Re: reboot bug fixed (NOT)
On Wed Mar 22, 2000 at 11:47:03PM -0500, Adam Di Carlo wrote:
>
> Exploiting a dbootstrap bug that causes it to crash (running non-DHCP
> netconfig more than once), which incidentally karlheg is fixing I
> hear, I am definately able to reproduce this init condition. After
> dbootstrap crashes, init seems freak.
>
> Erik Andersen <andersen@xmission.com> writes:
>
> > so the only thing I can think of would be if shutdown_system()
> > was to block. After running 'reboot' from the command line
> > are new processes being spawned from init (i.e. if you exit your
> > shell does a new shell get spawned)?
>
> No it's not.
>
> > If so then something _very_ strange is going on such that init
> > isn't getting the signal.
>
> Yup.
>
> > Also, what does 'kill 2 1' result in?
>
> Nothing.
Very odd. Can you give me a ps listing? We are not running as an initrd, so
init should be PID 1... Did you every see on the console the "The system is
going down NOW !!" message? That is important.
What I have seen so far suggests it is blocking in shutdown_system(). I did
notice last night that shutdown_system() is using kernelVersion, buy somewhere
in the last while, kernelVersion stopped getting initialized. I can't access
my work box from home till next week (we are moving to a bigger building, and
while I am already moved, the routing will be hosed until everybody else makes
the move on Friday). I'll be at work in about an hour and work on fixing
things up then. Meanwhile, could you try out the following patch to init.c?
+/* Figure out what kernel this is running */
+kernelVersion = get_kernel_revision();
+
/* Figure out where the default console should be */
console_init();
/* Close whatever files are open, and reset the console. */
-Erik
--
Erik B. Andersen Web: http://www.xmission.com/~andersen/
email: andersee@debian.org
--This message was written using 73% post-consumer electrons--
Reply to: