Path: sparky!uunet!mcsun!uknet!yorkohm!u8acb
From: u8...@ohm.york.ac.uk (Alex Barclay)
Newsgroups: comp.unix.bsd
Subject: Recompiling the bootcode. Fix for my PC
Message-ID: <1992Apr11.210239.2673@ohm.york.ac.uk>
Date: 11 Apr 92 21:02:39 GMT
Organization: Electronics Department, University of York, UK
Lines: 29

Could anyone out there in net land help me with a cure that I'm trying to
make BSD 386 boot on my PC (386SX AMI Bios, IDE H/D Seagate).

I can get the thing to boot off the floppy by... Wait for it...

1) Telling the BIOS that the HD isn't present so it doesn't try to boot from it

2) Waiting after switch on till the machine realises that there isn't a
   bootable device and aska for a boot disk.

3) Inserting the dist.fs disk and pressing the key as instructed. The OS
   now boot perfectly. And I can play with it, including the tty hangup
   demonstrated by the more command.

4) I can access the hard disk, label it, make a file system but the machine
   displays the same problem trying to boot off that as off the floppy. The
   strange bit is that the floppy drive light comes on when the HD boot fails.

This is definately strange and I'm am looking for a cure, involving going
through the chip datasheets that I have and checking what the code is doing.

The problem is that I have no way of testing my theories. Is there anyone
who could compile/assemble up some boot code that I send them on a working
386BSD system, then send the resulting binaries back so I can glue
them onto the front of a boot disk???

		Any help appreciated...

				Alex.

Newsgroups: comp.unix.bsd
Path: sparky!uunet!usc!sdd.hp.com!hp-col!bdale
From: bd...@col.hp.com (Bdale Garbee)
Subject: Re: Recompiling the bootcode. Fix for my PC
Sender: no...@col.hp.com (notes)
Message-ID: <1992Apr12.041455.25340@col.hp.com>
Date: Sun, 12 Apr 1992 04:14:55 GMT
References: <1992Apr11.210239.2673@ohm.york.ac.uk>
Organization: HP Colorado Springs Division
Lines: 23

u8...@ohm.york.ac.uk (Alex Barclay) writes:
> I can get the thing to boot off the floppy by... Wait for it...

Fascinating!  I tried your sequence of operations on the 16Mhz 386SX system 
I've been trying to boot 386BSD on, and it worked exactly as you document!.  My
system config is:

	- no-name 386SX 16Mhz motherboard with Intel 387SX16, 4MB RAM, AMI Bios
	- WD1006 with Maxtor XT-2190 and a 1.44M floppy drive
	- no-name hercules-clone video card and Northgate keyboard

This system had exactly one time booted 386bsd correctly from floppy, but on
every other attempt hung after the 'color' token as has been so often 
described.  As you noted, it still hangs that way when attempting to boot off
the hard disk.  I wonder what happens when a disk isn't found that is tickling
the kernel/hardware?  Maybe a double-initialization, or something?

I don't currently have any easy way to build a new kernel, so this still 
doesn't leave me with a useful 386BSD system...  

Bizarre, but thoroughly amusing!  :-)

Bdale

Path: sparky!uunet!munnari.oz.au!goanna!minyos.xx.rmit.oz.au!s871780
From: s87...@minyos.xx.rmit.oz.au (Antony Suter)
Newsgroups: comp.unix.bsd
Subject: Re: Recompiling the bootcode. Fix for my PC
Message-ID: <1992Apr13.104624.16982@minyos.xx.rmit.oz.au>
Date: 13 Apr 92 15:46:24 GMT
References: <1992Apr11.210239.2673@ohm.york.ac.uk>
Organization: RMIT Computer Centre, Melbourne Australia.
Lines: 19

u8...@ohm.york.ac.uk (Alex Barclay) writes:

>I can get the thing to boot off the floppy by... Wait for it...

>1) Telling the BIOS that the HD isn't present so it doesn't try to boot from it

I believe that Advanced AMI bioses have a CMOS config option specifying
which drives and the order in which to boot. You can set it to 'A: C:' or
'C:' etc. Look at yours, its probably 'C:' ! If you dont know how to get to
your CMOS setup... 1) Be careful changing it! 2) Press the 'DEL' key just
after the memory count before the rest of the boot is completed. (cold boot).
Enter advanced CMOS setup. 3) Be careful changing it! 4) Be very careful
changing it!


