[rfc/rft] Update remaining targets to value-based unwinding

[Daniel Jacobowitz]

On Thu, May 29, 2008 at 06:45:12PM +0200, Mark Kettenis wrote:
> > Date: Thu, 29 May 2008 12:40:03 -0400
> > From: Daniel Jacobowitz <drow@false.org>
> > 
> > On Sun, May 18, 2008 at 10:15:25PM +0200, Ulrich Weigand wrote:
> > Hi Ulrich,
> > 
> > At this point I think you've gotten all the likely testing; is it time
> > to commit this?
> 
> I'm already looking at m88k; if you commit it Ulrich, can you leave
> that bit out?

Mark, did you have a chance to finish m88k?  Ulrich, how do you feel
about committing the non-m88k bits now?  Or I can take care of it if
that's easier.

I'm working on removing some warts from the version of inlined
function support I posted.  This requires careful reasoning about each
of the frame_*_unwind and frame_unwind_* methods, which is much easier
when they are not used in any unwinders; so I'm working from a tree
with Ulrich's patch applied.

-- 
Daniel Jacobowitz
CodeSourcery