Bug 1862977

Summary: java 11 is crashing on armhfp
Product: [Fedora] Fedora Reporter: Nicolas Chauvet (kwizart) <kwizart>
Component: java-11-openjdkAssignee: jiri vanek <jvanek>
Status: CLOSED NOTABUG QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 33CC: ahughes, ashundi, jerboaa, jvanek, mike, pix, sgehwolf
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2020-08-17 21:05:19 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 1869030    
Bug Blocks: 245418    
Attachments:
Description Flags
koji crash log
none
crash for java --version none

Description Nicolas Chauvet (kwizart) 2020-08-03 11:33:31 UTC
Description of problem:
java failed to run with "java -version" on armhfp


Version-Release number of selected component (if applicable):
1:11.0.8.10-1.fc33  (last build with disabled lto as of today).

How reproducible:
always on a given hardware.

Steps to Reproduce:
1. run "java -version"

Actual results:
crash (see attachment)

Expected results:
no crash

Additional info:
This is a armhfp vm under aarch64 hw.
This is experienced in the context of a kodi build
See also http://koji.rpmfusion.org/koji/taskinfo?taskID=424238

Comment 1 Nicolas Chauvet (kwizart) 2020-08-03 11:34:00 UTC
Created attachment 1703282 [details]
koji crash log

Comment 2 Nicolas Chauvet (kwizart) 2020-08-03 20:35:12 UTC
Last known to work java build from kodi was 
java-1.8.0-openjdk-devel      armv7hl   1:1.8.0.262.b05-0.0.ea.fc33       build  7.4 M

Unfortunately, no attempt in-between since this earlier version.

Comment 3 Severin Gehwolf 2020-08-04 09:21:25 UTC
I'm confused, why is a rpmfusion build failure reported here (as a Fedora bug)? Is this reproducible outside rpmfusion's build system? The java-11-openjdk build in Fedora seems to be fine on armv7hl. the java-11-openjdk build itself is a bootcycle build which uses the just built JDK to build itself. So I'm doubtful java -version crashes. If it would, it would make the build fail.

https://koji.fedoraproject.org/koji/taskinfo?taskID=48134465 (passed on armv7hl)

Also, for rawhide we recently switched from JDK 8 to JDK 11 (as system JDK):
https://fedoraproject.org/wiki/Changes/Java11

Please tell us how to reproduce the bug in Fedora only otherwise, please report the bug to rpmfusion. Thanks!

Comment 4 Ben Cotton 2020-08-11 13:51:35 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 33 development cycle.
Changing version to 33.

Comment 5 Andrew John Hughes 2020-08-17 21:05:19 UTC
Closing as this does not seem to be a crash from the java-11-openjdk package.

As Severin says, java-11-openjdk has built with itself on armv7hl, so is known not to crash on 'java -version'.

Comment 6 Nicolas Chauvet (kwizart) 2020-08-21 11:22:44 UTC
Sorry for the lack of answer, I was in PTO and AFK


I've reproduced the issue on a native armv7 machine (jetson-tk1) running fedora 31 with official fedora kernel 5.7.15-100.fc31.armv7hl

mock -r fedora-rawhide-armhfp  install java-devel
mock -r fedora-rawhide-armhfp  shell
$ java --version

Crash with SIGSEGV
See the attached hs_err_pid5.log

Also I wonder if this information is current ?
https://fedoraproject.org/wiki/Java/StackTraces


As just seen on IRC, using mock -r fedora-rawhide-armhfp-rpmfusion_free  shell  --isolation=simple works as expected.
So I wonder if it's a side effect of https://github.com/systemd/systemd/pull/16739

Comment 7 Nicolas Chauvet (kwizart) 2020-08-21 11:24:47 UTC
Created attachment 1712160 [details]
crash for java --version

Comment 8 Nicolas Chauvet (kwizart) 2020-08-21 12:42:28 UTC
Related to #1869030