On Monday, 13 May 2002 at 22:59:07 -0400, Thor Lancelot Simon wrote:
> On Mon, May 13, 2002 at 08:39:45PM -0400, Perry E. Metzger wrote:
>>
>> Luke Mewburn <lukem@wasabisystems.com> writes:
>>> Actually, I don't think that this will help, at least for ffs file systems.
>>> You need to crank the sector size to 1024 bytes from 512 bytes to have an
>>> ffs file system larger than 1 TB (1024 GB), because ffs has a signed 32 bit
>>> quantity for disk block addresses, and 2^(31+9) == 1 TB.
>>
>> We really should fix this. The limitations in FFS are largely
>> artificial -- an ffs2 that did 64 bit quantities for all these things
>> would be very straightforward to produce.

FWIW, Kirk McKusick is currently working on UFS2, which will have 64
bit pointers.  Note that there's more than just a simple change to get
this to work: for one thing, it means that the on-disk data structures
are incompatible with UFS.

> If we were going to do *that*, it would make sense to rip out all
> the rotational placement and other obsolete complexity at the same
> time.

The FreeBSD project is doing a complete overhaul of UFS, in
cooperation with Kirk.  I'm not sure whether that stuff has been
removed yet, but if not, I'm sure it's on the list.

Greg
--
Finger grog@lemis.com for PGP public key
See complete headers for address and phone numbers