[CVP] Infer nneg on existing zext

llvm/lib/Transforms/Scalar/CorrelatedValuePropagation.cpp

@@ -93,6 +93,7 @@ STATISTIC(NumNonNull, "Number of function pointer arguments marked non-null");
 STATISTIC(NumMinMax, "Number of llvm.[us]{min,max} intrinsics removed");
 STATISTIC(NumUDivURemsNarrowedExpanded,
           "Number of bound udiv's/urem's expanded");
+STATISTIC(NumZExt, "Number of non-negative deductions");
 
 static bool processSelect(SelectInst *S, LazyValueInfo *LVI) {
   if (S->getType()->isVectorTy() || isa<Constant>(S->getCondition()))
@@ -1032,6 +1033,24 @@ static bool processSExt(SExtInst *SDI, LazyValueInfo *LVI) {
   return true;
 }
 
+static bool processZExt(ZExtInst *ZExt, LazyValueInfo *LVI) {
+  if (ZExt->getType()->isVectorTy())
+    return false;
+
+  if (ZExt->hasNonNeg())
+    return false;
+
+  const Use &Base = ZExt->getOperandUse(0);
+  if (!LVI->getConstantRangeAtUse(Base, /*UndefAllowed*/ false)
+           .isAllNonNegative())
+    return false;
+
+  ++NumZExt;
+  ZExt->setNonNeg();
+
+  return true;
+}
+
 static bool processBinOp(BinaryOperator *BinOp, LazyValueInfo *LVI) {
   using OBO = OverflowingBinaryOperator;
 
@@ -1162,6 +1181,9 @@ static bool runImpl(Function &F, LazyValueInfo *LVI, DominatorTree *DT,
       case Instruction::SExt:
         BBChanged |= processSExt(cast<SExtInst>(&II), LVI);
         break;
+      case Instruction::ZExt:
+        BBChanged |= processZExt(cast<ZExtInst>(&II), LVI);
+        break;
       case Instruction::Add:
       case Instruction::Sub:
       case Instruction::Mul:

