jbd2: fix infinite loop when recovering corrupt journal blocks
authorDarrick J. Wong <darrick.wong@oracle.com>
Wed, 27 Aug 2014 22:40:05 +0000 (18:40 -0400)
committerTheodore Ts'o <tytso@mit.edu>
Fri, 29 Aug 2014 02:22:28 +0000 (22:22 -0400)
When recovering the journal, don't fall into an infinite loop if we
encounter a corrupt journal block.  Instead, just skip the block and
return an error, which fails the mount and thus forces the user to run
a full filesystem fsck.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org
fs/jbd2/recovery.c

index 3b6bb19d60b17abceadec4c38df5d3570c0fcba9..00e9703d7dc6cf7c9d08bf36b9fffd3646ba0021 100644 (file)
@@ -426,6 +426,7 @@ static int do_one_pass(journal_t *journal,
        int                     tag_bytes = journal_tag_bytes(journal);
        __u32                   crc32_sum = ~0; /* Transactional Checksums */
        int                     descr_csum_size = 0;
+       int                     block_error = 0;
 
        /*
         * First thing is to establish what we expect to find in the log
@@ -598,7 +599,8 @@ static int do_one_pass(journal_t *journal,
                                                       "checksum recovering "
                                                       "block %llu in log\n",
                                                       blocknr);
-                                               continue;
+                                               block_error = 1;
+                                               goto skip_write;
                                        }
 
                                        /* Find a buffer for the new
@@ -797,7 +799,8 @@ static int do_one_pass(journal_t *journal,
                                success = -EIO;
                }
        }
-
+       if (block_error && success == 0)
+               success = -EIO;
        return success;
 
  failed: