[PATCH] D111410: [InstCombine] generalize fold for mask-with-signbit-splat

[Roman Lebedev via Phabricator via llvm-commits]

lebedev.ri accepted this revision.
lebedev.ri added a comment.
This revision is now accepted and ready to land.

Hm, looks fine to me.
I guess there's also the 'commutative' variant: https://alive2.llvm.org/ce/z/t8iHTy



================
Comment at: llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp:2067
 
-  // and(ashr(subNSW(Y, X), ScalarSizeInBits(Y)-1), X) --> X s> Y ? X : 0.
-  if (match(&I, m_c_And(m_OneUse(m_AShr(
-                            m_NSWSub(m_Value(Y), m_Value(X)),
-                            m_SpecificInt(Ty->getScalarSizeInBits() - 1))),
-                        m_Deferred(X)))) {
-    Value *NewICmpInst = Builder.CreateICmpSGT(X, Y);
-    return SelectInst::Create(NewICmpInst, X, ConstantInt::getNullValue(Ty));
+  // (iN X s>> (N-1)) & Y --> (X < 0) ? Y : 0
+  unsigned FullShift = Ty->getScalarSizeInBits() - 1;
----------------



CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111410/new/

https://reviews.llvm.org/D111410