Name: egR10015 Date: 05/17/2001



HotSpot 1.4.0beta-b64 fails to pass the following tests

nsk/jdi/EventRequestManager/deleteEventRequest/delevtreq001
nsk/jdi/EventRequestManager/deleteEventRequests/delevtreqs001

from testbase_nsk on all platforms. The target VM crashes
with the error ID 494E54455250524554455232554E54494D450E43505003C0
(interpreterRuntime.cpp, 960) when a requested AccessWatchpointEvent
occurs.

The test delevtreq001 mentioned above checks that the JDI method
EventRequestManager.deleteEventRequest()
properly disables and removes an event request. The test consists of
two parts: a debugger (please see the source file delevtreq001.java)
and a debuggee (the source file delevtreq001t.java). The debugger
creates several AccessWatchpointRequests for notification when
the contents of a field are accessed in the debuggee, and then it
verifies that appropriate events are generated by the target VM.
After that the debugger removes one request by deleteEventRequest() call.
Finally, the test checks that the event corresponding to the deleted
request, are not generated by the target VM any more. But when
the debuggee performes the very first access to the first testing field
(the byte field "byteFld"), its VM crashes with the error mentioned above.
I've observed the following. The crash never happens when the debuggee
accesses the fields of integer only type. Note also that HS 1.3.1-b24
successfully passes the test.

To reproduce the bug run with HS1.4

sh doit.sh $JAVA_HOME

in
/net/sqesvr.eng/export/vsn/GammaBase/Bugs/<this bug number>

and wait one minute for the target VM crash

--------------- output on Solsparc with HS 1.4.0-b64 -----------------
$ sh doit.sh ~/hotspot/jdk1.4/solsparc
java version "1.4.0-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.0-beta-b64)
Java HotSpot(TM) Client VM (build 1.4.0-beta-b64, mixed mode)
binder> Listening for a pipe connection to port: 7000
binder> Binding to pipe socket
binder> Listening to pipe server socket
binder> VirtualMachineManager: version 1.4
binder> FindingConnector: default
binder> LaunchingConnector:
binder> name: com.sun.jdi.CommandLineLaunch
binder> description: Launches target using Sun Java VM command line and attaches to it
binder> transport: dt_socket
binder> Connector arguments:
binder> home=/export/ld54/java/dest/jdk1.4.0beta-b64/solsparc
binder> vmexec=java
binder> options= -classpath ".:share/classes:/home/eug/hotspot/jdk1.4/solsparc/lib/tools.jar"
binder> main=delevtreq001t "-verbose" "-waittime=1"
binder> quote="
binder> suspend=true
binder> Launch debugee
test.pipe> Waiting for pipe connection
debugee.err> debugee.pipe> Making pipe connection to socket: localhost:7000
debugee.err> Debuggee: sending the command: ready
Creating AccessWatchpointRequest for the debuggee's field:
byte byteFld
Creating AccessWatchpointRequest for the debuggee's field:
short shortFld
Creating AccessWatchpointRequest for the debuggee's field:
int intFld
handleEvent: following JDI event occured: VMStartEvent in thread main

a) Getting all requested AccessWatchpointEvents...

1) Sending the command "run1" to a debuggee
debugee.err> Debuggee: received the command: run1
debugee.out> #
debugee.out> # HotSpot Virtual Machine Error, Internal Error
debugee.out> # Please report this error at
debugee.out> # http://java.sun.com/cgi-bin/bugreport.cgi
debugee.out> #
debugee.out> # Error ID: 494E54455250524554455232554E54494D450E43505003C0 01
debugee.out> #
debugee.out> # Problematic Thread: prio=5 tid=0x298e0 nid=0x1 runnable
debugee.out> #
handleEvent: following JDI event occured: VMDisconnectEvent
Notification about the AccessWatchpointEvent #0 received,
or time has elapsed
# ERROR: TEST FAILURE: debuggee's reply is empty
test.pipe> Closing pipe socket
binder> Closing pipe server socket
exit code = 97
----------------------------------------------------------------------

--------------- output on Solsparc with HS 1.3.1-b24 -----------------
$ sh doit.sh ~/hotspot/jdk1.3.1/solsparc
java version "1.3.1"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.1-b24)
Java HotSpot(TM) Client VM (build 1.3.1-b24, mixed mode)
binder> Listening for a pipe connection to port: 7000
binder> Binding to pipe socket
binder> Listening to pipe server socket
binder> VirtualMachineManager: version 1.3
binder> FindingConnector: default
binder> LaunchingConnector:
binder> name: com.sun.jdi.CommandLineLaunch
binder> description: Launches target using Sun Java VM command line and attaches to it
binder> transport: dt_socket
binder> Connector arguments:
binder> home=/export/ld54/java/dest/jdk1.3.1-b24/solsparc
binder> vmexec=java
binder> options= -classpath ".:share/classes:/home/eug/hotspot/jdk1.3.1/solsparc/lib/tools.jar"
binder> main=delevtreq001t "-verbose" "-waittime=1"
binder> quote="
binder> suspend=true
binder> Launch debugee
test.pipe> Waiting for pipe connection
debugee.err> debugee.pipe> Making pipe connection to socket: localhost:7000
debugee.err> Debuggee: sending the command: ready
Creating AccessWatchpointRequest for the debuggee's field:
byte byteFld
Creating AccessWatchpointRequest for the debuggee's field:
short shortFld
Creating AccessWatchpointRequest for the debuggee's field:
int intFld
handleEvent: following JDI event occured: VMStartEvent in thread main

a) Getting all requested AccessWatchpointEvents...

