Very Computer

    This changeset fixes reiserfs transactions replay code that sometimes
    may pick incorrect stuff to replay. Patch by Chris Mason.
    This changeset seems to be doomed. It was first submitted on April 17th, 2002.
    (and I thought it was accepted, anyway I cannot find another reason on why it
    was never resubmitted if it was not accepted. It was sent with some other changes,
    and some of those other changes are now in the tree which is even more strange).

    Anyway, here is this resurrected changeset, please pull from
    bk://namesys.com/bk/reiser3-linux-2.4-journal-replay-fix

Diffstat:
 journal.c |   20 ++++++++------------
 1 files changed, 8 insertions(+), 12 deletions(-)

Plain text patch:

# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
#                  ChangeSet    1.1101  -> 1.1102
#       fs/reiserfs/journal.c   1.26    -> 1.27  
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------

# reiserfs: Fix for journal replay process, to only replay transactions from last mount. By Chris Mason.
# --------------------------------------------
#
diff -Nru a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c
--- a/fs/reiserfs/journal.c     Thu Apr 17 14:09:48 2003

 }
 static int journal_read(struct super_block *p_s_sb) {
   struct reiserfs_journal_desc *desc ;
-  unsigned long last_flush_trans_id = 0 ;
   unsigned long oldest_trans_id = 0;
   unsigned long oldest_invalid_trans_id = 0 ;
   time_t start ;
-  unsigned long last_flush_start = 0;
   unsigned long oldest_start = 0;
   unsigned long cur_dblock = 0 ;

   if (le32_to_cpu(jh->j_first_unflushed_offset) >= 0 &&
       le32_to_cpu(jh->j_first_unflushed_offset) < JOURNAL_BLOCK_COUNT &&
       le32_to_cpu(jh->j_last_flush_trans_id) > 0) {
-    last_flush_start = reiserfs_get_journal_block(p_s_sb) +
+    oldest_start = reiserfs_get_journal_block(p_s_sb) +
                        le32_to_cpu(jh->j_first_unflushed_offset) ;
-    last_flush_trans_id = le32_to_cpu(jh->j_last_flush_trans_id) ;
+    oldest_trans_id = le32_to_cpu(jh->j_last_flush_trans_id) + 1;
+    newest_mount_id = le32_to_cpu(jh->j_mount_id);
     reiserfs_debug(p_s_sb, REISERFS_DEBUG_CODE, "journal-1153: found in "
                    "header: first_unflushed_offset %d, last_flushed_trans_id "
                   "%lu\n", le32_to_cpu(jh->j_first_unflushed_offset),
-                  last_flush_trans_id) ;
+                  le32_to_cpu(jh->j_last_flush_trans_id)) ;
     valid_journal_header = 1 ;

       continue_replay = 0 ;
     }
     brelse(d_bh) ;
+    goto start_log_replay;
   }

                      "newest_mount_id to %d\n", le32_to_cpu(desc->j_mount_id));
       }
       cur_dblock += le32_to_cpu(desc->j_len) + 2 ;
-    }
-    else {
+    } else {
       cur_dblock++ ;
     }
     brelse(d_bh) ;
   }
-  /* step three, starting at the oldest transaction, replay */
-  if (last_flush_start > 0) {
-    oldest_start = last_flush_start ;
-    oldest_trans_id = last_flush_trans_id + 1 ;
-  }
+
+start_log_replay:
   cur_dblock = oldest_start ;
   if (oldest_trans_id)  {
     reiserfs_debug(p_s_sb, REISERFS_DEBUG_CODE, "journal-1206: Starting replay "
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in

More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/