[rs6000] Clean up the option_mask defines (part 1)

  [PATCH, rs6000] Clean up the option_mask defines
    
Hi,

We have an assortment of MASK and OPTION_MASK #defines throughout
the rs6000 code, MASK_ALTIVEC and OPTION_MASK_ALTIVEC as an example.

We currently #define the MASK_<xxxx> entries to their OPTION_MASK_<xxxx>
equivalents so the two names could be used interchangeably.

The mapping is in place from when we switched from using
target_flags to rs6000_isa_flags via
commit 4d9675496a28ef6184f2a9c3ac5e6e3ea63606c1 in 2012.

This patch converts the references for most of the lingering MASK_*
values to OPTION_MASK_*  and removes the now redundant defines.

I have split this into multiple parts due to size.

Regtested OK on power10.  OK for trunk?

gcc/
	* rs6000.h (MASK_ALTIVEC, MASK_CMPB, MASK_CRYPTO
	MASK_DLMZB, MASK_EABI, MASK_FPRND, MASK_ISEL
	MASK_MFCRF, MASK_MULHW, MASK_POPCNTB, MASK_PPC_GFXOPT
	MASK_PPC_GPOPT):  Remove defines.
	(RS6000_BTM_ALTIVEC, RS6000_BTM_CMPB, RS6000_BTM_CRYPTO,
	RS6000_BTM_FRE, RS6000_BTM_FRES, RS6000_BTM_FRSQRTE,
	RS6000_BTM_FRSQRTES, RS6000_BTM_CELL) : Redefine using
	OPTION_MASK_<xxxx> instead of MASK_<xxxx>.
	* rs6000-cpus.def (RS6000_CPU) Update macro calls to use
	OPTION_MASK_<xxxx> instead of MASK_<xxxx>.
	* rs6000.cc (rs6000_darwin_file_start): Update mapping[] table
	entries to use OPTION_MASK_PPC_GPOPT, OPTION_MASK_MFCRF,
	OPTION_MASK_ALTIVEC instead of their MASK_<xxx> variants.
	* rs6000-c.cc : Update comment to reference OPTION_MASK_GFXOPT.
	* aix71.h (TARGET_DEFAULT): Update define to use OPTION_MASK_<xxxx>
	instead of MASK_<xxxx>.
	* darwin.h (TARGET_DEFAULT): Same.
	* darwin64-biarch.h (TARGET_DEFAULT): Same.
	* default64.h (TARGET_DEFAULT): Same.
	* eabi.h (TARGET_DEFAULT): Same.
	* eabialtivec.h (TARGET_DEFAULT): Same.
	* linuxaltivec.h (TARGET_DEFAULT): Same.
	* vxworks.h (TARGET_DEFAULT): Same.
  

[PATCH, rs6000] Clean up the option_mask defines (part 2)

Hi,
This patch reworks most of the lingering MASK_*
values to OPTION_MASK_* and removes the now redundant defines.

Regtested OK on power10.  OK for trunk?
    
gcc/
	* rs6000.h (RS6000_BTM_VSX, RS6000_BTM_P8_VECTOR, RS6000_BTM_P9_VECTOR,
	RS6000_BTM_P9_MISC, RS6000_BTM_HTM, RS6000_BTM_POPCNTD,
	RS6000_BTM_DFP, RS6000_BTM_HARD_FLOAT, RS6000_BTM_LDBL128,
	RS6000_BTM_FLOAT128, RS6000_BTM_FLOAT128_HW, RS6000_BTM_MMA,
	RS6000_BTM_P10): Rework defines to use OPTION_MASK_<xxxx>.
	(MASK_DFP, MASK_DIRECT_MOVE, MASK_FLOAT128_KEYWORD,
	MASK_FLOAT128_HW, MASK_P8_FUSION, MASK_HARD_FLOAT, MASK_HTM,
	MASK_MMA, MASK_MULTIPLE, MASK_NO_UPDATE, MASK_P8_VECTOR,
	MASK_P9_VECTOR, MASK_P9_MISC, MASK_POPCNTD, MASK_RECIP_PRECISION,
	MASK_SOFT_FLOAT, MASK_UPDATE, MASK_VSX, MASK_POWER10,
	MASK_P10_FUSION): Remove unused defines.
	* config/rs6000/rs6000-cpus.def (RS6000_CPU): Rework macro calls to
	use OPTION_MASK_<xxxx> defines.
	* config/rs6000/darwin.h (TARGET_DEFAULT) Update define to use
	OPTION_MASK_MULTIPLE.
	* config/rs6000/darwin64-biarch.h (TARGET_DEFAULT): Same.

diff --git a/gcc/config/rs6000/darwin.h b/gcc/config/rs6000/darwin.h
index 86556ccbbf58..6a8845eb3bb7 100644
--- a/gcc/config/rs6000/darwin.h
+++ b/gcc/config/rs6000/darwin.h
@@ -365,11 +365,11 @@
 /* Default target flag settings.  Despite the fact that STMW/LMW
    serializes, it's still a big code size win to use them.  Use FSEL by
    default as well.  */
 
