[6/7] ldbl-128ibm: Let long double files have specific compiler flags

  From: Tulio Magno Quites Machado Filho <tuliom@linux.vnet.ibm.com>

Soon, powerpc64le will need to provide extra compiler flags to the long
double files in order to continue to build using the IBM 128-bit
extended floating point type as long double.
---
 sysdeps/ieee754/ldbl-128ibm/Makefile | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)
  

Ping.  Will commit friday if no requests for improvement.

On 2/14/20 10:10 AM, Paul E. Murphy wrote:
> From: Tulio Magno Quites Machado Filho <tuliom@linux.vnet.ibm.com>
> 
> Soon, powerpc64le will need to provide extra compiler flags to the long
> double files in order to continue to build using the IBM 128-bit
> extended floating point type as long double.
> ---
>   sysdeps/ieee754/ldbl-128ibm/Makefile | 22 +++++++++++++++++++---
>   1 file changed, 19 insertions(+), 3 deletions(-)
> 
> diff --git a/sysdeps/ieee754/ldbl-128ibm/Makefile b/sysdeps/ieee754/ldbl-128ibm/Makefile
> index bdba6cc6b5..c38d4f6ac2 100644
> --- a/sysdeps/ieee754/ldbl-128ibm/Makefile
> +++ b/sysdeps/ieee754/ldbl-128ibm/Makefile
> @@ -9,8 +9,24 @@ tests += tst-strtold-ldbl-128ibm
>   $(objpfx)tst-strtold-ldbl-128ibm: $(libm)
>   endif
> 
> +ldbl-tests = test-fmodl-ldbl-128ibm test-remainderl-ldbl-128ibm \
> +	     test-remquol-ldbl-128ibm test-canonical-ldbl-128ibm \
> +	     test-totalorderl-ldbl-128ibm
> +
>   ifeq ($(subdir),math)
> -tests += test-fmodl-ldbl-128ibm test-remainderl-ldbl-128ibm \
> -	 test-remquol-ldbl-128ibm test-canonical-ldbl-128ibm \
> -	 test-totalorderl-ldbl-128ibm
> +tests += $(ldbl-tests)
>   endif
> +
> +# Long double files may need extra CFLAGS.
> +ldbl-128ibm-routines = s_nexttoward s_nexttowardf \
> +		       $(type-ldouble-routines) \
> +		       $(subst F,$(type-ldouble-suffix),$(libm-compat-calls)) \
> +		       $(subst F,$(type-ldouble-suffix),$(libm-calls)) \
> +		       $(subst F,$(type-ldouble-suffix),$(calls)) \
> +		       $(foreach f,$(libm-narrow-fns), \
> +				 $(subst F,$(f), \
> +					 $(libm-narrow-types-ldouble-yes)))
> +$(foreach suf,$(all-object-suffixes), \
> +	  $(foreach r,$(ldbl-128ibm-routines) $(ldbl-tests), \
> +		    $(objpfx)$(r)$(suf))): \
> +  CFLAGS += $(type-ldouble-CFLAGS)
>
  

And committed.

On 3/3/20 2:18 PM, Paul E Murphy wrote:
> Ping.  Will commit friday if no requests for improvement.
> 
> On 2/14/20 10:10 AM, Paul E. Murphy wrote:
>> From: Tulio Magno Quites Machado Filho <tuliom@linux.vnet.ibm.com>
>>
>> Soon, powerpc64le will need to provide extra compiler flags to the long
>> double files in order to continue to build using the IBM 128-bit
>> extended floating point type as long double.
>> ---
>>   sysdeps/ieee754/ldbl-128ibm/Makefile | 22 +++++++++++++++++++---
>>   1 file changed, 19 insertions(+), 3 deletions(-)