ext4: correct the start block of counting reserved clusters
authorZhang Yi <yi.zhang@huawei.com>
Thu, 24 Aug 2023 09:26:04 +0000 (17:26 +0800)
committerTheodore Ts'o <tytso@mit.edu>
Fri, 6 Oct 2023 02:32:14 +0000 (22:32 -0400)
When big allocate feature is enabled, we need to count and update
reserved clusters before removing a delayed only extent_status entry.
{init|count|get}_rsvd() have already done this, but the start block
number of this counting isn't correct in the following case.

  lblk            end
   |               |
   v               v
          -------------------------
          |                       | orig_es
          -------------------------
                   ^              ^
      len1 is 0    |     len2     |

If the start block of the orig_es entry founded is bigger than lblk, we
passed lblk as start block to count_rsvd(), but the length is correct,
finally, the range to be counted is offset. This patch fix this by
passing the start blocks to 'orig_es->lblk + len1'.

Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Cc: stable@kernel.org
Link: https://lore.kernel.org/r/20230824092619.1327976-2-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Jan Kara <jack@suse.cz>
fs/ext4/extents_status.c

index 6f7de14c0fa86f00906f2ce01fbc4905d22fc2c6..5e625ea4545dc4276ec2cb955b8f048daf51929c 100644 (file)
@@ -1405,8 +1405,8 @@ static int __es_remove_extent(struct inode *inode, ext4_lblk_t lblk,
                        }
                }
                if (count_reserved)
-                       count_rsvd(inode, lblk, orig_es.es_len - len1 - len2,
-                                  &orig_es, &rc);
+                       count_rsvd(inode, orig_es.es_lblk + len1,
+                                  orig_es.es_len - len1 - len2, &orig_es, &rc);
                goto out_get_reserved;
        }