 #undef  TARGET_DEFAULT
-#define TARGET_DEFAULT (MASK_MULTIPLE | OPTION_MASK_PPC_GFXOPT)
+#define TARGET_DEFAULT (OPTION_MASK_MULTIPLE | OPTION_MASK_PPC_GFXOPT)
 
 /* Darwin always uses IBM long double, never IEEE long double.  */
 #undef  TARGET_IEEEQUAD
 #define TARGET_IEEEQUAD 0
 
diff --git a/gcc/config/rs6000/darwin64-biarch.h b/gcc/config/rs6000/darwin64-biarch.h
index 6a700c61c4c2..6515bcc8bf5a 100644
--- a/gcc/config/rs6000/darwin64-biarch.h
+++ b/gcc/config/rs6000/darwin64-biarch.h
@@ -19,11 +19,11 @@
    along with GCC; see the file COPYING3.  If not see
    <http://www.gnu.org/licenses/>.  */
 
 #undef  TARGET_DEFAULT
 #define TARGET_DEFAULT (MASK_POWERPC64 | MASK_64BIT \
-			| MASK_MULTIPLE | OPTION_MASK_PPC_GFXOPT)
+			| OPTION_MASK_MULTIPLE | OPTION_MASK_PPC_GFXOPT)
 
 #undef DARWIN_ARCH_SPEC
 #define DARWIN_ARCH_SPEC "%{m32:ppc;:ppc64}"
 
 /* Actually, there's really only 970 as an active option.  */
diff --git a/gcc/config/rs6000/rs6000-cpus.def b/gcc/config/rs6000/rs6000-cpus.def
index ca78bd8cf89f..4301b1bcb120 100644
--- a/gcc/config/rs6000/rs6000-cpus.def
+++ b/gcc/config/rs6000/rs6000-cpus.def
@@ -174,29 +174,31 @@
 
    RS6000_CPU (NAME, CPU, FLAGS)
 
