Date   : Sun, 15 Apr 2007 01:00:43 +0100
From   : jgh@... (Jonathan Graham Harston)
Subject: Possibly available: SJ FDFS

>Message-ID: <9E3CD75D66D943F188FFC1C67B6DDCD4.MAI@...>
 
"Ian Wolstenholme" <BBCMailingList@...> wrote:
> If anyone is thinking of hacking the MDFS code, my request
 
Nothing I'm going to try until I get another MDFS up and running ;)
 
> list is:
> 1.  Bigger disc partitions than 60MB
 
Difficult, as the MDFS uses 1024-byte disk blocks addressed by
16-bit block numbers. That gives a maximum of 65536*1024 bytes -
64M. Some time ago I looked through the code and it looks at
though it would be simple enough to patch to allow 64M partitions
(actually, 64M-1K, &FFFF blocks). It looks like it might be as
simple as changing the right occurances of &F000 to &FFFF.
 
> 2.  More than 8 partitions per disc
 
Probably doable. I don't know how the MDFS paritions disks. Does
it just split the drive into 60M sections, and accessing drive x
accesses sector 60M*x, or does it have a partition table? Has
anybody imaged an MDFS drive and have the image available for
study? I've imaged floppies, see http://mdfs.net/Mirror/Image/SJ/MDFS
 
> 3.  More than 8 hard disc partitions available when the
>      MDFS is online.
 
Possible, but at a cost. The MDFS holds the disk allocation map
in RAM, and there's a finite amount of RAM. Though, considering
that the MDFS can have 8 hard drive partitions, four floppy drives,
and a tape drive mounted, that's 13 allocation maps in memory. 
 
> 4.  More than 8 SCSI IDs available for hard discs
 
Why would you want to plug more than eight drives in? Isn't one
enough? :) Anyway, I think 8 SCSI IDs is a characteristic of SCSI,
not the MDFS. SCSI block addressing is: b0-b20: block, b21-23: device
 
-- 
J.G.Harston - jgh@...                - mdfs.net/User/JGH
Jet Set Willy Resources - http://mdfs.net/Software/JSW