Merge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2:
  configfs: Fix race between configfs_readdir() and configfs_d_iput()
  configfs: Don't try to d_delete() negative dentries.
  ocfs2/dlm: Target node death during resource migration leads to thread spin
  ocfs2: Skip mount recovery for hard-ro mounts
  ocfs2/cluster: Heartbeat mismatch message improved
  ocfs2/cluster: Increase the live threshold for global heartbeat
  ocfs2/dlm: Use negotiated o2dlm protocol version
  ocfs2: skip existing hole when removing the last extent_rec in punching-hole codes.
  ocfs2: Initialize data_ac (might be used uninitialized)

7 files changed, 96 insertions, 27 deletions

diff --git a/fs/configfs/dir.c b/fs/configfs/dir.c
index 3313dd19f543..9a37a9b6de3a 100644
--- a/fs/configfs/dir.c
+++ b/fs/configfs/dir.c
@@ -53,11 +53,14 @@ DEFINE_SPINLOCK(configfs_dirent_lock);
 static void configfs_d_iput(struct dentry * dentry,
                             struct inode * inode)
 {
-        struct configfs_dirent * sd = dentry->d_fsdata;
+        struct configfs_dirent *sd = dentry->d_fsdata;
 
         if (sd) {
                 BUG_ON(sd->s_dentry != dentry);
+                /* Coordinate with configfs_readdir */
+                spin_lock(&configfs_dirent_lock);
                 sd->s_dentry = NULL;
+                spin_unlock(&configfs_dirent_lock);
                 configfs_put(sd);
         }
         iput(inode);
@@ -689,7 +692,8 @@ static int create_default_group(struct config_group *parent_group,
                         sd = child->d_fsdata;
                         sd->s_type |= CONFIGFS_USET_DEFAULT;
                 } else {
-                        d_delete(child);
+                        BUG_ON(child->d_inode);
+                        d_drop(child);
                         dput(child);
                 }
         }
@@ -1545,7 +1549,7 @@ static int configfs_readdir(struct file * filp, void * dirent, filldir_t filldir
         struct configfs_dirent * parent_sd = dentry->d_fsdata;
         struct configfs_dirent *cursor = filp->private_data;
         struct list_head *p, *q = &cursor->s_sibling;
-        ino_t ino;
+        ino_t ino = 0;
         int i = filp->f_pos;
 
         switch (i) {
@@ -1573,6 +1577,7 @@ static int configfs_readdir(struct file * filp, void * dirent, filldir_t filldir
                                 struct configfs_dirent *next;
                                 const char * name;
                                 int len;
+                                struct inode *inode = NULL;
 
                                 next = list_entry(p, struct configfs_dirent,
                                                    s_sibling);
@@ -1581,9 +1586,28 @@ static int configfs_readdir(struct file * filp, void * dirent, filldir_t filldir
 
                                 name = configfs_get_name(next);
                                 len = strlen(name);
-                                if (next->s_dentry)
-                                        ino = next->s_dentry->d_inode->i_ino;
-                                else
+
+                                /*
+                                 * We'll have a dentry and an inode for
+                                 * PINNED items and for open attribute
+                                 * files.  We lock here to prevent a race
+                                 * with configfs_d_iput() clearing
+                                 * s_dentry before calling iput().
+                                 *
+                                 * Why do we go to the trouble?  If
+                                 * someone has an attribute file open,
+                                 * the inode number should match until
+                                 * they close it.  Beyond that, we don't
+                                 * care.
+                                 */
+                                spin_lock(&configfs_dirent_lock);
+                                dentry = next->s_dentry;
+                                if (dentry)
+                                        inode = dentry->d_inode;
+                                if (inode)
+                                        ino = inode->i_ino;
+                                spin_unlock(&configfs_dirent_lock);
+                                if (!inode)
                                         ino = iunique(configfs_sb, 2);
 
