-
Bug
-
Resolution: Fixed
-
P3
-
7
-
b32
-
sparc
-
solaris
-
Verified
During PIT for JDK7 b22 2 tests:
nsk/jdi/stress/serial/monitorEvents001
nsk/jdi/MonitorContendedEnteredRequest/addClassExclusionFilter
failed because of unexpected NullPointerException from LocatableEventImpl.toString (looks like event has null location):
java.lang.NullPointerException
at com.sun.tools.jdi.EventSetImpl$LocatableEventImpl.toString(EventSetImpl.java:213)
at java.lang.String.valueOf(String.java:2839)
at java.lang.StringBuilder.append(StringBuilder.java:133)
at nsk.share.jdi.JDIEventsDebugger.eventToString(JDIEventsDebugger.java:292)
at nsk.share.jdi.JDIEventsDebugger.checkEvents(JDIEventsDebugger.java:223)
at nsk.share.jdi.JDIEventsDebugger$EventListener.eventReceived(JDIEventsDebugger.java:171)
at nsk.share.jdi.EventHandler.run(EventHandler.java:156)
at java.lang.Thread.run(Thread.java:675)
(failures logs are here: http://gtee.sfbay/gtee2.0/results/JDK7/NIGHTLY/VM/2007-10-05/Main_Baseline_JDK7/vm/solaris-sparcv9/server/mixed/vm-solaris-sparcv9_server_mixed_nsk.jdi.testlist2007-10-06-01-07-45/analysis.html
Note: log for monitorEvents001 also contains another error message, I'm investigating this error now:
# ERROR: Expected events was not generated:
# ERROR: expected event: com.sun.jdi.event.MonitorContendedEnteredEvent monitor: instance of nsk.share.jdi.MonitorEnterExecutor_1Subclass(id=408) thread: instance of nsk.share.jdi.JDIEventsDebuggee$EventActionsThread(name='Thread-243', id=405)
# ERROR: Debugger nsk.share.jdi.EventTestTemplates$ClassFilterTest_ClassName@4ce32802 finished with errors
)
This tests start several threads generating MonitorContendedEnteredEvents, debugger checks that all events were received. Failure reproduces at vm-ultra80-02.sfbay with solaris-sparc and solaris-sparcv9 JDKs when -server option is used. I have created minimized test reproducing this failure (debugee VM flags are specified as test arguments):
CLASSPATH=$JDK/lib/tools.jar
export CLASSPATH
$JDK/bin/java MonitorEvents_Debugger -server
Debuggee VM options: -server
ServerThread: ServerSocket.accept()
debuggee out>>>Try create socket (attempt 1)
debuggee out>>>Create socket
ServerThread: create streams
debuggee out>>>Create streams
ServerThread: finish
debuggee out>>>Streams created
debuggee out>>>Command received: COMMAND_START_TEST
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
Unexpected exception:
java.lang.NullPointerException
at com.sun.tools.jdi.EventSetImpl$LocatableEventImpl.toString(EventSetImpl.java:213)
at MonitorEvents_Debugger$EventListener.run(MonitorEvents_Debugger.java:314)
Unexpected exception:
java.lang.NullPointerException
at com.sun.tools.jdi.EventSetImpl$LocatableEventImpl.toString(EventSetImpl.java:213)
at MonitorEvents_Debugger$EventListener.run(MonitorEvents_Debugger.java:314)
debuggee out>>>Command received: QUIT
debuggee out>>>Exit
Stop listen events
Exit code: 95
TEST FAIL
We have observed several intermittent failures in nightly
that appear to be related to this bug. I'm adding my analysis
report entries here:
New nsk.quick_jdi failures (from 2007.10.05)
nsk/jdi/stress/serial/monitorEvents001
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC Client VM (machine naboovm). This
is the same failure mode as MonitorContendedEnteredRequest/
addClassFilter_ClassName below.
New nsk.quick_jdi failures (from 2007.10.03)
nsk/jdi/MonitorContendedEnterRequest/addClassFilter_ClassName
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC Server VM (machine vm-t1000-04).
This is the same failure mode as MonitorContendedEnteredRequest/
addClassFilter_ClassName below.
* nsk/jdi/stress/serial/monitorEvents002
This test failed due to "ERROR: Expected events was not
generated" on Win32 Client VM (machine vm-8450r-2). This is the
same failure mode as MonitorContendedEnteredRequest/
addClassFilter_ClassName below.
