Unlock RCU before running another loop iteration

Signed-off-by: Sergey Temerkhanov <[email protected]>
---
 fs/file.c | 34 ++++++++++++++++++++--------------
 1 file changed, 20 insertions(+), 14 deletions(-)

diff --git a/fs/file.c b/fs/file.c
index 4559b5fec3bd..49d57752e7a6 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -823,21 +823,27 @@ static struct file *__fget_files(struct files_struct 
*files, unsigned int fd,
                                 fmode_t mask, unsigned int refs)
 {
        struct file *file;
+       bool again = false;
 
-       rcu_read_lock();
-loop:
-       file = fcheck_files(files, fd);
-       if (file) {
-               /* File object ref couldn't be taken.
-                * dup2() atomicity guarantee is the reason
-                * we loop to catch the new file (or NULL pointer)
-                */
-               if (file->f_mode & mask)
-                       file = NULL;
-               else if (!get_file_rcu_many(file, refs))
-                       goto loop;
-       }
-       rcu_read_unlock();
+       do {
+               rcu_read_lock();
+
+               file = fcheck_files(files, fd);
+               if (file) {
+                       /* File object ref couldn't be taken.
+                        * dup2() atomicity guarantee is the reason
+                        * we loop to catch the new file (or NULL pointer)
+                        */
+                       if (file->f_mode & mask)
+                               file = NULL;
+                       else if (!get_file_rcu_many(file, refs))
+                               again = true;
+               }
+               rcu_read_unlock();
+
+               if (unlikely(again))
+                       schedule();
+       } while (again);
 
        return file;
 }
-- 
2.25.1

Reply via email to