Regression in IcedTea6 JTreg Results - hotspot

[Omair Majid]

Mark Wielaard wrote:
> Hi Omair,
> 
> On Tue, 2009-03-03 at 23:55 -0500, IcedTea6 JTreg wrote:
>> Recent commiters:
>> Mark Wielaard <mark at klomp.org>
>> Andrew John Hughes <ahughes at redhat.com>
>> Gary Benson <gbenson at redhat.com>
>>
>> Newly failing tests:
>> compiler/6775880/Test.java
> 
> And it is correct. I am the "guilty party". It would probably help to
> have the full hg description of the changes with the above list. And
> maybe add the suspects to the CC so they have a chance to defend
> themselves :)
> 
> So, the full failure is:
> 
> #section:main
> ----------messages:(3/353)----------
> command: main  -server -Xbatch -XX:+DoEscapeAnalysis -XX:+DeoptimizeALot -XX:CompileCommand=exclude,java.lang.AbstractStringBuilder::appendTest
> reason: User specified action: run main/othervm -server -Xbatch -XX:+DoEscapeAnalysis -XX:+DeoptimizeALot -XX:CompileCommand=exclude,java.lang.AbstractStringBuilder::append Test 
> elapsed time (seconds): 0.059
> ----------System.out:(0/0)----------
> ----------System.err:(2/84)----------
> Unrecognized VM option '+DeoptimizeALot'
> Could not create the Java virtual machine.
> result: Failed. Unexpected exit from test [exit code: 1]
> 
> This comes from:
> 
> changeset:   1412:f03bf9944657
> user:        Mark Wielaard <mark at klomp.org>
> date:        Mon Mar 02 22:19:24 2009 +0100
> files:       ChangeLog hotspot.map
> description:
> * hotspot.map (default): Set to rev fc6a5ae3fef5 (14.0-b10).
> 
> This has been "fixed" in HS15 with the following patch:
> 
> changeset:   564:323728917cf4
> user:        kvn
> date:        Thu Feb 05 13:38:52 2009 -0800
> files:       src/share/vm/runtime/arguments.cpp
> src/share/vm/runtime/globals.hpp
>  test/compiler/6775880/Test.java
> description:
> 6788376: allow to ignore unrecognized VM options
> Summary: Add new product flag -XX:+IgnoreUnrecognizedVMOptions
> Reviewed-by: ysr, xlu
> 
> Will backport.
> 
> Still looking where exactly the +DeoptimizeALot flag disappeared to,
> cannot immediately pinpoint the exact hotspot patch.

Looks like this option exists, but is only enabled in a debug build:
$ openjdk/control/build/linux-i586-debug/j2sdk-image/jre/bin/java
-XX:+DeoptimizeALot -version
VM option '+DeoptimizeALot'
Warning:  Cannot open log file: hotspot.log
Warning:  Forcing option -XX:LogFile=/tmp//hs_pid24007.log
java version "1.6.0_0-debug"
OpenJDK Runtime Environment (IcedTea6 1.5pre-r953caf041bef) (build
1.6.0_0-debug-b14)
OpenJDK Client VM (build 14.0-b10-jvmg, mixed mode)

$ openjdk/control/build/linux-i586/j2sdk-image/jre/bin/java
-XX:+DeoptimizeALot -version
Unrecognized VM option '+DeoptimizeALot'
Could not create the Java virtual machine.

This test was written to check for a bug in the debug build of the JVM.
I dont think it makes a lot of sense to run it against the production
build every night. Any thoughts?

> 
> Thanks,
> 
> Mark
> 

Cheers,
Omair