New nsk.quick_jdi failures (from 2007.10.02)
nsk/jdi/MonitorContendedEnterRequest/addClassExclusionFilter
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC-64 Server VM (machine vm-v215-03).
This is the same failure mode as addClassFilter_ClassName
below.
New nsk.quick_jdi failures (from 2007.10.01)
nsk/jdi/MonitorContendedEnteredRequest/addClassFilter_ClassName
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC-64 Server VM (machine
toasteroven).
nsk/jdi/MonitorContendedEnteredRequest/addClassFilter_ReferenceType
nsk/jdi/stress/serial/monitorEvents001
nsk/jdi/stress/serial/monitorEvents002
These tests failed the following assertion:
Internal Error (src/share/vm/runtime/jniHandles.hpp:188)
Error: assert(handle != 0L,"JNI handle should not be null")
on Solaris SPARC-64 Server VM (machine toasteroven).
Another intermittent failure from nightly:
New nsk.quick_jdi failures (from 2007.10.22)
nsk/jdi/MonitorContendedEnterRequest/addClassFilter_ReferenceType
This test failed due to "ERROR: Expected events was not
generated" on Linux AMD64 Server VM (machine vm-v20z-23). This
appears to be another occurrence of:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
Another intermittent failure from nightly:
New nsk.quick_jdi failures (from 2007.10.31)
* nsk/jdi/MonitorContendedEnteredRequest/addThreadFilter
This test failed due to "ERROR: Expected events was not
generated" on Linux IA32 Client VM (machine robodog). This
appears to be another occurrence of:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
Another failure mode for monitorEvents002 from nightly:
New nsk.quick_jdi failures (from 2007.09.28)
nsk/jdi/stress/serial/monitorEvents002
This test failed due to "ERROR: Unexpected exception:
java.lang.NullPointerException" on Linux IA32 Server VM
(machine sunfire002) and Solaris SPARC-64 Server VM (machine
vm-ultra80-02).
Update: Swamy says this is an occurrence of the following bug:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556. This test is already
on the fail_list due to 6614556, but is there for a
different failure mode. I will delete this entry in the
next report.
Another tests failing because of NullPointerException:
nsk/jdi/MonitorContendedEnteredRequest/MonitorContendedEnteredRequest001
(http://sqeweb/nfs/tools/gtee/results/JDK7/NIGHTLY/VM/2007-11-24/HSX_Baseline/vm/solaris-sparc-any/server/mixed/vm-solaris-sparc-any_server_mixed_nsk.jdi.testlist2007-11-25-01-56-38/ResultDir/MonitorContendedEnteredRequest001/MonitorContendedEnteredRequest001.log)
nsk/jdi/stress/MonitorEvents/MonitorEvents001
(http://sqeweb/nfs/tools/gtee/results/JDK7/NIGHTLY/VM/2007-11-24/HSX_Baseline/vm/solaris-sparc-any/server/mixed/vm-solaris-sparc-any_server_mixed_nsk.jdi.testlist2007-11-25-01-56-38/ResultDir/MonitorEvents001/MonitorEvents001.log)
nsk/jdi/stress/MonitorEvents/MonitorEvents002
(http://sqeweb/nfs/tools/gtee/results/JDK7/NIGHTLY/VM/2007-11-24/HSX_Baseline/vm/solaris-sparc-any/server/mixed/vm-solaris-sparc-any_server_mixed_nsk.jdi.testlist2007-11-25-01-56-38/ResultDir/MonitorEvents002/MonitorEvents002.log)
Another intermittent failure from nightly:
New nsk.quick_jdi failures (from 2007.11.21)
* nsk/jdi/MonitorContendedEnterRequest/addThreadFilter
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC Client VM (machine vm-t2000-01).
