Bug 478248 - crashing after opened a FileDialog or DirectoryDialog in Ubuntu14.04
Summary: crashing after opened a FileDialog or DirectoryDialog in Ubuntu14.04
Status: RESOLVED WORKSFORME
Alias: None
Product: Platform
Classification: Eclipse Project
Component: SWT (show other bugs)
Version: 3.7.2   Edit
Hardware: PC Linux
: P3 normal with 2 votes (vote)
Target Milestone: ---   Edit
Assignee: Platform-SWT-Inbox CLA
QA Contact:
URL:
Whiteboard:
Keywords: needinfo
Depends on:
Blocks:
 
Reported: 2015-09-23 23:26 EDT by Sheng Jun Wang CLA
Modified: 2018-03-23 09:49 EDT (History)
7 users (show)

See Also:


Attachments
The test code that could reproduce the crash (902 bytes, text/plain)
2015-10-13 03:50 EDT, df jk CLA
no flags Details
javacore file from the test code (94.78 KB, text/plain)
2015-11-02 03:48 EST, Sheng Jun Wang CLA
no flags Details
Snap file from the test code (57.12 KB, application/octet-stream)
2015-11-02 03:48 EST, Sheng Jun Wang CLA
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Sheng Jun Wang CLA 2015-09-23 23:26:32 EDT
After opened a FileDialog or DirectoryDialog in Ubuntu14.04, Notes client will hang for a while then crash and quit Notes.
Tested in Notes901FP5.
Comment 1 Sravan Kumar Lakkimsetti CLA 2015-09-24 04:59:34 EDT
Please attach any existing crash logs to this bug. Please share if you have any reproducible snippet or scenario on base eclipse sdk.
Comment 2 Eric Williams CLA 2015-09-24 08:44:08 EDT
(In reply to Sravan Kumar Lakkimsetti from comment #1)
> Please attach any existing crash logs to this bug. Please share if you have
> any reproducible snippet or scenario on base eclipse sdk.

In addition to this, can you please provide:
1) the version of Gtk2 and Gtk3 on your system, and
2) what version of Gtk (2 or 3) Eclipse is running on.
Comment 3 Arun Thondapu CLA 2015-09-24 09:28:00 EDT
(In reply to Eric Williams from comment #2)
> (In reply to Sravan Kumar Lakkimsetti from comment #1)
> > Please attach any existing crash logs to this bug. Please share if you have
> > any reproducible snippet or scenario on base eclipse sdk.
> 
> In addition to this, can you please provide:
> 1) the version of Gtk2 and Gtk3 on your system, and
> 2) what version of Gtk (2 or 3) Eclipse is running on.

Since it is Eclipse 3.7.2, it would definitely be gtk2 (should be 2.24.x on Ubuntu 14.04).
Comment 4 Sheng Jun Wang CLA 2015-09-25 04:28:37 EDT
As my test result, this issue only happens on Notes9.0.1 client. I didn't reproduce it with a sample based on SWT API. So I suspect it's a bug of Notes client.
Comment 5 Eclipse Genie CLA 2015-09-29 09:36:57 EDT
New Gerrit change created: https://git.eclipse.org/r/56950
Comment 6 Arun Thondapu CLA 2015-09-30 04:23:23 EDT
(In reply to Eclipse Genie from comment #5)
> New Gerrit change created: https://git.eclipse.org/r/56950

This seems to be a mistake, the gerrit review has nothing to do with this bug...
Comment 7 df jk CLA 2015-10-13 03:50:51 EDT
Created attachment 257232 [details]
The test code that could reproduce the crash
Comment 8 df jk CLA 2015-10-13 03:52:12 EDT
Comment on attachment 257232 [details]
The test code that could reproduce the crash

with Eclipse 372 on Ubuntu 14.04
Comment 9 df jk CLA 2015-10-16 04:57:17 EDT
With org.eclipse.swt.gtk.linux.x86_3.7.2.v3740f-RCP20150506-1514.jar
not sure if this is useful or not.
Comment 10 Eric Williams CLA 2015-10-22 14:59:45 EDT
I can't reproduce this on my system with:

