Re: WARNING: suspicious RCU usage in kernfs_iop_permission

[Amir Goldstein]

On Mon, Feb 1, 2021 at 11:06 AM syzbot
<syzbot+0e507d08417ca2d565bf@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit:    6642d600 Merge tag '5.11-rc5-smb3' of git://git.samba.org/..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=171405bf500000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=9408d1770a50819c
> dashboard link: https://syzkaller.appspot.com/bug?extid=0e507d08417ca2d565bf
> compiler:       clang version 11.0.1
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=13b8a330d00000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=16f3628cd00000
>
> The issue was bisected to:
>
> commit 89bdfaf93d9157499c3a0d61f489df66f2dead7f
> Author: Miklos Szeredi <mszeredi@xxxxxxxxxx>
> Date:   Mon Dec 14 14:26:14 2020 +0000
>
>     ovl: make ioctl() safe
>
> bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=11b85fb4d00000
> final oops:     https://syzkaller.appspot.com/x/report.txt?x=13b85fb4d00000
> console output: https://syzkaller.appspot.com/x/log.txt?x=15b85fb4d00000
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+0e507d08417ca2d565bf@xxxxxxxxxxxxxxxxxxxxxxxxx
> Fixes: 89bdfaf93d91 ("ovl: make ioctl() safe")
>
> =============================
> WARNING: suspicious RCU usage
> 5.11.0-rc5-syzkaller #0 Not tainted
> -----------------------------
> kernel/sched/core.c:7932 Illegal context switch in RCU-sched read-side critical section!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 2, debug_locks = 0
> no locks held by systemd/1.
>
> stack backtrace:
> CPU: 0 PID: 1 Comm: systemd Not tainted 5.11.0-rc5-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+0x137/0x1be lib/dump_stack.c:120
>  ___might_sleep+0xb4/0x530 kernel/sched/core.c:7932
>  __mutex_lock_common+0x4e/0x2f00 kernel/locking/mutex.c:935
>  __mutex_lock kernel/locking/mutex.c:1103 [inline]
>  mutex_lock_nested+0x1a/0x20 kernel/locking/mutex.c:1118
>  kernfs_iop_permission+0x66/0x2f0 fs/kernfs/inode.c:284
>  do_inode_permission fs/namei.c:398 [inline]
>  inode_permission+0x234/0x4a0 fs/namei.c:463
>  may_lookup fs/namei.c:1575 [inline]
>  link_path_walk+0x226/0xc10 fs/namei.c:2128
>  path_openat+0x1f5/0x37a0 fs/namei.c:3367
>  do_filp_open+0x191/0x3a0 fs/namei.c:3398
>  do_sys_openat2+0xba/0x380 fs/open.c:1172
>  do_sys_open fs/open.c:1188 [inline]
>  __do_sys_open fs/open.c:1196 [inline]
>  __se_sys_open fs/open.c:1192 [inline]
>  __x64_sys_open+0x1af/0x1e0 fs/open.c:1192
>  do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
>  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> RIP: 0033:0x7fdaf5f1370d
> Code: 30 2c 00 00 75 10 b8 02 00 00 00 0f 05 48 3d 01 f0 ff ff 73 31 c3 48 83 ec 08 e8 fe 9d 01 00 48 89 04 24 b8 02 00 00 00 0f 05 <48> 8b 3c 24 48 89 c2 e8 47 9e 01 00 48 89 d0 48 83 c4 08 48 3d 01
> RSP: 002b:00007ffedccb56f0 EFLAGS: 00000293 ORIG_RAX: 0000000000000002
> RAX: ffffffffffffffda RBX: 000055b998e80590 RCX: 00007fdaf5f1370d
> RDX: 00000000000001b6 RSI: 0000000000080000 RDI: 00007ffedccb57d0
> RBP: 0000000000000008 R08: 0000000000000008 R09: 0000000000000001
> R10: 0000000000080000 R11: 0000000000000293 R12: 00007fdaf764d7b4
> R13: 0000000000000001 R14: 000055b998d5ad60 R15: 00007ffedccb57d0
>
>

The bisection result is questionable.
The repro does ioctl NS_GET_OWNER_UID which should not be hitting
any code affected by the bisected commit.

If anything I would also try:

#syz fix: ovl: avoid deadlock on directory ioctl

Perhaps a lock left help by some ioctl that happened before the repro has
caused this report.

Thanks,
Amir.