patch: e500 change spu_const_offset_ok
Aldy Hernandez
aldyh@redhat.com
Thu Aug 1 20:01:00 GMT 2002
On Thu, Aug 01, 2002 at 05:52:59PM -0700, Richard Henderson wrote:
> On Thu, Aug 01, 2002 at 05:35:52PM -0700, Aldy Hernandez wrote:
> > > > [yes, I left 0xff and not 0xf8, because I want to catch any craziness
> > > > by gcc possibly not aligning stuff properly... It could happen ;-)]
> > >
> > > I do not think this is a good idea.
> >
> > Whoops. Then it's not obvious. I can back it out.
> >
> > Or would you rather me have 0xF8?
>
> You should have what the hardware supports.
Very well. Committing this.
Index: config/rs6000/rs6000.h
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/rs6000/rs6000.h,v
retrieving revision 1.217
diff -c -p -r1.217 rs6000.h
*** config/rs6000/rs6000.h 1 Aug 2002 21:18:34 -0000 1.217
--- config/rs6000/rs6000.h 2 Aug 2002 03:00:40 -0000
*************** typedef struct rs6000_args
*** 2030,2036 ****
#define TOC_RELATIVE_EXPR_P(X) (toc_relative_expr_p (X))
/* SPE offset addressing is limited to 5-bits worth of double words. */
! #define SPE_CONST_OFFSET_OK(x) (((x) & ~0xff) == 0)
#define LEGITIMATE_CONSTANT_POOL_ADDRESS_P(X) \
(TARGET_TOC \
--- 2030,2036 ----
#define TOC_RELATIVE_EXPR_P(X) (toc_relative_expr_p (X))
/* SPE offset addressing is limited to 5-bits worth of double words. */
! #define SPE_CONST_OFFSET_OK(x) (((x) & ~0xf8) == 0)
#define LEGITIMATE_CONSTANT_POOL_ADDRESS_P(X) \
(TARGET_TOC \
More information about the Gcc-patches
mailing list