[PATCH] D11437: Correct x86_64 fp128 calling convention

[Reid Kleckner]

rnk added inline comments.

================
Comment at: lib/CodeGen/TargetInfo.cpp:1861-1866
@@ -1860,8 +1860,8 @@
       Current = Integer;
     } else if ((k == BuiltinType::Float || k == BuiltinType::Double) ||
                (k == BuiltinType::LongDouble &&
                 getTarget().getTriple().isOSNaCl())) {
       Current = SSE;
     } else if (k == BuiltinType::LongDouble) {
       Lo = X87;
       Hi = X87Up;
----------------
Any reason we can't do the same fp classification here like we do below for complex types?

================
Comment at: lib/CodeGen/TargetInfo.cpp:1975-1976
@@ -1974,4 +1975,1 @@
-    else if (ET == getContext().DoubleTy ||
-             (ET == getContext().LongDoubleTy &&
-              getTarget().getTriple().isOSNaCl()))
       Lo = Hi = SSE;
----------------
Nice, I like this simplification.

================
Comment at: lib/CodeGen/TargetInfo.cpp:2528
@@ -2520,1 +2527,3 @@
 classifyReturnType(QualType RetTy) const {
+  // TODO: Simplify classify to handle f128 according to AMD64 ABI.
+  if (const BuiltinType *BT = RetTy->getAs<BuiltinType>()) {
----------------
Why can't we classify IEEEQuad long doubles as SSE in the usual `classify()` implementation?

================
Comment at: lib/CodeGen/TargetInfo.cpp:2667
@@ -2652,1 +2666,3 @@
 {
+  // TODO: Simplify classifyArgumentType to handle f128 according to AMD64 ABI.
+  if (const BuiltinType *BT = Ty->getAs<BuiltinType>()) {
----------------
ditto


http://reviews.llvm.org/D11437