btrfs_drop_extent fixe for inline items > 8K
authorYan <yanzheng@21cn.com>
Wed, 30 Jan 2008 19:39:54 +0000 (14:39 -0500)
committerChris Mason <chris.mason@oracle.com>
Thu, 25 Sep 2008 15:04:00 +0000 (11:04 -0400)
When truncating a inline extent, btrfs_drop_extents doesn't properly
handle the case "key.offset > inline_limit". This bug can only happen
when max line size is larger than 8K.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
fs/btrfs/file.c

index 7c10a90..7cbf8d8 100644 (file)
@@ -591,8 +591,7 @@ next_slot:
                                }
                        }
                        bookend = 1;
-                       if (found_inline && start <= key.offset &&
-                           inline_limit < extent_end)
+                       if (found_inline && start <= key.offset)
                                keep = 1;
                }
                /* truncate existing extent */
@@ -672,11 +671,10 @@ next_slot:
                        if (!bookend)
                                continue;
                }
-               if (bookend && found_inline && start <= key.offset &&
-                   inline_limit < extent_end && key.offset <= inline_limit) {
+               if (bookend && found_inline && start <= key.offset) {
                        u32 new_size;
                        new_size = btrfs_file_extent_calc_inline_size(
-                                                  extent_end - inline_limit);
+                                                  extent_end - end);
                        btrfs_truncate_item(trans, root, path, new_size, 0);
                }
                /* create bookend, splitting the extent in two */