[PATCH] D15054: [X86] Better support for the MCU psABI

Michael Kuperstein via llvm-commits llvm-commits at lists.llvm.org
Sun Nov 29 04:36:23 PST 2015


mkuper added a comment.

The clang part of this patch is http://reviews.llvm.org/D15055 .


================
Comment at: lib/Target/X86/X86CallingConv.h:46
@@ -44,1 +45,3 @@
 
+inline bool CC_X86_32_MCUInReg(unsigned &ValNo, MVT &ValVT,
+                                         MVT &LocVT,
----------------
I am not a big fan of this, but I couldn't think of a better to do it (and the idea here is stolen from similar code in the ARM codegen, although the set of constraints there is a bit different).
Other suggestions are welcome.

================
Comment at: lib/Target/X86/X86FastISel.cpp:1105
@@ -1104,4 +1104,3 @@
   // and into %rax. We also do the same with %eax for Win32.
-  if (F.hasStructRetAttr() &&
-      (Subtarget->is64Bit() || Subtarget->isTargetKnownWindowsMSVC())) {
+  if (F.hasStructRetAttr()) {
     unsigned Reg = X86MFInfo->getSRetReturnReg();
----------------
This is actually not an MCU-related issue - this code should fire for x86 regardless of bitness and windows-ness. The DAG-based selection does the right thing, but FastISel didn't. Of course, since the MCU is non-windows 32-bit, it also hits this.


http://reviews.llvm.org/D15054





More information about the llvm-commits mailing list