memcg: fix destination cgroup leak on task charges migration

We are supposed to take one css reference per each memory page and per
each swap entry accounted to a memory cgroup.  However, during task
charges migration we take a reference to the destination cgroup twice
per each swap entry: first in mem_cgroup_do_precharge()->try_charge()
and then in mem_cgroup_move_swap_account(), permanently leaking the
destination cgroup.

The hunk taking the second reference seems to be a leftover from the
pre-00501b531c472 ("mm: memcontrol: rewrite charge API") era.  Remove it
to fix the leak.

Fixes: e8ea14c (mm: memcontrol: take a css reference for each charged page)
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/memcontrol.c

@@ -3043,18 +3043,6 @@ static int mem_cgroup_move_swap_account(swp_entry_t entry,
 	if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) {
 		mem_cgroup_swap_statistics(from, false);
 		mem_cgroup_swap_statistics(to, true);
-		/*
-		 * This function is only called from task migration context now.
-		 * It postpones page_counter and refcount handling till the end
-		 * of task migration(mem_cgroup_clear_mc()) for performance
-		 * improvement. But we cannot postpone css_get(to)  because if
-		 * the process that has been moved to @to does swap-in, the
-		 * refcount of @to might be decreased to 0.
-		 *
-		 * We are in attach() phase, so the cgroup is guaranteed to be
-		 * alive, so we can just call css_get().
-		 */
-		css_get(&to->css);
 		return 0;
 	}
 	return -EINVAL;