[ARM NEON] Define vfms_f32 on ARM, and all vfms using vfma.

r259537 added vfma/vfms to armv7, but the builtin was only lowered
on the AArch64 side. Instead of supporting it on ARM, get rid of it.

The vfms builtin lowered to:
  %nb = fsub float -0.0, %b
  %r = @llvm.fma.f32(%a, %nb, %c)

Instead, define the operation in terms of vfma, and swap the
multiplicands. It now lowers to:
  %na = fsub float -0.0, %a
  %r = @llvm.fma.f32(%na, %b, %c)

This matches the instruction more closely, and lets current LLVM
generate the "natural" operand ordering:
  fmls.2s v0, v1, v2
instead of the crooked (but equivalent):
  fmls.2s v0, v2, v1
Except for theses changes, assembly is identical.

LLVM accepts both commutations, and the LLVM tests in:
  test/CodeGen/AArch64/arm64-fmadd.ll
  test/CodeGen/AArch64/fp-dp3.ll
  test/CodeGen/AArch64/neon-fma.ll
  test/CodeGen/ARM/fusedMAC.ll
already check either the new one only, or both.

Also verified against the test-suite unittests.

llvm-svn: 266807

1 files changed, 0 insertions, 16 deletions

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 2397356d1fb..ac22076d791 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -5319,22 +5319,6 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     Ops[2] = Builder.CreateExtractElement(Ops[2], Ops[3], "extract");
     return Builder.CreateCall(F, {Ops[1], Ops[2], Ops[0]});
   }
-  case NEON::BI__builtin_neon_vfms_v:
-  case NEON::BI__builtin_neon_vfmsq_v: {  // Only used for FP types
-    // FIXME: probably remove when we no longer support aarch64_simd.h
-    // (arm_neon.h delegates to vfma).
-
-    // The ARM builtins (and instructions) have the addend as the first
-    // operand, but the 'fma' intrinsics have it last. Swap it around here.
-    Value *Subtrahend = Ops[0];
-    Value *Multiplicand = Ops[2];
-    Ops[0] = Multiplicand;
-    Ops[2] = Subtrahend;
-    Ops[1] = Builder.CreateBitCast(Ops[1], VTy);
-    Ops[1] = Builder.CreateFNeg(Ops[1]);
-    Int = Intrinsic::fma;
-    return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "fmls");
-  }
   case NEON::BI__builtin_neon_vmull_v:
     // FIXME: improve sharing scheme to cope with 3 alternative LLVM intrinsics.
     Int = usgn ? Intrinsic::aarch64_neon_umull : Intrinsic::aarch64_neon_smull;