                                 if (filldir(dirent, name, len, filp->f_pos, ino,
@@ -1683,7 +1707,8 @@ int configfs_register_subsystem(struct configfs_subsystem *subsys)
                 err = configfs_attach_group(sd->s_element, &group->cg_item,
                                             dentry);
                 if (err) {
-                        d_delete(dentry);
+                        BUG_ON(dentry->d_inode);
+                        d_drop(dentry);
                         dput(dentry);
                 } else {
                         spin_lock(&configfs_dirent_lock);
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index 643720209a98..9a3e6bbff27b 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -539,25 +539,41 @@ static int o2hb_verify_crc(struct o2hb_region *reg,
 
 /* We want to make sure that nobody is heartbeating on top of us --
  * this will help detect an invalid configuration. */
-static int o2hb_check_last_timestamp(struct o2hb_region *reg)
+static void o2hb_check_last_timestamp(struct o2hb_region *reg)
 {
-        int node_num, ret;
         struct o2hb_disk_slot *slot;
         struct o2hb_disk_heartbeat_block *hb_block;
+        char *errstr;
 
-        node_num = o2nm_this_node();
-
-        ret = 1;
-        slot = &reg->hr_slots[node_num];
+        slot = &reg->hr_slots[o2nm_this_node()];
         /* Don't check on our 1st timestamp */
-        if (slot->ds_last_time) {
-                hb_block = slot->ds_raw_block;
+        if (!slot->ds_last_time)
+                return;
 
-                if (le64_to_cpu(hb_block->hb_seq) != slot->ds_last_time)
-                        ret = 0;
-        }
+        hb_block = slot->ds_raw_block;
+        if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time &&
+            le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation &&
+            hb_block->hb_node == slot->ds_node_num)
+                return;
 
-        return ret;
+#define ERRSTR1         "Another node is heartbeating on device"
+#define ERRSTR2         "Heartbeat generation mismatch on device"
+#define ERRSTR3         "Heartbeat sequence mismatch on device"
+
+        if (hb_block->hb_node != slot->ds_node_num)
+                errstr = ERRSTR1;
+        else if (le64_to_cpu(hb_block->hb_generation) !=
+                 slot->ds_last_generation)
+                errstr = ERRSTR2;
+        else
+                errstr = ERRSTR3;
+
+        mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), "
+             "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name,
+             slot->ds_node_num, (unsigned long long)slot->ds_last_generation,
+             (unsigned long long)slot->ds_last_time, hb_block->hb_node,
+             (unsigned long long)le64_to_cpu(hb_block->hb_generation),
+             (unsigned long long)le64_to_cpu(hb_block->hb_seq));
 }
 
 static inline void o2hb_prepare_block(struct o2hb_region *reg,
@@ -983,9 +999,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
         /* With an up to date view of the slots, we can check that no
          * other node has been improperly configured to heartbeat in
          * our slot. */
-        if (!o2hb_check_last_timestamp(reg))
-                mlog(ML_ERROR, "Device \"%s\": another node is heartbeating "
-                     "in our slot!\n", reg->hr_dev_name);
+        o2hb_check_last_timestamp(reg);
 
         /* fill in the proper info for our next heartbeat */
         o2hb_prepare_block(reg, reg->hr_generation);
@@ -999,8 +1013,8 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
         }
 
         i = -1;
-        while((i = find_next_bit(configured_nodes, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
-
+        while((i = find_next_bit(configured_nodes,
+                                 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
                 change |= o2hb_check_slot(reg, &reg->hr_slots[i]);
         }
 
@@ -1690,6 +1704,7 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
         struct file *filp = NULL;
         struct inode *inode = NULL;
         ssize_t ret = -EINVAL;
+        int live_threshold;
 
