[PATCH 33/59] x86: Fix modf{f} build on clang

[H.J. Lu]

On Sat, Oct 18, 2025 at 3:50 AM Adhemerval Zanella
<adhemerval.zanella@linaro.org> wrote:
>
> Similar to the generic implementation, also use NO_MATH_REDIRECT
> for the ifunc optimizations.

Does it change GCC codegen?

> ---
>  sysdeps/x86_64/fpu/multiarch/s_modf-avx.c     | 1 +
>  sysdeps/x86_64/fpu/multiarch/s_modf-sse4_1.c  | 1 +
>  sysdeps/x86_64/fpu/multiarch/s_modff-avx.c    | 1 +
>  sysdeps/x86_64/fpu/multiarch/s_modff-sse4_1.c | 1 +
>  4 files changed, 4 insertions(+)
>
> diff --git a/sysdeps/x86_64/fpu/multiarch/s_modf-avx.c b/sysdeps/x86_64/fpu/multiarch/s_modf-avx.c
> index 0982280d25..526262cd0c 100644
> --- a/sysdeps/x86_64/fpu/multiarch/s_modf-avx.c
> +++ b/sysdeps/x86_64/fpu/multiarch/s_modf-avx.c
> @@ -1,3 +1,4 @@
> +#define NO_MATH_REDIRECT
>  #include <math_private.h>
>
>  #define __modf __modf_avx
> diff --git a/sysdeps/x86_64/fpu/multiarch/s_modf-sse4_1.c b/sysdeps/x86_64/fpu/multiarch/s_modf-sse4_1.c
> index f6fb996f97..01cafcff7a 100644
> --- a/sysdeps/x86_64/fpu/multiarch/s_modf-sse4_1.c
> +++ b/sysdeps/x86_64/fpu/multiarch/s_modf-sse4_1.c
> @@ -1,3 +1,4 @@
> +#define NO_MATH_REDIRECT
>  #include <math_private.h>
>
>  #define __modf __modf_sse41
> diff --git a/sysdeps/x86_64/fpu/multiarch/s_modff-avx.c b/sysdeps/x86_64/fpu/multiarch/s_modff-avx.c
> index b2afe1efe3..13741e802a 100644
> --- a/sysdeps/x86_64/fpu/multiarch/s_modff-avx.c
> +++ b/sysdeps/x86_64/fpu/multiarch/s_modff-avx.c
> @@ -1,3 +1,4 @@
> +#define NO_MATH_REDIRECT
>  #include <math_private.h>
>
>  #define __modff __modff_avx
> diff --git a/sysdeps/x86_64/fpu/multiarch/s_modff-sse4_1.c b/sysdeps/x86_64/fpu/multiarch/s_modff-sse4_1.c
> index 0352c3ea4b..651bc8aa75 100644
> --- a/sysdeps/x86_64/fpu/multiarch/s_modff-sse4_1.c
> +++ b/sysdeps/x86_64/fpu/multiarch/s_modff-sse4_1.c
> @@ -1,3 +1,4 @@
> +#define NO_MATH_REDIRECT
>  #include <math_private.h>
>
>  #define __modff __modff_sse41
> --
> 2.43.0
>


-- 
H.J.