On Mon, Dec 16, 2002 at 11:56:23PM +0100, Aymeric Vincent wrote:
> Gunther Nikl <gni@gecko.de> writes:
> 
> >   Before I try to integrate my code into loadbsd, I would like to know who
> >   much interest in such a loadbsd exists.
> 
> there is definitely interest in having a "loadbsd" for ELF
> kernels, especially for installing new (e.g. "diskless")
> systems.

  Good, atleast one person besides me shows some interest :-)

> However, I think it would be a shame if we couldn't re-use loadfile().

  Well, loadfile is very tied to the NetBSD headers. Compiling it in a
  different environment isn't as trivial as it looks. I wasn't able to
  compile loadfile_aout.c with amigaos-gcc (its probably possible but
  I wasn't motivated to fix the header clashes). Anyway, compiling
  netbsd sources in an alien environment is hacky.

> Actually, I think a good solution would be to modify runbootblock so
> that it can get the bootblock in a file instead of on a partition.

  Yes, that would be a valuable enhancement to runbootblock.

> Having a complex stand-alone loadbsd, beside the fact that it
> potentially duplicates code, means that we need to have a working
> toolchain under AmigaOS to release new versions of that loader.

  You will need it anyway if you don't want/can't use the bootblocks
  directly.

> On the contrary, reading a bootblock file lets us compile it from
> under NetBSD, and makes automatic updates of said utility easier.

  Right, using the bootblocks is the best way to ensure compatibilty.
  I am aware of this. I tried the bootblock once and I did't like it.
  I prefer to pass arguments on the commandline.

> But anyway, keep your code around, I think that if you produce a
> working ELF loadbsd, many people will be happy to use it.

  If there is real interest, I will finish and donate the code.

  Gunther

-- 
Linux is only free if your time has no value
 - Jamie Zawinsky