Date: Sun, 25 Jun 1995 12:23:59 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> To: wpaul@skynet.ctr.columbia.edu (A boy and his worm gear) Cc: hackers@freebsd.org Subject: Re: 2.05R reboot hangs Message-ID: <199506251923.MAA02095@gndrsh.aac.dev.com> In-Reply-To: <199506251912.PAA07494@skynet.ctr.columbia.edu> from "A boy and his worm gear" at Jun 25, 95 03:12:54 pm
next in thread | previous in thread | raw e-mail | index | archive | help
>
> Of all the gin joints in all the world, Rodney W. Grimes had to walk
> into mine and say:
>
> [lots of discussion about failed system resets]
>
> > This problem has actually been around for a long time (machines that
> > would not detect a CPU shutdown cycle have always hung during the
> > reboot). I don't know if I have made the problem worse or better
> > though with my keyboard reset, as that now seems to hang machines
> > that did not hang before. Perhaps I should reverse the logic of
> > the #ifdef and only use the keyboard reset on machines that have
> > broken cpu shutdown detection.
> >
> > Comments?
> > --
> > Rod Grimes rgrimes@gndrsh.aac.dev.com
> > Accurate Automation Company Reliable computers for FreeBSD
>
> I have a thought. I may be barking up the wrong tree here, and for all I
> know this could be impossible, but I suppose it wouldn't hurt to ask.
>
> I seem to recall reading somewhere that there's a BIOS routine that
> can be used for rebooting the system. Now, I know using BIOS calls is a
> no-no without a VM86 interface of some kind, but what I'm wondering
> is if it's possible to force the system back into real mode just long
> enough to call the BIOS reboot routine.
It is allmost impossible at this time to go back to real mode, which
is what would be needed to call the BOIS reboot routine. The problem
is we have no page that is mapped physical==linear with kernel code
in it (that is what is required to go back to real mode).
> I've been reluctant to suggest this since I don't know a) how easy it
> would be to do a 'quick and dirty' switch back to real mode, b) if it's
> even _possible_ to switch back to real mode this late in the game,
> c) whether or not the BIOS reset call exists or if it's just a figment
> of my imagination or d) if it does exist, does it reset the system
> correctly (i.e. will it cause all hardware to be reset properly).
>
> My reasoning is that the BIOS reset function should be reasonably reliable
> ("It works in DOS" and all that crap). And since we're about to kill the
> system anyway, maybe it wouldn't be that hard to jam the system back into
> real mode for the purpose of this one call (we don't necessarily have
> to be that neat about it since we don't ever intend to return).
>
> So, am I nuts, or am I nuts?
Your not so nuts, but it is something very hard to do at this time.
Something that does need to be fixed though, we really do need a way
back to true real mode (not just VM86).
--
Rod Grimes rgrimes@gndrsh.aac.dev.com
Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199506251923.MAA02095>