SourceMage Grimoire Bug List

[bugzilla-daemon AT bugs.sourcemage.org]

http://bugs.sourcemage.org/show_bug.cgi?id=13114





------- Additional Comments From acedit AT armory.com  2007-02-27 02:34 -------
I think the error is meaningful because it prevents proper shutdown. 
sigsuspend
is supposed to wait for init to finish halting all running services. Instead 
it
is returning prematurely, then broadcasting kill signals, stomping on whatever
init was doing.

When sigsuspend returns, the code almost immediately arrives at the point 
where
it sends kill signals to all processes. Because the function does not block as
advertised (due to the error) this happens before init has finished shutting
down services. What should happen is sigsuspend blocks until init has halted 
all
running services. Then sigsuspend returns, and execution continues and sends 
the
system-wide kill signals.

It would be simple to test this theory by replacing the sigsuspend with a 
sleep
period long enough to allow init to shutdown all its services uninterrupted. I
think if that analysis is correct, stracing init will probably be inconclusive
as kill signals are broadcast while its trying to halt services. I think we
should disprove the above theory about non-blocking sigsuspend before going 
on a
potentially wild goose chase inside init.


-- 
Configure bugmail: http://bugs.sourcemage.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.