluster->cl_idle_timeout_ms    = O2NET_IDLE_TIMEOUT_MS_DEFAULT;
+	cluster->cl_keepalive_delay_ms = O2NET_KEEPALIVE_DELAY_MS_DEFAULT;
 
 	ret = &cluster->cl_group;
 	o2nm_single_cluster = cluster;

fs/ocfs2/cluster/nodemanager.h

@@ -53,6 +53,23 @@ struct o2nm_node {
 	unsigned long		nd_set_attributes;
 };
 
+struct o2nm_cluster {
+	struct config_group	cl_group;
+	unsigned		cl_has_local:1;
+	u8			cl_local_node;
+	rwlock_t		cl_nodes_lock;
+	struct o2nm_node  	*cl_nodes[O2NM_MAX_NODES];
+	struct rb_root		cl_node_ip_tree;
+	unsigned int		cl_idle_timeout_ms;
+	unsigned int		cl_keepalive_delay_ms;
+	unsigned int		cl_reconnect_delay_ms;
+
+	/* this bitmap is part of a hack for disk bitmap.. will go eventually. - zab */
+	unsigned long	cl_nodes_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
+};
+
+extern struct o2nm_cluster *o2nm_single_cluster;
+
 u8 o2nm_this_node(void);
 
 int o2nm_configured_node_map(unsigned long *map, unsigned bytes);

fs/ocfs2/cluster/tcp.h

@@ -54,6 +54,13 @@ typedef int (o2net_msg_handler_func)(struct o2net_msg *msg, u32 len, void *data)
 
 #define O2NET_MAX_PAYLOAD_BYTES  (4096 - sizeof(struct o2net_msg))
 
+/* same as hb delay, we're waiting for another node to recognize our hb */
+#define O2NET_RECONNECT_DELAY_MS_DEFAULT	2000
+
+#define O2NET_KEEPALIVE_DELAY_MS_DEFAULT	5000
+#define O2NET_IDLE_TIMEOUT_MS_DEFAULT		10000
+
+
 /* TODO: figure this out.... */
 static inline int o2net_link_down(int err, struct socket *sock)
 {
@@ -101,6 +108,7 @@ void o2net_unregister_hb_callbacks(void);
 int o2net_start_listening(struct o2nm_node *node);
 void o2net_stop_listening(struct o2nm_node *node);
 void o2net_disconnect_node(struct o2nm_node *node);
+int o2net_num_connected_peers(void);
 
 int o2net_init(void);
 void o2net_exit(void);

fs/ocfs2/cluster/tcp_internal.h

@@ -27,23 +27,20 @@
 #define O2NET_MSG_KEEP_REQ_MAGIC  ((u16)0xfa57)
 #define O2NET_MSG_KEEP_RESP_MAGIC ((u16)0xfa58)
 
-/* same as hb delay, we're waiting for another node to recognize our hb */
-#define O2NET_RECONNECT_DELAY_MS	O2HB_REGION_TIMEOUT_MS
-
 /* we're delaying our quorum decision so that heartbeat will have timed
  * out truly dead nodes by the time we come around to making decisions
  * on their number */
 #define O2NET_QUORUM_DELAY_MS	((o2hb_dead_threshold + 2) * O2HB_REGION_TIMEOUT_MS)
 
-#define O2NET_KEEPALIVE_DELAY_SECS	5
-#define O2NET_IDLE_TIMEOUT_SECS		10
-
 /* 
  * This version number represents quite a lot, unfortunately.  It not
  * only represents the raw network message protocol on the wire but also
  * locking semantics of the file system using the protocol.  It should 
  * be somewhere else, I'm sure, but right now it isn't.
  *
+ * New in version 5:
+ * 	- Network timeout checking protocol
+ *
  * New in version 4:
  * 	- Remove i_generation from lock names for better stat performance.
  *
@@ -54,10 +51,14 @@
  * 	- full 64 bit i_size in the metadata lock lvbs
  * 	- introduction of "rw" lock and pushing meta/data locking down
  */
-#define O2NET_PROTOCOL_VERSION 4ULL
+#define O2NET_PROTOCOL_VERSION 5ULL
 struct o2net_handshake {
 	__be64	protocol_version;
 	__be64	connector_id;
+	__be32  o2hb_heartbeat_timeout_ms;
+	__be32  o2net_idle_timeout_ms;
+	__be32  o2net_keepalive_delay_ms;
+	__be32  o2net_reconnect_delay_ms;
 };
 
