Re: PATCH [1/n] X32: Add initial -x32 support

[Uros Bizjak <ubizjak@gmail.com>]

On Tue, Jul 5, 2011 at 4:39 PM, H.J. Lu <hongjiu.lu@intel.com> wrote:

> I'd like to start submitting a series of patches to enable x32:
>
> https://sites.google.com/site/x32abi/
>
> The GCC x32 branch is very stable. There are no unexpected failures in
> C, C++, Fortran and Objective C testsuites.  SPEC CPU 2K/2006 compile
> and run correctly at -O2 and -O3.
>
> More than 90% of changes are in x86 backend.  I have submitted non-x86
> backend patches.  Most of them have been reviewed and checked in.  Only
> 4 patches are pending for review/approval.
>
> This is the first x86 backend patch to support x32.  By default, x32 is
> disabled and x32 run-time support isn't required.  OK for trunk?

Please strip out --enable-ia32 stuff, it complicates things ATM.  I
assume that --enable-x32 applies only to 64bit targets, so this part

> @@ -1232,7 +1232,17 @@ i[34567]86-*-linux* | i[34567]86-*-kfreebsd*-gnu | i[34567]86-*-knetbsd*-gnu | i
>                if test x$enable_targets = xall; then
>                        tm_file="${tm_file} i386/x86-64.h i386/gnu-user64.h i386/linux64.h"
>                        tm_defines="${tm_defines} TARGET_BI_ARCH=1"
> -                       tmake_file="${tmake_file} i386/t-linux64"
> +                       case x${enable_x32}${enable_ia32} in
> +                       xyesyes)
> +                               tmake_file="${tmake_file} i386/t-linux-x32"
> +                               ;;
> +                       xyesno)
> +                               tmake_file="${tmake_file} i386/t-linux64-x32"
> +                               ;;
> +                       *)
> +                               tmake_file="${tmake_file} i386/t-linux64"
> +                               ;;
> +                       esac

should be simplified to something:

if (enable_x32)
  tmake_file = "... i386/t-linux64-x32"
else
  tmake_file = "... i386/t-linux64"



> diff --git a/gcc/config/i386/gnu-user64.h b/gcc/config/i386/gnu-user64.h
> index b069975..954f3b2 100644
> --- a/gcc/config/i386/gnu-user64.h
> +++ b/gcc/config/i386/gnu-user64.h
> @@ -58,25 +58,31 @@ see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
>
>  #if TARGET_64BIT_DEFAULT
>  #define SPEC_32 "m32"
> -#define SPEC_64 "!m32"
> +#define SPEC_64 "m32|mx32:;"
> +#define SPEC_X32 "mx32"
>  #else
> -#define SPEC_32 "!m64"
> +#define SPEC_32 "m64|mx32:;"
>  #define SPEC_64 "m64"
> +#define SPEC_X32 "mx32"
>  #endif
>
>  #undef ASM_SPEC
> -#define ASM_SPEC "%{" SPEC_32 ":--32} %{" SPEC_64 ":--64} \
> +#define ASM_SPEC "%{" SPEC_32 ":--32} \
> + %{" SPEC_64 ":--64} \
> + %{" SPEC_X32 ":--x32} \
>  %{!mno-sse2avx:%{mavx:-msse2avx}} %{msse2avx:%{!mavx:-msse2avx}}"

Are you sure that above is correct? AFAICS, you are enabling SPEC_64
for m32 in case of TARGET_64BIT_DEFAULT.

>  /* SSE4.1 defines round instructions */
>  #define        OPTION_MASK_ISA_ROUND   OPTION_MASK_ISA_SSE4_1
> @@ -517,8 +519,8 @@ extern const char *host_detect_local_cpu (int argc, const char **argv);
>  #define OPT_ARCH64 "!m32"
>  #define OPT_ARCH32 "m32"
>  #else
> -#define OPT_ARCH64 "m64"
> -#define OPT_ARCH32 "!m64"
> +#define OPT_ARCH64 "m64|mx32"
> +#define OPT_ARCH32 "m64|mx32:;"
>  #endif

Same here.

Uros.