    where the arguments are the fields of struct rs6000_ptt.  */
 
-RS6000_CPU ("401", PROCESSOR_PPC403, MASK_SOFT_FLOAT)
-RS6000_CPU ("403", PROCESSOR_PPC403, MASK_SOFT_FLOAT | MASK_STRICT_ALIGN)
-RS6000_CPU ("405", PROCESSOR_PPC405, MASK_SOFT_FLOAT | OPTION_MASK_MULHW
-	    | OPTION_MASK_DLMZB)
+RS6000_CPU ("401", PROCESSOR_PPC403, OPTION_MASK_SOFT_FLOAT)
+RS6000_CPU ("403", PROCESSOR_PPC403, OPTION_MASK_SOFT_FLOAT | MASK_STRICT_ALIGN)
+RS6000_CPU ("405", PROCESSOR_PPC405, OPTION_MASK_SOFT_FLOAT
+	    | OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
 RS6000_CPU ("405fp", PROCESSOR_PPC405, OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
-RS6000_CPU ("440", PROCESSOR_PPC440, MASK_SOFT_FLOAT | OPTION_MASK_MULHW
+RS6000_CPU ("440", PROCESSOR_PPC440, OPTION_MASK_SOFT_FLOAT | OPTION_MASK_MULHW
 	    | OPTION_MASK_DLMZB)
 RS6000_CPU ("440fp", PROCESSOR_PPC440, OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
-RS6000_CPU ("464", PROCESSOR_PPC440, MASK_SOFT_FLOAT | OPTION_MASK_MULHW
+RS6000_CPU ("464", PROCESSOR_PPC440, OPTION_MASK_SOFT_FLOAT | OPTION_MASK_MULHW
 	    | OPTION_MASK_DLMZB)
 RS6000_CPU ("464fp", PROCESSOR_PPC440, OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
-RS6000_CPU ("476", PROCESSOR_PPC476, MASK_SOFT_FLOAT | OPTION_MASK_PPC_GFXOPT
-	    | OPTION_MASK_MFCRF | OPTION_MASK_POPCNTB | OPTION_MASK_FPRND
-	    | OPTION_MASK_CMPB | OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
-RS6000_CPU ("476fp", PROCESSOR_PPC476, OPTION_MASK_PPC_GFXOPT
-	    | OPTION_MASK_MFCRF | OPTION_MASK_POPCNTB | OPTION_MASK_FPRND
-	    | OPTION_MASK_CMPB | OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
+RS6000_CPU ("476", PROCESSOR_PPC476,
+	    OPTION_MASK_SOFT_FLOAT | OPTION_MASK_PPC_GFXOPT | OPTION_MASK_MFCRF
+	    | OPTION_MASK_POPCNTB | OPTION_MASK_FPRND | OPTION_MASK_CMPB
+	    | OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
+RS6000_CPU ("476fp", PROCESSOR_PPC476,
+	    OPTION_MASK_PPC_GFXOPT | OPTION_MASK_MFCRF | OPTION_MASK_POPCNTB
+	    | OPTION_MASK_FPRND | OPTION_MASK_CMPB | OPTION_MASK_MULHW
+	    | OPTION_MASK_DLMZB)
 RS6000_CPU ("505", PROCESSOR_MPCCORE, 0)
-RS6000_CPU ("601", PROCESSOR_PPC601, MASK_MULTIPLE)
+RS6000_CPU ("601", PROCESSOR_PPC601, OPTION_MASK_MULTIPLE)
 RS6000_CPU ("602", PROCESSOR_PPC603, OPTION_MASK_PPC_GFXOPT)
 RS6000_CPU ("603", PROCESSOR_PPC603, OPTION_MASK_PPC_GFXOPT)
 RS6000_CPU ("603e", PROCESSOR_PPC603, OPTION_MASK_PPC_GFXOPT)
 RS6000_CPU ("604", PROCESSOR_PPC604, OPTION_MASK_PPC_GFXOPT)
 RS6000_CPU ("604e", PROCESSOR_PPC604e, OPTION_MASK_PPC_GFXOPT)
@@ -204,35 +206,34 @@ RS6000_CPU ("620", PROCESSOR_PPC620, OPTION_MASK_PPC_GFXOPT | MASK_POWERPC64)
 RS6000_CPU ("630", PROCESSOR_PPC630, OPTION_MASK_PPC_GFXOPT | MASK_POWERPC64)
 RS6000_CPU ("740", PROCESSOR_PPC750, OPTION_MASK_PPC_GFXOPT)
 RS6000_CPU ("7400", PROCESSOR_PPC7400, POWERPC_7400_MASK)
 RS6000_CPU ("7450", PROCESSOR_PPC7450, POWERPC_7400_MASK)
 RS6000_CPU ("750", PROCESSOR_PPC750, OPTION_MASK_PPC_GFXOPT)
-RS6000_CPU ("801", PROCESSOR_MPCCORE, MASK_SOFT_FLOAT)
-RS6000_CPU ("821", PROCESSOR_MPCCORE, MASK_SOFT_FLOAT)
-RS6000_CPU ("823", PROCESSOR_MPCCORE, MASK_SOFT_FLOAT)
+RS6000_CPU ("801", PROCESSOR_MPCCORE, OPTION_MASK_SOFT_FLOAT)
+RS6000_CPU ("821", PROCESSOR_MPCCORE, OPTION_MASK_SOFT_FLOAT)
+RS6000_CPU ("823", PROCESSOR_MPCCORE, OPTION_MASK_SOFT_FLOAT)
 RS6000_CPU ("8540", PROCESSOR_PPC8540, MASK_STRICT_ALIGN | OPTION_MASK_ISEL)
 RS6000_CPU ("8548", PROCESSOR_PPC8548, MASK_STRICT_ALIGN | OPTION_MASK_ISEL)
 RS6000_CPU ("a2", PROCESSOR_PPCA2, OPTION_MASK_PPC_GFXOPT | MASK_POWERPC64
-	    | OPTION_MASK_POPCNTB | OPTION_MASK_CMPB | MASK_NO_UPDATE)
-RS6000_CPU ("e300c2", PROCESSOR_PPCE300C2, MASK_SOFT_FLOAT)
+	    | OPTION_MASK_POPCNTB | OPTION_MASK_CMPB | OPTION_MASK_NO_UPDATE)
+RS6000_CPU ("e300c2", PROCESSOR_PPCE300C2, OPTION_MASK_SOFT_FLOAT)
 RS6000_CPU ("e300c3", PROCESSOR_PPCE300C3, 0)
 RS6000_CPU ("e500mc", PROCESSOR_PPCE500MC, OPTION_MASK_PPC_GFXOPT
 	    | OPTION_MASK_ISEL)
 RS6000_CPU ("e500mc64", PROCESSOR_PPCE500MC64,
 	    MASK_POWERPC64 | OPTION_MASK_PPC_GFXOPT | OPTION_MASK_ISEL)
 RS6000_CPU ("e5500", PROCESSOR_PPCE5500,
 	    MASK_POWERPC64 | OPTION_MASK_PPC_GFXOPT | OPTION_MASK_ISEL)
 RS6000_CPU ("e6500", PROCESSOR_PPCE6500, POWERPC_7400_MASK | MASK_POWERPC64
 	    | OPTION_MASK_MFCRF | OPTION_MASK_ISEL)
-RS6000_CPU ("860", PROCESSOR_MPCCORE, MASK_SOFT_FLOAT)
-RS6000_CPU ("970", PROCESSOR_POWER4,
-	    POWERPC_7400_MASK | OPTION_MASK_PPC_GPOPT | OPTION_MASK_MFCRF
-	    | MASK_POWERPC64)
+RS6000_CPU ("860", PROCESSOR_MPCCORE, OPTION_MASK_SOFT_FLOAT)
+RS6000_CPU ("970", PROCESSOR_POWER4, POWERPC_7400_MASK | OPTION_MASK_PPC_GPOPT
+	   | OPTION_MASK_MFCRF | MASK_POWERPC64)
 RS6000_CPU ("cell", PROCESSOR_CELL,
-	    POWERPC_7400_MASK  | OPTION_MASK_PPC_GPOPT | OPTION_MASK_MFCRF
-	    | MASK_POWERPC64)
-RS6000_CPU ("ec603e", PROCESSOR_PPC603, MASK_SOFT_FLOAT)
+	    POWERPC_7400_MASK | OPTION_MASK_PPC_GPOPT | OPTION_MASK_MFCRF
+	   | MASK_POWERPC64)
+RS6000_CPU ("ec603e", PROCESSOR_PPC603, OPTION_MASK_SOFT_FLOAT)
 RS6000_CPU ("G3", PROCESSOR_PPC750, OPTION_MASK_PPC_GFXOPT)
 RS6000_CPU ("G4",  PROCESSOR_PPC7450, POWERPC_7400_MASK)
 RS6000_CPU ("G5", PROCESSOR_POWER4, POWERPC_7400_MASK | OPTION_MASK_PPC_GPOPT
 	    | OPTION_MASK_MFCRF | MASK_POWERPC64)
 RS6000_CPU ("titan", PROCESSOR_TITAN, OPTION_MASK_MULHW | OPTION_MASK_DLMZB)
@@ -245,16 +246,16 @@ RS6000_CPU ("power5", PROCESSOR_POWER5, MASK_POWERPC64 | OPTION_MASK_PPC_GPOPT
 RS6000_CPU ("power5+", PROCESSOR_POWER5, MASK_POWERPC64 | OPTION_MASK_PPC_GPOPT
 	    | OPTION_MASK_PPC_GFXOPT | OPTION_MASK_MFCRF | OPTION_MASK_POPCNTB
 	    | OPTION_MASK_FPRND)
 RS6000_CPU ("power6", PROCESSOR_POWER6, MASK_POWERPC64 | OPTION_MASK_PPC_GPOPT
 	    | OPTION_MASK_PPC_GFXOPT | OPTION_MASK_MFCRF | OPTION_MASK_POPCNTB
-	    | OPTION_MASK_FPRND | OPTION_MASK_CMPB | MASK_DFP
-	    | MASK_RECIP_PRECISION)
+	    | OPTION_MASK_FPRND | OPTION_MASK_CMPB | OPTION_MASK_DFP
+	    | OPTION_MASK_RECIP_PRECISION)
 RS6000_CPU ("power6x", PROCESSOR_POWER6, MASK_POWERPC64 | OPTION_MASK_PPC_GPOPT
 	    | OPTION_MASK_PPC_GFXOPT | OPTION_MASK_MFCRF | OPTION_MASK_POPCNTB
-	    | OPTION_MASK_FPRND | OPTION_MASK_CMPB | MASK_DFP
-	    | MASK_RECIP_PRECISION)
+	    | OPTION_MASK_FPRND | OPTION_MASK_CMPB | OPTION_MASK_DFP
+	    | OPTION_MASK_RECIP_PRECISION)
 RS6000_CPU ("power7", PROCESSOR_POWER7, MASK_POWERPC64 | ISA_2_6_MASKS_SERVER)
 RS6000_CPU ("power8", PROCESSOR_POWER8, MASK_POWERPC64 | ISA_2_7_MASKS_SERVER
 	    | OPTION_MASK_HTM)
 RS6000_CPU ("power9", PROCESSOR_POWER9, MASK_POWERPC64 | ISA_3_0_MASKS_SERVER
 	    | OPTION_MASK_HTM)
diff --git a/gcc/config/rs6000/rs6000.h b/gcc/config/rs6000/rs6000.h
index ef7f10e4efee..dcf632c1f1ad 100644
--- a/gcc/config/rs6000/rs6000.h
+++ b/gcc/config/rs6000/rs6000.h
@@ -505,31 +505,11 @@ extern int rs6000_vector_align[];
 			 && (TARGET_P9_MINMAX || !flag_trapping_math))
 
