Hello,

   I'm not subscribed to this list yet (will fix that), but just stumbled over 
it via
google. Sorry for the late reply.

I think the refactoring of libld into a cross linker should help with the ppc 
work.
I'm sure Brian will recognize his influence in all of this...  :-)

I didn't move the polaris-trunk stuff that Bill pointed me at into the main 
Solaris
branch though. Two reasons: (1) The polaris-trunk code is ~2 years old, and the
linker code has moved since then, and (2) The code seems unfinished, and it
would be better for someone who knows what's going on to do the moving.

(By unfinished, I mean I saw lots of "#if 0"s, and comments along the
lines of "do we need this?". Typical porting stuff)

The port of your work to the new framework ought to be straightforward
though, based on what I saw. Let me point you at the README I wrote,
as a starting point:

http://cvs.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/sgs/libld/common/README.XLINK

It's not a simple task, but it should make sense to the folks currently 
involved with
the ppc linker code. Once you've done that work, you should be able to build
Solaris for sparc or x86, and the resulting linker should be able to link your 
ppc
objects. This should eliminate the GNU ld impedance mismatch and let you get on
with things...

- Ali
 
 
This message posted from opensolaris.org

Reply via email to