]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
ext4: clean up inode bitmaps manipulation in ext4_free_inode
authorDmitry Monakhov <dmonakhov@openvz.org>
Sun, 16 May 2010 11:00:00 +0000 (07:00 -0400)
committerPaul Gortmaker <paul.gortmaker@windriver.com>
Thu, 6 Jan 2011 23:07:35 +0000 (18:07 -0500)
commit d17413c08cd2b1dd2bf2cfdbb0f7b736b2b2b15c upstream.

- Reorganize locking scheme to batch two atomic operation in to one.
  This also allow us to state what healthy group must obey following rule
  ext4_free_inodes_count(sb, gdp) == ext4_count_free(inode_bitmap, NUM);
- Fix possible undefined pointer dereference.
- Even if group descriptor stats aren't accessible we have to update
  inode bitmaps.
- Move non-group members update out of group_lock.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
fs/ext4/ialloc.c

index 57f6eef6ccd6c03afbd4c35d0fa364f4e350bbf8..52618d5a1773bf6b06b72cf5ff2f9c26478bde7d 100644 (file)
@@ -240,56 +240,49 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
        if (fatal)
                goto error_return;
 
-       /* Ok, now we can actually update the inode bitmaps.. */
-       cleared = ext4_clear_bit_atomic(ext4_group_lock_ptr(sb, block_group),
-                                       bit, bitmap_bh->b_data);
-       if (!cleared)
-               ext4_error(sb, "bit already cleared for inode %lu", ino);
-       else {
-               gdp = ext4_get_group_desc(sb, block_group, &bh2);
-
+       fatal = -ESRCH;
+       gdp = ext4_get_group_desc(sb, block_group, &bh2);
+       if (gdp) {
                BUFFER_TRACE(bh2, "get_write_access");
                fatal = ext4_journal_get_write_access(handle, bh2);
-               if (fatal) goto error_return;
-
-               if (gdp) {
-                       ext4_lock_group(sb, block_group);
-                       count = ext4_free_inodes_count(sb, gdp) + 1;
-                       ext4_free_inodes_set(sb, gdp, count);
-                       if (is_directory) {
-                               count = ext4_used_dirs_count(sb, gdp) - 1;
-                               ext4_used_dirs_set(sb, gdp, count);
-                               if (sbi->s_log_groups_per_flex) {
-                                       ext4_group_t f;
-
-                                       f = ext4_flex_group(sbi, block_group);
-                                       atomic_dec(&sbi->s_flex_groups[f].used_dirs);
-                               }
+       }
+       ext4_lock_group(sb, block_group);
+       cleared = ext4_clear_bit(bit, bitmap_bh->b_data);
+       if (fatal || !cleared) {
+               ext4_unlock_group(sb, block_group);
+               goto out;
+       }
 
-                       }
-                       gdp->bg_checksum = ext4_group_desc_csum(sbi,
-                                                       block_group, gdp);
-                       ext4_unlock_group(sb, block_group);
-                       percpu_counter_inc(&sbi->s_freeinodes_counter);
-                       if (is_directory)
-                               percpu_counter_dec(&sbi->s_dirs_counter);
-
-                       if (sbi->s_log_groups_per_flex) {
-                               ext4_group_t f;
-
-                               f = ext4_flex_group(sbi, block_group);
-                               atomic_inc(&sbi->s_flex_groups[f].free_inodes);
-                       }
-               }
-               BUFFER_TRACE(bh2, "call ext4_handle_dirty_metadata");
-               err = ext4_handle_dirty_metadata(handle, NULL, bh2);
-               if (!fatal) fatal = err;
+       count = ext4_free_inodes_count(sb, gdp) + 1;
+       ext4_free_inodes_set(sb, gdp, count);
+       if (is_directory) {
+               count = ext4_used_dirs_count(sb, gdp) - 1;
+               ext4_used_dirs_set(sb, gdp, count);
+               percpu_counter_dec(&sbi->s_dirs_counter);
        }
-       BUFFER_TRACE(bitmap_bh, "call ext4_handle_dirty_metadata");
-       err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
-       if (!fatal)
-               fatal = err;
-       sb->s_dirt = 1;
+       gdp->bg_checksum = ext4_group_desc_csum(sbi, block_group, gdp);
+       ext4_unlock_group(sb, block_group);
+
+       percpu_counter_inc(&sbi->s_freeinodes_counter);
+       if (sbi->s_log_groups_per_flex) {
+               ext4_group_t f = ext4_flex_group(sbi, block_group);
+
+               atomic_inc(&sbi->s_flex_groups[f].free_inodes);
+               if (is_directory)
+                       atomic_dec(&sbi->s_flex_groups[f].used_dirs);
+       }
+       BUFFER_TRACE(bh2, "call ext4_handle_dirty_metadata");
+       fatal = ext4_handle_dirty_metadata(handle, NULL, bh2);
+out:
+       if (cleared) {
+               BUFFER_TRACE(bitmap_bh, "call ext4_handle_dirty_metadata");
+               err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
+               if (!fatal)
+                       fatal = err;
+               sb->s_dirt = 1;
+       } else
+               ext4_error(sb, "bit already cleared for inode %lu", ino);
+
 error_return:
        brelse(bitmap_bh);
        ext4_std_error(sb, fatal);