 struct o2net_node {

fs/ocfs2/dlmglue.c

@@ -770,7 +770,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb,
 			     int dlm_flags)
 {
 	int ret = 0;
-	enum dlm_status status;
+	enum dlm_status status = DLM_NORMAL;
 	unsigned long flags;
 
 	mlog_entry_void();
@@ -1138,6 +1138,7 @@ int ocfs2_rw_lock(struct inode *inode, int write)
 {
 	int status, level;
 	struct ocfs2_lock_res *lockres;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	BUG_ON(!inode);
 
@@ -1147,6 +1148,9 @@ int ocfs2_rw_lock(struct inode *inode, int write)
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
+	if (ocfs2_mount_local(osb))
+		return 0;
+
 	lockres = &OCFS2_I(inode)->ip_rw_lockres;
 
 	level = write ? LKM_EXMODE : LKM_PRMODE;
@@ -1164,6 +1168,7 @@ void ocfs2_rw_unlock(struct inode *inode, int write)
 {
 	int level = write ? LKM_EXMODE : LKM_PRMODE;
 	struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_rw_lockres;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	mlog_entry_void();
 
@@ -1171,7 +1176,8 @@ void ocfs2_rw_unlock(struct inode *inode, int write)
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
-	ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
+	if (!ocfs2_mount_local(osb))
+		ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
 
 	mlog_exit_void();
 }
@@ -1182,6 +1188,7 @@ int ocfs2_data_lock_full(struct inode *inode,
 {
 	int status = 0, level;
 	struct ocfs2_lock_res *lockres;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	BUG_ON(!inode);
 
@@ -1201,6 +1208,9 @@ int ocfs2_data_lock_full(struct inode *inode,
 		goto out;
 	}
 
+	if (ocfs2_mount_local(osb))
+		goto out;
+
 	lockres = &OCFS2_I(inode)->ip_data_lockres;
 
 	level = write ? LKM_EXMODE : LKM_PRMODE;
@@ -1269,6 +1279,7 @@ void ocfs2_data_unlock(struct inode *inode,
 {
 	int level = write ? LKM_EXMODE : LKM_PRMODE;
 	struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_data_lockres;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	mlog_entry_void();
 
@@ -1276,7 +1287,8 @@ void ocfs2_data_unlock(struct inode *inode,
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
-	if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)))
+	if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)) &&
+	    !ocfs2_mount_local(osb))
 		ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
 
 	mlog_exit_void();
@@ -1467,8 +1479,9 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 {
 	int status = 0;
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
-	struct ocfs2_lock_res *lockres;
+	struct ocfs2_lock_res *lockres = NULL;
 	struct ocfs2_dinode *fe;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	mlog_entry_void();
 
@@ -1483,10 +1496,12 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 	}
 	spin_unlock(&oi->ip_lock);
 
-	lockres = &oi->ip_meta_lockres;
+	if (!ocfs2_mount_local(osb)) {
+		lockres = &oi->ip_meta_lockres;
 
-	if (!ocfs2_should_refresh_lock_res(lockres))
-		goto bail;
+		if (!ocfs2_should_refresh_lock_res(lockres))
+			goto bail;
+	}
 
 	/* This will discard any caching information we might have had
 	 * for the inode metadata. */
@@ -1496,7 +1511,7 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 	 * map (directories, bitmap files, etc) */
 	ocfs2_extent_map_trunc(inode, 0);
 
-	if (ocfs2_meta_lvb_is_trustable(inode, lockres)) {
+	if (lockres && ocfs2_meta_lvb_is_trustable(inode, lockres)) {
 		mlog(0, "Trusting LVB on inode %llu\n",
 		     (unsigned long long)oi->ip_blkno);
 		ocfs2_refresh_inode_from_lvb(inode);
@@ -1543,7 +1558,8 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 
 	status = 0;
 bail_refresh:
-	ocfs2_complete_lock_res_refresh(lockres, status);
+	if (lockres)
+		ocfs2_complete_lock_res_refresh(lockres, status);
 bail:
 	mlog_exit(status);
 	return status;
@@ -1585,7 +1601,7 @@ int ocfs2_meta_lock_full(struct inode *inode,
 			 int arg_flags)
 {
 	int status, level, dlm_flags, acquired;
-	struct ocfs2_lock_res *lockres;
+	struct ocfs2_lock_res *lockres = NULL;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct buffer_head *local_bh = NULL;
 
