When init receives SIGHUP, the state machine's requested_transition is changed to clean_ttys. Lets say SIGTERM is received while we are in clean_ttys, it will set requested_transition to death. But at the end of clean_ttys it will return multi_user directly which then resets requested_transition. So we loose the SIGTERM sent by shutdown
While returning, clean_ttys should check if requested_transition is
no longer clean_ttys and then return requested_transition
instead of multi_user
How-To-Repeat: Create a big /etc/ttys.
kill -1 1; shutdown -r now
Lot of times the machine will not reboot.
this is not just possible, it does happen frequently. I'm seeing this @ work several times a week.
If a script fires a signal to reboot and a few moments later, another instruction tells init to just shutdown, init will forget about the reboot.
I've created a patch for that problem, which solves that problem (stress tested my notebook with that):
Who's going to commit?
unowned PRs should not be in analyzed state
For bugs matching the following criteria:
Status: In Progress Changed: (is less than) 2014-06-01
Reset to default assignee and clear in-progress tags.
Mail being skipped