[PATCH] D98785: [Orc] Fix pending debug object tracking in DebugObjectManagerPlugin

Wed Mar 17 09:24:41 PDT 2021

sgraenitz added a comment.

This is hard to reproduce with lli, because single-threaded execution with LLLazyJIT will implicitly sequence materialization. LLJIT instead materializes module dependencies recursively, which makes this issue easy to reproduce. Unfortunately, we have no LLJIT tool that I could use for a test. It's not the first time this limitation comes up for me and I wonder if it's worth adding an `orc-greedy` kind to lli. What do you think?

================
Comment at: llvm/lib/ExecutionEngine/Orc/DebugObjectManagerPlugin.cpp:522
+    // is still in-flight. If we have pending debug objects for such MRs, we
+    // will remove them here.
+    // FIXME: Getting the resource key from a MaterializationResponsibility is
----------------
Can we guaranteed that affected in-flight MRs are still alive at this point? Running the code as is works, but I wonder if it's a robust assumption.

================
Comment at: llvm/lib/ExecutionEngine/Orc/DebugObjectManagerPlugin.cpp:530
+    for (const auto &KV : PendingObjs) {
+      if (K == getResourceKey(*KV.first))
+        PendingMRsForKey.push_back(KV.first);
----------------
It's a different issue, but I realized that this might be a dangerous use of resource keys. They are only guaranteed to be valid within the lambda passed into `withResourceKeyDo()` right?

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98785/new/

https://reviews.llvm.org/D98785