[llvm] [ValueTracking] ComputeNumSignBitsImpl - add basic handling of BITCAST nodes (PR #127218)

[Simon Pilgrim via llvm-commits]

================
@@ -3922,6 +3922,35 @@ static unsigned ComputeNumSignBitsImpl(const Value *V,
   if (auto *U = dyn_cast<Operator>(V)) {
     switch (Operator::getOpcode(V)) {
     default: break;
+    case Instruction::BitCast: {
+      Value *Src = U->getOperand(0);
+      Type *SrcTy = Src->getType();
+      Type *SrcScalarTy = SrcTy->getScalarType();
+
+      if (!SrcScalarTy->isIntegerTy())
+        break;
+
+      unsigned SrcBits = SrcTy->getScalarSizeInBits();
+
+      if ((SrcBits % TyBits) != 0)
+        break;
+
+      if (auto *DstVTy = dyn_cast<FixedVectorType>(Ty)) {
+        unsigned Scale = SrcBits / TyBits;
+
+        APInt SrcDemandedElts =
+            APInt::getSplat(DstVTy->getNumElements() / Scale, APInt(1, 1));
+
+        Tmp = ComputeNumSignBits(Src, SrcDemandedElts, Depth + 1, Q);
+        if (Tmp == SrcBits)
+          return TyBits;
+      } else {
+        Tmp = ComputeNumSignBits(Src, APInt(1, 1), Depth + 1, Q);
----------------
RKSimon wrote:

This isn't guaranteed to be a scalar here, so could the 1 bit demanded elts mask fail?

https://github.com/llvm/llvm-project/pull/127218