Index: include/llvm/Analysis/ScalarEvolutionExpressions.h
===================================================================
--- include/llvm/Analysis/ScalarEvolutionExpressions.h
+++ include/llvm/Analysis/ScalarEvolutionExpressions.h
@@ -350,9 +350,7 @@
 
     /// Return an expression representing the value of this expression
     /// one iteration of the loop ahead.
-    const SCEVAddRecExpr *getPostIncExpr(ScalarEvolution &SE) const {
-      return cast<SCEVAddRecExpr>(SE.getAddExpr(this, getStepRecurrence(SE)));
-    }
+    const SCEVAddRecExpr *getPostIncExpr(ScalarEvolution &SE) const;
 
     /// Methods for support type inquiry through isa, cast, and dyn_cast:
     static bool classof(const SCEV *S) {
Index: lib/Analysis/ScalarEvolution.cpp
===================================================================
--- lib/Analysis/ScalarEvolution.cpp
+++ lib/Analysis/ScalarEvolution.cpp
@@ -10215,6 +10215,29 @@
   return SE.getCouldNotCompute();
 }
 
+const SCEVAddRecExpr *
+SCEVAddRecExpr::getPostIncExpr(ScalarEvolution &SE) const {
+  assert(getNumOperands() > 1 && "AddRec with zero step?");
+  // There is a temptation to just call getAddExpr(this, getStepRecurrence(SE)),
+  // but in this case we cannot guarantee that the value returned will be an
+  // AddRec because SCEV does not have a fixed point where it stops
+  // simplification: it is legal to return ({rec1} + {rec2}). For example, it
+  // may happen if we reach arithmetic depth limit while simplifying. So we
+  // construct the returned value explicitly.
+  SmallVector<const SCEV *, 3> Ops;
+  // If this is {A,+,B,+,C,...,+,N}, then its step is {B,+,C,+,...,+,N}, and
+  // (this + Step) is {A+B,+,B+C,+...,+,N}.
+  for (unsigned i = 0, e = getNumOperands() - 1; i < e; ++i)
+    Ops.push_back(SE.getAddExpr(getOperand(i), getOperand(i + 1)));
+  // We know that the last operand is not a constant zero (otherwise it would
+  // have been popped out earlier). This guarantees us that if the result has
+  // the same last operand, then it will also not be popped out, meaning that
+  // the returned value will be an AddRec.
+  Ops.push_back(getOperand(getNumOperands() - 1));
+  return cast<SCEVAddRecExpr>(SE.getAddRecExpr(Ops, getLoop(),
+                                               SCEV::FlagAnyWrap));
+}
+
 // Return true when S contains at least an undef value.
 static inline bool containsUndefs(const SCEV *S) {
   return SCEVExprContains(S, [](const SCEV *S) {
Index: test/Analysis/ScalarEvolution/pr35890.ll
===================================================================
--- /dev/null
+++ test/Analysis/ScalarEvolution/pr35890.ll
@@ -0,0 +1,44 @@
+; RUN: opt < %s -scalar-evolution-max-arith-depth=0  -indvars  -S | FileCheck %s
+
+target datalayout = "e-m:e-i32:64-f80:128-n8:16:32:64-S128-ni:1"
+target triple = "x86_64-unknown-linux-gnu"
+
+; Check that it does not crash because SCEVAddRec's step is not an AddRec.
+
+define void @pr35890(i32* %inc_ptr, i32 %a) {
+
+; CHECK-LABEL: @pr35890(
+
+entry:
+  %inc = load i32, i32* %inc_ptr, !range !0
+  %ne.cond = icmp ne i32 %inc, 0
+  br i1 %ne.cond, label %loop, label %bail
+
+loop:
+  %iv = phi i32 [ 0, %entry ], [ %iv.next, %backedge ]
+  %a.1 = add i32 %a, 1
+  %iv.next = add i32 %iv, %a.1
+  %iv.wide = zext i32 %iv to i64
+  %iv.square = mul i64 %iv.wide, %iv.wide
+  %iv.cube = mul i64 %iv.square, %iv.wide
+  %brcond = icmp slt i64 %iv.wide, %iv.cube
+  br i1 %brcond, label %if.true, label %if.false
+
+if.true:
+  br label %backedge
+
+if.false:
+  br label %backedge
+
+backedge:
+  %loopcond = icmp slt i32 %iv, 200
+  br i1 %loopcond, label %loop, label %exit
+
+exit:
+  ret void
+
+bail:
+  ret void
+}
+
+!0 = !{i32 0, i32 100}