llvm/test/Transforms/CorrelatedValuePropagation/zext.ll

@@ -0,0 +1,173 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes=correlated-propagation -S | FileCheck %s
+
+declare void @use64(i64)
+
+define void @test1(i32 %n) {
+; CHECK-LABEL: @test1(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[FOR_COND:%.*]]
+; CHECK:       for.cond:
+; CHECK-NEXT:    [[A:%.*]] = phi i32 [ [[N:%.*]], [[ENTRY:%.*]] ], [ [[EXT:%.*]], [[FOR_BODY:%.*]] ]
+; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[A]], -1
+; CHECK-NEXT:    br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END:%.*]]
+; CHECK:       for.body:
+; CHECK-NEXT:    [[EXT_WIDE:%.*]] = zext nneg i32 [[A]] to i64
+; CHECK-NEXT:    call void @use64(i64 [[EXT_WIDE]])
+; CHECK-NEXT:    [[EXT]] = trunc i64 [[EXT_WIDE]] to i32
+; CHECK-NEXT:    br label [[FOR_COND]]
+; CHECK:       for.end:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %for.cond
+
+for.cond:                                         ; preds = %for.body, %entry
+  %a = phi i32 [ %n, %entry ], [ %ext, %for.body ]
+  %cmp = icmp sgt i32 %a, -1
+  br i1 %cmp, label %for.body, label %for.end
+
+for.body:                                         ; preds = %for.cond
+  %ext.wide = zext i32 %a to i64
+  call void @use64(i64 %ext.wide)
+  %ext = trunc i64 %ext.wide to i32
+  br label %for.cond
+
+for.end:                                          ; preds = %for.cond
+  ret void
+}
+
+;; Negative test to show transform doesn't happen unless n >= 0.
+define void @test2(i32 %n) {
+; CHECK-LABEL: @test2(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[FOR_COND:%.*]]
+; CHECK:       for.cond:
+; CHECK-NEXT:    [[A:%.*]] = phi i32 [ [[N:%.*]], [[ENTRY:%.*]] ], [ [[EXT:%.*]], [[FOR_BODY:%.*]] ]
+; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[A]], -2
+; CHECK-NEXT:    br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END:%.*]]
+; CHECK:       for.body:
+; CHECK-NEXT:    [[EXT_WIDE:%.*]] = zext i32 [[A]] to i64
+; CHECK-NEXT:    call void @use64(i64 [[EXT_WIDE]])
+; CHECK-NEXT:    [[EXT]] = trunc i64 [[EXT_WIDE]] to i32
+; CHECK-NEXT:    br label [[FOR_COND]]
+; CHECK:       for.end:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %for.cond
+
+for.cond:                                         ; preds = %for.body, %entry
+  %a = phi i32 [ %n, %entry ], [ %ext, %for.body ]
+  %cmp = icmp sgt i32 %a, -2
+  br i1 %cmp, label %for.body, label %for.end
+
+for.body:                                         ; preds = %for.cond
+  %ext.wide = zext i32 %a to i64
+  call void @use64(i64 %ext.wide)
+  %ext = trunc i64 %ext.wide to i32
+  br label %for.cond
+
+for.end:                                          ; preds = %for.cond
+  ret void
+}
+
+;; Non looping test case.
+define void @test3(i32 %n) {
+; CHECK-LABEL: @test3(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[N:%.*]], -1
+; CHECK-NEXT:    br i1 [[CMP]], label [[BB:%.*]], label [[EXIT:%.*]]
+; CHECK:       bb:
+; CHECK-NEXT:    [[EXT_WIDE:%.*]] = zext nneg i32 [[N]] to i64
+; CHECK-NEXT:    call void @use64(i64 [[EXT_WIDE]])
+; CHECK-NEXT:    [[EXT:%.*]] = trunc i64 [[EXT_WIDE]] to i32
+; CHECK-NEXT:    br label [[EXIT]]
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  %cmp = icmp sgt i32 %n, -1
+  br i1 %cmp, label %bb, label %exit
+
+bb:
+  %ext.wide = zext i32 %n to i64
+  call void @use64(i64 %ext.wide)
+  %ext = trunc i64 %ext.wide to i32
+  br label %exit
+
+exit:
+  ret void
+}
+
+;; Non looping negative test case.
+define void @test4(i32 %n) {
+; CHECK-LABEL: @test4(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[N:%.*]], -2
+; CHECK-NEXT:    br i1 [[CMP]], label [[BB:%.*]], label [[EXIT:%.*]]
+; CHECK:       bb:
+; CHECK-NEXT:    [[EXT_WIDE:%.*]] = zext i32 [[N]] to i64
+; CHECK-NEXT:    call void @use64(i64 [[EXT_WIDE]])
+; CHECK-NEXT:    [[EXT:%.*]] = trunc i64 [[EXT_WIDE]] to i32
+; CHECK-NEXT:    br label [[EXIT]]
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  %cmp = icmp sgt i32 %n, -2
+  br i1 %cmp, label %bb, label %exit
+
+bb:
+  %ext.wide = zext i32 %n to i64
+  call void @use64(i64 %ext.wide)
+  %ext = trunc i64 %ext.wide to i32
+  br label %exit
+
+exit:
+  ret void
+}
+
+define i64 @may_including_undef(i1 %c.1, i1 %c.2) {
+; CHECK-LABEL: @may_including_undef(
+; CHECK-NEXT:    br i1 [[C_1:%.*]], label [[TRUE_1:%.*]], label [[FALSE:%.*]]
+; CHECK:       true.1:
+; CHECK-NEXT:    br i1 [[C_2:%.*]], label [[TRUE_2:%.*]], label [[EXIT:%.*]]
+; CHECK:       true.2:
+; CHECK-NEXT:    br label [[EXIT]]
+; CHECK:       false:
+; CHECK-NEXT:    br label [[EXIT]]
+; CHECK:       exit:
+; CHECK-NEXT:    [[P:%.*]] = phi i32 [ 0, [[TRUE_1]] ], [ 1, [[TRUE_2]] ], [ undef, [[FALSE]] ]
+; CHECK-NEXT:    [[EXT:%.*]] = zext i32 [[P]] to i64
+; CHECK-NEXT:    ret i64 [[EXT]]
+;
+  br i1 %c.1, label %true.1, label %false
+
+true.1:
+  br i1 %c.2, label %true.2, label %exit
+
+true.2:
+  br label %exit
+
+false:
+  br label %exit
+
+exit:
+  %p = phi i32 [ 0, %true.1 ], [ 1, %true.2], [ undef, %false ]
+  %ext = zext i32 %p to i64
+  ret i64 %ext
+}
+
+define i64 @test_infer_at_use(i32 noundef %n) {
+; CHECK-LABEL: @test_infer_at_use(
+; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[N:%.*]], -1
+; CHECK-NEXT:    [[EXT:%.*]] = zext nneg i32 [[N]] to i64
+; CHECK-NEXT:    [[SELECT:%.*]] = select i1 [[CMP]], i64 [[EXT]], i64 0
+; CHECK-NEXT:    ret i64 [[SELECT]]
+;
+  %cmp = icmp sgt i32 %n, -1
+  %ext = zext i32 %n to i64
+  %select = select i1 %cmp, i64 %ext, i64 0
+  ret i64 %select
+}

