drochner@zel459.zel.kfa-juelich.de wrote:

> 
> simonb@telstra.com.au said:
> > > I can put my stuff up for FTP somewhere if you want.
> > Yes, please! 
> 
> I've put it to
> zelof1.zel.kfa-juelich.de:/pub/NetBSD/bootcode/pmax-netboot.tar.gz
> 
> It does probably not compile out-of-the box anymore due
> to header file rearrangements, but should be easy to do.

Ta, got it now.  I'll check it out.

> > At this stage, I disagree with this ("easily be called by a macro").  I
> > can't see a single place where I can say "#define bzero(a, b) memset(a,
> > 0, b)" that will affect all of libsa/libkern.  Proof the contrary gladly
> > accepted :-)
> 
> Perhaps a hook to an md include file should be added
> to "stand.h".

What I've already done is add defines for memset/memcpy only.  Perhaps
something like:

	#ifdef LIBSA_MACHINE_STAND_INCLUDE
	#include "machine_stand.h"
	#endif

would be cleaner - we could just add what we want.  I'll keep this in
mind if I need more MD-type changes to MI files.

> One could expand printf() and other
> PROM supplied functions into indirect calls without need
> for assembler stubs then.

Beleive it or not, it's cheaper space-wise to use the stubs - all the
pmax-specific code included <dec_prom.h> which defined a lot of routines
as indirect calls.  Changing these to function calls using the assembler
stubs saved quite a bit of space!

Simon.