Message-ID:	<c=US%a=_%p=SSG%l=SITE2S1-970204205754Z-23912@site2s1.sbservices.com>
From:	Jonathan Tarbox - SSG <JonatTar AT sbservices DOT com>
To:	"'Opendos AT mail DOT tacoma DOT net'" <Opendos AT mail DOT tacoma DOT net>
Subject:	RE: [opendos] OpenDOS + Win95 w/FAT32?
Date:	Tue, 4 Feb 1997 15:57:54 -0500
MIME-Version:	1.0
Sender:	owner-opendos AT mail DOT tacoma DOT net

	Actually, you are still wrong but looking in the right direction, the
superblock of the partition (say it was Ext2) would have the code to
supply read-only support for it's FS, and the IFS line enabled the full
read-write support for that and other IFS drives

	Example, (given that say HPFS and EXT2 IFS's were available) say I have
a 1.6 gig boot drive and a 540 meg secondary drive.. I would want the
1.6 gig to have ANY filesystem but FAT, thus I would make the first disk
with a 600 meg HPFS drive and a 1 gig EXT2 drive..  and possible FAT on
the 540. Now using Linux's LILO I can boot from partition 1 or partition
2, the root blocks of each partition tell it how to at least read that
file system used.  Now in the OpenDOS CONFIG.SYS I would put two IFS
lines, one for HPFS read/write support, and one for EXT2 read/write
support...  Without either the system could only read the HPFS, and
could read/write the FAT on the second disk, but not be able to see the
EXT2 since it doesn't know how to.

	But in that same exampe if I set it to load only the HPFS IFS it would
be able to read/write to the HPFS and FAT partitions, but still not
see/access the EXT2 partition.

	This may all sound reasonable, but I'm clueless on the actually
programming of it.. But it should be able to work, since MS-DOS uses the
same concept to the computer can see the system files on a FAT partition
and Linux does the same with it's EXT2 partitions..

Jonathan Tarbox
>----------
>From: 	Brian Dukes[SMTP:bdukes AT crox DOT demon DOT co DOT uk]
>Sent: 	Tuesday, February 04, 1997 2:11 PM
>To: 	opendos AT mail DOT tacoma DOT net
>Subject: 	Re: [opendos] OpenDOS + Win95 w/FAT32?
>
>On Tue, 4 Feb 1997 16:34:24 +0000, you wrote:
>
>>Brian Dukes writes:
>>
>>> >Hmm, there is an interesting question. If filesystem support is loaded
>>> >from CONFIG.SYS, and you want to have a 100% VFAT or even ext2fs
>>> >filesystem, how will DOS be able to read the filesystem to find
>>> >CONFIG.SYS so that it can load the file system driver that is required
>>> >to read things from the filesystem that CONFIG.SYS is held on....
>>> 
>>> [SNIP]
>>>
>>> To be honest, I can't see that its going to be practical to move from
>>> one IFS to another nilly-willy just by changing an instruction in the
>>> CONFIG.SYS, 
>>
>>Neither can I, that's not what I'm suggesting or indeed asking.
>>However, there is a very good reason why you'd want to not build all
>>filesystems into the DOS kernel.
>>
>>> for one the data stored under one IFS would be fairly incompatible
>>> with the data stored under another .. and therefore, in order to
>>> switch between filing systems you would probably ZAP the partition
>>> first!
>>
>>You have totally missed by point. I was wondering out loud about the
>>idea of not building the filesystem into the kernel of DOS but having
>>the filesystem loaded from (for example) CONFIG.SYS. No-one shifts
>>filesystem lightly and the fact that you will loose the data should be
>>pretty self evident to anyone.
>>
>>Besides, I can't see your point. Even if they *do* ZAP the partiton
>>(after taking a backup I'd hope), my question still stands.
>
>
>Hmmm, I seem to have either gotten out of the wrong side of the bed
>today or put my teeth in upside down.
>
>The original question/idea was to have an IFS= type command in
>CONFIG.SYS ... all I was saying is that its not practical because its
>too damn easy to change and to mess things up for one.
>
>I haven't missed your point Dave, I just didn't dwell on it.  Your
>point if I may paraphrase was that in order to read the IFS=
>instruction from CONFIG.SYS, the boot process would need to know
>something about the IFS in place in order to locate the config.sys and
>start processing it!   And what tells the boot process which IFS is
>installed?
>
>The fact that these instructions were in the CONFIG.SYS would mean
>that ANYONE would have the ability to alter the current IFS= without
>any pre-thought!   How many times have you had to go in and tidy up
>someones CONFIG.SYS after they messed!   What if someone took out the
>IFS= line by accident then rebooted?
>
>I think we should look at operating systems like OS/2 or Linux to see
>how they achieve the same goal ... personally I think the way Linux
>works would be far better.
>
>Bri, hoping i've made myself a little clearer now
>
>

- Raw text -