This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH v3] powerpc: Refactor fenvinline.h



On 25/02/2020 15:23, Tulio Magno Quites Machado Filho wrote:
> Rogerio Alves <rcardoso@linux.ibm.com> writes:
> 
>>  /* The weird 'i#*X' constraints on the following suppress a gcc
>>     warning when __excepts is not a constant.  Otherwise, they mean the
>> -   same as just plain 'i'.  */
>> +   same as just plain 'i'. This warning only happens in old GCC 
> 
> 2 spaces here ---------------^     trailing whitespace here ------^
> 
>> +   versions (gcc 3 or less). Otherwise plain 'i' works fine.  */
> 
> Likewise ----------------------^
> 
>> -	if (__e != 0)							      \
>> -	  __asm__ __volatile__ ("mtfsb1 %0"				      \
>> -				: : "i#*X" (__builtin_clz (__e)));	      \
>> -        __ret = 0;							      \
>> +	  __MTFSB1 ((__builtin_clz (__e)));				      \
> 
> Wrong indentation.
> 
>>          && __e != FE_INVALID)						      \
>>        {									      \
>> -	if (__e != 0)							      \
>> -	  __asm__ __volatile__ ("mtfsb0 %0"				      \
>> -				: : "i#*X" (__builtin_clz (__e)));	      \
>> -        __ret = 0;							      \
>> +	  __MTFSB0 ((__builtin_clz (__e)));				      \
> 
> Wrong indentation.
> 
>> @@ -80,15 +87,12 @@
>>  #   define feclearexcept(__excepts) \
>>    (__extension__  ({ 							      \
>>      int __e = __excepts;						      \
>> -    int __ret;								      \
>> +    int __ret = 0;							      \
>>      if (__builtin_constant_p (__e)					      \
>> -        && (__e & (__e - 1)) == 0					      \
>> +        && (__builtin_popcount (__e) == 1)				      \
> 
> I remember that Adhemerval disagreed with this particular change and I
> don't remember seeing any agreement there.

My reservation is the change is just an idiomatic one, it is not either
simplifying the code or optimizing it. I am ok with the usage of 
powerof2, as Florian suggested.

> Rogerio, if you agree with the removal of these lines I can remove this change
> from the patch, fix the cosmetic issues and push the patch.
> Does it look good for you?
> 


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