Re: possible deadlock in ovl_dir_real_file
From: Amir Goldstein
Date: Mon Feb 01 2021 - 04:40:15 EST
On Mon, Feb 1, 2021 at 11:21 AM syzbot
<syzbot+6a023cb2262c79301432@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=148aef78d00000
> kernel config: https://syzkaller.appspot.com/x/.config?x=9408d1770a50819c
> dashboard link: https://syzkaller.appspot.com/bug?extid=6a023cb2262c79301432
> compiler: clang version 11.0.1
>
> Unfortunately, I don't have any reproducer for this issue yet.
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+6a023cb2262c79301432@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> ============================================
> WARNING: possible recursive locking detected
> 5.11.0-rc5-syzkaller #0 Not tainted
> --------------------------------------------
> syz-executor.2/3639 is trying to acquire lock:
> ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: inode_lock include/linux/fs.h:773 [inline]
> ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: ovl_dir_real_file+0x20b/0x310 fs/overlayfs/readdir.c:886
>
> but task is already holding lock:
> ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: inode_lock include/linux/fs.h:773 [inline]
> ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: ovl_ioctl_set_flags fs/overlayfs/file.c:530 [inline]
> ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: ovl_ioctl+0x2fb/0x960 fs/overlayfs/file.c:569
>
> other info that might help us debug this:
> Possible unsafe locking scenario:
>
> CPU0
> ----
> lock(&ovl_i_mutex_dir_key[depth]);
> lock(&ovl_i_mutex_dir_key[depth]);
>
> *** DEADLOCK ***
>
> May be due to missing lock nesting notation
>
> 2 locks held by syz-executor.2/3639:
> #0: ffff88807a706460 (sb_writers#17){.+.+}-{0:0}, at: mnt_want_write_file+0x5a/0x250 fs/namespace.c:412
> #1: ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: inode_lock include/linux/fs.h:773 [inline]
> #1: ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: ovl_ioctl_set_flags fs/overlayfs/file.c:530 [inline]
> #1: ffff888084c0b5f0 (&ovl_i_mutex_dir_key[depth]){++++}-{3:3}, at: ovl_ioctl+0x2fb/0x960 fs/overlayfs/file.c:569
>
> stack backtrace:
> CPU: 1 PID: 3639 Comm: syz-executor.2 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
> __lock_acquire+0x2333/0x5e90 kernel/locking/lockdep.c:4670
> lock_acquire+0x114/0x5e0 kernel/locking/lockdep.c:5442
> down_write+0x56/0x120 kernel/locking/rwsem.c:1406
> inode_lock include/linux/fs.h:773 [inline]
> ovl_dir_real_file+0x20b/0x310 fs/overlayfs/readdir.c:886
> ovl_real_fdget fs/overlayfs/file.c:136 [inline]
> ovl_real_ioctl fs/overlayfs/file.c:499 [inline]
> ovl_ioctl_set_flags fs/overlayfs/file.c:545 [inline]
> ovl_ioctl+0x4de/0x960 fs/overlayfs/file.c:569
> vfs_ioctl fs/ioctl.c:48 [inline]
> __do_sys_ioctl fs/ioctl.c:753 [inline]
> __se_sys_ioctl+0xfb/0x170 fs/ioctl.c:739
> do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
> entry_SYSCALL_64_after_hwframe+0x44/0xa9
> RIP: 0033:0x45e219
> Code: 0d b4 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 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 0f 83 db b3 fb ff c3 66 2e 0f 1f 84 00 00 00 00
> RSP: 002b:00007f02ed677c68 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
> RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 000000000045e219
> RDX: 0000000000000000 RSI: 0000000040086602 RDI: 0000000000000003
> RBP: 000000000119bfc0 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000246 R12: 000000000119bf8c
> R13: 00007ffd373df6ef R14: 00007f02ed6789c0 R15: 000000000119bf8c
>
>
#syz fix: ovl: avoid deadlock on directory ioctl
Should be in linux-next tree already.
Thanks,
Amir.