Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!amdahl!drivax!alexande
From: alexande@drivax.UUCP (Mark Alexander)
Newsgroups: comp.os.minix
Subject: Re: Minix bug report (hard disk problems)
Message-ID: <1029@drivax.UUCP>
Date: Thu, 5-Mar-87 13:05:47 EST
Article-I.D.: drivax.1029
Posted: Thu Mar  5 13:05:47 1987
Date-Received: Sat, 7-Mar-87 23:38:29 EST
References: <495@ubu.warwick.UUCP> <237@inuxf.UUCP> <7731@utzoo.UUCP>
Reply-To: alexande@drivax.UUCP (Mark Alexander)
Organization: Kludges 'R' Us, Inc.
Lines: 21

In article <7731@utzoo.UUCP> henry@utzoo.UUCP (Henry Spencer) writes:
>Disk interrupts probably do not get exercised much by MSDOS software; it
>is quite possible that there's a design flaw there.  (Unix is notorious
>for discovering hardware flaws that nothing else finds.)

Actually, disk interrupts DO get used by MS-DOS, but only in the ROM BIOS
calls that perform disk I/O.  Take a look at the Fixed Disk BIOS listings
for the XT.  After issuing a command to the controller, the BIOS
goes into a loop waiting for an interrupt to occur.  The interrupt
handler itself is very stupid -- just outputs an EOI command to the 8259
interrupt controller and disables disk DMA.  I can't seen any reason
why the BIOS even needs to use interrupts in the first place, but there
it is.

There are a number of multitasking operating systems that seem to
run just fine on XTs, and which (presumably) use disk interrupts.
These include PC/IX, Xenix, RMX-86, and others.  This could be taken
as evidence against serious design flaws in the hard disk controller.
-- 
Mark Alexander	...{hplabs,seismo,sun,ihnp4}amdahl!drivax!alexande
"Lying is a way of taking personal responsibility for reality." -G. Keillor