[llvm] [InstCombine] Implemented missed optimization in muldivrem (PR #140916)

[Yingwei Zheng via llvm-commits]

================
@@ -1295,6 +1297,14 @@ Instruction *InstCombinerImpl::commonIDivRemTransforms(BinaryOperator &I) {
                                           /*FoldWithMultiUse*/ true))
       return R;
   }
+  if (match(Op0, m_OneUse(m_Intrinsic<Intrinsic::smul_fix>(m_APInt(C1), m_APInt(C2)))) &&
----------------
dtcxzyw wrote:

I don't think it is related to the original issue.


https://github.com/llvm/llvm-project/pull/140916