Re: [syzbot] KASAN: use-after-free Write in sk_psock_stop

[Cong Wang]

On Tue, Apr 6, 2021 at 6:01 AM syzbot
<syzbot+7b6548ae483d6f4c64ae@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
> ==================================================================
> BUG: KASAN: use-after-free in __lock_acquire+0x3e6f/0x54c0 kernel/locking/lockdep.c:4770
> Read of size 8 at addr ffff888024f66238 by task syz-executor.1/14202
>
> CPU: 0 PID: 14202 Comm: syz-executor.1 Not tainted 5.12.0-rc4-syzkaller #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
> Call Trace:
>  __dump_stack lib/dump_stack.c:79 [inline]
>  dump_stack+0x141/0x1d7 lib/dump_stack.c:120
>  print_address_description.constprop.0.cold+0x5b/0x2f8 mm/kasan/report.c:232
>  __kasan_report mm/kasan/report.c:399 [inline]
>  kasan_report.cold+0x7c/0xd8 mm/kasan/report.c:416
>  __lock_acquire+0x3e6f/0x54c0 kernel/locking/lockdep.c:4770
>  lock_acquire kernel/locking/lockdep.c:5510 [inline]
>  lock_acquire+0x1ab/0x740 kernel/locking/lockdep.c:5475
>  __raw_spin_lock_bh include/linux/spinlock_api_smp.h:135 [inline]
>  _raw_spin_lock_bh+0x2f/0x40 kernel/locking/spinlock.c:175
>  spin_lock_bh include/linux/spinlock.h:359 [inline]
>  sk_psock_stop+0x2f/0x4d0 net/core/skmsg.c:750
>  sock_map_close+0x172/0x390 net/core/sock_map.c:1534
>  inet_release+0x12e/0x280 net/ipv4/af_inet.c:431
>  __sock_release+0xcd/0x280 net/socket.c:599
>  sock_close+0x18/0x20 net/socket.c:1258
>  __fput+0x288/0x920 fs/file_table.c:280
>  task_work_run+0xdd/0x1a0 kernel/task_work.c:140
>  tracehook_notify_resume include/linux/tracehook.h:189 [inline]
>  exit_to_user_mode_loop kernel/entry/common.c:174 [inline]
>  exit_to_user_mode_prepare+0x249/0x250 kernel/entry/common.c:208
>  __syscall_exit_to_user_mode_work kernel/entry/common.c:290 [inline]
>  syscall_exit_to_user_mode+0x19/0x60 kernel/entry/common.c:301
>  entry_SYSCALL_64_after_hwframe+0x44/0xae
> RIP: 0033:0x466459
> Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 bc ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007f1bde3a3188 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
> RAX: 0000000000000000 RBX: 000000000056bf60 RCX: 0000000000466459
> RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000005
> RBP: 00000000004bf9fb R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000246 R12: 000000000056bf60
> R13: 00007ffe6eb13bbf R14: 00007f1bde3a3300 R15: 0000000000022000
[...]
> Second to last potentially related work creation:
>  kasan_save_stack+0x1b/0x40 mm/kasan/common.c:38
>  kasan_record_aux_stack+0xe5/0x110 mm/kasan/generic.c:345
>  __call_rcu kernel/rcu/tree.c:3039 [inline]
>  call_rcu+0xb1/0x740 kernel/rcu/tree.c:3114
>  queue_rcu_work+0x82/0xa0 kernel/workqueue.c:1753
>  sk_psock_put include/linux/skmsg.h:446 [inline]
>  sock_map_unref+0x109/0x190 net/core/sock_map.c:182
>  sock_hash_delete_from_link net/core/sock_map.c:918 [inline]
>  sock_map_unlink net/core/sock_map.c:1480 [inline]
>  sock_map_remove_links+0x389/0x530 net/core/sock_map.c:1492
>  sock_map_close+0x12f/0x390 net/core/sock_map.c:1532

Looks like the last refcnt can be gone before sk_psock_stop().

Technically, we can call sk_psock_stop() before
sock_map_remove_links(), the only barrier is the RCU read lock
there. Let me see if we can get rid of that RCU read lock.

Thanks.