[Loads] Check type size in bits during store to load forwarding

Rather than checking the rounded type store size, check the type
size in bits. We don't want to forward a store of i1 to a load
of i8 for example, even though they have the same type store size.
The padding bits have unspecified contents.

This is a partial fix for the issue reported at
https://reviews.llvm.org/D115924#inline-1179482,
the problem also needs to be addressed more generally in the
constant folding code.

llvm/lib/Analysis/Loads.cpp

@@ -504,8 +504,8 @@ static Value *getAvailableLoadStore(Instruction *Inst, const Value *Ptr,
     if (CastInst::isBitOrNoopPointerCastable(Val->getType(), AccessTy, DL))
       return Val;
 
-    TypeSize StoreSize = DL.getTypeStoreSize(Val->getType());
-    TypeSize LoadSize = DL.getTypeStoreSize(AccessTy);
+    TypeSize StoreSize = DL.getTypeSizeInBits(Val->getType());
+    TypeSize LoadSize = DL.getTypeSizeInBits(AccessTy);
     if (TypeSize::isKnownLE(LoadSize, StoreSize))
       if (auto *C = dyn_cast<Constant>(Val))
         return ConstantFoldLoadFromConst(C, AccessTy, DL);

llvm/test/Transforms/InstCombine/load-store-forward.ll

@@ -293,7 +293,9 @@ entry:
 define i8 @load_i8_store_i1(i1* %a) {
 ; CHECK-LABEL: @load_i8_store_i1(
 ; CHECK-NEXT:    store i1 true, i1* [[A:%.*]], align 1
-; CHECK-NEXT:    ret i8 -1
+; CHECK-NEXT:    [[A_I8:%.*]] = bitcast i1* [[A]] to i8*
+; CHECK-NEXT:    [[V:%.*]] = load i8, i8* [[A_I8]], align 1
+; CHECK-NEXT:    ret i8 [[V]]
 ;
   store i1 true, i1* %a
   %a.i8 = bitcast i1* %a to i8*