[DAGCombiner] Remove unneeded commonAlignment from reduceLoadWidth. (llvm#81707)

We already have the PtrOff factored into MachinePointerInfo. Any calls
to getAlign on the new load with do commonAlignment with the
MachinePointerInfo offset and the base alignment.

Author: topperc

llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

@@ -14382,7 +14382,6 @@ SDValue DAGCombiner::reduceLoadWidth(SDNode *N) {
       DAG.getDataLayout().isBigEndian() ? AdjustBigEndianShift(ShAmt) : ShAmt;
 
   uint64_t PtrOff = PtrAdjustmentInBits / 8;
-  Align NewAlign = commonAlignment(LN0->getAlign(), PtrOff);
   SDLoc DL(LN0);
   // The original load itself didn't wrap, so an offset within it doesn't.
   SDNodeFlags Flags;
@@ -14394,13 +14393,14 @@ SDValue DAGCombiner::reduceLoadWidth(SDNode *N) {
   SDValue Load;
   if (ExtType == ISD::NON_EXTLOAD)
     Load = DAG.getLoad(VT, DL, LN0->getChain(), NewPtr,
-                       LN0->getPointerInfo().getWithOffset(PtrOff), NewAlign,
+                       LN0->getPointerInfo().getWithOffset(PtrOff),
+                       LN0->getOriginalAlign(),
                        LN0->getMemOperand()->getFlags(), LN0->getAAInfo());
   else
     Load = DAG.getExtLoad(ExtType, DL, VT, LN0->getChain(), NewPtr,
                           LN0->getPointerInfo().getWithOffset(PtrOff), ExtVT,
-                          NewAlign, LN0->getMemOperand()->getFlags(),
-                          LN0->getAAInfo());
+                          LN0->getOriginalAlign(),
+                          LN0->getMemOperand()->getFlags(), LN0->getAAInfo());
 
   // Replace the old load's chain with the new load's chain.
   WorklistRemover DeadNodes(*this);

llvm/test/CodeGen/X86/fold-sext-trunc.ll

@@ -18,7 +18,7 @@ define void @int322(i32 %foo) !dbg !5 {
 entry:
   %val = load i64, ptr @g_10, !dbg !16
   %0 = load i32, ptr getelementptr inbounds (%struct.S1, ptr @g_10, i32 0, i32 1), align 4, !dbg !17
-; MIR: renamable {{\$r[a-z]+}} = MOVSX64rm32 {{.*}}, @g_10 + 4,{{.*}} debug-location !17 :: (dereferenceable load (s32) from @g_10 + 4)
+; MIR: renamable {{\$r[a-z]+}} = MOVSX64rm32 {{.*}}, @g_10 + 4,{{.*}} debug-location !17 :: (dereferenceable load (s32) from @g_10 + 4, basealign 8)
   %1 = sext i32 %0 to i64, !dbg !18
   %tmp4.i = lshr i64 %val, 32, !dbg !19
   %tmp5.i = trunc i64 %tmp4.i to i32, !dbg !20

llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/lanai_isel.ll.expected

@@ -53,7 +53,7 @@ define i64 @i16_test(i16 %i) nounwind readnone {
 ; CHECK-NEXT:    t14: ch,glue = CopyToReg t0, Register:i32 $rv, t33
 ; CHECK-NEXT:    t1: i32 = ADD_I_LO TargetFrameIndex:i32<-1>, TargetConstant:i32<0>
 ; CHECK-NEXT:    t21: i32 = OR_I_LO t1, TargetConstant:i32<2>
-; CHECK-NEXT:    t23: i32,ch = LDHz_RI<Mem:(load (s16) from %fixed-stack.0 + 2)> t21, TargetConstant:i32<0>, TargetConstant:i32<0>, t0
+; CHECK-NEXT:    t23: i32,ch = LDHz_RI<Mem:(load (s16) from %fixed-stack.0 + 2, basealign 4)> t21, TargetConstant:i32<0>, TargetConstant:i32<0>, t0
 ; CHECK-NEXT:    t22: i32,ch = LDHz_RI<Mem:(dereferenceable load (s16) from %ir.loc)> TargetFrameIndex:i32<0>, TargetConstant:i32<0>, TargetConstant:i32<0>, t0
 ; CHECK-NEXT:    t24: i32 = ADD_R t23, t22, TargetConstant:i32<0>
 ; CHECK-NEXT:    t27: i32 = AND_I_HI t24, TargetConstant:i32<0>
@@ -76,7 +76,7 @@ define i64 @i8_test(i8 %i) nounwind readnone {
 ; CHECK-NEXT:    t14: ch,glue = CopyToReg t0, Register:i32 $rv, t33
 ; CHECK-NEXT:    t1: i32 = ADD_I_LO TargetFrameIndex:i32<-1>, TargetConstant:i32<0>
 ; CHECK-NEXT:    t21: i32 = OR_I_LO t1, TargetConstant:i32<3>
-; CHECK-NEXT:    t23: i32,ch = LDBz_RI<Mem:(load (s8) from %fixed-stack.0 + 3)> t21, TargetConstant:i32<0>, TargetConstant:i32<0>, t0
+; CHECK-NEXT:    t23: i32,ch = LDBz_RI<Mem:(load (s8) from %fixed-stack.0 + 3, basealign 4)> t21, TargetConstant:i32<0>, TargetConstant:i32<0>, t0
 ; CHECK-NEXT:    t22: i32,ch = LDBz_RI<Mem:(dereferenceable load (s8) from %ir.loc)> TargetFrameIndex:i32<0>, TargetConstant:i32<0>, TargetConstant:i32<0>, t0
 ; CHECK-NEXT:    t24: i32 = ADD_R t23, t22, TargetConstant:i32<0>
 ; CHECK-NEXT:    t26: i32 = SLI TargetConstant:i32<255>