Eclipse Neon Nightly 20151001-2000
Gtk3.16.6
Fedora 22
Comment 11 Sheng Jun Wang CLA 2015-10-29 00:23:00 EDT
The newest update from tester:

Notes team says this crash could resolved by later update of Ubuntu 14.04
I just installed all the software updates for Ubuntu 14.04 (and restart machine), the crash from FileDialog in such many places are GONE.
Comment 12 Sheng Jun Wang CLA 2015-11-02 03:46:40 EST
Sorry, with another machine (Ubuntu 14.04) reboot, we see crash again, both in IBM Notes client and the test code.

The console output of the test code is:

Unhandled exception
Type=Segmentation error vmState=0x00040000
J9Generic_Signal_Number=00000004 Signal_Number=0000000b Error_Value=00000000 Signal_Code=00000080
Handler1=B742EE81 Handler2=B7403216 InaccessibleAddress=00000000
EDI=00000000 ESI=8F2258C0 EAX=FF9A9EAC EBX=93D6C000
ECX=8F27D510 EDX=00000010
EIP=93D51926 ES=007B DS=007B ESP=B7394548
EFlags=00010286 CS=0073 SS=007B EBP=00000000
Module=/usr/lib/i386-linux-gnu/libpixman-1.so.0
Module_base_address=93CC3000
Target=2_40_20150701_255724 (Linux 3.16.0-52-generic)
CPU=x86 (2 logical CPUs) (0x77bde000 RAM)
----------- Stack Backtrace -----------
(0xB73F6246 [libj9prt24.so+0xc246])
(0xB74027D7 [libj9prt24.so+0x187d7])
(0xB73F62D9 [libj9prt24.so+0xc2d9])
(0xB73F63FC [libj9prt24.so+0xc3fc])
(0xB73F6078 [libj9prt24.so+0xc078])
(0xB74027D7 [libj9prt24.so+0x187d7])
(0xB73F60EC [libj9prt24.so+0xc0ec])
(0xB742FC4D [libj9vm24.so+0xfc4d])
(0xB742F8CD [libj9vm24.so+0xf8cd])
(0xB74027D7 [libj9prt24.so+0x187d7])
(0xB742F0B2 [libj9vm24.so+0xf0b2])
(0xB742EF05 [libj9vm24.so+0xef05])
(0xB740338E [libj9prt24.so+0x1938e])
(0xB76EDD38)
---------------------------------------
JVMDUMP039I Processing dump event "gpf", detail "" at 2015/10/29 22:30:43 - please wait.
JVMDUMP032I JVM requested System dump using '/media/luomin/LIWENJIE/core.20151029.223043.5088.0001.dmp' in response to an event
JVMPORT030W /proc/sys/kernel/core_pattern setting "|/usr/share/apport/apport %p %s %c %P" specifies that the core dump is to be piped to an external program.  Attempting to rename either core or core.5123.

JVMDUMP010I System dump written to /media/luomin/LIWENJIE/core.20151029.223043.5088.0001.dmp
JVMDUMP032I JVM requested Java dump using '/media/luomin/LIWENJIE/javacore.20151029.223043.5088.0002.txt' in response to an event
JVMDUMP010I Java dump written to /media/luomin/LIWENJIE/javacore.20151029.223043.5088.0002.txt
JVMDUMP032I JVM requested Snap dump using '/media/luomin/LIWENJIE/Snap.20151029.223043.5088.0003.trc' in response to an event
JVMDUMP010I Snap dump written to /media/luomin/LIWENJIE/Snap.20151029.223043.5088.0003.trc
JVMDUMP013I Processed dump event "gpf", detail "".

I only attach the javacore*.txt and Snap*.rtc file, but not attach the too large core*.dump file
Comment 13 Sheng Jun Wang CLA 2015-11-02 03:48:00 EST
Created attachment 257674 [details]
javacore file from the test code
Comment 14 Sheng Jun Wang CLA 2015-11-02 03:48:52 EST
Created attachment 257675 [details]
Snap file from the test code
Comment 15 Sravan Kumar Lakkimsetti CLA 2015-11-05 01:48:12 EST
I tried this on Ubuntu 14.04(fully patched) I tried it on 32-bit and 64-bit with eclipse 3.7.2

I am not able to reproduce the problem.
Comment 16 Yu Wang CLA 2015-11-17 04:17:09 EST
Better install Notes 9.0.1 (we were using its FP4).

