I had a look at this problem, and as far as I an see, the kernel is the entity responsible for turning off the machine. At the end of init.d/halt, 'halt -d -f -i -p -h is called, and this end up with a syscall 'reboot(RB_POWER_OFF)'. I guess the problem is that this syscall do not turn off the power any more. I have no idea why.
Could it be that the kernel changed to try to call some user space programs using udev or similar before turning off the power? udev is dead at this point, so that is doomed to fail. To test this hypotesis, I tested in qemu with <URL:http://koltsoff.com/pub/reboot4fun/poweroff.c> (random link I found via google), and this did not shut down qemu, even if all user space programs were running as normal. I was using debian kernel 2.6.17-1-686. I also tested this on an old ubuntu installation with kernel 2.6.10-5-386, and there the machine powered down. Could this be apm/acpi related? I suspect it is. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
