[PATCH] [PPC64] Add 64-bit Vector Integer Arithmetic Instructions

[Kit Barton]

Addressed from previous review, plus cross-reference builtins with the new ABI document.

Notable changes:

1. Fixed return types for the predicate compare functions to always use int instead of long
2. Enabled vector bool long long when power8-vector is enabled (in addition to VSX)
3. Removed builtins that are not included in the new POWER ABI document
4. Removed the intrinsic for the vec_muluwm instruction as it can be generated using LLVM IR (multiplication of two v4i32 types)

The new interfaces added are slightly different from the existing interfaces in that they explicitly use the signed long long type (as opposed to implying it). This is done to be consistent with the new ABI document.


http://reviews.llvm.org/D8041

Files:
  include/clang/Basic/BuiltinsPPC.def
  include/clang/Basic/DiagnosticParseKinds.td
  lib/Headers/altivec.h
  lib/Sema/DeclSpec.cpp
  test/CodeGen/builtins-ppc-p8vector.c
  test/Headers/altivec-intrin.c
  test/Parser/altivec.c
  test/Parser/cxx-altivec.cpp

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D8041.21499.patch
Type: text/x-patch
Size: 59012 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20150309/54f03343/attachment.bin>