Not need to reproduce within Notes, just run the Test.java with java & swt from Notes.
Notes java is in "/opt/IBM/Notes/jvm", swt is in "/opt/IBM/Notes/framework/rcp/eclipse/plugins/org.eclipse.swt.**.jar".
  "path/to/java -cp path/to/swt.jar:path/of/compile com.lwjlq.Test".
Click the button in the dialog, crash.
Comment 17 Yu Wang CLA 2015-11-17 04:20:36 EST
If the problem can not be reproduced for the first time, please restart the Ubuntu system and retry. The tester reported they met the crash again after rebooting the system.
Comment 18 Sheng Jun Wang CLA 2015-12-17 19:38:06 EST
This problem doesn't occur in Oracle JRE. This indicates problem with IBM JRE.
When using IBM JRE, Crash only happens when clicking the home directory in File or Directory dialog. No crashing if clicking other directories except home directory.
Comment 19 Pallavi Singh CLA 2016-01-13 00:19:11 EST
Please provide an ETA for this defect. Thanks.
Comment 20 Arun Thondapu CLA 2016-01-14 07:31:31 EST
(In reply to Pallavi Singh from comment #19)
> Please provide an ETA for this defect. Thanks.

There is no ETA for a fix from Eclipse, the problem is specific to the IBM JRE and is expected to be fixed in the JRE itself. Please try with one of the latest Java 8 builds to confirm.
Comment 21 Sheng Jun Wang CLA 2016-01-27 00:45:49 EST
The update from IBM Java L3:

I could find an eclipse bug id # 178169 where similar crash issue was reported on clicking Home folder.
 
https://bugs.eclipse.org/bugs/show_bug.cgi?id=178169
 
The crashing stack for this pmr and the crashing stack for the eclipse bug reported are exactly the same.The java code is not involved in the crashing stack so fix needs to be provided by eclipse team.
Comment 22 Arun Thondapu CLA 2016-02-01 13:49:41 EST
(In reply to Sheng Jun Wang from comment #21)
> The update from IBM Java L3:
> 
> I could find an eclipse bug id # 178169 where similar crash issue was
> reported on clicking Home folder.
>  
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=178169
>  
> The crashing stack for this pmr and the crashing stack for the eclipse bug
> reported are exactly the same.The java code is not involved in the crashing
> stack so fix needs to be provided by eclipse team.

That bug is unrelated to this one, even though the stack traces look similar, the actual crash happens in completely different native modules. In fact, we cannot really go by that logic because most of the native crashes would have similar Java stack traces even though the actual root causes are different.

Coming back to this particular bug again, there definitely is something amiss when using the IBM JRE and the problem needs to be looked at more closely by the JVM team itself. If the root cause was in the eclipse libraries, it should have caused crashes with other JREs as well (Oracle/OpenJRE) which is not the case here.

One possibility which could be causing this crash is that there are some subtle interaction problems with the Java to native bridge while invoking the native libraries. We've recently seen a similar problem with 32-bit eclipse on RHEL when using the IBM JRE and it was figured that there were some memory alignment issues between the JRE and the OS native libraries which was causing crashes.

IIRC it was supposed to have been fixed in recent builds of IBM JRE 8 which is why I suggested trying with the latest JRE builds. I suspect this crash is also something similar to the one observed on RHEL with IBM JRE.

You could also probably try turning off the J9 mode of the IBM JVM to test the behavior, though I'm not too sure if it is possible or not.

In any case, unless we have a reproducible test case which crashes with Oracle JRE as well, we may not be able to do much from our side.

Thanks!
Comment 23 Eric Williams CLA 2018-03-23 09:49:02 EDT
Ubuntu 14.04 is no longer a supported platform. Please open a new bug with steps to reproduce the issue, should it occur again.