Merge branch 'wip-2.6.34' into old-private-masterarchived-private-master

70 files changed, 3092 insertions, 2260 deletions

diff --git a/fs/ocfs2/Kconfig b/fs/ocfs2/Kconfig
index 701b7a3a872e..0d840669698e 100644
--- a/fs/ocfs2/Kconfig
+++ b/fs/ocfs2/Kconfig
@@ -6,6 +6,7 @@ config OCFS2_FS
         select CRC32
         select QUOTA
         select QUOTA_TREE
+        select FS_POSIX_ACL
         help
           OCFS2 is a general purpose extent based shared disk cluster file
           system with many similarities to ext3. It supports 64 bit inode
@@ -74,12 +75,3 @@ config OCFS2_DEBUG_FS
           This option will enable expensive consistency checks. Enable
           this option for debugging only as it is likely to decrease
           performance of the filesystem.
-
-config OCFS2_FS_POSIX_ACL
-        bool "OCFS2 POSIX Access Control Lists"
-        depends on OCFS2_FS
-        select FS_POSIX_ACL
-        default n
-        help
-          Posix Access Control Lists (ACLs) support permissions for users and
-          groups beyond the owner/group/world scheme.
diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile
index 31f25ce32c97..791c0886c060 100644
--- a/fs/ocfs2/Makefile
+++ b/fs/ocfs2/Makefile
@@ -39,16 +39,14 @@ ocfs2-objs := \
         ver.o                   \
         quota_local.o           \
         quota_global.o          \
-        xattr.o
-
-ifeq ($(CONFIG_OCFS2_FS_POSIX_ACL),y)
-ocfs2-objs += acl.o
-endif
+        xattr.o                 \
+        acl.o
 
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
 ocfs2_stack_user-objs := stack_user.o
 
+obj-$(CONFIG_OCFS2_FS) += dlmfs/
 # cluster/ is always needed when OCFS2_FS for masklog support
 obj-$(CONFIG_OCFS2_FS) += cluster/
 obj-$(CONFIG_OCFS2_FS_O2CB) += dlm/
diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c
index fbeaec762103..e13fc9e8fcdc 100644
--- a/fs/ocfs2/acl.c
+++ b/fs/ocfs2/acl.c
@@ -21,6 +21,7 @@
 
 #include <linux/init.h>
 #include <linux/module.h>
+#include <linux/slab.h>
 #include <linux/string.h>
 
 #define MLOG_MASK_PREFIX ML_INODE
@@ -30,6 +31,8 @@
 #include "alloc.h"
 #include "dlmglue.h"
 #include "file.h"
+#include "inode.h"
+#include "journal.h"
 #include "ocfs2_fs.h"
 
 #include "xattr.h"
@@ -98,15 +101,11 @@ static struct posix_acl *ocfs2_get_acl_nolock(struct inode *inode,
                                               int type,
                                               struct buffer_head *di_bh)
 {
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
         int name_index;
         char *value = NULL;
         struct posix_acl *acl;
         int retval;
 
-        if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
-                return NULL;
-
         switch (type) {
         case ACL_TYPE_ACCESS:
                 name_index = OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS;
@@ -170,6 +169,60 @@ static struct posix_acl *ocfs2_get_acl(struct inode *inode, int type)
 }
 
 /*
+ * Helper function to set i_mode in memory and disk. Some call paths
+ * will not have di_bh or a journal handle to pass, in which case it
+ * will create it's own.
+ */
+static int ocfs2_acl_set_mode(struct inode *inode, struct buffer_head *di_bh,
+                              handle_t *handle, umode_t new_mode)
+{
+        int ret, commit_handle = 0;
+        struct ocfs2_dinode *di;
+
+        if (di_bh == NULL) {
+                ret = ocfs2_read_inode_block(inode, &di_bh);
+                if (ret) {
+                        mlog_errno(ret);
+                        goto out;
+                }
+        } else
+                get_bh(di_bh);
+
+        if (handle == NULL) {
+                handle = ocfs2_start_trans(OCFS2_SB(inode->i_sb),
+                                           OCFS2_INODE_UPDATE_CREDITS);
+                if (IS_ERR(handle)) {
+                        ret = PTR_ERR(handle);
+                        mlog_errno(ret);
+                        goto out_brelse;
+                }
+
+                commit_handle = 1;
+        }
+
+        di = (struct ocfs2_dinode *)di_bh->b_data;
+        ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
+                                      OCFS2_JOURNAL_ACCESS_WRITE);
+        if (ret) {
+                mlog_errno(ret);
+                goto out_commit;
+        }
+
+        inode->i_mode = new_mode;
+        di->i_mode = cpu_to_le16(inode->i_mode);
+
+        ocfs2_journal_dirty(handle, di_bh);
+
+out_commit:
+        if (commit_handle)
+                ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
+out_brelse:
+        brelse(di_bh);
+out:
+        return ret;
+}
+
+/*
  * Set the access or default ACL of an inode.
  */
 static int ocfs2_set_acl(handle_t *handle,
@@ -197,9 +250,14 @@ static int ocfs2_set_acl(handle_t *handle,
                         if (ret < 0)
                                 return ret;
                         else {
-                                inode->i_mode = mode;
                                 if (ret == 0)
                                         acl = NULL;
+
+                                ret = ocfs2_acl_set_mode(inode, di_bh,
+                                                         handle, mode);
+                                if (ret)
+                                        return ret;
+
                         }
                 }
                 break;
@@ -287,6 +345,7 @@ int ocfs2_init_acl(handle_t *handle,
         struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
         struct posix_acl *acl = NULL;
         int ret = 0;
+        mode_t mode;
 
         if (!S_ISLNK(inode->i_mode)) {
                 if (osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) {
@@ -295,12 +354,17 @@ int ocfs2_init_acl(handle_t *handle,
                         if (IS_ERR(acl))
                                 return PTR_ERR(acl);
                 }
-                if (!acl)
-                        inode->i_mode &= ~current_umask();
+                if (!acl) {
+                        mode = inode->i_mode & ~current_umask();
+                        ret = ocfs2_acl_set_mode(inode, di_bh, handle, mode);
+                        if (ret) {
+                                mlog_errno(ret);
+                                goto cleanup;
+                        }
+                }
         }
         if ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) && acl) {
                 struct posix_acl *clone;
-                mode_t mode;
 
                 if (S_ISDIR(inode->i_mode)) {
                         ret = ocfs2_set_acl(handle, inode, di_bh,
@@ -317,7 +381,7 @@ int ocfs2_init_acl(handle_t *handle,
                 mode = inode->i_mode;
                 ret = posix_acl_create_masq(clone, &mode);
                 if (ret >= 0) {
-                        inode->i_mode = mode;
+                        ret = ocfs2_acl_set_mode(inode, di_bh, handle, mode);
                         if (ret > 0) {
                                 ret = ocfs2_set_acl(handle, inode,
                                                     di_bh, ACL_TYPE_ACCESS,
@@ -331,13 +395,14 @@ cleanup:
         return ret;
 }
 
-static size_t ocfs2_xattr_list_acl_access(struct inode *inode,
+static size_t ocfs2_xattr_list_acl_access(struct dentry *dentry,
                                           char *list,
                                           size_t list_len,
                                           const char *name,
-                                          size_t name_len)
+                                          size_t name_len,
+                                          int type)
 {
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
         const size_t size = sizeof(POSIX_ACL_XATTR_ACCESS);
 
         if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
@@ -348,13 +413,14 @@ static size_t ocfs2_xattr_list_acl_access(struct inode *inode,
         return size;
 }
 
-static size_t ocfs2_xattr_list_acl_default(struct inode *inode,
+static size_t ocfs2_xattr_list_acl_default(struct dentry *dentry,
                                            char *list,
                                            size_t list_len,
                                            const char *name,
-                                           size_t name_len)
+                                           size_t name_len,
+                                           int type)
 {
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
         const size_t size = sizeof(POSIX_ACL_XATTR_DEFAULT);
 
         if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
@@ -365,19 +431,19 @@ static size_t ocfs2_xattr_list_acl_default(struct inode *inode,
         return size;
 }
 
-static int ocfs2_xattr_get_acl(struct inode *inode,
-                               int type,
-                               void *buffer,
-                               size_t size)
+static int ocfs2_xattr_get_acl(struct dentry *dentry, const char *name,
+                void *buffer, size_t size, int type)
 {
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
         struct posix_acl *acl;
         int ret;
 
+        if (strcmp(name, "") != 0)
+                return -EINVAL;
         if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
                 return -EOPNOTSUPP;
 
-        acl = ocfs2_get_acl(inode, type);
+        acl = ocfs2_get_acl(dentry->d_inode, type);
         if (IS_ERR(acl))
                 return PTR_ERR(acl);
         if (acl == NULL)
@@ -388,35 +454,16 @@ static int ocfs2_xattr_get_acl(struct inode *inode,
         return ret;
 }
 
-static int ocfs2_xattr_get_acl_access(struct inode *inode,
-                                      const char *name,
-                                      void *buffer,
-                                      size_t size)
-{
-        if (strcmp(name, "") != 0)
-                return -EINVAL;
-        return ocfs2_xattr_get_acl(inode, ACL_TYPE_ACCESS, buffer, size);
-}
-
-static int ocfs2_xattr_get_acl_default(struct inode *inode,
-                                       const char *name,
-                                       void *buffer,
-                                       size_t size)
-{
-        if (strcmp(name, "") != 0)
-                return -EINVAL;
-        return ocfs2_xattr_get_acl(inode, ACL_TYPE_DEFAULT, buffer, size);
-}
-
-static int ocfs2_xattr_set_acl(struct inode *inode,
-                               int type,
-                               const void *value,
-                               size_t size)
+static int ocfs2_xattr_set_acl(struct dentry *dentry, const char *name,
+                const void *value, size_t size, int flags, int type)
 {
+        struct inode *inode = dentry->d_inode;
         struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
         struct posix_acl *acl;
         int ret = 0;
 
+        if (strcmp(name, "") != 0)
+                return -EINVAL;
         if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
                 return -EOPNOTSUPP;
 
@@ -442,38 +489,18 @@ cleanup:
         return ret;
 }
 
-static int ocfs2_xattr_set_acl_access(struct inode *inode,
-                                      const char *name,
-                                      const void *value,
-                                      size_t size,
-                                      int flags)
-{
-        if (strcmp(name, "") != 0)
-                return -EINVAL;
-        return ocfs2_xattr_set_acl(inode, ACL_TYPE_ACCESS, value, size);
-}
-
-static int ocfs2_xattr_set_acl_default(struct inode *inode,
-                                       const char *name,
-                                       const void *value,
-                                       size_t size,
-                                       int flags)
-{
-        if (strcmp(name, "") != 0)
-                return -EINVAL;
-        return ocfs2_xattr_set_acl(inode, ACL_TYPE_DEFAULT, value, size);
-}
-
 struct xattr_handler ocfs2_xattr_acl_access_handler = {
         .prefix = POSIX_ACL_XATTR_ACCESS,
+        .flags  = ACL_TYPE_ACCESS,
         .list   = ocfs2_xattr_list_acl_access,
-        .get    = ocfs2_xattr_get_acl_access,
-        .set    = ocfs2_xattr_set_acl_access,
+        .get    = ocfs2_xattr_get_acl,
+        .set    = ocfs2_xattr_set_acl,
 };
 
 struct xattr_handler ocfs2_xattr_acl_default_handler = {
         .prefix = POSIX_ACL_XATTR_DEFAULT,
+        .flags  = ACL_TYPE_DEFAULT,
         .list   = ocfs2_xattr_list_acl_default,
-        .get    = ocfs2_xattr_get_acl_default,
-        .set    = ocfs2_xattr_set_acl_default,
+        .get    = ocfs2_xattr_get_acl,
+        .set    = ocfs2_xattr_set_acl,
 };
diff --git a/fs/ocfs2/acl.h b/fs/ocfs2/acl.h
index 8f6389ed4da5..5c5d31f05853 100644
--- a/fs/ocfs2/acl.h
+++ b/fs/ocfs2/acl.h
@@ -26,8 +26,6 @@ struct ocfs2_acl_entry {
         __le32 e_id;
 };
 
-#ifdef CONFIG_OCFS2_FS_POSIX_ACL
-
 extern int ocfs2_check_acl(struct inode *, int);
 extern int ocfs2_acl_chmod(struct inode *);
 extern int ocfs2_init_acl(handle_t *, struct inode *, struct inode *,
@@ -35,24 +33,4 @@ extern int ocfs2_init_acl(handle_t *, struct inode *, struct inode *,
                           struct ocfs2_alloc_context *,
                           struct ocfs2_alloc_context *);
 
-#else /* CONFIG_OCFS2_FS_POSIX_ACL*/
-
-#define ocfs2_check_acl NULL
-static inline int ocfs2_acl_chmod(struct inode *inode)
-{
-        return 0;
-}
-static inline int ocfs2_init_acl(handle_t *handle,
-                                 struct inode *inode,
-                                 struct inode *dir,
-                                 struct buffer_head *di_bh,
-                                 struct buffer_head *dir_bh,
-                                 struct ocfs2_alloc_context *meta_ac,
-                                 struct ocfs2_alloc_context *data_ac)
-{
-        return 0;
-}
-
-#endif /* CONFIG_OCFS2_FS_POSIX_ACL*/
-
 #endif /* OCFS2_ACL_H */
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 38a42f5d59ff..9f8bd913c51e 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -1050,7 +1050,8 @@ static int ocfs2_create_new_meta_bhs(handle_t *handle,
                         strcpy(eb->h_signature, OCFS2_EXTENT_BLOCK_SIGNATURE);
                         eb->h_blkno = cpu_to_le64(first_blkno);
                         eb->h_fs_generation = cpu_to_le32(osb->fs_generation);
-                        eb->h_suballoc_slot = cpu_to_le16(osb->slot_num);
+                        eb->h_suballoc_slot =
+                                cpu_to_le16(meta_ac->ac_alloc_slot);
                         eb->h_suballoc_bit = cpu_to_le16(suballoc_bit_start);
                         eb->h_list.l_count =
                                 cpu_to_le16(ocfs2_extent_recs_per_eb(osb->sb));
@@ -1765,9 +1766,9 @@ set_and_inc:
  *
  * The array index of the subtree root is passed back.
  */
-static int ocfs2_find_subtree_root(struct ocfs2_extent_tree *et,
-                                   struct ocfs2_path *left,
-                                   struct ocfs2_path *right)
+int ocfs2_find_subtree_root(struct ocfs2_extent_tree *et,
+                            struct ocfs2_path *left,
+                            struct ocfs2_path *right)
 {
         int i = 0;
 
@@ -2398,7 +2399,7 @@ static int ocfs2_leftmost_rec_contains(struct ocfs2_extent_list *el, u32 cpos)
  *
  * The array is assumed to be large enough to hold an entire path (tree depth).
  *
- * Upon succesful return from this function:
+ * Upon successful return from this function:
  *
  * - The 'right_path' array will contain a path to the leaf block
  *   whose range contains e_cpos.
@@ -2872,8 +2873,8 @@ out:
  * This looks similar, but is subtly different to
  * ocfs2_find_cpos_for_left_leaf().
  */
-static int ocfs2_find_cpos_for_right_leaf(struct super_block *sb,
-                                          struct ocfs2_path *path, u32 *cpos)
+int ocfs2_find_cpos_for_right_leaf(struct super_block *sb,
+                                   struct ocfs2_path *path, u32 *cpos)
 {
         int i, j, ret = 0;
         u64 blkno;
@@ -5712,7 +5713,7 @@ int ocfs2_remove_btree_range(struct inode *inode,
                 goto out;
         }
 
-        vfs_dq_free_space_nodirty(inode,
+        dquot_free_space_nodirty(inode,
                                   ocfs2_clusters_to_bytes(inode->i_sb, len));
 
         ret = ocfs2_remove_extent(handle, et, cpos, len, meta_ac, dealloc);
@@ -6037,7 +6038,7 @@ static void ocfs2_truncate_log_worker(struct work_struct *work)
         if (status < 0)
                 mlog_errno(status);
         else
-                ocfs2_init_inode_steal_slot(osb);
+                ocfs2_init_steal_slots(osb);
 
         mlog_exit(status);
 }
@@ -6935,7 +6936,7 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
                 goto bail;
         }
 
-        vfs_dq_free_space_nodirty(inode,
+        dquot_free_space_nodirty(inode,
                         ocfs2_clusters_to_bytes(osb->sb, clusters_to_del));
         spin_lock(&OCFS2_I(inode)->ip_lock);
         OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters) -
@@ -7190,8 +7191,8 @@ int ocfs2_zero_range_for_truncate(struct inode *inode, handle_t *handle,
          * wait on them - the truncate_inode_pages() call later will
          * do that for us.
          */
-        ret = do_sync_mapping_range(inode->i_mapping, range_start,
-                                    range_end - 1, SYNC_FILE_RANGE_WRITE);
+        ret = filemap_fdatawrite_range(inode->i_mapping, range_start,
+                                       range_end - 1);
         if (ret)
                 mlog_errno(ret);
 
@@ -7300,11 +7301,10 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
                 unsigned int page_end;
                 u64 phys;
 
-                if (vfs_dq_alloc_space_nodirty(inode,
-                                       ocfs2_clusters_to_bytes(osb->sb, 1))) {
-                        ret = -EDQUOT;
+                ret = dquot_alloc_space_nodirty(inode,
+                                       ocfs2_clusters_to_bytes(osb->sb, 1));
+                if (ret)
                         goto out_commit;
-                }
                 did_quota = 1;
 
                 ret = ocfs2_claim_clusters(osb, handle, data_ac, 1, &bit_off,
@@ -7380,7 +7380,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 
 out_commit:
         if (ret < 0 && did_quota)
-                vfs_dq_free_space_nodirty(inode,
+                dquot_free_space_nodirty(inode,
                                           ocfs2_clusters_to_bytes(osb->sb, 1));
 
         ocfs2_commit_trans(osb, handle);
diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h
index 9c122d574464..1db4359ccb90 100644
--- a/fs/ocfs2/alloc.h
+++ b/fs/ocfs2/alloc.h
@@ -317,4 +317,9 @@ int ocfs2_path_bh_journal_access(handle_t *handle,
 int ocfs2_journal_access_path(struct ocfs2_caching_info *ci,
                               handle_t *handle,
                               struct ocfs2_path *path);
+int ocfs2_find_cpos_for_right_leaf(struct super_block *sb,
+                                   struct ocfs2_path *path, u32 *cpos);
+int ocfs2_find_subtree_root(struct ocfs2_extent_tree *et,
+                            struct ocfs2_path *left,
+                            struct ocfs2_path *right);
 #endif /* OCFS2_ALLOC_H */
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index deb2b132ae5e..21441ddb5506 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -547,6 +547,9 @@ bail:
  *
  * called like this: dio->get_blocks(dio->inode, fs_startblk,
  *                                      fs_count, map_bh, dio->rw == WRITE);
+ *
+ * Note that we never bother to allocate blocks here, and thus ignore the
+ * create argument.
  */
 static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
                                      struct buffer_head *bh_result, int create)
@@ -563,14 +566,6 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
 
         inode_blocks = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
 
-        /*
-         * Any write past EOF is not allowed because we'd be extending.
-         */
-        if (create && (iblock + max_blocks) > inode_blocks) {
-                ret = -EIO;
-                goto bail;
-        }
-
         /* This figures out the size of the next contiguous block, and
          * our logical offset */
         ret = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno,
@@ -582,17 +577,9 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
                 goto bail;
         }
 
-        if (!ocfs2_sparse_alloc(OCFS2_SB(inode->i_sb)) && !p_blkno && create) {
-                ocfs2_error(inode->i_sb,
-                            "Inode %llu has a hole at block %llu\n",
-                            (unsigned long long)OCFS2_I(inode)->ip_blkno,
-                            (unsigned long long)iblock);
-                ret = -EROFS;
-                goto bail;
-        }
+        /* We should already CoW the refcounted extent in case of create. */
+        BUG_ON(create && (ext_flags & OCFS2_EXT_REFCOUNTED));
 
-        /* We should already CoW the refcounted extent. */
-        BUG_ON(ext_flags & OCFS2_EXT_REFCOUNTED);
         /*
          * get_more_blocks() expects us to describe a hole by clearing
          * the mapped bit on bh_result().
@@ -601,20 +588,8 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
          */
         if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
                 map_bh(bh_result, inode->i_sb, p_blkno);
-        else {
-                /*
-                 * ocfs2_prepare_inode_for_write() should have caught
-                 * the case where we'd be filling a hole and triggered
-                 * a buffered write instead.
-                 */
-                if (create) {
-                        ret = -EIO;
-                        mlog_errno(ret);
-                        goto bail;
-                }
-
+        else
                 clear_buffer_mapped(bh_result);
-        }
 
         /* make sure we don't map more than max_blocks blocks here as
            that's all the kernel will handle at this point. */
@@ -625,7 +600,7 @@ bail:
         return ret;
 }
 
-/* 
+/*
  * ocfs2_dio_end_io is called by the dio core when a dio is finished.  We're
  * particularly interested in the aio/dio case.  Like the core uses
  * i_alloc_sem, we use the rw_lock DLM lock to protect io on one node from
@@ -696,7 +671,7 @@ static ssize_t ocfs2_direct_IO(int rw,
 
         ret = blockdev_direct_IO_no_locking(rw, iocb, inode,
                                             inode->i_sb->s_bdev, iov, offset,
-                                            nr_segs, 
+                                            nr_segs,
                                             ocfs2_direct_IO_get_blocks,
                                             ocfs2_dio_end_io);
 
@@ -1789,10 +1764,11 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 
         wc->w_handle = handle;
 
-        if (clusters_to_alloc && vfs_dq_alloc_space_nodirty(inode,
-                        ocfs2_clusters_to_bytes(osb->sb, clusters_to_alloc))) {
-                ret = -EDQUOT;
-                goto out_commit;
+        if (clusters_to_alloc) {
+                ret = dquot_alloc_space_nodirty(inode,
+                        ocfs2_clusters_to_bytes(osb->sb, clusters_to_alloc));
+                if (ret)
+                        goto out_commit;
         }
         /*
          * We don't want this to fail in ocfs2_write_end(), so do it
@@ -1835,7 +1811,7 @@ success:
         return 0;
 out_quota:
         if (clusters_to_alloc)
-                vfs_dq_free_space(inode,
+                dquot_free_space(inode,
                           ocfs2_clusters_to_bytes(osb->sb, clusters_to_alloc));
 out_commit:
         ocfs2_commit_trans(osb, handle);
diff --git a/fs/ocfs2/blockcheck.c b/fs/ocfs2/blockcheck.c
index a1163b8b417c..b7428c5d0d3b 100644
--- a/fs/ocfs2/blockcheck.c
+++ b/fs/ocfs2/blockcheck.c
@@ -47,7 +47,7 @@
  * Calculate the bit offset in the hamming code buffer based on the bit's
  * offset in the data buffer.  Since the hamming code reserves all
  * power-of-two bits for parity, the data bit number and the code bit
- * number are offest by all the parity bits beforehand.
+ * number are offset by all the parity bits beforehand.
  *
  * Recall that bit numbers in hamming code are 1-based.  This function
  * takes the 0-based data bit from the caller.
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index d43d34a1dd31..f9d5d3ffc75a 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -25,7 +25,6 @@
 
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 
 #include <cluster/masklog.h>
@@ -368,7 +367,7 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
         }
         ocfs2_metadata_cache_io_unlock(ci);
 
-        mlog(ML_BH_IO, "block=(%llu), nr=(%d), cached=%s, flags=0x%x\n", 
+        mlog(ML_BH_IO, "block=(%llu), nr=(%d), cached=%s, flags=0x%x\n",
              (unsigned long long)block, nr,
              ((flags & OCFS2_BH_IGNORE_CACHE) || ignore_cache) ? "no" : "yes",
              flags);
@@ -407,6 +406,7 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
                                 struct buffer_head *bh)
 {
         int ret = 0;
+        struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
 
         mlog_entry_void();
 
@@ -426,6 +426,7 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
 
         get_bh(bh); /* for end_buffer_write_sync() */
         bh->b_end_io = end_buffer_write_sync;
+        ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &di->i_check);
         submit_bh(WRITE, bh);
 
         wait_on_buffer(bh);
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index c452d116b892..41d5f1f92d56 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -34,6 +34,7 @@
 #include <linux/crc32.h>
 #include <linux/time.h>
 #include <linux/debugfs.h>
+#include <linux/slab.h>
 
 #include "heartbeat.h"
 #include "tcp.h"
@@ -78,7 +79,7 @@ static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type);
 
 unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD;
 
-/* Only sets a new threshold if there are no active regions. 
+/* Only sets a new threshold if there are no active regions.
  *
  * No locking or otherwise interesting code is required for reading
  * o2hb_dead_threshold as it can't change once regions are active and
@@ -170,13 +171,14 @@ static void o2hb_write_timeout(struct work_struct *work)
 
         mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u "
              "milliseconds\n", reg->hr_dev_name,
-             jiffies_to_msecs(jiffies - reg->hr_last_timeout_start)); 
+             jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
         o2quo_disk_timeout();
 }
 
 static void o2hb_arm_write_timeout(struct o2hb_region *reg)
 {
-        mlog(0, "Queue write timeout for %u ms\n", O2HB_MAX_WRITE_TIMEOUT_MS);
+        mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
+             O2HB_MAX_WRITE_TIMEOUT_MS);
 
         cancel_delayed_work(&reg->hr_write_timeout_work);
         reg->hr_last_timeout_start = jiffies;
@@ -623,7 +625,7 @@ static int o2hb_check_slot(struct o2hb_region *reg,
              "seq %llu last %llu changed %u equal %u\n",
              slot->ds_node_num, (long long)slot->ds_last_generation,
              le32_to_cpu(hb_block->hb_cksum),
-             (unsigned long long)le64_to_cpu(hb_block->hb_seq), 
+             (unsigned long long)le64_to_cpu(hb_block->hb_seq),
              (unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
              slot->ds_equal_samples);
 
@@ -874,7 +876,8 @@ static int o2hb_thread(void *data)
                 do_gettimeofday(&after_hb);
                 elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb);
 
-                mlog(0, "start = %lu.%lu, end = %lu.%lu, msec = %u\n",
+                mlog(ML_HEARTBEAT,
+                     "start = %lu.%lu, end = %lu.%lu, msec = %u\n",
                      before_hb.tv_sec, (unsigned long) before_hb.tv_usec,
                      after_hb.tv_sec, (unsigned long) after_hb.tv_usec,
                      elapsed_msec);
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 1cd2934de615..3bb928a2bf7d 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -112,6 +112,7 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
         define_mask(XATTR),
         define_mask(QUOTA),
         define_mask(REFCOUNT),
+        define_mask(BASTS),
         define_mask(ERROR),
         define_mask(NOTICE),
         define_mask(KTHREAD),
@@ -135,7 +136,7 @@ static ssize_t mlog_store(struct kobject *obj, struct attribute *attr,
         return mlog_mask_store(mlog_attr->mask, buf, count);
 }
 
-static struct sysfs_ops mlog_attr_ops = {
+static const struct sysfs_ops mlog_attr_ops = {
         .show  = mlog_show,
         .store = mlog_store,
 };
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 9b4d11726cf2..3dfddbec32f2 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -114,6 +114,7 @@
 #define ML_XATTR        0x0000000020000000ULL /* ocfs2 extended attributes */
 #define ML_QUOTA        0x0000000040000000ULL /* ocfs2 quota operations */
 #define ML_REFCOUNT     0x0000000080000000ULL /* refcount tree operations */
+#define ML_BASTS        0x0000001000000000ULL /* dlmglue asts and basts */
 /* bits that are infrequently given and frequently matched in the high word */
 #define ML_ERROR        0x0000000100000000ULL /* sent to KERN_ERR */
 #define ML_NOTICE       0x0000000200000000ULL /* setn to KERN_NOTICE */
@@ -194,9 +195,9 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits;
  * previous token if args expands to nothing.
  */
 #define __mlog_printk(level, fmt, args...)                              \
-        printk(level "(%u,%lu):%s:%d " fmt, task_pid_nr(current),       \
-               __mlog_cpu_guess, __PRETTY_FUNCTION__, __LINE__ ,        \
-               ##args)
+        printk(level "(%s,%u,%lu):%s:%d " fmt, current->comm,           \
+               task_pid_nr(current), __mlog_cpu_guess,                  \
+               __PRETTY_FUNCTION__, __LINE__ , ##args)
 
 #define mlog(mask, fmt, args...) do {                                   \
         u64 __m = MLOG_MASK_PREFIX | (mask);                            \
diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c
index da794bc07a6c..a3f150e52b02 100644
--- a/fs/ocfs2/cluster/netdebug.c
+++ b/fs/ocfs2/cluster/netdebug.c
@@ -294,10 +294,10 @@ static int sc_seq_show(struct seq_file *seq, void *v)
                 if (sc->sc_sock) {
                         inet = inet_sk(sc->sc_sock->sk);
                         /* the stack's structs aren't sparse endian clean */
-                        saddr = (__force __be32)inet->saddr;
-                        daddr = (__force __be32)inet->daddr;
-                        sport = (__force __be16)inet->sport;
-                        dport = (__force __be16)inet->dport;
+                        saddr = (__force __be32)inet->inet_saddr;
+                        daddr = (__force __be32)inet->inet_daddr;
+                        sport = (__force __be16)inet->inet_sport;
+                        dport = (__force __be16)inet->inet_dport;
                 }
 
                 /* XXX sigh, inet-> doesn't have sparse annotation so any
diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c
index 7ee6188bc79a..ed0c9f367fed 100644
--- a/fs/ocfs2/cluster/nodemanager.c
+++ b/fs/ocfs2/cluster/nodemanager.c
@@ -19,6 +19,7 @@
  * Boston, MA 021110-1307, USA.
  */
 
+#include <linux/slab.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/configfs.h>
@@ -35,6 +36,10 @@
  * cluster references throughout where nodes are looked up */
 struct o2nm_cluster *o2nm_single_cluster = NULL;
 
+char *o2nm_fence_method_desc[O2NM_FENCE_METHODS] = {
+                "reset",        /* O2NM_FENCE_RESET */
+                "panic",        /* O2NM_FENCE_PANIC */
+};
 
 struct o2nm_node *o2nm_get_node_by_num(u8 node_num)
 {
@@ -579,6 +584,43 @@ static ssize_t o2nm_cluster_attr_reconnect_delay_ms_write(
         return o2nm_cluster_attr_write(page, count,
                                        &cluster->cl_reconnect_delay_ms);
 }
+
+static ssize_t o2nm_cluster_attr_fence_method_read(
+        struct o2nm_cluster *cluster, char *page)
+{
+        ssize_t ret = 0;
+
+        if (cluster)
+                ret = sprintf(page, "%s\n",
+                              o2nm_fence_method_desc[cluster->cl_fence_method]);
+        return ret;
+}
+
+static ssize_t o2nm_cluster_attr_fence_method_write(
+        struct o2nm_cluster *cluster, const char *page, size_t count)
+{
+        unsigned int i;
+
+        if (page[count - 1] != '\n')
+                goto bail;
+
+        for (i = 0; i < O2NM_FENCE_METHODS; ++i) {
+                if (count != strlen(o2nm_fence_method_desc[i]) + 1)
+                        continue;
+                if (strncasecmp(page, o2nm_fence_method_desc[i], count - 1))
+                        continue;
+                if (cluster->cl_fence_method != i) {
+                        printk(KERN_INFO "ocfs2: Changing fence method to %s\n",
+                               o2nm_fence_method_desc[i]);
+                        cluster->cl_fence_method = i;
+                }
+                return count;
+        }
+
+bail:
+        return -EINVAL;
+}
+
 static struct o2nm_cluster_attribute o2nm_cluster_attr_idle_timeout_ms = {
         .attr   = { .ca_owner = THIS_MODULE,
                     .ca_name = "idle_timeout_ms",
@@ -603,10 +645,19 @@ static struct o2nm_cluster_attribute o2nm_cluster_attr_reconnect_delay_ms = {
         .store  = o2nm_cluster_attr_reconnect_delay_ms_write,
 };
 
+static struct o2nm_cluster_attribute o2nm_cluster_attr_fence_method = {
+        .attr   = { .ca_owner = THIS_MODULE,
+                    .ca_name = "fence_method",
+                    .ca_mode = S_IRUGO | S_IWUSR },
+        .show   = o2nm_cluster_attr_fence_method_read,
+        .store  = o2nm_cluster_attr_fence_method_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,
+        &o2nm_cluster_attr_fence_method.attr,
         NULL,
 };
 static ssize_t o2nm_cluster_show(struct config_item *item,
@@ -778,6 +829,7 @@ static struct config_group *o2nm_cluster_group_make_group(struct config_group *g
         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;
+        cluster->cl_fence_method       = O2NM_FENCE_RESET;
 
         ret = &cluster->cl_group;
         o2nm_single_cluster = cluster;
diff --git a/fs/ocfs2/cluster/nodemanager.h b/fs/ocfs2/cluster/nodemanager.h
index c992ea0da4ad..09ea2d388bbb 100644
--- a/fs/ocfs2/cluster/nodemanager.h
+++ b/fs/ocfs2/cluster/nodemanager.h
@@ -33,6 +33,12 @@
 #include <linux/configfs.h>
 #include <linux/rbtree.h>
 
+enum o2nm_fence_method {
+        O2NM_FENCE_RESET        = 0,
+        O2NM_FENCE_PANIC,
+        O2NM_FENCE_METHODS,     /* Number of fence methods */
+};
+
 struct o2nm_node {
         spinlock_t              nd_lock;
         struct config_item      nd_item;
@@ -58,6 +64,7 @@ struct o2nm_cluster {
         unsigned int            cl_idle_timeout_ms;
         unsigned int            cl_keepalive_delay_ms;
         unsigned int            cl_reconnect_delay_ms;
+        enum o2nm_fence_method  cl_fence_method;
 
         /* 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)];
diff --git a/fs/ocfs2/cluster/quorum.c b/fs/ocfs2/cluster/quorum.c
index bbacf7da48a4..cf3e16696216 100644
--- a/fs/ocfs2/cluster/quorum.c
+++ b/fs/ocfs2/cluster/quorum.c
@@ -44,7 +44,6 @@
  * and if they're the last, they fire off the decision.
  */
 #include <linux/kernel.h>
-#include <linux/slab.h>
 #include <linux/workqueue.h>
 #include <linux/reboot.h>
 
@@ -74,8 +73,20 @@ static void o2quo_fence_self(void)
          * threads can still schedule, etc, etc */
         o2hb_stop_all_regions();
 
-        printk("ocfs2 is very sorry to be fencing this system by restarting\n");
-        emergency_restart();
+        switch (o2nm_single_cluster->cl_fence_method) {
+        case O2NM_FENCE_PANIC:
+                panic("*** ocfs2 is very sorry to be fencing this system by "
+                      "panicing ***\n");
+                break;
+        default:
+                WARN_ON(o2nm_single_cluster->cl_fence_method >=
+                        O2NM_FENCE_METHODS);
+        case O2NM_FENCE_RESET:
+                printk(KERN_ERR "*** ocfs2 is very sorry to be fencing this "
+                       "system by restarting ***\n");
+                emergency_restart();
+                break;
+        };
 }
 
 /* Indicate that a timeout occured on a hearbeat region write. The
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 334f231a422c..73e743eea2c8 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -72,9 +72,9 @@
 
 #include "tcp_internal.h"
 
-#define SC_NODEF_FMT "node %s (num %u) at %u.%u.%u.%u:%u"
+#define SC_NODEF_FMT "node %s (num %u) at %pI4:%u"
 #define SC_NODEF_ARGS(sc) sc->sc_node->nd_name, sc->sc_node->nd_num,    \
-                          NIPQUAD(sc->sc_node->nd_ipv4_address),        \
+                          &sc->sc_node->nd_ipv4_address,                \
                           ntohs(sc->sc_node->nd_ipv4_port)
 
 /*
@@ -485,7 +485,7 @@ static void o2net_set_nn_state(struct o2net_node *nn,
         }
 
         if (was_valid && !valid) {
-                printk(KERN_INFO "o2net: no longer connected to "
+                printk(KERN_NOTICE "o2net: no longer connected to "
                        SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc));
                 o2net_complete_nodes_nsw(nn);
         }
@@ -493,7 +493,7 @@ static void o2net_set_nn_state(struct o2net_node *nn,
         if (!was_valid && valid) {
                 o2quo_conn_up(o2net_num_from_nn(nn));
                 cancel_delayed_work(&nn->nn_connect_expired);
-                printk(KERN_INFO "o2net: %s " SC_NODEF_FMT "\n",
+                printk(KERN_NOTICE "o2net: %s " SC_NODEF_FMT "\n",
                        o2nm_this_node() > sc->sc_node->nd_num ?
                                 "connected to" : "accepted connection from",
                        SC_NODEF_ARGS(sc));
@@ -930,7 +930,7 @@ static void o2net_sendpage(struct o2net_sock_container *sc,
                         cond_resched();
                         continue;
                 }
-                mlog(ML_ERROR, "sendpage of size %zu to " SC_NODEF_FMT 
+                mlog(ML_ERROR, "sendpage of size %zu to " SC_NODEF_FMT
                      " failed with %zd\n", size, SC_NODEF_ARGS(sc), ret);
                 o2net_ensure_shutdown(nn, sc, 0);
                 break;
@@ -1476,14 +1476,14 @@ static void o2net_idle_timer(unsigned long data)
 
         do_gettimeofday(&now);
 
-        printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u "
+        printk(KERN_NOTICE "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u "
              "seconds, shutting it down.\n", SC_NODEF_ARGS(sc),
                      o2net_idle_timeout() / 1000,
                      o2net_idle_timeout() % 1000);
         mlog(ML_NOTICE, "here are some times that might help debug the "
              "situation: (tmr %ld.%ld now %ld.%ld dr %ld.%ld adv "
              "%ld.%ld:%ld.%ld func (%08x:%u) %ld.%ld:%ld.%ld)\n",
-             sc->sc_tv_timer.tv_sec, (long) sc->sc_tv_timer.tv_usec, 
+             sc->sc_tv_timer.tv_sec, (long) sc->sc_tv_timer.tv_usec,
              now.tv_sec, (long) now.tv_usec,
              sc->sc_tv_data_ready.tv_sec, (long) sc->sc_tv_data_ready.tv_usec,
              sc->sc_tv_advance_start.tv_sec,
diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h
index 8d58cfe410b1..96fa7ebc530c 100644
--- a/fs/ocfs2/cluster/tcp_internal.h
+++ b/fs/ocfs2/cluster/tcp_internal.h
@@ -32,10 +32,10 @@
  * on their number */
 #define O2NET_QUORUM_DELAY_MS   ((o2hb_dead_threshold + 2) * O2HB_REGION_TIMEOUT_MS)
 
-/* 
+/*
  * 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 
+ * locking semantics of the file system using the protocol.  It should
  * be somewhere else, I'm sure, but right now it isn't.
  *
  * With version 11, we separate out the filesystem locking portion.  The
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 28c3ec238796..efd77d071c80 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -2439,7 +2439,7 @@ static int ocfs2_dx_dir_attach_index(struct ocfs2_super *osb,
         dx_root = (struct ocfs2_dx_root_block *)dx_root_bh->b_data;
         memset(dx_root, 0, osb->sb->s_blocksize);
         strcpy(dx_root->dr_signature, OCFS2_DX_ROOT_SIGNATURE);
-        dx_root->dr_suballoc_slot = cpu_to_le16(osb->slot_num);
+        dx_root->dr_suballoc_slot = cpu_to_le16(meta_ac->ac_alloc_slot);
         dx_root->dr_suballoc_bit = cpu_to_le16(dr_suballoc_bit);
         dx_root->dr_fs_generation = cpu_to_le32(osb->fs_generation);
         dx_root->dr_blkno = cpu_to_le64(dr_blkno);
@@ -2964,12 +2964,10 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
                 goto out;
         }
 
-        if (vfs_dq_alloc_space_nodirty(dir,
-                                ocfs2_clusters_to_bytes(osb->sb,
-                                                        alloc + dx_alloc))) {
-                ret = -EDQUOT;
+        ret = dquot_alloc_space_nodirty(dir,
+                ocfs2_clusters_to_bytes(osb->sb, alloc + dx_alloc));
+        if (ret)
                 goto out_commit;
-        }
         did_quota = 1;
 
         if (ocfs2_supports_indexed_dirs(osb) && !dx_inline) {
@@ -3178,7 +3176,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 
 out_commit:
         if (ret < 0 && did_quota)
-                vfs_dq_free_space_nodirty(dir, bytes_allocated);
+                dquot_free_space_nodirty(dir, bytes_allocated);
 
         ocfs2_commit_trans(osb, handle);
 
@@ -3221,11 +3219,10 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
         if (extend) {
                 u32 offset = OCFS2_I(dir)->ip_clusters;
 
-                if (vfs_dq_alloc_space_nodirty(dir,
-                                        ocfs2_clusters_to_bytes(sb, 1))) {
-                        status = -EDQUOT;
+                status = dquot_alloc_space_nodirty(dir,
+                                        ocfs2_clusters_to_bytes(sb, 1));
+                if (status)
                         goto bail;
-                }
                 did_quota = 1;
 
                 status = ocfs2_add_inode_data(OCFS2_SB(sb), dir, &offset,
@@ -3254,7 +3251,7 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
         status = 0;
 bail:
         if (did_quota && status < 0)
-                vfs_dq_free_space_nodirty(dir, ocfs2_clusters_to_bytes(sb, 1));
+                dquot_free_space_nodirty(dir, ocfs2_clusters_to_bytes(sb, 1));
         mlog_exit(status);
         return status;
 }
@@ -3889,11 +3886,10 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
                 goto out;
         }
 
-        if (vfs_dq_alloc_space_nodirty(dir,
-                                       ocfs2_clusters_to_bytes(dir->i_sb, 1))) {
-                ret = -EDQUOT;
+        ret = dquot_alloc_space_nodirty(dir,
+                                       ocfs2_clusters_to_bytes(dir->i_sb, 1));
+        if (ret)
                 goto out_commit;
-        }
         did_quota = 1;
 
         ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir), dx_leaf_bh,
@@ -3983,7 +3979,7 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 
 out_commit:
         if (ret < 0 && did_quota)
-                vfs_dq_free_space_nodirty(dir,
+                dquot_free_space_nodirty(dir,
                                 ocfs2_clusters_to_bytes(dir->i_sb, 1));
 
         ocfs2_commit_trans(osb, handle);
@@ -4165,11 +4161,10 @@ static int ocfs2_expand_inline_dx_root(struct inode *dir,
                 goto out;
         }
 
-        if (vfs_dq_alloc_space_nodirty(dir,
-                                       ocfs2_clusters_to_bytes(osb->sb, 1))) {
-                ret = -EDQUOT;
+        ret = dquot_alloc_space_nodirty(dir,
+                                       ocfs2_clusters_to_bytes(osb->sb, 1));
+        if (ret)
                 goto out_commit;
-        }
         did_quota = 1;
 
         /*
@@ -4229,7 +4224,7 @@ static int ocfs2_expand_inline_dx_root(struct inode *dir,
 
 out_commit:
         if (ret < 0 && did_quota)
-                vfs_dq_free_space_nodirty(dir,
+                dquot_free_space_nodirty(dir,
                                           ocfs2_clusters_to_bytes(dir->i_sb, 1));
 
         ocfs2_commit_trans(osb, handle);
diff --git a/fs/ocfs2/dlm/Makefile b/fs/ocfs2/dlm/Makefile
index 190361375700..dcebf0d920fa 100644
--- a/fs/ocfs2/dlm/Makefile
+++ b/fs/ocfs2/dlm/Makefile
@@ -1,8 +1,7 @@
 EXTRA_CFLAGS += -Ifs/ocfs2
 
-obj-$(CONFIG_OCFS2_FS_O2CB) += ocfs2_dlm.o ocfs2_dlmfs.o
+obj-$(CONFIG_OCFS2_FS_O2CB) += ocfs2_dlm.o
 
 ocfs2_dlm-objs := dlmdomain.o dlmdebug.o dlmthread.o dlmrecovery.o \
         dlmmaster.o dlmast.o dlmconvert.o dlmlock.o dlmunlock.o dlmver.o
 
-ocfs2_dlmfs-objs := userdlm.o dlmfs.o dlmfsver.o
diff --git a/fs/ocfs2/dlm/dlmapi.h b/fs/ocfs2/dlm/dlmapi.h
index b5786a787fab..3cfa114aa391 100644
--- a/fs/ocfs2/dlm/dlmapi.h
+++ b/fs/ocfs2/dlm/dlmapi.h
@@ -95,7 +95,7 @@ const char *dlm_errname(enum dlm_status err);
                 mlog(ML_ERROR, "dlm status = %s\n", dlm_errname((st))); \
 } while (0)
 
-#define DLM_LKSB_UNUSED1           0x01  
+#define DLM_LKSB_UNUSED1           0x01
 #define DLM_LKSB_PUT_LVB           0x02
 #define DLM_LKSB_GET_LVB           0x04
 #define DLM_LKSB_UNUSED2           0x08
diff --git a/fs/ocfs2/dlm/dlmast.c b/fs/ocfs2/dlm/dlmast.c
index 01cf8cc3d286..12d5eb78a11a 100644
--- a/fs/ocfs2/dlm/dlmast.c
+++ b/fs/ocfs2/dlm/dlmast.c
@@ -28,7 +28,6 @@
 #include <linux/module.h>
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/init.h>
 #include <linux/sysctl.h>
@@ -123,7 +122,7 @@ static void __dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
                 dlm_lock_put(lock);
                 /* free up the reserved bast that we are cancelling.
                  * guaranteed that this will not be the last reserved
-                 * ast because *both* an ast and a bast were reserved 
+                 * ast because *both* an ast and a bast were reserved
                  * to get to this point.  the res->spinlock will not be
                  * taken here */
                 dlm_lockres_release_ast(dlm, res);
@@ -185,9 +184,8 @@ static void dlm_update_lvb(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
         BUG_ON(!lksb);
 
         /* only updates if this node masters the lockres */
+        spin_lock(&res->spinlock);
         if (res->owner == dlm->node_num) {
-
-                spin_lock(&res->spinlock);
                 /* check the lksb flags for the direction */
                 if (lksb->flags & DLM_LKSB_GET_LVB) {
                         mlog(0, "getting lvb from lockres for %s node\n",
@@ -202,8 +200,8 @@ static void dlm_update_lvb(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
                  * here. In the future we might want to clear it at the time
                  * the put is actually done.
                  */
-                spin_unlock(&res->spinlock);
         }
+        spin_unlock(&res->spinlock);
 
         /* reset any lvb flags on the lksb */
         lksb->flags &= ~(DLM_LKSB_PUT_LVB|DLM_LKSB_GET_LVB);
diff --git a/fs/ocfs2/dlm/dlmconvert.c b/fs/ocfs2/dlm/dlmconvert.c
index ca96bce50e18..90803b47cd8c 100644
--- a/fs/ocfs2/dlm/dlmconvert.c
+++ b/fs/ocfs2/dlm/dlmconvert.c
@@ -28,7 +28,6 @@
 #include <linux/module.h>
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/init.h>
 #include <linux/sysctl.h>
@@ -396,7 +395,7 @@ static enum dlm_status dlm_send_remote_convert_request(struct dlm_ctxt *dlm,
                         /* instead of logging the same network error over
                          * and over, sleep here and wait for the heartbeat
                          * to notice the node is dead.  times out after 5s. */
-                        dlm_wait_for_node_death(dlm, res->owner, 
+                        dlm_wait_for_node_death(dlm, res->owner,
                                                 DLM_NODE_DEATH_WAIT_MAX);
                         ret = DLM_RECOVERING;
                         mlog(0, "node %u died so returning DLM_RECOVERING "
diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c
index 42b0bad7a612..0cd24cf54396 100644
--- a/fs/ocfs2/dlm/dlmdebug.c
+++ b/fs/ocfs2/dlm/dlmdebug.c
@@ -102,7 +102,7 @@ void __dlm_print_one_lock_resource(struct dlm_lock_resource *res)
         assert_spin_locked(&res->spinlock);
 
         stringify_lockname(res->lockname.name, res->lockname.len,
-                           buf, sizeof(buf) - 1);
+                           buf, sizeof(buf));
         printk("lockres: %s, owner=%u, state=%u\n",
                buf, res->owner, res->state);
         printk("  last used: %lu, refcnt: %u, on purge list: %s\n",
diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c
index 0334000676d3..988c9055fd4e 100644
--- a/fs/ocfs2/dlm/dlmdomain.c
+++ b/fs/ocfs2/dlm/dlmdomain.c
@@ -816,7 +816,7 @@ static int dlm_query_join_handler(struct o2net_msg *msg, u32 len, void *data,
         }
 
