[v6.6] possible deadlock in smc_release (2)

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: eac8889a3a1c Linux 6.6.140
git tree: linux-6.6.y
console output: https://syzkaller.appspot.com/x/log.txt?x=14c9fa73980000
kernel config: https://syzkaller.appspot.com/x/.config?x=c5b35c4db8465904
dashboard link: https://syzkaller.appspot.com/bug?extid=df885f0959236be11223
compiler: Debian clang version 21.1.8 (++20251221033036+2078da43e25a-1~exp1~20251221153213.50), Debian LLD 21.1.8

Unfortunately, I don't have any reproducer for this issue yet.

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/284081ee4b1a/disk-eac8889a.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/6503c72409ee/vmlinux-eac8889a.xz
kernel image: https://storage.googleapis.com/syzbot-assets/5e529dbebdcb/bzImage-eac8889a.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+df885f...@syzkaller.appspotmail.com

======================================================
WARNING: possible circular locking dependency detected
syzkaller #0 Not tainted
------------------------------------------------------
syz.2.3071/17280 is trying to acquire lock:
ffff88807a861f98 ((work_completion)(&new_smc->smc_listen_work)){+.+.}-{0:0}, at: rcu_lock_acquire include/linux/rcupdate.h:334 [inline]
ffff88807a861f98 ((work_completion)(&new_smc->smc_listen_work)){+.+.}-{0:0}, at: rcu_read_lock include/linux/rcupdate.h:786 [inline]
ffff88807a861f98 ((work_completion)(&new_smc->smc_listen_work)){+.+.}-{0:0}, at: start_flush_work kernel/workqueue.c:3412 [inline]
ffff88807a861f98 ((work_completion)(&new_smc->smc_listen_work)){+.+.}-{0:0}, at: __flush_work+0x118/0xaa0 kernel/workqueue.c:3470

but task is already holding lock:
ffff88807a860c70 (sk_lock-AF_SMC/1){+.+.}-{0:0}, at: smc_release+0x255/0x530 net/smc/af_smc.c:-1

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (sk_lock-AF_SMC/1){+.+.}-{0:0}:
lock_sock_nested+0x48/0x100 net/core/sock.c:3535
smc_listen_out+0x109/0x3d0 net/smc/af_smc.c:1922
smc_listen_out_connected net/smc/af_smc.c:1942 [inline]
smc_listen_decline net/smc/af_smc.c:1977 [inline]
smc_listen_work+0x581/0xf40 net/smc/af_smc.c:2574
process_one_work kernel/workqueue.c:2653 [inline]
process_scheduled_works+0xa5d/0x15d0 kernel/workqueue.c:2730
worker_thread+0xa55/0xfc0 kernel/workqueue.c:2811
kthread+0x2fa/0x390 kernel/kthread.c:388
ret_from_fork+0x48/0x80 arch/x86/kernel/process.c:152
ret_from_fork_asm+0x11/0x20 arch/x86/entry/entry_64.S:293

-> #0 ((work_completion)(&new_smc->smc_listen_work)){+.+.}-{0:0}:
check_prev_add kernel/locking/lockdep.c:3134 [inline]
check_prevs_add kernel/locking/lockdep.c:3253 [inline]
validate_chain kernel/locking/lockdep.c:3869 [inline]
__lock_acquire+0x2df1/0x7d40 kernel/locking/lockdep.c:5137
lock_acquire+0x19e/0x420 kernel/locking/lockdep.c:5754
touch_work_lockdep_map kernel/workqueue.c:3186 [inline]
start_flush_work kernel/workqueue.c:3438 [inline]
__flush_work+0x64e/0xaa0 kernel/workqueue.c:3470
__cancel_work_timer+0x3f8/0x560 kernel/workqueue.c:3558
smc_clcsock_release+0x60/0xf0 net/smc/smc_close.c:29
__smc_release+0x661/0x7d0 net/smc/af_smc.c:311
smc_close_non_accepted+0xd5/0x1f0 net/smc/af_smc.c:1853
smc_close_cleanup_listen net/smc/smc_close.c:45 [inline]
smc_close_active+0xb61/0xf00 net/smc/smc_close.c:225
__smc_release+0x8d/0x7d0 net/smc/af_smc.c:287
smc_release+0x2ce/0x530 net/smc/af_smc.c:354
__sock_release net/socket.c:659 [inline]
sock_close+0xbd/0x230 net/socket.c:1420
__fput+0x234/0x970 fs/file_table.c:384
__do_sys_close fs/open.c:1573 [inline]
__se_sys_close+0x15f/0x220 fs/open.c:1558
do_syscall_x64 arch/x86/entry/common.c:46 [inline]
do_syscall_64+0x55/0xb0 arch/x86/entry/common.c:76
entry_SYSCALL_64_after_hwframe+0x68/0xd2

other info that might help us debug this:

