Mark Martin wrote:
> On 10/3/07, *Tom Riddle* <tom.riddle at sun.com 
> <mailto:tom.riddle at sun.com>> wrote:
>
>     It would be a great feature to have sparc hosting capability. Is
>     anyone interested in tackling this? Don't be shy we can assist
>     with a spin-up on the technical details. Tom
>
>
>
> I've been slowly getting to a point to do this.  My rough plan was as 
> follows:
>
> 1) Refresh my memory on why I thought Sun  'ld' wasn't usable in the 
> binutils toolchain or throw that notion right out
Sun "ld" is usable.  We've been using it.  We "have" to use it.  Too 
many Solaris things utilize features of the
Solaris "ld".  (One that comes to mind is the ability to bunch sections 
by attributes into a segment.)
> 2) Get GCC 3.4.3 and binutils-??? from gnu.org <http://gnu.org> 
> archives.  Does the specific revision matter as much?  I know GCC
I don't think it does, but we have been using one version rather than 
keeping up with the latest and greatest.
> 4.x brings in c++0x which should probably be avoided.
No specific knowledge here.
> 3) Get Sun ld + Gnu as cross assembling together (binutils).  
> Obviously associated tools to verify elf outputs.
> 4) Get Gnu cc cross compiling
> 5) (nice to have) document everything and capture what lacks in Gnu ld

Note: we did find that the gnu PowerPC C compiler generated unaligned 
(and incorrect) relocations for
some of the dwarf stuff.  If you don't use the dwarf (-gdwarf2) stuff, 
no matter, but if you do, you need to
fix the assembler.  If you wish I can help point out exactly where 
(assuming it is still broken in later releases of
the GNU stuff).
>
>
>
> -- 
> ------------------------------------------------------
> Born to the false world, the wanderer,
> Storyteller, The Pied Piper
> On a quest for immortality
> Gathering a troop to find the fantasy
> -- Nightwish
> ------------------------------------------------------------------------
>
> _______________________________________________
> powerpc-discuss mailing list
> powerpc-discuss at opensolaris.org
>   


Reply via email to