ext4: add action of moving index in ext4_ext_rm_idx for Punch Hole
authorRobin Dong <sanbai@taobao.com>
Thu, 28 Jul 2011 01:29:33 +0000 (21:29 -0400)
committerTheodore Ts'o <tytso@mit.edu>
Thu, 28 Jul 2011 01:29:33 +0000 (21:29 -0400)
The old function ext4_ext_rm_idx is used only for truncate case
because it just remove last index in extent-index-block. When punching
hole, it usually needed to remove "middle" index, therefore we must
move indexes which after it forward.

(I create a file with 1 depth extent tree and punch hole in the middle
of it, the last index in index-block strangly gone, so I find out this
bug)

Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
fs/ext4/extents.c

index 4d73e11ae883da6b73808d610ebf1869f64abc4a..a25bbdc7d493896b19ea4d0748b3397b7b97fb99 100644 (file)
@@ -2101,8 +2101,6 @@ ext4_ext_in_cache(struct inode *inode, ext4_lblk_t block,
 /*
  * ext4_ext_rm_idx:
  * removes index from the index block.
- * It's used in truncate case only, thus all requests are for
- * last index in the block only.
  */
 static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
                        struct ext4_ext_path *path)
@@ -2120,6 +2118,13 @@ static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
        err = ext4_ext_get_access(handle, inode, path);
        if (err)
                return err;
+
+       if (path->p_idx != EXT_LAST_INDEX(path->p_hdr)) {
+               int len = EXT_LAST_INDEX(path->p_hdr) - path->p_idx;
+               len *= sizeof(struct ext4_extent_idx);
+               memmove(path->p_idx, path->p_idx + 1, len);
+       }
+
        le16_add_cpu(&path->p_hdr->eh_entries, -1);
        err = ext4_ext_dirty(handle, inode, path);
        if (err)