Mailing List Archive

I opened https://bugs.openjdk.org/browse/JDK-8289127

Uwe

Am 24.06.2022 um 10:48 schrieb Uwe Schindler:
>
> This looks very stange:
>
> https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/artifact/lucene/join/build/tmp/tests-cwd/hs_err_pid3260852.log
>
> #
> # A fatal error has been detected by the Java Runtime Environment:
> #
> # Internal Error (macroAssembler_x86.cpp:845), pid=3260852, tid=3260993
> # fatal error: DEBUG MESSAGE: duplicated predicate failed which is impossible
> #
> # JRE version: OpenJDK Runtime Environment (19.0+27) (build 19-ea+27-2074)
> # Java VM: OpenJDK 64-Bit Server VM (19-ea+27-2074, mixed mode, sharing, tiered, compressed class ptrs, serial gc, linux-amd64)
> # Problematic frame:
> # V [libjvm.so+0xb6e5b1] MacroAssembler::debug64(char*, long, long*)+0x41
> #
>
> Not sure if I should open a bug report. Looks like a debug assettion,
> which should not happen in production code. I am quite sure that this
> is not a debug build!
>
> Uwe
>
>
> Am 24.06.2022 um 05:19 schrieb Policeman Jenkins Server:
>> Build:https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/
>> Java: 64bit/jdk-19-ea+27 -XX:-UseCompressedOops -XX:+UseSerialGC
>>
>> All tests passed
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail:builds-unsubscribe@lucene.apache.org
>> For additional commands, e-mail:builds-help@lucene.apache.org
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail:uwe@thetaphi.de

--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de

Hi,

I was now able to reproduce it locally with 100% probability, see
<https://bugs.openjdk.org/browse/JDK-8289127?focusedCommentId=14506741&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14506741>

$ cd /path/to/lucene-checkout $ export JAVA_HOME=/path/to/jdk17 $ export
RUNTIME_JAVA_HOME=/path/to/jdk19-ea+27

$ ./gradlew :lucene:join:beast -Ptests.multiplier=6 -Ptests.dups=100
-Ptests.jvmargs="-XX:CompileCommand=DumpReplay,org.apache.lucene.codecs.lucene90.ForUtil::decode14"
--tests TestBlockJoin.testMultiChildQueriesOfDiffParentLevels

Uwe

Am 24.06.2022 um 11:13 schrieb Uwe Schindler:
>
> I opened https://bugs.openjdk.org/browse/JDK-8289127
>
> Uwe
>
> Am 24.06.2022 um 10:48 schrieb Uwe Schindler:
>>
>> This looks very stange:
>>
>> https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/artifact/lucene/join/build/tmp/tests-cwd/hs_err_pid3260852.log
>>
>> #
>> # A fatal error has been detected by the Java Runtime Environment:
>> #
>> # Internal Error (macroAssembler_x86.cpp:845), pid=3260852, tid=3260993
>> # fatal error: DEBUG MESSAGE: duplicated predicate failed which is impossible
>> #
>> # JRE version: OpenJDK Runtime Environment (19.0+27) (build 19-ea+27-2074)
>> # Java VM: OpenJDK 64-Bit Server VM (19-ea+27-2074, mixed mode, sharing, tiered, compressed class ptrs, serial gc, linux-amd64)
>> # Problematic frame:
>> # V [libjvm.so+0xb6e5b1] MacroAssembler::debug64(char*, long, long*)+0x41
>> #
>>
>> Not sure if I should open a bug report. Looks like a debug assettion,
>> which should not happen in production code. I am quite sure that this
>> is not a debug build!
>>
>> Uwe
>>
>>
>> Am 24.06.2022 um 05:19 schrieb Policeman Jenkins Server:
>>> Build:https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/
>>> Java: 64bit/jdk-19-ea+27 -XX:-UseCompressedOops -XX:+UseSerialGC
>>>
>>> All tests passed
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail:builds-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail:builds-help@lucene.apache.org
>> --
>> Uwe Schindler
>> Achterdiek 19, D-28357 Bremen
>> https://www.thetaphi.de
>> eMail:uwe@thetaphi.de
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail:uwe@thetaphi.de

--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de

Hi,

the bug was fixed by Roland Westrelin and a new EA build was installed
on Jenkins. The changes will also be backported to JDK 17. It may also
happen there, but it if not fully known under which circumstances
unrolling the ForUtil#decode14() happens. It was just obvoius since JDK
19 because of other optimizations in JDK 19/20 that unlocked the faulty
optimization for Lucene.

I will keep you informed if there is any action needed like requiring a
minimum JDK-17 version enforced by our startup checks in Solr/Elasticsearch.

It is unknown if the optimization bug also affects Java 11 to 16.

