JDK-6215606 : CMS: JVM 1.4.2_06 crashes when testing NetBeans IDE 4.1 on Solaris sparc
  • Type: Bug
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 1.4.2_05,1.4.2_06
  • Priority: P4
  • Status: Closed
  • Resolution: Duplicate
  • OS: solaris_8,solaris_9
  • CPU: generic,sparc
  • Submitted: 2005-01-10
  • Updated: 2011-05-13
  • Resolved: 2006-01-18
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
Other
1.4.2Resolved
Related Reports
Duplicate :  
JDK-6263371 - SEGV in MarkFromRootsClosure::scanOopsInOop
Duplicate :  
JDK-6325837 - crash within sun.awt.font.GlyphList.discardData
Description
During automated tests execution on build of NetBeans IDE 4.1 we experienced several times JVM crash. It is run on Solaris 5.8 on sparc, JDK1.4.2_06. It might be connected to switch -J-XX:+UseConcMarkSweepGC which was added to NetBeans command line at time when it was started to fail. But we are not sure because JVM crashes occasinally. The log message is the following:

[ideTestRunner] Unexpected Signal : 10 occurred at PC=0xFE232F20
[ideTestRunner] Function=[Unknown. Nearest: JVM_IsSameClassPackage+0x880A0]
[ideTestRunner] Library=/opt/java/j2sdk1.4.2_06/jre/lib/sparc/client/libjvm.so


[ideTestRunner] Dynamic libraries:
[ideTestRunner] 0x10000 	/opt/java/j2sdk1.4.2_06/bin/java
[ideTestRunner] 0xff340000 	/usr/lib/libthread.so.1
[ideTestRunner] 0xff330000 	/usr/lib/libdl.so.1
[ideTestRunner] 0xff200000 	/usr/lib/libc.so.1
[ideTestRunner] 0xff380000 	/usr/platform/SUNW,Sun-Fire-480R/lib/libc_psr.so.1
[ideTestRunner] 0xfe000000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/client/libjvm.so
[ideTestRunner] 0xff2d0000 	/usr/lib/libCrun.so.1
[ideTestRunner] 0xff1d0000 	/usr/lib/libsocket.so.1
[ideTestRunner] 0xff100000 	/usr/lib/libnsl.so.1
[ideTestRunner] 0xff0d0000 	/usr/lib/libm.so.1
[ideTestRunner] 0xff0b0000 	/usr/lib/libsched.so.1
[ideTestRunner] 0xff090000 	/usr/lib/libmp.so.2
[ideTestRunner] 0xff060000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/native_threads/libhpi.so
[ideTestRunner] 0xfe7d0000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libverify.so
[ideTestRunner] 0xfe790000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libjava.so
[ideTestRunner] 0xff020000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libzip.so
[ideTestRunner] 0xfc2d0000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libnet.so
[ideTestRunner] 0xfc2b0000 	/tmp/netbeans/dev/trunkcb/nb_all/xtest/lib/lib.jnikill.solaris.sparc.so
[ideTestRunner] 0xe9900000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libawt.so
[ideTestRunner] 0xfc180000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libmlib_image.so
[ideTestRunner] 0xfb810000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/motif21/libmawt.so
[ideTestRunner] 0xe9180000 	/usr/dt/lib/libXm.so.4
[ideTestRunner] 0xfbb90000 	/usr/openwin/lib/libXt.so.4
[ideTestRunner] 0xfbad0000 	/usr/openwin/lib/libXext.so.0
[ideTestRunner] 0xfbab0000 	/usr/openwin/lib/libXtst.so.1
[ideTestRunner] 0xe9080000 	/usr/openwin/lib/libX11.so.4
[ideTestRunner] 0xfb9a0000 	/usr/openwin/lib/libdps.so.5
[ideTestRunner] 0xf97e0000 	/usr/openwin/lib/libSM.so.6
[ideTestRunner] 0xf97b0000 	/usr/openwin/lib/libICE.so.6
[ideTestRunner] 0xe8f80000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libfontmanager.so
[ideTestRunner] 0xe5990000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libnio.so
[ideTestRunner] 0xe5730000 	/usr/lib/librt.so.1
[ideTestRunner] 0xe5560000 	/usr/lib/libaio.so.1
[ideTestRunner] 0xe51e0000 	/usr/lib/libsendfile.so.1
[ideTestRunner] 0xe47c0000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libcmm.so
[ideTestRunner] 0xe4770000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libjpeg.so
[ideTestRunner] 0xe4730000 	/opt/java/j2sdk1.4.2_06/jre/lib/sparc/libdcpr.so

