Re: CVS commit: src/sbin/gpt

[David Brownlee <abs%absd.org@localhost>]

On 30 November 2015 at 13:34, Christos Zoulas <christos%zoulas.com@localhost> wrote:
>
> On Nov 29, 11:22pm, jnemeth%cue.bc.ca@localhost (John Nemeth) wrote:
> -- Subject: Re: CVS commit: src/sbin/gpt
>
> |      I have read the Makefile.  I think there is confusion between
> | the tool version and the crossbuilt version intended to go in the
> | release being built.  Obviously, the tool version will use the
> | system headers, and that may be the one that I saw fail (I don't
> | have the log anymore).
>
> The cross-built version is not affected either. In current it will just
> use the ioctls and be happy. In older versions it will use what it had,
> since it is not changed.
>
> |      Maybe so, but one could just lift the source code from -current
> | and plunk it on an older version of NetBSD and it would just work.
> | Now, one can not do that.  Having to update the kernel for a
> | relatively small userland utility is kind of annoying.
>
> That's the only disadvantage. Even if you wanted to do this, I am
> saying I would prefer that the pullup was done to the kernel to
> support the missing ioctls. And we are talking only about NetBSD-6
> where the utility is not that useful, since NetBSD-7 has them. Are
> you really planning to support gpt disks on NetBSD-6?

Uh, 3TB disks have been available since 2010, which predates NetBSD-7
by a little bit - and certainly >2TB raid and ccd setups have been
around for even longer.

While netbsd-7's gpt handling may be much improved, I'm assuming
netbsd-6's gpt support is not going to be dropped? :)