release/19.x: [IndVars] Check if WideInc available before trying to use it

[llvmbot]

Backport c9a5e1b

Requested by: @nikic

[llvmbot]

@llvm/pr-subscribers-llvm-transforms

Author: None (llvmbot)

Changes

Backport c9a5e1b

Requested by: @nikic

---

Full diff: https://github.com/llvm/llvm-project/pull/106892.diff

2 Files Affected:

• (modified) llvm/lib/Transforms/Utils/SimplifyIndVar.cpp (+17-11)
• (added) llvm/test/Transforms/IndVarSimplify/pr106239.ll (+36)

diff --git a/llvm/lib/Transforms/Utils/SimplifyIndVar.cpp b/llvm/lib/Transforms/Utils/SimplifyIndVar.cpp
index 5bda7c50c62c66..0b4a75e0bc52de 100644
--- a/llvm/lib/Transforms/Utils/SimplifyIndVar.cpp
+++ b/llvm/lib/Transforms/Utils/SimplifyIndVar.cpp
@@ -1928,18 +1928,24 @@ Instruction *WidenIV::widenIVUse(WidenIV::NarrowIVDefUse DU,
     if (!WideAddRec.first)
       return nullptr;
 
-    // Reuse the IV increment that SCEVExpander created. Recompute flags, unless
-    // the flags for both increments agree and it is safe to use the ones from
-    // the original inc. In that case, the new use of the wide increment won't
-    // be more poisonous.
-    bool NeedToRecomputeFlags =
-        !SCEVExpander::canReuseFlagsFromOriginalIVInc(OrigPhi, WidePhi,
-                                                      DU.NarrowUse, WideInc) ||
-        DU.NarrowUse->hasNoUnsignedWrap() != WideInc->hasNoUnsignedWrap() ||
-        DU.NarrowUse->hasNoSignedWrap() != WideInc->hasNoSignedWrap();
+    auto CanUseWideInc = [&]() {
+      if (!WideInc)
+        return false;
+      // Reuse the IV increment that SCEVExpander created. Recompute flags,
+      // unless the flags for both increments agree and it is safe to use the
+      // ones from the original inc. In that case, the new use of the wide
+      // increment won't be more poisonous.
+      bool NeedToRecomputeFlags =
+          !SCEVExpander::canReuseFlagsFromOriginalIVInc(
+              OrigPhi, WidePhi, DU.NarrowUse, WideInc) ||
+          DU.NarrowUse->hasNoUnsignedWrap() != WideInc->hasNoUnsignedWrap() ||
+          DU.NarrowUse->hasNoSignedWrap() != WideInc->hasNoSignedWrap();
+      return WideAddRec.first == WideIncExpr &&
+             Rewriter.hoistIVInc(WideInc, DU.NarrowUse, NeedToRecomputeFlags);
+    };
+
     Instruction *WideUse = nullptr;
-    if (WideAddRec.first == WideIncExpr &&
-        Rewriter.hoistIVInc(WideInc, DU.NarrowUse, NeedToRecomputeFlags))
+    if (CanUseWideInc())
       WideUse = WideInc;
     else {
       WideUse = cloneIVUser(DU, WideAddRec.first);
diff --git a/llvm/test/Transforms/IndVarSimplify/pr106239.ll b/llvm/test/Transforms/IndVarSimplify/pr106239.ll
new file mode 100644
index 00000000000000..8d5aa99539a5a7
--- /dev/null
+++ b/llvm/test/Transforms/IndVarSimplify/pr106239.ll
@@ -0,0 +1,36 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -passes=indvars < %s | FileCheck %s
+
+target datalayout = "n8:16:32:64"
+
+; Make sure it does not crash.
+
+define i32 @m() {
+; CHECK-LABEL: define i32 @m() {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    br label %[[FOR_BODY_I6:.*]]
+; CHECK:       [[FOR_BODY_I6]]:
+; CHECK-NEXT:    br i1 true, label %[[I_EXIT:.*]], label %[[IF_END_I:.*]]
+; CHECK:       [[IF_END_I]]:
+; CHECK-NEXT:    store i64 0, ptr null, align 8
+; CHECK-NEXT:    br label %[[FOR_BODY_I6]]
+; CHECK:       [[I_EXIT]]:
+; CHECK-NEXT:    ret i32 0
+;
+entry:
+  %div.i4 = sdiv i32 1, 0
+  br label %for.body.i6
+
+for.body.i6:                                      ; preds = %if.end.i, %entry
+  %add57.i = phi i32 [ %add.i7, %if.end.i ], [ 0, %entry ]
+  br i1 true, label %i.exit, label %if.end.i
+
+if.end.i:                                         ; preds = %for.body.i6
+  %add.i7 = add i32 %add57.i, %div.i4
+  %conv.i = zext i32 %add57.i to i64
+  store i64 %conv.i, ptr null, align 8
+  br label %for.body.i6
+
+i.exit:                                           ; preds = %for.body.i6
+  ret i32 0
+}

[tru]

@shafik @AaronBallman is this good to be merged?

[AaronBallman]

I think you may have intended to tag someone other than Shafik and I for this one, but the changes LGTM nonetheless.

[fhahn]

LGTM looks like a safe cherry-pick

[github-actions]

@nikic (or anyone else). If you would like to add a note about this fix in the release notes (completely optional). Please reply to this comment with a one or two sentence description of the fix. When you are done, please add the release:note label to this PR.