Bug 517740 – x11-misc/bumblebee - init script should rmmod bbswitch

• Product: Gentoo Linux
• Status: RESOLVED WORKSFORME

[Description Adam Feldman 2014-07-22 05:53:49 UTC]

Halting the system works without issue, however, when attempting to reboot the system, after input to the screen is terminated, the system hangs in that state until forcible power-down.
After much experimentation, I found that the issue is the bbswitch module, provided by sys-power/bbswitch. For some reason, "modprobe -R bbswitch" fails to unload the module, however, "rmmod bbswitch" works without fail.
As such, my recommendation for fixing this issue is to add a line to the bumblebee init script to rmmod bbswitch, fixing this weird issue for which I can't seem to get any real debugging information for.  Attached is a sample of what I have on my system.

Reproducible: Always

[Comment 1 Adam Feldman 2014-07-22 05:55:01 UTC]

Created attachment 381318 [details]
files/bumblebee.initd

Appends rmmod of bbswitch kernel module to the init script.

[Comment 2 Jeroen Roovers (RETIRED) 2014-07-22 20:18:19 UTC]

Comment on attachment 381318 [details]
files/bumblebee.initd

--- files/bumblebee.initd       2012-06-05 15:47:49.000000000 +0200
+++ -   2014-07-22 22:17:48.398035868 +0200
@@ -20,4 +20,7 @@
        ebegin "Stopping BumbleBee Daemon"
                start-stop-daemon -K -p "${PIDFILE}" -R SIGTERM/10
        eend $?
+       ebegin "Removing bbswitch module"
+               rmmod bbswitch
+       eend $?
 }

[Comment 3 Pacho Ramos 2014-07-23 07:59:51 UTC]

Do you have parallel build enabled in openRC? Maybe it's trying to kill something prematurely. We should also get the proper errors instead of adding rmmod to random places to force module unloading :/ @kernel team, do you know how could the reporter try to get more information?

[Comment 4 Adam Feldman 2014-07-24 04:03:10 UTC]

(In reply to Pacho Ramos from comment #3)
> Do you have parallel build enabled in openRC? Maybe it's trying to kill
> something prematurely. We should also get the proper errors instead of
> adding rmmod to random places to force module unloading :/ @kernel team, do
> you know how could the reporter try to get more information?

rc_parallel has not been set in my /etc/rc.conf file so it should be off by default.

And certainly let me know if you can figure out a way for me to collect more information for you.

[Comment 5 Adam Feldman 2017-05-11 11:33:27 UTC]

Going to close this since I don't run into this issue on my new computer and I can't reproduce the original issue.