tor 2007-06-07 klockan 17.59 skrev Thor Lancelot Simon:
> On Thu, Jun 07, 2007 at 05:41:37PM +0200, Johnny Billquist wrote:
> > tor 2007-06-07 klockan 17.29 skrev Jason Thorpe:
> > > 
> > > For systems whose firmware doesn't give a flying banana about the  
> > > partition table format, GPT will be just as ignored by the firmware as  
> > > BSD disklabels are.
> > 
> > I take it that this means you intend to switch them over to use GPTs. So
> > we'll "create" a format there, which I guess was designed for another
> > type of systems... Oh well, as long as it keeps as non-intrusive as
> > possible.
> 
> GPT is a nice, flexible partition table format -- much nicer and more
> flexible than BSD disklabel, and not really any more complex than it
> needs to be to overcome some of disklabel's limitations.
> 
> Since we *already* use a foreign partition table format almost everywhere
> (since that's what disklabel is) it seems eminently sensible to switch to
> a more modern and flexible one.  What disadvantage do you see to this?

Complexity. That's all I'm worried about.
Let's face it, I'm an old fart who likes to play with old hardware. And
I'm seeing bloat everywhere. :-)

I can't really say that I've seen any problems with limitations with the
current disk labels. But then again, I'm using a VAX to play with. >1 TB
disks aren't that common... :-)
And the current disk labels works excellent. I really can't see the need
for something more advanced here.
All it will add, from my point of view, is complexity without any gains.
Along with all the "excitement" of finding the bugs in the new code. And
there is always the question of if the design really will work well with
whatever odd hardware I have, which expects things to be in one way. And
the current disklabel design was done with that kind of hardware more in
mind than whatever people design for these days. :-)

	Johnny