         /* Once the dlm ctxt is marked as leaving then we don't want
-         * to be put in someone's domain map. 
+         * to be put in someone's domain map.
          * Also, explicitly disallow joining at certain troublesome
          * times (ie. during recovery). */
         if (dlm && dlm->dlm_state != DLM_CTXT_LEAVING) {
diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c
index 437698e9465f..733337772671 100644
--- a/fs/ocfs2/dlm/dlmlock.c
+++ b/fs/ocfs2/dlm/dlmlock.c
@@ -269,7 +269,7 @@ static enum dlm_status dlmlock_remote(struct dlm_ctxt *dlm,
                 }
                 dlm_revert_pending_lock(res, lock);
                 dlm_lock_put(lock);
-        } else if (dlm_is_recovery_lock(res->lockname.name, 
+        } else if (dlm_is_recovery_lock(res->lockname.name,
                                         res->lockname.len)) {
                 /* special case for the $RECOVERY lock.
                  * there will never be an AST delivered to put
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 83bcaf266b35..9289b4357d27 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -366,7 +366,7 @@ void dlm_hb_event_notify_attached(struct dlm_ctxt *dlm, int idx, int node_up)
         struct dlm_master_list_entry *mle;
 
         assert_spin_locked(&dlm->spinlock);
-        
+
         list_for_each_entry(mle, &dlm->mle_hb_events, hb_events) {
                 if (node_up)
                         dlm_mle_node_up(dlm, mle, NULL, idx);
@@ -833,7 +833,7 @@ lookup:
                 __dlm_insert_mle(dlm, mle);
 
                 /* still holding the dlm spinlock, check the recovery map
-                 * to see if there are any nodes that still need to be 
+                 * to see if there are any nodes that still need to be
                  * considered.  these will not appear in the mle nodemap
                  * but they might own this lockres.  wait on them. */
                 bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0);
@@ -883,7 +883,7 @@ redo_request:
                                 msleep(500);
                         }
                         continue;
-                } 
+                }
 
                 dlm_kick_recovery_thread(dlm);
                 msleep(1000);
@@ -939,8 +939,8 @@ wait:
                      res->lockname.name, blocked);
                 if (++tries > 20) {
                         mlog(ML_ERROR, "%s:%.*s: spinning on "
-                             "dlm_wait_for_lock_mastery, blocked=%d\n", 
-                             dlm->name, res->lockname.len, 
+                             "dlm_wait_for_lock_mastery, blocked=%d\n",
+                             dlm->name, res->lockname.len,
                              res->lockname.name, blocked);
                         dlm_print_one_lock_resource(res);
                         dlm_print_one_mle(mle);
@@ -1029,7 +1029,7 @@ recheck:
                 ret = dlm_restart_lock_mastery(dlm, res, mle, *blocked);
                 b = (mle->type == DLM_MLE_BLOCK);
                 if ((*blocked && !b) || (!*blocked && b)) {
-                        mlog(0, "%s:%.*s: status change: old=%d new=%d\n", 
+                        mlog(0, "%s:%.*s: status change: old=%d new=%d\n",
                              dlm->name, res->lockname.len, res->lockname.name,
                              *blocked, b);
                         *blocked = b;
@@ -1602,7 +1602,7 @@ send_response:
                 }
                 mlog(0, "%u is the owner of %.*s, cleaning everyone else\n",
                              dlm->node_num, res->lockname.len, res->lockname.name);
-                ret = dlm_dispatch_assert_master(dlm, res, 0, request->node_idx, 
+                ret = dlm_dispatch_assert_master(dlm, res, 0, request->node_idx,
                                                  DLM_ASSERT_MASTER_MLE_CLEANUP);
                 if (ret < 0) {
                         mlog(ML_ERROR, "failed to dispatch assert master work\n");
@@ -1701,7 +1701,7 @@ again:
 
                 if (r & DLM_ASSERT_RESPONSE_REASSERT) {
                         mlog(0, "%.*s: node %u create mles on other "
-                             "nodes and requests a re-assert\n", 
+                             "nodes and requests a re-assert\n",
                              namelen, lockname, to);
                         reassert = 1;
                 }
@@ -1812,7 +1812,7 @@ int dlm_assert_master_handler(struct o2net_msg *msg, u32 len, void *data,
                                 spin_unlock(&dlm->master_lock);
                                 spin_unlock(&dlm->spinlock);
                                 goto done;
-                        }       
+                        }
                 }
         }
         spin_unlock(&dlm->master_lock);
@@ -1875,7 +1875,6 @@ int dlm_assert_master_handler(struct o2net_msg *msg, u32 len, void *data,
 ok:
                 spin_unlock(&res->spinlock);
         }
-        spin_unlock(&dlm->spinlock);
 
         // mlog(0, "woo!  got an assert_master from node %u!\n",
         //           assert->node_idx);
@@ -1883,7 +1882,7 @@ ok:
                 int extra_ref = 0;
                 int nn = -1;
                 int rr, err = 0;
-                
+
                 spin_lock(&mle->spinlock);
                 if (mle->type == DLM_MLE_BLOCK || mle->type == DLM_MLE_MIGRATION)
                         extra_ref = 1;
@@ -1891,7 +1890,7 @@ ok:
                         /* MASTER mle: if any bits set in the response map
                          * then the calling node needs to re-assert to clear
                          * up nodes that this node contacted */
-                        while ((nn = find_next_bit (mle->response_map, O2NM_MAX_NODES, 
+                        while ((nn = find_next_bit (mle->response_map, O2NM_MAX_NODES,
                                                     nn+1)) < O2NM_MAX_NODES) {
                                 if (nn != dlm->node_num && nn != assert->node_idx)
                                         master_request = 1;
@@ -1926,7 +1925,6 @@ ok:
                 /* master is known, detach if not already detached.
                  * ensures that only one assert_master call will happen
                  * on this mle. */
-                spin_lock(&dlm->spinlock);
                 spin_lock(&dlm->master_lock);
 
                 rr = atomic_read(&mle->mle_refs.refcount);
@@ -1959,7 +1957,6 @@ ok:
                         __dlm_put_mle(mle);
                 }
                 spin_unlock(&dlm->master_lock);
-                spin_unlock(&dlm->spinlock);
         } else if (res) {
                 if (res->owner != assert->node_idx) {
                         mlog(0, "assert_master from %u, but current "
@@ -1967,6 +1964,7 @@ ok:
                              res->owner, namelen, name);
                 }
         }
+        spin_unlock(&dlm->spinlock);
 
 done:
         ret = 0;
@@ -2002,7 +2000,7 @@ kill:
         __dlm_print_one_lock_resource(res);
         spin_unlock(&res->spinlock);
         spin_unlock(&dlm->spinlock);
-        *ret_data = (void *)res; 
+        *ret_data = (void *)res;
         dlm_put(dlm);
         return -EINVAL;
 }
@@ -2040,10 +2038,10 @@ int dlm_dispatch_assert_master(struct dlm_ctxt *dlm,
         item->u.am.request_from = request_from;
         item->u.am.flags = flags;
 
-        if (ignore_higher) 
-                mlog(0, "IGNORE HIGHER: %.*s\n", res->lockname.len, 
+        if (ignore_higher)
+                mlog(0, "IGNORE HIGHER: %.*s\n", res->lockname.len,
                      res->lockname.name);
-                
+
         spin_lock(&dlm->work_lock);
         list_add_tail(&item->list, &dlm->work_list);
         spin_unlock(&dlm->work_lock);
@@ -2133,7 +2131,7 @@ put:
  * think that $RECOVERY is currently mastered by a dead node.  If so,
  * we wait a short time to allow that node to get notified by its own
  * heartbeat stack, then check again.  All $RECOVERY lock resources
- * mastered by dead nodes are purged when the hearbeat callback is 
+ * mastered by dead nodes are purged when the hearbeat callback is
  * fired, so we can know for sure that it is safe to continue once
  * the node returns a live node or no node.  */
 static int dlm_pre_master_reco_lockres(struct dlm_ctxt *dlm,
@@ -2174,7 +2172,7 @@ static int dlm_pre_master_reco_lockres(struct dlm_ctxt *dlm,
                                 ret = -EAGAIN;
                         }
                         spin_unlock(&dlm->spinlock);
-                        mlog(0, "%s: reco lock master is %u\n", dlm->name, 
+                        mlog(0, "%s: reco lock master is %u\n", dlm->name,
                              master);
                         break;
                 }
@@ -2586,7 +2584,7 @@ fail:
          * is complete everywhere.  if the target dies while this is
          * going on, some nodes could potentially see the target as the
          * master, so it is important that my recovery finds the migration
-         * mle and sets the master to UNKNONWN. */
+         * mle and sets the master to UNKNOWN. */
 
 
         /* wait for new node to assert master */
@@ -2602,7 +2600,7 @@ fail:
 
                         mlog(0, "%s:%.*s: timed out during migration\n",
                              dlm->name, res->lockname.len, res->lockname.name);
-                        /* avoid hang during shutdown when migrating lockres 
+                        /* avoid hang during shutdown when migrating lockres
                          * to a node which also goes down */
                         if (dlm_is_node_dead(dlm, target)) {
                                 mlog(0, "%s:%.*s: expected migration "
@@ -2738,7 +2736,7 @@ static int dlm_migration_can_proceed(struct dlm_ctxt *dlm,
         can_proceed = !!(res->state & DLM_LOCK_RES_MIGRATING);
         spin_unlock(&res->spinlock);
 
-        /* target has died, so make the caller break out of the 
+        /* target has died, so make the caller break out of the
          * wait_event, but caller must recheck the domain_map */
         spin_lock(&dlm->spinlock);
         if (!test_bit(mig_target, dlm->domain_map))
diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c
index d9fa3d22e17c..b4f99de2caf3 100644
--- a/fs/ocfs2/dlm/dlmrecovery.c
+++ b/fs/ocfs2/dlm/dlmrecovery.c
@@ -310,7 +310,7 @@ static int dlm_recovery_thread(void *data)
         mlog(0, "dlm thread running for %s...\n", dlm->name);
 
         while (!kthread_should_stop()) {
-                if (dlm_joined(dlm)) {
+                if (dlm_domain_fully_joined(dlm)) {
                         status = dlm_do_recovery(dlm);
                         if (status == -EAGAIN) {
                                 /* do not sleep, recheck immediately. */
@@ -1050,7 +1050,7 @@ static void dlm_move_reco_locks_to_list(struct dlm_ctxt *dlm,
                                 if (lock->ml.node == dead_node) {
                                         mlog(0, "AHA! there was "
                                              "a $RECOVERY lock for dead "
-                                             "node %u (%s)!\n", 
+                                             "node %u (%s)!\n",
                                              dead_node, dlm->name);
                                         list_del_init(&lock->list);
                                         dlm_lock_put(lock);
@@ -1164,6 +1164,39 @@ static void dlm_init_migratable_lockres(struct dlm_migratable_lockres *mres,
         mres->master = master;
 }
 
+static void dlm_prepare_lvb_for_migration(struct dlm_lock *lock,
+                                          struct dlm_migratable_lockres *mres,
+                                          int queue)
+{
+        if (!lock->lksb)
+               return;
+
+        /* Ignore lvb in all locks in the blocked list */
+        if (queue == DLM_BLOCKED_LIST)
+                return;
+
+        /* Only consider lvbs in locks with granted EX or PR lock levels */
+        if (lock->ml.type != LKM_EXMODE && lock->ml.type != LKM_PRMODE)
+                return;
+
+        if (dlm_lvb_is_empty(mres->lvb)) {
+                memcpy(mres->lvb, lock->lksb->lvb, DLM_LVB_LEN);
+                return;
+        }
+
+        /* Ensure the lvb copied for migration matches in other valid locks */
+        if (!memcmp(mres->lvb, lock->lksb->lvb, DLM_LVB_LEN))
+                return;
+
+        mlog(ML_ERROR, "Mismatched lvb in lock cookie=%u:%llu, name=%.*s, "
+             "node=%u\n",
+             dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)),
+             dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)),
+             lock->lockres->lockname.len, lock->lockres->lockname.name,
+             lock->ml.node);
+        dlm_print_one_lock_resource(lock->lockres);
+        BUG();
+}
 
 /* returns 1 if this lock fills the network structure,
  * 0 otherwise */
@@ -1181,20 +1214,7 @@ static int dlm_add_lock_to_array(struct dlm_lock *lock,
         ml->list = queue;
         if (lock->lksb) {
                 ml->flags = lock->lksb->flags;
-                /* send our current lvb */
-                if (ml->type == LKM_EXMODE ||
-                    ml->type == LKM_PRMODE) {
-                        /* if it is already set, this had better be a PR
-                         * and it has to match */
-                        if (!dlm_lvb_is_empty(mres->lvb) &&
-                            (ml->type == LKM_EXMODE ||
-                             memcmp(mres->lvb, lock->lksb->lvb, DLM_LVB_LEN))) {
-                                mlog(ML_ERROR, "mismatched lvbs!\n");
-                                dlm_print_one_lock_resource(lock->lockres);
-                                BUG();
-                        }
-                        memcpy(mres->lvb, lock->lksb->lvb, DLM_LVB_LEN);
-                }
+                dlm_prepare_lvb_for_migration(lock, mres, queue);
         }
         ml->node = lock->ml.node;
         mres->num_locks++;
@@ -1730,6 +1750,7 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm,
         struct dlm_lock *lock = NULL;
         u8 from = O2NM_MAX_NODES;
         unsigned int added = 0;
+        __be64 c;
 
         mlog(0, "running %d locks for this lockres\n", mres->num_locks);
         for (i=0; i<mres->num_locks; i++) {
@@ -1777,19 +1798,48 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm,
                         /* lock is always created locally first, and
                          * destroyed locally last.  it must be on the list */
                         if (!lock) {
-                                __be64 c = ml->cookie;
-                                mlog(ML_ERROR, "could not find local lock "
-                                               "with cookie %u:%llu!\n",
+                                c = ml->cookie;
+                                mlog(ML_ERROR, "Could not find local lock "
+                                               "with cookie %u:%llu, node %u, "
+                                               "list %u, flags 0x%x, type %d, "
+                                               "conv %d, highest blocked %d\n",
                                      dlm_get_lock_cookie_node(be64_to_cpu(c)),
-                                     dlm_get_lock_cookie_seq(be64_to_cpu(c)));
+                                     dlm_get_lock_cookie_seq(be64_to_cpu(c)),
+                                     ml->node, ml->list, ml->flags, ml->type,
+                                     ml->convert_type, ml->highest_blocked);
+                                __dlm_print_one_lock_resource(res);
+                                BUG();
+                        }
+
+                        if (lock->ml.node != ml->node) {
+                                c = lock->ml.cookie;
+                                mlog(ML_ERROR, "Mismatched node# in lock "
+                                     "cookie %u:%llu, name %.*s, node %u\n",
+                                     dlm_get_lock_cookie_node(be64_to_cpu(c)),
+                                     dlm_get_lock_cookie_seq(be64_to_cpu(c)),
+                                     res->lockname.len, res->lockname.name,
+                                     lock->ml.node);
+                                c = ml->cookie;
+                                mlog(ML_ERROR, "Migrate lock cookie %u:%llu, "
+                                     "node %u, list %u, flags 0x%x, type %d, "
+                                     "conv %d, highest blocked %d\n",
+                                     dlm_get_lock_cookie_node(be64_to_cpu(c)),
+                                     dlm_get_lock_cookie_seq(be64_to_cpu(c)),
+                                     ml->node, ml->list, ml->flags, ml->type,
+                                     ml->convert_type, ml->highest_blocked);
                                 __dlm_print_one_lock_resource(res);
                                 BUG();
                         }
-                        BUG_ON(lock->ml.node != ml->node);
 
                         if (tmpq != queue) {
-                                mlog(0, "lock was on %u instead of %u for %.*s\n",
-                                     j, ml->list, res->lockname.len, res->lockname.name);
+                                c = ml->cookie;
+                                mlog(0, "Lock cookie %u:%llu was on list %u "
+                                     "instead of list %u for %.*s\n",
+                                     dlm_get_lock_cookie_node(be64_to_cpu(c)),
+                                     dlm_get_lock_cookie_seq(be64_to_cpu(c)),
+                                     j, ml->list, res->lockname.len,
+                                     res->lockname.name);
+                                __dlm_print_one_lock_resource(res);
                                 spin_unlock(&res->spinlock);
                                 continue;
                         }
@@ -1839,7 +1889,7 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm,
                                  * the lvb. */
                                 memcpy(res->lvb, mres->lvb, DLM_LVB_LEN);
                         } else {
-                                /* otherwise, the node is sending its 
+                                /* otherwise, the node is sending its
                                  * most recent valid lvb info */
                                 BUG_ON(ml->type != LKM_EXMODE &&
                                        ml->type != LKM_PRMODE);
@@ -1886,7 +1936,7 @@ skip_lvb:
                 spin_lock(&res->spinlock);
                 list_for_each_entry(lock, queue, list) {
                         if (lock->ml.cookie == ml->cookie) {
-                                __be64 c = lock->ml.cookie;
+                                c = lock->ml.cookie;
                                 mlog(ML_ERROR, "%s:%.*s: %u:%llu: lock already "
                                      "exists on this lockres!\n", dlm->name,
                                      res->lockname.len, res->lockname.name,
@@ -2114,7 +2164,7 @@ static void dlm_revalidate_lvb(struct dlm_ctxt *dlm,
         assert_spin_locked(&res->spinlock);
 
         if (res->owner == dlm->node_num)
-                /* if this node owned the lockres, and if the dead node 
+                /* if this node owned the lockres, and if the dead node
                  * had an EX when he died, blank out the lvb */
                 search_node = dead_node;
         else {
@@ -2152,7 +2202,7 @@ static void dlm_free_dead_locks(struct dlm_ctxt *dlm,
 
         /* this node is the lockres master:
          * 1) remove any stale locks for the dead node
-         * 2) if the dead node had an EX when he died, blank out the lvb 
+         * 2) if the dead node had an EX when he died, blank out the lvb
          */
         assert_spin_locked(&dlm->spinlock);
         assert_spin_locked(&res->spinlock);
@@ -2193,7 +2243,12 @@ static void dlm_free_dead_locks(struct dlm_ctxt *dlm,
                 mlog(0, "%s:%.*s: freed %u locks for dead node %u, "
                      "dropping ref from lockres\n", dlm->name,
                      res->lockname.len, res->lockname.name, freed, dead_node);
-                BUG_ON(!test_bit(dead_node, res->refmap));
+                if(!test_bit(dead_node, res->refmap)) {
+                        mlog(ML_ERROR, "%s:%.*s: freed %u locks for dead node %u, "
+                             "but ref was not set\n", dlm->name,
+                             res->lockname.len, res->lockname.name, freed, dead_node);
+                        __dlm_print_one_lock_resource(res);
+                }
                 dlm_lockres_clear_refmap_bit(dead_node, res);
         } else if (test_bit(dead_node, res->refmap)) {
                 mlog(0, "%s:%.*s: dead node %u had a ref, but had "
@@ -2260,7 +2315,7 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node)
                                 }
                                 spin_unlock(&res->spinlock);
                                 continue;
-                        }                       
+                        }
                         spin_lock(&res->spinlock);
                         /* zero the lvb if necessary */
                         dlm_revalidate_lvb(dlm, res, dead_node);
@@ -2411,7 +2466,7 @@ static void dlm_reco_unlock_ast(void *astdata, enum dlm_status st)
  * this function on each node racing to become the recovery
  * master will not stop attempting this until either:
  * a) this node gets the EX (and becomes the recovery master),
- * or b) dlm->reco.new_master gets set to some nodenum 
+ * or b) dlm->reco.new_master gets set to some nodenum
  * != O2NM_INVALID_NODE_NUM (another node will do the reco).
  * so each time a recovery master is needed, the entire cluster
  * will sync at this point.  if the new master dies, that will
@@ -2424,7 +2479,7 @@ static int dlm_pick_recovery_master(struct dlm_ctxt *dlm)
 
         mlog(0, "starting recovery of %s at %lu, dead=%u, this=%u\n",
              dlm->name, jiffies, dlm->reco.dead_node, dlm->node_num);
-again:  
+again:
         memset(&lksb, 0, sizeof(lksb));
 
         ret = dlmlock(dlm, LKM_EXMODE, &lksb, LKM_NOQUEUE|LKM_RECOVERY,
@@ -2437,8 +2492,8 @@ again:
         if (ret == DLM_NORMAL) {
                 mlog(0, "dlm=%s dlmlock says I got it (this=%u)\n",
                      dlm->name, dlm->node_num);
-                
-                /* got the EX lock.  check to see if another node 
+
+                /* got the EX lock.  check to see if another node
                  * just became the reco master */
                 if (dlm_reco_master_ready(dlm)) {
                         mlog(0, "%s: got reco EX lock, but %u will "
@@ -2451,12 +2506,12 @@ again:
                         /* see if recovery was already finished elsewhere */
                         spin_lock(&dlm->spinlock);
                         if (dlm->reco.dead_node == O2NM_INVALID_NODE_NUM) {
-                                status = -EINVAL;       
+                                status = -EINVAL;
                                 mlog(0, "%s: got reco EX lock, but "
                                      "node got recovered already\n", dlm->name);
                                 if (dlm->reco.new_master != O2NM_INVALID_NODE_NUM) {
                                         mlog(ML_ERROR, "%s: new master is %u "
-                                             "but no dead node!\n", 
+                                             "but no dead node!\n",
                                              dlm->name, dlm->reco.new_master);
                                         BUG();
                                 }
@@ -2468,7 +2523,7 @@ again:
                  * set the master and send the messages to begin recovery */
                 if (!status) {
                         mlog(0, "%s: dead=%u, this=%u, sending "
-                             "begin_reco now\n", dlm->name, 
+                             "begin_reco now\n", dlm->name,
                              dlm->reco.dead_node, dlm->node_num);
                         status = dlm_send_begin_reco_message(dlm,
                                       dlm->reco.dead_node);
@@ -2501,7 +2556,7 @@ again:
                 mlog(0, "dlm=%s dlmlock says another node got it (this=%u)\n",
                      dlm->name, dlm->node_num);
                 /* another node is master. wait on
-                 * reco.new_master != O2NM_INVALID_NODE_NUM 
+                 * reco.new_master != O2NM_INVALID_NODE_NUM
                  * for at most one second */
                 wait_event_timeout(dlm->dlm_reco_thread_wq,
                                          dlm_reco_master_ready(dlm),
@@ -2589,9 +2644,23 @@ retry:
                              "begin reco msg (%d)\n", dlm->name, nodenum, ret);
                         ret = 0;
                 }
+
+                /*
+                 * Prior to commit aad1b15310b9bcd59fa81ab8f2b1513b59553ea8,
+                 * dlm_begin_reco_handler() returned EAGAIN and not -EAGAIN.
+                 * We are handling both for compatibility reasons.
+                 */
+                if (ret == -EAGAIN || ret == EAGAIN) {
+                        mlog(0, "%s: trying to start recovery of node "
+                             "%u, but node %u is waiting for last recovery "
+                             "to complete, backoff for a bit\n", dlm->name,
+                             dead_node, nodenum);
+                        msleep(100);
+                        goto retry;
+                }
                 if (ret < 0) {
                         struct dlm_lock_resource *res;
-                        /* this is now a serious problem, possibly ENOMEM 
+                        /* this is now a serious problem, possibly ENOMEM
                          * in the network stack.  must retry */
                         mlog_errno(ret);
                         mlog(ML_ERROR, "begin reco of dlm %s to node %u "
@@ -2604,18 +2673,10 @@ retry:
                         } else {
                                 mlog(ML_ERROR, "recovery lock not found\n");
                         }
-                        /* sleep for a bit in hopes that we can avoid 
+                        /* sleep for a bit in hopes that we can avoid
                          * another ENOMEM */
                         msleep(100);
                         goto retry;
-                } else if (ret == EAGAIN) {
-                        mlog(0, "%s: trying to start recovery of node "
-                             "%u, but node %u is waiting for last recovery "
-                             "to complete, backoff for a bit\n", dlm->name,
-                             dead_node, nodenum);
-                        /* TODO Look into replacing msleep with cond_resched() */
-                        msleep(100);
-                        goto retry;
                 }
         }
 
@@ -2639,7 +2700,7 @@ int dlm_begin_reco_handler(struct o2net_msg *msg, u32 len, void *data,
                      dlm->name, br->node_idx, br->dead_node,
                      dlm->reco.dead_node, dlm->reco.new_master);
                 spin_unlock(&dlm->spinlock);
-                return EAGAIN;
+                return -EAGAIN;
         }
         spin_unlock(&dlm->spinlock);
 
@@ -2664,7 +2725,7 @@ int dlm_begin_reco_handler(struct o2net_msg *msg, u32 len, void *data,
         }
         if (dlm->reco.dead_node != O2NM_INVALID_NODE_NUM) {
                 mlog(ML_NOTICE, "%s: dead_node previously set to %u, "
-                     "node %u changing it to %u\n", dlm->name, 
+                     "node %u changing it to %u\n", dlm->name,
                      dlm->reco.dead_node, br->node_idx, br->dead_node);
         }
         dlm_set_reco_master(dlm, br->node_idx);
@@ -2730,8 +2791,8 @@ stage2:
                 if (ret < 0) {
                         mlog_errno(ret);
                         if (dlm_is_host_down(ret)) {
-                                /* this has no effect on this recovery 
-                                 * session, so set the status to zero to 
+                                /* this has no effect on this recovery
+                                 * session, so set the status to zero to
                                  * finish out the last recovery */
                                 mlog(ML_ERROR, "node %u went down after this "
                                      "node finished recovery.\n", nodenum);
@@ -2768,7 +2829,7 @@ int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data,
         mlog(0, "%s: node %u finalizing recovery stage%d of "
              "node %u (%u:%u)\n", dlm->name, fr->node_idx, stage,
              fr->dead_node, dlm->reco.dead_node, dlm->reco.new_master);
- 
+
         spin_lock(&dlm->spinlock);
 
         if (dlm->reco.new_master != fr->node_idx) {
diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c
index 52ec020ea78b..11a6d1fd1d35 100644
--- a/fs/ocfs2/dlm/dlmthread.c
+++ b/fs/ocfs2/dlm/dlmthread.c
@@ -28,7 +28,6 @@
 #include <linux/module.h>
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/init.h>
 #include <linux/sysctl.h>
diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c
index 00f53b2aea76..b47c1b92b82b 100644
--- a/fs/ocfs2/dlm/dlmunlock.c
+++ b/fs/ocfs2/dlm/dlmunlock.c
@@ -28,7 +28,6 @@
 #include <linux/module.h>
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/init.h>
 #include <linux/sysctl.h>
@@ -190,8 +189,8 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm,
                         actions &= ~(DLM_UNLOCK_REMOVE_LOCK|
                                      DLM_UNLOCK_REGRANT_LOCK|
                                      DLM_UNLOCK_CLEAR_CONVERT_TYPE);
-                } else if (status == DLM_RECOVERING || 
-                           status == DLM_MIGRATING || 
+                } else if (status == DLM_RECOVERING ||
+                           status == DLM_MIGRATING ||
                            status == DLM_FORWARD) {
                         /* must clear the actions because this unlock
                          * is about to be retried.  cannot free or do
@@ -661,14 +660,14 @@ retry:
         if (call_ast) {
                 mlog(0, "calling unlockast(%p, %d)\n", data, status);
                 if (is_master) {
-                        /* it is possible that there is one last bast 
+                        /* it is possible that there is one last bast
                          * pending.  make sure it is flushed, then
                          * call the unlockast.
                          * not an issue if this is a mastered remotely,
                          * since this lock has been removed from the
                          * lockres queues and cannot be found. */
                         dlm_kick_thread(dlm, NULL);
-                        wait_event(dlm->ast_wq, 
+                        wait_event(dlm->ast_wq,
                                    dlm_lock_basts_flushed(dlm, lock));
                 }
                 (*unlockast)(data, status);
diff --git a/fs/ocfs2/dlmfs/Makefile b/fs/ocfs2/dlmfs/Makefile
new file mode 100644
index 000000000000..df69b4856d0d
--- /dev/null
+++ b/fs/ocfs2/dlmfs/Makefile
@@ -0,0 +1,5 @@
+EXTRA_CFLAGS += -Ifs/ocfs2
+
+obj-$(CONFIG_OCFS2_FS) += ocfs2_dlmfs.o
+
+ocfs2_dlmfs-objs := userdlm.o dlmfs.o dlmfsver.o
diff --git a/fs/ocfs2/dlm/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c
index 02bf17808bdc..b83d6107a1f5 100644
--- a/fs/ocfs2/dlm/dlmfs.c
+++ b/fs/ocfs2/dlmfs/dlmfs.c
@@ -43,24 +43,17 @@
 #include <linux/init.h>
 #include <linux/string.h>
 #include <linux/backing-dev.h>
+#include <linux/poll.h>
 
 #include <asm/uaccess.h>
 
-
-#include "cluster/nodemanager.h"
-#include "cluster/heartbeat.h"
-#include "cluster/tcp.h"
-
-#include "dlmapi.h"
-
+#include "stackglue.h"
 #include "userdlm.h"
-
 #include "dlmfsver.h"
 
 #define MLOG_MASK_PREFIX ML_DLMFS
 #include "cluster/masklog.h"
 
-#include "ocfs2_lockingver.h"
 
 static const struct super_operations dlmfs_ops;
 static const struct file_operations dlmfs_file_operations;
@@ -71,15 +64,46 @@ static struct kmem_cache *dlmfs_inode_cache;
 
 struct workqueue_struct *user_dlm_worker;
 
+
+
 /*
- * This is the userdlmfs locking protocol version.
+ * These are the ABI capabilities of dlmfs.
+ *
+ * Over time, dlmfs has added some features that were not part of the
+ * initial ABI.  Unfortunately, some of these features are not detectable
+ * via standard usage.  For example, Linux's default poll always returns
+ * POLLIN, so there is no way for a caller of poll(2) to know when dlmfs
+ * added poll support.  Instead, we provide this list of new capabilities.
+ *
+ * Capabilities is a read-only attribute.  We do it as a module parameter
+ * so we can discover it whether dlmfs is built in, loaded, or even not
+ * loaded.
  *
- * See fs/ocfs2/dlmglue.c for more details on locking versions.
+ * The ABI features are local to this machine's dlmfs mount.  This is
+ * distinct from the locking protocol, which is concerned with inter-node
+ * interaction.
+ *
+ * Capabilities:
+ * - bast       : POLLIN against the file descriptor of a held lock
+ *                signifies a bast fired on the lock.
  */
-static const struct dlm_protocol_version user_locking_protocol = {
-        .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
-        .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
-};
+#define DLMFS_CAPABILITIES "bast stackglue"
+extern int param_set_dlmfs_capabilities(const char *val,
+                                        struct kernel_param *kp)
+{
+        printk(KERN_ERR "%s: readonly parameter\n", kp->name);
+        return -EINVAL;
+}
+static int param_get_dlmfs_capabilities(char *buffer,
+                                        struct kernel_param *kp)
+{
+        return strlcpy(buffer, DLMFS_CAPABILITIES,
+                       strlen(DLMFS_CAPABILITIES) + 1);
+}
+module_param_call(capabilities, param_set_dlmfs_capabilities,
+                  param_get_dlmfs_capabilities, NULL, 0444);
+MODULE_PARM_DESC(capabilities, DLMFS_CAPABILITIES);
+
 
 /*
  * decodes a set of open flags into a valid lock level and a set of flags.
@@ -88,20 +112,20 @@ static const struct dlm_protocol_version user_locking_protocol = {
  * O_RDONLY -> PRMODE level
  * O_WRONLY -> EXMODE level
  *
- * O_NONBLOCK -> LKM_NOQUEUE
+ * O_NONBLOCK -> NOQUEUE
  */
 static int dlmfs_decode_open_flags(int open_flags,
                                    int *level,
                                    int *flags)
 {
         if (open_flags & (O_WRONLY|O_RDWR))
-                *level = LKM_EXMODE;
+                *level = DLM_LOCK_EX;
         else
-                *level = LKM_PRMODE;
+                *level = DLM_LOCK_PR;
 
         *flags = 0;
         if (open_flags & O_NONBLOCK)
-                *flags |= LKM_NOQUEUE;
+                *flags |= DLM_LKF_NOQUEUE;
 
         return 0;
 }
@@ -142,7 +166,7 @@ static int dlmfs_file_open(struct inode *inode,
                  * to be able userspace to be able to distinguish a
                  * valid lock request from one that simply couldn't be
                  * granted. */
-                if (flags & LKM_NOQUEUE && status == -EAGAIN)
+                if (flags & DLM_LKF_NOQUEUE && status == -EAGAIN)
                         status = -ETXTBSY;
                 kfree(fp);
                 goto bail;
@@ -169,7 +193,7 @@ static int dlmfs_file_release(struct inode *inode,
         status = 0;
         if (fp) {
                 level = fp->fp_lock_level;
-                if (level != LKM_IVMODE)
+                if (level != DLM_LOCK_IV)
                         user_dlm_cluster_unlock(&ip->ip_lockres, level);
 
                 kfree(fp);
@@ -179,13 +203,46 @@ static int dlmfs_file_release(struct inode *inode,
         return 0;
 }
 
+/*
+ * We do ->setattr() just to override size changes.  Our size is the size
+ * of the LVB and nothing else.
+ */
+static int dlmfs_file_setattr(struct dentry *dentry, struct iattr *attr)
+{
+        int error;
+        struct inode *inode = dentry->d_inode;
+
+        attr->ia_valid &= ~ATTR_SIZE;
+        error = inode_change_ok(inode, attr);
+        if (!error)
+                error = inode_setattr(inode, attr);
+
+        return error;
+}
+
+static unsigned int dlmfs_file_poll(struct file *file, poll_table *wait)
+{
+        int event = 0;
+        struct inode *inode = file->f_path.dentry->d_inode;
+        struct dlmfs_inode_private *ip = DLMFS_I(inode);
+
+        poll_wait(file, &ip->ip_lockres.l_event, wait);
+
+        spin_lock(&ip->ip_lockres.l_lock);
+        if (ip->ip_lockres.l_flags & USER_LOCK_BLOCKED)
+                event = POLLIN | POLLRDNORM;
+        spin_unlock(&ip->ip_lockres.l_lock);
+
+        return event;
+}
+
 static ssize_t dlmfs_file_read(struct file *filp,
                                char __user *buf,
                                size_t count,
                                loff_t *ppos)
 {
         int bytes_left;
-        ssize_t readlen;
+        ssize_t readlen, got;
         char *lvb_buf;
         struct inode *inode = filp->f_path.dentry->d_inode;
 
@@ -205,15 +262,19 @@ static ssize_t dlmfs_file_read(struct file *filp,
         if ((count + *ppos) > i_size_read(inode))
                 readlen = i_size_read(inode) - *ppos;
         else
-                readlen = count - *ppos;
+                readlen = count;
 
         lvb_buf = kmalloc(readlen, GFP_NOFS);
         if (!lvb_buf)
                 return -ENOMEM;
 
-        user_dlm_read_lvb(inode, lvb_buf, readlen);
-        bytes_left = __copy_to_user(buf, lvb_buf, readlen);
-        readlen -= bytes_left;
+        got = user_dlm_read_lvb(inode, lvb_buf, readlen);
+        if (got) {
+                BUG_ON(got != readlen);
+                bytes_left = __copy_to_user(buf, lvb_buf, readlen);
+                readlen -= bytes_left;
+        } else
+                readlen = 0;
 
         kfree(lvb_buf);
 
@@ -272,7 +333,7 @@ static void dlmfs_init_once(void *foo)
         struct dlmfs_inode_private *ip =
                 (struct dlmfs_inode_private *) foo;
 
-        ip->ip_dlm = NULL;
+        ip->ip_conn = NULL;
         ip->ip_parent = NULL;
 
         inode_init_once(&ip->ip_vfs_inode);
@@ -314,14 +375,14 @@ static void dlmfs_clear_inode(struct inode *inode)
                 goto clear_fields;
         }
 
-        mlog(0, "we're a directory, ip->ip_dlm = 0x%p\n", ip->ip_dlm);
+        mlog(0, "we're a directory, ip->ip_conn = 0x%p\n", ip->ip_conn);
         /* we must be a directory. If required, lets unregister the
          * dlm context now. */
-        if (ip->ip_dlm)
-                user_dlm_unregister_context(ip->ip_dlm);
+        if (ip->ip_conn)
+                user_dlm_unregister(ip->ip_conn);
 clear_fields:
         ip->ip_parent = NULL;
-        ip->ip_dlm = NULL;
+        ip->ip_conn = NULL;
 }
 
 static struct backing_dev_info dlmfs_backing_dev_info = {
@@ -371,7 +432,7 @@ static struct inode *dlmfs_get_inode(struct inode *parent,
         inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
         ip = DLMFS_I(inode);
-        ip->ip_dlm = DLMFS_I(parent)->ip_dlm;
+        ip->ip_conn = DLMFS_I(parent)->ip_conn;
 
         switch (mode & S_IFMT) {
         default:
@@ -425,13 +486,12 @@ static int dlmfs_mkdir(struct inode * dir,
         struct inode *inode = NULL;
         struct qstr *domain = &dentry->d_name;
         struct dlmfs_inode_private *ip;
-        struct dlm_ctxt *dlm;
-        struct dlm_protocol_version proto = user_locking_protocol;
+        struct ocfs2_cluster_connection *conn;
 
         mlog(0, "mkdir %.*s\n", domain->len, domain->name);
 
         /* verify that we have a proper domain */
-        if (domain->len >= O2NM_MAX_NAME_LEN) {
+        if (domain->len >= GROUP_NAME_MAX) {
                 status = -EINVAL;
                 mlog(ML_ERROR, "invalid domain name for directory.\n");
                 goto bail;
@@ -446,14 +506,14 @@ static int dlmfs_mkdir(struct inode * dir,
 
         ip = DLMFS_I(inode);
 
-        dlm = user_dlm_register_context(domain, &proto);
-        if (IS_ERR(dlm)) {
-                status = PTR_ERR(dlm);
+        conn = user_dlm_register(domain);
+        if (IS_ERR(conn)) {
+                status = PTR_ERR(conn);
                 mlog(ML_ERROR, "Error %d could not register domain \"%.*s\"\n",
                      status, domain->len, domain->name);
                 goto bail;
         }
-        ip->ip_dlm = dlm;
+        ip->ip_conn = conn;
 
         inc_nlink(dir);
         d_instantiate(dentry, inode);
@@ -549,6 +609,7 @@ static int dlmfs_fill_super(struct super_block * sb,
 static const struct file_operations dlmfs_file_operations = {
         .open           = dlmfs_file_open,
         .release        = dlmfs_file_release,
+        .poll           = dlmfs_file_poll,
         .read           = dlmfs_file_read,
         .write          = dlmfs_file_write,
 };
@@ -576,6 +637,7 @@ static const struct super_operations dlmfs_ops = {
 
 static const struct inode_operations dlmfs_file_inode_operations = {
         .getattr        = simple_getattr,
+        .setattr        = dlmfs_file_setattr,
 };
 
 static int dlmfs_get_sb(struct file_system_type *fs_type,
@@ -620,6 +682,7 @@ static int __init init_dlmfs_fs(void)
         }
         cleanup_worker = 1;
 
+        user_dlm_set_locking_protocol();
         status = register_filesystem(&dlmfs_fs_type);
 bail:
         if (status) {
diff --git a/fs/ocfs2/dlm/dlmfsver.c b/fs/ocfs2/dlmfs/dlmfsver.c
index a733b3321f83..a733b3321f83 100644
--- a/fs/ocfs2/dlm/dlmfsver.c
+++ b/fs/ocfs2/dlmfs/dlmfsver.c
diff --git a/fs/ocfs2/dlm/dlmfsver.h b/fs/ocfs2/dlmfs/dlmfsver.h
index f35eadbed25c..f35eadbed25c 100644
--- a/fs/ocfs2/dlm/dlmfsver.h
+++ b/fs/ocfs2/dlmfs/dlmfsver.h
diff --git a/fs/ocfs2/dlm/userdlm.c b/fs/ocfs2/dlmfs/userdlm.c
index 4cb1d3dae250..0499e3fb7bdb 100644
--- a/fs/ocfs2/dlm/userdlm.c
+++ b/fs/ocfs2/dlmfs/userdlm.c
@@ -34,18 +34,19 @@
 #include <linux/types.h>
 #include <linux/crc32.h>
 
-
-#include "cluster/nodemanager.h"
-#include "cluster/heartbeat.h"
-#include "cluster/tcp.h"
-
-#include "dlmapi.h"
-
+#include "ocfs2_lockingver.h"
+#include "stackglue.h"
 #include "userdlm.h"
 
 #define MLOG_MASK_PREFIX ML_DLMFS
 #include "cluster/masklog.h"
 
+
+static inline struct user_lock_res *user_lksb_to_lock_res(struct ocfs2_dlm_lksb *lksb)
+{
+        return container_of(lksb, struct user_lock_res, l_lksb);
+}
+
 static inline int user_check_wait_flag(struct user_lock_res *lockres,
                                        int flag)
 {
@@ -73,15 +74,15 @@ static inline void user_wait_on_blocked_lock(struct user_lock_res *lockres)
 }
 
 /* I heart container_of... */
-static inline struct dlm_ctxt *
-dlm_ctxt_from_user_lockres(struct user_lock_res *lockres)
+static inline struct ocfs2_cluster_connection *
+cluster_connection_from_user_lockres(struct user_lock_res *lockres)
 {
         struct dlmfs_inode_private *ip;
 
         ip = container_of(lockres,
                           struct dlmfs_inode_private,
                           ip_lockres);
-        return ip->ip_dlm;
+        return ip->ip_conn;
 }
 
 static struct inode *
@@ -103,9 +104,9 @@ static inline void user_recover_from_dlm_error(struct user_lock_res *lockres)
 }
 
 #define user_log_dlm_error(_func, _stat, _lockres) do {                 \
-        mlog(ML_ERROR, "Dlm error \"%s\" while calling %s on "          \
-                "resource %.*s: %s\n", dlm_errname(_stat), _func,       \
-                _lockres->l_namelen, _lockres->l_name, dlm_errmsg(_stat)); \
+        mlog(ML_ERROR, "Dlm error %d while calling %s on "              \
+                "resource %.*s\n", _stat, _func,                        \
+                _lockres->l_namelen, _lockres->l_name);                 \
 } while (0)
 
 /* WARNING: This function lives in a world where the only three lock
@@ -113,34 +114,35 @@ static inline void user_recover_from_dlm_error(struct user_lock_res *lockres)
  * lock types are added. */
 static inline int user_highest_compat_lock_level(int level)
 {
-        int new_level = LKM_EXMODE;
+        int new_level = DLM_LOCK_EX;
 
-        if (level == LKM_EXMODE)
-                new_level = LKM_NLMODE;
-        else if (level == LKM_PRMODE)
-                new_level = LKM_PRMODE;
+        if (level == DLM_LOCK_EX)
+                new_level = DLM_LOCK_NL;
+        else if (level == DLM_LOCK_PR)
+                new_level = DLM_LOCK_PR;
         return new_level;
 }
 
-static void user_ast(void *opaque)
+static void user_ast(struct ocfs2_dlm_lksb *lksb)
 {
-        struct user_lock_res *lockres = opaque;
-        struct dlm_lockstatus *lksb;
+        struct user_lock_res *lockres = user_lksb_to_lock_res(lksb);
+        int status;
 
-        mlog(0, "AST fired for lockres %.*s\n", lockres->l_namelen,
-             lockres->l_name);
+        mlog(ML_BASTS, "AST fired for lockres %.*s, level %d => %d\n",
+             lockres->l_namelen, lockres->l_name, lockres->l_level,
+             lockres->l_requested);
 
         spin_lock(&lockres->l_lock);
 
-        lksb = &(lockres->l_lksb);
-        if (lksb->status != DLM_NORMAL) {
+        status = ocfs2_dlm_lock_status(&lockres->l_lksb);
+        if (status) {
                 mlog(ML_ERROR, "lksb status value of %u on lockres %.*s\n",
-                     lksb->status, lockres->l_namelen, lockres->l_name);
+                     status, lockres->l_namelen, lockres->l_name);
                 spin_unlock(&lockres->l_lock);
                 return;
         }
 
-        mlog_bug_on_msg(lockres->l_requested == LKM_IVMODE,
+        mlog_bug_on_msg(lockres->l_requested == DLM_LOCK_IV,
                         "Lockres %.*s, requested ivmode. flags 0x%x\n",
                         lockres->l_namelen, lockres->l_name, lockres->l_flags);
 
@@ -148,13 +150,13 @@ static void user_ast(void *opaque)
         if (lockres->l_requested < lockres->l_level) {
                 if (lockres->l_requested <=
                     user_highest_compat_lock_level(lockres->l_blocking)) {
-                        lockres->l_blocking = LKM_NLMODE;
+                        lockres->l_blocking = DLM_LOCK_NL;
                         lockres->l_flags &= ~USER_LOCK_BLOCKED;
                 }
         }
 
         lockres->l_level = lockres->l_requested;
-        lockres->l_requested = LKM_IVMODE;
+        lockres->l_requested = DLM_LOCK_IV;
         lockres->l_flags |= USER_LOCK_ATTACHED;
         lockres->l_flags &= ~USER_LOCK_BUSY;
 
@@ -193,11 +195,11 @@ static void __user_dlm_cond_queue_lockres(struct user_lock_res *lockres)
                 return;
 
         switch (lockres->l_blocking) {
-        case LKM_EXMODE:
+        case DLM_LOCK_EX:
                 if (!lockres->l_ex_holders && !lockres->l_ro_holders)
                         queue = 1;
                 break;
-        case LKM_PRMODE:
+        case DLM_LOCK_PR:
                 if (!lockres->l_ex_holders)
                         queue = 1;
                 break;
@@ -209,12 +211,12 @@ static void __user_dlm_cond_queue_lockres(struct user_lock_res *lockres)
                 __user_dlm_queue_lockres(lockres);
 }
 
-static void user_bast(void *opaque, int level)
+static void user_bast(struct ocfs2_dlm_lksb *lksb, int level)
 {
-        struct user_lock_res *lockres = opaque;
+        struct user_lock_res *lockres = user_lksb_to_lock_res(lksb);
 
-        mlog(0, "Blocking AST fired for lockres %.*s. Blocking level %d\n",
-             lockres->l_namelen, lockres->l_name, level);
+        mlog(ML_BASTS, "BAST fired for lockres %.*s, blocking %d, level %d\n",
+             lockres->l_namelen, lockres->l_name, level, lockres->l_level);
 
         spin_lock(&lockres->l_lock);
         lockres->l_flags |= USER_LOCK_BLOCKED;
@@ -227,15 +229,15 @@ static void user_bast(void *opaque, int level)
         wake_up(&lockres->l_event);
 }
 
-static void user_unlock_ast(void *opaque, enum dlm_status status)
+static void user_unlock_ast(struct ocfs2_dlm_lksb *lksb, int status)
 {
-        struct user_lock_res *lockres = opaque;
+        struct user_lock_res *lockres = user_lksb_to_lock_res(lksb);
 
-        mlog(0, "UNLOCK AST called on lock %.*s\n", lockres->l_namelen,
-             lockres->l_name);
+        mlog(ML_BASTS, "UNLOCK AST fired for lockres %.*s, flags 0x%x\n",
+             lockres->l_namelen, lockres->l_name, lockres->l_flags);
 
-        if (status != DLM_NORMAL && status != DLM_CANCELGRANT)
-                mlog(ML_ERROR, "Dlm returns status %d\n", status);
+        if (status)
+                mlog(ML_ERROR, "dlm returns status %d\n", status);
 
         spin_lock(&lockres->l_lock);
         /* The teardown flag gets set early during the unlock process,
@@ -243,7 +245,7 @@ static void user_unlock_ast(void *opaque, enum dlm_status status)
          * for a concurrent cancel. */
         if (lockres->l_flags & USER_LOCK_IN_TEARDOWN
             && !(lockres->l_flags & USER_LOCK_IN_CANCEL)) {
-                lockres->l_level = LKM_IVMODE;
+                lockres->l_level = DLM_LOCK_IV;
         } else if (status == DLM_CANCELGRANT) {
                 /* We tried to cancel a convert request, but it was
                  * already granted. Don't clear the busy flag - the
@@ -254,7 +256,7 @@ static void user_unlock_ast(void *opaque, enum dlm_status status)
         } else {
                 BUG_ON(!(lockres->l_flags & USER_LOCK_IN_CANCEL));
                 /* Cancel succeeded, we want to re-queue */
-                lockres->l_requested = LKM_IVMODE; /* cancel an
+                lockres->l_requested = DLM_LOCK_IV; /* cancel an
                                                     * upconvert
                                                     * request. */
                 lockres->l_flags &= ~USER_LOCK_IN_CANCEL;
@@ -271,6 +273,21 @@ out_noclear:
         wake_up(&lockres->l_event);
 }
 
