Implement IRGen for the x86 vectorcall convention
The most complex aspect of the convention is the handling of homogeneous
vector and floating point aggregates. Reuse the homogeneous aggregate
classification code that we use on PPC64 and ARM for this.
This convention also has a C mangling, and we apparently implement that
in both Clang and LLVM.
Reviewed By: majnemer
Differential Revision: http://reviews.llvm.org/D6063
llvm-svn: 221006
diff --git a/clang/lib/CodeGen/MicrosoftCXXABI.cpp b/clang/lib/CodeGen/MicrosoftCXXABI.cpp
index 49907c9..428c7b1 100644
--- a/clang/lib/CodeGen/MicrosoftCXXABI.cpp
+++ b/clang/lib/CodeGen/MicrosoftCXXABI.cpp
@@ -617,8 +617,15 @@
if (RD->hasNonTrivialCopyConstructor())
return RAA_Indirect;
- // Win64 passes objects larger than 8 bytes indirectly.
- if (getContext().getTypeSize(RD->getTypeForDecl()) > 64)
+ // If an object has a destructor, we'd really like to pass it indirectly
+ // because it allows us to elide copies. Unfortunately, MSVC makes that
+ // impossible for small types, which it will pass in a single register or
+ // stack slot. Most objects with dtors are large-ish, so handle that early.
+ // We can't call out all large objects as being indirect because there are
+ // multiple x64 calling conventions and the C++ ABI code shouldn't dictate
+ // how we pass large POD types.
+ if (RD->hasNonTrivialDestructor() &&
+ getContext().getTypeSize(RD->getTypeForDecl()) > 64)
return RAA_Indirect;
// We have a trivial copy constructor or no copy constructors, but we have