ocfs2: Pass xs->bucket into ocfs2_add_new_xattr_bucket().
Joel Becker [Wed, 26 Nov 2008 23:25:41 +0000 (15:25 -0800)]
Pass the actual target bucket for insert through to
ocfs2_add_new_xattr_bucket().  Now growing a bucket has no buffer_head
knowledge.

ocfs2_add_new_xattr_bucket() leavs xs->bucket in the proper state for
insert.  However, it doesn't update the rest of the search fields in xs,
so we still have to relse() and re-find.  That's OK, because everything
is cached.

Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

fs/ocfs2/xattr.c

index 5a5a1bd..dfc51c3 100644 (file)
@@ -4314,43 +4314,42 @@ out:
 }
 
 /*
- * Add new xattr bucket in an extent record and adjust the buckets accordingly.
- * xb_bh is the ocfs2_xattr_block.
- * We will move all the buckets starting from header_bh to the next place. As
- * for this one, half num of its xattrs will be moved to the next one.
+ * Add new xattr bucket in an extent record and adjust the buckets
+ * accordingly.  xb_bh is the ocfs2_xattr_block, and target is the
+ * bucket we want to insert into.
  *
- * We will allocate a new cluster if current cluster is full.  The
- * underlying calls will make sure that there is space at the target
- * bucket, shifting buckets around if necessary.  'target' may be updated
- * by those calls.
+ * In the easy case, we will move all the buckets after target down by
+ * one. Half of target's xattrs will be moved to the next bucket.
+ *
+ * If current cluster is full, we'll allocate a new one.  This may not
+ * be contiguous.  The underlying calls will make sure that there is
+ * space for the insert, shifting buckets around if necessary.
+ * 'target' may be moved by those calls.
  */
 static int ocfs2_add_new_xattr_bucket(struct inode *inode,
                                      struct buffer_head *xb_bh,
-                                     struct buffer_head *header_bh,
+                                     struct ocfs2_xattr_bucket *target,
                                      struct ocfs2_xattr_set_ctxt *ctxt)
 {
        struct ocfs2_xattr_block *xb =
                        (struct ocfs2_xattr_block *)xb_bh->b_data;
        struct ocfs2_xattr_tree_root *xb_root = &xb->xb_attrs.xb_root;
        struct ocfs2_extent_list *el = &xb_root->xt_list;
-       struct ocfs2_xattr_header *xh =
-                       (struct ocfs2_xattr_header *)header_bh->b_data;
-       u32 name_hash = le32_to_cpu(xh->xh_entries[0].xe_name_hash);
+       u32 name_hash =
+               le32_to_cpu(bucket_xh(target)->xh_entries[0].xe_name_hash);
        struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
        int ret, num_buckets, extend = 1;
        u64 p_blkno;
        u32 e_cpos, num_clusters;
        /* The bucket at the front of the extent */
-       struct ocfs2_xattr_bucket *first, *target;
+       struct ocfs2_xattr_bucket *first;
 
-       mlog(0, "Add new xattr bucket starting form %llu\n",
-            (unsigned long long)header_bh->b_blocknr);
+       mlog(0, "Add new xattr bucket starting from %llu\n",
+            (unsigned long long)bucket_blkno(target));
 
        /* The first bucket of the original extent */
        first = ocfs2_xattr_bucket_new(inode);
-       /* The target bucket for insert */
-       target = ocfs2_xattr_bucket_new(inode);
-       if (!first || !target) {
+       if (!first) {
                ret = -ENOMEM;
                mlog_errno(ret);
                goto out;
@@ -4369,12 +4368,6 @@ static int ocfs2_add_new_xattr_bucket(struct inode *inode,
                goto out;
        }
 
-       ret = ocfs2_read_xattr_bucket(target, header_bh->b_blocknr);
-       if (ret) {
-               mlog_errno(ret);
-               goto out;
-       }
-
        num_buckets = ocfs2_xattr_buckets_per_cluster(osb) * num_clusters;
        if (num_buckets == le16_to_cpu(bucket_xh(first)->xh_num_buckets)) {
                /*
@@ -4407,7 +4400,6 @@ static int ocfs2_add_new_xattr_bucket(struct inode *inode,
 
 out:
        ocfs2_xattr_bucket_free(first);
-       ocfs2_xattr_bucket_free(target);
 
        return ret;
 }
@@ -5083,15 +5075,21 @@ try_again:
 
                ret = ocfs2_add_new_xattr_bucket(inode,
                                                 xs->xattr_bh,
-                                                xs->bucket->bu_bhs[0],
+                                                xs->bucket,
                                                 ctxt);
                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);