+/*
+ * This is the userdlmfs locking protocol version.
+ *
+ * See fs/ocfs2/dlmglue.c for more details on locking versions.
+ */
+static struct ocfs2_locking_protocol user_dlm_lproto = {
+        .lp_max_version = {
+                .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
+                .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
+        },
+        .lp_lock_ast            = user_ast,
+        .lp_blocking_ast        = user_bast,
+        .lp_unlock_ast          = user_unlock_ast,
+};
+
 static inline void user_dlm_drop_inode_ref(struct user_lock_res *lockres)
 {
         struct inode *inode;
@@ -283,10 +300,10 @@ static void user_dlm_unblock_lock(struct work_struct *work)
         int new_level, status;
         struct user_lock_res *lockres =
                 container_of(work, struct user_lock_res, l_work);
-        struct dlm_ctxt *dlm = dlm_ctxt_from_user_lockres(lockres);
+        struct ocfs2_cluster_connection *conn =
+                cluster_connection_from_user_lockres(lockres);
 
-        mlog(0, "processing lockres %.*s\n", lockres->l_namelen,
-             lockres->l_name);
+        mlog(0, "lockres %.*s\n", lockres->l_namelen, lockres->l_name);
 
         spin_lock(&lockres->l_lock);
 
@@ -304,17 +321,23 @@ static void user_dlm_unblock_lock(struct work_struct *work)
          * flag, and finally we might get another bast which re-queues
          * us before our ast for the downconvert is called. */
         if (!(lockres->l_flags & USER_LOCK_BLOCKED)) {
+                mlog(ML_BASTS, "lockres %.*s USER_LOCK_BLOCKED\n",
+                     lockres->l_namelen, lockres->l_name);
                 spin_unlock(&lockres->l_lock);
                 goto drop_ref;
         }
 
         if (lockres->l_flags & USER_LOCK_IN_TEARDOWN) {
+                mlog(ML_BASTS, "lockres %.*s USER_LOCK_IN_TEARDOWN\n",
+                     lockres->l_namelen, lockres->l_name);
                 spin_unlock(&lockres->l_lock);
                 goto drop_ref;
         }
 
         if (lockres->l_flags & USER_LOCK_BUSY) {
                 if (lockres->l_flags & USER_LOCK_IN_CANCEL) {
+                        mlog(ML_BASTS, "lockres %.*s USER_LOCK_IN_CANCEL\n",
+                             lockres->l_namelen, lockres->l_name);
                         spin_unlock(&lockres->l_lock);
                         goto drop_ref;
                 }
@@ -322,32 +345,31 @@ static void user_dlm_unblock_lock(struct work_struct *work)
                 lockres->l_flags |= USER_LOCK_IN_CANCEL;
                 spin_unlock(&lockres->l_lock);
 
-                status = dlmunlock(dlm,
-                                   &lockres->l_lksb,
-                                   LKM_CANCEL,
-                                   user_unlock_ast,
-                                   lockres);
-                if (status != DLM_NORMAL)
-                        user_log_dlm_error("dlmunlock", status, lockres);
+                status = ocfs2_dlm_unlock(conn, &lockres->l_lksb,
+                                          DLM_LKF_CANCEL);
+                if (status)
+                        user_log_dlm_error("ocfs2_dlm_unlock", status, lockres);
                 goto drop_ref;
         }
 
         /* If there are still incompat holders, we can exit safely
          * without worrying about re-queueing this lock as that will
          * happen on the last call to user_cluster_unlock. */
-        if ((lockres->l_blocking == LKM_EXMODE)
+        if ((lockres->l_blocking == DLM_LOCK_EX)
             && (lockres->l_ex_holders || lockres->l_ro_holders)) {
                 spin_unlock(&lockres->l_lock);
-                mlog(0, "can't downconvert for ex: ro = %u, ex = %u\n",
-                        lockres->l_ro_holders, lockres->l_ex_holders);
+                mlog(ML_BASTS, "lockres %.*s, EX/PR Holders %u,%u\n",
+                     lockres->l_namelen, lockres->l_name,
+                     lockres->l_ex_holders, lockres->l_ro_holders);
                 goto drop_ref;
         }
 
-        if ((lockres->l_blocking == LKM_PRMODE)
+        if ((lockres->l_blocking == DLM_LOCK_PR)
             && lockres->l_ex_holders) {
                 spin_unlock(&lockres->l_lock);
-                mlog(0, "can't downconvert for pr: ex = %u\n",
-                        lockres->l_ex_holders);
+                mlog(ML_BASTS, "lockres %.*s, EX Holders %u\n",
+                     lockres->l_namelen, lockres->l_name,
+                     lockres->l_ex_holders);
                 goto drop_ref;
         }
 
@@ -355,22 +377,17 @@ static void user_dlm_unblock_lock(struct work_struct *work)
         new_level = user_highest_compat_lock_level(lockres->l_blocking);
         lockres->l_requested = new_level;
         lockres->l_flags |= USER_LOCK_BUSY;
-        mlog(0, "Downconvert lock from %d to %d\n",
-                lockres->l_level, new_level);
+        mlog(ML_BASTS, "lockres %.*s, downconvert %d => %d\n",
+             lockres->l_namelen, lockres->l_name, lockres->l_level, new_level);
         spin_unlock(&lockres->l_lock);
 
         /* need lock downconvert request now... */
-        status = dlmlock(dlm,
-                         new_level,
-                         &lockres->l_lksb,
-                         LKM_CONVERT|LKM_VALBLK,
-                         lockres->l_name,
-                         lockres->l_namelen,
-                         user_ast,
-                         lockres,
-                         user_bast);
-        if (status != DLM_NORMAL) {
-                user_log_dlm_error("dlmlock", status, lockres);
+        status = ocfs2_dlm_lock(conn, new_level, &lockres->l_lksb,
+                                DLM_LKF_CONVERT|DLM_LKF_VALBLK,
+                                lockres->l_name,
+                                lockres->l_namelen);
+        if (status) {
+                user_log_dlm_error("ocfs2_dlm_lock", status, lockres);
                 user_recover_from_dlm_error(lockres);
         }
 
@@ -382,10 +399,10 @@ static inline void user_dlm_inc_holders(struct user_lock_res *lockres,
                                         int level)
 {
         switch(level) {
-        case LKM_EXMODE:
+        case DLM_LOCK_EX:
                 lockres->l_ex_holders++;
                 break;
-        case LKM_PRMODE:
+        case DLM_LOCK_PR:
                 lockres->l_ro_holders++;
                 break;
         default:
@@ -410,20 +427,19 @@ int user_dlm_cluster_lock(struct user_lock_res *lockres,
                           int lkm_flags)
 {
         int status, local_flags;
-        struct dlm_ctxt *dlm = dlm_ctxt_from_user_lockres(lockres);
+        struct ocfs2_cluster_connection *conn =
+                cluster_connection_from_user_lockres(lockres);
 
-        if (level != LKM_EXMODE &&
-            level != LKM_PRMODE) {
+        if (level != DLM_LOCK_EX &&
+            level != DLM_LOCK_PR) {
                 mlog(ML_ERROR, "lockres %.*s: invalid request!\n",
                      lockres->l_namelen, lockres->l_name);
                 status = -EINVAL;
                 goto bail;
         }
 
-        mlog(0, "lockres %.*s: asking for %s lock, passed flags = 0x%x\n",
-             lockres->l_namelen, lockres->l_name,
-             (level == LKM_EXMODE) ? "LKM_EXMODE" : "LKM_PRMODE",
-             lkm_flags);
+        mlog(ML_BASTS, "lockres %.*s, level %d, flags = 0x%x\n",
+             lockres->l_namelen, lockres->l_name, level, lkm_flags);
 
 again:
         if (signal_pending(current)) {
@@ -457,35 +473,26 @@ again:
         }
 
         if (level > lockres->l_level) {
-                local_flags = lkm_flags | LKM_VALBLK;
-                if (lockres->l_level != LKM_IVMODE)
-                        local_flags |= LKM_CONVERT;
+                local_flags = lkm_flags | DLM_LKF_VALBLK;
+                if (lockres->l_level != DLM_LOCK_IV)
+                        local_flags |= DLM_LKF_CONVERT;
 
                 lockres->l_requested = level;
                 lockres->l_flags |= USER_LOCK_BUSY;
                 spin_unlock(&lockres->l_lock);
 
-                BUG_ON(level == LKM_IVMODE);
-                BUG_ON(level == LKM_NLMODE);
+                BUG_ON(level == DLM_LOCK_IV);
+                BUG_ON(level == DLM_LOCK_NL);
 
                 /* call dlm_lock to upgrade lock now */
-                status = dlmlock(dlm,
-                                 level,
-                                 &lockres->l_lksb,
-                                 local_flags,
-                                 lockres->l_name,
-                                 lockres->l_namelen,
-                                 user_ast,
-                                 lockres,
-                                 user_bast);
-                if (status != DLM_NORMAL) {
-                        if ((lkm_flags & LKM_NOQUEUE) &&
-                            (status == DLM_NOTQUEUED))
-                                status = -EAGAIN;
-                        else {
-                                user_log_dlm_error("dlmlock", status, lockres);
-                                status = -EINVAL;
-                        }
+                status = ocfs2_dlm_lock(conn, level, &lockres->l_lksb,
+                                        local_flags, lockres->l_name,
+                                        lockres->l_namelen);
+                if (status) {
+                        if ((lkm_flags & DLM_LKF_NOQUEUE) &&
+                            (status != -EAGAIN))
+                                user_log_dlm_error("ocfs2_dlm_lock",
+                                                   status, lockres);
                         user_recover_from_dlm_error(lockres);
                         goto bail;
                 }
@@ -506,11 +513,11 @@ static inline void user_dlm_dec_holders(struct user_lock_res *lockres,
                                         int level)
 {
         switch(level) {
-        case LKM_EXMODE:
+        case DLM_LOCK_EX:
                 BUG_ON(!lockres->l_ex_holders);
                 lockres->l_ex_holders--;
                 break;
-        case LKM_PRMODE:
+        case DLM_LOCK_PR:
                 BUG_ON(!lockres->l_ro_holders);
                 lockres->l_ro_holders--;
                 break;
@@ -522,8 +529,8 @@ static inline void user_dlm_dec_holders(struct user_lock_res *lockres,
 void user_dlm_cluster_unlock(struct user_lock_res *lockres,
                              int level)
 {
-        if (level != LKM_EXMODE &&
-            level != LKM_PRMODE) {
+        if (level != DLM_LOCK_EX &&
+            level != DLM_LOCK_PR) {
                 mlog(ML_ERROR, "lockres %.*s: invalid request!\n",
                      lockres->l_namelen, lockres->l_name);
                 return;
@@ -540,33 +547,40 @@ void user_dlm_write_lvb(struct inode *inode,
                         unsigned int len)
 {
         struct user_lock_res *lockres = &DLMFS_I(inode)->ip_lockres;
-        char *lvb = lockres->l_lksb.lvb;
+        char *lvb;
 
         BUG_ON(len > DLM_LVB_LEN);
 
         spin_lock(&lockres->l_lock);
 
-        BUG_ON(lockres->l_level < LKM_EXMODE);
+        BUG_ON(lockres->l_level < DLM_LOCK_EX);
+        lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
         memcpy(lvb, val, len);
 
         spin_unlock(&lockres->l_lock);
 }
 
-void user_dlm_read_lvb(struct inode *inode,
-                       char *val,
-                       unsigned int len)
+ssize_t user_dlm_read_lvb(struct inode *inode,
+                          char *val,
+                          unsigned int len)
 {
         struct user_lock_res *lockres = &DLMFS_I(inode)->ip_lockres;
-        char *lvb = lockres->l_lksb.lvb;
+        char *lvb;
+        ssize_t ret = len;
 
         BUG_ON(len > DLM_LVB_LEN);
 
         spin_lock(&lockres->l_lock);
 
-        BUG_ON(lockres->l_level < LKM_PRMODE);
-        memcpy(val, lvb, len);
+        BUG_ON(lockres->l_level < DLM_LOCK_PR);
+        if (ocfs2_dlm_lvb_valid(&lockres->l_lksb)) {
+                lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+                memcpy(val, lvb, len);
+        } else
+                ret = 0;
 
         spin_unlock(&lockres->l_lock);
+        return ret;
 }
 
 void user_dlm_lock_res_init(struct user_lock_res *lockres,
@@ -576,9 +590,9 @@ void user_dlm_lock_res_init(struct user_lock_res *lockres,
 
         spin_lock_init(&lockres->l_lock);
         init_waitqueue_head(&lockres->l_event);
-        lockres->l_level = LKM_IVMODE;
-        lockres->l_requested = LKM_IVMODE;
-        lockres->l_blocking = LKM_IVMODE;
+        lockres->l_level = DLM_LOCK_IV;
+        lockres->l_requested = DLM_LOCK_IV;
+        lockres->l_blocking = DLM_LOCK_IV;
 
         /* should have been checked before getting here. */
         BUG_ON(dentry->d_name.len >= USER_DLM_LOCK_ID_MAX_LEN);
@@ -592,9 +606,10 @@ void user_dlm_lock_res_init(struct user_lock_res *lockres,
 int user_dlm_destroy_lock(struct user_lock_res *lockres)
 {
         int status = -EBUSY;
-        struct dlm_ctxt *dlm = dlm_ctxt_from_user_lockres(lockres);
+        struct ocfs2_cluster_connection *conn =
+                cluster_connection_from_user_lockres(lockres);
 
-        mlog(0, "asked to destroy %.*s\n", lockres->l_namelen, lockres->l_name);
+        mlog(ML_BASTS, "lockres %.*s\n", lockres->l_namelen, lockres->l_name);
 
         spin_lock(&lockres->l_lock);
         if (lockres->l_flags & USER_LOCK_IN_TEARDOWN) {
@@ -627,14 +642,9 @@ int user_dlm_destroy_lock(struct user_lock_res *lockres)
         lockres->l_flags |= USER_LOCK_BUSY;
         spin_unlock(&lockres->l_lock);
 
-        status = dlmunlock(dlm,
-                           &lockres->l_lksb,
-                           LKM_VALBLK,
-                           user_unlock_ast,
-                           lockres);
-        if (status != DLM_NORMAL) {
-                user_log_dlm_error("dlmunlock", status, lockres);
-                status = -EINVAL;
+        status = ocfs2_dlm_unlock(conn, &lockres->l_lksb, DLM_LKF_VALBLK);
+        if (status) {
+                user_log_dlm_error("ocfs2_dlm_unlock", status, lockres);
                 goto bail;
         }
 
@@ -645,32 +655,34 @@ bail:
         return status;
 }
 
-struct dlm_ctxt *user_dlm_register_context(struct qstr *name,
-                                           struct dlm_protocol_version *proto)
+static void user_dlm_recovery_handler_noop(int node_num,
+                                           void *recovery_data)
 {
-        struct dlm_ctxt *dlm;
-        u32 dlm_key;
-        char *domain;
-
-        domain = kmalloc(name->len + 1, GFP_NOFS);
-        if (!domain) {
-                mlog_errno(-ENOMEM);
-                return ERR_PTR(-ENOMEM);
-        }
+        /* We ignore recovery events */
+        return;
+}
 
-        dlm_key = crc32_le(0, name->name, name->len);
+void user_dlm_set_locking_protocol(void)
+{
+        ocfs2_stack_glue_set_max_proto_version(&user_dlm_lproto.lp_max_version);
+}
 
-        snprintf(domain, name->len + 1, "%.*s", name->len, name->name);
+struct ocfs2_cluster_connection *user_dlm_register(struct qstr *name)
+{
+        int rc;
+        struct ocfs2_cluster_connection *conn;
 
-        dlm = dlm_register_domain(domain, dlm_key, proto);
-        if (IS_ERR(dlm))
-                mlog_errno(PTR_ERR(dlm));
+        rc = ocfs2_cluster_connect_agnostic(name->name, name->len,
+                                            &user_dlm_lproto,
+                                            user_dlm_recovery_handler_noop,
+                                            NULL, &conn);
+        if (rc)
+                mlog_errno(rc);
 
-        kfree(domain);
-        return dlm;
+        return rc ? ERR_PTR(rc) : conn;
 }
 
-void user_dlm_unregister_context(struct dlm_ctxt *dlm)
+void user_dlm_unregister(struct ocfs2_cluster_connection *conn)
 {
-        dlm_unregister_domain(dlm);
+        ocfs2_cluster_disconnect(conn, 0);
 }
diff --git a/fs/ocfs2/dlm/userdlm.h b/fs/ocfs2/dlmfs/userdlm.h
index 0c3cc03c61fa..3b42d79531d7 100644
--- a/fs/ocfs2/dlm/userdlm.h
+++ b/fs/ocfs2/dlmfs/userdlm.h
@@ -57,7 +57,7 @@ struct user_lock_res {
         int                      l_level;
         unsigned int             l_ro_holders;
         unsigned int             l_ex_holders;
-        struct dlm_lockstatus    l_lksb;
+        struct ocfs2_dlm_lksb    l_lksb;
 
         int                      l_requested;
         int                      l_blocking;
@@ -80,15 +80,15 @@ void user_dlm_cluster_unlock(struct user_lock_res *lockres,
 void user_dlm_write_lvb(struct inode *inode,
                         const char *val,
                         unsigned int len);
-void user_dlm_read_lvb(struct inode *inode,
-                       char *val,
-                       unsigned int len);
-struct dlm_ctxt *user_dlm_register_context(struct qstr *name,
-                                           struct dlm_protocol_version *proto);
-void user_dlm_unregister_context(struct dlm_ctxt *dlm);
+ssize_t user_dlm_read_lvb(struct inode *inode,
+                          char *val,
+                          unsigned int len);
+struct ocfs2_cluster_connection *user_dlm_register(struct qstr *name);
+void user_dlm_unregister(struct ocfs2_cluster_connection *conn);
+void user_dlm_set_locking_protocol(void);
 
 struct dlmfs_inode_private {
-        struct dlm_ctxt             *ip_dlm;
+        struct ocfs2_cluster_connection *ip_conn;
 
         struct user_lock_res ip_lockres; /* unused for directories. */
         struct inode         *ip_parent;
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 0d38d67194cb..50c4ee805da4 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -297,6 +297,11 @@ static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres)
                 lockres->l_type == OCFS2_LOCK_TYPE_OPEN;
 }
 
+static inline struct ocfs2_lock_res *ocfs2_lksb_to_lock_res(struct ocfs2_dlm_lksb *lksb)
+{
+        return container_of(lksb, struct ocfs2_lock_res, l_lksb);
+}
+
 static inline struct inode *ocfs2_lock_res_inode(struct ocfs2_lock_res *lockres)
 {
         BUG_ON(!ocfs2_is_inode_lock(lockres));
@@ -875,6 +880,14 @@ static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lo
                 lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
 
         lockres->l_level = lockres->l_requested;
+
+        /*
+         * We set the OCFS2_LOCK_UPCONVERT_FINISHING flag before clearing
+         * the OCFS2_LOCK_BUSY flag to prevent the dc thread from
+         * downconverting the lock before the upconvert has fully completed.
+         */
+        lockres_or_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
+
         lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
 
         mlog_exit_void();
@@ -907,8 +920,6 @@ static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
 
         assert_spin_locked(&lockres->l_lock);
 
-        lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
-
         if (level > lockres->l_blocking) {
                 /* only schedule a downconvert if we haven't already scheduled
                  * one that goes low enough to satisfy the level we're
@@ -921,6 +932,13 @@ static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
                 lockres->l_blocking = level;
         }
 
+        mlog(ML_BASTS, "lockres %s, block %d, level %d, l_block %d, dwn %d\n",
+             lockres->l_name, level, lockres->l_level, lockres->l_blocking,
+             needs_downconvert);
+
+        if (needs_downconvert)
+                lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
+
         mlog_exit(needs_downconvert);
         return needs_downconvert;
 }
@@ -1031,18 +1049,17 @@ static unsigned int lockres_set_pending(struct ocfs2_lock_res *lockres)
         return lockres->l_pending_gen;
 }
 
-
-static void ocfs2_blocking_ast(void *opaque, int level)
+static void ocfs2_blocking_ast(struct ocfs2_dlm_lksb *lksb, int level)
 {
-        struct ocfs2_lock_res *lockres = opaque;
+        struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
         struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
         int needs_downconvert;
         unsigned long flags;
 
         BUG_ON(level <= DLM_LOCK_NL);
 
-        mlog(0, "BAST fired for lockres %s, blocking %d, level %d type %s\n",
-             lockres->l_name, level, lockres->l_level,
+        mlog(ML_BASTS, "BAST fired for lockres %s, blocking %d, level %d, "
+             "type %s\n", lockres->l_name, level, lockres->l_level,
              ocfs2_lock_type_string(lockres->l_type));
 
         /*
@@ -1063,9 +1080,9 @@ static void ocfs2_blocking_ast(void *opaque, int level)
         ocfs2_wake_downconvert_thread(osb);
 }
 
-static void ocfs2_locking_ast(void *opaque)
+static void ocfs2_locking_ast(struct ocfs2_dlm_lksb *lksb)
 {
-        struct ocfs2_lock_res *lockres = opaque;
+        struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
         struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
         unsigned long flags;
         int status;
@@ -1086,6 +1103,10 @@ static void ocfs2_locking_ast(void *opaque)
                 return;
         }
 
+        mlog(ML_BASTS, "AST fired for lockres %s, action %d, unlock %d, "
+             "level %d => %d\n", lockres->l_name, lockres->l_action,
+             lockres->l_unlock_action, lockres->l_level, lockres->l_requested);
+
         switch(lockres->l_action) {
         case OCFS2_AST_ATTACH:
                 ocfs2_generic_handle_attach_action(lockres);
@@ -1098,8 +1119,8 @@ static void ocfs2_locking_ast(void *opaque)
                 ocfs2_generic_handle_downconvert_action(lockres);
                 break;
         default:
-                mlog(ML_ERROR, "lockres %s: ast fired with invalid action: %u "
-                     "lockres flags = 0x%lx, unlock action: %u\n",
+                mlog(ML_ERROR, "lockres %s: AST fired with invalid action: %u, "
+                     "flags 0x%lx, unlock: %u\n",
                      lockres->l_name, lockres->l_action, lockres->l_flags,
                      lockres->l_unlock_action);
                 BUG();
@@ -1125,6 +1146,88 @@ out:
         spin_unlock_irqrestore(&lockres->l_lock, flags);
 }
 
+static void ocfs2_unlock_ast(struct ocfs2_dlm_lksb *lksb, int error)
+{
+        struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
+        unsigned long flags;
+
+        mlog_entry_void();
+
+        mlog(ML_BASTS, "UNLOCK AST fired for lockres %s, action = %d\n",
+             lockres->l_name, lockres->l_unlock_action);
+
+        spin_lock_irqsave(&lockres->l_lock, flags);
+        if (error) {
+                mlog(ML_ERROR, "Dlm passes error %d for lock %s, "
+                     "unlock_action %d\n", error, lockres->l_name,
+                     lockres->l_unlock_action);
+                spin_unlock_irqrestore(&lockres->l_lock, flags);
+                mlog_exit_void();
+                return;
+        }
+
+        switch(lockres->l_unlock_action) {
+        case OCFS2_UNLOCK_CANCEL_CONVERT:
+                mlog(0, "Cancel convert success for %s\n", lockres->l_name);
+                lockres->l_action = OCFS2_AST_INVALID;
+                /* Downconvert thread may have requeued this lock, we
+                 * need to wake it. */
+                if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
+                        ocfs2_wake_downconvert_thread(ocfs2_get_lockres_osb(lockres));
+                break;
+        case OCFS2_UNLOCK_DROP_LOCK:
+                lockres->l_level = DLM_LOCK_IV;
+                break;
+        default:
+                BUG();
+        }
+
+        lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
+        lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
+        wake_up(&lockres->l_event);
+        spin_unlock_irqrestore(&lockres->l_lock, flags);
+
+        mlog_exit_void();
+}
+
+/*
+ * This is the filesystem locking protocol.  It provides the lock handling
+ * hooks for the underlying DLM.  It has a maximum version number.
+ * The version number allows interoperability with systems running at
+ * the same major number and an equal or smaller minor number.
+ *
+ * Whenever the filesystem does new things with locks (adds or removes a
+ * lock, orders them differently, does different things underneath a lock),
+ * the version must be changed.  The protocol is negotiated when joining
+ * the dlm domain.  A node may join the domain if its major version is
+ * identical to all other nodes and its minor version is greater than
+ * or equal to all other nodes.  When its minor version is greater than
+ * the other nodes, it will run at the minor version specified by the
+ * other nodes.
+ *
+ * If a locking change is made that will not be compatible with older
+ * versions, the major number must be increased and the minor version set
+ * to zero.  If a change merely adds a behavior that can be disabled when
+ * speaking to older versions, the minor version must be increased.  If a
+ * change adds a fully backwards compatible change (eg, LVB changes that
+ * are just ignored by older versions), the version does not need to be
+ * updated.
+ */
+static struct ocfs2_locking_protocol lproto = {
+        .lp_max_version = {
+                .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
+                .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
+        },
+        .lp_lock_ast            = ocfs2_locking_ast,
+        .lp_blocking_ast        = ocfs2_blocking_ast,
+        .lp_unlock_ast          = ocfs2_unlock_ast,
+};
+
+void ocfs2_set_locking_protocol(void)
+{
+        ocfs2_stack_glue_set_max_proto_version(&lproto.lp_max_version);
+}
+
 static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
                                                 int convert)
 {
@@ -1133,6 +1236,7 @@ static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
         mlog_entry_void();
         spin_lock_irqsave(&lockres->l_lock, flags);
         lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
+        lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
         if (convert)
                 lockres->l_action = OCFS2_AST_INVALID;
         else
@@ -1179,8 +1283,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb,
                              &lockres->l_lksb,
                              dlm_flags,
                              lockres->l_name,
-                             OCFS2_LOCK_ID_MAX_LEN - 1,
-                             lockres);
+                             OCFS2_LOCK_ID_MAX_LEN - 1);
         lockres_clear_pending(lockres, gen, osb);
         if (ret) {
                 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
@@ -1323,13 +1426,13 @@ static int __ocfs2_cluster_lock(struct ocfs2_super *osb,
 again:
         wait = 0;
 
+        spin_lock_irqsave(&lockres->l_lock, flags);
+
         if (catch_signals && signal_pending(current)) {
                 ret = -ERESTARTSYS;
-                goto out;
+                goto unlock;
         }
 
-        spin_lock_irqsave(&lockres->l_lock, flags);
-
         mlog_bug_on_msg(lockres->l_flags & OCFS2_LOCK_FREEING,
                         "Cluster lock called on freeing lockres %s! flags "
                         "0x%lx\n", lockres->l_name, lockres->l_flags);
@@ -1346,6 +1449,25 @@ again:
                 goto unlock;
         }
 
+        if (lockres->l_flags & OCFS2_LOCK_UPCONVERT_FINISHING) {
+                /*
+                 * We've upconverted. If the lock now has a level we can
+                 * work with, we take it. If, however, the lock is not at the
+                 * required level, we go thru the full cycle. One way this could
+                 * happen is if a process requesting an upconvert to PR is
+                 * closely followed by another requesting upconvert to an EX.
+                 * If the process requesting EX lands here, we want it to
+                 * continue attempting to upconvert and let the process
+                 * requesting PR take the lock.
+                 * If multiple processes request upconvert to PR, the first one
+                 * here will take the lock. The others will have to go thru the
+                 * OCFS2_LOCK_BLOCKED check to ensure that there is no pending
+                 * downconvert request.
+                 */
+                if (level <= lockres->l_level)
+                        goto update_holders;
+        }
+
         if (lockres->l_flags & OCFS2_LOCK_BLOCKED &&
             !ocfs2_may_continue_on_blocked_lock(lockres, level)) {
                 /* is the lock is currently blocked on behalf of
@@ -1383,7 +1505,7 @@ again:
                 BUG_ON(level == DLM_LOCK_IV);
                 BUG_ON(level == DLM_LOCK_NL);
 
-                mlog(0, "lock %s, convert from %d to level = %d\n",
+                mlog(ML_BASTS, "lockres %s, convert from %d to %d\n",
                      lockres->l_name, lockres->l_level, level);
 
                 /* call dlm_lock to upgrade lock now */
@@ -1392,8 +1514,7 @@ again:
                                      &lockres->l_lksb,
                                      lkm_flags,
                                      lockres->l_name,
-                                     OCFS2_LOCK_ID_MAX_LEN - 1,
-                                     lockres);
+                                     OCFS2_LOCK_ID_MAX_LEN - 1);
                 lockres_clear_pending(lockres, gen, osb);
                 if (ret) {
                         if (!(lkm_flags & DLM_LKF_NOQUEUE) ||
@@ -1416,11 +1537,14 @@ again:
                 goto again;
         }
 
+update_holders:
         /* Ok, if we get here then we're good to go. */
         ocfs2_inc_holders(lockres, level);
 
         ret = 0;
 unlock:
+        lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
+
         spin_unlock_irqrestore(&lockres->l_lock, flags);
 out:
         /*
@@ -1757,7 +1881,7 @@ out:
  * ocfs2_file_lock() and ocfs2_file_unlock() map to a single pair of
  * flock() calls. The locking approach this requires is sufficiently
  * different from all other cluster lock types that we implement a
- * seperate path to the "low-level" dlm calls. In particular:
+ * separate path to the "low-level" dlm calls. In particular:
  *
  * - No optimization of lock levels is done - we take at exactly
  *   what's been requested.
@@ -1827,8 +1951,7 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock)
         spin_unlock_irqrestore(&lockres->l_lock, flags);
 
         ret = ocfs2_dlm_lock(osb->cconn, level, &lockres->l_lksb, lkm_flags,
-                             lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1,
-                             lockres);
+                             lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1);
         if (ret) {
                 if (!trylock || (ret != -EAGAIN)) {
                         ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
@@ -1855,7 +1978,7 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock)
                  * outstanding lock request, so a cancel convert is
                  * required. We intentionally overwrite 'ret' - if the
                  * cancel fails and the lock was granted, it's easier
-                 * to just bubble sucess back up to the user.
+                 * to just bubble success back up to the user.
                  */
                 ret = ocfs2_flock_handle_signal(lockres, level);
         } else if (!ret && (level > lockres->l_level)) {
@@ -2957,7 +3080,7 @@ int ocfs2_dlm_init(struct ocfs2_super *osb)
         status = ocfs2_cluster_connect(osb->osb_cluster_stack,
                                        osb->uuid_str,
                                        strlen(osb->uuid_str),
-                                       ocfs2_do_node_down, osb,
+                                       &lproto, ocfs2_do_node_down, osb,
                                        &conn);
         if (status) {
                 mlog_errno(status);
@@ -3024,50 +3147,6 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
         mlog_exit_void();
 }
 
-static void ocfs2_unlock_ast(void *opaque, int error)
-{
-        struct ocfs2_lock_res *lockres = opaque;
-        unsigned long flags;
-
-        mlog_entry_void();
-
-        mlog(0, "UNLOCK AST called on lock %s, action = %d\n", lockres->l_name,
-             lockres->l_unlock_action);
-
-        spin_lock_irqsave(&lockres->l_lock, flags);
-        if (error) {
-                mlog(ML_ERROR, "Dlm passes error %d for lock %s, "
-                     "unlock_action %d\n", error, lockres->l_name,
-                     lockres->l_unlock_action);
-                spin_unlock_irqrestore(&lockres->l_lock, flags);
-                mlog_exit_void();
-                return;
-        }
-
-        switch(lockres->l_unlock_action) {
-        case OCFS2_UNLOCK_CANCEL_CONVERT:
-                mlog(0, "Cancel convert success for %s\n", lockres->l_name);
-                lockres->l_action = OCFS2_AST_INVALID;
-                /* Downconvert thread may have requeued this lock, we
-                 * need to wake it. */
-                if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
-                        ocfs2_wake_downconvert_thread(ocfs2_get_lockres_osb(lockres));
-                break;
-        case OCFS2_UNLOCK_DROP_LOCK:
-                lockres->l_level = DLM_LOCK_IV;
-                break;
-        default:
-                BUG();
-        }
-
-        lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
-        lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
-        wake_up(&lockres->l_event);
-        spin_unlock_irqrestore(&lockres->l_lock, flags);
-
-        mlog_exit_void();
-}
-
 static int ocfs2_drop_lock(struct ocfs2_super *osb,
                            struct ocfs2_lock_res *lockres)
 {
@@ -3135,8 +3214,7 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb,
 
         mlog(0, "lock %s\n", lockres->l_name);
 
-        ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb, lkm_flags,
-                               lockres);
+        ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb, lkm_flags);
         if (ret) {
                 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres);
                 mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags);
@@ -3155,7 +3233,7 @@ out:
 /* Mark the lockres as being dropped. It will no longer be
  * queued if blocking, but we still may have to wait on it
  * being dequeued from the downconvert thread before we can consider
- * it safe to drop. 
+ * it safe to drop.
  *
  * You can *not* attempt to call cluster_lock on this lockres anymore. */
 void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres)
@@ -3244,13 +3322,20 @@ static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres,
         BUG_ON(lockres->l_blocking <= DLM_LOCK_NL);
 
         if (lockres->l_level <= new_level) {
-                mlog(ML_ERROR, "lockres->l_level (%d) <= new_level (%d)\n",
-                     lockres->l_level, new_level);
+                mlog(ML_ERROR, "lockres %s, lvl %d <= %d, blcklst %d, mask %d, "
+                     "type %d, flags 0x%lx, hold %d %d, act %d %d, req %d, "
+                     "block %d, pgen %d\n", lockres->l_name, lockres->l_level,
+                     new_level, list_empty(&lockres->l_blocked_list),
+                     list_empty(&lockres->l_mask_waiters), lockres->l_type,
+                     lockres->l_flags, lockres->l_ro_holders,
+                     lockres->l_ex_holders, lockres->l_action,
+                     lockres->l_unlock_action, lockres->l_requested,
+                     lockres->l_blocking, lockres->l_pending_gen);
                 BUG();
         }
 
-        mlog(0, "lock %s, new_level = %d, l_blocking = %d\n",
-             lockres->l_name, new_level, lockres->l_blocking);
+        mlog(ML_BASTS, "lockres %s, level %d => %d, blocking %d\n",
+             lockres->l_name, lockres->l_level, new_level, lockres->l_blocking);
 
         lockres->l_action = OCFS2_AST_DOWNCONVERT;
         lockres->l_requested = new_level;
@@ -3269,6 +3354,9 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
 
         mlog_entry_void();
 
+        mlog(ML_BASTS, "lockres %s, level %d => %d\n", lockres->l_name,
+             lockres->l_level, new_level);
+
         if (lvb)
                 dlm_flags |= DLM_LKF_VALBLK;
 
@@ -3277,8 +3365,7 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
                              &lockres->l_lksb,
                              dlm_flags,
                              lockres->l_name,
-                             OCFS2_LOCK_ID_MAX_LEN - 1,
-                             lockres);
+                             OCFS2_LOCK_ID_MAX_LEN - 1);
         lockres_clear_pending(lockres, generation, osb);
         if (ret) {
                 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
@@ -3299,14 +3386,12 @@ static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
         assert_spin_locked(&lockres->l_lock);
 
         mlog_entry_void();
-        mlog(0, "lock %s\n", lockres->l_name);
 
         if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) {
                 /* If we're already trying to cancel a lock conversion
                  * then just drop the spinlock and allow the caller to
                  * requeue this lock. */
-
-                mlog(0, "Lockres %s, skip convert\n", lockres->l_name);
+                mlog(ML_BASTS, "lockres %s, skip convert\n", lockres->l_name);
                 return 0;
         }
 
@@ -3321,6 +3406,8 @@ static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
                         "lock %s, invalid flags: 0x%lx\n",
                         lockres->l_name, lockres->l_flags);
 
+        mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
+
         return 1;
 }
 
@@ -3330,16 +3417,15 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb,
         int ret;
 
         mlog_entry_void();
-        mlog(0, "lock %s\n", lockres->l_name);
 
         ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb,
-                               DLM_LKF_CANCEL, lockres);
+                               DLM_LKF_CANCEL);
         if (ret) {
                 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres);
                 ocfs2_recover_from_dlm_error(lockres, 0);
         }
 
-        mlog(0, "lock %s return from ocfs2_dlm_unlock\n", lockres->l_name);
+        mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
 
         mlog_exit(ret);
         return ret;
@@ -3352,6 +3438,7 @@ static int ocfs2_unblock_lock(struct ocfs2_super *osb,
         unsigned long flags;
         int blocking;
         int new_level;
+        int level;
         int ret = 0;
         int set_lvb = 0;
         unsigned int gen;
@@ -3360,9 +3447,17 @@ static int ocfs2_unblock_lock(struct ocfs2_super *osb,
 
         spin_lock_irqsave(&lockres->l_lock, flags);
 
-        BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED));
-
 recheck:
+        /*
+         * Is it still blocking? If not, we have no more work to do.
+         */
+        if (!(lockres->l_flags & OCFS2_LOCK_BLOCKED)) {
+                BUG_ON(lockres->l_blocking != DLM_LOCK_NL);
+                spin_unlock_irqrestore(&lockres->l_lock, flags);
+                ret = 0;
+                goto leave;
+        }
+
         if (lockres->l_flags & OCFS2_LOCK_BUSY) {
                 /* XXX
                  * This is a *big* race.  The OCFS2_LOCK_PENDING flag
@@ -3387,8 +3482,11 @@ recheck:
                  * at the same time they set OCFS2_DLM_BUSY.  They must
                  * clear OCFS2_DLM_PENDING after dlm_lock() returns.
                  */
-                if (lockres->l_flags & OCFS2_LOCK_PENDING)
+                if (lockres->l_flags & OCFS2_LOCK_PENDING) {
+                        mlog(ML_BASTS, "lockres %s, ReQ: Pending\n",
+                             lockres->l_name);
                         goto leave_requeue;
+                }
 
                 ctl->requeue = 1;
                 ret = ocfs2_prepare_cancel_convert(osb, lockres);
@@ -3401,31 +3499,70 @@ recheck:
                 goto leave;
         }
 
+        /*
+         * This prevents livelocks. OCFS2_LOCK_UPCONVERT_FINISHING flag is
+         * set when the ast is received for an upconvert just before the
+         * OCFS2_LOCK_BUSY flag is cleared. Now if the fs received a bast
+         * on the heels of the ast, we want to delay the downconvert just
+         * enough to allow the up requestor to do its task. Because this
+         * lock is in the blocked queue, the lock will be downconverted
+         * as soon as the requestor is done with the lock.
+         */
+        if (lockres->l_flags & OCFS2_LOCK_UPCONVERT_FINISHING)
+                goto leave_requeue;
+
+        /*
+         * How can we block and yet be at NL?  We were trying to upconvert
+         * from NL and got canceled.  The code comes back here, and now
+         * we notice and clear BLOCKING.
+         */
+        if (lockres->l_level == DLM_LOCK_NL) {
+                BUG_ON(lockres->l_ex_holders || lockres->l_ro_holders);
+                mlog(ML_BASTS, "lockres %s, Aborting dc\n", lockres->l_name);
+                lockres->l_blocking = DLM_LOCK_NL;
+                lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED);
+                spin_unlock_irqrestore(&lockres->l_lock, flags);
+                goto leave;
+        }
+
         /* if we're blocking an exclusive and we have *any* holders,
          * then requeue. */
         if ((lockres->l_blocking == DLM_LOCK_EX)
-            && (lockres->l_ex_holders || lockres->l_ro_holders))
+            && (lockres->l_ex_holders || lockres->l_ro_holders)) {
+                mlog(ML_BASTS, "lockres %s, ReQ: EX/PR Holders %u,%u\n",
+                     lockres->l_name, lockres->l_ex_holders,
+                     lockres->l_ro_holders);
                 goto leave_requeue;
+        }
 
         /* If it's a PR we're blocking, then only
          * requeue if we've got any EX holders */
         if (lockres->l_blocking == DLM_LOCK_PR &&
-            lockres->l_ex_holders)
+            lockres->l_ex_holders) {
+                mlog(ML_BASTS, "lockres %s, ReQ: EX Holders %u\n",
+                     lockres->l_name, lockres->l_ex_holders);
                 goto leave_requeue;
+        }
 
         /*
          * Can we get a lock in this state if the holder counts are
          * zero? The meta data unblock code used to check this.
          */
         if ((lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
-            && (lockres->l_flags & OCFS2_LOCK_REFRESHING))
+            && (lockres->l_flags & OCFS2_LOCK_REFRESHING)) {
+                mlog(ML_BASTS, "lockres %s, ReQ: Lock Refreshing\n",
+                     lockres->l_name);
                 goto leave_requeue;
+        }
 
         new_level = ocfs2_highest_compat_lock_level(lockres->l_blocking);
 
         if (lockres->l_ops->check_downconvert
-            && !lockres->l_ops->check_downconvert(lockres, new_level))
+            && !lockres->l_ops->check_downconvert(lockres, new_level)) {
+                mlog(ML_BASTS, "lockres %s, ReQ: Checkpointing\n",
+                     lockres->l_name);
                 goto leave_requeue;
+        }
 
         /* If we get here, then we know that there are no more
          * incompatible holders (and anyone asking for an incompatible
@@ -3438,17 +3575,24 @@ recheck:
          * may sleep, so we save off a copy of what we're blocking as
          * it may change while we're not holding the spin lock. */
         blocking = lockres->l_blocking;
+        level = lockres->l_level;
         spin_unlock_irqrestore(&lockres->l_lock, flags);
 
         ctl->unblock_action = lockres->l_ops->downconvert_worker(lockres, blocking);
 
-        if (ctl->unblock_action == UNBLOCK_STOP_POST)
+        if (ctl->unblock_action == UNBLOCK_STOP_POST) {
+                mlog(ML_BASTS, "lockres %s, UNBLOCK_STOP_POST\n",
+                     lockres->l_name);
                 goto leave;
+        }
 
         spin_lock_irqsave(&lockres->l_lock, flags);
-        if (blocking != lockres->l_blocking) {
+        if ((blocking != lockres->l_blocking) || (level != lockres->l_level)) {
                 /* If this changed underneath us, then we can't drop
                  * it just yet. */
+                mlog(ML_BASTS, "lockres %s, block=%d:%d, level=%d:%d, "
+                     "Recheck\n", lockres->l_name, blocking,
+                     lockres->l_blocking, level, lockres->l_level);
                 goto recheck;
         }
 
@@ -3843,45 +3987,6 @@ void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex)
                 ocfs2_cluster_unlock(osb, lockres, level);
 }
 
-/*
- * This is the filesystem locking protocol.  It provides the lock handling
- * hooks for the underlying DLM.  It has a maximum version number.
- * The version number allows interoperability with systems running at
- * the same major number and an equal or smaller minor number.
- *
- * Whenever the filesystem does new things with locks (adds or removes a
- * lock, orders them differently, does different things underneath a lock),
- * the version must be changed.  The protocol is negotiated when joining
- * the dlm domain.  A node may join the domain if its major version is
- * identical to all other nodes and its minor version is greater than
- * or equal to all other nodes.  When its minor version is greater than
- * the other nodes, it will run at the minor version specified by the
- * other nodes.
- *
- * If a locking change is made that will not be compatible with older
- * versions, the major number must be increased and the minor version set
- * to zero.  If a change merely adds a behavior that can be disabled when
- * speaking to older versions, the minor version must be increased.  If a
- * change adds a fully backwards compatible change (eg, LVB changes that
- * are just ignored by older versions), the version does not need to be
- * updated.
- */
-static struct ocfs2_locking_protocol lproto = {
-        .lp_max_version = {
-                .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
-                .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
-        },
-        .lp_lock_ast            = ocfs2_locking_ast,
-        .lp_blocking_ast        = ocfs2_blocking_ast,
-        .lp_unlock_ast          = ocfs2_unlock_ast,
-};
-
-void ocfs2_set_locking_protocol(void)
-{
-        ocfs2_stack_glue_set_locking_protocol(&lproto);
-}
-
-
 static void ocfs2_process_blocked_lock(struct ocfs2_super *osb,
                                        struct ocfs2_lock_res *lockres)
 {
@@ -3898,7 +4003,7 @@ static void ocfs2_process_blocked_lock(struct ocfs2_super *osb,
         BUG_ON(!lockres);
         BUG_ON(!lockres->l_ops);
 
-        mlog(0, "lockres %s blocked.\n", lockres->l_name);
+        mlog(ML_BASTS, "lockres %s blocked\n", lockres->l_name);
 
         /* Detect whether a lock has been marked as going away while
          * the downconvert thread was processing other things. A lock can
@@ -3921,7 +4026,7 @@ unqueue:
         } else
                 ocfs2_schedule_blocked_lock(osb, lockres);
 
-        mlog(0, "lockres %s, requeue = %s.\n", lockres->l_name,
+        mlog(ML_BASTS, "lockres %s, requeue = %s.\n", lockres->l_name,
              ctl.requeue ? "yes" : "no");
         spin_unlock_irqrestore(&lockres->l_lock, flags);
 
@@ -3943,7 +4048,7 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
                 /* Do not schedule a lock for downconvert when it's on
                  * the way to destruction - any nodes wanting access
                  * to the resource will get it soon. */
-                mlog(0, "Lockres %s won't be scheduled: flags 0x%lx\n",
+                mlog(ML_BASTS, "lockres %s won't be scheduled: flags 0x%lx\n",
                      lockres->l_name, lockres->l_flags);
                 return;
         }
diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c
index 15713cbb865c..19ad145d2af3 100644
--- a/fs/ocfs2/export.c
+++ b/fs/ocfs2/export.c
@@ -239,7 +239,7 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len,
                 mlog(0, "Encoding parent: blkno: %llu, generation: %u\n",
                      (unsigned long long)blkno, generation);
         }
-        
+
         *max_len = len;
 
 bail:
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index 843db64e9d4a..09e3fdfa6d33 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -24,6 +24,7 @@
 
 #include <linux/fs.h>
 #include <linux/init.h>
+#include <linux/slab.h>
 #include <linux/types.h>
 #include <linux/fiemap.h>
 
@@ -37,6 +38,7 @@
 #include "extent_map.h"
 #include "inode.h"
 #include "super.h"
+#include "symlink.h"
 
 #include "buffer_head_io.h"
 
@@ -191,7 +193,7 @@ static int ocfs2_try_to_merge_extent_map(struct ocfs2_extent_map_item *emi,
                 emi->ei_clusters += ins->ei_clusters;
                 return 1;
         } else if ((ins->ei_phys + ins->ei_clusters) == emi->ei_phys &&
-                   (ins->ei_cpos + ins->ei_clusters) == emi->ei_phys &&
+                   (ins->ei_cpos + ins->ei_clusters) == emi->ei_cpos &&
                    ins->ei_flags == emi->ei_flags) {
                 emi->ei_phys = ins->ei_phys;
                 emi->ei_cpos = ins->ei_cpos;
@@ -452,7 +454,7 @@ static int ocfs2_get_clusters_nocache(struct inode *inode,
         if (i == -1) {
                 /*
                  * Holes can be larger than the maximum size of an
-                 * extent, so we return their lengths in a seperate
+                 * extent, so we return their lengths in a separate
                  * field.
                  */
                 if (hole_len) {
@@ -703,6 +705,12 @@ out:
         return ret;
 }
 
+/*
+ * The ocfs2_fiemap_inline() may be a little bit misleading, since
+ * it not only handles the fiemap for inlined files, but also deals
+ * with the fast symlink, cause they have no difference for extent
+ * mapping per se.
+ */
 static int ocfs2_fiemap_inline(struct inode *inode, struct buffer_head *di_bh,
                                struct fiemap_extent_info *fieinfo,
                                u64 map_start)
@@ -715,11 +723,18 @@ static int ocfs2_fiemap_inline(struct inode *inode, struct buffer_head *di_bh,
         struct ocfs2_inode_info *oi = OCFS2_I(inode);
 
         di = (struct ocfs2_dinode *)di_bh->b_data;
-        id_count = le16_to_cpu(di->id2.i_data.id_count);
+        if (ocfs2_inode_is_fast_symlink(inode))
+                id_count = ocfs2_fast_symlink_chars(inode->i_sb);
+        else
+                id_count = le16_to_cpu(di->id2.i_data.id_count);
 
         if (map_start < id_count) {
                 phys = oi->ip_blkno << inode->i_sb->s_blocksize_bits;
-                phys += offsetof(struct ocfs2_dinode, id2.i_data.id_data);
+                if (ocfs2_inode_is_fast_symlink(inode))
+                        phys += offsetof(struct ocfs2_dinode, id2.i_symlink);
+                else
+                        phys += offsetof(struct ocfs2_dinode,
+                                         id2.i_data.id_data);
 
                 ret = fiemap_fill_next_extent(fieinfo, 0, phys, id_count,
                                               flags);
@@ -756,9 +771,10 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
         down_read(&OCFS2_I(inode)->ip_alloc_sem);
 
         /*
-         * Handle inline-data separately.
+         * Handle inline-data and fast symlink separately.
          */
-        if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
+        if ((OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) ||
+            ocfs2_inode_is_fast_symlink(inode)) {
                 ret = ocfs2_fiemap_inline(inode, di_bh, fieinfo, map_start);
                 goto out_unlock;
         }
@@ -786,6 +802,8 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
                 fe_flags = 0;
                 if (rec.e_flags & OCFS2_EXT_UNWRITTEN)
                         fe_flags |= FIEMAP_EXTENT_UNWRITTEN;
+                if (rec.e_flags & OCFS2_EXT_REFCOUNTED)
+                        fe_flags |= FIEMAP_EXTENT_SHARED;
                 if (is_last)
                         fe_flags |= FIEMAP_EXTENT_LAST;
                 len_bytes = (u64)le16_to_cpu(rec.e_leaf_clusters) << osb->s_clustersize_bits;
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index de059f490586..a5fbd9cea968 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -107,6 +107,9 @@ static int ocfs2_file_open(struct inode *inode, struct file *file)
         mlog_entry("(0x%p, 0x%p, '%.*s')\n", inode, file,
                    file->f_path.dentry->d_name.len, file->f_path.dentry->d_name.name);
 
+        if (file->f_mode & FMODE_WRITE)
+                dquot_initialize(inode);
+
         spin_lock(&oi->ip_lock);
 
         /* Check that the inode hasn't been wiped from disk by another
@@ -629,11 +632,10 @@ restart_all:
         }
 
 restarted_transaction:
-        if (vfs_dq_alloc_space_nodirty(inode, ocfs2_clusters_to_bytes(osb->sb,
-            clusters_to_add))) {
-                status = -EDQUOT;
+        status = dquot_alloc_space_nodirty(inode,
+                        ocfs2_clusters_to_bytes(osb->sb, clusters_to_add));
+        if (status)
                 goto leave;
-        }
         did_quota = 1;
 
         /* reserve a write to the file entry early on - that we if we
@@ -674,7 +676,7 @@ restarted_transaction:
         clusters_to_add -= (OCFS2_I(inode)->ip_clusters - prev_clusters);
         spin_unlock(&OCFS2_I(inode)->ip_lock);
         /* Release unused quota reservation */
-        vfs_dq_free_space(inode,
+        dquot_free_space(inode,
                         ocfs2_clusters_to_bytes(osb->sb, clusters_to_add));
         did_quota = 0;
 
@@ -682,6 +684,7 @@ restarted_transaction:
                 if (why == RESTART_META) {
                         mlog(0, "restarting function.\n");
                         restart_func = 1;
+                        status = 0;
                 } else {
                         BUG_ON(why != RESTART_TRANS);
 
@@ -710,7 +713,7 @@ restarted_transaction:
 
 leave:
         if (status < 0 && did_quota)
-                vfs_dq_free_space(inode,
+                dquot_free_space(inode,
                         ocfs2_clusters_to_bytes(osb->sb, clusters_to_add));
         if (handle) {
                 ocfs2_commit_trans(osb, handle);
@@ -749,7 +752,7 @@ static int ocfs2_write_zero_page(struct inode *inode,
         int ret;
 
         offset = (size & (PAGE_CACHE_SIZE-1)); /* Within page */
-        /* ugh.  in prepare/commit_write, if from==to==start of block, we 
+        /* ugh.  in prepare/commit_write, if from==to==start of block, we
         ** skip the prepare.  make sure we never send an offset for the start
         ** of a block
         */
@@ -978,6 +981,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 
         size_change = S_ISREG(inode->i_mode) && attr->ia_valid & ATTR_SIZE;
         if (size_change) {
+                dquot_initialize(inode);
+
                 status = ocfs2_rw_lock(inode, 1);
                 if (status < 0) {
                         mlog_errno(status);
@@ -993,10 +998,9 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
         }
 
         if (size_change && attr->ia_size != i_size_read(inode)) {
-                if (attr->ia_size > sb->s_maxbytes) {
-                        status = -EFBIG;
+                status = inode_newsize_ok(inode, attr->ia_size);
+                if (status)
                         goto bail_unlock;
-                }
 
                 if (i_size_read(inode) > attr->ia_size) {
                         if (ocfs2_should_order_data(inode)) {
@@ -1021,7 +1025,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
                 /*
                  * Gather pointers to quota structures so that allocation /
                  * freeing of quota structures happens here and not inside
-                 * vfs_dq_transfer() where we have problems with lock ordering
+                 * dquot_transfer() where we have problems with lock ordering
                  */
                 if (attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid
                     && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
@@ -1054,7 +1058,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
                         mlog_errno(status);
                         goto bail_unlock;
                 }
-                status = vfs_dq_transfer(inode, attr) ? -EDQUOT : 0;
+                status = dquot_transfer(inode, attr);
                 if (status < 0)
                         goto bail_commit;
         } else {
@@ -1772,13 +1776,14 @@ static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
                                          loff_t *ppos,
                                          size_t count,
                                          int appending,
-                                         int *direct_io)
+                                         int *direct_io,
+                                         int *has_refcount)
 {
         int ret = 0, meta_level = 0;
         struct inode *inode = dentry->d_inode;
         loff_t saved_pos, end;
 
-        /* 
+        /*
          * We start with a read level meta lock and only jump to an ex
          * if we need to make modifications here.
          */
@@ -1833,6 +1838,10 @@ static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
                                                                saved_pos,
                                                                count,
                                                                &meta_level);
+                        if (has_refcount)
+                                *has_refcount = 1;
+                        if (direct_io)
+                                *direct_io = 0;
                 }
 
                 if (ret < 0) {
@@ -1899,7 +1908,7 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
                                     loff_t pos)
 {
         int ret, direct_io, appending, rw_level, have_alloc_sem  = 0;
-        int can_do_direct;
+        int can_do_direct, has_refcount = 0;
         ssize_t written = 0;
         size_t ocount;          /* original count */
         size_t count;           /* after file limit checks */
@@ -1942,7 +1951,7 @@ relock:
         can_do_direct = direct_io;
         ret = ocfs2_prepare_inode_for_write(file->f_path.dentry, ppos,
                                             iocb->ki_left, appending,
-                                            &can_do_direct);
+                                            &can_do_direct, &has_refcount);
         if (ret < 0) {
                 mlog_errno(ret);
                 goto out;
@@ -1973,18 +1982,18 @@ relock:
         /* communicate with ocfs2_dio_end_io */
         ocfs2_iocb_set_rw_locked(iocb, rw_level);
 
-        if (direct_io) {
-                ret = generic_segment_checks(iov, &nr_segs, &ocount,
-                                             VERIFY_READ);
-                if (ret)
-                        goto out_dio;
+        ret = generic_segment_checks(iov, &nr_segs, &ocount,
+                                     VERIFY_READ);
+        if (ret)
+                goto out_dio;
 
-                count = ocount;
-                ret = generic_write_checks(file, ppos, &count,
-                                           S_ISBLK(inode->i_mode));
-                if (ret)
-                        goto out_dio;
+        count = ocount;
+        ret = generic_write_checks(file, ppos, &count,
+                                   S_ISBLK(inode->i_mode));
+        if (ret)
+                goto out_dio;
 
+        if (direct_io) {
                 written = generic_file_direct_write(iocb, iov, &nr_segs, *ppos,
                                                     ppos, count, ocount);
                 if (written < 0) {
@@ -1999,21 +2008,26 @@ relock:
                         goto out_dio;
                 }
         } else {
-                written = __generic_file_aio_write(iocb, iov, nr_segs, ppos);
+                current->backing_dev_info = file->f_mapping->backing_dev_info;
+                written = generic_file_buffered_write(iocb, iov, nr_segs, *ppos,
+                                                      ppos, count, 0);
+                current->backing_dev_info = NULL;
         }
 
 out_dio:
         /* buffered aio wouldn't have proper lock coverage today */
         BUG_ON(ret == -EIOCBQUEUED && !(file->f_flags & O_DIRECT));
 
-        if ((file->f_flags & O_SYNC && !direct_io) || IS_SYNC(inode)) {
+        if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) ||
+            ((file->f_flags & O_DIRECT) && has_refcount)) {
                 ret = filemap_fdatawrite_range(file->f_mapping, pos,
                                                pos + count - 1);
                 if (ret < 0)
                         written = ret;
 
-                if (!ret && (old_size != i_size_read(inode) ||
-                    old_clusters != OCFS2_I(inode)->ip_clusters)) {
+                if (!ret && ((old_size != i_size_read(inode)) ||
+                             (old_clusters != OCFS2_I(inode)->ip_clusters) ||
+                             has_refcount)) {
                         ret = jbd2_journal_force_commit(osb->journal->j_journal);
                         if (ret < 0)
                                 written = ret;
@@ -2024,7 +2038,7 @@ out_dio:
                                                       pos + count - 1);
         }
 
-        /* 
+        /*
          * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
          * function pointer which is called when o_direct io completes so that
          * it can unlock our rw lock.  (it's the clustered equivalent of
@@ -2034,7 +2048,7 @@ out_dio:
          * async dio is going to do it in the future or an end_io after an
          * error has already done it.
          */
-        if (ret == -EIOCBQUEUED || !ocfs2_iocb_is_rw_locked(iocb)) {
+        if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
                 rw_level = -1;
                 have_alloc_sem = 0;
         }
@@ -2062,7 +2076,7 @@ static int ocfs2_splice_to_file(struct pipe_inode_info *pipe,
         int ret;
 
         ret = ocfs2_prepare_inode_for_write(out->f_path.dentry, &sd->pos,
-                                            sd->total_len, 0, NULL);
+                                            sd->total_len, 0, NULL, NULL);
         if (ret < 0) {
                 mlog_errno(ret);
                 return ret;
@@ -2189,7 +2203,7 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb,
                 goto bail;
         }
 
-        /* 
+        /*
          * buffered reads protect themselves in ->readpage().  O_DIRECT reads
          * need locks to protect pending reads from racing with truncate.
          */
@@ -2211,10 +2225,10 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb,
          * We're fine letting folks race truncates and extending
          * writes with read across the cluster, just like they can
          * locally. Hence no rw_lock during read.
-         * 
+         *
          * Take and drop the meta data lock to update inode fields
          * like i_size. This allows the checks down below
-         * generic_file_aio_read() a chance of actually working. 
+         * generic_file_aio_read() a chance of actually working.
          */
         ret = ocfs2_inode_lock_atime(inode, filp->f_vfsmnt, &lock_level);
         if (ret < 0) {
@@ -2239,7 +2253,7 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb,
 bail:
         if (have_alloc_sem)
                 up_read(&inode->i_alloc_sem);
-        if (rw_level != -1) 
+        if (rw_level != -1)
                 ocfs2_rw_unlock(inode, rw_level);
         mlog_exit(ret);
 
diff --git a/fs/ocfs2/heartbeat.c b/fs/ocfs2/heartbeat.c
index c6e7213db868..1aa863dd901f 100644
--- a/fs/ocfs2/heartbeat.c
+++ b/fs/ocfs2/heartbeat.c
@@ -26,7 +26,6 @@
 
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 
 #define MLOG_MASK_PREFIX ML_SUPER
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 0297fb8982b8..af189887201c 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -25,7 +25,6 @@
 
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
 #include <linux/quotaops.h>
@@ -475,7 +474,7 @@ static int ocfs2_read_locked_inode(struct inode *inode,
         if (args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY) {
                 status = ocfs2_try_open_lock(inode, 0);
                 if (status) {
-                        make_bad_inode(inode);  
+                        make_bad_inode(inode);
                         return status;
                 }
         }
@@ -559,6 +558,7 @@ static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
                 handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
                 if (IS_ERR(handle)) {
                         status = PTR_ERR(handle);
+                        handle = NULL;
                         mlog_errno(status);
                         goto out;
                 }
@@ -640,11 +640,13 @@ static int ocfs2_remove_inode(struct inode *inode,
                 goto bail_unlock;
         }
 
-        status = ocfs2_orphan_del(osb, handle, orphan_dir_inode, inode,
-                                  orphan_dir_bh);
-        if (status < 0) {
-                mlog_errno(status);
-                goto bail_commit;
+        if (!(OCFS2_I(inode)->ip_flags & OCFS2_INODE_SKIP_ORPHAN_DIR)) {
+                status = ocfs2_orphan_del(osb, handle, orphan_dir_inode, inode,
+                                          orphan_dir_bh);
+                if (status < 0) {
+                        mlog_errno(status);
+                        goto bail_commit;
+                }
         }
 
         /* set the inodes dtime */
@@ -665,7 +667,7 @@ static int ocfs2_remove_inode(struct inode *inode,
         }
 
         ocfs2_remove_from_cache(INODE_CACHE(inode), di_bh);
-        vfs_dq_free_inode(inode);
+        dquot_free_inode(inode);
 
         status = ocfs2_free_dinode(handle, inode_alloc_inode,
                                    inode_alloc_bh, di);
@@ -684,7 +686,7 @@ bail:
         return status;
 }
 
-/* 
+/*
  * Serialize with orphan dir recovery. If the process doing
  * recovery on this orphan dir does an iget() with the dir
  * i_mutex held, we'll deadlock here. Instead we detect this
@@ -723,38 +725,39 @@ static void ocfs2_signal_wipe_completion(struct ocfs2_super *osb,
 static int ocfs2_wipe_inode(struct inode *inode,
                             struct buffer_head *di_bh)
 {
-        int status, orphaned_slot;
+        int status, orphaned_slot = -1;
         struct inode *orphan_dir_inode = NULL;
         struct buffer_head *orphan_dir_bh = NULL;
         struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-        struct ocfs2_dinode *di;
+        struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data;
 
-        di = (struct ocfs2_dinode *) di_bh->b_data;
-        orphaned_slot = le16_to_cpu(di->i_orphaned_slot);
+        if (!(OCFS2_I(inode)->ip_flags & OCFS2_INODE_SKIP_ORPHAN_DIR)) {
+                orphaned_slot = le16_to_cpu(di->i_orphaned_slot);
 
-        status = ocfs2_check_orphan_recovery_state(osb, orphaned_slot);
-        if (status)
-                return status;
+                status = ocfs2_check_orphan_recovery_state(osb, orphaned_slot);
+                if (status)
+                        return status;
 
-        orphan_dir_inode = ocfs2_get_system_file_inode(osb,
-                                                       ORPHAN_DIR_SYSTEM_INODE,
-                                                       orphaned_slot);
-        if (!orphan_dir_inode) {
-                status = -EEXIST;
-                mlog_errno(status);
-                goto bail;
-        }
+                orphan_dir_inode = ocfs2_get_system_file_inode(osb,
+                                                               ORPHAN_DIR_SYSTEM_INODE,
+                                                               orphaned_slot);
+                if (!orphan_dir_inode) {
+                        status = -EEXIST;
+                        mlog_errno(status);
+                        goto bail;
+                }
 
-        /* Lock the orphan dir. The lock will be held for the entire
-         * delete_inode operation. We do this now to avoid races with
-         * recovery completion on other nodes. */
-        mutex_lock(&orphan_dir_inode->i_mutex);
-        status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
-        if (status < 0) {
-                mutex_unlock(&orphan_dir_inode->i_mutex);
+                /* Lock the orphan dir. The lock will be held for the entire
+                 * delete_inode operation. We do this now to avoid races with
+                 * recovery completion on other nodes. */
+                mutex_lock(&orphan_dir_inode->i_mutex);
+                status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
+                if (status < 0) {
+                        mutex_unlock(&orphan_dir_inode->i_mutex);
 
-                mlog_errno(status);
-                goto bail;
+                        mlog_errno(status);
+                        goto bail;
+                }
         }
 
         /* we do this while holding the orphan dir lock because we
@@ -795,6 +798,9 @@ static int ocfs2_wipe_inode(struct inode *inode,
                 mlog_errno(status);
 
 bail_unlock_dir:
+        if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SKIP_ORPHAN_DIR)
+                return status;
+
         ocfs2_inode_unlock(orphan_dir_inode, 1);
         mutex_unlock(&orphan_dir_inode->i_mutex);
         brelse(orphan_dir_bh);
@@ -890,7 +896,23 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
 
         /* Do some basic inode verification... */
         di = (struct ocfs2_dinode *) di_bh->b_data;
-        if (!(di->i_flags & cpu_to_le32(OCFS2_ORPHANED_FL))) {
+        if (!(di->i_flags & cpu_to_le32(OCFS2_ORPHANED_FL)) &&
+            !(oi->ip_flags & OCFS2_INODE_SKIP_ORPHAN_DIR)) {
+                /*
+                 * Inodes in the orphan dir must have ORPHANED_FL.  The only
+                 * inodes that come back out of the orphan dir are reflink
+                 * targets. A reflink target may be moved out of the orphan
+                 * dir between the time we scan the directory and the time we
+                 * process it. This would lead to HAS_REFCOUNT_FL being set but
+                 * ORPHANED_FL not.
+                 */
+                if (di->i_dyn_features & cpu_to_le16(OCFS2_HAS_REFCOUNT_FL)) {
+                        mlog(0, "Reflinked inode %llu is no longer orphaned.  "
+                             "it shouldn't be deleted\n",
+                             (unsigned long long)oi->ip_blkno);
+                        goto bail;
+                }
+
                 /* for lack of a better error? */
                 status = -EEXIST;
                 mlog(ML_ERROR,
@@ -971,6 +993,8 @@ void ocfs2_delete_inode(struct inode *inode)
                 goto bail;
         }
 
+        dquot_initialize(inode);
+
         if (!ocfs2_inode_is_valid_to_delete(inode)) {
                 /* It's probably not necessary to truncate_inode_pages
                  * here but we do it for safety anyway (it will most
@@ -1087,6 +1111,8 @@ void ocfs2_clear_inode(struct inode *inode)
         mlog_bug_on_msg(OCFS2_SB(inode->i_sb) == NULL,
                         "Inode=%lu\n", inode->i_ino);
 
+        dquot_drop(inode);
+
         /* To preven remote deletes we hold open lock before, now it
          * is time to unlock PR and EX open locks. */
         ocfs2_open_unlock(inode);
diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h
index ba4fe07b293c..0b28e1921a39 100644
--- a/fs/ocfs2/inode.h
+++ b/fs/ocfs2/inode.h
@@ -100,6 +100,8 @@ struct ocfs2_inode_info
 #define OCFS2_INODE_MAYBE_ORPHANED      0x00000020
 /* Does someone have the file open O_DIRECT */
 #define OCFS2_INODE_OPEN_DIRECT         0x00000040
+/* Tell the inode wipe code it's not in orphan dir */
+#define OCFS2_INODE_SKIP_ORPHAN_DIR     0x00000080
 
 static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode)
 {
diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c
index 31fbb0619510..7d9d9c132cef 100644
--- a/fs/ocfs2/ioctl.c
+++ b/fs/ocfs2/ioctl.c
@@ -7,6 +7,7 @@
 
 #include <linux/fs.h>
 #include <linux/mount.h>
+#include <linux/compat.h>
 
 #define MLOG_MASK_PREFIX ML_INODE
 #include <cluster/masklog.h>
@@ -181,6 +182,10 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 #ifdef CONFIG_COMPAT
 long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 {
+        bool preserve;
+        struct reflink_arguments args;
+        struct inode *inode = file->f_path.dentry->d_inode;
+
         switch (cmd) {
         case OCFS2_IOC32_GETFLAGS:
                 cmd = OCFS2_IOC_GETFLAGS;
@@ -195,8 +200,15 @@ long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg)
         case OCFS2_IOC_GROUP_EXTEND:
         case OCFS2_IOC_GROUP_ADD:
         case OCFS2_IOC_GROUP_ADD64:
-        case OCFS2_IOC_REFLINK:
                 break;
+        case OCFS2_IOC_REFLINK:
+                if (copy_from_user(&args, (struct reflink_arguments *)arg,
+                                   sizeof(args)))
+                        return -EFAULT;
+                preserve = (args.preserve != 0);
+
+                return ocfs2_reflink_ioctl(inode, compat_ptr(args.old_path),
+                                           compat_ptr(args.new_path), preserve);
         default:
                 return -ENOIOCTLCMD;
         }
diff --git a/fs/ocfs2/ioctl.h b/fs/ocfs2/ioctl.h
index cf9a5ee30fef..0cd5323bd3f0 100644
--- a/fs/ocfs2/ioctl.h
+++ b/fs/ocfs2/ioctl.h
@@ -7,10 +7,10 @@
  *
  */
 
-#ifndef OCFS2_IOCTL_H
-#define OCFS2_IOCTL_H
+#ifndef OCFS2_IOCTL_PROTO_H
+#define OCFS2_IOCTL_PROTO_H
 
 long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
 long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg);
 
-#endif /* OCFS2_IOCTL_H */
+#endif /* OCFS2_IOCTL_PROTO_H */
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 54c16b66327e..9336c60e3a36 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -659,7 +659,7 @@ static int __ocfs2_journal_access(handle_t *handle,
 
         default:
                 status = -EINVAL;
-                mlog(ML_ERROR, "Uknown access type!\n");
+                mlog(ML_ERROR, "Unknown access type!\n");
         }
         if (!status && ocfs2_meta_ecc(osb) && triggers)
                 jbd2_journal_set_triggers(bh, &triggers->ot_triggers);
@@ -2034,7 +2034,7 @@ static int ocfs2_queue_orphans(struct ocfs2_super *osb,
                 status = -ENOENT;
                 mlog_errno(status);
                 return status;
-        }       
+        }
 
         mutex_lock(&orphan_dir_inode->i_mutex);
         status = ocfs2_inode_lock(orphan_dir_inode, NULL, 0);
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index ac10f83edb95..c983715d8d8c 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -476,7 +476,7 @@ out_mutex:
 
 out:
         if (!status)
-                ocfs2_init_inode_steal_slot(osb);
+                ocfs2_init_steal_slots(osb);
         mlog_exit(status);
         return status;
 }
@@ -872,8 +872,10 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
                              (unsigned long long)la_start_blk,
                              (unsigned long long)blkno);
 
-                        status = ocfs2_free_clusters(handle, main_bm_inode,
-                                                     main_bm_bh, blkno, count);
+                        status = ocfs2_release_clusters(handle,
+                                                        main_bm_inode,
+                                                        main_bm_bh, blkno,
+                                                        count);
                         if (status < 0) {
                                 mlog_errno(status);
                                 goto bail;
@@ -984,8 +986,7 @@ static int ocfs2_local_alloc_reserve_for_window(struct ocfs2_super *osb,
         }
 
 retry_enospc:
-        (*ac)->ac_bits_wanted = osb->local_alloc_bits;
-
+        (*ac)->ac_bits_wanted = osb->local_alloc_default_bits;
         status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac);
         if (status == -ENOSPC) {
                 if (ocfs2_recalc_la_window(osb, OCFS2_LA_EVENT_ENOSPC) ==
@@ -1061,6 +1062,7 @@ retry_enospc:
                     OCFS2_LA_DISABLED)
                         goto bail;
 
+                ac->ac_bits_wanted = osb->local_alloc_default_bits;
                 status = ocfs2_claim_clusters(osb, handle, ac,
                                               osb->local_alloc_bits,
                                               &cluster_off,
diff --git a/fs/ocfs2/locks.c b/fs/ocfs2/locks.c
index 544ac6245175..b5cb3ede9408 100644
--- a/fs/ocfs2/locks.c
+++ b/fs/ocfs2/locks.c
@@ -133,7 +133,7 @@ int ocfs2_lock(struct file *file, int cmd, struct file_lock *fl)
 
         if (!(fl->fl_flags & FL_POSIX))
                 return -ENOLCK;
-        if (__mandatory_lock(inode))
+        if (__mandatory_lock(inode) && fl->fl_type != F_UNLCK)
                 return -ENOLCK;
 
         return ocfs2_plock(osb->cconn, OCFS2_I(inode)->ip_blkno, file, cmd, fl);
diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c
index 39737613424a..7898bd3a99f5 100644
--- a/fs/ocfs2/mmap.c
+++ b/fs/ocfs2/mmap.c
@@ -25,7 +25,6 @@
 
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
 #include <linux/uio.h>
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index f010b22b1c44..4cbb18f26c5f 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -84,7 +84,7 @@ static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
 static int ocfs2_orphan_add(struct ocfs2_super *osb,
                             handle_t *handle,
                             struct inode *inode,
-                            struct ocfs2_dinode *fe,
+                            struct buffer_head *fe_bh,
                             char *name,
                             struct ocfs2_dir_lookup_result *lookup,
                             struct inode *orphan_dir_inode);
@@ -212,7 +212,7 @@ static struct inode *ocfs2_get_init_inode(struct inode *dir, int mode)
         } else
                 inode->i_gid = current_fsgid();
         inode->i_mode = mode;
-        vfs_dq_init(inode);
+        dquot_initialize(inode);
         return inode;
 }
 
@@ -244,6 +244,8 @@ static int ocfs2_mknod(struct inode *dir,
                    (unsigned long)dev, dentry->d_name.len,
                    dentry->d_name.name);
 
+        dquot_initialize(dir);
+
         /* get our super block */
         osb = OCFS2_SB(dir->i_sb);
 
@@ -348,13 +350,9 @@ static int ocfs2_mknod(struct inode *dir,
                 goto leave;
         }
 
-        /* We don't use standard VFS wrapper because we don't want vfs_dq_init
-         * to be called. */
-        if (sb_any_quota_active(osb->sb) &&
-            osb->sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) {
-                status = -EDQUOT;
+        status = dquot_alloc_inode(inode);
+        if (status)
                 goto leave;
-        }
         did_quota_inode = 1;
 
         mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry,
@@ -410,28 +408,33 @@ static int ocfs2_mknod(struct inode *dir,
                 }
         }
 
-        status = ocfs2_add_entry(handle, dentry, inode,
-                                 OCFS2_I(inode)->ip_blkno, parent_fe_bh,
-                                 &lookup);
-        if (status < 0) {
+        /*
+         * Do this before adding the entry to the directory. We add
+         * also set d_op after success so that ->d_iput() will cleanup
+         * the dentry lock even if ocfs2_add_entry() fails below.
+         */
+        status = ocfs2_dentry_attach_lock(dentry, inode,
+                                          OCFS2_I(dir)->ip_blkno);
+        if (status) {
                 mlog_errno(status);
                 goto leave;
         }
+        dentry->d_op = &ocfs2_dentry_ops;
 
-        status = ocfs2_dentry_attach_lock(dentry, inode,
-                                          OCFS2_I(dir)->ip_blkno);
-        if (status) {
+        status = ocfs2_add_entry(handle, dentry, inode,
+                                 OCFS2_I(inode)->ip_blkno, parent_fe_bh,
+                                 &lookup);
+        if (status < 0) {
                 mlog_errno(status);
                 goto leave;
         }
 
         insert_inode_hash(inode);
-        dentry->d_op = &ocfs2_dentry_ops;
         d_instantiate(dentry, inode);
         status = 0;
 leave:
         if (status < 0 && did_quota_inode)
-                vfs_dq_free_inode(inode);
+                dquot_free_inode(inode);
         if (handle)
                 ocfs2_commit_trans(osb, handle);
 
@@ -447,11 +450,6 @@ leave:
 
         ocfs2_free_dir_lookup_result(&lookup);
 
-        if ((status < 0) && inode) {
-                clear_nlink(inode);
-                iput(inode);
-        }
-
         if (inode_ac)
                 ocfs2_free_alloc_context(inode_ac);
 
@@ -461,6 +459,17 @@ leave:
         if (meta_ac)
                 ocfs2_free_alloc_context(meta_ac);
 
+        /*
+         * We should call iput after the i_mutex of the bitmap been
+         * unlocked in ocfs2_free_alloc_context, or the
+         * ocfs2_delete_inode will mutex_lock again.
+         */
+        if ((status < 0) && inode) {
+                OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SKIP_ORPHAN_DIR;
+                clear_nlink(inode);
+                iput(inode);
+        }
+
         mlog_exit(status);
 
         return status;
@@ -636,6 +645,8 @@ static int ocfs2_link(struct dentry *old_dentry,
         if (S_ISDIR(inode->i_mode))
                 return -EPERM;
 
+        dquot_initialize(dir);
+
         err = ocfs2_inode_lock_nested(dir, &parent_fe_bh, 1, OI_LS_PARENT);
         if (err < 0) {
                 if (err != -ENOENT)
@@ -791,6 +802,8 @@ static int ocfs2_unlink(struct inode *dir,
         mlog_entry("(0x%p, 0x%p, '%.*s')\n", dir, dentry,
                    dentry->d_name.len, dentry->d_name.name);
 
+        dquot_initialize(dir);
+
         BUG_ON(dentry->d_parent->d_inode != dir);
 
         mlog(0, "ino = %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
@@ -877,7 +890,7 @@ static int ocfs2_unlink(struct inode *dir,
         fe = (struct ocfs2_dinode *) fe_bh->b_data;
 
         if (inode_is_unlinkable(inode)) {
-                status = ocfs2_orphan_add(osb, handle, inode, fe, orphan_name,
+                status = ocfs2_orphan_add(osb, handle, inode, fe_bh, orphan_name,
                                           &orphan_insert, orphan_dir);
                 if (status < 0) {
                         mlog_errno(status);
@@ -1051,6 +1064,9 @@ static int ocfs2_rename(struct inode *old_dir,
                    old_dentry->d_name.len, old_dentry->d_name.name,
                    new_dentry->d_name.len, new_dentry->d_name.name);
 
+        dquot_initialize(old_dir);
+        dquot_initialize(new_dir);
+
         osb = OCFS2_SB(old_dir->i_sb);
 
         if (new_inode) {
@@ -1295,7 +1311,7 @@ static int ocfs2_rename(struct inode *old_dir,
                 if (S_ISDIR(new_inode->i_mode) ||
                     (ocfs2_read_links_count(newfe) == 1)) {
                         status = ocfs2_orphan_add(osb, handle, new_inode,
-                                                  newfe, orphan_name,
+                                                  newfe_bh, orphan_name,
                                                   &orphan_insert, orphan_dir);
                         if (status < 0) {
                                 mlog_errno(status);
@@ -1599,6 +1615,8 @@ static int ocfs2_symlink(struct inode *dir,
         mlog_entry("(0x%p, 0x%p, symname='%s' actual='%.*s')\n", dir,
                    dentry, symname, dentry->d_name.len, dentry->d_name.name);
 
+        dquot_initialize(dir);
+
         sb = dir->i_sb;
         osb = OCFS2_SB(sb);
 
@@ -1688,13 +1706,9 @@ static int ocfs2_symlink(struct inode *dir,
                 goto bail;
         }
 
-        /* We don't use standard VFS wrapper because we don't want vfs_dq_init
-         * to be called. */
-        if (sb_any_quota_active(osb->sb) &&
-            osb->sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) {
-                status = -EDQUOT;
+        status = dquot_alloc_inode(inode);
+        if (status)
                 goto bail;
-        }
         did_quota_inode = 1;
 
         mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry,
@@ -1716,11 +1730,10 @@ static int ocfs2_symlink(struct inode *dir,
                 u32 offset = 0;
 
                 inode->i_op = &ocfs2_symlink_inode_operations;
-                if (vfs_dq_alloc_space_nodirty(inode,
-                    ocfs2_clusters_to_bytes(osb->sb, 1))) {
-                        status = -EDQUOT;
+                status = dquot_alloc_space_nodirty(inode,
+                    ocfs2_clusters_to_bytes(osb->sb, 1));
+                if (status)
                         goto bail;
-                }
                 did_quota = 1;
                 status = ocfs2_add_inode_data(osb, inode, &offset, 1, 0,
                                               new_fe_bh,
@@ -1769,29 +1782,34 @@ static int ocfs2_symlink(struct inode *dir,
                 }
         }
 
-        status = ocfs2_add_entry(handle, dentry, inode,
-                                 le64_to_cpu(fe->i_blkno), parent_fe_bh,
-                                 &lookup);
-        if (status < 0) {
+        /*
+         * Do this before adding the entry to the directory. We add
+         * also set d_op after success so that ->d_iput() will cleanup
+         * the dentry lock even if ocfs2_add_entry() fails below.
+         */
+        status = ocfs2_dentry_attach_lock(dentry, inode, OCFS2_I(dir)->ip_blkno);
+        if (status) {
                 mlog_errno(status);
                 goto bail;
         }
+        dentry->d_op = &ocfs2_dentry_ops;
 
-        status = ocfs2_dentry_attach_lock(dentry, inode, OCFS2_I(dir)->ip_blkno);
-        if (status) {
+        status = ocfs2_add_entry(handle, dentry, inode,
+                                 le64_to_cpu(fe->i_blkno), parent_fe_bh,
+                                 &lookup);
+        if (status < 0) {
                 mlog_errno(status);
                 goto bail;
         }
 
         insert_inode_hash(inode);
-        dentry->d_op = &ocfs2_dentry_ops;
         d_instantiate(dentry, inode);
 bail:
         if (status < 0 && did_quota)
-                vfs_dq_free_space_nodirty(inode,
+                dquot_free_space_nodirty(inode,
                                         ocfs2_clusters_to_bytes(osb->sb, 1));
         if (status < 0 && did_quota_inode)
-                vfs_dq_free_inode(inode);
+                dquot_free_inode(inode);
         if (handle)
                 ocfs2_commit_trans(osb, handle);
 
@@ -1809,6 +1827,7 @@ bail:
         if (xattr_ac)
                 ocfs2_free_alloc_context(xattr_ac);
         if ((status < 0) && inode) {
+                OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SKIP_ORPHAN_DIR;
                 clear_nlink(inode);
                 iput(inode);
         }
@@ -1909,7 +1928,7 @@ leave:
 static int ocfs2_orphan_add(struct ocfs2_super *osb,
                             handle_t *handle,
                             struct inode *inode,
-                            struct ocfs2_dinode *fe,
+                            struct buffer_head *fe_bh,
                             char *name,
                             struct ocfs2_dir_lookup_result *lookup,
                             struct inode *orphan_dir_inode)
@@ -1917,6 +1936,7 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
         struct buffer_head *orphan_dir_bh = NULL;
         int status = 0;
         struct ocfs2_dinode *orphan_fe;
+        struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
 
         mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);
 
@@ -1957,13 +1977,31 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
                 goto leave;
         }
 
+        /*
+         * We're going to journal the change of i_flags and i_orphaned_slot.
+         * It's safe anyway, though some callers may duplicate the journaling.
+         * Journaling within the func just make the logic look more
+         * straightforward.
+         */
+        status = ocfs2_journal_access_di(handle,
+                                         INODE_CACHE(inode),
+                                         fe_bh,
+                                         OCFS2_JOURNAL_ACCESS_WRITE);
+        if (status < 0) {
+                mlog_errno(status);
+                goto leave;
+        }
+
         le32_add_cpu(&fe->i_flags, OCFS2_ORPHANED_FL);
+        OCFS2_I(inode)->ip_flags &= ~OCFS2_INODE_SKIP_ORPHAN_DIR;
 
         /* Record which orphan dir our inode now resides
          * in. delete_inode will use this to determine which orphan
          * dir to lock. */
         fe->i_orphaned_slot = cpu_to_le16(osb->slot_num);
 
+        ocfs2_journal_dirty(handle, fe_bh);
+
         mlog(0, "Inode %llu orphaned in slot %d\n",
              (unsigned long long)OCFS2_I(inode)->ip_blkno, osb->slot_num);
 
@@ -2099,15 +2137,12 @@ int ocfs2_create_inode_in_orphan(struct inode *dir,
                 goto leave;
         }
 
-        /* We don't use standard VFS wrapper because we don't want vfs_dq_init
-         * to be called. */
-        if (sb_any_quota_active(osb->sb) &&
-            osb->sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) {
-                status = -EDQUOT;
+        status = dquot_alloc_inode(inode);
+        if (status)
                 goto leave;
-        }
         did_quota_inode = 1;
 
+        inode->i_nlink = 0;
         /* do the real work now. */
         status = ocfs2_mknod_locked(osb, dir, inode,
                                     0, &new_di_bh, parent_di_bh, handle,
@@ -2124,7 +2159,7 @@ int ocfs2_create_inode_in_orphan(struct inode *dir,
         }
 
         di = (struct ocfs2_dinode *)new_di_bh->b_data;
-        status = ocfs2_orphan_add(osb, handle, inode, di, orphan_name,
+        status = ocfs2_orphan_add(osb, handle, inode, new_di_bh, orphan_name,
                                   &orphan_insert, orphan_dir);
         if (status < 0) {
                 mlog_errno(status);
@@ -2136,9 +2171,10 @@ int ocfs2_create_inode_in_orphan(struct inode *dir,
         if (status < 0)
                 mlog_errno(status);
 
+        insert_inode_hash(inode);
 leave:
         if (status < 0 && did_quota_inode)
-                vfs_dq_free_inode(inode);
+                dquot_free_inode(inode);
         if (handle)
                 ocfs2_commit_trans(osb, handle);
 
@@ -2267,6 +2303,8 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
         di = (struct ocfs2_dinode *)di_bh->b_data;
         le32_add_cpu(&di->i_flags, -OCFS2_ORPHANED_FL);
         di->i_orphaned_slot = 0;
+        inode->i_nlink = 1;
+        ocfs2_set_links_count(di, inode->i_nlink);
         ocfs2_journal_dirty(handle, di_bh);
 
         status = ocfs2_add_entry(handle, dentry, inode,
@@ -2284,7 +2322,6 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
                 goto out_commit;
         }
 
-        insert_inode_hash(inode);
         dentry->d_op = &ocfs2_dentry_ops;
         d_instantiate(dentry, inode);
         status = 0;
@@ -2326,4 +2363,5 @@ const struct inode_operations ocfs2_dir_iops = {
         .getxattr       = generic_getxattr,
         .listxattr      = ocfs2_listxattr,
         .removexattr    = generic_removexattr,
+        .fiemap         = ocfs2_fiemap,
 };
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index d963d8638709..adf5e2ebc2c4 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -42,6 +42,7 @@
 
 #include "ocfs2_fs.h"
 #include "ocfs2_lockid.h"
+#include "ocfs2_ioctl.h"
 
 /* For struct ocfs2_blockcheck_stats */
 #include "blockcheck.h"
@@ -136,6 +137,10 @@ enum ocfs2_unlock_action {
 #define OCFS2_LOCK_PENDING       (0x00000400) /* This lockres is pending a
                                                  call to dlm_lock.  Only
                                                  exists with BUSY set. */
+#define OCFS2_LOCK_UPCONVERT_FINISHING (0x00000800) /* blocks the dc thread
+                                                     * from downconverting
+                                                     * before the upconvert
+                                                     * has completed */
 
 struct ocfs2_lock_res_ops;
 
@@ -155,7 +160,7 @@ struct ocfs2_lock_res {
         int                      l_level;
         unsigned int             l_ro_holders;
         unsigned int             l_ex_holders;
-        union ocfs2_dlm_lksb     l_lksb;
+        struct ocfs2_dlm_lksb    l_lksb;
 
         /* used from AST/BAST funcs. */
         enum ocfs2_ast_action    l_action;
@@ -245,9 +250,11 @@ enum ocfs2_mount_options
         OCFS2_MOUNT_LOCALFLOCKS = 1 << 5, /* No cluster aware user file locks */
         OCFS2_MOUNT_NOUSERXATTR = 1 << 6, /* No user xattr */
         OCFS2_MOUNT_INODE64 = 1 << 7,   /* Allow inode numbers > 2^32 */
-        OCFS2_MOUNT_POSIX_ACL = 1 << 8, /* POSIX access control lists */
-        OCFS2_MOUNT_USRQUOTA = 1 << 9, /* We support user quotas */
-        OCFS2_MOUNT_GRPQUOTA = 1 << 10, /* We support group quotas */
+        OCFS2_MOUNT_POSIX_ACL = 1 << 8, /* Force POSIX access control lists */
+        OCFS2_MOUNT_NO_POSIX_ACL = 1 << 9,      /* Disable POSIX access
+                                                   control lists */
+        OCFS2_MOUNT_USRQUOTA = 1 << 10, /* We support user quotas */
+        OCFS2_MOUNT_GRPQUOTA = 1 << 11, /* We support group quotas */
 };
 
 #define OCFS2_OSB_SOFT_RO                       0x0001
@@ -299,7 +306,9 @@ struct ocfs2_super
         u32 s_next_generation;
         unsigned long osb_flags;
         s16 s_inode_steal_slot;
+        s16 s_meta_steal_slot;
         atomic_t s_num_inodes_stolen;
+        atomic_t s_num_meta_stolen;
 
         unsigned long s_mount_opt;
         unsigned int s_atime_quantum;
@@ -754,35 +763,18 @@ static inline unsigned int ocfs2_megabytes_to_clusters(struct super_block *sb,
         return megs << (20 - OCFS2_SB(sb)->s_clustersize_bits);
 }
 
-static inline void ocfs2_init_inode_steal_slot(struct ocfs2_super *osb)
+static inline void _ocfs2_set_bit(unsigned int bit, unsigned long *bitmap)
 {
-        spin_lock(&osb->osb_lock);
-        osb->s_inode_steal_slot = OCFS2_INVALID_SLOT;
-        spin_unlock(&osb->osb_lock);
-        atomic_set(&osb->s_num_inodes_stolen, 0);
+        ext2_set_bit(bit, bitmap);
 }
+#define ocfs2_set_bit(bit, addr) _ocfs2_set_bit((bit), (unsigned long *)(addr))
 
-static inline void ocfs2_set_inode_steal_slot(struct ocfs2_super *osb,
-                                              s16 slot)
+static inline void _ocfs2_clear_bit(unsigned int bit, unsigned long *bitmap)
 {
-        spin_lock(&osb->osb_lock);
-        osb->s_inode_steal_slot = slot;
-        spin_unlock(&osb->osb_lock);
-}
-
-static inline s16 ocfs2_get_inode_steal_slot(struct ocfs2_super *osb)
-{
-        s16 slot;
-
-        spin_lock(&osb->osb_lock);
-        slot = osb->s_inode_steal_slot;
-        spin_unlock(&osb->osb_lock);
-
-        return slot;
+        ext2_clear_bit(bit, bitmap);
 }
+#define ocfs2_clear_bit(bit, addr) _ocfs2_clear_bit((bit), (unsigned long *)(addr))
 
-#define ocfs2_set_bit ext2_set_bit
-#define ocfs2_clear_bit ext2_clear_bit
 #define ocfs2_test_bit ext2_test_bit
 #define ocfs2_find_next_zero_bit ext2_find_next_zero_bit
 #define ocfs2_find_next_bit ext2_find_next_bit
diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h
index e9431e4a5e7c..bb37218a7978 100644
--- a/fs/ocfs2/ocfs2_fs.h
+++ b/fs/ocfs2/ocfs2_fs.h
@@ -254,63 +254,6 @@
                                                  * refcount tree */
 
 /*
- * ioctl commands
- */
-#define OCFS2_IOC_GETFLAGS      _IOR('f', 1, long)
-#define OCFS2_IOC_SETFLAGS      _IOW('f', 2, long)
-#define OCFS2_IOC32_GETFLAGS    _IOR('f', 1, int)
-#define OCFS2_IOC32_SETFLAGS    _IOW('f', 2, int)
-
-/*
- * Space reservation / allocation / free ioctls and argument structure
- * are designed to be compatible with XFS.
- *
- * ALLOCSP* and FREESP* are not and will never be supported, but are
- * included here for completeness.
- */
-struct ocfs2_space_resv {
-        __s16           l_type;
-        __s16           l_whence;
-        __s64           l_start;
-        __s64           l_len;          /* len == 0 means until end of file */
-        __s32           l_sysid;
-        __u32           l_pid;
-        __s32           l_pad[4];       /* reserve area                     */
-};
-
-#define OCFS2_IOC_ALLOCSP               _IOW ('X', 10, struct ocfs2_space_resv)
-#define OCFS2_IOC_FREESP                _IOW ('X', 11, struct ocfs2_space_resv)
-#define OCFS2_IOC_RESVSP                _IOW ('X', 40, struct ocfs2_space_resv)
-#define OCFS2_IOC_UNRESVSP      _IOW ('X', 41, struct ocfs2_space_resv)
-#define OCFS2_IOC_ALLOCSP64     _IOW ('X', 36, struct ocfs2_space_resv)
-#define OCFS2_IOC_FREESP64      _IOW ('X', 37, struct ocfs2_space_resv)
-#define OCFS2_IOC_RESVSP64      _IOW ('X', 42, struct ocfs2_space_resv)
-#define OCFS2_IOC_UNRESVSP64    _IOW ('X', 43, struct ocfs2_space_resv)
-
-/* Used to pass group descriptor data when online resize is done */
-struct ocfs2_new_group_input {
-        __u64 group;            /* Group descriptor's blkno. */
-        __u32 clusters;         /* Total number of clusters in this group */
-        __u32 frees;            /* Total free clusters in this group */
-        __u16 chain;            /* Chain for this group */
-        __u16 reserved1;
-        __u32 reserved2;
-};
-
-#define OCFS2_IOC_GROUP_EXTEND  _IOW('o', 1, int)
-#define OCFS2_IOC_GROUP_ADD     _IOW('o', 2,struct ocfs2_new_group_input)
-#define OCFS2_IOC_GROUP_ADD64   _IOW('o', 3,struct ocfs2_new_group_input)
-
-/* Used to pass 2 file names to reflink. */
-struct reflink_arguments {
-        __u64 old_path;
-        __u64 new_path;
-        __u64 preserve;
-};
-#define OCFS2_IOC_REFLINK       _IOW('o', 4, struct reflink_arguments)
-
-
-/*
  * Journal Flags (ocfs2_dinode.id1.journal1.i_flags)
  */
 #define OCFS2_JOURNAL_DIRTY_FL  (0x00000001)    /* Journal needs recovery */
@@ -1202,7 +1145,7 @@ struct ocfs2_local_disk_dqinfo {
 /* Header of one chunk of a quota file */
 struct ocfs2_local_disk_chunk {
         __le32 dqc_free;        /* Number of free entries in the bitmap */
-        u8 dqc_bitmap[0];       /* Bitmap of entries in the corresponding
+        __u8 dqc_bitmap[0];     /* Bitmap of entries in the corresponding
                                  * chunk of quota file */
 };
 
@@ -1417,9 +1360,16 @@ static inline int ocfs2_fast_symlink_chars(int blocksize)
         return blocksize - offsetof(struct ocfs2_dinode, id2.i_symlink);
 }
 
