[ScalarizeMaskedMemIntrin] Use pointer alignment from pointer of masked.compressstore/expandload. (#83519)

Previously we used Align(1) for all scalarized load/stores from
masked.compressstore/expandload.
For targets not supporting unaligned accesses, it make backend need to
split
aligned large width loads/stores to byte loads/stores.
To solve this performance issue, this patch preserves the alignment of
base
pointer after scalarizing.

Author: yetingk

llvm/lib/Transforms/Scalar/ScalarizeMaskedMemIntrin.cpp

@@ -627,6 +627,7 @@ static void scalarizeMaskedExpandLoad(const DataLayout &DL, CallInst *CI,
   Value *Ptr = CI->getArgOperand(0);
   Value *Mask = CI->getArgOperand(1);
   Value *PassThru = CI->getArgOperand(2);
+  Align Alignment = CI->getParamAlign(0).valueOrOne();
 
   auto *VecType = cast<FixedVectorType>(CI->getType());
 
@@ -644,6 +645,10 @@ static void scalarizeMaskedExpandLoad(const DataLayout &DL, CallInst *CI,
   // The result vector
   Value *VResult = PassThru;
 
+  // Adjust alignment for the scalar instruction.
+  const Align AdjustedAlignment =
+      commonAlignment(Alignment, EltTy->getPrimitiveSizeInBits() / 8);
+
   // Shorten the way if the mask is a vector of constants.
   // Create a build_vector pattern, with loads/poisons as necessary and then
   // shuffle blend with the pass through value.
@@ -659,7 +664,7 @@ static void scalarizeMaskedExpandLoad(const DataLayout &DL, CallInst *CI,
       } else {
         Value *NewPtr =
             Builder.CreateConstInBoundsGEP1_32(EltTy, Ptr, MemIndex);
-        InsertElt = Builder.CreateAlignedLoad(EltTy, NewPtr, Align(1),
+        InsertElt = Builder.CreateAlignedLoad(EltTy, NewPtr, AdjustedAlignment,
                                               "Load" + Twine(Idx));
         ShuffleMask[Idx] = Idx;
         ++MemIndex;
@@ -713,7 +718,7 @@ static void scalarizeMaskedExpandLoad(const DataLayout &DL, CallInst *CI,
     CondBlock->setName("cond.load");
 
     Builder.SetInsertPoint(CondBlock->getTerminator());
-    LoadInst *Load = Builder.CreateAlignedLoad(EltTy, Ptr, Align(1));
+    LoadInst *Load = Builder.CreateAlignedLoad(EltTy, Ptr, AdjustedAlignment);
     Value *NewVResult = Builder.CreateInsertElement(VResult, Load, Idx);
 
     // Move the pointer if there are more blocks to come.
@@ -755,6 +760,7 @@ static void scalarizeMaskedCompressStore(const DataLayout &DL, CallInst *CI,
   Value *Src = CI->getArgOperand(0);
   Value *Ptr = CI->getArgOperand(1);
   Value *Mask = CI->getArgOperand(2);
+  Align Alignment = CI->getParamAlign(1).valueOrOne();
 
   auto *VecType = cast<FixedVectorType>(Src->getType());
 
@@ -767,6 +773,10 @@ static void scalarizeMaskedCompressStore(const DataLayout &DL, CallInst *CI,
 
   Type *EltTy = VecType->getElementType();
 
+  // Adjust alignment for the scalar instruction.
+  const Align AdjustedAlignment =
+      commonAlignment(Alignment, EltTy->getPrimitiveSizeInBits() / 8);
+
   unsigned VectorWidth = VecType->getNumElements();
 
   // Shorten the way if the mask is a vector of constants.
@@ -778,7 +788,7 @@ static void scalarizeMaskedCompressStore(const DataLayout &DL, CallInst *CI,
       Value *OneElt =
           Builder.CreateExtractElement(Src, Idx, "Elt" + Twine(Idx));
       Value *NewPtr = Builder.CreateConstInBoundsGEP1_32(EltTy, Ptr, MemIndex);
-      Builder.CreateAlignedStore(OneElt, NewPtr, Align(1));
+      Builder.CreateAlignedStore(OneElt, NewPtr, AdjustedAlignment);
       ++MemIndex;
     }
     CI->eraseFromParent();
@@ -824,7 +834,7 @@ static void scalarizeMaskedCompressStore(const DataLayout &DL, CallInst *CI,
 
     Builder.SetInsertPoint(CondBlock->getTerminator());
     Value *OneElt = Builder.CreateExtractElement(Src, Idx);
-    Builder.CreateAlignedStore(OneElt, Ptr, Align(1));
+    Builder.CreateAlignedStore(OneElt, Ptr, AdjustedAlignment);
 
     // Move the pointer if there are more blocks to come.
     Value *NewPtr;