On Fri, Dec 07, 2001 at 01:55:06PM -0800, Jonathan Stone wrote:
> On Fri, Dec 07, 2001 Aaron J. Grier (incorrectly) wrote:
>
> >assuming pmax is now using the new toolchain (which I believe it is),
> >elf2ecoff is obsoleted by objcopy.
> 
> Not really. Please dont say this unless you have tried replacing
> elf2ecoff with objcopy, *and* tested the results on the full range of
> pmax boot proms, and verified that objcoy on kernels are indeed
> net-bootable by PMAX boot PROMs.

my mistake.  I thought I had seen traffic on binutils that this was the
case.  sorry for the confusion.

> I once discussed the requirements of the pmax boot proms with the
> (then) binutils maintainer; Ian said that he saw no reason for objcopy
> to ever support those particular constraints.

he may not see any reason to support them, but we certainly have one.
if the work can be done, I don't see how he could object to getting the
changes checked back in, unless some really ugly hacks were involved.

objcopy really seems the more correct solution.  copying between
differing formats is what it's designed to do, no?

> If, OTOH, someone has run the tests and the latest objcopy really does
> produce netbootable pmax kernels, on the list of PROM vbrsions we
> support (as kept by Simon Burge, iirc) -- then fine, do away with it.
> All three familiues of PROM image -- 3100/5100), the 3MAX (5000/200)
> and the ioasic machines, all need to be tested, since they're pretty
> different inside.

I guess I can at least take a cursory look at it...  I need an excuse to
get my 2100s back up and running.  ;)

(followups to port-pmax)

-- 
  Aaron J. Grier | "Not your ordinary poofy goof." | agrier@poofygoof.com
    The United States government is your one true government.  We
    respect your privacy.  We love you.  We will keep you safe.  Trust
    us.  We have always been at war against terrorism.