Cannot reboot from busybox

Rob Landley rob at landley.net
Fri Dec 23 01:21:05 UTC 2005


On Thursday 22 December 2005 09:46, IraqiGeek wrote:
> I am doing all the testing on a VMware virtual machine, no ACPI, and no
> buggy BIOS are involved. reboot doesnt hang, it just does nothing.

I prefer qemu myself.

Try this:

#include <sys/reboot.h>

int main(int argc, char *argv[])
{
  reboot(RB_POWER_OFF);
}

Compile it and run it as root.  That should halt and power off (meaning vmware 
should exit).  If that compiles and runs but doesn't work, it's probably 
vmware.  If it works, it's busybox init being horked again.

> I am running another Linux virtual machine, which is a copy of the one I am
> using for this testing, based on ttyLinux, only using the Busybox build of
> ttyLinux, instead of doing my own build and that one reboots like a charm
> (same boot scripts, and using the same vanilla 2.6.11.12 kernel that I
> built for the LFS system that I am using for all development and
> compilation work). I have checked the busybox.links file of that build
> against mine, and apart from a couple of irrelevant applets (hdparm,
> fdformat, for example), there arent any applets that are available on the
> ttyLinux build that are missing from mine (I went through the help of each
> missing applet to make sure).

No, it sounds like the attempt to communicate with init isn't working.  (Sends 
SIGUSR1 or some such.)  A rewrite of this entire area is a todo item of ours.

You said you tried 1.1.0-pre1?

> Finally, though I doubt its of any relevance, I am using wrappers instead
> of hard or soft links to call busybox. Just in case, I called reboot and
> halt directly by running "busybox reboot" and "busybox halt", and still the
> system refused to reboot.

Nah, shouldn't make too much of a difference.

Rob
-- 
Steve Ballmer: Innovation!  Inigo Montoya: You keep using that word.
I do not think it means what you think it means.



More information about the busybox mailing list