[NFC] Minor cleanup of usage of FloatModeKind with bitmask enums

[jolanta.jensen]

Removing a tab.

[shafik]

clang/include/clang/Basic/TargetInfo.h

894

Why not just cast directly to bool?

[tahonermann]

Looks good, thank you for following up @jolanta.jensen!

[jolanta.jensen]

clang/include/clang/Basic/TargetInfo.h

894

Yes, it will work. But as int is the underlying type I find this way a bit clearer. I hope to be forgiven if I keep it.

[tahonermann]

clang/include/clang/Basic/TargetInfo.h

894

I'm ok in either case. It wasn't clear to me what the rationale was for casting to int; your explanation was helpful. Since the underlying type of an enumeration tends to not be well understood (I had to go look up the rules again), it would be nice if we could use C++23 std::to_underlying():

return std::to_underlying((FloatModeKind)RealTypeUsesObjCFPRetMask & T);

LLVM's BitmaskEnum.h has Underlying(), but only exposed in the detail namespace and I don't think it is worth changing just for this. We could do:

return static_cast<std::underlying_type_t<FloatModeKind>>((FloatModeKind)RealTypeUsesObjCFPRetMask & T);

But I personally don't find that to be an improvement in this case. I think the code is fine as is.

clang/include/clang/Basic/TargetInfo.h

protected:
   TargetCXXABI TheCXXABI;
   const LangASMap *AddrSpaceMap;
   unsigned ProgramAddrSpace;
 
   mutable StringRef PlatformName;
   mutable VersionTuple PlatformMinVersion;
 
   unsigned HasAlignMac68kSupport : 1;
-  unsigned RealTypeUsesObjCFPRetMask
-      : llvm::BitmaskEnumDetail::bitWidth(
-            (int)FloatModeKind::LLVM_BITMASK_LARGEST_ENUMERATOR);
+  unsigned RealTypeUsesObjCFPRetMask : llvm::BitWidth<FloatModeKind>;
   unsigned ComplexLongDoubleUsesFP2Ret : 1;
 
   unsigned HasBuiltinMSVaList : 1;
 
   unsigned IsRenderScriptTarget : 1;
 
   unsigned HasAArch64SVETypes : 1;
 
public:
   /// integer type enum.
   ///
   /// For example, SignedLong -> "l".
   static const char *getTypeFormatModifier(IntType T);
 
   /// Check whether the given real type should use the "fpret" flavor of
   /// Objective-C message passing on this target.
   bool useObjCFPRetForRealType(FloatModeKind T) const {
-    return RealTypeUsesObjCFPRetMask & llvm::BitmaskEnumDetail::Underlying(T);
+    return (int)((FloatModeKind)RealTypeUsesObjCFPRetMask & T);
   }
 
   /// Check whether _Complex long double should use the "fp2ret" flavor
   /// of Objective-C message passing on this target.
   bool useObjCFP2RetForComplexLongDouble() const {
     return ComplexLongDoubleUsesFP2Ret;
   }
 

clang/lib/Basic/Targets/X86.h

resetDataLayout(
         Triple.isOSBinFormatMachO() ? "_" : "");
     SizeType = UnsignedInt;
     PtrDiffType = SignedInt;
     IntPtrType = SignedInt;
     RegParmMax = 3;
 
     // Use fpret for all types.
     RealTypeUsesObjCFPRetMask =
-        (int)(FloatModeKind::Float | FloatModeKind::Double |
+        (unsigned)(FloatModeKind::Float | FloatModeKind::Double |
                    FloatModeKind::LongDouble);
 
     // x86-32 has atomics up to 8 bytes
     MaxAtomicPromoteWidth = 64;
     MaxAtomicInlineWidth = 32;
   }
 
   BuiltinVaListKind getBuiltinVaListKind() const override {
     return TargetInfo::CharPtrBuiltinVaList;
X86_64TargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
     resetDataLayout(IsX32 ? "e-m:e-p:32:32-p270:32:32-p271:32:32-p272:64:64-"
                             "i64:64-f80:128-n8:16:32:64-S128"
                           : IsWinCOFF ? "e-m:w-p270:32:32-p271:32:32-p272:64:"
                                         "64-i64:64-f80:128-n8:16:32:64-S128"
                                       : "e-m:e-p270:32:32-p271:32:32-p272:64:"
                                         "64-i64:64-f80:128-n8:16:32:64-S128");
 
     // Use fpret only for long double.
-    RealTypeUsesObjCFPRetMask = (int)FloatModeKind::LongDouble;
+    RealTypeUsesObjCFPRetMask = (unsigned)FloatModeKind::LongDouble;
 
     // Use fp2ret for _Complex long double.
     ComplexLongDoubleUsesFP2Ret = true;
 
     // Make __builtin_ms_va_list available.
     HasBuiltinMSVaList = true;
 
     // x86-64 has atomics up to 16 bytes.