[PATCH] D103593: [Coroutine] Sink lifetime markers after switch of suspend blocks to avoid disturbing must tail calls

[John McCall via Phabricator via llvm-commits]

rjmccall added inline comments.


================
Comment at: llvm/lib/Transforms/Coroutines/CoroFrame.cpp:2343
            "should have split coro.suspend into its own block");
-    DomSet.insert(SuspendBlock->getSingleSuccessor());
+    BasicBlock *SwitchBlock = SuspendBlock->getSingleSuccessor();
+    for (auto *BB : llvm::successors(SwitchBlock))
----------------
Can we assert that this is a switch block somehow?  This is very dangerous if the code pattern isn't what we expect.


================
Comment at: llvm/lib/Transforms/Coroutines/CoroFrame.cpp:2404
+          NewLifetime->replaceUsesOfWith(NewLifetime->getOperand(1),
+                                         NewBitCast);
+          NewLifetime->insertAfter(NewBitCast);
----------------
Is this not just `NewLifetime->setOperand(1, NewBitCast)`?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D103593/new/

https://reviews.llvm.org/D103593