 /* In switching from using target_flags to using rs6000_isa_flags, the options
    machinery creates OPTION_MASK_<xxx> instead of MASK_<xxx>.  For now map
    OPTION_MASK_<xxx> back into MASK_<xxx>.  */
-#define MASK_DFP			OPTION_MASK_DFP
-#define MASK_DIRECT_MOVE		OPTION_MASK_DIRECT_MOVE
-#define MASK_FLOAT128_KEYWORD		OPTION_MASK_FLOAT128_KEYWORD
-#define MASK_FLOAT128_HW		OPTION_MASK_FLOAT128_HW
-#define MASK_P8_FUSION			OPTION_MASK_P8_FUSION
-#define MASK_HARD_FLOAT			OPTION_MASK_HARD_FLOAT
-#define MASK_HTM			OPTION_MASK_HTM
-#define MASK_MMA			OPTION_MASK_MMA
-#define MASK_MULTIPLE			OPTION_MASK_MULTIPLE
-#define MASK_NO_UPDATE			OPTION_MASK_NO_UPDATE
-#define MASK_P8_VECTOR			OPTION_MASK_P8_VECTOR
-#define MASK_P9_VECTOR			OPTION_MASK_P9_VECTOR
-#define MASK_P9_MISC			OPTION_MASK_P9_MISC
-#define MASK_POPCNTD			OPTION_MASK_POPCNTD
-#define MASK_RECIP_PRECISION		OPTION_MASK_RECIP_PRECISION
-#define MASK_SOFT_FLOAT			OPTION_MASK_SOFT_FLOAT
 #define MASK_STRICT_ALIGN		OPTION_MASK_STRICT_ALIGN
-#define MASK_UPDATE			OPTION_MASK_UPDATE
-#define MASK_VSX			OPTION_MASK_VSX
-#define MASK_POWER10			OPTION_MASK_POWER10
-#define MASK_P10_FUSION			OPTION_MASK_P10_FUSION
 
