[PATCH] D43618: [X86][MMX] Add support for MMX build vectors (PR29222)

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added inline comments.


================
Comment at: lib/Target/X86/X86ISelLowering.cpp:30863
+    if (V.getValueType().isFloatingPoint()) {
+      if (Subtarget.hasSSE1() && Subtarget.is64Bit() &&
+          !isa<ConstantFPSDNode>(V)) {
----------------
craig.topper wrote:
> Why does this require 64-bit?
I think it was just because we still make a mess of elts from consecutive loads for MMX and that messes with float args on i686 - I'll relax it and see what happens.


================
Comment at: lib/Target/X86/X86ISelLowering.cpp:30865
+          !isa<ConstantFPSDNode>(V)) {
+        V = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, MVT::v4f32,
+                        DAG.getUNDEF(MVT::v4f32), V,
----------------
craig.topper wrote:
> Should this be SCALAR_TO_VECTOR?
Yes, it should be.


Repository:
  rL LLVM

https://reviews.llvm.org/D43618