This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC/CFT] Hookize TARGET_UNWIND_INFO and related macros
On Mon, 2010-09-27 at 15:26 -0700, Richard Henderson wrote:
> On 09/27/2010 12:27 PM, Richard Henderson wrote:
> > On 09/27/2010 09:09 AM, Steve Ellcey wrote:
> >> I tried using this patch and the stage2/stage3 comparision failed on
> >> most files, for both ia64 HP-UX and Linux. If I compile cfg.o to a .s
> >> file and do a diff I see:
> >
> > Got it. The real difference is:
> >
> > alloc r34 = ar.pfs, 1, 3, 1, 0
> > +.LCFI0:
> > addl r14 = @ltoffx(_xexit_cleanup#), r1
> >
> > i.e. the dwarf2 cfi stuff isn't creating debug labels,
> > which means we break the bundle and add nops where we
> > didn't want em.
> >
> > Hopefully I'll have a new patch this afternoon so you
> > can do overnight tests.
>
> Whee. Tested along with the T_U_I patch on ia64-linux
> and committed.
>
>
> r~
I got good bootstraps on IA64 HP-UX and Linux but I don't know if there
are any regressions because I am getting lots of failures with
/proj/opensrc/nightly/src/trunk/gcc/testsuite/gcc.c-torture/compile/20000105-1.c:25:1: sorry, unimplemented: gimple bytecode streams do not support machine specific builtin functions on this target
This is unrelated to your patch but is due to Richard Guenther's patch
to always use TRANSLATION_UNIT_DECL as context.
Steve Ellcey
sje@cup.hp.com