[PATCH] D112754: X86: Fold masked merge pattern when and-not is not available

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added a comment.

A minor style comments

Does anyone else have any comments?



================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:46388
+    std::swap(And1_R, And1_L);
+  }
+  // (~(NotOp) & And0_R) | (NotOp & And1_R)
----------------
(style) remove unnecessary braces


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:46391
+  // --> ((And0_R ^ And1_R) & NotOp) ^ And1_R
+  if (NotOp == And1_L) {
+    EVT VT = And1_L->getValueType(0);
----------------
(style)
```
if (NotOp != And1_L)
  return SDValue();
```


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:46406
+/// "and-not" operation. This function is intended to be called from a
+/// `TargetLowering::PerformDAGCombing` callback on `ISD::OR` nodes.
+static SDValue foldMaskedMerge(SDNode *Node, SelectionDAG &DAG) {
----------------
PerformDAGCombing ?


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:46410
+  // normalized to OR by InstCombine so we only check for OR.
+  assert(Node->getOpcode() == ISD::OR);
+  SDValue N0 = Node->getOperand(0);
----------------
assert message


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:46531
+      return R;
+  }
+
----------------
(style) remove braces


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D112754/new/

https://reviews.llvm.org/D112754