[llvm] [InstCombine] Improve folding of `icmp pred (and X, Mask/~Mask), Y)` (PR #81562)

Nikita Popov via llvm-commits llvm-commits at lists.llvm.org
Sat Mar 9 13:15:12 PST 2024


================
@@ -4068,11 +4068,101 @@ Instruction *InstCombinerImpl::foldSelectICmp(ICmpInst::Predicate Pred,
   return nullptr;
 }
 
+// Returns whether V is a Mask ((X + 1) & X == 0) or ~Mask (-Pow2OrZero)
+static bool isMaskOrZero(const Value *V, bool Not, const SimplifyQuery &Q,
+                         unsigned Depth = 0) {
+  if (Not ? match(V, m_NegatedPower2OrZero()) : match(V, m_LowBitMaskOrZero()))
+    return true;
+  if (V->getType()->getScalarSizeInBits() == 1)
+    return true;
+  if (Depth++ >= MaxAnalysisRecursionDepth)
+    return false;
+  Value *X;
+  const Operator *I = dyn_cast<Operator>(V);
----------------
nikic wrote:

Use `Instruction` instead of `Operator` here. I don't see a good reason to make this deal with constant expressions.

https://github.com/llvm/llvm-project/pull/81562


More information about the llvm-commits mailing list