Re: [PATCH] Add Blackfin support in newlib (try 3) (was Re: [PATCH] Add Blackfin support in newlib (try 2))

[Jie Zhang <jzhang918 at gmail dot com>]

Sorry for the delay of this reply.

Joel Sherrill wrote:
> Jie Zhang wrote:
> > Hi all,
> >
> > Here is a third patch for adding Blackfin support in newlib. I should 
> > send it out sooner. But I was waiting for Bernd's opinion without 
> > knowing he's on vocation.
> >
> > This patch has fixed these more issues:
> It looks like it is getting there.  The license is OK for newlib now and 
> you have setjmp/longjmp.
> AFAIK machine/bfin should not have access.c or sys/syscall.h.  Those are 
> part of the target OS
> which in your case is libgloss.  For RTEMS, we could use setjmp/longjmp 
> but would
> avoid those since we really have a functional access and those syscall's 
> are meaningless.
So is it OK to move syscall.h into libgloss and remove access.c?

And further, is it a good idea to reuse the common syscall.h in 
libgloss? I did some work in the last week to do so. And I found we need 
three new syscalls to be added into the common syscall.h to support argc 
and argv: SYS_argc, SYS_argnlen and SYS_argn. Is it OK to add them to 
the common libgloss syscall.h?

Thanks,
Jie