Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!cbatt!ihnp4!inuxc!inuxf!matt
From: matt@inuxf.UUCP
Newsgroups: comp.os.minix
Subject: Disk Scheduling (& note to Dr T.)
Message-ID: <236@inuxf.UUCP>
Date: Thu, 19-Feb-87 20:33:16 EST
Article-I.D.: inuxf.236
Posted: Thu Feb 19 20:33:16 1987
Date-Received: Sun, 22-Feb-87 15:38:51 EST
Distribution: net
Organization: AT&T Consumer Products, Indianapolis
Lines: 51


I am planning on adding true disk scheduling to MINIX as part of an over-all
study of operating system efficiencies.  The existing method in MINIX is a 
First-come-first-served pipeline with no disk i/o queue.  I would like to
through out a straw man proposal for a method so that the alert among you
folks can poke holes, make suggestions, etc.

My first shot goes something like this:

	1) Alter floppy_task() on line 2540 of 'the book' so that all it does 
is receive a message, check for panic and queue up the request in a table
of pending process-message pairs.  It then loops back to receive another
message.

	2) Create a new procedure, say floppy_sched(), that examines the
queued requestes for the shortest seek time request and issues a do_rdwt()
for that request.

	3) floppy_sched() then will prepare a return message for the proper
process and send() a reply to that process.

	4) floppy_sched() re-examines the queue and the process continues.

A couple of questions come to mind.  How do I get floppy_sched() up and
running in the first place?  Make it PUBLIC (external or global) and allow
init() to call it as the system is coming up?  This seems to violate the
whole task concept where the user processes (ie init()) are not supposed to
have access to task calls.  There must be a clean method for doing this
but I have not run across it (I've only had the book three days now).

This seems to be fairly straight forward but I am not sure of the possible
race conditions or other hard to debug situations I may be headed for.
Is it not true that floppy_task() can only receive from one process at a time?
If so than I dont think the queue could be screwed up by simultaneous calls
to floppy_task().

Any comments?
---------------------------------------------------------
Dr Tanenbaum:
	I have tried three different paths to mail my thanks to you with
no success.  I did reach Mr. Fagen at PH and he gave his OK.  I have talked
to Prof Ganesh (dont remember his last name :-) and I should receive his
package shortly.  Again thanks for the info - my whole class appreciates it.


 Matt Verner   				UUCP:  ...ihnp4!inuxc!matt
 AT&T Graphics Software Labs		AT&T:  (317) 844-4364
 Indianapolis,  IN

           "The whole point of this sentence is to clearly
             explain the point this sentence is making."