5 files changed, 151 insertions, 19 deletions

diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index 979113479c66..2bd7f788cf34 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -1335,6 +1335,7 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
         ret = wait_event_interruptible(o2hb_steady_queue,
                                 atomic_read(&reg->hr_steady_iterations) == 0);
         if (ret) {
+                /* We got interrupted (hello ptrace!).  Clean up */
                 spin_lock(&o2hb_live_lock);
                 hb_task = reg->hr_task;
                 reg->hr_task = NULL;
@@ -1345,7 +1346,16 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
                 goto out;
         }
 
-        ret = count;
+        /* Ok, we were woken.  Make sure it wasn't by drop_item() */
+        spin_lock(&o2hb_live_lock);
+        hb_task = reg->hr_task;
+        spin_unlock(&o2hb_live_lock);
+
+        if (hb_task)
+                ret = count;
+        else
+                ret = -EIO;
+
 out:
         if (filp)
                 fput(filp);
@@ -1523,6 +1533,15 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group,
         if (hb_task)
                 kthread_stop(hb_task);
 
+        /*
+         * If we're racing a dev_write(), we need to wake them.  They will
+         * check reg->hr_task
+         */
+        if (atomic_read(&reg->hr_steady_iterations) != 0) {
+                atomic_set(&reg->hr_steady_iterations, 0);
+                wake_up(&o2hb_steady_queue);
+        }
+
         config_item_put(item);
 }
 
@@ -1665,7 +1684,67 @@ void o2hb_setup_callback(struct o2hb_callback_func *hc,
 }
 EXPORT_SYMBOL_GPL(o2hb_setup_callback);
 
-int o2hb_register_callback(struct o2hb_callback_func *hc)
+static struct o2hb_region *o2hb_find_region(const char *region_uuid)
+{
+        struct o2hb_region *p, *reg = NULL;
+
+        assert_spin_locked(&o2hb_live_lock);
+
+        list_for_each_entry(p, &o2hb_all_regions, hr_all_item) {
+                if (!strcmp(region_uuid, config_item_name(&p->hr_item))) {
+                        reg = p;
+                        break;
+                }
+        }
+
+        return reg;
+}
+
+static int o2hb_region_get(const char *region_uuid)
+{
+        int ret = 0;
+        struct o2hb_region *reg;
+
+        spin_lock(&o2hb_live_lock);
+
+        reg = o2hb_find_region(region_uuid);
+        if (!reg)
+                ret = -ENOENT;
+        spin_unlock(&o2hb_live_lock);
+
+        if (ret)
+                goto out;
+
+        ret = o2nm_depend_this_node();
+        if (ret)
+                goto out;
+
+        ret = o2nm_depend_item(&reg->hr_item);
+        if (ret)
+                o2nm_undepend_this_node();
+
+out:
+        return ret;
+}
+
+static void o2hb_region_put(const char *region_uuid)
+{
+        struct o2hb_region *reg;
+
+        spin_lock(&o2hb_live_lock);
+
+        reg = o2hb_find_region(region_uuid);
+
+        spin_unlock(&o2hb_live_lock);
+
+        if (reg) {
+                o2nm_undepend_item(&reg->hr_item);
+                o2nm_undepend_this_node();
+        }
+}
+
+int o2hb_register_callback(const char *region_uuid,
+                           struct o2hb_callback_func *hc)
 {
         struct o2hb_callback_func *tmp;
         struct list_head *iter;
@@ -1681,6 +1760,12 @@ int o2hb_register_callback(struct o2hb_callback_func *hc)
                 goto out;
         }
 
+        if (region_uuid) {
+                ret = o2hb_region_get(region_uuid);
+                if (ret)
+                        goto out;
+        }
+
         down_write(&o2hb_callback_sem);
 
         list_for_each(iter, &hbcall->list) {
@@ -1702,16 +1787,21 @@ out:
 }
 EXPORT_SYMBOL_GPL(o2hb_register_callback);
 
-void o2hb_unregister_callback(struct o2hb_callback_func *hc)
+void o2hb_unregister_callback(const char *region_uuid,
+                              struct o2hb_callback_func *hc)
 {
         BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
 
         mlog(ML_HEARTBEAT, "on behalf of %p for funcs %p\n",
              __builtin_return_address(0), hc);
 
+        /* XXX Can this happen _with_ a region reference? */
         if (list_empty(&hc->hc_item))
                 return;
 
+        if (region_uuid)
+                o2hb_region_put(region_uuid);
+
         down_write(&o2hb_callback_sem);
 
         list_del_init(&hc->hc_item);
diff --git a/fs/ocfs2/cluster/heartbeat.h b/fs/ocfs2/cluster/heartbeat.h
index cc6d40b39771..35397dd5ecdb 100644
--- a/fs/ocfs2/cluster/heartbeat.h
+++ b/fs/ocfs2/cluster/heartbeat.h
@@ -69,8 +69,10 @@ void o2hb_setup_callback(struct o2hb_callback_func *hc,
                          o2hb_cb_func *func,
                          void *data,
                          int priority);
-int o2hb_register_callback(struct o2hb_callback_func *hc);
-void o2hb_unregister_callback(struct o2hb_callback_func *hc);
+int o2hb_register_callback(const char *region_uuid,
+                           struct o2hb_callback_func *hc);
+void o2hb_unregister_callback(const char *region_uuid,
+                              struct o2hb_callback_func *hc);
 void o2hb_fill_node_map(unsigned long *map,
                         unsigned bytes);
 void o2hb_init(void);
diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c
index 9f5ad0f01ce0..af2070da308b 100644
--- a/fs/ocfs2/cluster/nodemanager.c
+++ b/fs/ocfs2/cluster/nodemanager.c
@@ -900,6 +900,46 @@ static struct o2nm_cluster_group o2nm_cluster_group = {
         },
 };
 
