[clang] [IRGen][AArch64][RISCV] Generalize bitcast between i1 predicate vector and i8 fixed vector. (PR #76548)

[Sander de Smalen via cfe-commits]

================
@@ -2136,14 +2136,16 @@ Value *ScalarExprEmitter::VisitCastExpr(CastExpr *CE) {
     // bitcast.
     if (const auto *FixedSrc = dyn_cast<llvm::FixedVectorType>(SrcTy)) {
       if (const auto *ScalableDst = dyn_cast<llvm::ScalableVectorType>(DstTy)) {
-        // If we are casting a fixed i8 vector to a scalable 16 x i1 predicate
+        // If we are casting a fixed i8 vector to a scalable i1 predicate
         // vector, use a vector insert and bitcast the result.
         bool NeedsBitCast = false;
-        auto PredType = llvm::ScalableVectorType::get(Builder.getInt1Ty(), 16);
         llvm::Type *OrigType = DstTy;
-        if (ScalableDst == PredType &&
-            FixedSrc->getElementType() == Builder.getInt8Ty()) {
-          DstTy = llvm::ScalableVectorType::get(Builder.getInt8Ty(), 2);
+        if (ScalableDst->getElementType()->isIntegerTy(1) &&
----------------
sdesmalen-arm wrote:

nit: While you're touching this code, maybe refactor this whole block to remove the control flow by `NeedsBitCast`, by something like this:

```
if (const auto *FixedSrc = dyn_cast<llvm::FixedVectorType>(SrcTy)) {
  if (const auto *ScalableDst = dyn_cast<llvm::ScalableVectorType>(DstTy)) {
    llvm::Value *Zero = llvm::Constant::getNullValue(CGF.CGM.Int64Ty);
    if (ScalableDst->getElementType()->isIntegerTy(1) &&
        ScalableDst->getElementCount().isKnownMultipleOf(8) &&
        FixedSrc->getElementType()->isIntegerTy(8)) {
      auto InsertTy = llvm::ScalableVectorType::get(
          FixedSrc->getElementType(), FixedSrc->getNumElements());
      llvm::Value *Insert = Builder.CreateInsertVector(
          InsertTy, llvm::UndefValue::get(InsertTy), Src, Zero,
          "cast.scalable");
      return Builder.CreateBitCast(Insert, DstTy);
    } else if (FixedSrc->getElementType() == ScalableDst->getElementType())
      return Builder.CreateInsertVector(DstTy, llvm::UndefValue::get(DstTy),
                                        Src, Zero, "cast.scalable");
  }
}
```

What do you think?

(similar suggestion for the other case below for scalable -> fixed)

https://github.com/llvm/llvm-project/pull/76548