[llvm] [RemoveDIs][DebugInfo] Add support for DPValues to LoopStrengthReduce (PR #78706)

[Jeremy Morse via llvm-commits]

================
@@ -6647,48 +6671,56 @@ static void DbgGatherSalvagableDVI(
     SmallSet<AssertingVH<DbgValueInst>, 2> &DVIHandles) {
   for (const auto &B : L->getBlocks()) {
     for (auto &I : *B) {
-      auto DVI = dyn_cast<DbgValueInst>(&I);
-      if (!DVI)
-        continue;
-      // Ensure that if any location op is undef that the dbg.vlue is not
-      // cached.
-      if (DVI->isKillLocation())
-        continue;
-
-      // Check that the location op SCEVs are suitable for translation to
-      // DIExpression.
-      const auto &HasTranslatableLocationOps =
-          [&](const DbgValueInst *DVI) -> bool {
-        for (const auto LocOp : DVI->location_ops()) {
-          if (!LocOp)
-            return false;
-
-          if (!SE.isSCEVable(LocOp->getType()))
-            return false;
-
-          const SCEV *S = SE.getSCEV(LocOp);
-          if (SE.containsUndefs(S))
-            return false;
+      auto ProcessDbgValue = [&](auto *DbgVal) -> bool {
+        // Ensure that if any location op is undef that the dbg.vlue is not
+        // cached.
+        if (DbgVal->isKillLocation())
+          return false;
+
+        // Check that the location op SCEVs are suitable for translation to
+        // DIExpression.
+        const auto &HasTranslatableLocationOps =
+            [&](const auto *DbgVal) -> bool {
----------------
jmorse wrote:

I'd recommend against shadowing the outer lambdas argument name, in case of future confusion. It's what the existing code is doing, but we're adding yet another layer of abstraction in, so best to protect against more opportunities for confusion.

https://github.com/llvm/llvm-project/pull/78706