Re: [PATCH 2/2] gdb: Enable finish command and inferior calls for _Float16 on amd64 and i386.

[Luis Machado <luis.machado@linaro.org>]

On 7/30/21 4:52 AM, Willgerodt, Felix wrote:
>> Is _Float16 equivalent to fp16? If so, the presence of yet another name for a
>> fp16 type is a bit unfortunate.
>>
> 
> They are the same usually, but on ARM the story is apparently a bit more complex:
> https://gcc.gnu.org/onlinedocs/gcc/Half-Precision.html

It is. :-)

> 
> But to cite the link:
> " The __fp16 type may only be used as an argument to intrinsics defined in
> <arm_fp16.h>, or as a storage format."
> 
> And:
> "It is recommended that portable code use the _Float16 type defined by
> ISO/IEC TS 18661-3:2015."
> 
> Therefore I don't think it is that unfortunate and not really negotiable by now.
> Both GCC and clang know about both already.

The unfortunate bit is the lack of enough DWARF information to make the 
distinction between different float formats of the same size. If the 
compilers are OK with the naming and already support it, it's not a big 
deal then.

> 
>>> -  if (type->code () == TYPE_CODE_FLT)
>>> +  /* _FLoat16 and _Float16 _Complex values are returned via xmm0.  */
>>
>> _Float16
>>
> 
> Thanks for spotting this. I fixed it locally.
> 
>>> +/* Test program for bfloat16 of AVX 512 registers.
>>
>> float16 and not bfloat16, right?
>>
> 
> Thanks. I reworded it to say:
> "Test program for _Float16 parameters and return values."
> 
> Regards,
> Felix
> Intel Deutschland GmbH
> Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
> Tel: +49 89 99 8853-0, www.intel.de <http://www.intel.de>
> Managing Directors: Christin Eisenschmid, Sharon Heck, Tiffany Doon Silva
> Chairperson of the Supervisory Board: Nicole Lau
> Registered Office: Munich
> Commercial Register: Amtsgericht Muenchen HRB 186928
>