[syslinux] isohybrid question

[H. Peter Anvin]

On 03/13/2010 03:30 AM, Thomas Schmitt wrote:
> Hi,
> 
> I wrote about isohybrid MBR and sdb1 offset:
>> The change is not necessarily bound to built-in
>> MBR generation.
> 
> It comes to me that the offset and start address
> of the second set of meta data in the ISO image
> has to match the offset of partition 1 in the
> table of the MBR.
> 
> Another point is that the offset must be at least
> 38 kB, better 64 kB.
> 
> So it is indeed desirable to produce MBR and
> ISO image in a coordinated way.
> 

Yes, at the very least they need to agree on what the offset is supposed
to be.

My preference would be if we can do the integrated generation without
having to hard-code the actual MBR sequence; i.e. to produce a data file
out of the Syslinux build that can be fed into your tool.

The partition table -- as opposed to the MBR *code* -- probably should
be generated by xorriso/libisofs, which takes care of the "offset"
problem.  The only other thing that needs patching is the entry of the
LBA of isolinux.bin into offset 432; we should be able to keep that
stable.  The handling of forcehd0/ctrlhd0/partok can be dealt with by
selecting different MBR files -- in fact, these are different MBR files
which are already built, and this is how isohybrid handles it.

So all in all I think the only "essential" piece here is that the MBR
code should come out of the mbr/ files in the Syslinux build, instead of
being compiled into xorriso/libisofs, unless, of course, isolinux.bin
itself is also compiled in.

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.