 #ifndef IN_LIBGCC2
 #define MASK_POWERPC64			OPTION_MASK_POWERPC64
 #endif
 
@@ -2241,31 +2221,31 @@ extern int frame_pointer_needed;
    target flags, and pick a random bit for ldbl128, which isn't in
    target_flags.  */
 #define RS6000_BTM_ALWAYS	0		/* Always enabled.  */
 #define RS6000_BTM_ALTIVEC	OPTION_MASK_ALTIVEC	/* VMX/altivec vectors.  */
 #define RS6000_BTM_CMPB		OPTION_MASK_CMPB	/* ISA 2.05: compare bytes.  */
-#define RS6000_BTM_VSX		MASK_VSX	/* VSX (vector/scalar).  */
-#define RS6000_BTM_P8_VECTOR	MASK_P8_VECTOR	/* ISA 2.07 vector.  */
-#define RS6000_BTM_P9_VECTOR	MASK_P9_VECTOR	/* ISA 3.0 vector.  */
-#define RS6000_BTM_P9_MISC	MASK_P9_MISC	/* ISA 3.0 misc. non-vector */
+#define RS6000_BTM_VSX		OPTION_MASK_VSX	/* VSX (vector/scalar).  */
+#define RS6000_BTM_P8_VECTOR	OPTION_MASK_P8_VECTOR	/* ISA 2.07 vector.  */
+#define RS6000_BTM_P9_VECTOR	OPTION_MASK_P9_VECTOR	/* ISA 3.0 vector.  */
+#define RS6000_BTM_P9_MISC	OPTION_MASK_P9_MISC	/* ISA 3.0 misc. non-vector */
 #define RS6000_BTM_CRYPTO	OPTION_MASK_CRYPTO	/* crypto funcs.  */
-#define RS6000_BTM_HTM		MASK_HTM	/* hardware TM funcs.  */
+#define RS6000_BTM_HTM		OPTION_MASK_HTM	/* hardware TM funcs.  */
 #define RS6000_BTM_FRE		OPTION_MASK_POPCNTB	/* FRE instruction.  */
 #define RS6000_BTM_FRES		OPTION_MASK_PPC_GFXOPT	/* FRES instruction.  */
 #define RS6000_BTM_FRSQRTE	OPTION_MASK_PPC_GFXOPT	/* FRSQRTE instruction.  */
 #define RS6000_BTM_FRSQRTES	OPTION_MASK_POPCNTB	/* FRSQRTES instruction.  */
-#define RS6000_BTM_POPCNTD	MASK_POPCNTD	/* Target supports ISA 2.06.  */
+#define RS6000_BTM_POPCNTD	OPTION_MASK_POPCNTD	/* Target supports ISA 2.06.  */
 #define RS6000_BTM_CELL		OPTION_MASK_FPRND	/* Target is cell powerpc.  */
-#define RS6000_BTM_DFP		MASK_DFP	/* Decimal floating point.  */
-#define RS6000_BTM_HARD_FLOAT	MASK_SOFT_FLOAT	/* Hardware floating point.  */
-#define RS6000_BTM_LDBL128	MASK_MULTIPLE	/* 128-bit long double.  */
+#define RS6000_BTM_DFP		OPTION_MASK_DFP	/* Decimal floating point.  */
+#define RS6000_BTM_HARD_FLOAT	OPTION_MASK_SOFT_FLOAT	/* Hardware floating point.  */
+#define RS6000_BTM_LDBL128	OPTION_MASK_MULTIPLE	/* 128-bit long double.  */
 #define RS6000_BTM_64BIT	MASK_64BIT	/* 64-bit addressing.  */
 #define RS6000_BTM_POWERPC64	MASK_POWERPC64	/* 64-bit registers.  */
-#define RS6000_BTM_FLOAT128	MASK_FLOAT128_KEYWORD /* IEEE 128-bit float.  */
-#define RS6000_BTM_FLOAT128_HW	MASK_FLOAT128_HW /* IEEE 128-bit float h/w.  */
-#define RS6000_BTM_MMA		MASK_MMA	/* ISA 3.1 MMA.  */
-#define RS6000_BTM_P10		MASK_POWER10
+#define RS6000_BTM_FLOAT128	OPTION_MASK_FLOAT128_KEYWORD /* IEEE 128-bit float.  */
+#define RS6000_BTM_FLOAT128_HW	OPTION_MASK_FLOAT128_HW /* IEEE 128-bit float h/w.  */
+#define RS6000_BTM_MMA		OPTION_MASK_MMA	/* ISA 3.1 MMA.  */
+#define RS6000_BTM_P10		OPTION_MASK_POWER10
 
 #define RS6000_BTM_COMMON	(RS6000_BTM_ALTIVEC			\
 				 | RS6000_BTM_VSX			\
 				 | RS6000_BTM_P8_VECTOR			\
 				 | RS6000_BTM_P9_VECTOR			\
  

Hi Will,

on 2022/5/26 04:25, will schmidt via Gcc-patches wrote:
> [PATCH, rs6000] Clean up the option_mask defines
>     
> Hi,
> 
> We have an assortment of MASK and OPTION_MASK #defines throughout
> the rs6000 code, MASK_ALTIVEC and OPTION_MASK_ALTIVEC as an example.
> 
> We currently #define the MASK_<xxxx> entries to their OPTION_MASK_<xxxx>
> equivalents so the two names could be used interchangeably.
> 
> The mapping is in place from when we switched from using
> target_flags to rs6000_isa_flags via
> commit 4d9675496a28ef6184f2a9c3ac5e6e3ea63606c1 in 2012.
> 
> This patch converts the references for most of the lingering MASK_*
> values to OPTION_MASK_*  and removes the now redundant defines.
> 

Nice, thanks for the cleanup!

