fs: push rcu_barrier() from deactivate_locked_super() to filesystems

There's no reason to call rcu_barrier() on every
deactivate_locked_super().  We only need to make sure that all delayed rcu
free inodes are flushed before we destroy related cache.

Removing rcu_barrier() from deactivate_locked_super() affects some fast
paths.  E.g.  on my machine exit_group() of a last process in IPC
namespace takes 0.07538s.  rcu_barrier() takes 0.05188s of that time.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

2 files changed, 11 insertions, 0 deletions

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 4c878476bb91..b08ea4717e9d 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -107,6 +107,12 @@ void extent_io_exit(void)
                 list_del(&eb->leak_list);
                 kmem_cache_free(extent_buffer_cache, eb);
         }
+
+        /*
+         * Make sure all delayed rcu free are flushed before we
+         * destroy caches.
+         */
+        rcu_barrier();
         if (extent_state_cache)
                 kmem_cache_destroy(extent_state_cache);
         if (extent_buffer_cache)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index ec154f954646..cf03a91d806f 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -7076,6 +7076,11 @@ static void init_once(void *foo)
 
 void btrfs_destroy_cachep(void)
 {
+        /*
+         * Make sure all delayed rcu free inodes are flushed before we
+         * destroy cache.
+         */
+        rcu_barrier();
         if (btrfs_inode_cachep)
                 kmem_cache_destroy(btrfs_inode_cachep);
         if (btrfs_trans_handle_cachep)