-static inline int ocfs2_max_inline_data(int blocksize)
+static inline int ocfs2_max_inline_data_with_xattr(int blocksize,
+                                                   struct ocfs2_dinode *di)
 {
-        return blocksize - offsetof(struct ocfs2_dinode, id2.i_data.id_data);
+        if (di && (di->i_dyn_features & OCFS2_INLINE_XATTR_FL))
+                return blocksize -
+                        offsetof(struct ocfs2_dinode, id2.i_data.id_data) -
+                        di->i_xattr_inline_size;
+        else
+                return blocksize -
+                        offsetof(struct ocfs2_dinode, id2.i_data.id_data);
 }
 
 static inline int ocfs2_extent_recs_per_inode(int blocksize)
diff --git a/fs/ocfs2/ocfs2_ioctl.h b/fs/ocfs2/ocfs2_ioctl.h
new file mode 100644
index 000000000000..2d3420af1a83
--- /dev/null
+++ b/fs/ocfs2/ocfs2_ioctl.h
@@ -0,0 +1,79 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * ocfs2_ioctl.h
+ *
+ * Defines OCFS2 ioctls.
+ *
+ * Copyright (C) 2010 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#ifndef OCFS2_IOCTL_H
+#define OCFS2_IOCTL_H
+
+/*
+ * ioctl commands
+ */
+#define OCFS2_IOC_GETFLAGS      _IOR('f', 1, long)
+#define OCFS2_IOC_SETFLAGS      _IOW('f', 2, long)
+#define OCFS2_IOC32_GETFLAGS    _IOR('f', 1, int)
+#define OCFS2_IOC32_SETFLAGS    _IOW('f', 2, int)
+
+/*
+ * Space reservation / allocation / free ioctls and argument structure
+ * are designed to be compatible with XFS.
+ *
+ * ALLOCSP* and FREESP* are not and will never be supported, but are
+ * included here for completeness.
+ */
+struct ocfs2_space_resv {
+        __s16           l_type;
+        __s16           l_whence;
+        __s64           l_start;
+        __s64           l_len;          /* len == 0 means until end of file */
+        __s32           l_sysid;
+        __u32           l_pid;
+        __s32           l_pad[4];       /* reserve area                     */
+};
+
+#define OCFS2_IOC_ALLOCSP               _IOW ('X', 10, struct ocfs2_space_resv)
+#define OCFS2_IOC_FREESP                _IOW ('X', 11, struct ocfs2_space_resv)
+#define OCFS2_IOC_RESVSP                _IOW ('X', 40, struct ocfs2_space_resv)
+#define OCFS2_IOC_UNRESVSP      _IOW ('X', 41, struct ocfs2_space_resv)
+#define OCFS2_IOC_ALLOCSP64     _IOW ('X', 36, struct ocfs2_space_resv)
+#define OCFS2_IOC_FREESP64      _IOW ('X', 37, struct ocfs2_space_resv)
+#define OCFS2_IOC_RESVSP64      _IOW ('X', 42, struct ocfs2_space_resv)
+#define OCFS2_IOC_UNRESVSP64    _IOW ('X', 43, struct ocfs2_space_resv)
+
+/* Used to pass group descriptor data when online resize is done */
+struct ocfs2_new_group_input {
+        __u64 group;            /* Group descriptor's blkno. */
+        __u32 clusters;         /* Total number of clusters in this group */
+        __u32 frees;            /* Total free clusters in this group */
+        __u16 chain;            /* Chain for this group */
+        __u16 reserved1;
+        __u32 reserved2;
+};
+
+#define OCFS2_IOC_GROUP_EXTEND  _IOW('o', 1, int)
+#define OCFS2_IOC_GROUP_ADD     _IOW('o', 2,struct ocfs2_new_group_input)
+#define OCFS2_IOC_GROUP_ADD64   _IOW('o', 3,struct ocfs2_new_group_input)
+
+/* Used to pass 2 file names to reflink. */
+struct reflink_arguments {
+        __u64 old_path;
+        __u64 new_path;
+        __u64 preserve;
+};
+#define OCFS2_IOC_REFLINK       _IOW('o', 4, struct reflink_arguments)
+
+#endif /* OCFS2_IOCTL_H */
diff --git a/fs/ocfs2/ocfs2_lockingver.h b/fs/ocfs2/ocfs2_lockingver.h
index 82d5eeac0fff..2e45c8d2ea7e 100644
--- a/fs/ocfs2/ocfs2_lockingver.h
+++ b/fs/ocfs2/ocfs2_lockingver.h
@@ -23,6 +23,8 @@
 /*
  * The protocol version for ocfs2 cluster locking.  See dlmglue.c for
  * more details.
+ *
+ * 1.0 - Initial locking version from ocfs2 1.4.
  */
 #define OCFS2_LOCKING_PROTOCOL_MAJOR 1
 #define OCFS2_LOCKING_PROTOCOL_MINOR 0
diff --git a/fs/ocfs2/quota.h b/fs/ocfs2/quota.h
index e5df9d170b0c..123bc520a2c0 100644
--- a/fs/ocfs2/quota.h
+++ b/fs/ocfs2/quota.h
@@ -17,10 +17,6 @@
 
 #include "ocfs2.h"
 
-/* Common stuff */
-/* id number of quota format */
-#define QFMT_OCFS2 3
-
 /*
  * In-memory structures
  */
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index b437dc0c4cad..ab42a74c7539 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -3,6 +3,7 @@
  */
 #include <linux/spinlock.h>
 #include <linux/fs.h>
+#include <linux/slab.h>
 #include <linux/quota.h>
 #include <linux/quotaops.h>
 #include <linux/dqblk_qtree.h>
@@ -851,13 +852,6 @@ static void ocfs2_destroy_dquot(struct dquot *dquot)
 }
 
 const struct dquot_operations ocfs2_quota_operations = {
-        .initialize     = dquot_initialize,
-        .drop           = dquot_drop,
-        .alloc_space    = dquot_alloc_space,
-        .alloc_inode    = dquot_alloc_inode,
-        .free_space     = dquot_free_space,
-        .free_inode     = dquot_free_inode,
-        .transfer       = dquot_transfer,
         .write_dquot    = ocfs2_write_dquot,
         .acquire_dquot  = ocfs2_acquire_dquot,
         .release_dquot  = ocfs2_release_dquot,
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index 1a2c50a759fa..9ad49305f450 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -3,6 +3,7 @@
  */
 
 #include <linux/fs.h>
+#include <linux/slab.h>
 #include <linux/quota.h>
 #include <linux/quotaops.h>
 #include <linux/module.h>
@@ -457,7 +458,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
                         break;
                 }
                 dchunk = (struct ocfs2_local_disk_chunk *)hbh->b_data;
-                for_each_bit(bit, rchunk->rc_bitmap, ol_chunk_entries(sb)) {
+                for_each_set_bit(bit, rchunk->rc_bitmap, ol_chunk_entries(sb)) {
                         qbh = NULL;
                         status = ocfs2_read_quota_block(lqinode,
                                                 ol_dqblk_block(sb, chunk, bit),
@@ -1325,7 +1326,7 @@ out:
         return status;
 }
 
-static struct quota_format_ops ocfs2_format_ops = {
+static const struct quota_format_ops ocfs2_format_ops = {
         .check_quota_file       = ocfs2_local_check_quota_file,
         .read_file_info         = ocfs2_local_read_info,
         .write_file_info        = ocfs2_global_write_info,
diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c
index 3a0df7a1b810..5cbcd0f008fc 100644
--- a/fs/ocfs2/refcounttree.c
+++ b/fs/ocfs2/refcounttree.c
@@ -37,7 +37,6 @@
 
 #include <linux/bio.h>
 #include <linux/blkdev.h>
-#include <linux/gfp.h>
 #include <linux/slab.h>
 #include <linux/writeback.h>
 #include <linux/pagevec.h>
@@ -276,7 +275,7 @@ static void ocfs2_erase_refcount_tree_from_list(struct ocfs2_super *osb,
         spin_unlock(&osb->osb_lock);
 }
 
-void ocfs2_kref_remove_refcount_tree(struct kref *kref)
+static void ocfs2_kref_remove_refcount_tree(struct kref *kref)
 {
         struct ocfs2_refcount_tree *tree =
                 container_of(kref, struct ocfs2_refcount_tree, rf_getcnt);
@@ -524,23 +523,6 @@ out:
         return ret;
 }
 
-int ocfs2_lock_refcount_tree_by_inode(struct inode *inode, int rw,
-                                      struct ocfs2_refcount_tree **ret_tree,
-                                      struct buffer_head **ref_bh)
-{
-        int ret;
-        u64 ref_blkno;
-
-        ret = ocfs2_get_refcount_block(inode, &ref_blkno);
-        if (ret) {
-                mlog_errno(ret);
-                return ret;
-        }
-
-        return ocfs2_lock_refcount_tree(OCFS2_SB(inode->i_sb), ref_blkno,
-                                        rw, ret_tree, ref_bh);
-}
-
 void ocfs2_unlock_refcount_tree(struct ocfs2_super *osb,
                                 struct ocfs2_refcount_tree *tree, int rw)
 {
@@ -643,7 +625,7 @@ static int ocfs2_create_refcount_tree(struct inode *inode,
         rb = (struct ocfs2_refcount_block *)new_bh->b_data;
         memset(rb, 0, inode->i_sb->s_blocksize);
         strcpy((void *)rb, OCFS2_REFCOUNT_BLOCK_SIGNATURE);
-        rb->rf_suballoc_slot = cpu_to_le16(osb->slot_num);
+        rb->rf_suballoc_slot = cpu_to_le16(meta_ac->ac_alloc_slot);
         rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
         rb->rf_fs_generation = cpu_to_le32(osb->fs_generation);
         rb->rf_blkno = cpu_to_le64(first_blkno);
@@ -969,6 +951,103 @@ out:
 }
 
 /*
+ * Find the end range for a leaf refcount block indicated by
+ * el->l_recs[index].e_blkno.
+ */
+static int ocfs2_get_refcount_cpos_end(struct ocfs2_caching_info *ci,
+                                       struct buffer_head *ref_root_bh,
+                                       struct ocfs2_extent_block *eb,
+                                       struct ocfs2_extent_list *el,
+                                       int index,  u32 *cpos_end)
+{
+        int ret, i, subtree_root;
+        u32 cpos;
+        u64 blkno;
+        struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+        struct ocfs2_path *left_path = NULL, *right_path = NULL;
+        struct ocfs2_extent_tree et;
+        struct ocfs2_extent_list *tmp_el;
+
+        if (index < le16_to_cpu(el->l_next_free_rec) - 1) {
+                /*
+                 * We have a extent rec after index, so just use the e_cpos
+                 * of the next extent rec.
+                 */
+                *cpos_end = le32_to_cpu(el->l_recs[index+1].e_cpos);
+                return 0;
+        }
+
+        if (!eb || (eb && !eb->h_next_leaf_blk)) {
+                /*
+                 * We are the last extent rec, so any high cpos should
+                 * be stored in this leaf refcount block.
+                 */
+                *cpos_end = UINT_MAX;
+                return 0;
+        }
+
+        /*
+         * If the extent block isn't the last one, we have to find
+         * the subtree root between this extent block and the next
+         * leaf extent block and get the corresponding e_cpos from
+         * the subroot. Otherwise we may corrupt the b-tree.
+         */
+        ocfs2_init_refcount_extent_tree(&et, ci, ref_root_bh);
+
+        left_path = ocfs2_new_path_from_et(&et);
+        if (!left_path) {
+                ret = -ENOMEM;
+                mlog_errno(ret);
+                goto out;
+        }
+
+        cpos = le32_to_cpu(eb->h_list.l_recs[index].e_cpos);
+        ret = ocfs2_find_path(ci, left_path, cpos);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
+        }
+
+        right_path = ocfs2_new_path_from_path(left_path);
+        if (!right_path) {
+                ret = -ENOMEM;
+                mlog_errno(ret);
+                goto out;
+        }
+
+        ret = ocfs2_find_cpos_for_right_leaf(sb, left_path, &cpos);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
+        }
+
+        ret = ocfs2_find_path(ci, right_path, cpos);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
+        }
+
+        subtree_root = ocfs2_find_subtree_root(&et, left_path,
+                                               right_path);
+
+        tmp_el = left_path->p_node[subtree_root].el;
+        blkno = left_path->p_node[subtree_root+1].bh->b_blocknr;
+        for (i = 0; i < le32_to_cpu(tmp_el->l_next_free_rec); i++) {
+                if (le64_to_cpu(tmp_el->l_recs[i].e_blkno) == blkno) {
+                        *cpos_end = le32_to_cpu(tmp_el->l_recs[i+1].e_cpos);
+                        break;
+                }
+        }
+
+        BUG_ON(i == le32_to_cpu(tmp_el->l_next_free_rec));
+
+out:
+        ocfs2_free_path(left_path);
+        ocfs2_free_path(right_path);
+        return ret;
+}
+
+/*
  * Given a cpos and len, try to find the refcount record which contains cpos.
  * 1. If cpos can be found in one refcount record, return the record.
  * 2. If cpos can't be found, return a fake record which start from cpos
@@ -983,10 +1062,10 @@ static int ocfs2_get_refcount_rec(struct ocfs2_caching_info *ci,
                                   struct buffer_head **ret_bh)
 {
         int ret = 0, i, found;
-        u32 low_cpos;
+        u32 low_cpos, uninitialized_var(cpos_end);
         struct ocfs2_extent_list *el;
-        struct ocfs2_extent_rec *tmp, *rec = NULL;
-        struct ocfs2_extent_block *eb;
+        struct ocfs2_extent_rec *rec = NULL;
+        struct ocfs2_extent_block *eb = NULL;
         struct buffer_head *eb_bh = NULL, *ref_leaf_bh = NULL;
         struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
         struct ocfs2_refcount_block *rb =
@@ -1034,12 +1113,16 @@ static int ocfs2_get_refcount_rec(struct ocfs2_caching_info *ci,
                 }
         }
 
-        /* adjust len when we have ocfs2_extent_rec after it. */
-        if (found && i < le16_to_cpu(el->l_next_free_rec) - 1) {
-                tmp = &el->l_recs[i+1];
+        if (found) {
+                ret = ocfs2_get_refcount_cpos_end(ci, ref_root_bh,
+                                                  eb, el, i, &cpos_end);
+                if (ret) {
+                        mlog_errno(ret);
+                        goto out;
+                }
 
-                if (le32_to_cpu(tmp->e_cpos) < cpos + len)
-                        len = le32_to_cpu(tmp->e_cpos) - cpos;
+                if (cpos_end < low_cpos + len)
+                        len = cpos_end - low_cpos;
         }
 
         ret = ocfs2_read_refcount_block(ci, le64_to_cpu(rec->e_blkno),
@@ -1246,7 +1329,7 @@ static int ocfs2_expand_inline_ref_root(handle_t *handle,
         memcpy(new_bh->b_data, ref_root_bh->b_data, sb->s_blocksize);
 
         new_rb = (struct ocfs2_refcount_block *)new_bh->b_data;
-        new_rb->rf_suballoc_slot = cpu_to_le16(OCFS2_SB(sb)->slot_num);
+        new_rb->rf_suballoc_slot = cpu_to_le16(meta_ac->ac_alloc_slot);
         new_rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
         new_rb->rf_blkno = cpu_to_le64(blkno);
         new_rb->rf_cpos = cpu_to_le32(0);
@@ -1418,7 +1501,7 @@ static int ocfs2_divide_leaf_refcount_block(struct buffer_head *ref_leaf_bh,
 
         /* change old and new rl_used accordingly. */
         le16_add_cpu(&rl->rl_used, -num_moved);
-        new_rl->rl_used = cpu_to_le32(num_moved);
+        new_rl->rl_used = cpu_to_le16(num_moved);
 
         sort(&rl->rl_recs, le16_to_cpu(rl->rl_used),
              sizeof(struct ocfs2_refcount_rec),
@@ -1492,7 +1575,7 @@ static int ocfs2_new_leaf_refcount_block(handle_t *handle,
         new_rb = (struct ocfs2_refcount_block *)new_bh->b_data;
         memset(new_rb, 0, sb->s_blocksize);
         strcpy((void *)new_rb, OCFS2_REFCOUNT_BLOCK_SIGNATURE);
-        new_rb->rf_suballoc_slot = cpu_to_le16(OCFS2_SB(sb)->slot_num);
+        new_rb->rf_suballoc_slot = cpu_to_le16(meta_ac->ac_alloc_slot);
         new_rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
         new_rb->rf_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
         new_rb->rf_blkno = cpu_to_le64(blkno);
@@ -1797,7 +1880,8 @@ static int ocfs2_split_refcount_rec(handle_t *handle,
                 recs_need++;
 
         /* If the leaf block don't have enough record, expand it. */
-        if (le16_to_cpu(rf_list->rl_used) + recs_need > rf_list->rl_count) {
+        if (le16_to_cpu(rf_list->rl_used) + recs_need >
+                                         le16_to_cpu(rf_list->rl_count)) {
                 struct ocfs2_refcount_rec tmp_rec;
                 u64 cpos = le64_to_cpu(orig_rec->r_cpos);
                 len = le32_to_cpu(orig_rec->r_clusters);
@@ -1859,7 +1943,7 @@ static int ocfs2_split_refcount_rec(handle_t *handle,
                 memcpy(tail_rec, orig_rec, sizeof(struct ocfs2_refcount_rec));
                 le64_add_cpu(&tail_rec->r_cpos,
                              le32_to_cpu(tail_rec->r_clusters) - len);
-                tail_rec->r_clusters = le32_to_cpu(len);
+                tail_rec->r_clusters = cpu_to_le32(len);
         }
 
         /*
@@ -2431,7 +2515,7 @@ out:
  * we gonna touch and whether we need to create new blocks.
  *
  * Normally the refcount blocks store these refcount should be
- * continguous also, so that we can get the number easily.
+ * contiguous also, so that we can get the number easily.
  * As for meta_ac, we will at most add split 2 refcount record and
  * 2 more refcount block, so just check it in a rough way.
  *
@@ -2860,7 +2944,7 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle,
 
         while (offset < end) {
                 page_index = offset >> PAGE_CACHE_SHIFT;
-                map_end = (page_index + 1) << PAGE_CACHE_SHIFT;
+                map_end = ((loff_t)page_index + 1) << PAGE_CACHE_SHIFT;
                 if (map_end > end)
                         map_end = end;
 
@@ -2872,8 +2956,12 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle,
 
                 page = grab_cache_page(mapping, page_index);
 
-                /* This page can't be dirtied before we CoW it out. */
-                BUG_ON(PageDirty(page));
+                /*
+                 * In case PAGE_CACHE_SIZE <= CLUSTER_SIZE, This page
+                 * can't be dirtied before we CoW it out.
+                 */
+                if (PAGE_CACHE_SIZE <= OCFS2_SB(sb)->s_clustersize)
+                        BUG_ON(PageDirty(page));
 
                 if (!PageUptodate(page)) {
                         ret = block_read_full_page(page, ocfs2_get_block);
@@ -3085,7 +3173,7 @@ static int ocfs2_cow_sync_writeback(struct super_block *sb,
 
         while (offset < end) {
                 page_index = offset >> PAGE_CACHE_SHIFT;
-                map_end = (page_index + 1) << PAGE_CACHE_SHIFT;
+                map_end = ((loff_t)page_index + 1) << PAGE_CACHE_SHIFT;
                 if (map_end > end)
                         map_end = end;
 
@@ -3840,8 +3928,7 @@ static int ocfs2_add_refcounted_extent(struct inode *inode,
         }
 
         ret = ocfs2_insert_extent(handle, et, cpos,
-                        cpu_to_le64(ocfs2_clusters_to_blocks(inode->i_sb,
-                                                             p_cluster)),
+                        ocfs2_clusters_to_blocks(inode->i_sb, p_cluster),
                         num_clusters, ext_flags, meta_ac);
         if (ret) {
                 mlog_errno(ret);
@@ -3987,6 +4074,7 @@ static int ocfs2_complete_reflink(struct inode *s_inode,
         OCFS2_I(t_inode)->ip_dyn_features = OCFS2_I(s_inode)->ip_dyn_features;
         spin_unlock(&OCFS2_I(t_inode)->ip_lock);
         i_size_write(t_inode, size);
+        t_inode->i_blocks = s_inode->i_blocks;
 
         di->i_xattr_inline_size = s_di->i_xattr_inline_size;
         di->i_clusters = s_di->i_clusters;
@@ -3995,6 +4083,9 @@ static int ocfs2_complete_reflink(struct inode *s_inode,
         di->i_attr = s_di->i_attr;
 
         if (preserve) {
+                t_inode->i_uid = s_inode->i_uid;
+                t_inode->i_gid = s_inode->i_gid;
+                t_inode->i_mode = s_inode->i_mode;
                 di->i_uid = s_di->i_uid;
                 di->i_gid = s_di->i_gid;
                 di->i_mode = s_di->i_mode;
@@ -4253,8 +4344,8 @@ static int ocfs2_user_path_parent(const char __user *path,
  * @new_dentry:        target dentry
  * @preserve:  if true, preserve all file attributes
  */
-int ocfs2_vfs_reflink(struct dentry *old_dentry, struct inode *dir,
-                      struct dentry *new_dentry, bool preserve)
+static int ocfs2_vfs_reflink(struct dentry *old_dentry, struct inode *dir,
+                             struct dentry *new_dentry, bool preserve)
 {
         struct inode *inode = old_dentry->d_inode;
         int error;
@@ -4302,7 +4393,7 @@ int ocfs2_vfs_reflink(struct dentry *old_dentry, struct inode *dir,
         }
 
         mutex_lock(&inode->i_mutex);
-        vfs_dq_init(dir);
+        dquot_initialize(dir);
         error = ocfs2_reflink(old_dentry, dir, new_dentry, preserve);
         mutex_unlock(&inode->i_mutex);
         if (!error)
diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c
index e49c41050264..0d3049f696c5 100644
--- a/fs/ocfs2/stack_o2cb.c
+++ b/fs/ocfs2/stack_o2cb.c
@@ -19,6 +19,7 @@
 
 #include <linux/kernel.h>
 #include <linux/crc32.h>
+#include <linux/slab.h>
 #include <linux/module.h>
 
 /* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */
@@ -161,24 +162,23 @@ static int dlm_status_to_errno(enum dlm_status status)
 
 static void o2dlm_lock_ast_wrapper(void *astarg)
 {
-        BUG_ON(o2cb_stack.sp_proto == NULL);
+        struct ocfs2_dlm_lksb *lksb = astarg;
 
-        o2cb_stack.sp_proto->lp_lock_ast(astarg);
+        lksb->lksb_conn->cc_proto->lp_lock_ast(lksb);
 }
 
 static void o2dlm_blocking_ast_wrapper(void *astarg, int level)
 {
-        BUG_ON(o2cb_stack.sp_proto == NULL);
+        struct ocfs2_dlm_lksb *lksb = astarg;
 
-        o2cb_stack.sp_proto->lp_blocking_ast(astarg, level);
+        lksb->lksb_conn->cc_proto->lp_blocking_ast(lksb, level);
 }
 
 static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status)
 {
+        struct ocfs2_dlm_lksb *lksb = astarg;
         int error = dlm_status_to_errno(status);
 
-        BUG_ON(o2cb_stack.sp_proto == NULL);
-
         /*
          * In o2dlm, you can get both the lock_ast() for the lock being
          * granted and the unlock_ast() for the CANCEL failing.  A
@@ -193,16 +193,15 @@ static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status)
         if (status == DLM_CANCELGRANT)
                 return;
 
-        o2cb_stack.sp_proto->lp_unlock_ast(astarg, error);
+        lksb->lksb_conn->cc_proto->lp_unlock_ast(lksb, error);
 }
 
 static int o2cb_dlm_lock(struct ocfs2_cluster_connection *conn,
                          int mode,
-                         union ocfs2_dlm_lksb *lksb,
+                         struct ocfs2_dlm_lksb *lksb,
                          u32 flags,
                          void *name,
-                         unsigned int namelen,
-                         void *astarg)
+                         unsigned int namelen)
 {
         enum dlm_status status;
         int o2dlm_mode = mode_to_o2dlm(mode);
@@ -211,28 +210,27 @@ static int o2cb_dlm_lock(struct ocfs2_cluster_connection *conn,
 
         status = dlmlock(conn->cc_lockspace, o2dlm_mode, &lksb->lksb_o2dlm,
                          o2dlm_flags, name, namelen,
-                         o2dlm_lock_ast_wrapper, astarg,
+                         o2dlm_lock_ast_wrapper, lksb,
                          o2dlm_blocking_ast_wrapper);
         ret = dlm_status_to_errno(status);
         return ret;
 }
 
 static int o2cb_dlm_unlock(struct ocfs2_cluster_connection *conn,
-                           union ocfs2_dlm_lksb *lksb,
-                           u32 flags,
-                           void *astarg)
+                           struct ocfs2_dlm_lksb *lksb,
+                           u32 flags)
 {
         enum dlm_status status;
         int o2dlm_flags = flags_to_o2dlm(flags);
         int ret;
 
         status = dlmunlock(conn->cc_lockspace, &lksb->lksb_o2dlm,
-                           o2dlm_flags, o2dlm_unlock_ast_wrapper, astarg);
+                           o2dlm_flags, o2dlm_unlock_ast_wrapper, lksb);
         ret = dlm_status_to_errno(status);
         return ret;
 }
 
-static int o2cb_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
+static int o2cb_dlm_lock_status(struct ocfs2_dlm_lksb *lksb)
 {
         return dlm_status_to_errno(lksb->lksb_o2dlm.status);
 }
@@ -242,17 +240,17 @@ static int o2cb_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
  * contents, it will zero out the LVB.  Thus the caller can always trust
  * the contents.
  */
-static int o2cb_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb)
+static int o2cb_dlm_lvb_valid(struct ocfs2_dlm_lksb *lksb)
 {
         return 1;
 }
 
-static void *o2cb_dlm_lvb(union ocfs2_dlm_lksb *lksb)
+static void *o2cb_dlm_lvb(struct ocfs2_dlm_lksb *lksb)
 {
         return (void *)(lksb->lksb_o2dlm.lvb);
 }
 
-static void o2cb_dump_lksb(union ocfs2_dlm_lksb *lksb)
+static void o2cb_dump_lksb(struct ocfs2_dlm_lksb *lksb)
 {
         dlm_print_one_lock(lksb->lksb_o2dlm.lockid);
 }
@@ -277,10 +275,10 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn)
         u32 dlm_key;
         struct dlm_ctxt *dlm;
         struct o2dlm_private *priv;
-        struct dlm_protocol_version dlm_version;
+        struct dlm_protocol_version fs_version;
 
         BUG_ON(conn == NULL);
-        BUG_ON(o2cb_stack.sp_proto == NULL);
+        BUG_ON(conn->cc_proto == NULL);
 
         /* for now we only have one cluster/node, make sure we see it
          * in the heartbeat universe */
@@ -304,18 +302,18 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn)
         /* used by the dlm code to make message headers unique, each
          * node in this domain must agree on this. */
         dlm_key = crc32_le(0, conn->cc_name, conn->cc_namelen);
-        dlm_version.pv_major = conn->cc_version.pv_major;
-        dlm_version.pv_minor = conn->cc_version.pv_minor;
+        fs_version.pv_major = conn->cc_version.pv_major;
+        fs_version.pv_minor = conn->cc_version.pv_minor;
 
-        dlm = dlm_register_domain(conn->cc_name, dlm_key, &dlm_version);
+        dlm = dlm_register_domain(conn->cc_name, dlm_key, &fs_version);
         if (IS_ERR(dlm)) {
                 rc = PTR_ERR(dlm);
                 mlog_errno(rc);
                 goto out_free;
         }
 
-        conn->cc_version.pv_major = dlm_version.pv_major;
-        conn->cc_version.pv_minor = dlm_version.pv_minor;
+        conn->cc_version.pv_major = fs_version.pv_major;
+        conn->cc_version.pv_minor = fs_version.pv_minor;
         conn->cc_lockspace = dlm;
 
         dlm_register_eviction_cb(dlm, &priv->op_eviction_cb);
diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c
index ff4c798a5635..2dc57bca0688 100644
--- a/fs/ocfs2/stack_user.c
+++ b/fs/ocfs2/stack_user.c
@@ -21,11 +21,11 @@
 #include <linux/fs.h>
 #include <linux/miscdevice.h>
 #include <linux/mutex.h>
+#include <linux/slab.h>
 #include <linux/smp_lock.h>
 #include <linux/reboot.h>
 #include <asm/uaccess.h>
 
-#include "ocfs2.h"  /* For struct ocfs2_lock_res */
 #include "stackglue.h"
 
 #include <linux/dlm_plock.h>
@@ -63,8 +63,8 @@
  * negotiated by the client.  The client negotiates based on the maximum
  * version advertised in /sys/fs/ocfs2/max_locking_protocol.  The major
  * number from the "SETV" message must match
- * ocfs2_user_plugin.sp_proto->lp_max_version.pv_major, and the minor number
- * must be less than or equal to ...->lp_max_version.pv_minor.
+ * ocfs2_user_plugin.sp_max_proto.pv_major, and the minor number
+ * must be less than or equal to ...sp_max_version.pv_minor.
  *
  * Once this information has been set, mounts will be allowed.  From this
  * point on, the "DOWN" message can be sent for node down notification.
@@ -401,7 +401,7 @@ static int ocfs2_control_do_setversion_msg(struct file *file,
         char *ptr = NULL;
         struct ocfs2_control_private *p = file->private_data;
         struct ocfs2_protocol_version *max =
-                &ocfs2_user_plugin.sp_proto->lp_max_version;
+                &ocfs2_user_plugin.sp_max_proto;
 
         if (ocfs2_control_get_handshake_state(file) !=
             OCFS2_CONTROL_HANDSHAKE_PROTOCOL)
@@ -664,18 +664,10 @@ static void ocfs2_control_exit(void)
                        -rc);
 }
 
-static struct dlm_lksb *fsdlm_astarg_to_lksb(void *astarg)
-{
-        struct ocfs2_lock_res *res = astarg;
-        return &res->l_lksb.lksb_fsdlm;
-}
-
 static void fsdlm_lock_ast_wrapper(void *astarg)
 {
-        struct dlm_lksb *lksb = fsdlm_astarg_to_lksb(astarg);
-        int status = lksb->sb_status;
-
-        BUG_ON(ocfs2_user_plugin.sp_proto == NULL);
+        struct ocfs2_dlm_lksb *lksb = astarg;
+        int status = lksb->lksb_fsdlm.sb_status;
 
         /*
          * For now we're punting on the issue of other non-standard errors
@@ -688,25 +680,24 @@ static void fsdlm_lock_ast_wrapper(void *astarg)
          */
 
         if (status == -DLM_EUNLOCK || status == -DLM_ECANCEL)
-                ocfs2_user_plugin.sp_proto->lp_unlock_ast(astarg, 0);
+                lksb->lksb_conn->cc_proto->lp_unlock_ast(lksb, 0);
         else
-                ocfs2_user_plugin.sp_proto->lp_lock_ast(astarg);
+                lksb->lksb_conn->cc_proto->lp_lock_ast(lksb);
 }
 
 static void fsdlm_blocking_ast_wrapper(void *astarg, int level)
 {
-        BUG_ON(ocfs2_user_plugin.sp_proto == NULL);
+        struct ocfs2_dlm_lksb *lksb = astarg;
 
-        ocfs2_user_plugin.sp_proto->lp_blocking_ast(astarg, level);
+        lksb->lksb_conn->cc_proto->lp_blocking_ast(lksb, level);
 }
 
 static int user_dlm_lock(struct ocfs2_cluster_connection *conn,
                          int mode,
-                         union ocfs2_dlm_lksb *lksb,
+                         struct ocfs2_dlm_lksb *lksb,
                          u32 flags,
                          void *name,
-                         unsigned int namelen,
-                         void *astarg)
+                         unsigned int namelen)
 {
         int ret;
 
@@ -716,36 +707,35 @@ static int user_dlm_lock(struct ocfs2_cluster_connection *conn,
 
         ret = dlm_lock(conn->cc_lockspace, mode, &lksb->lksb_fsdlm,
                        flags|DLM_LKF_NODLCKWT, name, namelen, 0,
-                       fsdlm_lock_ast_wrapper, astarg,
+                       fsdlm_lock_ast_wrapper, lksb,
                        fsdlm_blocking_ast_wrapper);
         return ret;
 }
 
 static int user_dlm_unlock(struct ocfs2_cluster_connection *conn,
-                           union ocfs2_dlm_lksb *lksb,
-                           u32 flags,
-                           void *astarg)
+                           struct ocfs2_dlm_lksb *lksb,
+                           u32 flags)
 {
         int ret;
 
         ret = dlm_unlock(conn->cc_lockspace, lksb->lksb_fsdlm.sb_lkid,
-                         flags, &lksb->lksb_fsdlm, astarg);
+                         flags, &lksb->lksb_fsdlm, lksb);
         return ret;
 }
 
-static int user_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
+static int user_dlm_lock_status(struct ocfs2_dlm_lksb *lksb)
 {
         return lksb->lksb_fsdlm.sb_status;
 }
 
-static int user_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb)
+static int user_dlm_lvb_valid(struct ocfs2_dlm_lksb *lksb)
 {
         int invalid = lksb->lksb_fsdlm.sb_flags & DLM_SBF_VALNOTVALID;
 
         return !invalid;
 }
 
-static void *user_dlm_lvb(union ocfs2_dlm_lksb *lksb)
+static void *user_dlm_lvb(struct ocfs2_dlm_lksb *lksb)
 {
         if (!lksb->lksb_fsdlm.sb_lvbptr)
                 lksb->lksb_fsdlm.sb_lvbptr = (char *)lksb +
@@ -753,7 +743,7 @@ static void *user_dlm_lvb(union ocfs2_dlm_lksb *lksb)
         return (void *)(lksb->lksb_fsdlm.sb_lvbptr);
 }
 
-static void user_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb)
+static void user_dlm_dump_lksb(struct ocfs2_dlm_lksb *lksb)
 {
 }
 
@@ -814,7 +804,7 @@ static int fs_protocol_compare(struct ocfs2_protocol_version *existing,
 static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
 {
         dlm_lockspace_t *fsdlm;
-        struct ocfs2_live_connection *control;
+        struct ocfs2_live_connection *uninitialized_var(control);
         int rc = 0;
 
         BUG_ON(conn == NULL);
diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c
index 3f2f1c45b7b6..39abf89697ed 100644
--- a/fs/ocfs2/stackglue.c
+++ b/fs/ocfs2/stackglue.c
@@ -36,7 +36,7 @@
 #define OCFS2_STACK_PLUGIN_USER         "user"
 #define OCFS2_MAX_HB_CTL_PATH           256
 
-static struct ocfs2_locking_protocol *lproto;
+static struct ocfs2_protocol_version locking_max_version;
 static DEFINE_SPINLOCK(ocfs2_stack_lock);
 static LIST_HEAD(ocfs2_stack_list);
 static char cluster_stack_name[OCFS2_STACK_LABEL_LEN + 1];
@@ -176,7 +176,7 @@ int ocfs2_stack_glue_register(struct ocfs2_stack_plugin *plugin)
         spin_lock(&ocfs2_stack_lock);
         if (!ocfs2_stack_lookup(plugin->sp_name)) {
                 plugin->sp_count = 0;
-                plugin->sp_proto = lproto;
+                plugin->sp_max_proto = locking_max_version;
                 list_add(&plugin->sp_list, &ocfs2_stack_list);
                 printk(KERN_INFO "ocfs2: Registered cluster interface %s\n",
                        plugin->sp_name);
@@ -213,77 +213,76 @@ void ocfs2_stack_glue_unregister(struct ocfs2_stack_plugin *plugin)
 }
 EXPORT_SYMBOL_GPL(ocfs2_stack_glue_unregister);
 
-void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto)
+void ocfs2_stack_glue_set_max_proto_version(struct ocfs2_protocol_version *max_proto)
 {
         struct ocfs2_stack_plugin *p;
 
-        BUG_ON(proto == NULL);
-
         spin_lock(&ocfs2_stack_lock);
-        BUG_ON(active_stack != NULL);
+        if (memcmp(max_proto, &locking_max_version,
+                   sizeof(struct ocfs2_protocol_version))) {
+                BUG_ON(locking_max_version.pv_major != 0);
 
-        lproto = proto;
-        list_for_each_entry(p, &ocfs2_stack_list, sp_list) {
-                p->sp_proto = lproto;
+                locking_max_version = *max_proto;
+                list_for_each_entry(p, &ocfs2_stack_list, sp_list) {
+                        p->sp_max_proto = locking_max_version;
+                }
         }
-
         spin_unlock(&ocfs2_stack_lock);
 }
-EXPORT_SYMBOL_GPL(ocfs2_stack_glue_set_locking_protocol);
+EXPORT_SYMBOL_GPL(ocfs2_stack_glue_set_max_proto_version);
 
 
 /*
- * The ocfs2_dlm_lock() and ocfs2_dlm_unlock() functions take
- * "struct ocfs2_lock_res *astarg" instead of "void *astarg" because the
- * underlying stack plugins need to pilfer the lksb off of the lock_res.
- * If some other structure needs to be passed as an astarg, the plugins
- * will need to be given a different avenue to the lksb.
+ * The ocfs2_dlm_lock() and ocfs2_dlm_unlock() functions take no argument
+ * for the ast and bast functions.  They will pass the lksb to the ast
+ * and bast.  The caller can wrap the lksb with their own structure to
+ * get more information.
  */
 int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn,
                    int mode,
-                   union ocfs2_dlm_lksb *lksb,
+                   struct ocfs2_dlm_lksb *lksb,
                    u32 flags,
                    void *name,
-                   unsigned int namelen,
-                   struct ocfs2_lock_res *astarg)
+                   unsigned int namelen)
 {
-        BUG_ON(lproto == NULL);
-
+        if (!lksb->lksb_conn)
+                lksb->lksb_conn = conn;
+        else
+                BUG_ON(lksb->lksb_conn != conn);
         return active_stack->sp_ops->dlm_lock(conn, mode, lksb, flags,
-                                              name, namelen, astarg);
+                                              name, namelen);
 }
 EXPORT_SYMBOL_GPL(ocfs2_dlm_lock);
 
 int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn,
-                     union ocfs2_dlm_lksb *lksb,
-                     u32 flags,
-                     struct ocfs2_lock_res *astarg)
+                     struct ocfs2_dlm_lksb *lksb,
+                     u32 flags)
 {
-        BUG_ON(lproto == NULL);
+        BUG_ON(lksb->lksb_conn == NULL);
 
-        return active_stack->sp_ops->dlm_unlock(conn, lksb, flags, astarg);
+        return active_stack->sp_ops->dlm_unlock(conn, lksb, flags);
 }
 EXPORT_SYMBOL_GPL(ocfs2_dlm_unlock);
 
-int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
+int ocfs2_dlm_lock_status(struct ocfs2_dlm_lksb *lksb)
 {
         return active_stack->sp_ops->lock_status(lksb);
 }
 EXPORT_SYMBOL_GPL(ocfs2_dlm_lock_status);
 
-int ocfs2_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb)
+int ocfs2_dlm_lvb_valid(struct ocfs2_dlm_lksb *lksb)
 {
         return active_stack->sp_ops->lvb_valid(lksb);
 }
 EXPORT_SYMBOL_GPL(ocfs2_dlm_lvb_valid);
 
-void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb)
+void *ocfs2_dlm_lvb(struct ocfs2_dlm_lksb *lksb)
 {
         return active_stack->sp_ops->lock_lvb(lksb);
 }
 EXPORT_SYMBOL_GPL(ocfs2_dlm_lvb);
 
-void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb)
+void ocfs2_dlm_dump_lksb(struct ocfs2_dlm_lksb *lksb)
 {
         active_stack->sp_ops->dump_lksb(lksb);
 }
@@ -312,6 +311,7 @@ EXPORT_SYMBOL_GPL(ocfs2_plock);
 int ocfs2_cluster_connect(const char *stack_name,
                           const char *group,
                           int grouplen,
+                          struct ocfs2_locking_protocol *lproto,
                           void (*recovery_handler)(int node_num,
                                                    void *recovery_data),
                           void *recovery_data,
@@ -329,6 +329,12 @@ int ocfs2_cluster_connect(const char *stack_name,
                 goto out;
         }
 
+        if (memcmp(&lproto->lp_max_version, &locking_max_version,
+                   sizeof(struct ocfs2_protocol_version))) {
+                rc = -EINVAL;
+                goto out;
+        }
+
         new_conn = kzalloc(sizeof(struct ocfs2_cluster_connection),
                            GFP_KERNEL);
         if (!new_conn) {
@@ -341,6 +347,7 @@ int ocfs2_cluster_connect(const char *stack_name,
         new_conn->cc_recovery_handler = recovery_handler;
         new_conn->cc_recovery_data = recovery_data;
 
+        new_conn->cc_proto = lproto;
         /* Start the new connection at our maximum compatibility level */
         new_conn->cc_version = lproto->lp_max_version;
 
@@ -366,6 +373,24 @@ out:
 }
 EXPORT_SYMBOL_GPL(ocfs2_cluster_connect);
 
+/* The caller will ensure all nodes have the same cluster stack */
+int ocfs2_cluster_connect_agnostic(const char *group,
+                                   int grouplen,
+                                   struct ocfs2_locking_protocol *lproto,
+                                   void (*recovery_handler)(int node_num,
+                                                            void *recovery_data),
+                                   void *recovery_data,
+                                   struct ocfs2_cluster_connection **conn)
+{
+        char *stack_name = NULL;
+
+        if (cluster_stack_name[0])
+                stack_name = cluster_stack_name;
+        return ocfs2_cluster_connect(stack_name, group, grouplen, lproto,
+                                     recovery_handler, recovery_data, conn);
+}
+EXPORT_SYMBOL_GPL(ocfs2_cluster_connect_agnostic);
+
 /* If hangup_pending is 0, the stack driver will be dropped */
 int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn,
                              int hangup_pending)
@@ -453,10 +478,10 @@ static ssize_t ocfs2_max_locking_protocol_show(struct kobject *kobj,
         ssize_t ret = 0;
 
         spin_lock(&ocfs2_stack_lock);
-        if (lproto)
+        if (locking_max_version.pv_major)
                 ret = snprintf(buf, PAGE_SIZE, "%u.%u\n",
-                               lproto->lp_max_version.pv_major,
-                               lproto->lp_max_version.pv_minor);
+                               locking_max_version.pv_major,
+                               locking_max_version.pv_minor);
         spin_unlock(&ocfs2_stack_lock);
 
         return ret;
@@ -620,51 +645,46 @@ error:
 
 static ctl_table ocfs2_nm_table[] = {
         {
-                .ctl_name       = 1,
                 .procname       = "hb_ctl_path",
                 .data           = ocfs2_hb_ctl_path,
                 .maxlen         = OCFS2_MAX_HB_CTL_PATH,
                 .mode           = 0644,
-                .proc_handler   = &proc_dostring,
-                .strategy       = &sysctl_string,
+                .proc_handler   = proc_dostring,
         },
-        { .ctl_name = 0 }
+        { }
 };
 
 static ctl_table ocfs2_mod_table[] = {
         {
-                .ctl_name       = FS_OCFS2_NM,
                 .procname       = "nm",
                 .data           = NULL,
                 .maxlen         = 0,
                 .mode           = 0555,
                 .child          = ocfs2_nm_table
         },
-        { .ctl_name = 0}
+        { }
 };
 
 static ctl_table ocfs2_kern_table[] = {
         {
-                .ctl_name       = FS_OCFS2,
                 .procname       = "ocfs2",
                 .data           = NULL,
                 .maxlen         = 0,
                 .mode           = 0555,
                 .child          = ocfs2_mod_table
         },
-        { .ctl_name = 0}
+        { }
 };
 
 static ctl_table ocfs2_root_table[] = {
         {
-                .ctl_name       = CTL_FS,
                 .procname       = "fs",
                 .data           = NULL,
                 .maxlen         = 0,
                 .mode           = 0555,
                 .child          = ocfs2_kern_table
         },
-        { .ctl_name = 0 }
+        { }
 };
 
 static struct ctl_table_header *ocfs2_table_header = NULL;
@@ -690,7 +710,10 @@ static int __init ocfs2_stack_glue_init(void)
 
 static void __exit ocfs2_stack_glue_exit(void)
 {
-        lproto = NULL;
+        memset(&locking_max_version, 0,
+               sizeof(struct ocfs2_protocol_version));
+        locking_max_version.pv_major = 0;
+        locking_max_version.pv_minor = 0;
         ocfs2_sysfs_exit();
         if (ocfs2_table_header)
                 unregister_sysctl_table(ocfs2_table_header);
diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h
index 03a44d60eac9..8ce7398ae1d2 100644
--- a/fs/ocfs2/stackglue.h
+++ b/fs/ocfs2/stackglue.h
@@ -56,17 +56,6 @@ struct ocfs2_protocol_version {
 };
 
 /*
- * The ocfs2_locking_protocol defines the handlers called on ocfs2's behalf.
- */
-struct ocfs2_locking_protocol {
-        struct ocfs2_protocol_version lp_max_version;
-        void (*lp_lock_ast)(void *astarg);
-        void (*lp_blocking_ast)(void *astarg, int level);
-        void (*lp_unlock_ast)(void *astarg, int error);
-};
-
-
-/*
  * The dlm_lockstatus struct includes lvb space, but the dlm_lksb struct only
  * has a pointer to separately allocated lvb space.  This struct exists only to
  * include in the lksb union to make space for a combined dlm_lksb and lvb.
@@ -81,12 +70,27 @@ struct fsdlm_lksb_plus_lvb {
  * size of the union is known.  Lock status structures are embedded in
  * ocfs2 inodes.
  */
-union ocfs2_dlm_lksb {
-        struct dlm_lockstatus lksb_o2dlm;
-        struct dlm_lksb lksb_fsdlm;
-        struct fsdlm_lksb_plus_lvb padding;
+struct ocfs2_cluster_connection;
+struct ocfs2_dlm_lksb {
+         union {
+                 struct dlm_lockstatus lksb_o2dlm;
+                 struct dlm_lksb lksb_fsdlm;
+                 struct fsdlm_lksb_plus_lvb padding;
+         };
+         struct ocfs2_cluster_connection *lksb_conn;
+};
+
+/*
+ * The ocfs2_locking_protocol defines the handlers called on ocfs2's behalf.
+ */
+struct ocfs2_locking_protocol {
+        struct ocfs2_protocol_version lp_max_version;
+        void (*lp_lock_ast)(struct ocfs2_dlm_lksb *lksb);
+        void (*lp_blocking_ast)(struct ocfs2_dlm_lksb *lksb, int level);
+        void (*lp_unlock_ast)(struct ocfs2_dlm_lksb *lksb, int error);
 };
 
+
 /*
  * A cluster connection.  Mostly opaque to ocfs2, the connection holds
  * state for the underlying stack.  ocfs2 does use cc_version to determine
@@ -96,6 +100,7 @@ struct ocfs2_cluster_connection {
         char cc_name[GROUP_NAME_MAX];
         int cc_namelen;
         struct ocfs2_protocol_version cc_version;
+        struct ocfs2_locking_protocol *cc_proto;
         void (*cc_recovery_handler)(int node_num, void *recovery_data);
         void *cc_recovery_data;
         void *cc_lockspace;
@@ -155,27 +160,29 @@ struct ocfs2_stack_operations {
          *
          * ast and bast functions are not part of the call because the
          * stack will likely want to wrap ast and bast calls before passing
-         * them to stack->sp_proto.
+         * them to stack->sp_proto.  There is no astarg.  The lksb will
+         * be passed back to the ast and bast functions.  The caller can
+         * use this to find their object.
          */
         int (*dlm_lock)(struct ocfs2_cluster_connection *conn,
                         int mode,
-                        union ocfs2_dlm_lksb *lksb,
+                        struct ocfs2_dlm_lksb *lksb,
                         u32 flags,
                         void *name,
-                        unsigned int namelen,
-                        void *astarg);
+                        unsigned int namelen);
 
         /*
          * Call the underlying dlm unlock function.  The ->dlm_unlock()
          * function should convert the flags as appropriate.
          *
          * The unlock ast is not passed, as the stack will want to wrap
-         * it before calling stack->sp_proto->lp_unlock_ast().
+         * it before calling stack->sp_proto->lp_unlock_ast().  There is
+         * no astarg.  The lksb will be passed back to the unlock ast
+         * function.  The caller can use this to find their object.
          */
         int (*dlm_unlock)(struct ocfs2_cluster_connection *conn,
-                          union ocfs2_dlm_lksb *lksb,
-                          u32 flags,
-                          void *astarg);
+                          struct ocfs2_dlm_lksb *lksb,
+                          u32 flags);
 
         /*
          * Return the status of the current lock status block.  The fs
@@ -183,17 +190,17 @@ struct ocfs2_stack_operations {
          * callback pulls out the stack-specific lksb, converts the status
          * to a proper errno, and returns it.
          */
-        int (*lock_status)(union ocfs2_dlm_lksb *lksb);
+        int (*lock_status)(struct ocfs2_dlm_lksb *lksb);
 
         /*
          * Return non-zero if the LVB is valid.
          */
-        int (*lvb_valid)(union ocfs2_dlm_lksb *lksb);
+        int (*lvb_valid)(struct ocfs2_dlm_lksb *lksb);
 
         /*
          * Pull the lvb pointer off of the stack-specific lksb.
          */
-        void *(*lock_lvb)(union ocfs2_dlm_lksb *lksb);
+        void *(*lock_lvb)(struct ocfs2_dlm_lksb *lksb);
 
         /*
          * Cluster-aware posix locks
@@ -210,7 +217,7 @@ struct ocfs2_stack_operations {
          * This is an optoinal debugging hook.  If provided, the
          * stack can dump debugging information about this lock.
          */
-        void (*dump_lksb)(union ocfs2_dlm_lksb *lksb);
+        void (*dump_lksb)(struct ocfs2_dlm_lksb *lksb);
 };
 
 /*
@@ -226,7 +233,7 @@ struct ocfs2_stack_plugin {
         /* These are managed by the stackglue code. */
         struct list_head sp_list;
         unsigned int sp_count;
-        struct ocfs2_locking_protocol *sp_proto;
+        struct ocfs2_protocol_version sp_max_proto;
 };
 