         if (reg->hr_bdev)
                 goto out;
@@ -1766,8 +1781,18 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
          * A node is considered live after it has beat LIVE_THRESHOLD
          * times.  We're not steady until we've given them a chance
          * _after_ our first read.
+         * The default threshold is bare minimum so as to limit the delay
+         * during mounts. For global heartbeat, the threshold doubled for the
+         * first region.
          */
-        atomic_set(&reg->hr_steady_iterations, O2HB_LIVE_THRESHOLD + 1);
+        live_threshold = O2HB_LIVE_THRESHOLD;
+        if (o2hb_global_heartbeat_active()) {
+                spin_lock(&o2hb_live_lock);
+                if (o2hb_pop_count(&o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1)
+                        live_threshold <<= 1;
+                spin_unlock(&o2hb_live_lock);
+        }
+        atomic_set(&reg->hr_steady_iterations, live_threshold + 1);
 
         hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s",
                               reg->hr_item.ci_name);
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 9fe5b8fd658f..8582e3f4f120 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -2868,7 +2868,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
                 bytes = blocks_wanted << sb->s_blocksize_bits;
         struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
         struct ocfs2_inode_info *oi = OCFS2_I(dir);
-        struct ocfs2_alloc_context *data_ac;
+        struct ocfs2_alloc_context *data_ac = NULL;
         struct ocfs2_alloc_context *meta_ac = NULL;
         struct buffer_head *dirdata_bh = NULL;
         struct buffer_head *dx_root_bh = NULL;
diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c
index 7540a492eaba..3b179d6cbde0 100644
--- a/fs/ocfs2/dlm/dlmdomain.c
+++ b/fs/ocfs2/dlm/dlmdomain.c
@@ -1614,7 +1614,8 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm)
         spin_unlock(&dlm->spinlock);
 
         /* Support for global heartbeat and node info was added in 1.1 */
-        if (dlm_protocol.pv_major > 1 || dlm_protocol.pv_minor > 0) {
+        if (dlm->dlm_locking_proto.pv_major > 1 ||
+            dlm->dlm_locking_proto.pv_minor > 0) {
                 status = dlm_send_nodeinfo(dlm, ctxt->yes_resp_map);
                 if (status) {
                         mlog_errno(status);
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index fede57ed005f..84d166328cf7 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -2574,6 +2574,9 @@ fail:
                 res->state &= ~DLM_LOCK_RES_MIGRATING;
                 wake = 1;
                 spin_unlock(&res->spinlock);
+                if (dlm_is_host_down(ret))
+                        dlm_wait_for_node_death(dlm, target,
+                                                DLM_NODE_DEATH_WAIT_MAX);
                 goto leave;
         }
 
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 41565ae52856..89659d6dc206 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -1607,6 +1607,9 @@ static void ocfs2_calc_trunc_pos(struct inode *inode,
         range = le32_to_cpu(rec->e_cpos) + ocfs2_rec_clusters(el, rec);
 
         if (le32_to_cpu(rec->e_cpos) >= trunc_start) {
+                /*
+                 * remove an entire extent record.
+                 */
                 *trunc_cpos = le32_to_cpu(rec->e_cpos);
                 /*
                  * Skip holes if any.
@@ -1617,7 +1620,16 @@ static void ocfs2_calc_trunc_pos(struct inode *inode,
                 *blkno = le64_to_cpu(rec->e_blkno);
                 *trunc_end = le32_to_cpu(rec->e_cpos);
         } else if (range > trunc_start) {
+                /*
+                 * remove a partial extent record, which means we're
+                 * removing the last extent record.
+                 */
                 *trunc_cpos = trunc_start;
+                /*
+                 * skip hole if any.
+                 */
+                if (range < *trunc_end)
+                        *trunc_end = range;
                 *trunc_len = *trunc_end - trunc_start;
                 coff = trunc_start - le32_to_cpu(rec->e_cpos);
                 *blkno = le64_to_cpu(rec->e_blkno) +
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index b141a44605ca..295d56454e8b 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -1260,6 +1260,9 @@ void ocfs2_complete_mount_recovery(struct ocfs2_super *osb)
 {
         struct ocfs2_journal *journal = osb->journal;
 
+        if (ocfs2_is_hard_readonly(osb))
+                return;
+
         /* No need to queue up our truncate_log as regular cleanup will catch
          * that */
         ocfs2_queue_recovery_completion(journal, osb->slot_num,