+int o2nm_depend_item(struct config_item *item)
+{
+        return configfs_depend_item(&o2nm_cluster_group.cs_subsys, item);
+}
+
+void o2nm_undepend_item(struct config_item *item)
+{
+        configfs_undepend_item(&o2nm_cluster_group.cs_subsys, item);
+}
+
+int o2nm_depend_this_node(void)
+{
+        int ret = 0;
+        struct o2nm_node *local_node;
+
+        local_node = o2nm_get_node_by_num(o2nm_this_node());
+        if (!local_node) {
+                ret = -EINVAL;
+                goto out;
+        }
+
+        ret = o2nm_depend_item(&local_node->nd_item);
+        o2nm_node_put(local_node);
+
+out:
+        return ret;
+}
+
+void o2nm_undepend_this_node(void)
+{
+        struct o2nm_node *local_node;
+
+        local_node = o2nm_get_node_by_num(o2nm_this_node());
+        BUG_ON(!local_node);
+
+        o2nm_undepend_item(&local_node->nd_item);
+        o2nm_node_put(local_node);
+}
+
+
 static void __exit exit_o2nm(void)
 {
         if (ocfs2_table_header)
@@ -934,7 +974,7 @@ static int __init init_o2nm(void)
                 goto out_sysctl;
 
         config_group_init(&o2nm_cluster_group.cs_subsys.su_group);
-        init_MUTEX(&o2nm_cluster_group.cs_subsys.su_sem);
+        mutex_init(&o2nm_cluster_group.cs_subsys.su_mutex);
         ret = configfs_register_subsystem(&o2nm_cluster_group.cs_subsys);
         if (ret) {
                 printk(KERN_ERR "nodemanager: Registration returned %d\n", ret);
diff --git a/fs/ocfs2/cluster/nodemanager.h b/fs/ocfs2/cluster/nodemanager.h
index 070522138ae2..7c860361b8dd 100644
--- a/fs/ocfs2/cluster/nodemanager.h
+++ b/fs/ocfs2/cluster/nodemanager.h
@@ -77,4 +77,9 @@ struct o2nm_node *o2nm_get_node_by_ip(__be32 addr);
 void o2nm_node_get(struct o2nm_node *node);
 void o2nm_node_put(struct o2nm_node *node);
 
+int o2nm_depend_item(struct config_item *item);
+void o2nm_undepend_item(struct config_item *item);
+int o2nm_depend_this_node(void);
+void o2nm_undepend_this_node(void);
+
 #endif /* O2CLUSTER_NODEMANAGER_H */
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 0b229a9c7952..f0bdfd944c44 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -261,14 +261,12 @@ out:
 
 static void o2net_complete_nodes_nsw(struct o2net_node *nn)
 {
-        struct list_head *iter, *tmp;
+        struct o2net_status_wait *nsw, *tmp;
         unsigned int num_kills = 0;
-        struct o2net_status_wait *nsw;
 
         assert_spin_locked(&nn->nn_lock);
 
-        list_for_each_safe(iter, tmp, &nn->nn_status_list) {
-                nsw = list_entry(iter, struct o2net_status_wait, ns_node_item);
+        list_for_each_entry_safe(nsw, tmp, &nn->nn_status_list, ns_node_item) {
                 o2net_complete_nsw_locked(nn, nsw, O2NET_ERR_DIED, 0);
                 num_kills++;
         }
@@ -764,13 +762,10 @@ EXPORT_SYMBOL_GPL(o2net_register_handler);
 
 void o2net_unregister_handler_list(struct list_head *list)
 {
-        struct list_head *pos, *n;
-        struct o2net_msg_handler *nmh;
+        struct o2net_msg_handler *nmh, *n;
 
         write_lock(&o2net_handler_lock);
-        list_for_each_safe(pos, n, list) {
-                nmh = list_entry(pos, struct o2net_msg_handler,
-                                 nh_unregister_item);
+        list_for_each_entry_safe(nmh, n, list, nh_unregister_item) {
                 mlog(ML_TCP, "unregistering handler func %p type %u key %08x\n",
                      nmh->nh_func, nmh->nh_msg_type, nmh->nh_key);
                 rb_erase(&nmh->nh_node, &o2net_handler_tree);
@@ -1638,8 +1633,8 @@ static void o2net_hb_node_up_cb(struct o2nm_node *node, int node_num,
 
 void o2net_unregister_hb_callbacks(void)
 {
-        o2hb_unregister_callback(&o2net_hb_up);
-        o2hb_unregister_callback(&o2net_hb_down);
+        o2hb_unregister_callback(NULL, &o2net_hb_up);
+        o2hb_unregister_callback(NULL, &o2net_hb_down);
 }
 
 int o2net_register_hb_callbacks(void)
@@ -1651,9 +1646,9 @@ int o2net_register_hb_callbacks(void)
         o2hb_setup_callback(&o2net_hb_up, O2HB_NODE_UP_CB,
                             o2net_hb_node_up_cb, NULL, O2NET_HB_PRI);
 
-        ret = o2hb_register_callback(&o2net_hb_up);
+        ret = o2hb_register_callback(NULL, &o2net_hb_up);
         if (ret == 0)
-                ret = o2hb_register_callback(&o2net_hb_down);
+                ret = o2hb_register_callback(NULL, &o2net_hb_down);
 
         if (ret)
                 o2net_unregister_hb_callbacks();