[PATCH 5.14 756/849] mm/zsmalloc.c: close race window between zs_pool_dec_isolated() and zs_unregister_migration()

From: Greg Kroah-Hartman
Date: Mon Nov 15 2021 - 20:16:25 EST


From: Miaohe Lin <linmiaohe@xxxxxxxxxx>

[ Upstream commit afe8605ca45424629fdddfd85984b442c763dc47 ]

There is one possible race window between zs_pool_dec_isolated() and
zs_unregister_migration() because wait_for_isolated_drain() checks the
isolated count without holding class->lock and there is no order inside
zs_pool_dec_isolated(). Thus the below race window could be possible:

zs_pool_dec_isolated zs_unregister_migration
check pool->destroying != 0
pool->destroying = true;
smp_mb();
wait_for_isolated_drain()
wait for pool->isolated_pages == 0
atomic_long_dec(&pool->isolated_pages);
atomic_long_read(&pool->isolated_pages) == 0

Since we observe the pool->destroying (false) before atomic_long_dec()
for pool->isolated_pages, waking pool->migration_wait up is missed.

Fix this by ensure checking pool->destroying happens after the
atomic_long_dec(&pool->isolated_pages).

Link: https://lkml.kernel.org/r/20210708115027.7557-1-linmiaohe@xxxxxxxxxx
Fixes: 701d678599d0 ("mm/zsmalloc.c: fix race condition in zs_destroy_pool")
Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
Cc: Minchan Kim <minchan@xxxxxxxxxx>
Cc: Sergey Senozhatsky <senozhatsky@xxxxxxxxxxxx>
Cc: Henry Burns <henryburns@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
mm/zsmalloc.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c
index 68e8831068f4b..b897ce3b399a1 100644
--- a/mm/zsmalloc.c
+++ b/mm/zsmalloc.c
@@ -1830,10 +1830,11 @@ static inline void zs_pool_dec_isolated(struct zs_pool *pool)
VM_BUG_ON(atomic_long_read(&pool->isolated_pages) <= 0);
atomic_long_dec(&pool->isolated_pages);
/*
- * There's no possibility of racing, since wait_for_isolated_drain()
- * checks the isolated count under &class->lock after enqueuing
- * on migration_wait.
+ * Checking pool->destroying must happen after atomic_long_dec()
+ * for pool->isolated_pages above. Paired with the smp_mb() in
+ * zs_unregister_migration().
*/
+ smp_mb__after_atomic();
if (atomic_long_read(&pool->isolated_pages) == 0 && pool->destroying)
wake_up_all(&pool->migration_wait);
}
--
2.33.0