ext4: prevent fsreentrance deadlock for inline_data

ext4_da_convert_inline_data_to_extent() invokes
grab_cache_page_write_begin().  grab_cache_page_write_begin performs
memory allocation, so fs-reentrance should be prohibited because we
are inside journal transaction.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>

1 files changed, 6 insertions, 5 deletions

diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 3ea62695abce..efdcede40c22 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -870,6 +870,12 @@ retry_journal:
                         goto out_journal;
         }
 
+        /*
+         * We cannot recurse into the filesystem as the transaction
+         * is already started.
+         */
+        flags |= AOP_FLAG_NOFS;
+
         if (ret == -ENOSPC) {
                 ret = ext4_da_convert_inline_data_to_extent(mapping,
                                                             inode,
@@ -882,11 +888,6 @@ retry_journal:
                 goto out;
         }
 
-        /*
-         * We cannot recurse into the filesystem as the transaction
-         * is already started.
-         */
-        flags |= AOP_FLAG_NOFS;
 
         page = grab_cache_page_write_begin(mapping, 0, flags);
         if (!page) {