[ 1390.785437] ======================================================
[ 1390.791654] WARNING: possible circular locking dependency detected
[ 1390.797874] 6.0.0-rc7 #1 Tainted: G        W  OE     
[ 1390.802958] ------------------------------------------------------
[ 1390.809177] kworker/7:4/13932 is trying to acquire lock:
[ 1390.814524] ffff8881775abeb0 (sk_lock-AF_INET){+.+.}-{0:0}, at: tcp_sendpage+0x23/0x90
[ 1390.822513] 
               but task is already holding lock:
[ 1390.828381] ffff8881273b8b60 (&psock->work_mutex){+.+.}-{3:3}, at: sk_psock_backlog+0x2c/0x630
[ 1390.837061] 
               which lock already depends on the new lock.

[ 1390.845287] 
               the existing dependency chain (in reverse order) is:
[ 1390.852813] 
               -> #2 (&psock->work_mutex){+.+.}-{3:3}:
[ 1390.859217]        __mutex_lock+0x12b/0xcf0
[ 1390.863433]        sk_psock_backlog+0x2c/0x630
[ 1390.867906]        process_one_work+0x53e/0x9b0
[ 1390.872470]        worker_thread+0x8f/0x630
[ 1390.876683]        kthread+0x17b/0x1b0
[ 1390.880460]        ret_from_fork+0x1f/0x30
[ 1390.884586] 
               -> #1 ((work_completion)(&psock->work)){+.+.}-{0:0}:
[ 1390.892126]        __flush_work+0xfb/0x6b0
[ 1390.896251]        __cancel_work_timer+0x21b/0x2d0
[ 1390.901074]        sock_map_close+0x1a5/0x4b0
[ 1390.905460]        inet_release+0x8c/0xf0
[ 1390.909498]        __sock_release+0x6f/0x130
[ 1390.913797]        sock_close+0x14/0x20
[ 1390.917660]        __fput+0x11a/0x400
[ 1390.921350]        task_work_run+0x88/0xe0
[ 1390.925477]        exit_to_user_mode_prepare+0x22c/0x250
[ 1390.930831]        syscall_exit_to_user_mode+0x16/0x50
[ 1390.936007]        do_syscall_64+0x67/0x80
[ 1390.940134]        entry_SYSCALL_64_after_hwframe+0x63/0xcd
[ 1390.945747] 
               -> #0 (sk_lock-AF_INET){+.+.}-{0:0}:
[ 1390.951890]        __lock_acquire+0x1a82/0x2e10
[ 1390.956453]        lock_acquire+0x16f/0x410
[ 1390.960666]        lock_sock_nested+0x37/0x90
[ 1390.965055]        tcp_sendpage+0x23/0x90
[ 1390.969094]        inet_sendpage+0x6f/0xc0
[ 1390.973218]        kernel_sendpage+0xdd/0x220
[ 1390.977606]        __skb_send_sock+0x33f/0x540
[ 1390.982083]        sk_psock_backlog+0x116/0x630
[ 1390.986645]        process_one_work+0x53e/0x9b0
[ 1390.991207]        worker_thread+0x8f/0x630
[ 1390.995422]        kthread+0x17b/0x1b0
[ 1390.999200]        ret_from_fork+0x1f/0x30
[ 1391.003329] 
               other info that might help us debug this:

[ 1391.011382] Chain exists of:
                 sk_lock-AF_INET --> (work_completion)(&psock->work) --> &psock->work_mutex

[ 1391.023908]  Possible unsafe locking scenario:

[ 1391.029863]        CPU0                    CPU1
[ 1391.034420]        ----                    ----
[ 1391.038978]   lock(&psock->work_mutex);
[ 1391.042842]                                lock((work_completion)(&psock->work));
[ 1391.050374]                                lock(&psock->work_mutex);
[ 1391.056772]   lock(sk_lock-AF_INET);
[ 1391.060375] 
                *** DEADLOCK ***

