$ uname -a Linux nettle 2.6.20-1.2307.fc5 #1 Sun Mar 18 20:44:48 EDT 2007 i686 i686 i386 GNU/Linux testStepSigLongJmp(frysk.rt.TestStepping)junit.framework.AssertionFailedError: event loop run explictly stopped (Attempting to add observer) at frysk.proc.TestLib.assertRunUntilStop(TestRunner) at frysk.proc.TestLib.assertRunUntilStop(TestRunner) at frysk.rt.TestStepping.testStepSigLongJmp(TestRunner) at frysk.junit.Runner.runCases(TestRunner) at frysk.junit.Runner.runArchCases(TestRunner) at frysk.junit.Runner.runTestCases(TestRunner) at TestRunner.main(TestRunner)
Does this also happen with an older (pre-trace) 2.6.17 kernel or a newer 2.6.20 (2933 or higher) kernel with utrace fixes?
Its very likely this failure is just a result of the enormous amount of instruction stepping required for this test. I'll look into it soon.
I've been looking in to this, and it looks like after a lot of instruction stepping (which is required at this point, and will be the case until step-out works, pending libunwind), then the following line simply refuses to process: Task.requestUnblock(TaskObserver.Instruction); The call is made, but never returns and instead hangs. Then the testcase times out and fails.
Fixed in latest version of TestStepping. 2006-07-27 Mike Cvet <mcvet@redhat.com> * InstructionStepState.java: Moved here from frysk.rt * LineStepState.java: Ditto. * NextInstructionStepState.java: Ditto. * NextInstructionStepTestState.java: Ditto. * RunningState.java: Ditto. * State.java: Ditto. * StepAdvanceState.java: Ditto. * StepOutState.java: Ditto. * StepOverState.java: Ditto. * SteppingEngine.java: Ditto. * StoppedState.java: Ditto. * TaskStepEngine.java: Ditto. * TestSteppingEngine.java: Ditto. * TestStepping.java: New file. Fixes #4853. * TestStepping.java: Marked a couple of tests as broken, via 4855. * TestStepping.java (testStepSigLongJmp): Enabled test.