Tetsuo, can you look yet another lockdep message. This message occurs every time when I pair bluetooth headset. Is this message false positive too? [ 6829.212232] ============================================ [ 6829.212234] WARNING: possible recursive locking detected [ 6829.212236] 4.18.0-0.rc7.git1.1.fc29.x86_64 #1 Not tainted [ 6829.212237] -------------------------------------------- [ 6829.212239] kworker/u17:2/28441 is trying to acquire lock: [ 6829.212242] 000000004025b723 (sk_lock-AF_BLUETOOTH-BTPROTO_L2CAP){+.+.}, at: bt_accept_enqueue+0x3c/0xb0 [bluetooth] [ 6829.212260] but task is already holding lock: [ 6829.212262] 000000004cb71eef (sk_lock-AF_BLUETOOTH-BTPROTO_L2CAP){+.+.}, at: l2cap_sock_new_connection_cb+0x18/0xa0 [bluetooth] [ 6829.212278] other info that might help us debug this: [ 6829.212279] Possible unsafe locking scenario: [ 6829.212281] CPU0 [ 6829.212282] ---- [ 6829.212284] lock(sk_lock-AF_BLUETOOTH-BTPROTO_L2CAP); [ 6829.212286] lock(sk_lock-AF_BLUETOOTH-BTPROTO_L2CAP); [ 6829.212288] *** DEADLOCK *** [ 6829.212290] May be due to missing lock nesting notation [ 6829.212293] 5 locks held by kworker/u17:2/28441: [ 6829.212294] #0: 000000009af6a4dc ((wq_completion)"%s"hdev->name#2){+.+.}, at: process_one_work+0x1f3/0x650 [ 6829.212301] #1: 000000006f7488f4 ((work_completion)(&hdev->rx_work)){+.+.}, at: process_one_work+0x1f3/0x650 [ 6829.212306] #2: 000000003dba8333 (&conn->chan_lock){+.+.}, at: l2cap_connect+0x8f/0x5a0 [bluetooth] [ 6829.212321] #3: 00000000aaa813b9 (&chan->lock/2){+.+.}, at: l2cap_connect+0xa9/0x5a0 [bluetooth] [ 6829.212335] #4: 000000004cb71eef (sk_lock-AF_BLUETOOTH-BTPROTO_L2CAP){+.+.}, at: l2cap_sock_new_connection_cb+0x18/0xa0 [bluetooth] [ 6829.212350] stack backtrace: [ 6829.212354] CPU: 6 PID: 28441 Comm: kworker/u17:2 Not tainted 4.18.0-0.rc7.git1.1.fc29.x86_64 #1 [ 6829.212355] Hardware name: Gigabyte Technology Co., Ltd. Z87M-D3H/Z87M-D3H, BIOS F11 08/12/2014 [ 6829.212365] Workqueue: hci0 hci_rx_work [bluetooth] [ 6829.212367] Call Trace: [ 6829.212373] dump_stack+0x85/0xc0 [ 6829.212377] __lock_acquire.cold.64+0x158/0x227 [ 6829.212381] ? mark_held_locks+0x57/0x80 [ 6829.212384] lock_acquire+0x9e/0x1b0 [ 6829.212394] ? bt_accept_enqueue+0x3c/0xb0 [bluetooth] [ 6829.212398] lock_sock_nested+0x72/0xa0 [ 6829.212407] ? bt_accept_enqueue+0x3c/0xb0 [bluetooth] [ 6829.212417] bt_accept_enqueue+0x3c/0xb0 [bluetooth] [ 6829.212429] l2cap_sock_new_connection_cb+0x5d/0xa0 [bluetooth] [ 6829.212441] l2cap_connect+0x110/0x5a0 [bluetooth] [ 6829.212454] ? l2cap_recv_frame+0x6d0/0x2cb0 [bluetooth] [ 6829.212458] ? __mutex_unlock_slowpath+0x4b/0x2b0 [ 6829.212470] l2cap_recv_frame+0x6e8/0x2cb0 [bluetooth] [ 6829.212474] ? __mutex_unlock_slowpath+0x4b/0x2b0 [ 6829.212484] hci_rx_work+0x1c6/0x5d0 [bluetooth] [ 6829.212488] process_one_work+0x27d/0x650 [ 6829.212492] worker_thread+0x3c/0x390 [ 6829.212494] ? process_one_work+0x650/0x650 [ 6829.212498] kthread+0x120/0x140 [ 6829.212501] ? kthread_create_worker_on_cpu+0x70/0x70 [ 6829.212504] ret_from_fork+0x3a/0x50 [ 6829.285343] BUG: sleeping function called from invalid context at net/core/sock.c:2833 [ 6829.285349] in_atomic(): 1, irqs_disabled(): 0, pid: 1743, name: krfcommd [ 6829.285351] INFO: lockdep is turned off. [ 6829.285355] CPU: 6 PID: 1743 Comm: krfcommd Not tainted 4.18.0-0.rc7.git1.1.fc29.x86_64 #1 [ 6829.285358] Hardware name: Gigabyte Technology Co., Ltd. Z87M-D3H/Z87M-D3H, BIOS F11 08/12/2014 [ 6829.285360] Call Trace: [ 6829.285368] dump_stack+0x85/0xc0 [ 6829.285373] ___might_sleep.cold.72+0xac/0xbc [ 6829.285378] lock_sock_nested+0x29/0xa0 [ 6829.285394] bt_accept_enqueue+0x3c/0xb0 [bluetooth] [ 6829.285401] rfcomm_connect_ind+0x21b/0x260 [rfcomm] [ 6829.285406] rfcomm_run+0x1611/0x1820 [rfcomm] [ 6829.285411] ? do_wait_intr_irq+0xb0/0xb0 [ 6829.285416] ? rfcomm_check_accept+0x90/0x90 [rfcomm] [ 6829.285419] kthread+0x120/0x140 [ 6829.285422] ? kthread_create_worker_on_cpu+0x70/0x70 [ 6829.285426] ret_from_fork+0x3a/0x50 [ 6829.476282] input: 04:5D:4B:5F:34:57 as /devices/virtual/input/input35 [ 7273.090391] show_signal_msg: 23 callbacks suppressed [ 7273.090393] CFileWriterThre[29422]: segfault at 7f078bfe7240 ip 00007f079137843c sp 00007f078bb8dcf0 error 4 in steamclient.so[7f0790880000+14d2000] [ 7273.090404] Code: 89 df ff d2 8b 45 00 83 f8 02 0f 84 9e 00 00 00 83 f8 03 0f 84 55 05 00 00 83 f8 01 74 48 31 ed 4d 85 e4 74 11 48 85 db 74 0c <48> 8b 03 4c 89 e6 48 89 df ff 50 10 48 8b b4 24 e8 00 00 00 64 48 [ 7755.656023] rfkill: input handler enabled [ 7773.439895] rfkill: input handler disabled [ 8075.232946] BUG: sleeping function called from invalid context at net/core/sock.c:2833 [ 8075.232951] in_atomic(): 1, irqs_disabled(): 0, pid: 1743, name: krfcommd [ 8075.232952] INFO: lockdep is turned off. [ 8075.232956] CPU: 5 PID: 1743 Comm: krfcommd Tainted: G W 4.18.0-0.rc7.git1.1.fc29.x86_64 #1 [ 8075.232957] Hardware name: Gigabyte Technology Co., Ltd. Z87M-D3H/Z87M-D3H, BIOS F11 08/12/2014 [ 8075.232959] Call Trace: [ 8075.232965] dump_stack+0x85/0xc0 [ 8075.232969] ___might_sleep.cold.72+0xac/0xbc [ 8075.232973] lock_sock_nested+0x29/0xa0 [ 8075.232987] bt_accept_enqueue+0x3c/0xb0 [bluetooth] [ 8075.232992] rfcomm_connect_ind+0x21b/0x260 [rfcomm] [ 8075.232997] rfcomm_run+0x1611/0x1820 [rfcomm] [ 8075.233001] ? do_wait_intr_irq+0xb0/0xb0 [ 8075.233005] ? rfcomm_check_accept+0x90/0x90 [rfcomm] [ 8075.233008] kthread+0x120/0x140 [ 8075.233011] ? kthread_create_worker_on_cpu+0x70/0x70 [ 8075.233014] ret_from_fork+0x3a/0x50 [ 8075.413187] input: 04:5D:4B:5F:34:57 as /devices/virtual/input/input36 -- Best Regards, Mike Gavrilov. On Sat, 4 Aug 2018 at 15:18, Tetsuo Handa wrote: > > On 2018/08/04 17:18, Mikhail Gavrilov wrote: > > Hi guys I see this warning already several times when use latest 4.18 kernel. > > May be this is well know issue, I don't know. > > Hope this can fixed soon. > > https://marc.info/?l=linux-xfs&m=152954339303385 > > And since I can't afford fixing lockdep, we can't fix this soon...