Automerge: [AArch64][GlobalISel] Fix legalization for <4 x i1> vector stores.

This case is different from the earlier <8 x i1> case handled because it triggers
a legalization failure in lowerStore() that's intended for scalar code.

It also was triggering incorrect bitcast actions in the AArch64 rules that weren't
expecting truncating stores.

With these two fixed, more cases are handled. The code is still bad, including
some missing load promotion in our combiners that result in dead stores hanging
around at the end of codegen. Again, we can fix these in separate changes.

Reviewers: davemgreen, madhur13490, topperc, arsenm

Reviewed By: davemgreen

Pull Request: llvm/llvm-project#121185

Author: aemerson

llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp

@@ -4118,10 +4118,7 @@ LegalizerHelper::LegalizeResult LegalizerHelper::lowerStore(GStore &StoreMI) {
   unsigned StoreWidth = MemTy.getSizeInBits();
   unsigned StoreSizeInBits = 8 * MemTy.getSizeInBytes();
 
-  if (StoreWidth != StoreSizeInBits) {
-    if (SrcTy.isVector())
-      return UnableToLegalize;
-
+  if (StoreWidth != StoreSizeInBits && !SrcTy.isVector()) {
     // Promote to a byte-sized store with upper bits zero if not
     // storing an integral number of bytes.  For example, promote
     // TRUNCSTORE:i1 X -> TRUNCSTORE:i8 (and X, 1)

llvm/lib/Target/AArch64/GISel/AArch64LegalizerInfo.cpp

@@ -467,7 +467,11 @@ AArch64LegalizerInfo::AArch64LegalizerInfo(const AArch64Subtarget &ST)
       .clampMaxNumElements(0, p0, 2)
       .lowerIfMemSizeNotPow2()
       // TODO: Use BITCAST for v2i8, v2i16 after G_TRUNC gets sorted out
-      .bitcastIf(typeInSet(0, {v4s8}),
+      .bitcastIf(all(typeInSet(0, {v4s8}),
+                     LegalityPredicate([=](const LegalityQuery &Query) {
+                       return Query.Types[0].getSizeInBits() ==
+                              Query.MMODescrs[0].MemoryTy.getSizeInBits();
+                     })),
                  [=](const LegalityQuery &Query) {
                    const LLT VecTy = Query.Types[0];
                    return std::pair(0, LLT::scalar(VecTy.getSizeInBits()));