This appears to be another occurrence of:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
Another occurrence of this failure mode in nightly:
New nsk.quick-jdi failures (from 2008.05.21)
* nsk/jdi/stress/serial/mixed002
This test failed due to:
ERROR: Expected events was not generated:
ERROR: expected event: com.sun.jdi.event.MonitorContendedEnterEvent
on Win32 Server VM (machine vm-x2200-01). This looks like
an occurrence of the following bug:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
nsk/jdi/stress/serial/monitorEvents001
nsk/jdi/MonitorContendedEnteredRequest/addClassExclusionFilter
failed because of unexpected NullPointerException from LocatableEventImpl.toString (looks like event has null location):
java.lang.NullPointerException
at com.sun.tools.jdi.EventSetImpl$LocatableEventImpl.toString(EventSetImpl.java:213)
at java.lang.String.valueOf(String.java:2839)
at java.lang.StringBuilder.append(StringBuilder.java:133)
at nsk.share.jdi.JDIEventsDebugger.eventToString(JDIEventsDebugger.java:292)
at nsk.share.jdi.JDIEventsDebugger.checkEvents(JDIEventsDebugger.java:223)
at nsk.share.jdi.JDIEventsDebugger$EventListener.eventReceived(JDIEventsDebugger.java:171)
at nsk.share.jdi.EventHandler.run(EventHandler.java:156)
at java.lang.Thread.run(Thread.java:675)
(failures logs are here: http://gtee.sfbay/gtee2.0/results/JDK7/NIGHTLY/VM/2007-10-05/Main_Baseline_JDK7/vm/solaris-sparcv9/server/mixed/vm-solaris-sparcv9_server_mixed_nsk.jdi.testlist2007-10-06-01-07-45/analysis.html
Note: log for monitorEvents001 also contains another error message, I'm investigating this error now:
# ERROR: Expected events was not generated:
# ERROR: expected event: com.sun.jdi.event.MonitorContendedEnteredEvent monitor: instance of nsk.share.jdi.MonitorEnterExecutor_1Subclass(id=408) thread: instance of nsk.share.jdi.JDIEventsDebuggee$EventActionsThread(name='Thread-243', id=405)
# ERROR: Debugger nsk.share.jdi.EventTestTemplates$ClassFilterTest_ClassName@4ce32802 finished with errors
)
This tests start several threads generating MonitorContendedEnteredEvents, debugger checks that all events were received. Failure reproduces at vm-ultra80-02.sfbay with solaris-sparc and solaris-sparcv9 JDKs when -server option is used. I have created minimized test reproducing this failure (debugee VM flags are specified as test arguments):
CLASSPATH=$JDK/lib/tools.jar
export CLASSPATH
$JDK/bin/java MonitorEvents_Debugger -server
Debuggee VM options: -server
ServerThread: ServerSocket.accept()
debuggee out>>>Try create socket (attempt 1)
debuggee out>>>Create socket
ServerThread: create streams
debuggee out>>>Create streams
ServerThread: finish
debuggee out>>>Streams created
debuggee out>>>Command received: COMMAND_START_TEST
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
debuggee out>>>Thread starts test
Unexpected exception:
java.lang.NullPointerException
at com.sun.tools.jdi.EventSetImpl$LocatableEventImpl.toString(EventSetImpl.java:213)
at MonitorEvents_Debugger$EventListener.run(MonitorEvents_Debugger.java:314)
Unexpected exception:
java.lang.NullPointerException
at com.sun.tools.jdi.EventSetImpl$LocatableEventImpl.toString(EventSetImpl.java:213)
at MonitorEvents_Debugger$EventListener.run(MonitorEvents_Debugger.java:314)
debuggee out>>>Command received: QUIT
debuggee out>>>Exit
Stop listen events
Exit code: 95
TEST FAIL
We have observed several intermittent failures in nightly
that appear to be related to this bug. I'm adding my analysis
report entries here:
New nsk.quick_jdi failures (from 2007.10.05)
nsk/jdi/stress/serial/monitorEvents001
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC Client VM (machine naboovm). This
is the same failure mode as MonitorContendedEnteredRequest/
addClassFilter_ClassName below.
New nsk.quick_jdi failures (from 2007.10.03)
nsk/jdi/MonitorContendedEnterRequest/addClassFilter_ClassName
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC Server VM (machine vm-t1000-04).
This is the same failure mode as MonitorContendedEnteredRequest/
addClassFilter_ClassName below.
* nsk/jdi/stress/serial/monitorEvents002
This test failed due to "ERROR: Expected events was not
generated" on Win32 Client VM (machine vm-8450r-2). This is the
same failure mode as MonitorContendedEnteredRequest/
addClassFilter_ClassName below.
New nsk.quick_jdi failures (from 2007.10.02)
nsk/jdi/MonitorContendedEnterRequest/addClassExclusionFilter
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC-64 Server VM (machine vm-v215-03).
