[PATCH mm-unstable v1] mm: multi-gen LRU: fix crash during cgroup migration

[PATCH mm-unstable v1] mm: multi-gen LRU: fix crash during cgroup migration

[Yu Zhao]

lru_gen_migrate_mm() assumes lru_gen_add_mm() runs prior to itself.
This isn't true for the following scenario:

    CPU 1                         CPU 2

  clone()
    cgroup_can_fork()
                                cgroup_procs_write()
    cgroup_post_fork()
                                  task_lock()
                                  lru_gen_migrate_mm()
                                  task_unlock()
    task_lock()
    lru_gen_add_mm()
    task_unlock()

And when the above happens, kernel crashes because of linked list
corruption (mm_struct->lru_gen.list).

Link: https://lore.kernel.org/r/20230115134651.30028-1-msizanoen@qtmlabs.xyz/
Reported-by: msizanoen <msizanoen@qtmlabs.xyz>
Tested-by: msizanoen <msizanoen@qtmlabs.xyz>
Fixes: bd74fdaea146 ("mm: multi-gen LRU: support page table walks")
Cc: stable@vger.kernel.org # v6.1+
Signed-off-by: Yu Zhao <yuzhao@google.com>
---
 mm/vmscan.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index cdf96aec39dc..394ff4962cbc 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3335,13 +3335,16 @@ void lru_gen_migrate_mm(struct mm_struct *mm)
 	if (mem_cgroup_disabled())
 		return;
 
+	/* migration can happen before addition */
+	if (!mm->lru_gen.memcg)
+		return;
+
 	rcu_read_lock();
 	memcg = mem_cgroup_from_task(task);
 	rcu_read_unlock();
 	if (memcg == mm->lru_gen.memcg)
 		return;
 
-	VM_WARN_ON_ONCE(!mm->lru_gen.memcg);
 	VM_WARN_ON_ONCE(list_empty(&mm->lru_gen.list));
 
 	lru_gen_del_mm(mm);
-- 
2.39.0.314.g84b9a713c41-goog