Possible unsafe locking scenario:

CPU0 CPU1
---- ----
lock(sk_lock-AF_SMC/1);
lock((work_completion)(&new_smc->smc_listen_work));
lock(sk_lock-AF_SMC/1);
lock((work_completion)(&new_smc->smc_listen_work));

*** DEADLOCK ***

3 locks held by syz.2.3071/17280:
#0: ffff88805e189420 (&sb->s_type->i_mutex_key#11){+.+.}-{3:3}, at: inode_lock include/linux/fs.h:804 [inline]
#0: ffff88805e189420 (&sb->s_type->i_mutex_key#11){+.+.}-{3:3}, at: __sock_release net/socket.c:658 [inline]
#0: ffff88805e189420 (&sb->s_type->i_mutex_key#11){+.+.}-{3:3}, at: sock_close+0x9b/0x230 net/socket.c:1420
#1: ffff88807a860c70 (sk_lock-AF_SMC/1){+.+.}-{0:0}, at: smc_release+0x255/0x530 net/smc/af_smc.c:-1
#2: ffffffff8d132160 (rcu_read_lock){....}-{1:2}, at: rcu_lock_acquire include/linux/rcupdate.h:334 [inline]
#2: ffffffff8d132160 (rcu_read_lock){....}-{1:2}, at: rcu_read_lock include/linux/rcupdate.h:786 [inline]
#2: ffffffff8d132160 (rcu_read_lock){....}-{1:2}, at: start_flush_work kernel/workqueue.c:3412 [inline]
#2: ffffffff8d132160 (rcu_read_lock){....}-{1:2}, at: __flush_work+0x118/0xaa0 kernel/workqueue.c:3470

stack backtrace:
CPU: 0 PID: 17280 Comm: syz.2.3071 Not tainted syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 04/18/2026
Call Trace:
<TASK>
dump_stack_lvl+0x18c/0x250 lib/dump_stack.c:106
check_noncircular+0x2fc/0x400 kernel/locking/lockdep.c:2187
check_prev_add kernel/locking/lockdep.c:3134 [inline]
check_prevs_add kernel/locking/lockdep.c:3253 [inline]
validate_chain kernel/locking/lockdep.c:3869 [inline]
__lock_acquire+0x2df1/0x7d40 kernel/locking/lockdep.c:5137
lock_acquire+0x19e/0x420 kernel/locking/lockdep.c:5754
touch_work_lockdep_map kernel/workqueue.c:3186 [inline]
start_flush_work kernel/workqueue.c:3438 [inline]
__flush_work+0x64e/0xaa0 kernel/workqueue.c:3470
__cancel_work_timer+0x3f8/0x560 kernel/workqueue.c:3558
smc_clcsock_release+0x60/0xf0 net/smc/smc_close.c:29
__smc_release+0x661/0x7d0 net/smc/af_smc.c:311
smc_close_non_accepted+0xd5/0x1f0 net/smc/af_smc.c:1853
smc_close_cleanup_listen net/smc/smc_close.c:45 [inline]
smc_close_active+0xb61/0xf00 net/smc/smc_close.c:225
__smc_release+0x8d/0x7d0 net/smc/af_smc.c:287
smc_release+0x2ce/0x530 net/smc/af_smc.c:354
__sock_release net/socket.c:659 [inline]
sock_close+0xbd/0x230 net/socket.c:1420
__fput+0x234/0x970 fs/file_table.c:384
__do_sys_close fs/open.c:1573 [inline]
__se_sys_close+0x15f/0x220 fs/open.c:1558
do_syscall_x64 arch/x86/entry/common.c:46 [inline]
do_syscall_64+0x55/0xb0 arch/x86/entry/common.c:76
entry_SYSCALL_64_after_hwframe+0x68/0xd2
RIP: 0033:0x7fb4c2f9ce59
Code: ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 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 e8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fb4c3e66028 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
RAX: ffffffffffffffda RBX: 00007fb4c3215fa0 RCX: 00007fb4c2f9ce59
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000003
RBP: 00007fb4c3032d6f R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007fb4c3216038 R14: 00007fb4c3215fa0 R15: 00007ffe7e105a98
</TASK>


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

If the report is already addressed, let syzbot know by replying with:
#syz fix: exact-commit-title

If you want to overwrite report's subsystems, reply with:
#syz set subsystems: new-subsystem
(See the list of subsystem names on the web dashboard)

If the report is a duplicate of another one, reply with:
#syz dup: exact-subject-of-another-report

If you want to undo deduplication, reply with:
#syz undup