llvm/test/Transforms/LoopVectorize/X86/float-induction-x86.ll

@@ -20,11 +20,11 @@ define void @fp_iv_loop1(ptr noalias nocapture %A, i32 %N) #0 {
 ; AUTO_VEC-NEXT:    [[CMP4:%.*]] = icmp sgt i32 [[N:%.*]], 0
 ; AUTO_VEC-NEXT:    br i1 [[CMP4]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]]
 ; AUTO_VEC:       for.body.preheader:
-; AUTO_VEC-NEXT:    [[ZEXT:%.*]] = zext i32 [[N]] to i64
+; AUTO_VEC-NEXT:    [[ZEXT:%.*]] = zext nneg i32 [[N]] to i64
 ; AUTO_VEC-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[N]], 32
 ; AUTO_VEC-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[FOR_BODY:%.*]], label [[VECTOR_PH:%.*]]
 ; AUTO_VEC:       vector.ph:
-; AUTO_VEC-NEXT:    [[N_VEC:%.*]] = and i64 [[ZEXT]], 4294967264
+; AUTO_VEC-NEXT:    [[N_VEC:%.*]] = and i64 [[ZEXT]], 2147483616
 ; AUTO_VEC-NEXT:    [[DOTCAST:%.*]] = sitofp i64 [[N_VEC]] to float
 ; AUTO_VEC-NEXT:    [[TMP0:%.*]] = fmul fast float [[DOTCAST]], 5.000000e-01
 ; AUTO_VEC-NEXT:    [[IND_END:%.*]] = fadd fast float [[TMP0]], 1.000000e+00
@@ -103,12 +103,12 @@ define void @fp_iv_loop2(ptr noalias nocapture %A, i32 %N) {
 ; AUTO_VEC-NEXT:    [[CMP4:%.*]] = icmp sgt i32 [[N:%.*]], 0
 ; AUTO_VEC-NEXT:    br i1 [[CMP4]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]]
 ; AUTO_VEC:       for.body.preheader:
-; AUTO_VEC-NEXT:    [[ZEXT:%.*]] = zext i32 [[N]] to i64
+; AUTO_VEC-NEXT:    [[ZEXT:%.*]] = zext nneg i32 [[N]] to i64
 ; AUTO_VEC-NEXT:    [[XTRAITER:%.*]] = and i64 [[ZEXT]], 7
 ; AUTO_VEC-NEXT:    [[TMP0:%.*]] = icmp ult i32 [[N]], 8
 ; AUTO_VEC-NEXT:    br i1 [[TMP0]], label [[FOR_END_LOOPEXIT_UNR_LCSSA:%.*]], label [[FOR_BODY_PREHEADER_NEW:%.*]]
 ; AUTO_VEC:       for.body.preheader.new:
-; AUTO_VEC-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[ZEXT]], 4294967288
+; AUTO_VEC-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[ZEXT]], 2147483640
 ; AUTO_VEC-NEXT:    br label [[FOR_BODY:%.*]]
 ; AUTO_VEC:       for.body:
 ; AUTO_VEC-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ 0, [[FOR_BODY_PREHEADER_NEW]] ], [ [[INDVARS_IV_NEXT_7:%.*]], [[FOR_BODY]] ]

llvm/test/Transforms/PhaseOrdering/X86/excessive-unrolling.ll

@@ -171,11 +171,11 @@ define void @test_runtime_trip_count(i32 %N) {
 ; CHECK-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[N:%.*]], 0
 ; CHECK-NEXT:    br i1 [[CMP1]], label [[FOR_BODY_PREHEADER:%.*]], label [[EXIT:%.*]]
 ; CHECK:       for.body.preheader:
-; CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[N]] to i64
+; CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext nneg i32 [[N]] to i64
 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[N]], 4
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[FOR_BODY_PREHEADER7:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
-; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[WIDE_TRIP_COUNT]], 4294967292
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[WIDE_TRIP_COUNT]], 2147483644
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]

llvm/test/Transforms/PhaseOrdering/X86/pixel-splat.ll

@@ -24,11 +24,11 @@ define void @loop_or(ptr noalias %pIn, ptr noalias %pOut, i32 %s) {
 ; CHECK-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[S:%.*]], 0
 ; CHECK-NEXT:    br i1 [[CMP1]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]]
 ; CHECK:       for.body.preheader:
-; CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[S]] to i64
+; CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext nneg i32 [[S]] to i64
 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[S]], 8
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[FOR_BODY_PREHEADER5:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
-; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[WIDE_TRIP_COUNT]], 4294967288
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[WIDE_TRIP_COUNT]], 2147483640
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]

llvm/test/Transforms/PhaseOrdering/X86/vdiv.ll

@@ -26,7 +26,7 @@ define void @vdiv(ptr %x, ptr %y, double %a, i32 %N) #0 {
 ; CHECK-NEXT:    [[OR_COND:%.*]] = select i1 [[MIN_ITERS_CHECK]], i1 true, i1 [[DIFF_CHECK]]
 ; CHECK-NEXT:    br i1 [[OR_COND]], label [[FOR_BODY_PREHEADER9:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
-; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[WIDE_TRIP_COUNT]], 4294967280
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[WIDE_TRIP_COUNT]], 2147483632
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x double> poison, double [[A:%.*]], i64 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x double> [[BROADCAST_SPLATINSERT]], <4 x double> poison, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP1:%.*]] = fdiv fast <4 x double> <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>, [[BROADCAST_SPLAT]]