@@ -1607,6 +1623,9 @@ int ocfs2_meta_lock_full(struct inode *inode,
 		goto bail;
 	}
 
+	if (ocfs2_mount_local(osb))
+		goto local;
+
 	if (!(arg_flags & OCFS2_META_LOCK_RECOVERY))
 		wait_event(osb->recovery_event,
 			   ocfs2_node_map_is_empty(osb, &osb->recovery_map));
@@ -1636,6 +1655,7 @@ int ocfs2_meta_lock_full(struct inode *inode,
 		wait_event(osb->recovery_event,
 			   ocfs2_node_map_is_empty(osb, &osb->recovery_map));
 
+local:
 	/*
 	 * We only see this flag if we're being called from
 	 * ocfs2_read_locked_inode(). It means we're locking an inode
@@ -1644,7 +1664,8 @@ int ocfs2_meta_lock_full(struct inode *inode,
 	 */
 	if (inode->i_state & I_NEW) {
 		status = 0;
-		ocfs2_complete_lock_res_refresh(lockres, 0);
+		if (lockres)
+			ocfs2_complete_lock_res_refresh(lockres, 0);
 		goto bail;
 	}
 
@@ -1767,6 +1788,7 @@ void ocfs2_meta_unlock(struct inode *inode,
 {
 	int level = ex ? LKM_EXMODE : LKM_PRMODE;
 	struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_meta_lockres;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	mlog_entry_void();
 
@@ -1774,7 +1796,8 @@ void ocfs2_meta_unlock(struct inode *inode,
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     ex ? "EXMODE" : "PRMODE");
 
-	if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)))
+	if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)) &&
+	    !ocfs2_mount_local(osb))
 		ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
 
 	mlog_exit_void();
@@ -1783,7 +1806,7 @@ void ocfs2_meta_unlock(struct inode *inode,
 int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex)
 {
-	int status;
+	int status = 0;
 	int level = ex ? LKM_EXMODE : LKM_PRMODE;
 	struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
 	struct buffer_head *bh;
@@ -1794,6 +1817,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
 	if (ocfs2_is_hard_readonly(osb))
 		return -EROFS;
 
+	if (ocfs2_mount_local(osb))
+		goto bail;
+
 	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1832,7 +1858,8 @@ void ocfs2_super_unlock(struct ocfs2_super *osb,
 	int level = ex ? LKM_EXMODE : LKM_PRMODE;
 	struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
 
-	ocfs2_cluster_unlock(osb, lockres, level);
+	if (!ocfs2_mount_local(osb))
+		ocfs2_cluster_unlock(osb, lockres, level);
 }
 
 int ocfs2_rename_lock(struct ocfs2_super *osb)
@@ -1843,6 +1870,9 @@ int ocfs2_rename_lock(struct ocfs2_super *osb)
 	if (ocfs2_is_hard_readonly(osb))
 		return -EROFS;
 
+	if (ocfs2_mount_local(osb))
+		return 0;
+
 	status = ocfs2_cluster_lock(osb, lockres, LKM_EXMODE, 0, 0);
 	if (status < 0)
 		mlog_errno(status);
@@ -1854,7 +1884,8 @@ void ocfs2_rename_unlock(struct ocfs2_super *osb)
 {
 	struct ocfs2_lock_res *lockres = &osb->osb_rename_lockres;
 
-	ocfs2_cluster_unlock(osb, lockres, LKM_EXMODE);
+	if (!ocfs2_mount_local(osb))
+		ocfs2_cluster_unlock(osb, lockres, LKM_EXMODE);
 }
 
 int ocfs2_dentry_lock(struct dentry *dentry, int ex)
@@ -1869,6 +1900,9 @@ int ocfs2_dentry_lock(struct dentry *dentry, int ex)
 	if (ocfs2_is_hard_readonly(osb))
 		return -EROFS;
 
+	if (ocfs2_mount_local(osb))
+		return 0;
+
 	ret = ocfs2_cluster_lock(osb, &dl->dl_lockres, level, 0, 0);
 	if (ret < 0)
 		mlog_errno(ret);
@@ -1882,7 +1916,8 @@ void ocfs2_dentry_unlock(struct dentry *dentry, int ex)
 	struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
 	struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
 
