You could certainly redefine the BPB; the result would not be a FAT
filesystem. You would need to have your own format *and* chkdsk utility
set and make sure that no existing FAT utilities, Microsoft or
otherwise, would think your new format looked enough like FAT that they
would “fix” it. I’m sure that would require more changes to the BPB.
There is no 512B/s limit. That’s just what virtually all current
magnetic spinning media exposes these days. Rumors are afoot that will
be changing in the next few years, but even now MO and DVD-RAM media
exist with 1KB and 2KB sectors. These work perfectly fine with our
current FAT.
If there is a limit its more like 4KB, but that’s a whole different
story.
I should note that in this case especially I’m no legal eagle, and
anything I say here is not an indication that the IFSKit license covers
such reuse of the code and FAT specification. You should check this
through your IFSKit contact directly if you do intend to go down this
road.
–
This posting is provided “AS IS” with no warranties, and confers no
rights. You assume all risk for your use.
-----Original Message-----
From: darragh.jones@cs.tcd.ie [mailto:darragh.jones@cs.tcd.ie]
Sent: Monday, January 14, 2002 5:00 PM
To: File Systems Developers
Subject: [ntfsd] RE: Increasing sector / cluster size in fat32
Ok, now i see where the 64k limit on cluster size comes from.
However, would it be possible to alter the Bios Parameter Block
structure
to increase the SectorsPerCluster type to USHORT? (If I made the changes
in
both the fastfat source code and format utility source code?)
Where is the 512 byte limit on sectors imposed? Is this limit imposed by
hardware, or a lower level driver?
On 01/14/02, ““Daniel Lovinger” ” wrote:
> SectorsPerCluster in the volume’s “BIOS” parameter block (FATs
> superblock equivalent) is UCHAR. Clusters are always 2^n sectors,
> therefore SPC maxes at 128 and physically caps cluster size at 64KB
for
> 512B/s media. The theoretical 1MB cluster would require at least 8KB
> sectors.
>
> No, it isn’t possible, and I’m actually not aware of an existing
> filesystem which would allow that large an allocation unit. UDF
> (http://www.osta.org) could physically represent up to a 2GB cluster,
> but carries implementation restrictions which limit it to sector size
=
> =3D
> cluster size.
>
> –
> This posting is provided “AS IS” with no warranties, and confers no
> rights. You assume all risk for your use.
>
> -----Original Message-----
> From: Darragh Jones [mailto:darragh.jones@cs.tcd.ie]=20
> Sent: Monday, January 14, 2002 8:40 AM
> To: File Systems Developers
> Subject: [ntfsd] Increasing sector / cluster size in fat32
>
> Here’s what I would like to do:
>
> Format a drive with the FAT32 file system and large allocation units
> (clusters).
>
> format /? reveals FAT32 supports 512,1024,2048,4096,8192,16k,32k,64k
> (128k,256k for sector size > 512 bytes).
> Also, 65526 < Number or clusters < 268435446
>
> So, if I wanted to use cluster sizes of 1MB what would be involved?
>
> I have the W2K IFS kit with the fastfat sample code.
> I do not however have the source code for the Microsoft format
utility.
> I do however, have the source to FormatX (www.sysinternals.com) and
> mkdosfs
> (linux utility for creating FAT32 file system).
> FormatX, however seems to simply call the (undocumented, AFAIK)
FormatEx
> function exported from the fmifs.dll library.
>
> Is what I am trying to do possible?
> Could someone please point me in the right direction if they have any
> ideas?
>
> Thanks,
> Darragh
>
> n.b. AFAIK =3D as fas as i know 
—
You are currently subscribed to ntfsd as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-ntfsd-$subst(‘Recip.MemberIDChar’)@lists.osr.com