United StatesChange Country, Oracle Worldwide Web Sites Communities I am a... I want to...
Bug ID: JDK-7102300 compiler/5091921/Test6890943.java is incorrect
JDK-7102300 : compiler/5091921/Test6890943.java is incorrect

Details
Type:
Bug
Submit Date:
2011-10-18
Status:
Resolved
Updated Date:
2013-09-04
Project Name:
JDK
Resolved Date:
2013-02-19
Component:
hotspot
OS:
generic
Sub-Component:
compiler
CPU:
generic
Priority:
P4
Resolution:
Fixed
Affected Versions:
hs24,8
Fixed Versions:
hs25 (b20)

Related Reports
Backport:
Backport:
Backport:
Backport:

Sub Tasks

Description
This test invokes java program via script and then compares the output with gold result, if different, it fails. However, if a VM issues a warning, the test will report failure even actually there was no failure.

                                    

Comments
So next warning is expected behavior?

Java HotSpot(TM) Server VM warning: Performance bug: SystemDictionary lookup_count=19343 lookup_length=20430 average=1.056196 load=0.500496

We will not catch such cases if we filter such warnings. Should we instead fix VM to avoid these warnings?

                                     
2013-02-09
Note, I agree with adding -XX:+IgnoreUnrecognizedVMOptions  -XX:-PrintVMOptions. But I am concern about  -XX:-PrintWarnings.

                                     
2013-02-09
If this test will always tickle this performance warning then we at least need to exclude this warning from consideration.
                                     
2013-02-09
Got it. We will add 2 flags to the test and filter only THIS particular warning.
                                     
2013-02-09
URL:   http://hg.openjdk.java.net/hsx/hotspot-comp/hotspot/rev/a2bc322ca273
User:  kvn
Date:  2013-02-19 00:17:20 +0000

                                     
2013-02-19
URL:   http://hg.openjdk.java.net/hsx/hsx25/hotspot/rev/a2bc322ca273
User:  amurillo
Date:  2013-02-22 20:02:50 +0000

                                     
2013-02-22
Better to have the script filter out warning messages I think.
                                     
2013-02-06
EVALUATION

This test passes now.  I think we want it to start failing again if there are warnings from the VM.  I'm going to close this as a not a bug (anymore).
                                     
2012-08-02
Failing again.
See 
* run http://aurora.ru.oracle.com/functional/faces/RunDetails.xhtml?names=164276.JAVASE.PIT.VM-145&show-limit=2000&filter=
* log http://aurora-ds.ru.oracle.com:9500/runs/164276.JAVASE.PIT.VM-145/results/workDir/compiler/5091921/Test6890943.jtr
                                     
2013-02-05
EVALUATION

The code generating the warning:

src/share/vm/utilities/hashtable.cpp:

225 void BasicHashtable::verify_lookup_length(double load) {
226   if ((double)_lookup_length / (double)_lookup_count > load * 2.0) {
227     warning("Performance bug: SystemDictionary lookup_count=%d "
228             "lookup_length=%d average=%lf load=%f",
229             _lookup_count, _lookup_length,
230             (double) _lookup_length / _lookup_count, load);
231   }
232 }


Tests that do golden file comparisons need to be run
with the following options:

    -XX:+IgnoreUnrecognizedVMOptions
    -XX:-PrintVMOptions
    -XX:-PrintWarnings
                                     
2011-12-09



Hardware and Software, Engineered to Work Together