 
@@ -234,10 +241,22 @@ struct ocfs2_stack_plugin {
 int ocfs2_cluster_connect(const char *stack_name,
                           const char *group,
                           int grouplen,
+                          struct ocfs2_locking_protocol *lproto,
                           void (*recovery_handler)(int node_num,
                                                    void *recovery_data),
                           void *recovery_data,
                           struct ocfs2_cluster_connection **conn);
+/*
+ * Used by callers that don't store their stack name.  They must ensure
+ * all nodes have the same stack.
+ */
+int ocfs2_cluster_connect_agnostic(const char *group,
+                                   int grouplen,
+                                   struct ocfs2_locking_protocol *lproto,
+                                   void (*recovery_handler)(int node_num,
+                                                            void *recovery_data),
+                                   void *recovery_data,
+                                   struct ocfs2_cluster_connection **conn);
 int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn,
                              int hangup_pending);
 void ocfs2_cluster_hangup(const char *group, int grouplen);
@@ -246,26 +265,24 @@ int ocfs2_cluster_this_node(unsigned int *node);
 struct ocfs2_lock_res;
 int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn,
                    int mode,
-                   union ocfs2_dlm_lksb *lksb,
+                   struct ocfs2_dlm_lksb *lksb,
                    u32 flags,
                    void *name,
-                   unsigned int namelen,
-                   struct ocfs2_lock_res *astarg);
+                   unsigned int namelen);
 int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn,
-                     union ocfs2_dlm_lksb *lksb,
-                     u32 flags,
-                     struct ocfs2_lock_res *astarg);
+                     struct ocfs2_dlm_lksb *lksb,
+                     u32 flags);
 
-int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb);
-int ocfs2_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb);
-void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb);
-void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb);
+int ocfs2_dlm_lock_status(struct ocfs2_dlm_lksb *lksb);
+int ocfs2_dlm_lvb_valid(struct ocfs2_dlm_lksb *lksb);
+void *ocfs2_dlm_lvb(struct ocfs2_dlm_lksb *lksb);
+void ocfs2_dlm_dump_lksb(struct ocfs2_dlm_lksb *lksb);
 
 int ocfs2_stack_supports_plocks(void);
 int ocfs2_plock(struct ocfs2_cluster_connection *conn, u64 ino,
                 struct file *file, int cmd, struct file_lock *fl);
 
-void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto);
+void ocfs2_stack_glue_set_max_proto_version(struct ocfs2_protocol_version *max_proto);
 
 
 /* Used by stack plugins */
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index c30b644d9572..19ba00f28547 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -51,7 +51,7 @@
 #define ALLOC_NEW_GROUP                 0x1
 #define ALLOC_GROUPS_FROM_GLOBAL        0x2
 
-#define OCFS2_MAX_INODES_TO_STEAL       1024
+#define OCFS2_MAX_TO_STEAL              1024
 
 static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg);
 static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe);
@@ -95,13 +95,6 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
                                              struct buffer_head *group_bh,
                                              unsigned int bit_off,
                                              unsigned int num_bits);
-static inline int ocfs2_block_group_clear_bits(handle_t *handle,
-                                               struct inode *alloc_inode,
-                                               struct ocfs2_group_desc *bg,
-                                               struct buffer_head *group_bh,
-                                               unsigned int bit_off,
-                                               unsigned int num_bits);
-
 static int ocfs2_relink_block_group(handle_t *handle,
                                     struct inode *alloc_inode,
                                     struct buffer_head *fe_bh,
@@ -152,7 +145,7 @@ static u32 ocfs2_bits_per_group(struct ocfs2_chain_list *cl)
 
 #define do_error(fmt, ...)                                              \
         do{                                                             \
-                if (clean_error)                                        \
+                if (resize)                                     \
                         mlog(ML_ERROR, fmt "\n", ##__VA_ARGS__);        \
                 else                                                    \
                         ocfs2_error(sb, fmt, ##__VA_ARGS__);            \
@@ -160,7 +153,7 @@ static u32 ocfs2_bits_per_group(struct ocfs2_chain_list *cl)
 
 static int ocfs2_validate_gd_self(struct super_block *sb,
                                   struct buffer_head *bh,
-                                  int clean_error)
+                                  int resize)
 {
         struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
 
@@ -211,7 +204,7 @@ static int ocfs2_validate_gd_self(struct super_block *sb,
 static int ocfs2_validate_gd_parent(struct super_block *sb,
                                     struct ocfs2_dinode *di,
                                     struct buffer_head *bh,
-                                    int clean_error)
+                                    int resize)
 {
         unsigned int max_bits;
         struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
@@ -233,8 +226,11 @@ static int ocfs2_validate_gd_parent(struct super_block *sb,
                 return -EINVAL;
         }
 
-        if (le16_to_cpu(gd->bg_chain) >=
-            le16_to_cpu(di->id2.i_chain.cl_next_free_rec)) {
+        /* In resize, we may meet the case bg_chain == cl_next_free_rec. */
+        if ((le16_to_cpu(gd->bg_chain) >
+             le16_to_cpu(di->id2.i_chain.cl_next_free_rec)) ||
+            ((le16_to_cpu(gd->bg_chain) ==
+             le16_to_cpu(di->id2.i_chain.cl_next_free_rec)) && !resize)) {
                 do_error("Group descriptor #%llu has bad chain %u",
                          (unsigned long long)bh->b_blocknr,
                          le16_to_cpu(gd->bg_chain));
@@ -637,12 +633,113 @@ bail:
         return status;
 }
 
+static void ocfs2_init_inode_steal_slot(struct ocfs2_super *osb)
+{
+        spin_lock(&osb->osb_lock);
+        osb->s_inode_steal_slot = OCFS2_INVALID_SLOT;
+        spin_unlock(&osb->osb_lock);
+        atomic_set(&osb->s_num_inodes_stolen, 0);
+}
+
+static void ocfs2_init_meta_steal_slot(struct ocfs2_super *osb)
+{
+        spin_lock(&osb->osb_lock);
+        osb->s_meta_steal_slot = OCFS2_INVALID_SLOT;
+        spin_unlock(&osb->osb_lock);
+        atomic_set(&osb->s_num_meta_stolen, 0);
+}
+
+void ocfs2_init_steal_slots(struct ocfs2_super *osb)
+{
+        ocfs2_init_inode_steal_slot(osb);
+        ocfs2_init_meta_steal_slot(osb);
+}
+
+static void __ocfs2_set_steal_slot(struct ocfs2_super *osb, int slot, int type)
+{
+        spin_lock(&osb->osb_lock);
+        if (type == INODE_ALLOC_SYSTEM_INODE)
+                osb->s_inode_steal_slot = slot;
+        else if (type == EXTENT_ALLOC_SYSTEM_INODE)
+                osb->s_meta_steal_slot = slot;
+        spin_unlock(&osb->osb_lock);
+}
+
+static int __ocfs2_get_steal_slot(struct ocfs2_super *osb, int type)
+{
+        int slot = OCFS2_INVALID_SLOT;
+
+        spin_lock(&osb->osb_lock);
+        if (type == INODE_ALLOC_SYSTEM_INODE)
+                slot = osb->s_inode_steal_slot;
+        else if (type == EXTENT_ALLOC_SYSTEM_INODE)
+                slot = osb->s_meta_steal_slot;
+        spin_unlock(&osb->osb_lock);
+
+        return slot;
+}
+
+static int ocfs2_get_inode_steal_slot(struct ocfs2_super *osb)
+{
+        return __ocfs2_get_steal_slot(osb, INODE_ALLOC_SYSTEM_INODE);
+}
+
+static int ocfs2_get_meta_steal_slot(struct ocfs2_super *osb)
+{
+        return __ocfs2_get_steal_slot(osb, EXTENT_ALLOC_SYSTEM_INODE);
+}
+
+static int ocfs2_steal_resource(struct ocfs2_super *osb,
+                                struct ocfs2_alloc_context *ac,
+                                int type)
+{
+        int i, status = -ENOSPC;
+        int slot = __ocfs2_get_steal_slot(osb, type);
+
+        /* Start to steal resource from the first slot after ours. */
+        if (slot == OCFS2_INVALID_SLOT)
+                slot = osb->slot_num + 1;
+
+        for (i = 0; i < osb->max_slots; i++, slot++) {
+                if (slot == osb->max_slots)
+                        slot = 0;
+
+                if (slot == osb->slot_num)
+                        continue;
+
+                status = ocfs2_reserve_suballoc_bits(osb, ac,
+                                                     type,
+                                                     (u32)slot, NULL,
+                                                     NOT_ALLOC_NEW_GROUP);
+                if (status >= 0) {
+                        __ocfs2_set_steal_slot(osb, slot, type);
+                        break;
+                }
+
+                ocfs2_free_ac_resource(ac);
+        }
+
+        return status;
+}
+
+static int ocfs2_steal_inode(struct ocfs2_super *osb,
+                             struct ocfs2_alloc_context *ac)
+{
+        return ocfs2_steal_resource(osb, ac, INODE_ALLOC_SYSTEM_INODE);
+}
+
+static int ocfs2_steal_meta(struct ocfs2_super *osb,
+                            struct ocfs2_alloc_context *ac)
+{
+        return ocfs2_steal_resource(osb, ac, EXTENT_ALLOC_SYSTEM_INODE);
+}
+
 int ocfs2_reserve_new_metadata_blocks(struct ocfs2_super *osb,
                                       int blocks,
                                       struct ocfs2_alloc_context **ac)
 {
         int status;
-        u32 slot;
+        int slot = ocfs2_get_meta_steal_slot(osb);
 
         *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
         if (!(*ac)) {
@@ -653,12 +750,34 @@ int ocfs2_reserve_new_metadata_blocks(struct ocfs2_super *osb,
 
         (*ac)->ac_bits_wanted = blocks;
         (*ac)->ac_which = OCFS2_AC_USE_META;
-        slot = osb->slot_num;
         (*ac)->ac_group_search = ocfs2_block_group_search;
 
+        if (slot != OCFS2_INVALID_SLOT &&
+                atomic_read(&osb->s_num_meta_stolen) < OCFS2_MAX_TO_STEAL)
+                goto extent_steal;
+
+        atomic_set(&osb->s_num_meta_stolen, 0);
         status = ocfs2_reserve_suballoc_bits(osb, (*ac),
                                              EXTENT_ALLOC_SYSTEM_INODE,
-                                             slot, NULL, ALLOC_NEW_GROUP);
+                                             (u32)osb->slot_num, NULL,
+                                             ALLOC_NEW_GROUP);
+
+
+        if (status >= 0) {
+                status = 0;
+                if (slot != OCFS2_INVALID_SLOT)
+                        ocfs2_init_meta_steal_slot(osb);
+                goto bail;
+        } else if (status < 0 && status != -ENOSPC) {
+                mlog_errno(status);
+                goto bail;
+        }
+
+        ocfs2_free_ac_resource(*ac);
+
+extent_steal:
+        status = ocfs2_steal_meta(osb, *ac);
+        atomic_inc(&osb->s_num_meta_stolen);
         if (status < 0) {
                 if (status != -ENOSPC)
                         mlog_errno(status);
@@ -685,43 +804,11 @@ int ocfs2_reserve_new_metadata(struct ocfs2_super *osb,
                                         ac);
 }
 
-static int ocfs2_steal_inode_from_other_nodes(struct ocfs2_super *osb,
-                                              struct ocfs2_alloc_context *ac)
-{
-        int i, status = -ENOSPC;
-        s16 slot = ocfs2_get_inode_steal_slot(osb);
-
-        /* Start to steal inodes from the first slot after ours. */
-        if (slot == OCFS2_INVALID_SLOT)
-                slot = osb->slot_num + 1;
-
-        for (i = 0; i < osb->max_slots; i++, slot++) {
-                if (slot == osb->max_slots)
-                        slot = 0;
-
-                if (slot == osb->slot_num)
-                        continue;
-
-                status = ocfs2_reserve_suballoc_bits(osb, ac,
-                                                     INODE_ALLOC_SYSTEM_INODE,
-                                                     slot, NULL,
-                                                     NOT_ALLOC_NEW_GROUP);
-                if (status >= 0) {
-                        ocfs2_set_inode_steal_slot(osb, slot);
-                        break;
-                }
-
-                ocfs2_free_ac_resource(ac);
-        }
-
-        return status;
-}
-
 int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
                             struct ocfs2_alloc_context **ac)
 {
         int status;
-        s16 slot = ocfs2_get_inode_steal_slot(osb);
+        int slot = ocfs2_get_inode_steal_slot(osb);
         u64 alloc_group;
 
         *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
@@ -754,14 +841,14 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
          * need to check our slots to see whether there is some space for us.
          */
         if (slot != OCFS2_INVALID_SLOT &&
-            atomic_read(&osb->s_num_inodes_stolen) < OCFS2_MAX_INODES_TO_STEAL)
+            atomic_read(&osb->s_num_inodes_stolen) < OCFS2_MAX_TO_STEAL)
                 goto inode_steal;
 
         atomic_set(&osb->s_num_inodes_stolen, 0);
         alloc_group = osb->osb_inode_alloc_group;
         status = ocfs2_reserve_suballoc_bits(osb, *ac,
                                              INODE_ALLOC_SYSTEM_INODE,
-                                             osb->slot_num,
+                                             (u32)osb->slot_num,
                                              &alloc_group,
                                              ALLOC_NEW_GROUP |
                                              ALLOC_GROUPS_FROM_GLOBAL);
@@ -789,7 +876,7 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
         ocfs2_free_ac_resource(*ac);
 
 inode_steal:
-        status = ocfs2_steal_inode_from_other_nodes(osb, *ac);
+        status = ocfs2_steal_inode(osb, *ac);
         atomic_inc(&osb->s_num_inodes_stolen);
         if (status < 0) {
                 if (status != -ENOSPC)
@@ -1884,18 +1971,18 @@ int ocfs2_claim_clusters(struct ocfs2_super *osb,
                                       bits_wanted, cluster_start, num_clusters);
 }
 
-static inline int ocfs2_block_group_clear_bits(handle_t *handle,
-                                               struct inode *alloc_inode,
-                                               struct ocfs2_group_desc *bg,
-                                               struct buffer_head *group_bh,
-                                               unsigned int bit_off,
-                                               unsigned int num_bits)
+static int ocfs2_block_group_clear_bits(handle_t *handle,
+                                        struct inode *alloc_inode,
+                                        struct ocfs2_group_desc *bg,
+                                        struct buffer_head *group_bh,
+                                        unsigned int bit_off,
+                                        unsigned int num_bits,
+                                        void (*undo_fn)(unsigned int bit,
+                                                        unsigned long *bmap))
 {
         int status;
         unsigned int tmp;
-        int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
         struct ocfs2_group_desc *undo_bg = NULL;
-        int cluster_bitmap = 0;
 
         mlog_entry_void();
 
@@ -1905,20 +1992,18 @@ static inline int ocfs2_block_group_clear_bits(handle_t *handle,
 
         mlog(0, "off = %u, num = %u\n", bit_off, num_bits);
 
-        if (ocfs2_is_cluster_bitmap(alloc_inode))
-                journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
-
+        BUG_ON(undo_fn && !ocfs2_is_cluster_bitmap(alloc_inode));
         status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode),
-                                         group_bh, journal_type);
+                                         group_bh,
+                                         undo_fn ?
+                                         OCFS2_JOURNAL_ACCESS_UNDO :
+                                         OCFS2_JOURNAL_ACCESS_WRITE);
         if (status < 0) {
                 mlog_errno(status);
                 goto bail;
         }
 
-        if (ocfs2_is_cluster_bitmap(alloc_inode))
-                cluster_bitmap = 1;
-
-        if (cluster_bitmap) {
+        if (undo_fn) {
                 jbd_lock_bh_state(group_bh);
                 undo_bg = (struct ocfs2_group_desc *)
                                         bh2jh(group_bh)->b_committed_data;
@@ -1929,13 +2014,13 @@ static inline int ocfs2_block_group_clear_bits(handle_t *handle,
         while(tmp--) {
                 ocfs2_clear_bit((bit_off + tmp),
                                 (unsigned long *) bg->bg_bitmap);
-                if (cluster_bitmap)
-                        ocfs2_set_bit(bit_off + tmp,
-                                      (unsigned long *) undo_bg->bg_bitmap);
+                if (undo_fn)
+                        undo_fn(bit_off + tmp,
+                                (unsigned long *) undo_bg->bg_bitmap);
         }
         le16_add_cpu(&bg->bg_free_bits_count, num_bits);
 
-        if (cluster_bitmap)
+        if (undo_fn)
                 jbd_unlock_bh_state(group_bh);
 
         status = ocfs2_journal_dirty(handle, group_bh);
@@ -1948,12 +2033,14 @@ bail:
 /*
  * expects the suballoc inode to already be locked.
  */
-int ocfs2_free_suballoc_bits(handle_t *handle,
-                             struct inode *alloc_inode,
-                             struct buffer_head *alloc_bh,
-                             unsigned int start_bit,
-                             u64 bg_blkno,
-                             unsigned int count)
+static int _ocfs2_free_suballoc_bits(handle_t *handle,
+                                     struct inode *alloc_inode,
+                                     struct buffer_head *alloc_bh,
+                                     unsigned int start_bit,
+                                     u64 bg_blkno,
+                                     unsigned int count,
+                                     void (*undo_fn)(unsigned int bit,
+                                                     unsigned long *bitmap))
 {
         int status = 0;
         u32 tmp_used;
@@ -1988,7 +2075,7 @@ int ocfs2_free_suballoc_bits(handle_t *handle,
 
         status = ocfs2_block_group_clear_bits(handle, alloc_inode,
                                               group, group_bh,
-                                              start_bit, count);
+                                              start_bit, count, undo_fn);
         if (status < 0) {
                 mlog_errno(status);
                 goto bail;
@@ -2019,6 +2106,17 @@ bail:
         return status;
 }
 
+int ocfs2_free_suballoc_bits(handle_t *handle,
+                             struct inode *alloc_inode,
+                             struct buffer_head *alloc_bh,
+                             unsigned int start_bit,
+                             u64 bg_blkno,
+                             unsigned int count)
+{
+        return _ocfs2_free_suballoc_bits(handle, alloc_inode, alloc_bh,
+                                         start_bit, bg_blkno, count, NULL);
+}
+
 int ocfs2_free_dinode(handle_t *handle,
                       struct inode *inode_alloc_inode,
                       struct buffer_head *inode_alloc_bh,
@@ -2032,11 +2130,13 @@ int ocfs2_free_dinode(handle_t *handle,
                                         inode_alloc_bh, bit, bg_blkno, 1);
 }
 
-int ocfs2_free_clusters(handle_t *handle,
-                       struct inode *bitmap_inode,
-                       struct buffer_head *bitmap_bh,
-                       u64 start_blk,
-                       unsigned int num_clusters)
+static int _ocfs2_free_clusters(handle_t *handle,
+                                struct inode *bitmap_inode,
+                                struct buffer_head *bitmap_bh,
+                                u64 start_blk,
+                                unsigned int num_clusters,
+                                void (*undo_fn)(unsigned int bit,
+                                                unsigned long *bitmap))
 {
         int status;
         u16 bg_start_bit;
@@ -2063,9 +2163,9 @@ int ocfs2_free_clusters(handle_t *handle,
         mlog(0, "bg_blkno = %llu, bg_start_bit = %u\n",
              (unsigned long long)bg_blkno, bg_start_bit);
 
-        status = ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
-                                          bg_start_bit, bg_blkno,
-                                          num_clusters);
+        status = _ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
+                                           bg_start_bit, bg_blkno,
+                                           num_clusters, undo_fn);
         if (status < 0) {
                 mlog_errno(status);
                 goto out;
@@ -2079,6 +2179,32 @@ out:
         return status;
 }
 
+int ocfs2_free_clusters(handle_t *handle,
+                        struct inode *bitmap_inode,
+                        struct buffer_head *bitmap_bh,
+                        u64 start_blk,
+                        unsigned int num_clusters)
+{
+        return _ocfs2_free_clusters(handle, bitmap_inode, bitmap_bh,
+                                    start_blk, num_clusters,
+                                    _ocfs2_set_bit);
+}
+
+/*
+ * Give never-used clusters back to the global bitmap.  We don't need
+ * to protect these bits in the undo buffer.
+ */
+int ocfs2_release_clusters(handle_t *handle,
+                           struct inode *bitmap_inode,
+                           struct buffer_head *bitmap_bh,
+                           u64 start_blk,
+                           unsigned int num_clusters)
+{
+        return _ocfs2_free_clusters(handle, bitmap_inode, bitmap_bh,
+                                    start_blk, num_clusters,
+                                    _ocfs2_clear_bit);
+}
+
 static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg)
 {
         printk("Block Group:\n");
diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h
index 8c9a78a43164..e0f46df357e6 100644
--- a/fs/ocfs2/suballoc.h
+++ b/fs/ocfs2/suballoc.h
@@ -56,6 +56,7 @@ struct ocfs2_alloc_context {
                                  is the same as ~0 - unlimited */
 };
 
+void ocfs2_init_steal_slots(struct ocfs2_super *osb);
 void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac);
 static inline int ocfs2_alloc_context_bits_left(struct ocfs2_alloc_context *ac)
 {
@@ -126,6 +127,11 @@ int ocfs2_free_clusters(handle_t *handle,
                         struct buffer_head *bitmap_bh,
                         u64 start_blk,
                         unsigned int num_clusters);
+int ocfs2_release_clusters(handle_t *handle,
+                           struct inode *bitmap_inode,
+                           struct buffer_head *bitmap_bh,
+                           u64 start_blk,
+                           unsigned int num_clusters);
 
 static inline u64 ocfs2_which_suballoc_group(u64 block, unsigned int bit)
 {
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 14f47d2bfe02..dee03197a494 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -69,6 +69,7 @@
 #include "xattr.h"
 #include "quota.h"
 #include "refcounttree.h"
+#include "suballoc.h"
 
 #include "buffer_head_io.h"
 
@@ -100,6 +101,8 @@ struct mount_options
 static int ocfs2_parse_options(struct super_block *sb, char *options,
                                struct mount_options *mopt,
                                int is_remount);
+static int ocfs2_check_set_options(struct super_block *sb,
+                                   struct mount_options *options);
 static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt);
 static void ocfs2_put_super(struct super_block *sb);
 static int ocfs2_mount_volume(struct super_block *sb);
@@ -299,9 +302,12 @@ static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len)
 
         spin_lock(&osb->osb_lock);
         out += snprintf(buf + out, len - out,
-                        "%10s => Slot: %d  NumStolen: %d\n", "Steal",
+                        "%10s => InodeSlot: %d  StolenInodes: %d, "
+                        "MetaSlot: %d  StolenMeta: %d\n", "Steal",
                         osb->s_inode_steal_slot,
-                        atomic_read(&osb->s_num_inodes_stolen));
+                        atomic_read(&osb->s_num_inodes_stolen),
+                        osb->s_meta_steal_slot,
+                        atomic_read(&osb->s_num_meta_stolen));
         spin_unlock(&osb->osb_lock);
 
         out += snprintf(buf + out, len - out, "OrphanScan => ");
@@ -600,7 +606,8 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data)
 
         lock_kernel();
 
-        if (!ocfs2_parse_options(sb, data, &parsed_options, 1)) {
+        if (!ocfs2_parse_options(sb, data, &parsed_options, 1) ||
+            !ocfs2_check_set_options(sb, &parsed_options)) {
                 ret = -EINVAL;
                 goto out;
         }
@@ -691,8 +698,6 @@ unlock_osb:
         if (!ret) {
                 /* Only save off the new mount options in case of a successful
                  * remount. */
-                if (!(osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_XATTR))
-                        parsed_options.mount_opt &= ~OCFS2_MOUNT_POSIX_ACL;
                 osb->s_mount_opt = parsed_options.mount_opt;
                 osb->s_atime_quantum = parsed_options.atime_quantum;
                 osb->preferred_slot = parsed_options.slot;
@@ -701,6 +706,10 @@ unlock_osb:
 
                 if (!ocfs2_is_hard_readonly(osb))
                         ocfs2_set_journal_params(osb);
+
+                sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
+                        ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) ?
+                                                        MS_POSIXACL : 0);
         }
 out:
         unlock_kernel();
@@ -1011,31 +1020,16 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
         brelse(bh);
         bh = NULL;
 
-        if (!(osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_XATTR))
-                parsed_options.mount_opt &= ~OCFS2_MOUNT_POSIX_ACL;
-
+        if (!ocfs2_check_set_options(sb, &parsed_options)) {
+                status = -EINVAL;
+                goto read_super_error;
+        }
         osb->s_mount_opt = parsed_options.mount_opt;
         osb->s_atime_quantum = parsed_options.atime_quantum;
         osb->preferred_slot = parsed_options.slot;
         osb->osb_commit_interval = parsed_options.commit_interval;
         osb->local_alloc_default_bits = ocfs2_megabytes_to_clusters(sb, parsed_options.localalloc_opt);
         osb->local_alloc_bits = osb->local_alloc_default_bits;
-        if (osb->s_mount_opt & OCFS2_MOUNT_USRQUOTA &&
-            !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
-                                         OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
-                status = -EINVAL;
-                mlog(ML_ERROR, "User quotas were requested, but this "
-                     "filesystem does not have the feature enabled.\n");
-                goto read_super_error;
-        }
-        if (osb->s_mount_opt & OCFS2_MOUNT_GRPQUOTA &&
-            !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
-                                         OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
-                status = -EINVAL;
-                mlog(ML_ERROR, "Group quotas were requested, but this "
-                     "filesystem does not have the feature enabled.\n");
-                goto read_super_error;
-        }
 
         status = ocfs2_verify_userspace_stack(osb, &parsed_options);
         if (status)
@@ -1072,7 +1066,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
                                      "file system, but write access is "
                                      "unavailable.\n");
                         else
-                                mlog_errno(status);                     
+                                mlog_errno(status);
                         goto read_super_error;
                 }
 
@@ -1245,6 +1239,40 @@ static struct file_system_type ocfs2_fs_type = {
         .next           = NULL
 };
 
+static int ocfs2_check_set_options(struct super_block *sb,
+                                   struct mount_options *options)
+{
+        if (options->mount_opt & OCFS2_MOUNT_USRQUOTA &&
+            !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+                                         OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
+                mlog(ML_ERROR, "User quotas were requested, but this "
+                     "filesystem does not have the feature enabled.\n");
+                return 0;
+        }
+        if (options->mount_opt & OCFS2_MOUNT_GRPQUOTA &&
+            !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+                                         OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
+                mlog(ML_ERROR, "Group quotas were requested, but this "
+                     "filesystem does not have the feature enabled.\n");
+                return 0;
+        }
+        if (options->mount_opt & OCFS2_MOUNT_POSIX_ACL &&
+            !OCFS2_HAS_INCOMPAT_FEATURE(sb, OCFS2_FEATURE_INCOMPAT_XATTR)) {
+                mlog(ML_ERROR, "ACL support requested but extended attributes "
+                     "feature is not enabled\n");
+                return 0;
+        }
+        /* No ACL setting specified? Use XATTR feature... */
+        if (!(options->mount_opt & (OCFS2_MOUNT_POSIX_ACL |
+                                    OCFS2_MOUNT_NO_POSIX_ACL))) {
+                if (OCFS2_HAS_INCOMPAT_FEATURE(sb, OCFS2_FEATURE_INCOMPAT_XATTR))
+                        options->mount_opt |= OCFS2_MOUNT_POSIX_ACL;
+                else
+                        options->mount_opt |= OCFS2_MOUNT_NO_POSIX_ACL;
+        }
+        return 1;
+}
+
 static int ocfs2_parse_options(struct super_block *sb,
                                char *options,
                                struct mount_options *mopt,
@@ -1392,40 +1420,19 @@ static int ocfs2_parse_options(struct super_block *sb,
                         mopt->mount_opt |= OCFS2_MOUNT_INODE64;
                         break;
                 case Opt_usrquota:
-                        /* We check only on remount, otherwise features
-                         * aren't yet initialized. */
-                        if (is_remount && !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
-                            OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
-                                mlog(ML_ERROR, "User quota requested but "
-                                     "filesystem feature is not set\n");
-                                status = 0;
-                                goto bail;
-                        }
                         mopt->mount_opt |= OCFS2_MOUNT_USRQUOTA;
                         break;
                 case Opt_grpquota:
-                        if (is_remount && !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
-                            OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
-                                mlog(ML_ERROR, "Group quota requested but "
-                                     "filesystem feature is not set\n");
-                                status = 0;
-                                goto bail;
-                        }
                         mopt->mount_opt |= OCFS2_MOUNT_GRPQUOTA;
                         break;
-#ifdef CONFIG_OCFS2_FS_POSIX_ACL
                 case Opt_acl:
                         mopt->mount_opt |= OCFS2_MOUNT_POSIX_ACL;
+                        mopt->mount_opt &= ~OCFS2_MOUNT_NO_POSIX_ACL;
                         break;
                 case Opt_noacl:
+                        mopt->mount_opt |= OCFS2_MOUNT_NO_POSIX_ACL;
                         mopt->mount_opt &= ~OCFS2_MOUNT_POSIX_ACL;
                         break;
-#else
-                case Opt_acl:
-                case Opt_noacl:
-                        printk(KERN_INFO "ocfs2 (no)acl options not supported\n");
-                        break;
-#endif
                 default:
                         mlog(ML_ERROR,
                              "Unrecognized mount option \"%s\" "
@@ -1502,12 +1509,10 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
         if (opts & OCFS2_MOUNT_INODE64)
                 seq_printf(s, ",inode64");
 
-#ifdef CONFIG_OCFS2_FS_POSIX_ACL
         if (opts & OCFS2_MOUNT_POSIX_ACL)
                 seq_printf(s, ",acl");
         else
                 seq_printf(s, ",noacl");
-#endif
 
         return 0;
 }
@@ -1996,7 +2001,7 @@ static int ocfs2_initialize_super(struct super_block *sb,
         osb->blocked_lock_count = 0;
         spin_lock_init(&osb->osb_lock);
         spin_lock_init(&osb->osb_xattr_lock);
-        ocfs2_init_inode_steal_slot(osb);
+        ocfs2_init_steal_slots(osb);
 
         atomic_set(&osb->alloc_stats.moves, 0);
         atomic_set(&osb->alloc_stats.local_data, 0);
diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c
index e3421030a69f..32499d213fc4 100644
--- a/fs/ocfs2/symlink.c
+++ b/fs/ocfs2/symlink.c
@@ -137,20 +137,20 @@ static void *ocfs2_fast_follow_link(struct dentry *dentry,
         }
 
         memcpy(link, target, len);
-        nd_set_link(nd, link);
 
 bail:
+        nd_set_link(nd, status ? ERR_PTR(status) : link);
         brelse(bh);
 
         mlog_exit(status);
-        return status ? ERR_PTR(status) : link;
+        return NULL;
 }
 
 static void ocfs2_fast_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie)
 {
-        char *link = cookie;
-
-        kfree(link);
+        char *link = nd_get_link(nd);
+        if (!IS_ERR(link))
+                kfree(link);
 }
 
 const struct inode_operations ocfs2_symlink_inode_operations = {
@@ -163,6 +163,7 @@ const struct inode_operations ocfs2_symlink_inode_operations = {
         .getxattr       = generic_getxattr,
         .listxattr      = ocfs2_listxattr,
         .removexattr    = generic_removexattr,
+        .fiemap         = ocfs2_fiemap,
 };
 const struct inode_operations ocfs2_fast_symlink_inode_operations = {
         .readlink       = ocfs2_readlink,
@@ -174,4 +175,5 @@ const struct inode_operations ocfs2_fast_symlink_inode_operations = {
         .getxattr       = generic_getxattr,
         .listxattr      = ocfs2_listxattr,
         .removexattr    = generic_removexattr,
+        .fiemap         = ocfs2_fiemap,
 };
diff --git a/fs/ocfs2/sysfile.c b/fs/ocfs2/sysfile.c
index 40e53702948c..bfe7190cdbf1 100644
--- a/fs/ocfs2/sysfile.c
+++ b/fs/ocfs2/sysfile.c
@@ -25,7 +25,6 @@
 
 #include <linux/fs.h>
 #include <linux/types.h>
-#include <linux/slab.h>
 #include <linux/highmem.h>
 
 #define MLOG_MASK_PREFIX ML_INODE
diff --git a/fs/ocfs2/uptodate.c b/fs/ocfs2/uptodate.c
index c61369342a27..a0a120e82b97 100644
--- a/fs/ocfs2/uptodate.c
+++ b/fs/ocfs2/uptodate.c
@@ -267,8 +267,8 @@ static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
 }
 
 /* Warning: even if it returns true, this does *not* guarantee that
- * the block is stored in our inode metadata cache. 
- * 
+ * the block is stored in our inode metadata cache.
+ *
  * This can be called under lock_buffer()
  */
 int ocfs2_buffer_uptodate(struct ocfs2_caching_info *ci,
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index fe3419068df2..3e7773089b96 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -98,10 +98,8 @@ static struct ocfs2_xattr_def_value_root def_xv = {
 
 struct xattr_handler *ocfs2_xattr_handlers[] = {
         &ocfs2_xattr_user_handler,
-#ifdef CONFIG_OCFS2_FS_POSIX_ACL
         &ocfs2_xattr_acl_access_handler,
         &ocfs2_xattr_acl_default_handler,
-#endif
         &ocfs2_xattr_trusted_handler,
         &ocfs2_xattr_security_handler,
         NULL
@@ -109,21 +107,20 @@ struct xattr_handler *ocfs2_xattr_handlers[] = {
 
 static struct xattr_handler *ocfs2_xattr_handler_map[OCFS2_XATTR_MAX] = {
         [OCFS2_XATTR_INDEX_USER]        = &ocfs2_xattr_user_handler,
-#ifdef CONFIG_OCFS2_FS_POSIX_ACL
         [OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS]
                                         = &ocfs2_xattr_acl_access_handler,
         [OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT]
                                         = &ocfs2_xattr_acl_default_handler,
-#endif
         [OCFS2_XATTR_INDEX_TRUSTED]     = &ocfs2_xattr_trusted_handler,
         [OCFS2_XATTR_INDEX_SECURITY]    = &ocfs2_xattr_security_handler,
 };
 
 struct ocfs2_xattr_info {
-        int name_index;
-        const char *name;
-        const void *value;
-        size_t value_len;
+        int             xi_name_index;
+        const char      *xi_name;
+        int             xi_name_len;
+        const void      *xi_value;
+        size_t          xi_value_len;
 };
 
 struct ocfs2_xattr_search {
@@ -141,6 +138,115 @@ struct ocfs2_xattr_search {
         int not_found;
 };
 
+/* Operations on struct ocfs2_xa_entry */
+struct ocfs2_xa_loc;
+struct ocfs2_xa_loc_operations {
+        /*
+         * Journal functions
+         */
+        int (*xlo_journal_access)(handle_t *handle, struct ocfs2_xa_loc *loc,
+                                  int type);
+        void (*xlo_journal_dirty)(handle_t *handle, struct ocfs2_xa_loc *loc);
+
+        /*
+         * Return a pointer to the appropriate buffer in loc->xl_storage
+         * at the given offset from loc->xl_header.
+         */
+        void *(*xlo_offset_pointer)(struct ocfs2_xa_loc *loc, int offset);
+
+        /* Can we reuse the existing entry for the new value? */
+        int (*xlo_can_reuse)(struct ocfs2_xa_loc *loc,
+                             struct ocfs2_xattr_info *xi);
+
+        /* How much space is needed for the new value? */
+        int (*xlo_check_space)(struct ocfs2_xa_loc *loc,
+                               struct ocfs2_xattr_info *xi);
+
+        /*
+         * Return the offset of the first name+value pair.  This is
+         * the start of our downward-filling free space.
+         */
+        int (*xlo_get_free_start)(struct ocfs2_xa_loc *loc);
+
+        /*
+         * Remove the name+value at this location.  Do whatever is
+         * appropriate with the remaining name+value pairs.
+         */
+        void (*xlo_wipe_namevalue)(struct ocfs2_xa_loc *loc);
+
+        /* Fill xl_entry with a new entry */
+        void (*xlo_add_entry)(struct ocfs2_xa_loc *loc, u32 name_hash);
+
+        /* Add name+value storage to an entry */
+        void (*xlo_add_namevalue)(struct ocfs2_xa_loc *loc, int size);
+
+        /*
+         * Initialize the value buf's access and bh fields for this entry.
+         * ocfs2_xa_fill_value_buf() will handle the xv pointer.
+         */
+        void (*xlo_fill_value_buf)(struct ocfs2_xa_loc *loc,
+                                   struct ocfs2_xattr_value_buf *vb);
+};
+
+/*
+ * Describes an xattr entry location.  This is a memory structure
+ * tracking the on-disk structure.
+ */
+struct ocfs2_xa_loc {
+        /* This xattr belongs to this inode */
+        struct inode *xl_inode;
+
+        /* The ocfs2_xattr_header inside the on-disk storage. Not NULL. */
+        struct ocfs2_xattr_header *xl_header;
+
+        /* Bytes from xl_header to the end of the storage */
+        int xl_size;
+
+        /*
+         * The ocfs2_xattr_entry this location describes.  If this is
+         * NULL, this location describes the on-disk structure where it
+         * would have been.
+         */
+        struct ocfs2_xattr_entry *xl_entry;
+
+        /*
+         * Internal housekeeping
+         */
+
+        /* Buffer(s) containing this entry */
+        void *xl_storage;
+
+        /* Operations on the storage backing this location */
+        const struct ocfs2_xa_loc_operations *xl_ops;
+};
+
+/*
+ * Convenience functions to calculate how much space is needed for a
+ * given name+value pair
+ */
+static int namevalue_size(int name_len, uint64_t value_len)
+{
+        if (value_len > OCFS2_XATTR_INLINE_SIZE)
+                return OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
+        else
+                return OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_SIZE(value_len);
+}
+
+static int namevalue_size_xi(struct ocfs2_xattr_info *xi)
+{
+        return namevalue_size(xi->xi_name_len, xi->xi_value_len);
+}
+
+static int namevalue_size_xe(struct ocfs2_xattr_entry *xe)
+{
+        u64 value_len = le64_to_cpu(xe->xe_value_size);
+
+        BUG_ON((value_len > OCFS2_XATTR_INLINE_SIZE) &&
+               ocfs2_xattr_is_local(xe));
+        return namevalue_size(xe->xe_name_len, value_len);
+}
+
+
 static int ocfs2_xattr_bucket_get_name_value(struct super_block *sb,
                                              struct ocfs2_xattr_header *xh,
                                              int index,
@@ -205,8 +311,6 @@ static int ocfs2_get_xattr_tree_value_root(struct super_block *sb,
                                            int offset,
                                            struct ocfs2_xattr_value_root **xv,
                                            struct buffer_head **bh);
-static int ocfs2_xattr_security_set(struct inode *inode, const char *name,
-                                    const void *value, size_t size, int flags);
 
 static inline u16 ocfs2_xattr_buckets_per_cluster(struct ocfs2_super *osb)
 {
@@ -218,14 +322,6 @@ static inline u16 ocfs2_blocks_per_xattr_bucket(struct super_block *sb)
         return OCFS2_XATTR_BUCKET_SIZE / (1 << sb->s_blocksize_bits);
 }
 
-static inline u16 ocfs2_xattr_max_xe_in_bucket(struct super_block *sb)
-{
-        u16 len = sb->s_blocksize -
-                 offsetof(struct ocfs2_xattr_header, xh_entries);
-
-        return len / sizeof(struct ocfs2_xattr_entry);
-}
-
 #define bucket_blkno(_b) ((_b)->bu_bhs[0]->b_blocknr)
 #define bucket_block(_b, _n) ((_b)->bu_bhs[(_n)]->b_data)
 #define bucket_xh(_b) ((struct ocfs2_xattr_header *)bucket_block((_b), 0))
@@ -469,35 +565,22 @@ static u32 ocfs2_xattr_name_hash(struct inode *inode,
         return hash;
 }
 
-/*
- * ocfs2_xattr_hash_entry()
- *
- * Compute the hash of an extended attribute.
- */
-static void ocfs2_xattr_hash_entry(struct inode *inode,
-                                   struct ocfs2_xattr_header *header,
-                                   struct ocfs2_xattr_entry *entry)
+static int ocfs2_xattr_entry_real_size(int name_len, size_t value_len)
 {
-        u32 hash = 0;
-        char *name = (char *)header + le16_to_cpu(entry->xe_name_offset);
-
-        hash = ocfs2_xattr_name_hash(inode, name, entry->xe_name_len);
-        entry->xe_name_hash = cpu_to_le32(hash);
-
-        return;
+        return namevalue_size(name_len, value_len) +
+                sizeof(struct ocfs2_xattr_entry);
 }
 
-static int ocfs2_xattr_entry_real_size(int name_len, size_t value_len)
+static int ocfs2_xi_entry_usage(struct ocfs2_xattr_info *xi)
 {
-        int size = 0;
-
-        if (value_len <= OCFS2_XATTR_INLINE_SIZE)
-                size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_SIZE(value_len);
-        else
-                size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
-        size += sizeof(struct ocfs2_xattr_entry);
+        return namevalue_size_xi(xi) +
+                sizeof(struct ocfs2_xattr_entry);
+}
 
-        return size;
+static int ocfs2_xe_entry_usage(struct ocfs2_xattr_entry *xe)
+{
+        return namevalue_size_xe(xe) +
+                sizeof(struct ocfs2_xattr_entry);
 }
 
 int ocfs2_calc_security_init(struct inode *dir,
@@ -1314,452 +1397,897 @@ out:
         return ret;
 }
 
-static int ocfs2_xattr_cleanup(struct inode *inode,
-                               handle_t *handle,
-                               struct ocfs2_xattr_info *xi,
-                               struct ocfs2_xattr_search *xs,
-                               struct ocfs2_xattr_value_buf *vb,
-                               size_t offs)
+static int ocfs2_xa_check_space_helper(int needed_space, int free_start,
+                                       int num_entries)
 {
-        int ret = 0;
-        size_t name_len = strlen(xi->name);
-        void *val = xs->base + offs;
-        size_t size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
+        int free_space;
 
-        ret = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
-                            OCFS2_JOURNAL_ACCESS_WRITE);
-        if (ret) {
-                mlog_errno(ret);
-                goto out;
-        }
-        /* Decrease xattr count */
-        le16_add_cpu(&xs->header->xh_count, -1);
-        /* Remove the xattr entry and tree root which has already be set*/
-        memset((void *)xs->here, 0, sizeof(struct ocfs2_xattr_entry));
-        memset(val, 0, size);
+        if (!needed_space)
+                return 0;
 
-        ret = ocfs2_journal_dirty(handle, vb->vb_bh);
-        if (ret < 0)
-                mlog_errno(ret);
-out:
-        return ret;
+        free_space = free_start -
+                sizeof(struct ocfs2_xattr_header) -
+                (num_entries * sizeof(struct ocfs2_xattr_entry)) -
+                OCFS2_XATTR_HEADER_GAP;
+        if (free_space < 0)
+                return -EIO;
+        if (free_space < needed_space)
+                return -ENOSPC;
+
+        return 0;
 }
 
-static int ocfs2_xattr_update_entry(struct inode *inode,
-                                    handle_t *handle,
-                                    struct ocfs2_xattr_info *xi,
-                                    struct ocfs2_xattr_search *xs,
-                                    struct ocfs2_xattr_value_buf *vb,
-                                    size_t offs)
+static int ocfs2_xa_journal_access(handle_t *handle, struct ocfs2_xa_loc *loc,
+                                   int type)
 {
-        int ret;
+        return loc->xl_ops->xlo_journal_access(handle, loc, type);
+}
 
-        ret = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
-                            OCFS2_JOURNAL_ACCESS_WRITE);
-        if (ret) {
-                mlog_errno(ret);
-                goto out;
-        }
+static void ocfs2_xa_journal_dirty(handle_t *handle, struct ocfs2_xa_loc *loc)
+{
+        loc->xl_ops->xlo_journal_dirty(handle, loc);
+}
 
-        xs->here->xe_name_offset = cpu_to_le16(offs);
-        xs->here->xe_value_size = cpu_to_le64(xi->value_len);
-        if (xi->value_len <= OCFS2_XATTR_INLINE_SIZE)
-                ocfs2_xattr_set_local(xs->here, 1);
-        else
-                ocfs2_xattr_set_local(xs->here, 0);
-        ocfs2_xattr_hash_entry(inode, xs->header, xs->here);
+/* Give a pointer into the storage for the given offset */
+static void *ocfs2_xa_offset_pointer(struct ocfs2_xa_loc *loc, int offset)
+{
+        BUG_ON(offset >= loc->xl_size);
+        return loc->xl_ops->xlo_offset_pointer(loc, offset);
+}
 
-        ret = ocfs2_journal_dirty(handle, vb->vb_bh);
-        if (ret < 0)
-                mlog_errno(ret);
-out:
-        return ret;
+/*
+ * Wipe the name+value pair and allow the storage to reclaim it.  This
+ * must be followed by either removal of the entry or a call to
+ * ocfs2_xa_add_namevalue().
+ */
+static void ocfs2_xa_wipe_namevalue(struct ocfs2_xa_loc *loc)
+{
+        loc->xl_ops->xlo_wipe_namevalue(loc);
 }
 
 /*
- * ocfs2_xattr_set_value_outside()
- *
- * Set large size value in B tree.
+ * Find lowest offset to a name+value pair.  This is the start of our
+ * downward-growing free space.
  */
-static int ocfs2_xattr_set_value_outside(struct inode *inode,
-                                         struct ocfs2_xattr_info *xi,
-                                         struct ocfs2_xattr_search *xs,
-                                         struct ocfs2_xattr_set_ctxt *ctxt,
-                                         struct ocfs2_xattr_value_buf *vb,
-                                         size_t offs)
+static int ocfs2_xa_get_free_start(struct ocfs2_xa_loc *loc)
 {
-        size_t name_len = strlen(xi->name);
-        void *val = xs->base + offs;
-        struct ocfs2_xattr_value_root *xv = NULL;
-        size_t size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
-        int ret = 0;
+        return loc->xl_ops->xlo_get_free_start(loc);
+}
 
