-
Bug
-
Resolution: Duplicate
-
P3
-
None
-
1.4.1
-
x86
-
linux, windows_2000
Name: egR10015 Date: 04/25/2002
Hprof agent shipped with HotSpot 1.4.1-b10 fails to pass the following test
nsk/hprof/regression/b4500875
from testbase_nsk on Linux and Win2k platforms.
The mentioned test is a regression one written to cover the following bug:
4500875 Very few CPU sampling due to failure in wait_for_suspend_completion
which has been closed as fixed verified. However, the described issue is
observed again with Hopper's build 10. Moreover, HPROF being run on Linux
in server mode takes no CPU sampling at all.
Note that HotSpot 1.4.1-b08 as well as HotSpot 1.4.0-b92 successfully pass
the test.
The test works as follows. Java thread is blocked on an OS system call (via
calling the Java method ServerSocket.accept()) for a time specified in
the internal variable DELAY, then the Perl part of the test checks that
HPROF has taken enough CPU sampling.
To reproduce the bug run with HS1.4.1
sh doit.sh $JAVA_HOME [JAVA_OPTS]
on Linux, Win2k platforms in
/net/sqesvr.sfbay/export/vsn/GammaBase/Bugs/<this bug number>
where JAVA_HOME should point to jdk1.4.1
JAVA_OPTS is an optional value specified VM mode
The profiling information is written out to the file b4500875.hprof.out.
------------------- output on Linux with HS 1.4.1-b10: --------------------
$ sh doit.sh ~/hotspot/jdk1.4.1/linux
java version "1.4.1-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.1-beta-b10)
Java HotSpot(TM) Client VM (build 1.4.1-beta-b10, mixed mode)
Executing the test ...
Dumping CPU usage by sampling running threads ... done.
TEST FAILED: very few CPU sampling 2 taken by HPROF
expected that it should be more than 300
exit code = 97
$ grep "total =" b4500875.hprof.out
CPU SAMPLES BEGIN (total = 2) Thu Apr 25 13:30:03 2002
$
$ sh doit.sh ~/hotspot/jdk1.4.1/linux -server
java version "1.4.1-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.1-beta-b10)
Java HotSpot(TM) Server VM (build 1.4.1-beta-b10, mixed mode)
Executing the test ...
Dumping CPU usage by sampling running threads ... done.
TEST FAILED: very few CPU sampling 0 taken by HPROF
expected that it should be more than 300
exit code = 97
$ grep "total =" b4500875.hprof.out
CPU SAMPLES BEGIN (total = 0) Thu Apr 25 13:31:39 2002
$
---------------------------------------------------------------------------
------------------- output on Linux with HS 1.4.1-b08: --------------------
$ sh doit.sh /export/ld59/java/dest/jdk1.4.1-b08/linux-i586
java version "1.4.1-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.1-beta-b08)
Java HotSpot(TM) Client VM (build 1.4.1-beta-b08, mixed mode)
Executing the test ...
Dumping CPU usage by sampling running threads ... done.
exit code = 95
$ grep "total =" b4500875.hprof.out
CPU SAMPLES BEGIN (total = 2998) Thu Apr 25 13:01:27 2002
$
---------------------------------------------------------------------------
------------------- output on Linux with HS 1.4.1-b92: --------------------
$ sh doit.sh ~/hotspot/jdk1.4/linux
java version "1.4.0"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.0-b92)
Java HotSpot(TM) Client VM (build 1.4.0-b92, mixed mode)
Executing the test ...
Dumping CPU usage by sampling running threads ... done.
exit code = 95
$ grep "total =" b4500875.hprof.out
CPU SAMPLES BEGIN (total = 862) Thu Apr 25 12:57:29 2002
$
---------------------------------------------------------------------------
======================================================================
- duplicates
-
JDK-4678415 Reg-test MangleStepTest.java Failing
- Closed
- relates to
-
JDK-4500875 Very few CPU sampling due to failure in wait_for_suspend_completion
- Closed
-
JDK-4683006 3rd party memory managers risk deadlock with JVMDI/JVMPI/JVM_Suspend
- Closed