Path: utzoo!attcan!uunet!munnari!basser!steve
From: steve@basser.oz (Stephen Russell)
Newsgroups: comp.os.minix
Subject: Re: Request for advice on PS/2 problems
Message-ID: <1726@basser.oz>
Date: 25 Jan 89 14:09:45 GMT
References: <1944@ast.cs.vu.nl>
Organization: Dept of Comp Sci, Uni of Sydney, Australia
Lines: 84

In article <1944@ast.cs.vu.nl> ast@cs.vu.nl (Andy Tanenbaum) writes:
>
>[ ... ]  I think the original
>design is fairly simple and don't want to muck it up because of the way
>the PS/2 does things.)

This won't be the first time that a simple system had to be abandoned
because of badly-designed hardware :-)

>Another point, I am still trying to figure out exactly how to tell if 
>a machine is (1) PS/2 and (2) Microchannel.  Several people have pointed
>out that INT 15 is supposed to tell you,

Only ref I have is in Thom Hogan's "The Programmer's PC Sourcebook", pp234.
INT 0x15, function AH=0xC0 returns a pointer in ES:BX to a "System
Descriptor Table". The layout is as follows:

	Offset	Length	Description		Value

	0	Word	Size of table in bytes	>= 8
	2	Byte	Model byte		See below
	3	Byte	Submodel byte		"
	4	Byte	BIOS Revision level	00 = first release
	5	Byte	Features		bit 7 HD BIOS uses DMA 3
						bit 6 2nd interrupt chip present
						bit 5 RTC present
						bit 4 keyboard intercept called
						bit 3 wait for ext event called
						bit 2 extended BIOS area alloc'd
						bit 1 PS/2-type I/O channel
						bit 0 (Reserved)

The rest is reserved. The interesting things are the model bytes and bit 1
of byte 5. Note: According to Hogan, this stuff applies to AT's after
1985 Nov 15, XTs after 1986 Jan 10, all XT286s, Convertibles and PS/2s.
His ref is IBM PS/2 and PC BIOS Interface Tech Ref, pp 2-94 to 2-96.

On page 197 of Hogan, the model number bytes are described as follows:

	Model Byte  Submodel Revision BIOS Version  Machine

	    FF        --        --        All	    IBM PC
	    FE	      --        --      8 Nov 82    PC/XT and Portable PC
	    FD        --        --        All       PCjr
	    FC        --        --     10 Jan 84    AT
		      00        01     10 Jun 85    AT
		      01        00     15 Nov 85    AT
		      04	00     Initial      PS/2 Model 50
		      05        00     Initial      PS/2 Model 60
		      02        00       All        PC/XT286
	    FB	      00	01     10 Jan 86    XT
		      00        02      9 May 86    XT
	    FA        00	00      2 Aug 86    PS/2 Model 30
	    F9        00	00     13 Aug 85    PC Convertible
	    F8	      00	00      Initial     PS/2 Model 80
		      01	00	Initial     PS/2 Model 80
	    FE	      NA	NA		    Compaq DeskPro
	    2D	      NA	NA		    Compaq Portable
	    9A	      NA	NA		    Compaq Portable Plus
	    4B	      NA	NA		    MegaBIOS ROM
	    B6	      NA	NA		    HP110 Portable
	
The machine model byte can also be found at address F000:FFFE. Ref:
PS/2 and PC BIOS Tech Ref, page 4-18.

So, the strategy to find the machine might be something like this:

- Check the model byte at F000:FFFE.
- If it's F[BDEF], then it's a PC, XT or Jr.
- If it's F8 or FA, it's a PS/2.
- If it's FC, could be an AT, XT286 or PS/2 Model 30/50. Early model
AT's will not support INT 0x15:0xC0 (carry set on return?). The rest
can be differentiated by the submodel/revision bytes.
- If it's none of the above, assume a PC clone, and hope for the best!

- For PS/2 machines, check bit 1 in byte +5 in configuration block for
PS/2 style I/O channels.

Now I don't have enough machines around here to test this out, but
others on the net may be able to check the details for their machine.
Of course, who knows what the clone makers have done. Fortunately, there
aren't a lot of PS/2 MCA clones (yet), but let's hope they get the
above details right.  One worry: why are there 2 model 80's for F8?
Mistake in Hogan's table?