Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 10 Sep 2005 11:47:52 -0600
From:      Scott Long <scottl@samsco.org>
To:        Poul-Henning Kamp <phk@phk.freebsd.dk>
Cc:        Jung-uk Kim <jkim@freebsd.org>, freebsd-arch@freebsd.org
Subject:   Re: Bigger boot block size?
Message-ID:  <43231C48.80008@samsco.org>
In-Reply-To: <8350.1126341497@phk.freebsd.dk>
References:  <8350.1126341497@phk.freebsd.dk>

next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp wrote:

> In message <200509091214.41429.jkim@FreeBSD.org>, Jung-uk Kim writes:
> 
> 
>>Sigh...  But bsdlabel(8) should be able to handle this case when '-B' 
>>option is given and first partition of the slice is UFS1, i. e., we 
>>keep 'historical' boot1/boot2 for a while and drop the support 
>>later. ;-)
> 
> 
> Guys, bsdlabel has no future, we need to migrate to something
> that is
> 	A: 64 capable
> 	B: supports more than 7/8 partitions
> 	C: understands that metadata should not be exposed in-band.
> 

Yes, and you've been waving your hands about this for a long time =-)
GPT is a nice sucessor, but our support for it is not ready for prime
time, IMHO.  What is required to actually make the switch:

1) appropriate support in all stages of the bootloader for i386 and
amd64.
2) appropriate support in sysinstall for all platforms.
3) docs that describe GPT on FreeBSD, how to install a system to it, and
how to upgrade a system to it.
4) visual editor for GPT.  You might laugh that bsdlabel/disklabel only
uses vi, but it's worlds better than what the gpt utility provides now.

Scott



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?43231C48.80008>