Subject: Re: Changes to libexec/ld.aout_so/arch/sparc/md.c
To: Bill Studenmund <wrstuden@nas.nasa.gov>
From: Eduardo E. Horvath <eeh@one-o.com>
List: tech-toolchain
Date: 01/17/2000 14:54:24
> 
> On Sat, 15 Jan 2000, Matt Thomas wrote:
> 
> > You need to include the iflush code if RTLD is defined or a cross-built 
> > ld.aout_so will be wrong.
> 
> Right.
> 
> These patches do
> 
> #ifndef CROSS_LINKER
> 	iflush(where)
> #endif
> 
> So they should be fine for a cross-built ld.aout_so (CROSS_LINKER is only
> defined during the building of ld as a cross_linker AFAIK).

I don't understand this.  ld.*_so is a dynamic linker.  While it can
be cross-compiled, how do you expect it to cross-link?  If it's
cross-compiled, it still needs to issue the proper iflush instructions
there is potential corruption.

If you can generate a SPARC binary containing SPARC instructions, why
can't you generate an iflush instruction?  If you're not generating a
SPARC binary with SPARC instructions, how do you expect it to work as
a dynamic linker?

=========================================================================
Eduardo Horvath				eeh@netbsd.org
	"I need to find a pithy new quote." -- me