[SOLVED] Update changed behavior of power switch (systemd update?)

Trilby · 2012-09-28 17:53:02

I have a binding for the power button of my netbook to run a dmenu script which gives me options to logout/reboot/shutdown/cancel. But after an update this is bypassed and a single press of the button immediately shuts the computer down.

I do not have acpid installed, nor do I want any other process managing this button. My window manager and scripts do what I want it to do.

I checked my pacman logs to look for suspects. The following packages were updated:

upgraded feh (2.6.1-1 -> 2.6.3-1)
upgraded libtasn1 (2.13-1 -> 2.14-1)
upgraded gnutls (3.1.1-1 -> 3.1.2-1)
upgraded imagemagick (6.7.9.4-1 -> 6.7.9.8-1)
upgraded midori (0.4.7-1 -> 0.4.7-2)
upgraded mkinitcpio-busybox (1.20.1-1 -> 1.20.2-1)
upgraded ncdu (1.8-1 -> 1.9-1)
warning: /etc/systemd/journald.conf installed as /etc/systemd/journald.conf.pacnew
upgraded systemd (189-4 -> 192-1)
upgraded systemd-sysvcompat (189-4 -> 192-1)
upgraded xkeyboard-config (2.6-1 -> 2.7-1)
upgraded youtube-dl (2012.02.27-1 -> 2012.09.27-1)

The only suspects seem to be xkeyboard-config and the two systemd packages. Investigating xkeyboard-config (pacman -Ql) suggests that it is only for internationalization and should not be relevant. systemd however owns "/usr/lib/udev/rules.d/70-power-switch.rules" which seems a likely cause of this behavior.

That rule file contains the following (plus some comments)

ACTION=="remove", GOTO="power_switch_end"

SUBSYSTEM=="input", KERNEL=="event*", SUBSYSTEMS=="acpi", TAG+="power-switch"
SUBSYSTEM=="input", KERNEL=="event*", KERNELS=="thinkpad_acpi", TAG+="power-switch"

LABEL="power_switch_end"

I've only dabbled with udev rules, and these are of a different form than those I've used. I'm not sure exactly what they do as there is no script nor executable called, they simply add a "TAG" to a list if I read them right. Would this cause the undesired behavior? If so, would deleting this file be advisable? Further, does anyone know why such a behavior was added to udev's defaults - this does not seem to be a useful change.

Alternately, is this seemingly relevant rule just a redherring and should I keep looking for other causes?

EDIT: additionally, I see that since udev became part of systemd (or whatever the appropriate terminology would be), an abundnce of other rule files have been added (40 in total) when I only had one or two before. I don't mean to gripe or complain, but I also don't like not knowing what is on my system or why it is there. If anyone could point me to the appropriate documentation or reading material to learn about if/why these udev rules are necessary or whether I can streamline them, I'd appreciate it.

Last edited by Trilby (2012-09-28 20:03:20)

Into the Pit · 2012-09-28 18:04:55

Maybe it's related to the changes of /etc/systemd/logind.conf?

nc · 2012-09-28 18:09:57

I had the same problem. Just change the Handle*Key options in /etc/systemd/logind.conf to ignore. wiki

Scimmia · 2012-09-28 18:10:09

Yep. atleast the 4th thread about it in the last 24 hours. Systemd 190+ will handle the power events itself unless you change it in logind.conf. I added a couple of notes in the wiki, hopefully it makes sense.

Last edited by Scimmia (2012-09-28 18:11:00)

Trilby · 2012-09-28 20:02:59

Thanks all, sorry if I missed other relevant threads - my google-fu must have been weak.

Arch Linux

#1 2012-09-28 17:53:02

[SOLVED] Update changed behavior of power switch (systemd update?)

#2 2012-09-28 18:04:55

Re: [SOLVED] Update changed behavior of power switch (systemd update?)

#3 2012-09-28 18:09:57

Re: [SOLVED] Update changed behavior of power switch (systemd update?)

#4 2012-09-28 18:10:09

Re: [SOLVED] Update changed behavior of power switch (systemd update?)

#5 2012-09-28 20:02:59

Re: [SOLVED] Update changed behavior of power switch (systemd update?)

Board footer