-	ocfs2_cluster_unlock(osb, &dl->dl_lockres, level);
+	if (!ocfs2_mount_local(osb))
+		ocfs2_cluster_unlock(osb, &dl->dl_lockres, level);
 }
 
 /* Reference counting of the dlm debug structure. We want this because
@@ -2145,12 +2180,15 @@ static void ocfs2_dlm_shutdown_debug(struct ocfs2_super *osb)
 
 int ocfs2_dlm_init(struct ocfs2_super *osb)
 {
-	int status;
+	int status = 0;
 	u32 dlm_key;
-	struct dlm_ctxt *dlm;
+	struct dlm_ctxt *dlm = NULL;
 
 	mlog_entry_void();
 
+	if (ocfs2_mount_local(osb))
+		goto local;
+
 	status = ocfs2_dlm_init_debug(osb);
 	if (status < 0) {
 		mlog_errno(status);
@@ -2178,11 +2216,12 @@ int ocfs2_dlm_init(struct ocfs2_super *osb)
 		goto bail;
 	}
 
+	dlm_register_eviction_cb(dlm, &osb->osb_eviction_cb);
+
+local:
 	ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
 	ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
 
-	dlm_register_eviction_cb(dlm, &osb->osb_eviction_cb);
-
 	osb->dlm = dlm;
 
 	status = 0;

fs/ocfs2/heartbeat.c

@@ -154,6 +154,9 @@ int ocfs2_register_hb_callbacks(struct ocfs2_super *osb)
 {
 	int status;
 
+	if (ocfs2_mount_local(osb))
+		return 0;
+
 	status = o2hb_register_callback(&osb->osb_hb_down);
 	if (status < 0) {
 		mlog_errno(status);
@@ -172,6 +175,9 @@ void ocfs2_clear_hb_callbacks(struct ocfs2_super *osb)
 {
 	int status;
 
+	if (ocfs2_mount_local(osb))
+		return;
+
 	status = o2hb_unregister_callback(&osb->osb_hb_down);
 	if (status < 0)
 		mlog_errno(status);
@@ -186,6 +192,9 @@ void ocfs2_stop_heartbeat(struct ocfs2_super *osb)
 	int ret;
 	char *argv[5], *envp[3];
 
+	if (ocfs2_mount_local(osb))
+		return;
+
 	if (!osb->uuid_str) {
 		/* This can happen if we don't get far enough in mount... */
 		mlog(0, "No UUID with which to stop heartbeat!\n\n");

fs/ocfs2/inode.c

@@ -423,7 +423,8 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 	 * cluster lock before trusting anything anyway.
 	 */
 	can_lock = !(args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
-		&& !(args->fi_flags & OCFS2_FI_FLAG_NOLOCK);
+		&& !(args->fi_flags & OCFS2_FI_FLAG_NOLOCK)
+		&& !ocfs2_mount_local(osb);
 
 	/*
 	 * To maintain backwards compatibility with older versions of

fs/ocfs2/journal.c

@@ -144,8 +144,10 @@ handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs)
 			ocfs2_abort(osb->sb, "Detected aborted journal");
 			handle = ERR_PTR(-EROFS);
 		}
-	} else
-		atomic_inc(&(osb->journal->j_num_trans));
+	} else {
+		if (!ocfs2_mount_local(osb))
+			atomic_inc(&(osb->journal->j_num_trans));
+	}
 
 	return handle;
 }
@@ -507,9 +509,23 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb)
 
 	BUG_ON(atomic_read(&(osb->journal->j_num_trans)) != 0);
 
-	status = ocfs2_journal_toggle_dirty(osb, 0);
-	if (status < 0)
-		mlog_errno(status);
+	if (ocfs2_mount_local(osb)) {
+		journal_lock_updates(journal->j_journal);
+		status = journal_flush(journal->j_journal);
+		journal_unlock_updates(journal->j_journal);
+		if (status < 0)
+			mlog_errno(status);
+	}
+
+	if (status == 0) {
+		/*
+		 * Do not toggle if flush was unsuccessful otherwise
+		 * will leave dirty metadata in a "clean" journal
+		 */
+		status = ocfs2_journal_toggle_dirty(osb, 0);
+		if (status < 0)
+			mlog_errno(status);
+	}
 