[ideTestRunner] Heap at VM Abort:
[ideTestRunner] Heap
[ideTestRunner]  par new generation   total 4032K, used 1264K [0xed400000, 0xed800000, 0xedc00000)
[ideTestRunner]   eden space 3968K,  31% used [0xed400000, 0xed53c2c8, 0xed7e0000)
[ideTestRunner]   from space 64K,   0% used [0xed7f0000, 0xed7f0000, 0xed800000)
[ideTestRunner]   to   space 64K,   0% used [0xed7e0000, 0xed7e0000, 0xed7f0000)
[ideTestRunner]  concurrent mark-sweep generation total 50664K, used 34686K [0xedc00000, 0xf0d7a000, 0xf5400000)
[ideTestRunner]  concurrent-mark-sweep perm gen total 61240K, used 46225K [0xf5400000, 0xf8fce000, 0xf9400000)

[ideTestRunner] Local Time = Thu Jan  6 23:41:23 2005
[ideTestRunner] Elapsed Time = 196
[ideTestRunner] #
[ideTestRunner] # HotSpot Virtual Machine Error : 10
[ideTestRunner] # Error ID : 4F530E43505002EF 01
[ideTestRunner] # Please report this error at
[ideTestRunner] # http://java.sun.com/cgi-bin/bugreport.cgi
[ideTestRunner] #
[ideTestRunner] # Java VM: Java HotSpot(TM) Client VM (1.4.2_06-b03 mixed mode)
[ideTestRunner] #
[ideTestRunner] # An error report file has been saved as hs_err_pid21435.log.
[ideTestRunner] # Please refer to the file for further information.
[ideTestRunner] #
[ideTestRunner] my pid = 21435
[ideTestRunner] Abort - core dumped
[ideTestRunner] Result: 134

###@###.### 2005-1-10 10:45:06 GMT
Comments
EVALUATION The customer reported (in the comments section) that the problem is specific to 1.4.2-XX where they are not using this collector. (They report that in Tiger they are having good success using this collector; I assume that means that the crash has not been reproduced using Tiger. Please correct if that is not true.) Based on this, I'd like to downgrade this bug. Please escalate if you need any fixes for bugs in 1.4.2_XX.
29-08-2005
EVALUATION It is neccessary to know if this works with Tiger. According to stack attached, the error in a native method sun.awt.font.GlyphList.discardData. 2D guis should know more about it. ###@###.### 2005-1-14 08:36:38 GMT =================== I agree - please try to reproduce this on tiger. The native method where this crashes does not exist in 1.5 - the code is substantially different. If such a bug can be reproduced in 1.5 that strongly suggests the problem isn't in libfontmanger. Also it seems highly unlikely that its a concurrency problem in this code in 1.4.x - every single access to all native font code and font rendering is locked on a single global lock. There should be no way that two threads can be creating an inconsistent state. That architecture and its details have been in place since JDK 1.2 so it would be surprising if this were the first discovery of such a problem. We know there is at least one other MT-access bug in 2D which could cause crashes but I have no evidence that nb does the kind of MT access to a Graphics which could trigger it. I think there's little point investigating this from a 2D perspective unless it can be reproduced on 1.5 update 2 build 05 - which includes a fix for that MT-access crash (which is unrelated to libfontmanager) ###@###.### 2005-1-14 20:02:19 GMT ========================================= The two core dumps and log files just attached seem to point to a VM problem. ###@###.### 2005-1-19 18:08:02 GMT
14-01-2005