Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!van-bc!seac!wain
From: wain@seac.UUCP (Wain Dobson)
Newsgroups: comp.unix.xenix
Subject: Re: XENIX 2.2.3 and >1024 h.d. cylinders: possible?
Message-ID: <5564@seac.UUCP>
Date: 14 Sep 89 18:16:07 GMT
References: <5563@seac.UUCP> <4593@cps3xx.UUCP>
Reply-To: wain@seac.UUCP (Wain Dobson)
Organization: SEAC Software Engineering, Vancouver, B.C.
Lines: 60

In article <4593@cps3xx.UUCP> usenet@cps3xx.UUCP (Usenet file owner) writes:
>in article <5563@seac.UUCP>, wain@seac.UUCP (Wain Dobson) says:
>> Keywords: xenix, hard disk, cylinders, ESDI
>> 
>> If it don't work, it's the manufacturers problem, not SCO's. After all,
>> which comes first, hardware selection, or software selection.
>
>For people migrating up from a DOS machine with a big disk to XENIX it's
>hardware selection.....
>
>Why did the 1024 limit start? Well, the MSDOS defined partition table
>layout only allows 10bit cylinder numbers. So I could see if the
>machine was going to have DOS and XENIX partitions using the disk to
>stick with the MSDOS partition table layout. However, if someone wnated
>a XENIX only machine shouldn't XENIX be easily able to do its own
>hard disk boot sector to handle a partition table layout that is
>different? And since XENIX is contained within one physical partition
>that then has logical partitions....
>

Was not trying to pose the 'Which comes first, the chicken or the egg?'
question; nor trying to get into the technical reasons for why the 1024
limit is a reality. I agree with you, SCO should be able to do things
itself. My posting was in-reply to the assertion that SCO UNIX 3.2's
documentation should be clear on whether or not it supports greater than
1024 cylinders. And, I simply pointed out that it is quite clear that SCO
UNIX 3.2 does not support greater than 1024 cylinders. To quote from page
A-12 of the Release Notes for SCO UNIX 3.2:

	"SCO supports UNIX on hard disks with up to the 1024
     cylinders only. Some hard disks have more than 1024
     cylinders. Due to the non-standard nature of the hard
     disk controllers that support disks with more than
	 1024 cylinders, and due to the need for DOS co-residency,
     SCO does not support UNIX on these disks. However, UNIX
	 has been reported to work on disks with more than 1024
	 cylinders. For example, with the DPT PM3011 controller,
     UNIX works with disks containing 2048 cylinders, 16 heads,
     and 63 sectors, for a total maximum disk capacity of 1
     gigabyte."

If this quote is not clear as to what SCO UNIX 3.2 does and does not support,
then I will have to declare myself 'illiterate.' (Probably should, anyways,
if I keep getting myself into discussions that seemingly go round and round.)

As to which comes first, the harware or the software, in this case, like it
or not, buy the software first, then the hardware. SCO states, very succinctly
that because it considers controllers that support greater than 1024 to be
non-standard and because it supports DOS co-residency, it does not support 
greater than 1024 cylinders.  

Having dispensed with what SCO UNIX 3.2 does and does not support with respect
to cylinder count, SCO UNIX 3.2 documentation at page A-23 states that they 
have 'used' a number of controllers, some of which provide a means to go beyond 
the 1024 limit in some manner. To support and to have used, from my
perspective, are very different. That something 'works' has a different 
meaning, as well.  
-- 
Wain Dobson, Vancouver, B.C.
	...!{uunet,ubc-cs}!van-bc!seac!wain