> I have split this into multiple parts due to size.
> 

I guess they can be bootstrapped & regressed incrementally?

I found there are still some masks left:

MASK_POWERPC64, MASK_64BIT and MASK_LITTLE_ENDIAN.

Is there one part 4 for them?  Or is there some particular reason
not to clean up them?

BR,
Kewen
  

on 2022/5/26 14:12, Kewen.Lin via Gcc-patches wrote:
> Hi Will,
> 
> on 2022/5/26 04:25, will schmidt via Gcc-patches wrote:
>> [PATCH, rs6000] Clean up the option_mask defines
>>     
>> Hi,
>>
>> We have an assortment of MASK and OPTION_MASK #defines throughout
>> the rs6000 code, MASK_ALTIVEC and OPTION_MASK_ALTIVEC as an example.
>>
>> We currently #define the MASK_<xxxx> entries to their OPTION_MASK_<xxxx>
>> equivalents so the two names could be used interchangeably.
>>
>> The mapping is in place from when we switched from using
>> target_flags to rs6000_isa_flags via
>> commit 4d9675496a28ef6184f2a9c3ac5e6e3ea63606c1 in 2012.
>>
>> This patch converts the references for most of the lingering MASK_*
>> values to OPTION_MASK_*  and removes the now redundant defines.
>>
> 
> Nice, thanks for the cleanup!
> 
>> I have split this into multiple parts due to size.
>>
> 
> I guess they can be bootstrapped & regressed incrementally?
> 
> I found there are still some masks left:
> 
> MASK_POWERPC64, MASK_64BIT and MASK_LITTLE_ENDIAN.
> 
> Is there one part 4 for them?  Or is there some particular reason
> not to clean up them?
> 

aha, I see.  Those three are conditional definitions, I agree it's better
to leave them alone. :)


>> diff --git a/gcc/config/rs6000/rs6000.cc b/gcc/config/rs6000/rs6000.cc
>> index d4defc855d02..200bef3f822e 100644
>> --- a/gcc/config/rs6000/rs6000.cc
>> +++ b/gcc/config/rs6000/rs6000.cc
>> @@ -20727,15 +20727,15 @@ rs6000_darwin_file_start (void)
>>      const char *arg;
>>      const char *name;
>>      HOST_WIDE_INT if_set;
>>    } mapping[] = {
>>      { "ppc64", "ppc64", MASK_64BIT },
>> -    { "970", "ppc970", MASK_PPC_GPOPT | MASK_MFCRF | MASK_POWERPC64 },
>> +    { "970", "ppc970", OPTION_MASK_PPC_GPOPT | OPTION_MASK_MFCRF | MASK_POWERPC64 },

Nit: This line is too long.

>>      { "power4", "ppc970", 0 },
>>      { "G5", "ppc970", 0 },
>>      { "7450", "ppc7450", 0 },
>> -    { "7400", "ppc7400", MASK_ALTIVEC },
>> +    { "7400", "ppc7400", OPTION_MASK_ALTIVEC },
>>      { "G4", "ppc7400", 0 },
>>      { "750", "ppc750", 0 },
>>      { "740", "ppc750", 0 },
>>      { "G3", "ppc750", 0 },
>>      { "604e", "ppc604e", 0 },
...
>>  /* Builtin targets.  For now, we reuse the masks for those options that are in
>>     target flags, and pick a random bit for ldbl128, which isn't in
>>     target_flags.  */

Nit: Some of these BTM lines below exceed 80 characters, a few already existed
previously.

