Subzero: Fix frame size for floating-point register out-args.

The code that calculates maximum out-arg stack space was neglecting the fact that on x86-64, the first N scalar floating-point arguments are passed through xmm registers, not the stack. As a result, stack frames were sometimes larger than necessary. BUG= none R=jpp@chromium.org Review URL: https://codereview.chromium.org/2076663006 .

Subzero: Fix frame size for floating-point register out-args.
c577727f · Jim Stichnoth · e450656d · c577727f
Commit c577727f authored Jun 20, 2016 by Jim Stichnoth
Hide whitespace changes
Inline Side-by-side

Showing with 3 additions and 0 deletions

IceTargetLoweringX86BaseImpl.h src/IceTargetLoweringX86BaseImpl.h +3 -0

No files found.
--- a/src/IceTargetLoweringX86BaseImpl.h
+++ b/src/IceTargetLoweringX86BaseImpl.h
@@ -7020,6 +7020,9 @@ uint32_t TargetX86Base<TraitsType>::getCallStackArgumentsSizeBytes(
    assert(typeWidthInBytes(Ty) >= 4);
    if (isVectorType(Ty) && XmmArgCount < Traits::X86_MAX_XMM_ARGS) {
      ++XmmArgCount;
+    } else if (isScalarFloatingType(Ty) && Traits::X86_PASS_SCALAR_FP_IN_XMM &&
+               XmmArgCount < Traits::X86_MAX_XMM_ARGS) {
+      ++XmmArgCount;
    } else if (isScalarIntegerType(Ty) &&
               GprArgCount < Traits::X86_MAX_GPR_ARGS) {
      // The 64 bit ABI allows some integers to be passed in GPRs.