 	/* Shutdown the kernel journal system */
 	journal_destroy(journal->j_journal);
@@ -549,7 +565,7 @@ static void ocfs2_clear_journal_error(struct super_block *sb,
 	}
 }
 
-int ocfs2_journal_load(struct ocfs2_journal *journal)
+int ocfs2_journal_load(struct ocfs2_journal *journal, int local)
 {
 	int status = 0;
 	struct ocfs2_super *osb;
@@ -576,14 +592,18 @@ int ocfs2_journal_load(struct ocfs2_journal *journal)
 	}
 
 	/* Launch the commit thread */
-	osb->commit_task = kthread_run(ocfs2_commit_thread, osb, "ocfs2cmt");
-	if (IS_ERR(osb->commit_task)) {
-		status = PTR_ERR(osb->commit_task);
+	if (!local) {
+		osb->commit_task = kthread_run(ocfs2_commit_thread, osb,
+					       "ocfs2cmt");
+		if (IS_ERR(osb->commit_task)) {
+			status = PTR_ERR(osb->commit_task);
+			osb->commit_task = NULL;
+			mlog(ML_ERROR, "unable to launch ocfs2commit thread, "
+			     "error=%d", status);
+			goto done;
+		}
+	} else
 		osb->commit_task = NULL;
-		mlog(ML_ERROR, "unable to launch ocfs2commit thread, error=%d",
-		     status);
-		goto done;
-	}
 
 done:
 	mlog_exit(status);

fs/ocfs2/journal.h

@@ -157,7 +157,7 @@ int    ocfs2_journal_init(struct ocfs2_journal *journal,
 void   ocfs2_journal_shutdown(struct ocfs2_super *osb);
 int    ocfs2_journal_wipe(struct ocfs2_journal *journal,
 			  int full);
-int    ocfs2_journal_load(struct ocfs2_journal *journal);
+int    ocfs2_journal_load(struct ocfs2_journal *journal, int local);
 int    ocfs2_check_journals_nolocks(struct ocfs2_super *osb);
 void   ocfs2_recovery_thread(struct ocfs2_super *osb,
 			     int node_num);
@@ -174,6 +174,9 @@ static inline void ocfs2_checkpoint_inode(struct inode *inode)
 {
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
+	if (ocfs2_mount_local(osb))
+		return;
+
 	if (!ocfs2_inode_fully_checkpointed(inode)) {
 		/* WARNING: This only kicks off a single
 		 * checkpoint. If someone races you and adds more

fs/ocfs2/mmap.c

@@ -83,10 +83,12 @@ static struct vm_operations_struct ocfs2_file_vm_ops = {
 int ocfs2_mmap(struct file *file, struct vm_area_struct *vma)
 {
 	int ret = 0, lock_level = 0;
+	struct ocfs2_super *osb = OCFS2_SB(file->f_dentry->d_inode->i_sb);
 
 	/* We don't want to support shared writable mappings yet. */
-	if (((vma->vm_flags & VM_SHARED) || (vma->vm_flags & VM_MAYSHARE))
-	    && ((vma->vm_flags & VM_WRITE) || (vma->vm_flags & VM_MAYWRITE))) {
+	if (!ocfs2_mount_local(osb) &&
+	    ((vma->vm_flags & VM_SHARED) || (vma->vm_flags & VM_MAYSHARE)) &&
+	    ((vma->vm_flags & VM_WRITE) || (vma->vm_flags & VM_MAYWRITE))) {
 		mlog(0, "disallow shared writable mmaps %lx\n", vma->vm_flags);
 		/* This is -EINVAL because generic_file_readonly_mmap
 		 * returns it in a similar situation. */

fs/ocfs2/namei.c

@@ -587,9 +587,11 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	}
 
 	ocfs2_inode_set_new(osb, inode);
-	status = ocfs2_create_new_inode_locks(inode);
-	if (status < 0)
-		mlog_errno(status);
+	if (!ocfs2_mount_local(osb)) {
+		status = ocfs2_create_new_inode_locks(inode);
+		if (status < 0)
+			mlog_errno(status);
+	}
 
 	status = 0; /* error in ocfs2_create_new_inode_locks is not
 		     * critical */

fs/ocfs2/ocfs2.h

@@ -349,6 +349,11 @@ static inline int ocfs2_is_soft_readonly(struct ocfs2_super *osb)
 	return ret;
 }
 