>>  #define RS6000_BTM_ALWAYS	0		/* Always enabled.  */
>> -#define RS6000_BTM_ALTIVEC	MASK_ALTIVEC	/* VMX/altivec vectors.  */
>> -#define RS6000_BTM_CMPB		MASK_CMPB	/* ISA 2.05: compare bytes.  */
>> +#define RS6000_BTM_ALTIVEC	OPTION_MASK_ALTIVEC	/* VMX/altivec vectors.  */
>> +#define RS6000_BTM_CMPB		OPTION_MASK_CMPB	/* ISA 2.05: compare bytes.  */
>>  #define RS6000_BTM_VSX		MASK_VSX	/* VSX (vector/scalar).  */
>>  #define RS6000_BTM_P8_VECTOR	MASK_P8_VECTOR	/* ISA 2.07 vector.  */
>>  #define RS6000_BTM_P9_VECTOR	MASK_P9_VECTOR	/* ISA 3.0 vector.  */
>>  #define RS6000_BTM_P9_MISC	MASK_P9_MISC	/* ISA 3.0 misc. non-vector */
>> -#define RS6000_BTM_CRYPTO	MASK_CRYPTO	/* crypto funcs.  */
>> +#define RS6000_BTM_CRYPTO	OPTION_MASK_CRYPTO	/* crypto funcs.  */
>>  #define RS6000_BTM_HTM		MASK_HTM	/* hardware TM funcs.  */
>> -#define RS6000_BTM_FRE		MASK_POPCNTB	/* FRE instruction.  */
>> -#define RS6000_BTM_FRES		MASK_PPC_GFXOPT	/* FRES instruction.  */
>> -#define RS6000_BTM_FRSQRTE	MASK_PPC_GFXOPT	/* FRSQRTE instruction.  */
>> -#define RS6000_BTM_FRSQRTES	MASK_POPCNTB	/* FRSQRTES instruction.  */
>> +#define RS6000_BTM_FRE		OPTION_MASK_POPCNTB	/* FRE instruction.  */
>> +#define RS6000_BTM_FRES		OPTION_MASK_PPC_GFXOPT	/* FRES instruction.  */
>> +#define RS6000_BTM_FRSQRTE	OPTION_MASK_PPC_GFXOPT	/* FRSQRTE instruction.  */
>> +#define RS6000_BTM_FRSQRTES	OPTION_MASK_POPCNTB	/* FRSQRTES instruction.  */
>>  #define RS6000_BTM_POPCNTD	MASK_POPCNTD	/* Target supports ISA 2.06.  */
>> -#define RS6000_BTM_CELL		MASK_FPRND	/* Target is cell powerpc.  */
>> +#define RS6000_BTM_CELL		OPTION_MASK_FPRND	/* Target is cell powerpc.  */
>>  #define RS6000_BTM_DFP		MASK_DFP	/* Decimal floating point.  */
>>  #define RS6000_BTM_HARD_FLOAT	MASK_SOFT_FLOAT	/* Hardware floating point.  */
>>  #define RS6000_BTM_LDBL128	MASK_MULTIPLE	/* 128-bit long double.  */
>>  #define RS6000_BTM_64BIT	MASK_64BIT	/* 64-bit addressing.  */
>>  #define RS6000_BTM_POWERPC64	MASK_POWERPC64	/* 64-bit registers.  */


BR,
Kewen

BR,
Kewen
  

Hi!

On Thu, May 26, 2022 at 03:01:37PM +0800, Kewen.Lin wrote:
> on 2022/5/26 14:12, Kewen.Lin via Gcc-patches wrote:
> > on 2022/5/26 04:25, will schmidt via Gcc-patches wrote:
> >> We have an assortment of MASK and OPTION_MASK #defines throughout
> >> the rs6000 code, MASK_ALTIVEC and OPTION_MASK_ALTIVEC as an example.
> >>
> >> We currently #define the MASK_<xxxx> entries to their OPTION_MASK_<xxxx>
> >> equivalents so the two names could be used interchangeably.
> >>
> >> The mapping is in place from when we switched from using
> >> target_flags to rs6000_isa_flags via
> >> commit 4d9675496a28ef6184f2a9c3ac5e6e3ea63606c1 in 2012.
> >>
> >> This patch converts the references for most of the lingering MASK_*
> >> values to OPTION_MASK_*  and removes the now redundant defines.
> > 
> > Nice, thanks for the cleanup!

+1

> > I found there are still some masks left:
> > 
> > MASK_POWERPC64, MASK_64BIT and MASK_LITTLE_ENDIAN.
> > 
> > Is there one part 4 for them?  Or is there some particular reason
> > not to clean up them?
> 
> aha, I see.  Those three are conditional definitions, I agree it's better
> to leave them alone. :)

It is much better to untangle this mess, and fix it :-)  But that is
(potentially) a bigger job, of course, so let's not balloon this patch.

> >> -    { "970", "ppc970", MASK_PPC_GPOPT | MASK_MFCRF | MASK_POWERPC64 },
> >> +    { "970", "ppc970", OPTION_MASK_PPC_GPOPT | OPTION_MASK_MFCRF | MASK_POWERPC64 },
> 
> Nit: This line is too long.

Yeah, the longer names are a bit annoying in any case.  We'll get used
to it (if those long lines are fixed ;-) )

> Nit: Some of these BTM lines below exceed 80 characters, a few already existed
> previously.

Yes, and it is easily avoidable in this case.  Most of these comments
have no content at all, and the rest could just be on separate lines.

But, are those builtin masks still used at all?  Can't we just use the
option masks where they still are?  The builtins do not use them
anymore :-)


Segher
  

On Thu, 2022-05-26 at 05:47 -0500, Segher Boessenkool wrote:
> Hi!
> 

Hi, 
Thanks Kewen and Segher for the reviews.  Additional comments below.


