[SOLVED] Netctl won't shut down on poweroff

Philbo__Baggins · 2017-10-01 03:12:25

Hello everyone,

I recently upgraded my Arch with pacman -Syu, which seems to have worked perfectly except for one thing. When I shut down the system, it tells me that a stop job is running for the netctl service. It keeps doing this and says it will timeout in 1 min 30 seconds, then it extends that to 3 minutes and keeps extending it. If I disable netctl with systemctl, it shuts down fine. I can manually start netctl with "sudo netctl start wlp3s0.etc," but when I try to stop it with "sudo netctl stop wlp3s0.etc" the command just hangs, and I get strange side effects like not being able to use sudo or start chrome or firefox. The same thing happens if I use netctl-auto. Even if I disable netctl and install and enable NetworkManager, then NetworkManager has a "stop job" that won't let my pc shut down.

This situation also causes problems when I close my laptop lid and it tries to suspend.
Any vague idea of what might be happening would be much appreciated, and I will gladly post any logs that might be of use.

Thanks,
Phil

Last edited by Philbo__Baggins (2017-10-01 22:51:02)

jasonwryan · 2017-10-01 03:24:50

Please paste the output of `netctl status wlp3s0.etc` and `hostnamectl`

Philbo__Baggins · 2017-10-01 03:30:01

Here is the output of netctl status wlp3s0.etc:

netctl@wlp3s0\x2dPITT\x2dGAMING.service - Networking for netctl profile wlp3s0-PITT-GAMING
   Loaded: loaded (/usr/lib/systemd/system/netctl@.service; static; vendor preset: disabled)
   Active: active (exited) since Sat 2017-09-30 23:21:43 UTC; 4min 50s ago
     Docs: man:netctl.profile(5)
  Process: 447 ExecStart=/usr/lib/netctl/network start wlp3s0-PITT-GAMING (code=exited, status=0/SUCCESS)
 Main PID: 447 (code=exited, status=0/SUCCESS)
   Status: "online"
    Tasks: 2 (limit: 4915)
   CGroup: /system.slice/system-netctl.slice/netctl@wlp3s0\x2dPITT\x2dGAMING.service
           ├─458 wpa_supplicant -q -B -P /run/wpa_supplicant-wlp3s0.pid -i wlp3s0 -D nl80211,wext -c/run/netctl/wpa_supplicant-wlp3s0.conf
           └─652 dhcpcd -4 -q -t 30 -L wlp3s0

And here is hostnamectl:

   Static hostname: Urist
         Icon name: computer-laptop
           Chassis: laptop
        Machine ID: bd313026084547bcbf7dc325f99eb4a1
           Boot ID: 334e1a081c134000ae615c677b5c9f9a
  Operating System: Arch Linux
            Kernel: Linux 4.13.3-1-ARCH
      Architecture: x86-64

seth · 2017-10-01 08:28:10

Try to connect manually, then unwind the connection and see whether/which step fails: https://wiki.archlinux.org/index.php/Wi … nual_setup

loqs · 2017-10-01 11:00:47

Also the output of journalctl -k from one of the affected shutdowns in case is it https://bugs.archlinux.org/task/55645

Philbo__Baggins · 2017-10-01 15:26:59

Ah, so when I look at journalctl, I get messages that look suspiciously like that bug that loqs suggested, with the kworker and mac80211. When I start netctl, it works, but gives me a strange PCIe bus error. Then, every 2 minutes, some kworker that has something to do with mac80211 seems to be getting blocked. I believe I am using kernel 4.13.3, so this is probably the bug that I am experiencing: https://bugs.archlinux.org/task/55645

Does it look like I am experiencing this bug? If so, I will now attempt to fix this with the PKGBUILD that Kevin Dodd supplies in the comments of that bug report, unless anyone has any other suggestions.
Many thanks to loqs and seth.

When I start netctl, I get this:

Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2: AER: Multiple Corrected error received: id=00ea
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2: PCIe Bus Error: severity=Corrected, type=Physical Layer, id=00ea(Receiver ID)
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2:   device [8086:9d1a] error status/mask=00000081/00002000
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2:    [ 0] Receiver Error         (First)
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2:    [ 7] Bad DLLP              
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2: AER: Multiple Corrected error received: id=00ea
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2: PCIe Bus Error: severity=Corrected, type=Physical Layer, id=00ea(Receiver ID)
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2:   device [8086:9d1a] error status/mask=00002041/00002000
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2:    [ 0] Receiver Error         (First)
Okt 01 11:12:25 Urist kernel: pcieport 0000:00:1d.2:    [ 6] Bad TLP               
Okt 01 11:12:25 Urist kernel: IPv6: ADDRCONF(NETDEV_UP): wlp3s0: link is not ready
Okt 01 11:12:30 Urist kernel: wlp3s0: authenticate with 00:a2:89:0f:21:4b
Okt 01 11:12:30 Urist kernel: wlp3s0: send auth to 00:a2:89:0f:21:4b (try 1/3)
Okt 01 11:12:30 Urist kernel: wlp3s0: authenticated
Okt 01 11:12:30 Urist kernel: wlp3s0: associate with 00:a2:89:0f:21:4b (try 1/3)
Okt 01 11:12:30 Urist kernel: wlp3s0: RX AssocResp from 00:a2:89:0f:21:4b (capab=0x101 status=0 aid=1)
Okt 01 11:12:30 Urist kernel: wlp3s0: associated
Okt 01 11:12:30 Urist kernel: IPv6: ADDRCONF(NETDEV_CHANGE): wlp3s0: link becomes ready
Okt 01 11:12:30 Urist kernel: wlp3s0: Limiting TX power to 24 (24 - 0) dBm as advertised by 00:a2:89:0f:21:4b