+static inline int ocfs2_mount_local(struct ocfs2_super *osb)
+{
+	return (osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT);
+}
+
 #define OCFS2_IS_VALID_DINODE(ptr)					\
 	(!strcmp((ptr)->i_signature, OCFS2_INODE_SIGNATURE))
 

fs/ocfs2/ocfs2_fs.h

@@ -86,7 +86,7 @@
 	OCFS2_SB(sb)->s_feature_incompat &= ~(mask)
 
 #define OCFS2_FEATURE_COMPAT_SUPP	0
-#define OCFS2_FEATURE_INCOMPAT_SUPP	0
+#define OCFS2_FEATURE_INCOMPAT_SUPP	OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT
 #define OCFS2_FEATURE_RO_COMPAT_SUPP	0
 
 /*
@@ -96,6 +96,18 @@
  */
 #define OCFS2_FEATURE_INCOMPAT_HEARTBEAT_DEV	0x0002
 
+/*
+ * tunefs sets this incompat flag before starting the resize and clears it
+ * at the end. This flag protects users from inadvertently mounting the fs
+ * after an aborted run without fsck-ing.
+ */
+#define OCFS2_FEATURE_INCOMPAT_RESIZE_INPROG    0x0004
+
+/* Used to denote a non-clustered volume */
+#define OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT	0x0008
+
+/* Support for sparse allocation in b-trees */
+#define OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC	0x0010
 
 /*
  * Flags on ocfs2_dinode.i_flags

fs/ocfs2/super.c

@@ -508,6 +508,27 @@ static int ocfs2_sb_probe(struct super_block *sb,
 	return status;
 }
 
+static int ocfs2_verify_heartbeat(struct ocfs2_super *osb)
+{
+	if (ocfs2_mount_local(osb)) {
+		if (osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) {
+			mlog(ML_ERROR, "Cannot heartbeat on a locally "
+			     "mounted device.\n");
+			return -EINVAL;
+		}
+	}
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL)) {
+		if (!ocfs2_mount_local(osb) && !ocfs2_is_hard_readonly(osb)) {
+			mlog(ML_ERROR, "Heartbeat has to be started to mount "
+			     "a read-write clustered device.\n");
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
 static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 {
 	struct dentry *root;
@@ -516,16 +537,24 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	struct inode *inode = NULL;
 	struct ocfs2_super *osb = NULL;
 	struct buffer_head *bh = NULL;
+	char nodestr[8];
 
 	mlog_entry("%p, %p, %i", sb, data, silent);
 
-	/* for now we only have one cluster/node, make sure we see it
-	 * in the heartbeat universe */
-	if (!o2hb_check_local_node_heartbeating()) {
+	if (!ocfs2_parse_options(sb, data, &parsed_opt, 0)) {
 		status = -EINVAL;
 		goto read_super_error;
 	}
 
+	/* for now we only have one cluster/node, make sure we see it
+	 * in the heartbeat universe */
+	if (parsed_opt & OCFS2_MOUNT_HB_LOCAL) {
+		if (!o2hb_check_local_node_heartbeating()) {
+			status = -EINVAL;
+			goto read_super_error;
+		}
+	}
+
 	/* probe for superblock */
 	status = ocfs2_sb_probe(sb, &bh, &sector_size);
 	if (status < 0) {
@@ -541,11 +570,6 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	}
 	brelse(bh);
 	bh = NULL;
-
-	if (!ocfs2_parse_options(sb, data, &parsed_opt, 0)) {
-		status = -EINVAL;
-		goto read_super_error;
-	}
 	osb->s_mount_opt = parsed_opt;
 
 	sb->s_magic = OCFS2_SUPER_MAGIC;
@@ -588,21 +612,16 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	}
 
 	if (!ocfs2_is_hard_readonly(osb)) {
-		/* If this isn't a hard readonly mount, then we need
-		 * to make sure that heartbeat is in a valid state,
-		 * and that we mark ourselves soft readonly is -oro
-		 * was specified. */
-		if (!(osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL)) {
-			mlog(ML_ERROR, "No heartbeat for device (%s)\n",
-			     sb->s_id);
-			status = -EINVAL;
-			goto read_super_error;
-		}
-
 		if (sb->s_flags & MS_RDONLY)
 			ocfs2_set_ro_flag(osb, 0);
 	}
 
