[llvm] [SimplifyCFG] Simplify switch instruction that has duplicate arms (PR #114262)

[Yingwei Zheng via llvm-commits]

================
@@ -7436,6 +7437,110 @@ static bool simplifySwitchOfCmpIntrinsic(SwitchInst *SI, IRBuilderBase &Builder,
   return true;
 }
 
+bool SimplifyCFGOpt::simplifyDuplicateSwitchArms(SwitchInst *SI) {
+  // Simplify the case where multiple arms contain only a terminator, the
+  // terminators are the same, and their sucessor PHIS incoming values are the
+  // same.
+
+  // Find BBs that are candidates for simplification.
+  SmallPtrSet<BasicBlock *, 8> BBs;
+  for (auto &Case : SI->cases()) {
+    BasicBlock *BB = Case.getCaseSuccessor();
+
+    // FIXME: This case needs some extra care because the terminators other than
+    // SI need to be updated.
+    if (!BB->hasNPredecessors(1))
+      continue;
+
+    // FIXME: Relax that the terminator is a BranchInst by checking for equality
+    // on other kinds of terminators.
+    Instruction *T = BB->getTerminator();
+    if (T && isa<BranchInst>(T))
+      BBs.insert(BB);
+  }
+
+  auto IsBranchEq = [](BranchInst *A, BranchInst *B) {
+    if (A->isConditional() != B->isConditional())
+      return false;
+
+    if (A->isConditional()) {
+      // If the conditions are instructions, check equality up to commutativity.
+      // Otherwise, check that the two Values are the same.
+      Value *AC = A->getCondition();
+      Value *BC = B->getCondition();
+      auto *ACI = dyn_cast<Instruction>(AC);
+      auto *BCI = dyn_cast<Instruction>(BC);
+      if ((ACI && BCI && !areIdenticalUpToCommutativity(ACI, BCI)) && AC != BC)
+        return false;
+    }
+
+    if (A->getNumSuccessors() != B->getNumSuccessors())
+      return false;
+
+    for (unsigned I = 0; I < A->getNumSuccessors(); ++I) {
+      BasicBlock *ASucc = A->getSuccessor(I);
+      if (ASucc != B->getSuccessor(I))
+        return false;
+      // Need to check that PHIs in sucessors have matching values
+      for (PHINode &Phi : ASucc->phis())
+        if (Phi.getIncomingValueForBlock(A->getParent()) !=
+            Phi.getIncomingValueForBlock(B->getParent()))
+          return false;
+    }
+
+    return true;
+  };
+
+  auto IsBBEqualTo = [&IsBranchEq](BasicBlock *A, BasicBlock *B) {
+    // FIXME: Support more than just a single BranchInst. One way we could do
+    // this is by taking a hashing approach.
+    if (A->size() != 1 || B->size() != 1)
+      return false;
+
+    return IsBranchEq(cast<BranchInst>(A->getTerminator()),
+                      cast<BranchInst>(B->getTerminator()));
+  };
+
+  // Construct a map from candidate basic block to an equivalent basic block
+  // to replace it with. All equivalent basic blocks should be replaced with
+  // the same basic block. To do this, if there is no equivalent BB in the map,
+  // then insert into the map BB -> BB. Otherwise, we should check only elements
+  // in the map for equivalence to ensure that all equivalent BB get replaced
+  // by the BB in the map. Replacing BB with BB has no impact, so we skip
+  // a call to setSuccessor when we do the actual replacement.
+  DenseMap<BasicBlock *, BasicBlock *> ReplaceWith;
+  for (BasicBlock *BB : BBs) {
+    bool Inserted = false;
+    for (auto KV : ReplaceWith) {
+      if (IsBBEqualTo(BB, KV.first)) {
+        ReplaceWith[BB] = KV.first;
+        Inserted = true;
+        break;
+      }
+    }
+    if (!Inserted)
+      ReplaceWith[BB] = BB;
+  }
----------------
dtcxzyw wrote:

I agree we need a hash-based O(n) method to merge these blocks.

Compile-time impact:
```
Top 5 regressions:
  llvm/TokenKinds.cpp.ll 151977954 -> 792848805 +421.69%
  cvc5/kind.cpp.ll 230032379 -> 1046824473 +355.08%
  llvm/ASTCommon.cpp.ll 519331147 -> 2229199828 +329.24%
  llvm/X86EncodingOptimization.cpp.ll 424207707 -> 1707051386 +302.41%
  rust-analyzer-rs/4ifo5x52byu175vr.ll 177537147 -> 557775939 +214.17%
```

https://github.com/llvm/llvm-project/pull/114262