[PATCH v3] x86: Define __HAVE_FLOAT128 for Clang and use __builtin_*f128 code path

[H.J. Lu]

On Fri, Dec 27, 2024, 11:26 AM Khem Raj <raj.khem@gmail.com> wrote:

> I am seeing compiler errors with this patch when compiling the mpfr
> package with clang. see
> https://0x0.st/8sHM.73658


Please create a glibc bug with a small testcase.


>
> On Fri, Dec 20, 2024 at 1:23 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > From: Fangrui Song <i@maskray.me>
> >
> > Clang supports __builtin_fabsf128 (despite not supporting _Float128) but
> > it does not support __builtin_fabsq.  Fallback to back to
> > `typedef __float128 _Float128;` it clang is used.
> >
> > Originally developed by Fangrui Song <maskray@google.com>.
> > Co-Authored-By: H.J. Lu <hjl.tools@gmail.com>
> > ---
> >  sysdeps/x86/bits/floatn.h | 24 ++++++++++++++++--------
> >  1 file changed, 16 insertions(+), 8 deletions(-)
> >
> > diff --git a/sysdeps/x86/bits/floatn.h b/sysdeps/x86/bits/floatn.h
> > index e661abaea4..ba030d270a 100644
> > --- a/sysdeps/x86/bits/floatn.h
> > +++ b/sysdeps/x86/bits/floatn.h
> > @@ -28,7 +28,8 @@
> >     support, for x86_64 and x86.  */
> >  #if (defined __x86_64__
>         \
> >       ? __GNUC_PREREQ (4, 3)                                            \
> > -     : (defined __GNU__ ? __GNUC_PREREQ (4, 5) : __GNUC_PREREQ (4, 4)))
> > +     : (defined __GNU__ ? __GNUC_PREREQ (4, 5) : __GNUC_PREREQ (4, 4)))
> \
> > +    || __glibc_clang_prereq (3, 4)
> >  # define __HAVE_FLOAT128 1
> >  #else
> >  # define __HAVE_FLOAT128 0
> > @@ -58,7 +59,9 @@
> >  /* Defined to concatenate the literal suffix to be used with _Float128
> >     types, if __HAVE_FLOAT128 is 1. */
> >  # if __HAVE_FLOAT128
> > -#  if !__GNUC_PREREQ (7, 0) || (defined __cplusplus && !__GNUC_PREREQ
> (13, 0))
> > +#  if !__GNUC_PREREQ (7, 0) \
> > +      || (defined __cplusplus && !__GNUC_PREREQ (13, 0)) \
> > +      || defined __clang__
> >  /* The literal suffix f128 exists only since GCC 7.0.  */
> >  #   define __f128(x) x##q
> >  #  else
> > @@ -68,7 +71,9 @@
> >
> >  /* Defined to a complex binary128 type if __HAVE_FLOAT128 is 1.  */
> >  # if __HAVE_FLOAT128
> > -#  if !__GNUC_PREREQ (7, 0) || (defined __cplusplus && !__GNUC_PREREQ
> (13, 0))
> > +#  if !__GNUC_PREREQ (7, 0) \
> > +      || (defined __cplusplus && !__GNUC_PREREQ (13, 0)) \
> > +      || defined __clang__
> >  /* Add a typedef for older GCC compilers which don't natively support
> >     _Complex _Float128.  */
> >  typedef _Complex float __cfloat128 __attribute__ ((__mode__ (__TC__)));
> > @@ -82,12 +87,14 @@ typedef _Complex float __cfloat128 __attribute__
> ((__mode__ (__TC__)));
> >  # if __HAVE_FLOAT128
> >
> >  /* The type _Float128 exists only since GCC 7.0.  */
> > -#  if !__GNUC_PREREQ (7, 0) || (defined __cplusplus && !__GNUC_PREREQ
> (13, 0))
> > +#  if !__GNUC_PREREQ (7, 0) \
> > +      || (defined __cplusplus && !__GNUC_PREREQ (13, 0)) \
> > +      || __glibc_clang_prereq (3, 4)
> >  typedef __float128 _Float128;
> >  #  endif
> >
> > -/* __builtin_huge_valf128 doesn't exist before GCC 7.0.  */
> > -#  if !__GNUC_PREREQ (7, 0)
> > +/* __builtin_huge_valf128 doesn't exist before GCC 7.0 nor Clang 7.0.
> */
> > +#  if !__GNUC_PREREQ (7, 0) && !__glibc_clang_prereq (7, 0)
> >  #   define __builtin_huge_valf128() ((_Float128) __builtin_huge_val ())
> >  #  endif
> >
> > @@ -96,7 +103,7 @@ typedef __float128 _Float128;
> >     Converting a narrower sNaN to _Float128 produces a quiet NaN, so
> >     attempts to use _Float128 sNaNs will not work properly with older
> >     compilers.  */
> > -#  if !__GNUC_PREREQ (7, 0)
> > +#  if !__GNUC_PREREQ (7, 0) && !defined __clang__
> >  #   define __builtin_copysignf128 __builtin_copysignq
> >  #   define __builtin_fabsf128 __builtin_fabsq
> >  #   define __builtin_inff128() ((_Float128) __builtin_inf ())
> > @@ -107,7 +114,8 @@ typedef __float128 _Float128;
> >  /* In math/math.h, __MATH_TG will expand signbit to __builtin_signbit*,
> >     e.g.: __builtin_signbitf128, before GCC 6.  However, there has never
> >     been a __builtin_signbitf128 in GCC and the type-generic builtin is
> > -   only available since GCC 6.  */
> > +   only available since GCC 6.  signbit is expanded to __builtin_signbit
> > +   after Clang 3.3.  */
> >  #  if !__GNUC_PREREQ (6, 0)
> >  #   define __builtin_signbitf128 __signbitf128
> >  #  endif
> > --
> > 2.47.1
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://sourceware.org/pipermail/libc-alpha/attachments/20241227/fab49c0b/attachment.htm>