+	status = ocfs2_verify_heartbeat(osb);
+	if (status < 0) {
+		mlog_errno(status);
+		goto read_super_error;
+	}
+
 	osb->osb_debug_root = debugfs_create_dir(osb->uuid_str,
 						 ocfs2_debugfs_root);
 	if (!osb->osb_debug_root) {
@@ -635,9 +654,14 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 
 	ocfs2_complete_mount_recovery(osb);
 
-	printk(KERN_INFO "ocfs2: Mounting device (%s) on (node %d, slot %d) "
+	if (ocfs2_mount_local(osb))
+		snprintf(nodestr, sizeof(nodestr), "local");
+	else
+		snprintf(nodestr, sizeof(nodestr), "%d", osb->node_num);
+
+	printk(KERN_INFO "ocfs2: Mounting device (%s) on (node %s, slot %d) "
 	       "with %s data mode.\n",
-	       osb->dev_str, osb->node_num, osb->slot_num,
+	       osb->dev_str, nodestr, osb->slot_num,
 	       osb->s_mount_opt & OCFS2_MOUNT_DATA_WRITEBACK ? "writeback" :
 	       "ordered");
 
@@ -999,7 +1023,11 @@ static int ocfs2_fill_local_node_info(struct ocfs2_super *osb)
 
 	/* XXX hold a ref on the node while mounte?  easy enough, if
 	 * desirable. */
-	osb->node_num = o2nm_this_node();
+	if (ocfs2_mount_local(osb))
+		osb->node_num = 0;
+	else
+		osb->node_num = o2nm_this_node();
+
 	if (osb->node_num == O2NM_MAX_NODES) {
 		mlog(ML_ERROR, "could not find this host's node number\n");
 		status = -ENOENT;
@@ -1084,6 +1112,9 @@ static int ocfs2_mount_volume(struct super_block *sb)
 		goto leave;
 	}
 
+	if (ocfs2_mount_local(osb))
+		goto leave;
+
 	/* This should be sent *after* we recovered our journal as it
 	 * will cause other nodes to unmark us as needing
 	 * recovery. However, we need to send it *before* dropping the
@@ -1114,6 +1145,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 {
 	int tmp;
 	struct ocfs2_super *osb = NULL;
+	char nodestr[8];
 
 	mlog_entry("(0x%p)\n", sb);
 
@@ -1177,8 +1209,13 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	atomic_set(&osb->vol_state, VOLUME_DISMOUNTED);
 
-	printk(KERN_INFO "ocfs2: Unmounting device (%s) on (node %d)\n",
-	       osb->dev_str, osb->node_num);
+	if (ocfs2_mount_local(osb))
+		snprintf(nodestr, sizeof(nodestr), "local");
+	else
+		snprintf(nodestr, sizeof(nodestr), "%d", osb->node_num);
+
+	printk(KERN_INFO "ocfs2: Unmounting device (%s) on (node %s)\n",
+	       osb->dev_str, nodestr);
 
 	ocfs2_delete_osb(osb);
 	kfree(osb);
@@ -1536,6 +1573,7 @@ static int ocfs2_check_volume(struct ocfs2_super *osb)
 {
 	int status = 0;
 	int dirty;
+	int local;
 	struct ocfs2_dinode *local_alloc = NULL; /* only used if we
 						  * recover
 						  * ourselves. */
@@ -1563,8 +1601,10 @@ static int ocfs2_check_volume(struct ocfs2_super *osb)
 		     "recovering volume.\n");
 	}
 
+	local = ocfs2_mount_local(osb);
+
 	/* will play back anything left in the journal. */
-	ocfs2_journal_load(osb->journal);
+	ocfs2_journal_load(osb->journal, local);
 
 	if (dirty) {
 		/* recover my local alloc if we didn't unmount cleanly. */

fs/ocfs2/vote.c

@@ -1000,6 +1000,9 @@ int ocfs2_register_net_handlers(struct ocfs2_super *osb)
 {
 	int status = 0;
 
+	if (ocfs2_mount_local(osb))
+		return 0;
+
 	status = o2net_register_handler(OCFS2_MESSAGE_TYPE_RESPONSE,
 					osb->net_key,
 					sizeof(struct ocfs2_response_msg),