[ 1391.066330] 3 locks held by kworker/7:4/13932:
[ 1391.070802]  #0: ffff88810007e338 ((wq_completion)events){+.+.}-{0:0}, at: process_one_work+0x46a/0x9b0
[ 1391.080274]  #1: ffff888115f1fde8 ((work_completion)(&psock->work)){+.+.}-{0:0}, at: process_one_work+0x46a/0x9b0
[ 1391.090618]  #2: ffff8881273b8b60 (&psock->work_mutex){+.+.}-{3:3}, at: sk_psock_backlog+0x2c/0x630
[ 1391.099739] 
               stack backtrace:
[ 1391.104124] CPU: 7 PID: 13932 Comm: kworker/7:4 Tainted: G        W  OE      6.0.0-rc7 #1
[ 1391.112358] Hardware name: HPE ProLiant DL360 Gen10/ProLiant DL360 Gen10, BIOS U32 03/08/2022
[ 1391.120936] Workqueue: events sk_psock_backlog
[ 1391.125415] Call Trace:
[ 1391.127878]  <TASK>
[ 1391.129992]  dump_stack_lvl+0x5b/0x77
[ 1391.133688]  check_noncircular+0x1f2/0x230
[ 1391.137816]  ? print_circular_bug+0x110/0x110
[ 1391.142207]  ? preempt_count_sub+0x15/0x100
[ 1391.146425]  ? unwind_next_frame+0x9f3/0xbc0
[ 1391.150727]  ? __orc_find+0x6c/0xd0
[ 1391.154247]  ? lockdep_lock+0xb8/0x170
[ 1391.158028]  ? call_rcu_zapped+0x90/0x90
[ 1391.161983]  ? add_chain_block+0x3b5/0x3e0
[ 1391.166117]  __lock_acquire+0x1a82/0x2e10
[ 1391.170171]  ? lockdep_hardirqs_on_prepare+0x230/0x230
[ 1391.175360]  lock_acquire+0x16f/0x410
[ 1391.179050]  ? tcp_sendpage+0x23/0x90
[ 1391.182747]  ? lock_downgrade+0x3c0/0x3c0
[ 1391.186790]  ? lockdep_hardirqs_on+0x7d/0x100
[ 1391.191188]  ? lockdep_lock+0xb8/0x170
[ 1391.194968]  ? __list_add_valid+0x3f/0x70
[ 1391.199015]  ? add_lock_to_list+0x129/0x170
[ 1391.203239]  lock_sock_nested+0x37/0x90
[ 1391.207105]  ? tcp_sendpage+0x23/0x90
[ 1391.210799]  tcp_sendpage+0x23/0x90
[ 1391.214320]  inet_sendpage+0x6f/0xc0
[ 1391.217928]  kernel_sendpage+0xdd/0x220
[ 1391.221802]  __skb_send_sock+0x33f/0x540
[ 1391.225758]  ? sendmsg_unlocked+0x60/0x60
[ 1391.229799]  ? sock_dequeue_err_skb+0x1b0/0x1b0
[ 1391.234375]  ? sendpage_unlocked+0x60/0x60
[ 1391.238504]  ? lock_release+0x233/0x470
[ 1391.242370]  ? skb_dequeue+0x92/0xb0
[ 1391.245973]  ? reacquire_held_locks+0x270/0x270
[ 1391.250542]  ? mark_held_locks+0x65/0x90
[ 1391.254502]  ? preempt_count_sub+0xb7/0x100
[ 1391.258719]  ? _raw_spin_unlock_irqrestore+0x40/0x60
[ 1391.263727]  sk_psock_backlog+0x116/0x630
[ 1391.267769]  ? lock_is_held_type+0xe3/0x140
[ 1391.271993]  process_one_work+0x53e/0x9b0
[ 1391.276043]  ? pwq_dec_nr_in_flight+0x100/0x100
[ 1391.280613]  ? rwlock_bug.part.0+0x60/0x60
[ 1391.284751]  worker_thread+0x8f/0x630
[ 1391.288454]  ? process_one_work+0x9b0/0x9b0
[ 1391.292670]  kthread+0x17b/0x1b0
[ 1391.295926]  ? kthread_complete_and_exit+0x20/0x20
[ 1391.300758]  ret_from_fork+0x1f/0x30
[ 1391.304378]  </TASK>