1) Sending the command "run1" to a debuggee
debugee.err> Debuggee: received the command: run1
handleEvent: following JDI event occured: ###@###.###(java.lang.String[])+71 in thread
main
handleEvent: AccessWatchpointEvent for the debuggee's field #0:
byte byteFld
handleEvent: notifying about the received event #0
Notification about the AccessWatchpointEvent #0 received,
or time has elapsed
debugee.err> Debuggee: access to the field "byteFld" is done
debugee.err> Debuggee: sending the command: ready
Debuggee's reply received: ready
Got expected AccessWatchpointEvent for the following debuggee's field:
byte byteFld

2) Sending the command "run2" to a debuggee
debugee.err> Debuggee: received the command: run2
handleEvent: following JDI event occured: ###@###.###(java.lang.String[])+98 in thread
main
handleEvent: AccessWatchpointEvent for the debuggee's field #1:
short shortFld
handleEvent: notifying about the received event #1
Notification about the AccessWatchpointEvent #1 received,
or time has elapsed
Debuggee's reply received: ready
Got expected AccessWatchpointEvent for the following debuggee's field:
short shortFld

3) Sending the command "run3" to a debuggee
handleEvent: following JDI event occured: ###@###.###(java.lang.String[])+126 in thread
main
handleEvent: AccessWatchpointEvent for the debuggee's field #2:
int intFld
handleEvent: notifying about the received event #2
Notification about the AccessWatchpointEvent #2 received,
or time has elapsed
debugee.err> Debuggee: access to the field "shortFld" is done
debugee.err> Debuggee: sending the command: ready
debugee.err> Debuggee: received the command: run3
Debuggee's reply received: ready
Got expected AccessWatchpointEvent for the following debuggee's field:
int intFld

b) Getting AccessWatchpointEvents without request #1...

Removing AccessWatchpointRequest #1 for the debuggee's field:
short shortFld
debugee.err> Debuggee: access to the field "intFld" is done
debugee.err> Debuggee: sending the command: ready

1) Sending the command "run1" to a debuggee
debugee.err> Debuggee: received the command: run1
handleEvent: following JDI event occured: ###@###.###(java.lang.String[])+71 in thread
main
handleEvent: AccessWatchpointEvent for the debuggee's field #0:
byte byteFld
handleEvent: notifying about the received event #0
Notification about the AccessWatchpointEvent #0 received,
or time has elapsed
Debuggee's reply received: ready
Got expected AccessWatchpointEvent for the following debuggee's field:
byte byteFld

2) Sending the command "run2" to a debuggee
debugee.err> Debuggee: access to the field "byteFld" is done
debugee.err> Debuggee: sending the command: ready
debugee.err> Debuggee: received the command: run2
debugee.err> Debuggee: access to the field "shortFld" is done
debugee.err> Debuggee: sending the command: ready
Notification about the AccessWatchpointEvent #1 received,
or time has elapsed
Debuggee's reply received: ready

TEST PASSED: no event for the deleted AccessWatchpointRequest #1

3) Sending the command "run3" to a debuggee
handleEvent: following JDI event occured: ###@###.###(java.lang.String[])+126 in thread
main
handleEvent: AccessWatchpointEvent for the debuggee's field #2:
int intFld
handleEvent: notifying about the received event #2
Notification about the AccessWatchpointEvent #2 received,
or time has elapsed
debugee.err> Debuggee: received the command: run3
Debuggee's reply received: ready
Got expected AccessWatchpointEvent for the following debuggee's field:
int intFld
handleEvent: following JDI event occured: VMDeathEvent
debugee.err> Debuggee: access to the field "intFld" is done
debugee.err> Debuggee: sending the command: ready
debugee.err> Debuggee: received the command: quit
debugee.err> Debuggee: exiting
handleEvent: following JDI event occured: VMDisconnectEvent
test.pipe> Closing pipe socket
binder> Closing pipe server socket
exit code = 95
----------------------------------------------------------------------

======================================================================

Name: egR10015 Date: 05/22/2001


The bug also affects the following test from testbase_nsk:

nsk/jdi/EventRequestManager/createAccessWatchpointRequest/craccwtchpreq001

======================================================================