ocfs2: do not BUG if buffer not uptodate in __ocfs2_journal_access

When storage network is unstable, it may trigger the BUG in
__ocfs2_journal_access because of buffer not uptodate.  We can retry the
write in this case or return error instead of BUG.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Reported-by: Zhangguanghui <zhang.guanghui@h3c.com>
Tested-by: Zhangguanghui <zhang.guanghui@h3c.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 5e56268..3bfd36a 100644 (file)
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -668,7 +668,23 @@ static int __ocfs2_journal_access(handle_t *handle,
                mlog(ML_ERROR, "giving me a buffer that's not uptodate!\n");
                mlog(ML_ERROR, "b_blocknr=%llu\n",
                     (unsigned long long)bh->b_blocknr);
-               BUG();
+
+               lock_buffer(bh);
+               /*
+                * A previous attempt to write this buffer head failed.
+                * Nothing we can do but to retry the write and hope for
+                * the best.
+                */
+               if (buffer_write_io_error(bh) && !buffer_uptodate(bh)) {
+                       clear_buffer_write_io_error(bh);
+                       set_buffer_uptodate(bh);
+               }
+
+               if (!buffer_uptodate(bh)) {
+                       unlock_buffer(bh);
+                       return -EIO;
+               }
+               unlock_buffer(bh);
        }
 
        /* Set the current transaction information on the ci so