vfs: make evict() use smp_mb__after_spinlock instead of smp_mb

It literally directly follows a spin_lock() call.

This whacks an explicit barrier on x86-64.

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/r/20241113155103.4194099-1-mjguzik@gmail.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>

fs/inode.c

@@ -746,7 +746,7 @@ static void evict(struct inode *inode)
 	 * ___wait_var_event() either sees the bit cleared or
 	 * waitqueue_active() check in wake_up_var() sees the waiter.
 	 */
-	smp_mb();
+	smp_mb__after_spinlock();
 	inode_wake_up_bit(inode, __I_NEW);
 	BUG_ON(inode->i_state != (I_FREEING | I_CLEAR));
 	spin_unlock(&inode->i_lock);