Re: kern/45179: NetBSD disklabel does not support devices larger 2 TByte

[Martin Husemann <martin%duskware.de@localhost>]

The following reply was made to PR kern/45179; it has been noted by GNATS.

From: Martin Husemann <martin%duskware.de@localhost>
To: Wolfgang Stukenbrock <Wolfgang.Stukenbrock%nagler-company.com@localhost>
Cc: gnats-bugs%NetBSD.org@localhost
Subject: Re: kern/45179: NetBSD disklabel does not support devices larger 2 
TByte
Date: Wed, 27 Jul 2011 12:10:25 +0200

 On Wed, Jul 27, 2011 at 11:35:59AM +0200, Wolfgang Stukenbrock wrote:
 > The current hack in the bootloader code just try to get a valid 
 > disklabel 64 block's later for booting from a raid1 device. That is a 
 > very simple sollution.
 
 It is a tiny bit more complex: the GPT entry is of type
 GPT_ENT_TYPE_NETBSD_RAIDFRAME instead of GPT_ENT_TYPE_NETBSD_FFS,
 so the whole loop would need to match both GUIDs and if it is a RAIDFRAME
 match, offset the superblock-LBA by 64.
 
 > If disklabel is supposed to retire, than the sysinst should "always" 
 > install systems useing gtp as soon as possible.
 
 Yes, but it is not quite trivial, because for example Windows 7 (even SP1)
 can not share the boot disk if it is partitioned with GPT.
 For a NetBSD only setups, I agree (and intend do make it so).
 
 > On the other hand, the extensions to the disklabel structure does not 
 > sound very expensive to me - I only have no time to do it (and test on 
 > "all" architectures).
 
 disklabel has to die, ASAP!
 Besides, quite a few archs can not extend the on-disk version of the disklabel,
 as that would make them incompatible with their firmware.
 
 Martin