In-proc event listener listening to native runtime event source events crashes on Mono Linux

[kouvel]

In PR #66765 I tried to use an in-proc EventListener to listen to a thread pool event. On Mono, this appears to work fine on Windows but is crashing at the following stack on Linux. Also crashed on OSX in the PR's CI.

The test is in System.Threading.ThreadPool tests named 'CooperativeBlockingCanCreateThreadsFaster'. For now, I'll disable the test on Mono in that PR.

  =================================================================
        Managed Stacktrace:
  =================================================================
          at <unknown> <0xffffffff>
          at System.Runtime.InteropServices.GCHandle:InternalGet <0x00079>
          at System.Runtime.InteropServices.GCHandle:get_Target <0x00047>
          at HandleManager:FromHandle <0x000ab>
          at System.Threading.WaitSubsystem:SetEvent <0x0002b>
          at System.Threading.EventWaitHandle:Set <0x0004b>
          at System.Diagnostics.Tracing.EventPipeEventDispatcher:StopDispatchTask <0x0009b>
          at System.Diagnostics.Tracing.EventPipeEventDispatcher:CommitDispatchConfiguration <0x0007b>
          at System.Diagnostics.Tracing.EventPipeEventDispatcher:RemoveEventListener <0x000cb>
          at System.Diagnostics.Tracing.EventListener:RemoveReferencesToListenerInEventSources <0x002eb>
          at System.Diagnostics.Tracing.EventListener:Dispose <0x0010f>
          at <>c:<CooperativeBlockingCanCreateThreadsFaster>b__38_0 <0x0052f>
          at System.Object:runtime_invoke_void__this__ <0x00091>
          at <unknown> <0xffffffff>
          at System.Reflection.RuntimeMethodInfo:InternalInvoke <0x000b9>
          at System.Reflection.RuntimeMethodInfo:InvokeWorker <0x0009f>
          at System.Reflection.RuntimeMethodInfo:Invoke <0x001d3>
          at System.Reflection.MethodBase:Invoke <0x00049>
          at Microsoft.DotNet.RemoteExecutor.Program:Main <0x00313>
          at <Module>:runtime_invoke_int_object <0x00091>
  =================================================================
      System.Threading.ThreadPools.Tests.ThreadPoolTests.CooperativeBlockingCanCreateThreadsFaster [FAIL]
        Exit code was 134 but it should have been 42
        Expected: True
        Actual:   False
        Stack Trace:
          /_/src/Microsoft.DotNet.RemoteExecutor/src/RemoteInvokeHandle.cs(239,0): at Microsoft.DotNet.RemoteExecutor.RemoteInvokeHandle.Dispose(Boolean disposing)
          /_/src/Microsoft.DotNet.RemoteExecutor/src/RemoteInvokeHandle.cs(57,0): at Microsoft.DotNet.RemoteExecutor.RemoteInvokeHandle.Dispose()
          /home/user/runtime/src/libraries/System.Threading.ThreadPool/tests/ThreadPoolTests.cs(914,0): at System.Threading.ThreadPools.Tests.ThreadPoolTests.CooperativeBlockingCanCreateThreadsFaster()
          /home/user/runtime/src/mono/System.Private.CoreLib/src/System/Reflection/RuntimeMethodInfo.Mono.cs(386,0): at System.Reflection.RuntimeMethodInfo.InvokeWorker(Object obj, BindingFlags invokeAttr, Span`1 parameters)
    Finished:    System.Threading.ThreadPool.Tests

[lateralusX]

Most likely fixed by #68320. @kouvel is the test still disabled? I couldn't find any reference in the test to this issue anymore. If not, then I guess this issue could be closed.

[lateralusX]

I run the CooperativeBlockingCanCreateThreadsFaster from ThreadPoolTests.cs test suite. Test runs and pass on OSX x64, so test is no longer disabled and works, closing issue.