In my analysis for r56076 I missed the case where the original

multiplication overflows.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@56082 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Transforms/Scalar/InstructionCombining.cpp b/lib/Transforms/Scalar/InstructionCombining.cpp
index 8f578200d659a73039196c97ae8824e9c48ffda7..46ac762b0a78349d3e9e6d71e91e5a3d4fabf655 100644 (file)
--- a/lib/Transforms/Scalar/InstructionCombining.cpp
+++ b/lib/Transforms/Scalar/InstructionCombining.cpp
@@ -5789,7 +5789,8 @@ Instruction *InstCombiner::FoldICmpDivCst(ICmpInst &ICI, BinaryOperator *DivI,
       // e.g. X/-5 op -3  --> [15, 20)
       LoBound = Prod;
       LoOverflow = HiOverflow = ProdOV ? 1 : 0;
-      HiOverflow = SubWithOverflow(HiBound, Prod, DivRHS, true);
+      if (!HiOverflow)
+        HiOverflow = SubWithOverflow(HiBound, Prod, DivRHS, true);
     }
     
     // Dividing by a negative swaps the condition.  LT <-> GT