Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!seismo!sundc!hqda-ai!merlin
From: merlin@hqda-ai.UUCP
Newsgroups: comp.os.minix
Subject: MINIX extensions
Message-ID: <252@hqda-ai.UUCP>
Date: Sat, 24-Jan-87 05:45:49 EST
Article-I.D.: hqda-ai.252
Posted: Sat Jan 24 05:45:49 1987
Date-Received: Sat, 24-Jan-87 16:04:25 EST
Organization: Army AI Center, Pentagon
Lines: 108
Keywords: shared swapped bernoulli jobs
Summary: Suggestions for features to be added to MINIX. New device drivers, disk formatting, shared text, job control, and process swapping.


     Well, Tanenbaum's book hasn't even had time to let the ink
dry, and here I am looking for extensions.  I really love the idea
of a more-or-less public domain OS, but it needs some work to be
really useful.  Here's a few suggestions.

     COMMENTS  -->  postnews
     FLAMES    -->  mail
     (I do read flame mail, but there's no reason everyone else has to.)


1. Device driver for the IOMEGA Bernoulli Box:

     I have a twin 20-meg bernoulli here.  Sure would be nice to
be able to use it...  Seriously, the bernoulli is good, reliable,
has a large capacity, and is relatively fast.  Good candidate for
a device driver.

     I'd like to have partitions (Sys5 slices) on the drive.  With
a floppy, there's no sense in partitioning the thing.  It's just
not big enough to worry about.  But 20 megs is quite enough to do
it with, and I want partitions for another extension.


2. Provide a way to format disks without going into MS-DOS.

     My system isn't used for MS-DOS, or won't be, as soon as I
can get a hold of MINIX.  I'd like to be able to format disks.  I
think the disk drivers should have another IOCTL for this.

     A format ioctl could chew up lots of space in the kernel, and
it really won't get used that often.  I think an ioctl to write a
raw track, including all the format bytes, would be in order.  It
would be relatively small.  The real work of figuring out how to
format a disk can be left to a user-level program.


3. Share program text

     I see no reason why the same program should have to be loaded
multiple times.  The code segment is assumed to be read-only.  I
know there's no hardware protection for this, but in general I
think it's a safe bet, and it would free up memory.  Memory seems
to be the limiting factor in MINIX.


4. Swap processes

     This is why I wanted partitions.  At present, when MINIX runs
out of core, no more process creations are allowed.  Two things
can be done about this:

     a) Rearrange processes in memory, to free up the unused space
that may be lurking between processes.  This may work, provided
that MINIX processes are allowed to fiddle with their segment
registers.  Really, they can't even *refer* to the segment
registers.  The only reason for doing that is to reference
something via the segment register.  If the segment register might
change later, that would invalidate the reference.

     b)  Swap processes, or portions thereof, out to disk.  Not so
good on floppy, but acceptable on a bernoulli or winchester.  This
isn't really feasible unless (a) above is also ok.  

     Assuming that we can reassign processes, we need a place to
put them.  The partition scheme provides an efficient place (a
block device) for this storage.  I would prefer to swap the text
portion of a process out first.  Only swap the data/stack segment
if the text didn't free up enough memory.  Reason:  the stack area
is read-write, so we must write it to swap space, and read it back
in.  The code segment is (presumably) read-only, so there's no
need to write it out.  It can be reloaded from the executable
file.


5. Provide SIGSTOP/SIGTSTP and job control, ala 4.xBSD

     Modifications required to the terminal/keyboard drivers.
Also, the shell would have to be told about this, to make it
useful.  This is not a particularly great idea unless 4 is also
implemented.  No point in background jobs unless there's memory to
hold them.  (God, I miss VM hardware :-( )


Well, that's my list so far.  I should say that my machine is an
AT.  This set of extensions is probably too much for the
garden-variety PC, but I've got the CPU cycles to support this
sort of stuff (I hope).  

     Why bother with all this?  Why not just buy Sys-V for the PC?
BECAUSE I WANT THE SOURCE, AND I DON'T LIKE AT&T.  Also, I intend
to run uucp mail and usenet news on this beast, and it'll need all
the help it can get.

Points for discussion:

a)   Are these useful?
b)   Can they be done?
c)   Does anyone want them?
d)   If they are done, how do I publish them without stepping all
     over Tanenbaum's rights?  I think its OK to publish only the new
     stuff (you'd still have to pay Prentice-Hall for the basic OS),
     but I'd like a second, third, and fourth opinion.
-- 
	David S. Hayes, The Merlin of Avalon
	PhoneNet:	(202) 694-6900
	ARPA:		merlin%hqda-ai.uucp@brl-smoke
	UUCP:		...!seismo!sundc!hqda-ai!merlin