This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: SH: declare fpul_operands


> 
> --=-=-=
> 
> On Nov 24, 2000, Joern Rennecke <amylaar@redhat.com> wrote:
> 
> >> Any preference among the existing ones?  Or should it be a new header
> >> file?  Or should we just standardize the common practice of declaring
> >> the predicates in <mach>-protos.h and stop emitting the declarations
> >> in insn-output.c?
> 
> > No particular preference.
> 
> Then, how about this?

In general, fine with me.
Of course, you need approval of the non-SH parts from someone else.

Two minor points:

> --- gcc/machmode.h 2000/09/18 17:55:38 1.24
> +++ gcc/machmode.h 2000/11/24 22:44:30
> @@ -96,16 +96,21 @@ extern const unsigned int mode_unit_size
>  extern const unsigned int mode_bitsize[];
>  #define GET_MODE_BITSIZE(MODE)  (mode_bitsize[(int) (MODE)])
>  
> -#ifdef HOST_WIDE_INT
> +#endif

This should be
#endif /* not HAVE_MACHINE_MODES */

> --- gcc/config/sh/sh.c 2000/11/24 22:06:30 1.76
> +++ gcc/config/sh/sh.c 2000/11/24 22:44:37
> @@ -4815,6 +4815,7 @@ fpscr_operand (op, mode)
>  int
>  fpul_operand (op, mode)
>       rtx op;
> +     enum machine_mode mode ATTRIBUTE_UNUSED;

As I said before, mode is not unused.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]