[CodeGen] Properly support the half FP type with non-native operations.

On AArch64, the -fallow-half-args-and-returns option is the default.
With it, the half type is considered legal (rather than the i16 used
normally for __fp16), but no operation is, except conversions and
load/stores and such.

The previous behavior was tantamount to saying LangOpts.NativeHalfType
was implied by LangOpts.HalfArgsAndReturns, which isn't true.
Instead, teach the various parts of CodeGen that already know about
half (using the intrinsics or not) about this weird in-between case,
where the "half" type is legal, but operations on it aren't.

This is a smaller intermediate step to the end-goal of removing the
intrinsic, always using "half", and letting the backend legalize.

Builds on r232968.
rdar://20045970, rdar://17468714
Differential Revision: http://reviews.llvm.org/D8367

llvm-svn: 232971

1 files changed, 60 insertions, 34 deletions

diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
index 85a0de2b919..11d10d27c39 100644
--- a/clang/lib/CodeGen/CGExprScalar.cpp
+++ b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -751,20 +751,29 @@ Value *ScalarExprEmitter::EmitScalarConversion(Value *Src, QualType SrcType,
 
   llvm::Type *DstTy = ConvertType(DstType);
 
-  // Cast from storage-only half FP using the special intrinsic.
-  if (SrcType->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType &&
-      !CGF.getContext().getLangOpts().HalfArgsAndReturns) {
-    if (DstTy->isFloatingPointTy())
-      return Builder.CreateCall(
-          CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_from_fp16, DstTy), Src);
-
-    // If this isn't an FP->FP conversion, go through float.
-    Src = Builder.CreateCall(
-        CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_from_fp16,
-                             CGF.CGM.FloatTy),
-        Src);
-    SrcType = CGF.getContext().FloatTy;
-    SrcTy = CGF.FloatTy;
+  // Cast from half through float if half isn't a native type.
+  if (SrcType->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType) {
+    // Cast to FP using the intrinsic if the half type itself isn't supported.
+    if (DstTy->isFloatingPointTy()) {
+      if (!CGF.getContext().getLangOpts().HalfArgsAndReturns)
+        return Builder.CreateCall(
+            CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_from_fp16, DstTy),
+            Src);
+    } else {
+      // Cast to other types through float, using either the intrinsic or FPExt,
+      // depending on whether the half type itself is supported
+      // (as opposed to operations on half, available with NativeHalfType).
+      if (!CGF.getContext().getLangOpts().HalfArgsAndReturns) {
+        Src = Builder.CreateCall(
+            CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_from_fp16,
+                                 CGF.CGM.FloatTy),
+            Src);
+      } else {
+        Src = Builder.CreateFPExt(Src, CGF.CGM.FloatTy, "conv");
+      }
+      SrcType = CGF.getContext().FloatTy;
+      SrcTy = CGF.FloatTy;
+    }
   }
 
   // Ignore conversions like int -> uint.
@@ -823,12 +832,18 @@ Value *ScalarExprEmitter::EmitScalarConversion(Value *Src, QualType SrcType,
     EmitFloatConversionCheck(OrigSrc, OrigSrcType, Src, SrcType, DstType,
                              DstTy);
 
-  // Cast to half using the intrinsic if from FP type, through float otherwise.
-  if (DstType->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType &&
-      !CGF.getContext().getLangOpts().HalfArgsAndReturns) {
-    if (SrcTy->isFloatingPointTy())
-      return Builder.CreateCall(
-          CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_to_fp16, SrcTy), Src);
+  // Cast to half through float if half isn't a native type.
+  if (DstType->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType) {
+    // Make sure we cast in a single step if from another FP type.
+    if (SrcTy->isFloatingPointTy()) {
+      // Use the intrinsic if the half type itself isn't supported
+      // (as opposed to operations on half, available with NativeHalfType).
+      if (!CGF.getContext().getLangOpts().HalfArgsAndReturns)
+        return Builder.CreateCall(
+            CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_to_fp16, SrcTy), Src);
+      // If the half type is supported, just use an fptrunc.
+      return Builder.CreateFPTrunc(Src, DstTy);
+    }
     DstTy = CGF.FloatTy;
   }
 
@@ -856,10 +871,14 @@ Value *ScalarExprEmitter::EmitScalarConversion(Value *Src, QualType SrcType,
   }
 
   if (DstTy != ResTy) {
-    assert(ResTy->isIntegerTy(16) && "Only half FP requires extra conversion");
-    Res = Builder.CreateCall(
+    if (!CGF.getContext().getLangOpts().HalfArgsAndReturns) {
+      assert(ResTy->isIntegerTy(16) && "Only half FP requires extra conversion");
+      Res = Builder.CreateCall(
         CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_to_fp16, CGF.CGM.FloatTy),
         Res);
+    } else {
+      Res = Builder.CreateFPTrunc(Res, ResTy, "conv");
+    }
   }
 
   return Res;
@@ -1762,13 +1781,16 @@ ScalarExprEmitter::EmitScalarPrePostIncDec(const UnaryOperator *E, LValue LV,
     // Add the inc/dec to the real part.
     llvm::Value *amt;
 
-    if (type->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType &&
-        !CGF.getContext().getLangOpts().HalfArgsAndReturns) {
+    if (type->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType) {
       // Another special case: half FP increment should be done via float
-      value = Builder.CreateCall(
-          CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_from_fp16,
-                               CGF.CGM.FloatTy),
-          input);
+      if (!CGF.getContext().getLangOpts().HalfArgsAndReturns) {
+        value = Builder.CreateCall(
+            CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_from_fp16,
+                                 CGF.CGM.FloatTy),
+            input, "incdec.conv");
+      } else {
+        value = Builder.CreateFPExt(input, CGF.CGM.FloatTy, "incdec.conv");
+      }
     }
 
     if (value->getType()->isFloatTy())
@@ -1786,12 +1808,16 @@ ScalarExprEmitter::EmitScalarPrePostIncDec(const UnaryOperator *E, LValue LV,
     }
     value = Builder.CreateFAdd(value, amt, isInc ? "inc" : "dec");
 
-    if (type->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType &&
-        !CGF.getContext().getLangOpts().HalfArgsAndReturns)
-      value = Builder.CreateCall(
-          CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_to_fp16,
-                               CGF.CGM.FloatTy),
-          value);
+    if (type->isHalfType() && !CGF.getContext().getLangOpts().NativeHalfType) {
+      if (!CGF.getContext().getLangOpts().HalfArgsAndReturns) {
+        value = Builder.CreateCall(
+            CGF.CGM.getIntrinsic(llvm::Intrinsic::convert_to_fp16,
+                                 CGF.CGM.FloatTy),
+            value, "incdec.conv");
+      } else {
+        value = Builder.CreateFPTrunc(value, input->getType(), "incdec.conv");
+      }
+    }
 
   // Objective-C pointer types.
   } else {