> On Thu, May 26, 2022 at 03:01:37PM +0800, Kewen.Lin wrote:
> > on 2022/5/26 14:12, Kewen.Lin via Gcc-patches wrote:
> > > on 2022/5/26 04:25, will schmidt via Gcc-patches wrote:
> > > > We have an assortment of MASK and OPTION_MASK #defines
> > > > throughout
> > > > the rs6000 code, MASK_ALTIVEC and OPTION_MASK_ALTIVEC as an
> > > > example.
> > > > 
> > > > We currently #define the MASK_<xxxx> entries to their
> > > > OPTION_MASK_<xxxx>
> > > > equivalents so the two names could be used interchangeably.
> > > > 
> > > > The mapping is in place from when we switched from using
> > > > target_flags to rs6000_isa_flags via
> > > > commit 4d9675496a28ef6184f2a9c3ac5e6e3ea63606c1 in 2012.
> > > > 
> > > > This patch converts the references for most of the lingering
> > > > MASK_*
> > > > values to OPTION_MASK_*  and removes the now redundant defines.
> > > 
> > > Nice, thanks for the cleanup!
> 
> +1
> 
> > > I found there are still some masks left:
> > > 
> > > MASK_POWERPC64, MASK_64BIT and MASK_LITTLE_ENDIAN.
> > > 
> > > Is there one part 4 for them?  Or is there some particular reason
> > > not to clean up them?
> > 
> > aha, I see.  Those three are conditional definitions, I agree it's
> > better
> > to leave them alone. :)
> 
> It is much better to untangle this mess, and fix it :-)  But that is
> (potentially) a bigger job, of course, so let's not balloon this
> patch.

Right.  I have looked briefly at those, and was not convinced those
three would be trivial to rework.  In the interest if incremental
progress I didn't address those in this set.   :-)   If anything I'll
address those in a later patch, whether could be part4 but more likely
a different patchset.

> 
> > > > -    { "970", "ppc970", MASK_PPC_GPOPT | MASK_MFCRF |
> > > > MASK_POWERPC64 },
> > > > +    { "970", "ppc970", OPTION_MASK_PPC_GPOPT |
> > > > OPTION_MASK_MFCRF | MASK_POWERPC64 },
> > 
> > Nit: This line is too long.
> 
Yup, I missed that one. :-)

> Yeah, the longer names are a bit annoying in any case.  We'll get
> used
> to it (if those long lines are fixed ;-) )

Agree.  I would not be opposed to somewhat shorter names for these, but
naming is hard, and the long names are existing and sufficient for the
moment.

> 
> > Nit: Some of these BTM lines below exceed 80 characters, a few
> > already existed
> > previously.
> 
> Yes, and it is easily avoidable in this case.  Most of these comments
> have no content at all, and the rest could just be on separate lines.
> 
> But, are those builtin masks still used at all?  Can't we just use
> the
> option masks where they still are?  The builtins do not use them
> anymore :-)

They are still referenced in rs6000_builtin_mask_calculate() function,
which is used to assign a value to rs6000_builtin_mask, which is still
in use.  I had not yet dug deeper there, but agree it appears that is
only used to print the current options, so could probably be safely
eliminated.  I'll dig a bit more, but would handle that in a separate
patch.

Thanks
-Will


> 
> 
> Segher
  

On Thu, May 26, 2022 at 09:40:18AM -0500, will schmidt wrote:
> On Thu, 2022-05-26 at 05:47 -0500, Segher Boessenkool wrote:
> > Yeah, the longer names are a bit annoying in any case.  We'll get
> > used
> > to it (if those long lines are fixed ;-) )
> 
> Agree.  I would not be opposed to somewhat shorter names for these, but
> naming is hard, and the long names are existing and sufficient for the
> moment.

The longer name is there no matter what we do...  uphill battle, I'd
rather not fight another one :-)

> > But, are those builtin masks still used at all?  Can't we just use
> > the
> > option masks where they still are?  The builtins do not use them
> > anymore :-)
> 
> They are still referenced in rs6000_builtin_mask_calculate() function,
> which is used to assign a value to rs6000_builtin_mask, which is still
> in use.

But we can easily use the OPTION_MASK things directly there.

> I had not yet dug deeper there, but agree it appears that is
> only used to print the current options, so could probably be safely
> eliminated.

And/or that, yes.

> I'll dig a bit more, but would handle that in a separate
> patch.

Can you please make a new patch series that just does everything?  This
is so much easier to handle for everyone, even you yourself :-)

First some small preparatory patches; then the long *boring* patches
that are the meat of the matter, but are completely mechanical
(formatting notwithstanding), so are easy to review; and then some more
small patches to do final cleanup.

So each patch will be easy to write, write a commit message for, write a
changelog for, and easy to review as well.  Long patches are no problem
at all if they are completely boring!


Segher
  

On Thu, 2022-05-26 at 13:31 -0500, Segher Boessenkool wrote:
> > > 

<snip>

> On Thu, May 26, 2022 at 09:40:18AM -0500, will schmidt wrote:
> > On Thu, 2022-05-26 at 05:47 -0500, Segher Boessenkool wrote:
> 
> > I'll dig a bit more, but would handle that in a separate
> > patch.
> 
> Can you please make a new patch series that just does everything?  This
> is so much easier to handle for everyone, even you yourself :-)

Yes, willdo.  Thanks


-Will


> 
> First some small preparatory patches; then the long *boring* patches
> that are the meat of the matter, but are completely mechanical
> (formatting notwithstanding), so are easy to review; and then some more
> small patches to do final cleanup.
> 
> So each patch will be easy to write, write a commit message for, write a
> changelog for, and easy to review as well.  Long patches are no problem
> at all if they are completely boring!
> 
> 
> Segher