JDK-8028313 : SEGV during java.net.SocketInputStream.socketread0 on mac osx
  • Type: Bug
  • Component: core-libs
  • Sub-Component: java.net
  • Affected Version: 7u40
  • Priority: P2
  • Status: Resolved
  • Resolution: Duplicate
  • OS: other
  • CPU: generic
  • Submitted: 2013-11-13
  • Updated: 2014-11-21
  • Resolved: 2014-03-04
Related Reports
Duplicate :  
JDK-8035897 - Better memory allocation for file descriptors greater than 1024 on macosx
Relates :  
JDK-8024045 - SEGV in java.net.SocketInputStream.socketRead0
Description
See BugDB # 17315456 for detail.

When running java version "1.7.0_17" on OS/X and running jmx enabled cache
server in active persistence mode, causes SIGSEGV in DefaultCacheServer when
connecting from jconsole.
.
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x0000000105dab34c, pid=3839, tid=93187
#
# JRE version: 7.0_17-b02
# Java VM: Java HotSpot(TM) 64-Bit Server VM (23.7-b01 mixed mode bsd-amd64
compressed oops)
# Problematic frame:
# C  [libjava.dylib+0x934c]  JNU_NewStringPlatform+0x1c8

 COUNTER INDICATORS:
 TRIGGERS:
 Attaching to a JMX process via jconsole will cause the crash. Attaching via JMC also
 shows the the coherence process aborting.
 KNOWN WORKAROUND:  N/A
 PRESENT SINCE:  Seems to be an issue in the jdk7u port of Mac OSX
 HOW TO VERIFY:
 NOTES FOR SE:  A testcase is available via BugDB report. I'm seeing some other issues while  trying to reproduce the SEGV :

java(417,0x19fdef000) malloc: *** error for object 0x19fdde430: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug
cache-server.sh: line 22:   417 Abort trap: 6           $JAVA_HOME/bin/java


There are similar reports of JNI network exception crash/issue already in JIRA.
Comments
Duplicate of JDK-8035897
04-03-2014
Michael, JDK-8024045 and JI-9005835 report similar crashes. There's an interesting comment at https://bugs.openjdk.java.net/browse/JDK-8024045?focusedCommentId=13427741&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13427741 However, I've no proof yet that this issue is a networking bug.
25-11-2013
Sean. Do you have a list of bugids that might be similar to this one? I don't see anything mac specific in the native code for SocketInputStream. Is it the case also, there is external native code associated with this app?
22-11-2013
Is there a patch for this? This may be something to get into 8 and let bake before thinking about 7u60.
13-11-2013