FS#55456 - [netctl] 1.13-2 won't connect to wireless network
Attached to Project:
Arch Linux
Opened by Alin Porcic (onixion) - Friday, 01 September 2017, 14:18 GMT
Last edited by Jouke Witteveen (jouke) - Sunday, 08 October 2017, 07:53 GMT
Opened by Alin Porcic (onixion) - Friday, 01 September 2017, 14:18 GMT
Last edited by Jouke Witteveen (jouke) - Sunday, 08 October 2017, 07:53 GMT
|
Details
Description:
After upgrading the system the wireless netctl scripts stopped working. I first thought that the issue only affected the iwlwifi driver, but rt5370 (rt2800usb) (usb wireless adapter) also didn't work. After downgrading the netctl package to netctl 1.12-2 my scripts worked again. So something is up with the package netctl 1.13-2 (at least on my machine). Additional info: netctl 1.13-2 linux-firmware 20170622.7d2c913-1 Linux archlinux 4.9.46-1-lts #1 SMP Wed Aug 30 17:49:09 CEST 2017 x86_64 GNU/Linux Attached files: home.conf - Netctl config file for home hotspot.conf - Netctl config file for hotspot home.txt - Is the dmesg output after trying to connect with 'home.conf' (using iwlwifi driver) hotspot.txt - Is the journal output after trying to connect with 'hotspot.conf' (using rt2800usb driver) Steps to reproduce: (remove '.conf' ending and place config files in /etc/netctl/ folder) netctl start home netctl start hotspot |
This task depends upon
edit.: I also tried disabling IPv6 support with linux cmd, still not working.
netctl stop-all
pacman -U /var/cache/pacman/pkg/netctl-1.13-2-any.pkg.tar.xz
systemctl daemon-reload
(edited the config; see new_config.txt)
netctl reenable home
netctl start home
The I got this (see new.txt). No other packages where upgraded or downgraded.
When switching back to 1.12, I have to "ip link set wlp3s0 down" because 1.13 leaves the network interface "up".
Thanks for the great support :)
----
Description=test
Interface=test
Connection=dummy
IP=no
ExecUpPost='sleep 3'
WaitOnline=yes
NETCTL_DEBUG=yes
----
This should not really 'do' anything, other than waiting 3 seconds. If it also gets a timeout, than somehow systemd is configured with a very short TimeoutStartSec. If it does not crash, then probably the driver problems somehow convince systemd to trigger a timeout.
1.12 seems to work (see 1_12.txt) (note: I renamed the interface to 'test2').
DefaultTimeoutStartSec=1s
DefaultTimeoutStopSec=1s
When I uncomment this and make a 'daemon-reload', 1.13 works again ...
I just change it to 5 seconds or something.
Thanks Jouke for wasting your time helping me. You can close this bug. :D
Thanks for getting back with the root cause of the problem and your input in debugging this issue!