Re: [PATCH,RS6000 2/5] Rework the RS6000_BTM defines.

["Kewen.Lin" <linkw@linux.ibm.com>]

Hi Will,

The whole series looks good to me, thanks!  IMHO one place can be
further refactored, not sure if it's worth to updating together in
this series, it's ...

on 2022/6/7 06:05, will schmidt wrote:
> [PATCH,RS6000 2/5) Rework the RS6000_BTM defines.
> 
> The RS6000_BTM_<xxxx> definitions are mostly unused after the rs6000
> builtin code was reworked.  The remaining references can be replaced
> with the OPTION_MASK_<xxxx> and MASK_<xxxx> equivalents.
> 
> This patch remvoes the defines:
> RS6000_BTM_FRES, RS6000_BTM_FRSQRTE, RS6000_BTM_FRSQRTES,
> RS6000_BTM_POPCNTD, RS6000_BTM_CELL, RS6000_BTM_DFP,
> RS6000_BTM_HARD_FLOAT, RS6000_BTM_LDBL128, RS6000_BTM_64BIT,
> RS6000_BTM_POWERPC64, RS6000_BTM_FLOAT128, RS6000_BTM_FLOAT128_HW
> RS6000_BTM_MMA, RS6000_BTM_P10.
> 
> I note that the BTM -> OPTION_MASK mappings are not always 1-to-1.
> in particular the BTM_FRES and BTM_FRSQRTE values were both mapped to
> OPTION_MASK_PPC_GFXOPT, while the BTM_FRE and BTM_FRSQRTES both mapped
> to OPTION_MASK_POPCNTB.  In total I spent quite a bit of time
> double-checking these since it looked like copy/paste errors.  I split
> some of these changes out into a subsequent patch to limit the amount
> of potential confusion in any particular patch.
> 
> gcc/
> 	* config/rs6000/rs6000-c.cc: Update comments.
> 	* config/rs6000/rs6000.cc (RS6000_BTM_FRES, RS6000_BTM_FRSQRTE,
> 	RS6000_BTM_FRSQRTES, RS6000_BTM_POPCNTD, RS6000_BTM_CELL,
> 	RS6000_BTM_64BIT, RS6000_BTM_POWERPC64, RS6000_BTM_DFP,
> 	RS6000_BTM_HARD_FLOAT,RS6000_BTM_LDBL128, RS6000_BTM_FLOAT128,
> 	RS6000_BTM_FLOAT128_HW, RS6000_BTM_MMA, RS6000_BTM_P10): Replace
> 	with OPTION_MASK_PPC_GFXOPT, OPTION_MASK_PPC_GFXOPT,
> 	OPTION_MASK_POPCNTB, OPTION_MASK_POPCNTD,
> 	OPTION_MASK_FPRND, MASK_64BIT, MASK_POWERPC64,
> 	OPTION_MASK_DFP, OPTION_MASK_SOFT_FLOAT, OPTION_MASK_MULTIPLE,
> 	OPTION_MASK_FLOAT128_KEYWORD, OPTION_MASK_FLOAT128_HW,
> 	OPTION_MASK_MMA, OPTION_MASK_POWER10.
> 	* config/rs6000/rs6000.h (RS6000_BTM_FRES, RS6000_BTM_FRSQRTE,
> 	RS6000_BTM_FRSQRTES, RS6000_BTM_POPCNTD, RS6000_BTM_CELL,
> 	RS6000_BTM_DFP, RS6000_BTM_HARD_FLOAT, RS6000_BTM_LDBL128,
> 	RS6000_BTM_64BIT, RS6000_BTM_POWERPC64, RS6000_BTM_FLOAT128,
> 	RS6000_BTM_FLOAT128_HW, RS6000_BTM_MMA, RS6000_BTM_P10): Delete.
> 
> diff --git a/gcc/config/rs6000/rs6000-c.cc b/gcc/config/rs6000/rs6000-c.cc
> index 9c8cbd7a66e4..4c99afc761ae 100644
> --- a/gcc/config/rs6000/rs6000-c.cc
> +++ b/gcc/config/rs6000/rs6000-c.cc
> @@ -594,13 +594,13 @@ rs6000_target_modify_macros (bool define_p, HOST_WIDE_INT flags,
>       via the target attribute/pragma.  */
>    if ((flags & OPTION_MASK_FLOAT128_HW) != 0)
>      rs6000_define_or_undefine_macro (define_p, "__FLOAT128_HARDWARE__");
>  
>    /* options from the builtin masks.  */
> -  /* Note that RS6000_BTM_CELL is enabled only if (rs6000_cpu ==
> -     PROCESSOR_CELL) (e.g. -mcpu=cell).  */
> -  if ((bu_mask & RS6000_BTM_CELL) != 0)
> +  /* Note that OPTION_MASK_FPRND is enabled only if
> +     (rs6000_cpu == PROCESSOR_CELL) (e.g. -mcpu=cell).  */
> +  if ((bu_mask & OPTION_MASK_FPRND) != 0)
>      rs6000_define_or_undefine_macro (define_p, "__PPU__");
>  

... here.  In function rs6000_target_modify_macros, bu_mask is used by
two places, the beginning debug outputting and the above OPTION_MASK_FPRND
check.  I wonder if we can get rid of bu_mask and just use sth. like:

(rs6000_cpu == PROCESSOR_CELL) && (flags & OPTION_MASK_FPRND)

// the others are using "flags &", it's passed by rs6000_isa_flags,
// should be the same as just using OPTION_MASK_FPRND.

If we drop bu_mask in function rs6000_target_modify_macros, function
rs6000_builtin_mask_calculate will have only one use place in function
rs6000_option_override_internal.  IMHO this function
rs6000_builtin_mask_calculate also becomes stale after built-in function
rewriting and needs some updates with new bif framework later.

BR,
Kewen