-- 
Antony Suter			s87...@minyos.xx.rmit.oz.au
Melbourne, Australia		ant...@werple.pub.uu.oz.au
"Call me Jack, stranger things have happened!"

Path: sparky!uunet!mcsun!uknet!yorkohm!u8acb
From: u8...@ohm.york.ac.uk (Alex Barclay)
Newsgroups: comp.unix.bsd
Subject: Re: Recompiling the bootcode. Fix for my PC
Message-ID: <1992Apr13.111123.5981@ohm.york.ac.uk>
Date: 13 Apr 92 11:11:23 GMT
References: <1992Apr11.210239.2673@ohm.york.ac.uk> 
<1992Apr13.104624.16982@minyos.xx.rmit.oz.au>
Organization: Electronics Department, University of York, UK
Lines: 20

In <1992Apr13.1...@minyos.xx.rmit.oz.au> 
s87...@minyos.xx.rmit.oz.au (Antony Suter) writes:

>I believe that Advanced AMI bioses have a CMOS config option specifying
>which drives and the order in which to boot. You can set it to 'A: C:' or
>'C:' etc. Look at yours, its probably 'C:' ! If you dont know how to get to
>your CMOS setup... 1) Be careful changing it! 2) Press the 'DEL' key just
>after the memory count before the rest of the boot is completed. (cold boot).
>Enter advanced CMOS setup. 3) Be careful changing it! 4) Be very careful
>changing it!

Well from what you describe I haven't got an advanced AMI BIOS. The boot
sequence is always A: then C:. The point of disabling the hard drive was
so that I could get the BIOS to barf up the insert boot disk message. BSD
doesn't need the bios to tell it about the hard drive so it's a case of the
BIOS behaving differently on booting first time of a 3.5in floppy and
booting after the message. It's very odd but I'm looking to have a go at
fixing it, by first printing numbers to the screen from the boot sequence
at various points. When the numbers stop then I've found the problem!

			Alex.

Path: sparky!uunet!mcsun!uknet!icdoc!news!lmjm
From: lm...@doc.ic.ac.uk (Lee M J McLoughlin)
Newsgroups: comp.unix.bsd
Subject: Re: Recompiling the bootcode. Fix for my PC
Message-ID: <LMJM.92Apr14145237@raquel.doc.ic.ac.uk>
Date: 14 Apr 92 14:52:37 GMT
References: <1992Apr11.210239.2673@ohm.york.ac.uk>
Sender: use...@doc.ic.ac.uk (News System)
Organization: Dept. of Computing, Imperial College, London. U.K.
Lines: 29
In-Reply-To: u8acb@ohm.york.ac.uk's message of 11 Apr 92 21:02:39 GMT
Nntp-Posting-Host: raquel.doc.ic.ac.uk


Have you tried the locore.s patch?  A copy of my kernel with this
patch applied along with the recent collection of patches to get
kern_execve to work is in:
	src.doc.ic.ac.uk:incoming/386bsd.lmjm

The two patches applied to this kernel are given below.  I also
rebuilt it to allow NFS.  I cannot promise that these are the only
patches.  I've been hit by an odd problem with login and have been
bouncing in and out of RCS revisions.

From: ja...@raid.dell.com (James Van Artsdalen)
Subject: a20 bug in locore.s
Message-ID: <1992Mar31....@raid.dell.com>
Date: 31 Mar 92 02:28:30 GMT

From: ting...@plains.NoDak.edu (Mark Tinguely)
Subject: Re: Fixed problem with long command lines
Message-ID: <16...@plains.NoDak.edu>
Date: 9 Apr 92 02:50:02 GMT

Copies are in, respectively:
	src.doc.ic.ac.uk:386BSD/comp.unix.bsd/volume92/Apr/920401.11.Z
	src.doc.ic.ac.uk:386BSD/comp.unix.bsd/volume92/Apr/920409.20.Z
--
--
Lee McLoughlin.                          Phone: +44 71 589 5111 X 5085
Dept of Computing, Imperial College,     Fax: +44 71 581 8024
180 Queens Gate, London, SW7 2BZ, UK.    Email: L.McLo...@doc.ic.ac.uk