Bootscript poweroff options

Declan Moriarty junk_mail at
Wed Aug 22 01:01:21 PDT 2007

On Tue, 2007-08-21 at 15:25 -0500, Kevin Day wrote:

> > halt, poweroff, shutdown -<anything> etc. But it does respond to
> > whatever the bootscripts do when you get a drive error.
> At some point in the process of changing kernel/binutils/gcc/uClibc,
> the poweroff just stopped "powering off" for me.  It would shutdown
> the harddrive and say your os is off, meanwhile my entire machine and
> moniter is running, but is unresponsive because the kernel is
> unloaded. (maybe an initng version ended up breaking this??)
> This happens on every hardware I try and it pisses me off as I cannot
> find the fix.  I've swapped versions of binutils, gcc, uclibc, initng,
> etc..

I am pretty sure it's an apm/acpi thing. It does come up in the kernel
Just looking at kernel configs, you have nearly all the ACPI options set
in the one you sent to me, but my hlfs config only has the ACPI poweroff
settings set. But shutdown -<any> now does _NOT_ work in HLFS. Only the
bootscript manages, which intrigues me
> > If I fill /etc/fstab with random nonsense, it will fail to mount, and
> > power off until I fix it. HOW? That's the question. I did look at the
> > scripts, but .... cannot answer that for myself.
> Elaborate.
Okay, I recently added a drive, then removed one. On bootup with lines
in fstab for /dev/hdcN (which didn't exist) lfs-bootscripts let the
first one go, and said 'that's a mistake'. On the second one, it decided
'I'm out of here' and successfully powered the box off, and the box went
off - electrically.  I'd _love_ to know how it did that! No other halt
command on any system of that vintage or later manages to power the box
off electrically. BTW, windows can do power it off :-(, but I don't have
windows any more.

> Try to forcefully boot into the os by specifying init=/bin/bash in the

This is not the point. The point is I want to find how the box turns
off. I can get always going if needs be with a bit of B.F.+I. :-D.
        With Best Regards,

        Declan Moriarty.

More information about the hlfs-dev mailing list