-        memset(val, 0, size);
-        memcpy(val, xi->name, name_len);
-        xv = (struct ocfs2_xattr_value_root *)
-                (val + OCFS2_XATTR_SIZE(name_len));
-        xv->xr_clusters = 0;
-        xv->xr_last_eb_blk = 0;
-        xv->xr_list.l_tree_depth = 0;
-        xv->xr_list.l_count = cpu_to_le16(1);
-        xv->xr_list.l_next_free_rec = 0;
-        vb->vb_xv = xv;
-
-        ret = ocfs2_xattr_value_truncate(inode, vb, xi->value_len, ctxt);
-        if (ret < 0) {
-                mlog_errno(ret);
-                return ret;
+/* Can we reuse loc->xl_entry for xi? */
+static int ocfs2_xa_can_reuse_entry(struct ocfs2_xa_loc *loc,
+                                    struct ocfs2_xattr_info *xi)
+{
+        return loc->xl_ops->xlo_can_reuse(loc, xi);
+}
+
+/* How much free space is needed to set the new value */
+static int ocfs2_xa_check_space(struct ocfs2_xa_loc *loc,
+                                struct ocfs2_xattr_info *xi)
+{
+        return loc->xl_ops->xlo_check_space(loc, xi);
+}
+
+static void ocfs2_xa_add_entry(struct ocfs2_xa_loc *loc, u32 name_hash)
+{
+        loc->xl_ops->xlo_add_entry(loc, name_hash);
+        loc->xl_entry->xe_name_hash = cpu_to_le32(name_hash);
+        /*
+         * We can't leave the new entry's xe_name_offset at zero or
+         * add_namevalue() will go nuts.  We set it to the size of our
+         * storage so that it can never be less than any other entry.
+         */
+        loc->xl_entry->xe_name_offset = cpu_to_le16(loc->xl_size);
+}
+
+static void ocfs2_xa_add_namevalue(struct ocfs2_xa_loc *loc,
+                                   struct ocfs2_xattr_info *xi)
+{
+        int size = namevalue_size_xi(xi);
+        int nameval_offset;
+        char *nameval_buf;
+
+        loc->xl_ops->xlo_add_namevalue(loc, size);
+        loc->xl_entry->xe_value_size = cpu_to_le64(xi->xi_value_len);
+        loc->xl_entry->xe_name_len = xi->xi_name_len;
+        ocfs2_xattr_set_type(loc->xl_entry, xi->xi_name_index);
+        ocfs2_xattr_set_local(loc->xl_entry,
+                              xi->xi_value_len <= OCFS2_XATTR_INLINE_SIZE);
+
+        nameval_offset = le16_to_cpu(loc->xl_entry->xe_name_offset);
+        nameval_buf = ocfs2_xa_offset_pointer(loc, nameval_offset);
+        memset(nameval_buf, 0, size);
+        memcpy(nameval_buf, xi->xi_name, xi->xi_name_len);
+}
+
+static void ocfs2_xa_fill_value_buf(struct ocfs2_xa_loc *loc,
+                                    struct ocfs2_xattr_value_buf *vb)
+{
+        int nameval_offset = le16_to_cpu(loc->xl_entry->xe_name_offset);
+        int name_size = OCFS2_XATTR_SIZE(loc->xl_entry->xe_name_len);
+
+        /* Value bufs are for value trees */
+        BUG_ON(ocfs2_xattr_is_local(loc->xl_entry));
+        BUG_ON(namevalue_size_xe(loc->xl_entry) !=
+               (name_size + OCFS2_XATTR_ROOT_SIZE));
+
+        loc->xl_ops->xlo_fill_value_buf(loc, vb);
+        vb->vb_xv =
+                (struct ocfs2_xattr_value_root *)ocfs2_xa_offset_pointer(loc,
+                                                        nameval_offset +
+                                                        name_size);
+}
+
+static int ocfs2_xa_block_journal_access(handle_t *handle,
+                                         struct ocfs2_xa_loc *loc, int type)
+{
+        struct buffer_head *bh = loc->xl_storage;
+        ocfs2_journal_access_func access;
+
+        if (loc->xl_size == (bh->b_size -
+                             offsetof(struct ocfs2_xattr_block,
+                                      xb_attrs.xb_header)))
+                access = ocfs2_journal_access_xb;
+        else
+                access = ocfs2_journal_access_di;
+        return access(handle, INODE_CACHE(loc->xl_inode), bh, type);
+}
+
+static void ocfs2_xa_block_journal_dirty(handle_t *handle,
+                                         struct ocfs2_xa_loc *loc)
+{
+        struct buffer_head *bh = loc->xl_storage;
+
+        ocfs2_journal_dirty(handle, bh);
+}
+
+static void *ocfs2_xa_block_offset_pointer(struct ocfs2_xa_loc *loc,
+                                           int offset)
+{
+        return (char *)loc->xl_header + offset;
+}
+
+static int ocfs2_xa_block_can_reuse(struct ocfs2_xa_loc *loc,
+                                    struct ocfs2_xattr_info *xi)
+{
+        /*
+         * Block storage is strict.  If the sizes aren't exact, we will
+         * remove the old one and reinsert the new.
+         */
+        return namevalue_size_xe(loc->xl_entry) ==
+                namevalue_size_xi(xi);
+}
+
+static int ocfs2_xa_block_get_free_start(struct ocfs2_xa_loc *loc)
+{
+        struct ocfs2_xattr_header *xh = loc->xl_header;
+        int i, count = le16_to_cpu(xh->xh_count);
+        int offset, free_start = loc->xl_size;
+
+        for (i = 0; i < count; i++) {
+                offset = le16_to_cpu(xh->xh_entries[i].xe_name_offset);
+                if (offset < free_start)
+                        free_start = offset;
         }
-        ret = ocfs2_xattr_update_entry(inode, ctxt->handle, xi, xs, vb, offs);
-        if (ret < 0) {
-                mlog_errno(ret);
-                return ret;
+
+        return free_start;
+}
+
+static int ocfs2_xa_block_check_space(struct ocfs2_xa_loc *loc,
+                                      struct ocfs2_xattr_info *xi)
+{
+        int count = le16_to_cpu(loc->xl_header->xh_count);
+        int free_start = ocfs2_xa_get_free_start(loc);
+        int needed_space = ocfs2_xi_entry_usage(xi);
+
+        /*
+         * Block storage will reclaim the original entry before inserting
+         * the new value, so we only need the difference.  If the new
+         * entry is smaller than the old one, we don't need anything.
+         */
+        if (loc->xl_entry) {
+                /* Don't need space if we're reusing! */
+                if (ocfs2_xa_can_reuse_entry(loc, xi))
+                        needed_space = 0;
+                else
+                        needed_space -= ocfs2_xe_entry_usage(loc->xl_entry);
         }
-        ret = __ocfs2_xattr_set_value_outside(inode, ctxt->handle, vb,
-                                              xi->value, xi->value_len);
-        if (ret < 0)
-                mlog_errno(ret);
+        if (needed_space < 0)
+                needed_space = 0;
+        return ocfs2_xa_check_space_helper(needed_space, free_start, count);
+}
 
-        return ret;
+/*
+ * Block storage for xattrs keeps the name+value pairs compacted.  When
+ * we remove one, we have to shift any that preceded it towards the end.
+ */
+static void ocfs2_xa_block_wipe_namevalue(struct ocfs2_xa_loc *loc)
+{
+        int i, offset;
+        int namevalue_offset, first_namevalue_offset, namevalue_size;
+        struct ocfs2_xattr_entry *entry = loc->xl_entry;
+        struct ocfs2_xattr_header *xh = loc->xl_header;
+        int count = le16_to_cpu(xh->xh_count);
+
+        namevalue_offset = le16_to_cpu(entry->xe_name_offset);
+        namevalue_size = namevalue_size_xe(entry);
+        first_namevalue_offset = ocfs2_xa_get_free_start(loc);
+
+        /* Shift the name+value pairs */
+        memmove((char *)xh + first_namevalue_offset + namevalue_size,
+                (char *)xh + first_namevalue_offset,
+                namevalue_offset - first_namevalue_offset);
+        memset((char *)xh + first_namevalue_offset, 0, namevalue_size);
+
+        /* Now tell xh->xh_entries about it */
+        for (i = 0; i < count; i++) {
+                offset = le16_to_cpu(xh->xh_entries[i].xe_name_offset);
+                if (offset <= namevalue_offset)
+                        le16_add_cpu(&xh->xh_entries[i].xe_name_offset,
+                                     namevalue_size);
+        }
+
+        /*
+         * Note that we don't update xh_free_start or xh_name_value_len
+         * because they're not used in block-stored xattrs.
+         */
+}
+
+static void ocfs2_xa_block_add_entry(struct ocfs2_xa_loc *loc, u32 name_hash)
+{
+        int count = le16_to_cpu(loc->xl_header->xh_count);
+        loc->xl_entry = &(loc->xl_header->xh_entries[count]);
+        le16_add_cpu(&loc->xl_header->xh_count, 1);
+        memset(loc->xl_entry, 0, sizeof(struct ocfs2_xattr_entry));
+}
+
+static void ocfs2_xa_block_add_namevalue(struct ocfs2_xa_loc *loc, int size)
+{
+        int free_start = ocfs2_xa_get_free_start(loc);
+
+        loc->xl_entry->xe_name_offset = cpu_to_le16(free_start - size);
+}
+
+static void ocfs2_xa_block_fill_value_buf(struct ocfs2_xa_loc *loc,
+                                          struct ocfs2_xattr_value_buf *vb)
+{
+        struct buffer_head *bh = loc->xl_storage;
+
+        if (loc->xl_size == (bh->b_size -
+                             offsetof(struct ocfs2_xattr_block,
+                                      xb_attrs.xb_header)))
+                vb->vb_access = ocfs2_journal_access_xb;
+        else
+                vb->vb_access = ocfs2_journal_access_di;
+        vb->vb_bh = bh;
 }
 
 /*
- * ocfs2_xattr_set_entry_local()
- *
- * Set, replace or remove extended attribute in local.
+ * Operations for xattrs stored in blocks.  This includes inline inode
+ * storage and unindexed ocfs2_xattr_blocks.
  */
-static void ocfs2_xattr_set_entry_local(struct inode *inode,
-                                        struct ocfs2_xattr_info *xi,
-                                        struct ocfs2_xattr_search *xs,
-                                        struct ocfs2_xattr_entry *last,
-                                        size_t min_offs)
+static const struct ocfs2_xa_loc_operations ocfs2_xa_block_loc_ops = {
+        .xlo_journal_access     = ocfs2_xa_block_journal_access,
+        .xlo_journal_dirty      = ocfs2_xa_block_journal_dirty,
+        .xlo_offset_pointer     = ocfs2_xa_block_offset_pointer,
+        .xlo_check_space        = ocfs2_xa_block_check_space,
+        .xlo_can_reuse          = ocfs2_xa_block_can_reuse,
+        .xlo_get_free_start     = ocfs2_xa_block_get_free_start,
+        .xlo_wipe_namevalue     = ocfs2_xa_block_wipe_namevalue,
+        .xlo_add_entry          = ocfs2_xa_block_add_entry,
+        .xlo_add_namevalue      = ocfs2_xa_block_add_namevalue,
+        .xlo_fill_value_buf     = ocfs2_xa_block_fill_value_buf,
+};
+
+static int ocfs2_xa_bucket_journal_access(handle_t *handle,
+                                          struct ocfs2_xa_loc *loc, int type)
 {
-        size_t name_len = strlen(xi->name);
-        int i;
+        struct ocfs2_xattr_bucket *bucket = loc->xl_storage;
 
-        if (xi->value && xs->not_found) {
-                /* Insert the new xattr entry. */
-                le16_add_cpu(&xs->header->xh_count, 1);
-                ocfs2_xattr_set_type(last, xi->name_index);
-                ocfs2_xattr_set_local(last, 1);
-                last->xe_name_len = name_len;
-        } else {
-                void *first_val;
-                void *val;
-                size_t offs, size;
-
-                first_val = xs->base + min_offs;
-                offs = le16_to_cpu(xs->here->xe_name_offset);
-                val = xs->base + offs;
-
-                if (le64_to_cpu(xs->here->xe_value_size) >
-                    OCFS2_XATTR_INLINE_SIZE)
-                        size = OCFS2_XATTR_SIZE(name_len) +
-                                OCFS2_XATTR_ROOT_SIZE;
+        return ocfs2_xattr_bucket_journal_access(handle, bucket, type);
+}
+
+static void ocfs2_xa_bucket_journal_dirty(handle_t *handle,
+                                          struct ocfs2_xa_loc *loc)
+{
+        struct ocfs2_xattr_bucket *bucket = loc->xl_storage;
+
+        ocfs2_xattr_bucket_journal_dirty(handle, bucket);
+}
+
+static void *ocfs2_xa_bucket_offset_pointer(struct ocfs2_xa_loc *loc,
+                                            int offset)
+{
+        struct ocfs2_xattr_bucket *bucket = loc->xl_storage;
+        int block, block_offset;
+
+        /* The header is at the front of the bucket */
+        block = offset >> loc->xl_inode->i_sb->s_blocksize_bits;
+        block_offset = offset % loc->xl_inode->i_sb->s_blocksize;
+
+        return bucket_block(bucket, block) + block_offset;
+}
+
+static int ocfs2_xa_bucket_can_reuse(struct ocfs2_xa_loc *loc,
+                                     struct ocfs2_xattr_info *xi)
+{
+        return namevalue_size_xe(loc->xl_entry) >=
+                namevalue_size_xi(xi);
+}
+
+static int ocfs2_xa_bucket_get_free_start(struct ocfs2_xa_loc *loc)
+{
+        struct ocfs2_xattr_bucket *bucket = loc->xl_storage;
+        return le16_to_cpu(bucket_xh(bucket)->xh_free_start);
+}
+
+static int ocfs2_bucket_align_free_start(struct super_block *sb,
+                                         int free_start, int size)
+{
+        /*
+         * We need to make sure that the name+value pair fits within
+         * one block.
+         */
+        if (((free_start - size) >> sb->s_blocksize_bits) !=
+            ((free_start - 1) >> sb->s_blocksize_bits))
+                free_start -= free_start % sb->s_blocksize;
+
+        return free_start;
+}
+
+static int ocfs2_xa_bucket_check_space(struct ocfs2_xa_loc *loc,
+                                       struct ocfs2_xattr_info *xi)
+{
+        int rc;
+        int count = le16_to_cpu(loc->xl_header->xh_count);
+        int free_start = ocfs2_xa_get_free_start(loc);
+        int needed_space = ocfs2_xi_entry_usage(xi);
+        int size = namevalue_size_xi(xi);
+        struct super_block *sb = loc->xl_inode->i_sb;
+
+        /*
+         * Bucket storage does not reclaim name+value pairs it cannot
+         * reuse.  They live as holes until the bucket fills, and then
+         * the bucket is defragmented.  However, the bucket can reclaim
+         * the ocfs2_xattr_entry.
+         */
+        if (loc->xl_entry) {
+                /* Don't need space if we're reusing! */
+                if (ocfs2_xa_can_reuse_entry(loc, xi))
+                        needed_space = 0;
                 else
-                        size = OCFS2_XATTR_SIZE(name_len) +
-                        OCFS2_XATTR_SIZE(le64_to_cpu(xs->here->xe_value_size));
-
-                if (xi->value && size == OCFS2_XATTR_SIZE(name_len) +
-                                OCFS2_XATTR_SIZE(xi->value_len)) {
-                        /* The old and the new value have the
-                           same size. Just replace the value. */
-                        ocfs2_xattr_set_local(xs->here, 1);
-                        xs->here->xe_value_size = cpu_to_le64(xi->value_len);
-                        /* Clear value bytes. */
-                        memset(val + OCFS2_XATTR_SIZE(name_len),
-                               0,
-                               OCFS2_XATTR_SIZE(xi->value_len));
-                        memcpy(val + OCFS2_XATTR_SIZE(name_len),
-                               xi->value,
-                               xi->value_len);
-                        return;
-                }
-                /* Remove the old name+value. */
-                memmove(first_val + size, first_val, val - first_val);
-                memset(first_val, 0, size);
-                xs->here->xe_name_hash = 0;
-                xs->here->xe_name_offset = 0;
-                ocfs2_xattr_set_local(xs->here, 1);
-                xs->here->xe_value_size = 0;
-
-                min_offs += size;
-
-                /* Adjust all value offsets. */
-                last = xs->header->xh_entries;
-                for (i = 0 ; i < le16_to_cpu(xs->header->xh_count); i++) {
-                        size_t o = le16_to_cpu(last->xe_name_offset);
-
-                        if (o < offs)
-                                last->xe_name_offset = cpu_to_le16(o + size);
-                        last += 1;
-                }
+                        needed_space -= sizeof(struct ocfs2_xattr_entry);
+        }
+        BUG_ON(needed_space < 0);
 
-                if (!xi->value) {
-                        /* Remove the old entry. */
-                        last -= 1;
-                        memmove(xs->here, xs->here + 1,
-                                (void *)last - (void *)xs->here);
-                        memset(last, 0, sizeof(struct ocfs2_xattr_entry));
-                        le16_add_cpu(&xs->header->xh_count, -1);
-                }
+        if (free_start < size) {
+                if (needed_space)
+                        return -ENOSPC;
+        } else {
+                /*
+                 * First we check if it would fit in the first place.
+                 * Below, we align the free start to a block.  This may
+                 * slide us below the minimum gap.  By checking unaligned
+                 * first, we avoid that error.
+                 */
+                rc = ocfs2_xa_check_space_helper(needed_space, free_start,
+                                                 count);
+                if (rc)
+                        return rc;
+                free_start = ocfs2_bucket_align_free_start(sb, free_start,
+                                                           size);
         }
-        if (xi->value) {
-                /* Insert the new name+value. */
-                size_t size = OCFS2_XATTR_SIZE(name_len) +
-                                OCFS2_XATTR_SIZE(xi->value_len);
-                void *val = xs->base + min_offs - size;
+        return ocfs2_xa_check_space_helper(needed_space, free_start, count);
+}
+
+static void ocfs2_xa_bucket_wipe_namevalue(struct ocfs2_xa_loc *loc)
+{
+        le16_add_cpu(&loc->xl_header->xh_name_value_len,
+                     -namevalue_size_xe(loc->xl_entry));
+}
+
+static void ocfs2_xa_bucket_add_entry(struct ocfs2_xa_loc *loc, u32 name_hash)
+{
+        struct ocfs2_xattr_header *xh = loc->xl_header;
+        int count = le16_to_cpu(xh->xh_count);
+        int low = 0, high = count - 1, tmp;
+        struct ocfs2_xattr_entry *tmp_xe;
 
-                xs->here->xe_name_offset = cpu_to_le16(min_offs - size);
-                memset(val, 0, size);
-                memcpy(val, xi->name, name_len);
-                memcpy(val + OCFS2_XATTR_SIZE(name_len),
-                       xi->value,
-                       xi->value_len);
-                xs->here->xe_value_size = cpu_to_le64(xi->value_len);
-                ocfs2_xattr_set_local(xs->here, 1);
-                ocfs2_xattr_hash_entry(inode, xs->header, xs->here);
+        /*
+         * We keep buckets sorted by name_hash, so we need to find
+         * our insert place.
+         */
+        while (low <= high && count) {
+                tmp = (low + high) / 2;
+                tmp_xe = &xh->xh_entries[tmp];
+
+                if (name_hash > le32_to_cpu(tmp_xe->xe_name_hash))
+                        low = tmp + 1;
+                else if (name_hash < le32_to_cpu(tmp_xe->xe_name_hash))
+                        high = tmp - 1;
+                else {
+                        low = tmp;
+                        break;
+                }
         }
 
-        return;
+        if (low != count)
+                memmove(&xh->xh_entries[low + 1],
+                        &xh->xh_entries[low],
+                        ((count - low) * sizeof(struct ocfs2_xattr_entry)));
+
+        le16_add_cpu(&xh->xh_count, 1);
+        loc->xl_entry = &xh->xh_entries[low];
+        memset(loc->xl_entry, 0, sizeof(struct ocfs2_xattr_entry));
+}
+
+static void ocfs2_xa_bucket_add_namevalue(struct ocfs2_xa_loc *loc, int size)
+{
+        int free_start = ocfs2_xa_get_free_start(loc);
+        struct ocfs2_xattr_header *xh = loc->xl_header;
+        struct super_block *sb = loc->xl_inode->i_sb;
+        int nameval_offset;
+
+        free_start = ocfs2_bucket_align_free_start(sb, free_start, size);
+        nameval_offset = free_start - size;
+        loc->xl_entry->xe_name_offset = cpu_to_le16(nameval_offset);
+        xh->xh_free_start = cpu_to_le16(nameval_offset);
+        le16_add_cpu(&xh->xh_name_value_len, size);
+
+}
+
+static void ocfs2_xa_bucket_fill_value_buf(struct ocfs2_xa_loc *loc,
+                                           struct ocfs2_xattr_value_buf *vb)
+{
+        struct ocfs2_xattr_bucket *bucket = loc->xl_storage;
+        struct super_block *sb = loc->xl_inode->i_sb;
+        int nameval_offset = le16_to_cpu(loc->xl_entry->xe_name_offset);
+        int size = namevalue_size_xe(loc->xl_entry);
+        int block_offset = nameval_offset >> sb->s_blocksize_bits;
+
+        /* Values are not allowed to straddle block boundaries */
+        BUG_ON(block_offset !=
+               ((nameval_offset + size - 1) >> sb->s_blocksize_bits));
+        /* We expect the bucket to be filled in */
+        BUG_ON(!bucket->bu_bhs[block_offset]);
+
+        vb->vb_access = ocfs2_journal_access;
+        vb->vb_bh = bucket->bu_bhs[block_offset];
+}
+
+/* Operations for xattrs stored in buckets. */
+static const struct ocfs2_xa_loc_operations ocfs2_xa_bucket_loc_ops = {
+        .xlo_journal_access     = ocfs2_xa_bucket_journal_access,
+        .xlo_journal_dirty      = ocfs2_xa_bucket_journal_dirty,
+        .xlo_offset_pointer     = ocfs2_xa_bucket_offset_pointer,
+        .xlo_check_space        = ocfs2_xa_bucket_check_space,
+        .xlo_can_reuse          = ocfs2_xa_bucket_can_reuse,
+        .xlo_get_free_start     = ocfs2_xa_bucket_get_free_start,
+        .xlo_wipe_namevalue     = ocfs2_xa_bucket_wipe_namevalue,
+        .xlo_add_entry          = ocfs2_xa_bucket_add_entry,
+        .xlo_add_namevalue      = ocfs2_xa_bucket_add_namevalue,
+        .xlo_fill_value_buf     = ocfs2_xa_bucket_fill_value_buf,
+};
+
+static unsigned int ocfs2_xa_value_clusters(struct ocfs2_xa_loc *loc)
+{
+        struct ocfs2_xattr_value_buf vb;
+
+        if (ocfs2_xattr_is_local(loc->xl_entry))
+                return 0;
+
+        ocfs2_xa_fill_value_buf(loc, &vb);
+        return le32_to_cpu(vb.vb_xv->xr_clusters);
+}
+
+static int ocfs2_xa_value_truncate(struct ocfs2_xa_loc *loc, u64 bytes,
+                                   struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int trunc_rc, access_rc;
+        struct ocfs2_xattr_value_buf vb;
+
+        ocfs2_xa_fill_value_buf(loc, &vb);
+        trunc_rc = ocfs2_xattr_value_truncate(loc->xl_inode, &vb, bytes,
+                                              ctxt);
+
+        /*
+         * The caller of ocfs2_xa_value_truncate() has already called
+         * ocfs2_xa_journal_access on the loc.  However, The truncate code
+         * calls ocfs2_extend_trans().  This may commit the previous
+         * transaction and open a new one.  If this is a bucket, truncate
+         * could leave only vb->vb_bh set up for journaling.  Meanwhile,
+         * the caller is expecting to dirty the entire bucket.  So we must
+         * reset the journal work.  We do this even if truncate has failed,
+         * as it could have failed after committing the extend.
+         */
+        access_rc = ocfs2_xa_journal_access(ctxt->handle, loc,
+                                            OCFS2_JOURNAL_ACCESS_WRITE);
+
+        /* Errors in truncate take precedence */
+        return trunc_rc ? trunc_rc : access_rc;
+}
+
+static void ocfs2_xa_remove_entry(struct ocfs2_xa_loc *loc)
+{
+        int index, count;
+        struct ocfs2_xattr_header *xh = loc->xl_header;
+        struct ocfs2_xattr_entry *entry = loc->xl_entry;
+
+        ocfs2_xa_wipe_namevalue(loc);
+        loc->xl_entry = NULL;
+
+        le16_add_cpu(&xh->xh_count, -1);
+        count = le16_to_cpu(xh->xh_count);
+
+        /*
+         * Only zero out the entry if there are more remaining.  This is
+         * important for an empty bucket, as it keeps track of the
+         * bucket's hash value.  It doesn't hurt empty block storage.
+         */
+        if (count) {
+                index = ((char *)entry - (char *)&xh->xh_entries) /
+                        sizeof(struct ocfs2_xattr_entry);
+                memmove(&xh->xh_entries[index], &xh->xh_entries[index + 1],
+                        (count - index) * sizeof(struct ocfs2_xattr_entry));
+                memset(&xh->xh_entries[count], 0,
+                       sizeof(struct ocfs2_xattr_entry));
+        }
 }
 
 /*
- * ocfs2_xattr_set_entry()
+ * If we have a problem adjusting the size of an external value during
+ * ocfs2_xa_prepare_entry() or ocfs2_xa_remove(), we may have an xattr
+ * in an intermediate state.  For example, the value may be partially
+ * truncated.
  *
- * Set extended attribute entry into inode or block.
+ * If the value tree hasn't changed, the extend/truncate went nowhere.
+ * We have nothing to do.  The caller can treat it as a straight error.
  *
- * If extended attribute value size > OCFS2_XATTR_INLINE_SIZE,
- * We first insert tree root(ocfs2_xattr_value_root) with set_entry_local(),
- * then set value in B tree with set_value_outside().
+ * If the value tree got partially truncated, we now have a corrupted
+ * extended attribute.  We're going to wipe its entry and leak the
+ * clusters.  Better to leak some storage than leave a corrupt entry.
+ *
+ * If the value tree grew, it obviously didn't grow enough for the
+ * new entry.  We're not going to try and reclaim those clusters either.
+ * If there was already an external value there (orig_clusters != 0),
+ * the new clusters are attached safely and we can just leave the old
+ * value in place.  If there was no external value there, we remove
+ * the entry.
+ *
+ * This way, the xattr block we store in the journal will be consistent.
+ * If the size change broke because of the journal, no changes will hit
+ * disk anyway.
  */
-static int ocfs2_xattr_set_entry(struct inode *inode,
-                                 struct ocfs2_xattr_info *xi,
-                                 struct ocfs2_xattr_search *xs,
-                                 struct ocfs2_xattr_set_ctxt *ctxt,
-                                 int flag)
-{
-        struct ocfs2_xattr_entry *last;
-        struct ocfs2_inode_info *oi = OCFS2_I(inode);
-        struct ocfs2_dinode *di = (struct ocfs2_dinode *)xs->inode_bh->b_data;
-        size_t min_offs = xs->end - xs->base, name_len = strlen(xi->name);
-        size_t size_l = 0;
-        handle_t *handle = ctxt->handle;
-        int free, i, ret;
-        struct ocfs2_xattr_info xi_l = {
-                .name_index = xi->name_index,
-                .name = xi->name,
-                .value = xi->value,
-                .value_len = xi->value_len,
-        };
-        struct ocfs2_xattr_value_buf vb = {
-                .vb_bh = xs->xattr_bh,
-                .vb_access = ocfs2_journal_access_di,
-        };
+static void ocfs2_xa_cleanup_value_truncate(struct ocfs2_xa_loc *loc,
+                                            const char *what,
+                                            unsigned int orig_clusters)
+{
+        unsigned int new_clusters = ocfs2_xa_value_clusters(loc);
+        char *nameval_buf = ocfs2_xa_offset_pointer(loc,
+                                le16_to_cpu(loc->xl_entry->xe_name_offset));
+
+        if (new_clusters < orig_clusters) {
+                mlog(ML_ERROR,
+                     "Partial truncate while %s xattr %.*s.  Leaking "
+                     "%u clusters and removing the entry\n",
+                     what, loc->xl_entry->xe_name_len, nameval_buf,
+                     orig_clusters - new_clusters);
+                ocfs2_xa_remove_entry(loc);
+        } else if (!orig_clusters) {
+                mlog(ML_ERROR,
+                     "Unable to allocate an external value for xattr "
+                     "%.*s safely.  Leaking %u clusters and removing the "
+                     "entry\n",
+                     loc->xl_entry->xe_name_len, nameval_buf,
+                     new_clusters - orig_clusters);
+                ocfs2_xa_remove_entry(loc);
+        } else if (new_clusters > orig_clusters)
+                mlog(ML_ERROR,
+                     "Unable to grow xattr %.*s safely.  %u new clusters "
+                     "have been added, but the value will not be "
+                     "modified\n",
+                     loc->xl_entry->xe_name_len, nameval_buf,
+                     new_clusters - orig_clusters);
+}
+
+static int ocfs2_xa_remove(struct ocfs2_xa_loc *loc,
+                           struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int rc = 0;
+        unsigned int orig_clusters;
+
+        if (!ocfs2_xattr_is_local(loc->xl_entry)) {
+                orig_clusters = ocfs2_xa_value_clusters(loc);
+                rc = ocfs2_xa_value_truncate(loc, 0, ctxt);
+                if (rc) {
+                        mlog_errno(rc);
+                        /*
+                         * Since this is remove, we can return 0 if
+                         * ocfs2_xa_cleanup_value_truncate() is going to
+                         * wipe the entry anyway.  So we check the
+                         * cluster count as well.
+                         */
+                        if (orig_clusters != ocfs2_xa_value_clusters(loc))
+                                rc = 0;
+                        ocfs2_xa_cleanup_value_truncate(loc, "removing",
+                                                        orig_clusters);
+                        if (rc)
+                                goto out;
+                }
+        }
 
-        if (!(flag & OCFS2_INLINE_XATTR_FL)) {
-                BUG_ON(xs->xattr_bh == xs->inode_bh);
-                vb.vb_access = ocfs2_journal_access_xb;
-        } else
-                BUG_ON(xs->xattr_bh != xs->inode_bh);
+        ocfs2_xa_remove_entry(loc);
 
-        /* Compute min_offs, last and free space. */
-        last = xs->header->xh_entries;
+out:
+        return rc;
+}
 
-        for (i = 0 ; i < le16_to_cpu(xs->header->xh_count); i++) {
-                size_t offs = le16_to_cpu(last->xe_name_offset);
-                if (offs < min_offs)
-                        min_offs = offs;
-                last += 1;
-        }
+static void ocfs2_xa_install_value_root(struct ocfs2_xa_loc *loc)
+{
+        int name_size = OCFS2_XATTR_SIZE(loc->xl_entry->xe_name_len);
+        char *nameval_buf;
 
-        free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP;
-        if (free < 0)
-                return -EIO;
+        nameval_buf = ocfs2_xa_offset_pointer(loc,
+                                le16_to_cpu(loc->xl_entry->xe_name_offset));
+        memcpy(nameval_buf + name_size, &def_xv, OCFS2_XATTR_ROOT_SIZE);
+}
 
-        if (!xs->not_found) {
-                size_t size = 0;
-                if (ocfs2_xattr_is_local(xs->here))
-                        size = OCFS2_XATTR_SIZE(name_len) +
-                        OCFS2_XATTR_SIZE(le64_to_cpu(xs->here->xe_value_size));
-                else
-                        size = OCFS2_XATTR_SIZE(name_len) +
-                                OCFS2_XATTR_ROOT_SIZE;
-                free += (size + sizeof(struct ocfs2_xattr_entry));
-        }
-        /* Check free space in inode or block */
-        if (xi->value && xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
-                if (free < sizeof(struct ocfs2_xattr_entry) +
-                           OCFS2_XATTR_SIZE(name_len) +
-                           OCFS2_XATTR_ROOT_SIZE) {
-                        ret = -ENOSPC;
-                        goto out;
+/*
+ * Take an existing entry and make it ready for the new value.  This
+ * won't allocate space, but it may free space.  It should be ready for
+ * ocfs2_xa_prepare_entry() to finish the work.
+ */
+static int ocfs2_xa_reuse_entry(struct ocfs2_xa_loc *loc,
+                                struct ocfs2_xattr_info *xi,
+                                struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int rc = 0;
+        int name_size = OCFS2_XATTR_SIZE(xi->xi_name_len);
+        unsigned int orig_clusters;
+        char *nameval_buf;
+        int xe_local = ocfs2_xattr_is_local(loc->xl_entry);
+        int xi_local = xi->xi_value_len <= OCFS2_XATTR_INLINE_SIZE;
+
+        BUG_ON(OCFS2_XATTR_SIZE(loc->xl_entry->xe_name_len) !=
+               name_size);
+
+        nameval_buf = ocfs2_xa_offset_pointer(loc,
+                                le16_to_cpu(loc->xl_entry->xe_name_offset));
+        if (xe_local) {
+                memset(nameval_buf + name_size, 0,
+                       namevalue_size_xe(loc->xl_entry) - name_size);
+                if (!xi_local)
+                        ocfs2_xa_install_value_root(loc);
+        } else {
+                orig_clusters = ocfs2_xa_value_clusters(loc);
+                if (xi_local) {
+                        rc = ocfs2_xa_value_truncate(loc, 0, ctxt);
+                        if (rc < 0)
+                                mlog_errno(rc);
+                        else
+                                memset(nameval_buf + name_size, 0,
+                                       namevalue_size_xe(loc->xl_entry) -
+                                       name_size);
+                } else if (le64_to_cpu(loc->xl_entry->xe_value_size) >
+                           xi->xi_value_len) {
+                        rc = ocfs2_xa_value_truncate(loc, xi->xi_value_len,
+                                                     ctxt);
+                        if (rc < 0)
+                                mlog_errno(rc);
                 }
-                size_l = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
-                xi_l.value = (void *)&def_xv;
-                xi_l.value_len = OCFS2_XATTR_ROOT_SIZE;
-        } else if (xi->value) {
-                if (free < sizeof(struct ocfs2_xattr_entry) +
-                           OCFS2_XATTR_SIZE(name_len) +
-                           OCFS2_XATTR_SIZE(xi->value_len)) {
-                        ret = -ENOSPC;
+
+                if (rc) {
+                        ocfs2_xa_cleanup_value_truncate(loc, "reusing",
+                                                        orig_clusters);
                         goto out;
                 }
         }
 
-        if (!xs->not_found) {
-                /* For existing extended attribute */
-                size_t size = OCFS2_XATTR_SIZE(name_len) +
-                        OCFS2_XATTR_SIZE(le64_to_cpu(xs->here->xe_value_size));
-                size_t offs = le16_to_cpu(xs->here->xe_name_offset);
-                void *val = xs->base + offs;
+        loc->xl_entry->xe_value_size = cpu_to_le64(xi->xi_value_len);
+        ocfs2_xattr_set_local(loc->xl_entry, xi_local);
 
-                if (ocfs2_xattr_is_local(xs->here) && size == size_l) {
-                        /* Replace existing local xattr with tree root */
-                        ret = ocfs2_xattr_set_value_outside(inode, xi, xs,
-                                                            ctxt, &vb, offs);
-                        if (ret < 0)
-                                mlog_errno(ret);
-                        goto out;
-                } else if (!ocfs2_xattr_is_local(xs->here)) {
-                        /* For existing xattr which has value outside */
-                        vb.vb_xv = (struct ocfs2_xattr_value_root *)
-                                (val + OCFS2_XATTR_SIZE(name_len));
+out:
+        return rc;
+}
 
-                        if (xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
-                                /*
-                                 * If new value need set outside also,
-                                 * first truncate old value to new value,
-                                 * then set new value with set_value_outside().
-                                 */
-                                ret = ocfs2_xattr_value_truncate(inode,
-                                                                 &vb,
-                                                                 xi->value_len,
-                                                                 ctxt);
-                                if (ret < 0) {
-                                        mlog_errno(ret);
-                                        goto out;
-                                }
+/*
+ * Prepares loc->xl_entry to receive the new xattr.  This includes
+ * properly setting up the name+value pair region.  If loc->xl_entry
+ * already exists, it will take care of modifying it appropriately.
+ *
+ * Note that this modifies the data.  You did journal_access already,
+ * right?
+ */
+static int ocfs2_xa_prepare_entry(struct ocfs2_xa_loc *loc,
+                                  struct ocfs2_xattr_info *xi,
+                                  u32 name_hash,
+                                  struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int rc = 0;
+        unsigned int orig_clusters;
+        __le64 orig_value_size = 0;
 
-                                ret = ocfs2_xattr_update_entry(inode,
-                                                               handle,
-                                                               xi,
-                                                               xs,
-                                                               &vb,
-                                                               offs);
-                                if (ret < 0) {
-                                        mlog_errno(ret);
-                                        goto out;
-                                }
+        rc = ocfs2_xa_check_space(loc, xi);
+        if (rc)
+                goto out;
 
-                                ret = __ocfs2_xattr_set_value_outside(inode,
-                                                                handle,
-                                                                &vb,
-                                                                xi->value,
-                                                                xi->value_len);
-                                if (ret < 0)
-                                        mlog_errno(ret);
+        if (loc->xl_entry) {
+                if (ocfs2_xa_can_reuse_entry(loc, xi)) {
+                        orig_value_size = loc->xl_entry->xe_value_size;
+                        rc = ocfs2_xa_reuse_entry(loc, xi, ctxt);
+                        if (rc)
+                                goto out;
+                        goto alloc_value;
+                }
+
+                if (!ocfs2_xattr_is_local(loc->xl_entry)) {
+                        orig_clusters = ocfs2_xa_value_clusters(loc);
+                        rc = ocfs2_xa_value_truncate(loc, 0, ctxt);
+                        if (rc) {
+                                mlog_errno(rc);
+                                ocfs2_xa_cleanup_value_truncate(loc,
+                                                                "overwriting",
+                                                                orig_clusters);
                                 goto out;
-                        } else {
-                                /*
-                                 * If new value need set in local,
-                                 * just trucate old value to zero.
-                                 */
-                                 ret = ocfs2_xattr_value_truncate(inode,
-                                                                  &vb,
-                                                                  0,
-                                                                  ctxt);
-                                if (ret < 0)
-                                        mlog_errno(ret);
                         }
                 }
+                ocfs2_xa_wipe_namevalue(loc);
+        } else
+                ocfs2_xa_add_entry(loc, name_hash);
+
+        /*
+         * If we get here, we have a blank entry.  Fill it.  We grow our
+         * name+value pair back from the end.
+         */
+        ocfs2_xa_add_namevalue(loc, xi);
+        if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE)
+                ocfs2_xa_install_value_root(loc);
+
+alloc_value:
+        if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) {
+                orig_clusters = ocfs2_xa_value_clusters(loc);
+                rc = ocfs2_xa_value_truncate(loc, xi->xi_value_len, ctxt);
+                if (rc < 0) {
+                        /*
+                         * If we tried to grow an existing external value,
+                         * ocfs2_xa_cleanuP-value_truncate() is going to
+                         * let it stand.  We have to restore its original
+                         * value size.
+                         */
+                        loc->xl_entry->xe_value_size = orig_value_size;
+                        ocfs2_xa_cleanup_value_truncate(loc, "growing",
+                                                        orig_clusters);
+                        mlog_errno(rc);
+                }
         }
 
-        ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), xs->inode_bh,
+out:
+        return rc;
+}
+
+/*
+ * Store the value portion of the name+value pair.  This will skip
+ * values that are stored externally.  Their tree roots were set up
+ * by ocfs2_xa_prepare_entry().
+ */
+static int ocfs2_xa_store_value(struct ocfs2_xa_loc *loc,
+                                struct ocfs2_xattr_info *xi,
+                                struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int rc = 0;
+        int nameval_offset = le16_to_cpu(loc->xl_entry->xe_name_offset);
+        int name_size = OCFS2_XATTR_SIZE(xi->xi_name_len);
+        char *nameval_buf;
+        struct ocfs2_xattr_value_buf vb;
+
+        nameval_buf = ocfs2_xa_offset_pointer(loc, nameval_offset);
+        if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) {
+                ocfs2_xa_fill_value_buf(loc, &vb);
+                rc = __ocfs2_xattr_set_value_outside(loc->xl_inode,
+                                                     ctxt->handle, &vb,
+                                                     xi->xi_value,
+                                                     xi->xi_value_len);
+        } else
+                memcpy(nameval_buf + name_size, xi->xi_value, xi->xi_value_len);
+
+        return rc;
+}
+
+static int ocfs2_xa_set(struct ocfs2_xa_loc *loc,
+                        struct ocfs2_xattr_info *xi,
+                        struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int ret;
+        u32 name_hash = ocfs2_xattr_name_hash(loc->xl_inode, xi->xi_name,
+                                              xi->xi_name_len);
+
+        ret = ocfs2_xa_journal_access(ctxt->handle, loc,
                                       OCFS2_JOURNAL_ACCESS_WRITE);
         if (ret) {
                 mlog_errno(ret);
                 goto out;
         }
 
-        if (!(flag & OCFS2_INLINE_XATTR_FL)) {
-                ret = vb.vb_access(handle, INODE_CACHE(inode), vb.vb_bh,
-                                   OCFS2_JOURNAL_ACCESS_WRITE);
-                if (ret) {
-                        mlog_errno(ret);
-                        goto out;
-                }
-        }
-
         /*
-         * Set value in local, include set tree root in local.
-         * This is the first step for value size >INLINE_SIZE.
+         * From here on out, everything is going to modify the buffer a
+         * little.  Errors are going to leave the xattr header in a
+         * sane state.  Thus, even with errors we dirty the sucker.
          */
-        ocfs2_xattr_set_entry_local(inode, &xi_l, xs, last, min_offs);
 
-        if (!(flag & OCFS2_INLINE_XATTR_FL)) {
-                ret = ocfs2_journal_dirty(handle, xs->xattr_bh);
-                if (ret < 0) {
-                        mlog_errno(ret);
-                        goto out;
-                }
+        /* Don't worry, we are never called with !xi_value and !xl_entry */
+        if (!xi->xi_value) {
+                ret = ocfs2_xa_remove(loc, ctxt);
+                goto out_dirty;
         }
 
-        if (!(oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL) &&
-            (flag & OCFS2_INLINE_XATTR_FL)) {
-                struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-                unsigned int xattrsize = osb->s_xattr_inline_size;
-
-                /*
-                 * Adjust extent record count or inline data size
-                 * to reserve space for extended attribute.
-                 */
-                if (oi->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
-                        struct ocfs2_inline_data *idata = &di->id2.i_data;
-                        le16_add_cpu(&idata->id_count, -xattrsize);
-                } else if (!(ocfs2_inode_is_fast_symlink(inode))) {
-                        struct ocfs2_extent_list *el = &di->id2.i_list;
-                        le16_add_cpu(&el->l_count, -(xattrsize /
-                                        sizeof(struct ocfs2_extent_rec)));
-                }
-                di->i_xattr_inline_size = cpu_to_le16(xattrsize);
+        ret = ocfs2_xa_prepare_entry(loc, xi, name_hash, ctxt);
+        if (ret) {
+                if (ret != -ENOSPC)
+                        mlog_errno(ret);
+                goto out_dirty;
         }
-        /* Update xattr flag */
-        spin_lock(&oi->ip_lock);
-        oi->ip_dyn_features |= flag;
-        di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
-        spin_unlock(&oi->ip_lock);
 
-        ret = ocfs2_journal_dirty(handle, xs->inode_bh);
-        if (ret < 0)
+        ret = ocfs2_xa_store_value(loc, xi, ctxt);
+        if (ret)
                 mlog_errno(ret);
 
-        if (!ret && xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
-                /*
-                 * Set value outside in B tree.
-                 * This is the second step for value size > INLINE_SIZE.
-                 */
-                size_t offs = le16_to_cpu(xs->here->xe_name_offset);
-                ret = ocfs2_xattr_set_value_outside(inode, xi, xs, ctxt,
-                                                    &vb, offs);
-                if (ret < 0) {
-                        int ret2;
+out_dirty:
+        ocfs2_xa_journal_dirty(ctxt->handle, loc);
 
-                        mlog_errno(ret);
-                        /*
-                         * If set value outside failed, we have to clean
-                         * the junk tree root we have already set in local.
-                         */
-                        ret2 = ocfs2_xattr_cleanup(inode, ctxt->handle,
-                                                   xi, xs, &vb, offs);
-                        if (ret2 < 0)
-                                mlog_errno(ret2);
-                }
-        }
 out:
         return ret;
 }
 
+static void ocfs2_init_dinode_xa_loc(struct ocfs2_xa_loc *loc,
+                                     struct inode *inode,
+                                     struct buffer_head *bh,
+                                     struct ocfs2_xattr_entry *entry)
+{
+        struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+        BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_XATTR_FL));
+
+        loc->xl_inode = inode;
+        loc->xl_ops = &ocfs2_xa_block_loc_ops;
+        loc->xl_storage = bh;
+        loc->xl_entry = entry;
+        loc->xl_size = le16_to_cpu(di->i_xattr_inline_size);
+        loc->xl_header =
+                (struct ocfs2_xattr_header *)(bh->b_data + bh->b_size -
+                                              loc->xl_size);
+}
+
+static void ocfs2_init_xattr_block_xa_loc(struct ocfs2_xa_loc *loc,
+                                          struct inode *inode,
+                                          struct buffer_head *bh,
+                                          struct ocfs2_xattr_entry *entry)
+{
+        struct ocfs2_xattr_block *xb =
+                (struct ocfs2_xattr_block *)bh->b_data;
+
+        BUG_ON(le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED);
+
+        loc->xl_inode = inode;
+        loc->xl_ops = &ocfs2_xa_block_loc_ops;
+        loc->xl_storage = bh;
+        loc->xl_header = &(xb->xb_attrs.xb_header);
+        loc->xl_entry = entry;
+        loc->xl_size = bh->b_size - offsetof(struct ocfs2_xattr_block,
+                                             xb_attrs.xb_header);
+}
+
+static void ocfs2_init_xattr_bucket_xa_loc(struct ocfs2_xa_loc *loc,
+                                           struct ocfs2_xattr_bucket *bucket,
+                                           struct ocfs2_xattr_entry *entry)
+{
+        loc->xl_inode = bucket->bu_inode;
+        loc->xl_ops = &ocfs2_xa_bucket_loc_ops;
+        loc->xl_storage = bucket;
+        loc->xl_header = bucket_xh(bucket);
+        loc->xl_entry = entry;
+        loc->xl_size = OCFS2_XATTR_BUCKET_SIZE;
+}
+
 /*
  * In xattr remove, if it is stored outside and refcounted, we may have
  * the chance to split the refcount tree. So need the allocators.
@@ -2155,6 +2683,55 @@ static int ocfs2_xattr_ibody_find(struct inode *inode,
         return 0;
 }
 
+static int ocfs2_xattr_ibody_init(struct inode *inode,
+                                  struct buffer_head *di_bh,
+                                  struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int ret;
+        struct ocfs2_inode_info *oi = OCFS2_I(inode);
+        struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        unsigned int xattrsize = osb->s_xattr_inline_size;
+
+        if (!ocfs2_xattr_has_space_inline(inode, di)) {
+                ret = -ENOSPC;
+                goto out;
+        }
+
+        ret = ocfs2_journal_access_di(ctxt->handle, INODE_CACHE(inode), di_bh,
+                                      OCFS2_JOURNAL_ACCESS_WRITE);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
+        }
+
+        /*
+         * Adjust extent record count or inline data size
+         * to reserve space for extended attribute.
+         */
+        if (oi->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
+                struct ocfs2_inline_data *idata = &di->id2.i_data;
+                le16_add_cpu(&idata->id_count, -xattrsize);
+        } else if (!(ocfs2_inode_is_fast_symlink(inode))) {
+                struct ocfs2_extent_list *el = &di->id2.i_list;
+                le16_add_cpu(&el->l_count, -(xattrsize /
+                                             sizeof(struct ocfs2_extent_rec)));
+        }
+        di->i_xattr_inline_size = cpu_to_le16(xattrsize);
+
+        spin_lock(&oi->ip_lock);
+        oi->ip_dyn_features |= OCFS2_INLINE_XATTR_FL|OCFS2_HAS_XATTR_FL;
+        di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
+        spin_unlock(&oi->ip_lock);
+
+        ret = ocfs2_journal_dirty(ctxt->handle, di_bh);
+        if (ret < 0)
+                mlog_errno(ret);
+
+out:
+        return ret;
+}
+
 /*
  * ocfs2_xattr_ibody_set()
  *
@@ -2166,9 +2743,10 @@ static int ocfs2_xattr_ibody_set(struct inode *inode,
                                  struct ocfs2_xattr_search *xs,
                                  struct ocfs2_xattr_set_ctxt *ctxt)
 {
+        int ret;
         struct ocfs2_inode_info *oi = OCFS2_I(inode);
         struct ocfs2_dinode *di = (struct ocfs2_dinode *)xs->inode_bh->b_data;
-        int ret;
+        struct ocfs2_xa_loc loc;
 
         if (inode->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE)
                 return -ENOSPC;
@@ -2181,8 +2759,25 @@ static int ocfs2_xattr_ibody_set(struct inode *inode,
                 }
         }
 
-        ret = ocfs2_xattr_set_entry(inode, xi, xs, ctxt,
-                                (OCFS2_INLINE_XATTR_FL | OCFS2_HAS_XATTR_FL));
+        if (!(oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL)) {
+                ret = ocfs2_xattr_ibody_init(inode, xs->inode_bh, ctxt);
+                if (ret) {
+                        if (ret != -ENOSPC)
+                                mlog_errno(ret);
+                        goto out;
+                }
+        }
+
+        ocfs2_init_dinode_xa_loc(&loc, inode, xs->inode_bh,
+                                 xs->not_found ? NULL : xs->here);
+        ret = ocfs2_xa_set(&loc, xi, ctxt);
+        if (ret) {
+                if (ret != -ENOSPC)
+                        mlog_errno(ret);
+                goto out;
+        }
+        xs->here = loc.xl_entry;
+
 out:
         up_write(&oi->ip_alloc_sem);
 
@@ -2242,12 +2837,11 @@ cleanup:
         return ret;
 }
 
-static int ocfs2_create_xattr_block(handle_t *handle,
-                                    struct inode *inode,
+static int ocfs2_create_xattr_block(struct inode *inode,
                                     struct buffer_head *inode_bh,
-                                    struct ocfs2_alloc_context *meta_ac,
-                                    struct buffer_head **ret_bh,
-                                    int indexed)
+                                    struct ocfs2_xattr_set_ctxt *ctxt,
+                                    int indexed,
+                                    struct buffer_head **ret_bh)
 {
         int ret;
         u16 suballoc_bit_start;
@@ -2258,14 +2852,14 @@ static int ocfs2_create_xattr_block(handle_t *handle,
         struct buffer_head *new_bh = NULL;
         struct ocfs2_xattr_block *xblk;
 
-        ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), inode_bh,
-                                      OCFS2_JOURNAL_ACCESS_CREATE);
+        ret = ocfs2_journal_access_di(ctxt->handle, INODE_CACHE(inode),
+                                      inode_bh, OCFS2_JOURNAL_ACCESS_CREATE);
         if (ret < 0) {
                 mlog_errno(ret);
                 goto end;
         }
 
-        ret = ocfs2_claim_metadata(osb, handle, meta_ac, 1,
+        ret = ocfs2_claim_metadata(osb, ctxt->handle, ctxt->meta_ac, 1,
                                    &suballoc_bit_start, &num_got,
                                    &first_blkno);
         if (ret < 0) {
@@ -2276,7 +2870,7 @@ static int ocfs2_create_xattr_block(handle_t *handle,
         new_bh = sb_getblk(inode->i_sb, first_blkno);
         ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), new_bh);
 
-        ret = ocfs2_journal_access_xb(handle, INODE_CACHE(inode),
+        ret = ocfs2_journal_access_xb(ctxt->handle, INODE_CACHE(inode),
                                       new_bh,
                                       OCFS2_JOURNAL_ACCESS_CREATE);
         if (ret < 0) {
@@ -2288,11 +2882,10 @@ static int ocfs2_create_xattr_block(handle_t *handle,
         xblk = (struct ocfs2_xattr_block *)new_bh->b_data;
         memset(xblk, 0, inode->i_sb->s_blocksize);
         strcpy((void *)xblk, OCFS2_XATTR_BLOCK_SIGNATURE);
-        xblk->xb_suballoc_slot = cpu_to_le16(osb->slot_num);
+        xblk->xb_suballoc_slot = cpu_to_le16(ctxt->meta_ac->ac_alloc_slot);
         xblk->xb_suballoc_bit = cpu_to_le16(suballoc_bit_start);
         xblk->xb_fs_generation = cpu_to_le32(osb->fs_generation);
         xblk->xb_blkno = cpu_to_le64(first_blkno);
-
         if (indexed) {
                 struct ocfs2_xattr_tree_root *xr = &xblk->xb_attrs.xb_root;
                 xr->xt_clusters = cpu_to_le32(1);
@@ -2303,14 +2896,17 @@ static int ocfs2_create_xattr_block(handle_t *handle,
                 xr->xt_list.l_next_free_rec = cpu_to_le16(1);
                 xblk->xb_flags = cpu_to_le16(OCFS2_XATTR_INDEXED);
         }
+        ocfs2_journal_dirty(ctxt->handle, new_bh);
 
-        ret = ocfs2_journal_dirty(handle, new_bh);
-        if (ret < 0) {
-                mlog_errno(ret);
-                goto end;
-        }
+        /* Add it to the inode */
         di->i_xattr_loc = cpu_to_le64(first_blkno);
-        ocfs2_journal_dirty(handle, inode_bh);
+
+        spin_lock(&OCFS2_I(inode)->ip_lock);
+        OCFS2_I(inode)->ip_dyn_features |= OCFS2_HAS_XATTR_FL;
+        di->i_dyn_features = cpu_to_le16(OCFS2_I(inode)->ip_dyn_features);
+        spin_unlock(&OCFS2_I(inode)->ip_lock);
+
+        ocfs2_journal_dirty(ctxt->handle, inode_bh);
 
         *ret_bh = new_bh;
         new_bh = NULL;
@@ -2332,13 +2928,13 @@ static int ocfs2_xattr_block_set(struct inode *inode,
                                  struct ocfs2_xattr_set_ctxt *ctxt)
 {
         struct buffer_head *new_bh = NULL;
-        handle_t *handle = ctxt->handle;
         struct ocfs2_xattr_block *xblk = NULL;
         int ret;
+        struct ocfs2_xa_loc loc;
 
         if (!xs->xattr_bh) {
-                ret = ocfs2_create_xattr_block(handle, inode, xs->inode_bh,
-                                               ctxt->meta_ac, &new_bh, 0);
+                ret = ocfs2_create_xattr_block(inode, xs->inode_bh, ctxt,
+                                               0, &new_bh);
                 if (ret) {
                         mlog_errno(ret);
                         goto end;
@@ -2354,21 +2950,25 @@ static int ocfs2_xattr_block_set(struct inode *inode,
                 xblk = (struct ocfs2_xattr_block *)xs->xattr_bh->b_data;
 
         if (!(le16_to_cpu(xblk->xb_flags) & OCFS2_XATTR_INDEXED)) {
-                /* Set extended attribute into external block */
-                ret = ocfs2_xattr_set_entry(inode, xi, xs, ctxt,
-                                            OCFS2_HAS_XATTR_FL);
-                if (!ret || ret != -ENOSPC)
-                        goto end;
+                ocfs2_init_xattr_block_xa_loc(&loc, inode, xs->xattr_bh,
+                                              xs->not_found ? NULL : xs->here);
 
-                ret = ocfs2_xattr_create_index_block(inode, xs, ctxt);
-                if (ret)
+                ret = ocfs2_xa_set(&loc, xi, ctxt);
+                if (!ret)
+                        xs->here = loc.xl_entry;
+                else if (ret != -ENOSPC)
                         goto end;
+                else {
+                        ret = ocfs2_xattr_create_index_block(inode, xs, ctxt);
+                        if (ret)
+                                goto end;
+                }
         }
 
-        ret = ocfs2_xattr_set_entry_index_block(inode, xi, xs, ctxt);
+        if (le16_to_cpu(xblk->xb_flags) & OCFS2_XATTR_INDEXED)
+                ret = ocfs2_xattr_set_entry_index_block(inode, xi, xs, ctxt);
 
 end:
-
         return ret;
 }
 
@@ -2377,7 +2977,6 @@ static int ocfs2_xattr_can_be_in_inode(struct inode *inode,
                                        struct ocfs2_xattr_info *xi,
                                        struct ocfs2_xattr_search *xs)
 {
-        u64 value_size;
         struct ocfs2_xattr_entry *last;
         int free, i;
         size_t min_offs = xs->end - xs->base;
@@ -2400,13 +2999,7 @@ static int ocfs2_xattr_can_be_in_inode(struct inode *inode,
 
         BUG_ON(!xs->not_found);
 
-        if (xi->value_len > OCFS2_XATTR_INLINE_SIZE)
-                value_size = OCFS2_XATTR_ROOT_SIZE;
-        else
-                value_size = OCFS2_XATTR_SIZE(xi->value_len);
-
-        if (free >= sizeof(struct ocfs2_xattr_entry) +
-                   OCFS2_XATTR_SIZE(strlen(xi->name)) + value_size)
+        if (free >= (sizeof(struct ocfs2_xattr_entry) + namevalue_size_xi(xi)))
                 return 1;
 
         return 0;
@@ -2430,7 +3023,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode,
         char *base = NULL;
         int name_offset, name_len = 0;
         u32 new_clusters = ocfs2_clusters_for_bytes(inode->i_sb,
-                                                    xi->value_len);
+                                                    xi->xi_value_len);
         u64 value_size;
 
         /*
@@ -2438,14 +3031,14 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode,
          * No matter whether we replace an old one or add a new one,
          * we need this for writing.
          */
-        if (xi->value_len > OCFS2_XATTR_INLINE_SIZE)
+        if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE)
                 credits += new_clusters *
                            ocfs2_clusters_to_blocks(inode->i_sb, 1);
 
         if (xis->not_found && xbs->not_found) {
                 credits += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 
-                if (xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
+                if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) {
                         clusters_add += new_clusters;
                         credits += ocfs2_calc_extend_credits(inode->i_sb,
                                                         &def_xv.xv.xr_list,
@@ -2490,7 +3083,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode,
          * The credits for removing the value tree will be extended
          * by ocfs2_remove_extent itself.
          */
-        if (!xi->value) {
+        if (!xi->xi_value) {
                 if (!ocfs2_xattr_is_local(xe))
                         credits += ocfs2_remove_extent_credits(inode->i_sb);
 
@@ -2520,7 +3113,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode,
                 }
         }
 
-        if (xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
+        if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) {
                 /* the new values will be stored outside. */
                 u32 old_clusters = 0;
 
@@ -2553,9 +3146,10 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode,
                  * value, we don't need any allocation, otherwise we have
                  * to guess metadata allocation.
                  */
-                if ((ocfs2_xattr_is_local(xe) && value_size >= xi->value_len) ||
+                if ((ocfs2_xattr_is_local(xe) &&
+                     (value_size >= xi->xi_value_len)) ||
                     (!ocfs2_xattr_is_local(xe) &&
-                     OCFS2_XATTR_ROOT_SIZE >= xi->value_len))
+                     OCFS2_XATTR_ROOT_SIZE >= xi->xi_value_len))
                         goto out;
         }
 
@@ -2645,7 +3239,7 @@ static int ocfs2_init_xattr_set_ctxt(struct inode *inode,
 
         meta_add += extra_meta;
         mlog(0, "Set xattr %s, reserve meta blocks = %d, clusters = %d, "
-             "credits = %d\n", xi->name, meta_add, clusters_add, *credits);
+             "credits = %d\n", xi->xi_name, meta_add, clusters_add, *credits);
 
         if (meta_add) {
                 ret = ocfs2_reserve_new_metadata_blocks(osb, meta_add,
@@ -2685,7 +3279,7 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
 {
         int ret = 0, credits, old_found;
 
-        if (!xi->value) {
+        if (!xi->xi_value) {
                 /* Remove existing extended attribute */
                 if (!xis->not_found)
                         ret = ocfs2_xattr_ibody_set(inode, xi, xis, ctxt);
@@ -2699,8 +3293,8 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
                          * If succeed and that extended attribute existing in
                          * external block, then we will remove it.
                          */
-                        xi->value = NULL;
-                        xi->value_len = 0;
+                        xi->xi_value = NULL;
+                        xi->xi_value_len = 0;
 
                         old_found = xis->not_found;
                         xis->not_found = -ENODATA;
@@ -2728,8 +3322,8 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
                 } else if (ret == -ENOSPC) {
                         if (di->i_xattr_loc && !xbs->xattr_bh) {
                                 ret = ocfs2_xattr_block_find(inode,
-                                                             xi->name_index,
-                                                             xi->name, xbs);
+                                                             xi->xi_name_index,
+                                                             xi->xi_name, xbs);
                                 if (ret)
                                         goto out;
 
@@ -2768,8 +3362,8 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
                                  * If succeed and that extended attribute
                                  * existing in inode, we will remove it.
                                  */
-                                xi->value = NULL;
-                                xi->value_len = 0;
+                                xi->xi_value = NULL;
+                                xi->xi_value_len = 0;
                                 xbs->not_found = -ENODATA;
                                 ret = ocfs2_calc_xattr_set_need(inode,
                                                                 di,
@@ -2835,10 +3429,11 @@ int ocfs2_xattr_set_handle(handle_t *handle,
         int ret;
 
         struct ocfs2_xattr_info xi = {
-                .name_index = name_index,
-                .name = name,
-                .value = value,
-                .value_len = value_len,
+                .xi_name_index = name_index,
+                .xi_name = name,
+                .xi_name_len = strlen(name),
+                .xi_value = value,
+                .xi_value_len = value_len,
         };
 
         struct ocfs2_xattr_search xis = {
@@ -2918,10 +3513,11 @@ int ocfs2_xattr_set(struct inode *inode,
         struct ocfs2_refcount_tree *ref_tree = NULL;
 
         struct ocfs2_xattr_info xi = {
-                .name_index = name_index,
-                .name = name,
-                .value = value,
-                .value_len = value_len,
+                .xi_name_index = name_index,
+                .xi_name = name,
+                .xi_name_len = strlen(name),
+                .xi_value = value,
+                .xi_value_len = value_len,
         };
 
         struct ocfs2_xattr_search xis = {
@@ -3765,7 +4361,7 @@ static int ocfs2_defrag_xattr_bucket(struct inode *inode,
                                      struct ocfs2_xattr_bucket *bucket)
 {
         int ret, i;
-        size_t end, offset, len, value_len;
+        size_t end, offset, len;
         struct ocfs2_xattr_header *xh;
         char *entries, *buf, *bucket_buf = NULL;
         u64 blkno = bucket_blkno(bucket);
@@ -3819,12 +4415,7 @@ static int ocfs2_defrag_xattr_bucket(struct inode *inode,
         end = OCFS2_XATTR_BUCKET_SIZE;
         for (i = 0; i < le16_to_cpu(xh->xh_count); i++, xe++) {
                 offset = le16_to_cpu(xe->xe_name_offset);
-                if (ocfs2_xattr_is_local(xe))
-                        value_len = OCFS2_XATTR_SIZE(
-                                        le64_to_cpu(xe->xe_value_size));
-                else
-                        value_len = OCFS2_XATTR_ROOT_SIZE;
-                len = OCFS2_XATTR_SIZE(xe->xe_name_len) + value_len;
+                len = namevalue_size_xe(xe);
 
                 /*
                  * We must make sure that the name/value pair
@@ -4013,7 +4604,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
                                     int new_bucket_head)
 {
         int ret, i;
-        int count, start, len, name_value_len = 0, xe_len, name_offset = 0;
+        int count, start, len, name_value_len = 0, name_offset = 0;
         struct ocfs2_xattr_bucket *s_bucket = NULL, *t_bucket = NULL;
         struct ocfs2_xattr_header *xh;
         struct ocfs2_xattr_entry *xe;
@@ -4104,13 +4695,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
         name_value_len = 0;
         for (i = 0; i < start; i++) {
                 xe = &xh->xh_entries[i];
-                xe_len = OCFS2_XATTR_SIZE(xe->xe_name_len);
-                if (ocfs2_xattr_is_local(xe))
-                        xe_len +=
-                           OCFS2_XATTR_SIZE(le64_to_cpu(xe->xe_value_size));
-                else
-                        xe_len += OCFS2_XATTR_ROOT_SIZE;
-                name_value_len += xe_len;
+                name_value_len += namevalue_size_xe(xe);
                 if (le16_to_cpu(xe->xe_name_offset) < name_offset)
                         name_offset = le16_to_cpu(xe->xe_name_offset);
         }
@@ -4140,12 +4725,6 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
         xh->xh_free_start = cpu_to_le16(OCFS2_XATTR_BUCKET_SIZE);
         for (i = 0; i < le16_to_cpu(xh->xh_count); i++) {
                 xe = &xh->xh_entries[i];
-                xe_len = OCFS2_XATTR_SIZE(xe->xe_name_len);
-                if (ocfs2_xattr_is_local(xe))
-                        xe_len +=
-                           OCFS2_XATTR_SIZE(le64_to_cpu(xe->xe_value_size));
-                else
-                        xe_len += OCFS2_XATTR_ROOT_SIZE;
                 if (le16_to_cpu(xe->xe_name_offset) <
                     le16_to_cpu(xh->xh_free_start))
                         xh->xh_free_start = xe->xe_name_offset;
@@ -4757,195 +5336,6 @@ static inline char *ocfs2_xattr_bucket_get_val(struct inode *inode,
 }
 
 /*
- * Handle the normal xattr set, including replace, delete and new.
- *
- * Note: "local" indicates the real data's locality. So we can't
- * just its bucket locality by its length.
- */
-static void ocfs2_xattr_set_entry_normal(struct inode *inode,
-                                         struct ocfs2_xattr_info *xi,
-                                         struct ocfs2_xattr_search *xs,
-                                         u32 name_hash,
-                                         int local)
-{
-        struct ocfs2_xattr_entry *last, *xe;
-        int name_len = strlen(xi->name);
-        struct ocfs2_xattr_header *xh = xs->header;
-        u16 count = le16_to_cpu(xh->xh_count), start;
-        size_t blocksize = inode->i_sb->s_blocksize;
-        char *val;
-        size_t offs, size, new_size;
-
-        last = &xh->xh_entries[count];
-        if (!xs->not_found) {
-                xe = xs->here;
-                offs = le16_to_cpu(xe->xe_name_offset);
-                if (ocfs2_xattr_is_local(xe))
-                        size = OCFS2_XATTR_SIZE(name_len) +
-                        OCFS2_XATTR_SIZE(le64_to_cpu(xe->xe_value_size));
-                else
-                        size = OCFS2_XATTR_SIZE(name_len) +
-                        OCFS2_XATTR_SIZE(OCFS2_XATTR_ROOT_SIZE);
-
-                /*
-                 * If the new value will be stored outside, xi->value has been
-                 * initalized as an empty ocfs2_xattr_value_root, and the same
-                 * goes with xi->value_len, so we can set new_size safely here.
-                 * See ocfs2_xattr_set_in_bucket.
-                 */
-                new_size = OCFS2_XATTR_SIZE(name_len) +
-                           OCFS2_XATTR_SIZE(xi->value_len);
-
-                le16_add_cpu(&xh->xh_name_value_len, -size);
-                if (xi->value) {
-                        if (new_size > size)
-                                goto set_new_name_value;
-
-                        /* Now replace the old value with new one. */
-                        if (local)
-                                xe->xe_value_size = cpu_to_le64(xi->value_len);
-                        else
-                                xe->xe_value_size = 0;
-
-                        val = ocfs2_xattr_bucket_get_val(inode,
-                                                         xs->bucket, offs);
-                        memset(val + OCFS2_XATTR_SIZE(name_len), 0,
-                               size - OCFS2_XATTR_SIZE(name_len));
-                        if (OCFS2_XATTR_SIZE(xi->value_len) > 0)
-                                memcpy(val + OCFS2_XATTR_SIZE(name_len),
-                                       xi->value, xi->value_len);
-
-                        le16_add_cpu(&xh->xh_name_value_len, new_size);
-                        ocfs2_xattr_set_local(xe, local);
-                        return;
-                } else {
-                        /*
-                         * Remove the old entry if there is more than one.
-                         * We don't remove the last entry so that we can
-                         * use it to indicate the hash value of the empty
-                         * bucket.
-                         */
-                        last -= 1;
-                        le16_add_cpu(&xh->xh_count, -1);
-                        if (xh->xh_count) {
-                                memmove(xe, xe + 1,
-                                        (void *)last - (void *)xe);
-                                memset(last, 0,
-                                       sizeof(struct ocfs2_xattr_entry));
-                        } else
-                                xh->xh_free_start =
-                                        cpu_to_le16(OCFS2_XATTR_BUCKET_SIZE);
-
-                        return;
-                }
-        } else {
-                /* find a new entry for insert. */
-                int low = 0, high = count - 1, tmp;
-                struct ocfs2_xattr_entry *tmp_xe;
-
-                while (low <= high && count) {
-                        tmp = (low + high) / 2;
-                        tmp_xe = &xh->xh_entries[tmp];
-
-                        if (name_hash > le32_to_cpu(tmp_xe->xe_name_hash))
-                                low = tmp + 1;
-                        else if (name_hash <
-                                 le32_to_cpu(tmp_xe->xe_name_hash))
-                                high = tmp - 1;
-                        else {
-                                low = tmp;
-                                break;
-                        }
-                }
-
-                xe = &xh->xh_entries[low];
-                if (low != count)
-                        memmove(xe + 1, xe, (void *)last - (void *)xe);
-
-                le16_add_cpu(&xh->xh_count, 1);
-                memset(xe, 0, sizeof(struct ocfs2_xattr_entry));
-                xe->xe_name_hash = cpu_to_le32(name_hash);
-                xe->xe_name_len = name_len;
-                ocfs2_xattr_set_type(xe, xi->name_index);
-        }
-
-set_new_name_value:
-        /* Insert the new name+value. */
-        size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_SIZE(xi->value_len);
-
-        /*
-         * We must make sure that the name/value pair
-         * exists in the same block.
-         */
-        offs = le16_to_cpu(xh->xh_free_start);
-        start = offs - size;
-
-        if (start >> inode->i_sb->s_blocksize_bits !=
-            (offs - 1) >> inode->i_sb->s_blocksize_bits) {
-                offs = offs - offs % blocksize;
-                xh->xh_free_start = cpu_to_le16(offs);
-        }
-
-        val = ocfs2_xattr_bucket_get_val(inode, xs->bucket, offs - size);
-        xe->xe_name_offset = cpu_to_le16(offs - size);
-
-        memset(val, 0, size);
-        memcpy(val, xi->name, name_len);
-        memcpy(val + OCFS2_XATTR_SIZE(name_len), xi->value, xi->value_len);
-
-        xe->xe_value_size = cpu_to_le64(xi->value_len);
-        ocfs2_xattr_set_local(xe, local);
-        xs->here = xe;
-        le16_add_cpu(&xh->xh_free_start, -size);
-        le16_add_cpu(&xh->xh_name_value_len, size);
-
-        return;
-}
-
-/*
- * Set the xattr entry in the specified bucket.
- * The bucket is indicated by xs->bucket and it should have the enough
- * space for the xattr insertion.
- */
-static int ocfs2_xattr_set_entry_in_bucket(struct inode *inode,
-                                           handle_t *handle,
-                                           struct ocfs2_xattr_info *xi,
-                                           struct ocfs2_xattr_search *xs,
-                                           u32 name_hash,
-                                           int local)
-{
-        int ret;
-        u64 blkno;
-
-        mlog(0, "Set xattr entry len = %lu index = %d in bucket %llu\n",
-             (unsigned long)xi->value_len, xi->name_index,
-             (unsigned long long)bucket_blkno(xs->bucket));
-
-        if (!xs->bucket->bu_bhs[1]) {
-                blkno = bucket_blkno(xs->bucket);
-                ocfs2_xattr_bucket_relse(xs->bucket);
-                ret = ocfs2_read_xattr_bucket(xs->bucket, blkno);
-                if (ret) {
-                        mlog_errno(ret);
-                        goto out;
-                }
-        }
-
-        ret = ocfs2_xattr_bucket_journal_access(handle, xs->bucket,
-                                                OCFS2_JOURNAL_ACCESS_WRITE);
-        if (ret < 0) {
-                mlog_errno(ret);
-                goto out;
-        }
-
-        ocfs2_xattr_set_entry_normal(inode, xi, xs, name_hash, local);
-        ocfs2_xattr_bucket_journal_dirty(handle, xs->bucket);
-
-out:
-        return ret;
-}
-
-/*
  * Truncate the specified xe_off entry in xattr bucket.
  * bucket is indicated by header_bh and len is the new length.
  * Both the ocfs2_xattr_value_root and the entry will be updated here.
@@ -5015,66 +5405,6 @@ out:
         return ret;
 }
 
-static int ocfs2_xattr_bucket_value_truncate_xs(struct inode *inode,
-                                        struct ocfs2_xattr_search *xs,
-                                        int len,
-                                        struct ocfs2_xattr_set_ctxt *ctxt)
-{
-        int ret, offset;
-        struct ocfs2_xattr_entry *xe = xs->here;
-        struct ocfs2_xattr_header *xh = (struct ocfs2_xattr_header *)xs->base;
-
-        BUG_ON(!xs->bucket->bu_bhs[0] || !xe || ocfs2_xattr_is_local(xe));
-
-        offset = xe - xh->xh_entries;
-        ret = ocfs2_xattr_bucket_value_truncate(inode, xs->bucket,
-                                                offset, len, ctxt);
-        if (ret)
-                mlog_errno(ret);
-
-        return ret;
-}
-
-static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode,
-                                                handle_t *handle,
-                                                struct ocfs2_xattr_search *xs,
-                                                char *val,
-                                                int value_len)
-{
-        int ret, offset, block_off;
-        struct ocfs2_xattr_value_root *xv;
-        struct ocfs2_xattr_entry *xe = xs->here;
-        struct ocfs2_xattr_header *xh = bucket_xh(xs->bucket);
-        void *base;
-        struct ocfs2_xattr_value_buf vb = {
-                .vb_access = ocfs2_journal_access,
-        };
-
-        BUG_ON(!xs->base || !xe || ocfs2_xattr_is_local(xe));
-
-        ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb, xh,
-                                                xe - xh->xh_entries,
-                                                &block_off,
-                                                &offset);
-        if (ret) {
-                mlog_errno(ret);
-                goto out;
-        }
-
-        base = bucket_block(xs->bucket, block_off);
-        xv = (struct ocfs2_xattr_value_root *)(base + offset +
-                 OCFS2_XATTR_SIZE(xe->xe_name_len));
-
-        vb.vb_xv = xv;
-        vb.vb_bh = xs->bucket->bu_bhs[block_off];
-        ret = __ocfs2_xattr_set_value_outside(inode, handle,
-                                              &vb, val, value_len);
-        if (ret)
-                mlog_errno(ret);
-out:
-        return ret;
-}
-
 static int ocfs2_rm_xattr_cluster(struct inode *inode,
                                   struct buffer_head *root_bh,
                                   u64 blkno,
@@ -5173,128 +5503,6 @@ out:
         return ret;
 }
 
-static void ocfs2_xattr_bucket_remove_xs(struct inode *inode,
-                                         handle_t *handle,
-                                         struct ocfs2_xattr_search *xs)
-{
-        struct ocfs2_xattr_header *xh = bucket_xh(xs->bucket);
-        struct ocfs2_xattr_entry *last = &xh->xh_entries[
-                                                le16_to_cpu(xh->xh_count) - 1];
-        int ret = 0;
-
-        ret = ocfs2_xattr_bucket_journal_access(handle, xs->bucket,
-                                                OCFS2_JOURNAL_ACCESS_WRITE);
-        if (ret) {
-                mlog_errno(ret);
-                return;
-        }
-
-        /* Remove the old entry. */
-        memmove(xs->here, xs->here + 1,
-                (void *)last - (void *)xs->here);
-        memset(last, 0, sizeof(struct ocfs2_xattr_entry));
-        le16_add_cpu(&xh->xh_count, -1);
-
-        ocfs2_xattr_bucket_journal_dirty(handle, xs->bucket);
-}
-
-/*
- * Set the xattr name/value in the bucket specified in xs.
- *
- * As the new value in xi may be stored in the bucket or in an outside cluster,
- * we divide the whole process into 3 steps:
- * 1. insert name/value in the bucket(ocfs2_xattr_set_entry_in_bucket)
- * 2. truncate of the outside cluster(ocfs2_xattr_bucket_value_truncate_xs)
- * 3. Set the value to the outside cluster(ocfs2_xattr_bucket_set_value_outside)
- * 4. If the clusters for the new outside value can't be allocated, we need
- *    to free the xattr we allocated in set.
- */
-static int ocfs2_xattr_set_in_bucket(struct inode *inode,
-                                     struct ocfs2_xattr_info *xi,
-                                     struct ocfs2_xattr_search *xs,
-                                     struct ocfs2_xattr_set_ctxt *ctxt)
-{
-        int ret, local = 1;
-        size_t value_len;
-        char *val = (char *)xi->value;
-        struct ocfs2_xattr_entry *xe = xs->here;
-        u32 name_hash = ocfs2_xattr_name_hash(inode, xi->name,
-                                              strlen(xi->name));
-
-        if (!xs->not_found && !ocfs2_xattr_is_local(xe)) {
-                /*
-                 * We need to truncate the xattr storage first.
-                 *
-                 * If both the old and new value are stored to
-                 * outside block, we only need to truncate
-                 * the storage and then set the value outside.
-                 *
-                 * If the new value should be stored within block,
-                 * we should free all the outside block first and
-                 * the modification to the xattr block will be done
-                 * by following steps.
-                 */
-                if (xi->value_len > OCFS2_XATTR_INLINE_SIZE)
-                        value_len = xi->value_len;
-                else
-                        value_len = 0;
-
-                ret = ocfs2_xattr_bucket_value_truncate_xs(inode, xs,
-                                                           value_len,
-                                                           ctxt);
-                if (ret)
-                        goto out;
-
-                if (value_len)
-                        goto set_value_outside;
-        }
-
-        value_len = xi->value_len;
-        /* So we have to handle the inside block change now. */
-        if (value_len > OCFS2_XATTR_INLINE_SIZE) {
-                /*
-                 * If the new value will be stored outside of block,
-                 * initalize a new empty value root and insert it first.
-                 */
-                local = 0;
-                xi->value = &def_xv;
-                xi->value_len = OCFS2_XATTR_ROOT_SIZE;
-        }
-
-        ret = ocfs2_xattr_set_entry_in_bucket(inode, ctxt->handle, xi, xs,
-                                              name_hash, local);
-        if (ret) {
-                mlog_errno(ret);
-                goto out;
-        }
-
-        if (value_len <= OCFS2_XATTR_INLINE_SIZE)
-                goto out;
-
-        /* allocate the space now for the outside block storage. */
-        ret = ocfs2_xattr_bucket_value_truncate_xs(inode, xs,
-                                                   value_len, ctxt);
-        if (ret) {
-                mlog_errno(ret);
-
-                if (xs->not_found) {
-                        /*
-                         * We can't allocate enough clusters for outside
-                         * storage and we have allocated xattr already,
-                         * so need to remove it.
-                         */
-                        ocfs2_xattr_bucket_remove_xs(inode, ctxt->handle, xs);
-                }
-                goto out;
-        }
-
-set_value_outside:
-        ret = ocfs2_xattr_bucket_set_value_outside(inode, ctxt->handle,
-                                                   xs, val, value_len);
-out:
-        return ret;
-}
-
 /*
  * check whether the xattr bucket is filled up with the same hash value.
  * If we want to insert the xattr with the same hash, return -ENOSPC.
@@ -5323,156 +5531,116 @@ static int ocfs2_check_xattr_bucket_collision(struct inode *inode,
         return 0;
 }
 
-static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
-                                             struct ocfs2_xattr_info *xi,
-                                             struct ocfs2_xattr_search *xs,
-                                             struct ocfs2_xattr_set_ctxt *ctxt)
+/*
+ * Try to set the entry in the current bucket.  If we fail, the caller
+ * will handle getting us another bucket.
+ */
+static int ocfs2_xattr_set_entry_bucket(struct inode *inode,
+                                        struct ocfs2_xattr_info *xi,
+                                        struct ocfs2_xattr_search *xs,
+                                        struct ocfs2_xattr_set_ctxt *ctxt)
 {
-        struct ocfs2_xattr_header *xh;
-        struct ocfs2_xattr_entry *xe;
-        u16 count, header_size, xh_free_start;
-        int free, max_free, need, old;
-        size_t value_size = 0, name_len = strlen(xi->name);
-        size_t blocksize = inode->i_sb->s_blocksize;
-        int ret, allocation = 0;
-
-        mlog_entry("Set xattr %s in xattr index block\n", xi->name);
-
-try_again:
-        xh = xs->header;
-        count = le16_to_cpu(xh->xh_count);
-        xh_free_start = le16_to_cpu(xh->xh_free_start);
-        header_size = sizeof(struct ocfs2_xattr_header) +
-                        count * sizeof(struct ocfs2_xattr_entry);
-        max_free = OCFS2_XATTR_BUCKET_SIZE - header_size -
-                le16_to_cpu(xh->xh_name_value_len) - OCFS2_XATTR_HEADER_GAP;
-
-        mlog_bug_on_msg(header_size > blocksize, "bucket %llu has header size "
-                        "of %u which exceed block size\n",
-                        (unsigned long long)bucket_blkno(xs->bucket),
-                        header_size);
+        int ret;
+        struct ocfs2_xa_loc loc;
 
-        if (xi->value && xi->value_len > OCFS2_XATTR_INLINE_SIZE)
-                value_size = OCFS2_XATTR_ROOT_SIZE;
-        else if (xi->value)
-                value_size = OCFS2_XATTR_SIZE(xi->value_len);
+        mlog_entry("Set xattr %s in xattr bucket\n", xi->xi_name);
 
-        if (xs->not_found)
-                need = sizeof(struct ocfs2_xattr_entry) +
-                        OCFS2_XATTR_SIZE(name_len) + value_size;
-        else {
-                need = value_size + OCFS2_XATTR_SIZE(name_len);
+        ocfs2_init_xattr_bucket_xa_loc(&loc, xs->bucket,
+                                       xs->not_found ? NULL : xs->here);
+        ret = ocfs2_xa_set(&loc, xi, ctxt);
+        if (!ret) {
+                xs->here = loc.xl_entry;
+                goto out;
+        }
+        if (ret != -ENOSPC) {
+                mlog_errno(ret);
+                goto out;
+        }
 
-                /*
-                 * We only replace the old value if the new length is smaller
-                 * than the old one. Otherwise we will allocate new space in the
-                 * bucket to store it.
-                 */
-                xe = xs->here;
-                if (ocfs2_xattr_is_local(xe))
-                        old = OCFS2_XATTR_SIZE(le64_to_cpu(xe->xe_value_size));
-                else
-                        old = OCFS2_XATTR_SIZE(OCFS2_XATTR_ROOT_SIZE);
+        /* Ok, we need space.  Let's try defragmenting the bucket. */
+        ret = ocfs2_defrag_xattr_bucket(inode, ctxt->handle,
+                                        xs->bucket);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
+        }
 
-                if (old >= value_size)
-                        need = 0;
+        ret = ocfs2_xa_set(&loc, xi, ctxt);
+        if (!ret) {
+                xs->here = loc.xl_entry;
+                goto out;
         }
+        if (ret != -ENOSPC)
+                mlog_errno(ret);
 
-        free = xh_free_start - header_size - OCFS2_XATTR_HEADER_GAP;
-        /*
-         * We need to make sure the new name/value pair
-         * can exist in the same block.
-         */
-        if (xh_free_start % blocksize < need)
-                free -= xh_free_start % blocksize;
-
-        mlog(0, "xs->not_found = %d, in xattr bucket %llu: free = %d, "
-             "need = %d, max_free = %d, xh_free_start = %u, xh_name_value_len ="
-             " %u\n", xs->not_found,
-             (unsigned long long)bucket_blkno(xs->bucket),
-             free, need, max_free, le16_to_cpu(xh->xh_free_start),
-             le16_to_cpu(xh->xh_name_value_len));
-
-        if (free < need ||
-            (xs->not_found &&
-             count == ocfs2_xattr_max_xe_in_bucket(inode->i_sb))) {
-                if (need <= max_free &&
-                    count < ocfs2_xattr_max_xe_in_bucket(inode->i_sb)) {
-                        /*
-                         * We can create the space by defragment. Since only the
-                         * name/value will be moved, the xe shouldn't be changed
-                         * in xs.
-                         */
-                        ret = ocfs2_defrag_xattr_bucket(inode, ctxt->handle,
-                                                        xs->bucket);
-                        if (ret) {
-                                mlog_errno(ret);
-                                goto out;
-                        }
 
-                        xh_free_start = le16_to_cpu(xh->xh_free_start);
-                        free = xh_free_start - header_size
-                                - OCFS2_XATTR_HEADER_GAP;
-                        if (xh_free_start % blocksize < need)
-                                free -= xh_free_start % blocksize;
+out:
+        mlog_exit(ret);
+        return ret;
+}
 
-                        if (free >= need)
-                                goto xattr_set;
+static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
+                                             struct ocfs2_xattr_info *xi,
+                                             struct ocfs2_xattr_search *xs,
+                                             struct ocfs2_xattr_set_ctxt *ctxt)
+{
+        int ret;
 
-                        mlog(0, "Can't get enough space for xattr insert by "
-                             "defragment. Need %u bytes, but we have %d, so "
-                             "allocate new bucket for it.\n", need, free);
-                }
+        mlog_entry("Set xattr %s in xattr index block\n", xi->xi_name);
 
-                /*
-                 * We have to add new buckets or clusters and one
-                 * allocation should leave us enough space for insert.
-                 */
-                BUG_ON(allocation);
+        ret = ocfs2_xattr_set_entry_bucket(inode, xi, xs, ctxt);
+        if (!ret)
+                goto out;
+        if (ret != -ENOSPC) {
+                mlog_errno(ret);
+                goto out;
+        }
 
-                /*
-                 * We do not allow for overlapping ranges between buckets. And
-                 * the maximum number of collisions we will allow for then is
-                 * one bucket's worth, so check it here whether we need to
-                 * add a new bucket for the insert.
-                 */
-                ret = ocfs2_check_xattr_bucket_collision(inode,
-                                                         xs->bucket,
-                                                         xi->name);
-                if (ret) {
-                        mlog_errno(ret);
-                        goto out;
-                }
+        /* Ack, need more space.  Let's try to get another bucket! */
 
-                ret = ocfs2_add_new_xattr_bucket(inode,
-                                                 xs->xattr_bh,
+        /*
+         * We do not allow for overlapping ranges between buckets. And
+         * the maximum number of collisions we will allow for then is
+         * one bucket's worth, so check it here whether we need to
+         * add a new bucket for the insert.
+         */
+        ret = ocfs2_check_xattr_bucket_collision(inode,
                                                  xs->bucket,
-                                                 ctxt);
-                if (ret) {
-                        mlog_errno(ret);
-                        goto out;
-                }
+                                                 xi->xi_name);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
+        }
 
-                /*
-                 * ocfs2_add_new_xattr_bucket() will have updated
-                 * xs->bucket if it moved, but it will not have updated
-                 * any of the other search fields.  Thus, we drop it and
-                 * re-search.  Everything should be cached, so it'll be
-                 * quick.
-                 */
-                ocfs2_xattr_bucket_relse(xs->bucket);
-                ret = ocfs2_xattr_index_block_find(inode, xs->xattr_bh,
-                                                   xi->name_index,
-                                                   xi->name, xs);
-                if (ret && ret != -ENODATA)
-                        goto out;
-                xs->not_found = ret;
-                allocation = 1;
-                goto try_again;
+        ret = ocfs2_add_new_xattr_bucket(inode,
+                                         xs->xattr_bh,
+                                         xs->bucket,
+                                         ctxt);
+        if (ret) {
+                mlog_errno(ret);
+                goto out;
         }
 
-xattr_set:
-        ret = ocfs2_xattr_set_in_bucket(inode, xi, xs, ctxt);
+        /*
+         * ocfs2_add_new_xattr_bucket() will have updated
+         * xs->bucket if it moved, but it will not have updated
+         * any of the other search fields.  Thus, we drop it and
+         * re-search.  Everything should be cached, so it'll be
+         * quick.
+         */
+        ocfs2_xattr_bucket_relse(xs->bucket);
+        ret = ocfs2_xattr_index_block_find(inode, xs->xattr_bh,
+                                           xi->xi_name_index,
+                                           xi->xi_name, xs);
+        if (ret && ret != -ENODATA)
+                goto out;
+        xs->not_found = ret;
+
+        /* Ok, we have a new bucket, let's try again */
+        ret = ocfs2_xattr_set_entry_bucket(inode, xi, xs, ctxt);
+        if (ret && (ret != -ENOSPC))
+                mlog_errno(ret);
+
 out:
         mlog_exit(ret);
         return ret;
@@ -5684,7 +5852,7 @@ static int ocfs2_prepare_refcount_xattr(struct inode *inode,
          * refcount tree, and make the original extent become 3. So we will need
          * 2 * cluster more extent recs at most.
          */
-        if (!xi->value || xi->value_len <= OCFS2_XATTR_INLINE_SIZE) {
+        if (!xi->xi_value || xi->xi_value_len <= OCFS2_XATTR_INLINE_SIZE) {
 
                 ret = ocfs2_refcounted_xattr_delete_need(inode,
                                                          &(*ref_tree)->rf_ci,
@@ -6066,7 +6234,7 @@ static int ocfs2_value_metas_in_xattr_header(struct super_block *sb,
                  * to the extent block, so just calculate a maximum record num.
                  */
                 if (!xv->xr_list.l_tree_depth)
-                        *num_recs += xv->xr_list.l_next_free_rec;
+                        *num_recs += le16_to_cpu(xv->xr_list.l_next_free_rec);
                 else
                         *num_recs += ocfs2_clusters_for_bytes(sb,
                                                               XATTR_SIZE_MAX);
@@ -6360,33 +6528,33 @@ static int ocfs2_create_empty_xattr_block(struct inode *inode,
                                           int indexed)
 {
         int ret;
-        handle_t *handle;
-        struct ocfs2_alloc_context *meta_ac;
         struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_xattr_set_ctxt ctxt;
 
-        ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
+        memset(&ctxt, 0, sizeof(ctxt));
+        ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &ctxt.meta_ac);
         if (ret < 0) {
                 mlog_errno(ret);
                 return ret;
         }
 
-        handle = ocfs2_start_trans(osb, OCFS2_XATTR_BLOCK_CREATE_CREDITS);
-        if (IS_ERR(handle)) {
-                ret = PTR_ERR(handle);
+        ctxt.handle = ocfs2_start_trans(osb, OCFS2_XATTR_BLOCK_CREATE_CREDITS);
+        if (IS_ERR(ctxt.handle)) {
+                ret = PTR_ERR(ctxt.handle);
                 mlog_errno(ret);
                 goto out;
         }
 
         mlog(0, "create new xattr block for inode %llu, index = %d\n",
              (unsigned long long)fe_bh->b_blocknr, indexed);
-        ret = ocfs2_create_xattr_block(handle, inode, fe_bh,
-                                       meta_ac, ret_bh, indexed);
+        ret = ocfs2_create_xattr_block(inode, fe_bh, &ctxt, indexed,
+                                       ret_bh);
         if (ret)
                 mlog_errno(ret);
 
-        ocfs2_commit_trans(osb, handle);
+        ocfs2_commit_trans(osb, ctxt.handle);
 out:
-        ocfs2_free_alloc_context(meta_ac);
+        ocfs2_free_alloc_context(ctxt.meta_ac);
         return ret;
 }
 
@@ -6978,9 +7146,9 @@ int ocfs2_init_security_and_acl(struct inode *dir,
 
         ret = ocfs2_init_security_get(inode, dir, &si);
         if (!ret) {
-                ret = ocfs2_xattr_security_set(inode, si.name,
-                                               si.value, si.value_len,
-                                               XATTR_CREATE);
+                ret = ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY,
+                                      si.name, si.value, si.value_len,
+                                      XATTR_CREATE);
                 if (ret) {
                         mlog_errno(ret);
                         goto leave;
@@ -7008,9 +7176,9 @@ leave:
 /*
  * 'security' attributes support
  */
-static size_t ocfs2_xattr_security_list(struct inode *inode, char *list,
+static size_t ocfs2_xattr_security_list(struct dentry *dentry, char *list,
                                         size_t list_size, const char *name,
-                                        size_t name_len)
+                                        size_t name_len, int type)
 {
         const size_t prefix_len = XATTR_SECURITY_PREFIX_LEN;
         const size_t total_len = prefix_len + name_len + 1;
@@ -7023,23 +7191,23 @@ static size_t ocfs2_xattr_security_list(struct inode *inode, char *list,
         return total_len;
 }
 
-static int ocfs2_xattr_security_get(struct inode *inode, const char *name,
-                                    void *buffer, size_t size)
+static int ocfs2_xattr_security_get(struct dentry *dentry, const char *name,
+                                    void *buffer, size_t size, int type)
 {
         if (strcmp(name, "") == 0)
                 return -EINVAL;
-        return ocfs2_xattr_get(inode, OCFS2_XATTR_INDEX_SECURITY, name,
-                               buffer, size);
+        return ocfs2_xattr_get(dentry->d_inode, OCFS2_XATTR_INDEX_SECURITY,
+                               name, buffer, size);
 }
 
-static int ocfs2_xattr_security_set(struct inode *inode, const char *name,
-                                    const void *value, size_t size, int flags)
+static int ocfs2_xattr_security_set(struct dentry *dentry, const char *name,
+                const void *value, size_t size, int flags, int type)
 {
         if (strcmp(name, "") == 0)
                 return -EINVAL;
 
-        return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, name, value,
-                               size, flags);
+        return ocfs2_xattr_set(dentry->d_inode, OCFS2_XATTR_INDEX_SECURITY,
+                               name, value, size, flags);
 }
 
 int ocfs2_init_security_get(struct inode *inode,
@@ -7076,9 +7244,9 @@ struct xattr_handler ocfs2_xattr_security_handler = {
 /*
  * 'trusted' attributes support
  */
-static size_t ocfs2_xattr_trusted_list(struct inode *inode, char *list,
+static size_t ocfs2_xattr_trusted_list(struct dentry *dentry, char *list,
                                        size_t list_size, const char *name,
-                                       size_t name_len)
+                                       size_t name_len, int type)
 {
         const size_t prefix_len = XATTR_TRUSTED_PREFIX_LEN;
         const size_t total_len = prefix_len + name_len + 1;
@@ -7091,23 +7259,23 @@ static size_t ocfs2_xattr_trusted_list(struct inode *inode, char *list,
         return total_len;
 }
 
-static int ocfs2_xattr_trusted_get(struct inode *inode, const char *name,
-                                   void *buffer, size_t size)
+static int ocfs2_xattr_trusted_get(struct dentry *dentry, const char *name,
+                void *buffer, size_t size, int type)
 {
         if (strcmp(name, "") == 0)
                 return -EINVAL;
-        return ocfs2_xattr_get(inode, OCFS2_XATTR_INDEX_TRUSTED, name,
-                               buffer, size);
+        return ocfs2_xattr_get(dentry->d_inode, OCFS2_XATTR_INDEX_TRUSTED,
+                               name, buffer, size);
 }
 
-static int ocfs2_xattr_trusted_set(struct inode *inode, const char *name,
-                                   const void *value, size_t size, int flags)
+static int ocfs2_xattr_trusted_set(struct dentry *dentry, const char *name,
+                const void *value, size_t size, int flags, int type)
 {
         if (strcmp(name, "") == 0)
                 return -EINVAL;
 
-        return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_TRUSTED, name, value,
-                               size, flags);
+        return ocfs2_xattr_set(dentry->d_inode, OCFS2_XATTR_INDEX_TRUSTED,
+                               name, value, size, flags);
 }
 
 struct xattr_handler ocfs2_xattr_trusted_handler = {
@@ -7120,13 +7288,13 @@ struct xattr_handler ocfs2_xattr_trusted_handler = {
 /*
  * 'user' attributes support
  */
-static size_t ocfs2_xattr_user_list(struct inode *inode, char *list,
+static size_t ocfs2_xattr_user_list(struct dentry *dentry, char *list,
                                     size_t list_size, const char *name,
-                                    size_t name_len)
+                                    size_t name_len, int type)
 {
         const size_t prefix_len = XATTR_USER_PREFIX_LEN;
         const size_t total_len = prefix_len + name_len + 1;
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
 
         if (osb->s_mount_opt & OCFS2_MOUNT_NOUSERXATTR)
                 return 0;
@@ -7139,31 +7307,31 @@ static size_t ocfs2_xattr_user_list(struct inode *inode, char *list,
         return total_len;
 }
 
-static int ocfs2_xattr_user_get(struct inode *inode, const char *name,
-                                void *buffer, size_t size)
+static int ocfs2_xattr_user_get(struct dentry *dentry, const char *name,
+                void *buffer, size_t size, int type)
 {
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
 
         if (strcmp(name, "") == 0)
                 return -EINVAL;
         if (osb->s_mount_opt & OCFS2_MOUNT_NOUSERXATTR)
                 return -EOPNOTSUPP;
-        return ocfs2_xattr_get(inode, OCFS2_XATTR_INDEX_USER, name,
+        return ocfs2_xattr_get(dentry->d_inode, OCFS2_XATTR_INDEX_USER, name,
                                buffer, size);
 }
 
-static int ocfs2_xattr_user_set(struct inode *inode, const char *name,
-                                const void *value, size_t size, int flags)
+static int ocfs2_xattr_user_set(struct dentry *dentry, const char *name,
+                const void *value, size_t size, int flags, int type)
 {
-        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+        struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
 
         if (strcmp(name, "") == 0)
                 return -EINVAL;
         if (osb->s_mount_opt & OCFS2_MOUNT_NOUSERXATTR)
                 return -EOPNOTSUPP;
 
-        return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_USER, name, value,
-                               size, flags);
+        return ocfs2_xattr_set(dentry->d_inode, OCFS2_XATTR_INDEX_USER,
+                               name, value, size, flags);
 }
 
 struct xattr_handler ocfs2_xattr_user_handler = {
diff --git a/fs/ocfs2/xattr.h b/fs/ocfs2/xattr.h
index 08e36389f56d..abd72a47f520 100644
--- a/fs/ocfs2/xattr.h
+++ b/fs/ocfs2/xattr.h
@@ -40,10 +40,8 @@ struct ocfs2_security_xattr_info {
 extern struct xattr_handler ocfs2_xattr_user_handler;
 extern struct xattr_handler ocfs2_xattr_trusted_handler;
 extern struct xattr_handler ocfs2_xattr_security_handler;
-#ifdef CONFIG_OCFS2_FS_POSIX_ACL
 extern struct xattr_handler ocfs2_xattr_acl_access_handler;
 extern struct xattr_handler ocfs2_xattr_acl_default_handler;
-#endif
 extern struct xattr_handler *ocfs2_xattr_handlers[];
 
 ssize_t ocfs2_listxattr(struct dentry *, char *, size_t);