This is the same failure mode as addClassFilter_ClassName
below.
New nsk.quick_jdi failures (from 2007.10.01)
nsk/jdi/MonitorContendedEnteredRequest/addClassFilter_ClassName
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC-64 Server VM (machine
toasteroven).
nsk/jdi/MonitorContendedEnteredRequest/addClassFilter_ReferenceType
nsk/jdi/stress/serial/monitorEvents001
nsk/jdi/stress/serial/monitorEvents002
These tests failed the following assertion:
Internal Error (src/share/vm/runtime/jniHandles.hpp:188)
Error: assert(handle != 0L,"JNI handle should not be null")
on Solaris SPARC-64 Server VM (machine toasteroven).
Another intermittent failure from nightly:
New nsk.quick_jdi failures (from 2007.10.22)
nsk/jdi/MonitorContendedEnterRequest/addClassFilter_ReferenceType
This test failed due to "ERROR: Expected events was not
generated" on Linux AMD64 Server VM (machine vm-v20z-23). This
appears to be another occurrence of:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
Another intermittent failure from nightly:
New nsk.quick_jdi failures (from 2007.10.31)
* nsk/jdi/MonitorContendedEnteredRequest/addThreadFilter
This test failed due to "ERROR: Expected events was not
generated" on Linux IA32 Client VM (machine robodog). This
appears to be another occurrence of:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
Another failure mode for monitorEvents002 from nightly:
New nsk.quick_jdi failures (from 2007.09.28)
nsk/jdi/stress/serial/monitorEvents002
This test failed due to "ERROR: Unexpected exception:
java.lang.NullPointerException" on Linux IA32 Server VM
(machine sunfire002) and Solaris SPARC-64 Server VM (machine
vm-ultra80-02).
Update: Swamy says this is an occurrence of the following bug:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556. This test is already
on the fail_list due to 6614556, but is there for a
different failure mode. I will delete this entry in the
next report.
Another tests failing because of NullPointerException:
nsk/jdi/MonitorContendedEnteredRequest/MonitorContendedEnteredRequest001
(http://sqeweb/nfs/tools/gtee/results/JDK7/NIGHTLY/VM/2007-11-24/HSX_Baseline/vm/solaris-sparc-any/server/mixed/vm-solaris-sparc-any_server_mixed_nsk.jdi.testlist2007-11-25-01-56-38/ResultDir/MonitorContendedEnteredRequest001/MonitorContendedEnteredRequest001.log)
nsk/jdi/stress/MonitorEvents/MonitorEvents001
(http://sqeweb/nfs/tools/gtee/results/JDK7/NIGHTLY/VM/2007-11-24/HSX_Baseline/vm/solaris-sparc-any/server/mixed/vm-solaris-sparc-any_server_mixed_nsk.jdi.testlist2007-11-25-01-56-38/ResultDir/MonitorEvents001/MonitorEvents001.log)
nsk/jdi/stress/MonitorEvents/MonitorEvents002
(http://sqeweb/nfs/tools/gtee/results/JDK7/NIGHTLY/VM/2007-11-24/HSX_Baseline/vm/solaris-sparc-any/server/mixed/vm-solaris-sparc-any_server_mixed_nsk.jdi.testlist2007-11-25-01-56-38/ResultDir/MonitorEvents002/MonitorEvents002.log)
Another intermittent failure from nightly:
New nsk.quick_jdi failures (from 2007.11.21)
* nsk/jdi/MonitorContendedEnterRequest/addThreadFilter
This test failed due to "ERROR: Expected events was not
generated" on Solaris SPARC Client VM (machine vm-t2000-01).
This appears to be another occurrence of:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
Another occurrence of this failure mode in nightly:
New nsk.quick-jdi failures (from 2008.05.21)
* nsk/jdi/stress/serial/mixed002
This test failed due to:
ERROR: Expected events was not generated:
ERROR: expected event: com.sun.jdi.event.MonitorContendedEnterEvent
on Win32 Server VM (machine vm-x2200-01). This looks like
an occurrence of the following bug:
6614556 3/4 null location for MonitorContendedEnteredEvents
I will add this entry to 6614556.
- relates to
-
JDK-6749729 vm crashes due to bad jni handle passed from backend when monitor events are enabled.
-
- Closed
-