Uwe

Am 24.06.2022 um 11:13 schrieb Uwe Schindler:
>
> I opened https://bugs.openjdk.org/browse/JDK-8289127
>
> Uwe
>
> Am 24.06.2022 um 10:48 schrieb Uwe Schindler:
>>
>> This looks very stange:
>>
>> https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/artifact/lucene/join/build/tmp/tests-cwd/hs_err_pid3260852.log
>>
>> #
>> # A fatal error has been detected by the Java Runtime Environment:
>> #
>> # Internal Error (macroAssembler_x86.cpp:845), pid=3260852, tid=3260993
>> # fatal error: DEBUG MESSAGE: duplicated predicate failed which is impossible
>> #
>> # JRE version: OpenJDK Runtime Environment (19.0+27) (build 19-ea+27-2074)
>> # Java VM: OpenJDK 64-Bit Server VM (19-ea+27-2074, mixed mode, sharing, tiered, compressed class ptrs, serial gc, linux-amd64)
>> # Problematic frame:
>> # V [libjvm.so+0xb6e5b1] MacroAssembler::debug64(char*, long, long*)+0x41
>> #
>>
>> Not sure if I should open a bug report. Looks like a debug assettion,
>> which should not happen in production code. I am quite sure that this
>> is not a debug build!
>>
>> Uwe
>>
>>
>> Am 24.06.2022 um 05:19 schrieb Policeman Jenkins Server:
>>> Build:https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/
>>> Java: 64bit/jdk-19-ea+27 -XX:-UseCompressedOops -XX:+UseSerialGC
>>>
>>> All tests passed
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail:builds-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail:builds-help@lucene.apache.org
>> --
>> Uwe Schindler
>> Achterdiek 19, D-28357 Bremen
>> https://www.thetaphi.de
>> eMail:uwe@thetaphi.de
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail:uwe@thetaphi.de

--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de

Hi,

For all who are interested: This is the fix:

https://github.com/openjdk/jdk/commit/4f3f74c14121d0a80f0dcf1d593b4cf1c3e4a64c

The most funny thing is the test case added by Roland. It is just the
faulty method ForUtil#decode14 running in a loop with empty (0L) arrays.
I don't think the license change to GPL is an issue because the copied
code in Lucene was autogenerated! Or should we care?

Uwe

Am 22.07.2022 um 10:09 schrieb Uwe Schindler:
>
> Hi,
>
> the bug was fixed by Roland Westrelin and a new EA build was installed
> on Jenkins. The changes will also be backported to JDK 17. It may also
> happen there, but it if not fully known under which circumstances
> unrolling the ForUtil#decode14() happens. It was just obvoius since
> JDK 19 because of other optimizations in JDK 19/20 that unlocked the
> faulty optimization for Lucene.
>
> I will keep you informed if there is any action needed like requiring
> a minimum JDK-17 version enforced by our startup checks in
> Solr/Elasticsearch.
>
> It is unknown if the optimization bug also affects Java 11 to 16.
>
> Uwe
>
> Am 24.06.2022 um 11:13 schrieb Uwe Schindler:
>>
>> I opened https://bugs.openjdk.org/browse/JDK-8289127
>>
>> Uwe
>>
>> Am 24.06.2022 um 10:48 schrieb Uwe Schindler:
>>>
>>> This looks very stange:
>>>
>>> https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/artifact/lucene/join/build/tmp/tests-cwd/hs_err_pid3260852.log
>>>
>>> #
>>> # A fatal error has been detected by the Java Runtime Environment:
>>> #
>>> # Internal Error (macroAssembler_x86.cpp:845), pid=3260852, tid=3260993
>>> # fatal error: DEBUG MESSAGE: duplicated predicate failed which is impossible
>>> #
>>> # JRE version: OpenJDK Runtime Environment (19.0+27) (build 19-ea+27-2074)
>>> # Java VM: OpenJDK 64-Bit Server VM (19-ea+27-2074, mixed mode, sharing, tiered, compressed class ptrs, serial gc, linux-amd64)
>>> # Problematic frame:
>>> # V [libjvm.so+0xb6e5b1] MacroAssembler::debug64(char*, long, long*)+0x41
>>> #
>>>
>>> Not sure if I should open a bug report. Looks like a debug
>>> assettion, which should not happen in production code. I am quite
>>> sure that this is not a debug build!
>>>
>>> Uwe
>>>
>>>
>>> Am 24.06.2022 um 05:19 schrieb Policeman Jenkins Server:
>>>> Build:https://jenkins.thetaphi.de/job/Lucene-MMAPv2-Linux/114/
>>>> Java: 64bit/jdk-19-ea+27 -XX:-UseCompressedOops -XX:+UseSerialGC
>>>>
>>>> All tests passed
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail:builds-unsubscribe@lucene.apache.org
>>>> For additional commands, e-mail:builds-help@lucene.apache.org
>>> --
>>> Uwe Schindler
>>> Achterdiek 19, D-28357 Bremen
>>> https://www.thetaphi.de
>>> eMail:uwe@thetaphi.de
>> --
>> Uwe Schindler
>> Achterdiek 19, D-28357 Bremen
>> https://www.thetaphi.de
>> eMail:uwe@thetaphi.de
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail:uwe@thetaphi.de

--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de