Then, every 2 minutes, I get this message:

Okt 01 11:14:56 Urist kernel: INFO: task kworker/u8:0:5 blocked for more than 120 seconds.
Okt 01 11:14:56 Urist kernel:       Tainted: G           O    4.13.3-1-ARCH #1
Okt 01 11:14:56 Urist kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Okt 01 11:14:56 Urist kernel: kworker/u8:0    D    0     5      2 0x00000000
Okt 01 11:14:56 Urist kernel: Workqueue: phy0 ieee80211_ba_session_work [mac80211]
Okt 01 11:14:56 Urist kernel: Call Trace:
Okt 01 11:14:56 Urist kernel:  __schedule+0x239/0x890
Okt 01 11:14:56 Urist kernel:  schedule+0x3d/0x90
Okt 01 11:14:56 Urist kernel:  schedule_preempt_disabled+0x15/0x20
Okt 01 11:14:56 Urist kernel:  __mutex_lock.isra.2+0x295/0x520
Okt 01 11:14:56 Urist kernel:  ? __slab_free+0x14c/0x2d0
Okt 01 11:14:56 Urist kernel:  __mutex_lock_slowpath+0x13/0x20
Okt 01 11:14:56 Urist kernel:  ? __mutex_lock_slowpath+0x13/0x20
Okt 01 11:14:56 Urist kernel:  mutex_lock+0x25/0x30
Okt 01 11:14:56 Urist kernel:  __ieee80211_start_rx_ba_session+0x153/0x520 [mac80211]
Okt 01 11:14:56 Urist kernel:  ? dequeue_entity+0xed/0x4d0
Okt 01 11:14:56 Urist kernel:  ? kfree+0x175/0x190
Okt 01 11:14:56 Urist kernel:  ieee80211_ba_session_work+0x164/0x250 [mac80211]
Okt 01 11:14:56 Urist kernel:  process_one_work+0x1de/0x430
Okt 01 11:14:56 Urist kernel:  worker_thread+0x47/0x3f0
Okt 01 11:14:56 Urist kernel:  kthread+0x125/0x140
Okt 01 11:14:56 Urist kernel:  ? process_one_work+0x430/0x430
Okt 01 11:14:56 Urist kernel:  ? kthread_create_on_node+0x70/0x70
Okt 01 11:14:56 Urist kernel:  ret_from_fork+0x25/0x30

seth · 2017-10-01 18:46:00

Does it look like I am experiencing this bug?

Yup - https://bugzilla.kernel.org/attachment.cgi?id=258351 shows the exact same blocks and call traces.

Philbo__Baggins · 2017-10-01 18:55:23

Is this kind of thing that would typically be fixed in the next release? Apparently it is still present in 4.13.4.

seth · 2017-10-01 19:19:37

https://bugzilla.kernel.org/show_bug.cgi?id=196927#c6
I assume it'll be in 4.14 but not in the 4.13 series (it's part of a larger commit series but I don't know about the policy here either.

You could install the lts kernel for the time being (so you won't have a partial upgrade about the kernel)

loqs · 2017-10-01 19:33:53

https://git.kernel.org/pub/scm/linux/ke … 08ae18c828
includes a cc to stable@vger.kernel.org see https://git.kernel.org/pub/scm/linux/ke … -rules.rst for the process that follows and reasons it may have been rejected.
see also https://git.kernel.org/pub/scm/linux/ke … .txt?#n145 as it is a networking patch.

Philbo__Baggins · 2017-10-01 22:52:32

It was indeed that bug, I used Arch Build System to rebuild the kernel with the patches included. Thanks everyone!

Arch Linux

#1 2017-10-01 03:12:25

[SOLVED] Netctl won't shut down on poweroff

#2 2017-10-01 03:24:50

Re: [SOLVED] Netctl won't shut down on poweroff

#3 2017-10-01 03:30:01

Re: [SOLVED] Netctl won't shut down on poweroff

#4 2017-10-01 08:28:10

Re: [SOLVED] Netctl won't shut down on poweroff

#5 2017-10-01 11:00:47

Re: [SOLVED] Netctl won't shut down on poweroff

#6 2017-10-01 15:26:59

Re: [SOLVED] Netctl won't shut down on poweroff

#7 2017-10-01 18:46:00

Re: [SOLVED] Netctl won't shut down on poweroff

#8 2017-10-01 18:55:23

Re: [SOLVED] Netctl won't shut down on poweroff

#9 2017-10-01 19:19:37

Re: [SOLVED] Netctl won't shut down on poweroff

#10 2017-10-01 19:33:53

Re: [SOLVED] Netctl won't shut down on poweroff

#11 2017-10-01 22:52:32

Re: [SOLVED] Netctl won't shut down on poweroff

Board footer