ocfs2: fix race between crashed dio and rm

There is a race case between crashed dio and rm, which will lead to
OCFS2_VALID_FL not set read-only.

  N1                              N2
  ------------------------------------------------------------------------
  dd with direct flag
                                  rm file
  crashed with an dio entry left
  in orphan dir
                                  clear OCFS2_VALID_FL in
                                  ocfs2_remove_inode
                                  recover N1 and read the corrupted inode,
                                  and set filesystem read-only

So we skip the inode deletion this time and wait for dio entry recovered
first.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
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/inode.c b/fs/ocfs2/inode.c
index 7868f7e7c455bb60fac5e79c231c5b8314270ca3..fe4b3f7db245531f1f2be8c7e363c8a346a222c7 100644 (file)
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -971,6 +971,7 @@ static void ocfs2_delete_inode(struct inode *inode)
        int wipe, status;
        sigset_t oldset;
        struct buffer_head *di_bh = NULL;
+       struct ocfs2_dinode *di = NULL;
 
        trace_ocfs2_delete_inode(inode->i_ino,
                                 (unsigned long long)OCFS2_I(inode)->ip_blkno,
@@ -1025,6 +1026,14 @@ static void ocfs2_delete_inode(struct inode *inode)
                goto bail_unlock_nfs_sync;
        }
 
+       di = (struct ocfs2_dinode *)di_bh->b_data;
+       /* Skip inode deletion and wait for dio orphan entry recovered
+        * first */
+       if (unlikely(di->i_flags & cpu_to_le32(OCFS2_DIO_ORPHANED_FL))) {
+               ocfs2_cleanup_delete_inode(inode, 0);
+               goto bail_unlock_inode;
+       }
+
        /* Query the cluster. This will be the final decision made
         * before we go ahead and wipe the inode. */
        status = ocfs2_query_inode_wipe(inode, di_bh, &wipe);

diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 3bfd36a23e401b403e09aa8f5fbde2f3242f4011..52948af646b63ea29701f40f3f99f336bfef5717 100644 (file)
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -2210,7 +2210,9 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb,
                         * ocfs2_delete_inode. */
                        oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED;
                        spin_unlock(&oi->ip_lock);
-               } else if ((orphan_reco_type == ORPHAN_NEED_TRUNCATE) &&
+               }
+
+               if ((orphan_reco_type == ORPHAN_NEED_TRUNCATE) &&
                                (di->i_flags & cpu_to_le32(OCFS2_DIO_ORPHANED_FL))) {
                        ret = ocfs2_truncate_file(inode, di_bh,
                                        i_size_read(inode));