diff --git a/llvm/lib/Transforms/InstCombine/InstCombineAddSub.cpp b/llvm/lib/Transforms/InstCombine/InstCombineAddSub.cpp
--- a/llvm/lib/Transforms/InstCombine/InstCombineAddSub.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineAddSub.cpp
@@ -2352,12 +2352,28 @@
     return BinaryOperator::CreateNeg(USub);
   }
 
-  // C - ctpop(X) => ctpop(~X) if C is bitwidth
-  if (match(Op0, m_SpecificInt(Ty->getScalarSizeInBits())) &&
-      match(Op1, m_OneUse(m_Intrinsic<Intrinsic::ctpop>(m_Value(X)))))
-    return replaceInstUsesWith(
-        I, Builder.CreateIntrinsic(Intrinsic::ctpop, {I.getType()},
-                                   {Builder.CreateNot(X)}));
+  // C - X if C is bitwidth
+  if (match(Op0, m_SpecificInt(Ty->getScalarSizeInBits()))) {
+    // C - ctpop(X) => ctpop(~X) if C is bitwidth
+    if (match(Op1, m_OneUse(m_Intrinsic<Intrinsic::ctpop>(m_Value(X)))))
+      return replaceInstUsesWith(
+          I, Builder.CreateIntrinsic(Intrinsic::ctpop, {I.getType()},
+                                     {Builder.CreateNot(X)}));
+
+    // C - cttz(X) => ctlz(X) if C is bitwidth
+    if (match(Op1, m_OneUse(m_Intrinsic<Intrinsic::cttz>(m_Value(X)))))
+      return replaceInstUsesWith(
+          I, Builder.CreateIntrinsic(
+                 Intrinsic::ctlz, {I.getType()},
+                 {X, cast<IntrinsicInst>(Op1)->getArgOperand(1)}));
+
+    // C - ctlz(X) => cttz(X) if C is bitwidth
+    if (match(Op1, m_OneUse(m_Intrinsic<Intrinsic::ctlz>(m_Value(X)))))
+      return replaceInstUsesWith(
+          I, Builder.CreateIntrinsic(
+                 Intrinsic::cttz, {I.getType()},
+                 {X, cast<IntrinsicInst>(Op1)->getArgOperand(1)}));
+  }
 
   return TryToNarrowDeduceFlags();
 }
diff --git a/llvm/test/Transforms/InstCombine/ctpop-cttz.ll b/llvm/test/Transforms/InstCombine/ctpop-cttz.ll
--- a/llvm/test/Transforms/InstCombine/ctpop-cttz.ll
+++ b/llvm/test/Transforms/InstCombine/ctpop-cttz.ll
@@ -20,9 +20,8 @@
 
 define <2 x i32> @ctpop1v(<2 x i32> %0) {
 ; CHECK-LABEL: @ctpop1v(
-; CHECK-NEXT:    [[TMP2:%.*]] = call <2 x i32> @llvm.cttz.v2i32(<2 x i32> [[TMP0:%.*]], i1 false)
-; CHECK-NEXT:    [[TMP3:%.*]] = sub nuw nsw <2 x i32> <i32 32, i32 32>, [[TMP2]]
-; CHECK-NEXT:    ret <2 x i32> [[TMP3]]
+; CHECK-NEXT:    [[TMP2:%.*]] = call <2 x i32> @llvm.ctlz.v2i32(<2 x i32> [[TMP0:%.*]], i1 false)
+; CHECK-NEXT:    ret <2 x i32> [[TMP2]]
 ;
   %2 = sub <2 x i32> zeroinitializer, %0
   %3 = or <2 x i32> %2, %0
diff --git a/llvm/test/Transforms/InstCombine/fls-i16.ll b/llvm/test/Transforms/InstCombine/fls-i16.ll
--- a/llvm/test/Transforms/InstCombine/fls-i16.ll
+++ b/llvm/test/Transforms/InstCombine/fls-i16.ll
@@ -16,9 +16,8 @@
 ; CHECK-LABEL: @fold_fls(
 ; CHECK-NEXT:    call void @sink(i16 0)
 ; CHECK-NEXT:    call void @sink(i16 1)
-; CHECK-NEXT:    [[CTLZ:%.*]] = call i16 @llvm.ctlz.i16(i16 [[X:%.*]], i1 false), !range [[RNG0:![0-9]+]]
-; CHECK-NEXT:    [[TMP1:%.*]] = sub nuw nsw i16 16, [[CTLZ]]
-; CHECK-NEXT:    call void @sink(i16 [[TMP1]])
+; CHECK-NEXT:    [[NX:%.*]] = call i16 @llvm.cttz.i16(i16 [[X:%.*]], i1 false), !range [[RNG0:![0-9]+]]
+; CHECK-NEXT:    call void @sink(i16 [[NX]])
 ; CHECK-NEXT:    ret void
 ;
   %n0 = call i16 @fls(i16 0)
diff --git a/llvm/test/Transforms/InstCombine/fls.ll b/llvm/test/Transforms/InstCombine/fls.ll
--- a/llvm/test/Transforms/InstCombine/fls.ll
+++ b/llvm/test/Transforms/InstCombine/fls.ll
@@ -31,9 +31,8 @@
 
 define i32 @flsnotconst(i64 %z) {
 ; CHECK-LABEL: @flsnotconst(
-; CHECK-NEXT:    [[CTLZ:%.*]] = call i64 @llvm.ctlz.i64(i64 [[Z:%.*]], i1 false), !range [[RNG0:![0-9]+]]
-; CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[CTLZ]] to i32
-; CHECK-NEXT:    [[GOO:%.*]] = sub nsw i32 64, [[TMP1]]
+; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.cttz.i64(i64 [[Z:%.*]], i1 false), !range [[RNG0:![0-9]+]]
+; CHECK-NEXT:    [[GOO:%.*]] = trunc i64 [[TMP1]] to i32
 ; CHECK-NEXT:    ret i32 [[GOO]]
 ;
   %goo = call i32 @flsl(i64 %z)