Re: CVS commit: src/usr.sbin/makefs

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

> On Tue, May 24, 2011 at 09:52:18PM +0900, Izumi Tsutsui wrote:
> > It looks mandatory to make makefs(8) handle partition maps for CD images.
> 
> I'm not sure I understand exactly what you mean and why makefs is the right
> tool here. Postprocessing the image is fine 

I think the problem is that the final ISO9660 output including
embeded MD boot fs and partition maps still has to be iso9660 compliant.
There is no public space to store MD parameters for postprocessing
in output image, and I don't think we can simply use existing
fdisk/pdisk tools to fill embeded MD (MBR or APM etc) partitions
and iso9660 parameters. (but I could be wrong)

> (I'm sure you know how sgimips
> works with the volume header, the sparc64 case is even simpler by just
> postprocessing the image filling in the sun label.)

On sgimips case, sgimipscd.iso has SGI specific format
and it can't be mounted by non-SGI machines, IIRC.
(while we can mount hybrid mac68kcd.iso and macppccd.iso on i386)

No idea about sparc64, but maybe it works because
it appends MD bootfs at the end of iso9660 fs and
Sun's label which has offsets of the bootfs is small enough
to put it into unused part at the top of iso9660.

> What is the content of the pre-generated HFS piece? Is it relatively static?

I have not checked FreeBSD's one, but it may have at least ofwboot.elf
and one kernel, which should be updated during build.

---
Izumi Tsutsui