JDK-6681372 : 64-bit VM CompilerThread received SEGV in ciObjectFactory::find_non_perm()
  • Type: Bug
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 6u4,6u14
  • Priority: P3
  • Status: Closed
  • Resolution: Cannot Reproduce
  • OS: linux_redhat_4.0,linux_redhat_5.0
  • CPU: x86
  • Submitted: 2008-03-28
  • Updated: 2019-05-08
  • Resolved: 2010-11-30
Related Reports
Relates :  
Relates :  
Relates :  
Relates :  
64-bit VM CompilerThread crashed in ciObjectFactory::find_non_perm().

OS: Red Hat Enterprise Linux AS release 4 (Nahant Update 6)
vm_info: Java HotSpot(TM) 64-Bit Server VM (10.0-b19) for linux-amd64 JRE (1.6.0_04-b12)

Please find the hs_err_pid.log file attached.
Here is the stack trace (mangled and unmangled):

Stack: [0x000000004123c000,0x000000004133d000],  sp=0x000000004133a7b0,  free space=1017k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0x2099b4] _ZN15ciObjectFactory13find_non_permEP7oopDesc+0x34
V  [libjvm.so+0x20898b] _ZN15ciObjectFactory3getEP7oopDesc+0x8b
V  [libjvm.so+0x205f08] _ZN18ciReceiverTypeData28translate_receiver_data_fromEP11ProfileData+0x98
V  [libjvm.so+0x206a49] _ZN18ciReceiverTypeData14translate_fromEP11ProfileData+0x9
V  [libjvm.so+0x205e1f] _ZN12ciMethodData9load_dataEv+0x1af
V  [libjvm.so+0x201473] _ZN8ciMethod17build_method_dataE12methodHandle+0xb3
V  [libjvm.so+0x2015fe] _ZN8ciMethod17build_method_dataEv+0x11e
V  [libjvm.so+0x240b6d] _ZN7CompileC1EP5ciEnvP10C2CompilerP8ciMethodib+0x2fd
V  [libjvm.so+0x1e05c7] _ZN10C2Compiler14compile_methodEP5ciEnvP8ciMethodi+0x67
V  [libjvm.so+0x248ec8] _ZN13CompileBroker25invoke_compiler_on_methodEP11CompileTask+0x298
V  [libjvm.so+0x248866] _ZN13CompileBroker20compiler_thread_loopEv+0x306
V  [libjvm.so+0x62a3f9] _Z21compiler_thread_entryP10JavaThreadP6Thread+0x9
V  [libjvm.so+0x6246a1] _ZN10JavaThread3runEv+0x111
V  [libjvm.so+0x505eea] _Z10java_startP6Thread+0x14a
C  [libpthread.so.0+0x6137]
C  0x0000000000000000

V [libjvm.so+0x2099b4] ciObjectFactory::find_non_perm(oopDesc*)+0x34
V [libjvm.so+0x20898b] ciObjectFactory::get(oopDesc*)+0x8b
V [libjvm.so+0x205f08] ciReceiverTypeData::translate_receiver_data_from(ProfileData*)+0x98
V [libjvm.so+0x206a49] ciReceiverTypeData::translate_from(ProfileData*)+0x9
V [libjvm.so+0x205e1f] ciMethodData::load_data()+0x1af
V [libjvm.so+0x201473] ciMethod::build_method_data(methodHandle)+0xb3
V [libjvm.so+0x2015fe] ciMethod::build_method_data()+0x11e
V [libjvm.so+0x240b6d] Compile::Compile(ciEnv*, C2Compiler*, ciMethod*, int, bool)+0x2fd
V [libjvm.so+0x1e05c7] C2Compiler::compile_method(ciEnv*, ciMethod*, int)+0x67
V [libjvm.so+0x248ec8] CompileBroker::invoke_compiler_on_method(CompileTask*)+0x298
V [libjvm.so+0x248866] CompileBroker::compiler_thread_loop()+0x306
V [libjvm.so+0x62a3f9] compiler_thread_entry(JavaThread*, Thread*)+0x9
V [libjvm.so+0x6246a1] JavaThread::run()+0x111
V [libjvm.so+0x505eea] java_start(Thread*)+0x14a

EVALUATION Closing as "Not Reproducible"; looks like a case of a trail gone cold, perhaps also at the customer. If issue is still present, I advise opening a nice, fresh, new bug, since this one is by now old, stale and dead.

WORK AROUND Tyr -XX:-UseCompressedOops?


EVALUATION Without further data from the customer, it appears difficult to proceed further. As such we are closing this bug as "not reproducible". We suggest that the customer run with the fixes for 6722112, 6722113 and 6722116. If the issue is still reproducible after that, please reopen this bug and file an escalation along with complete command line options, the core file and associated libraries (a la pkgcore), and the hs_err_<pid> file.

EVALUATION This looks similar to other CMS induced crashes when accessing the perm gen.