JDK-8077314 : Assertion fails on reference in free region in SATB buffer
  • Type: Bug
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 9
  • Priority: P2
  • Status: Closed
  • Resolution: Duplicate
  • Submitted: 2015-04-09
  • Updated: 2015-04-09
  • Resolved: 2015-04-09
Related Reports
Duplicate :  
Relates :  
Description
Failed on Windows in GC nightlies.

# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (\hotspot\src\share\vm\gc_implementation\g1\concurrentMark.cpp:3811), pid=4308, tid=11288
#  assert(!_g1h->is_on_master_free_list( _g1h->heap_region_containing((HeapWord*) obj))) failed: invariant
#
# JRE version: Java(TM) SE Runtime Environment (9.0) (build 1.9.0-internal-fastdebug-20150407085837.tschatzl.8058354-b00)
# Java VM: Java HotSpot(TM) Client VM (1.9.0-internal-20150407085837.tschatzl.8058354-b00 mixed mode, sharing windows-x86 )


Stack: [0x044d0000,0x04520000],  sp=0x0451efb0,  free space=315k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [jvm.dll+0x2b4535]  VMError::report_and_die+0x3d5;;  ?report_and_die@VMError@@QAEXXZ+0x3d5
V  [jvm.dll+0x2a7ed5]  report_vm_error+0x45;;  ?report_vm_error@@YAXPBDH00@Z+0x45
V  [jvm.dll+0x3d93ef]  CMTask::drain_local_queue+0x19f;;  ?drain_local_queue@CMTask@@QAEX_N@Z+0x19f
V  [jvm.dll+0x3dc8a3]  CMTask::do_marking_step+0x763;;  ?do_marking_step@CMTask@@QAEXN_N0@Z+0x763
V  [jvm.dll+0x3dcf93]  CMConcurrentMarkingTask::work+0x153;;  ?work@CMConcurrentMarkingTask@@UAEXI@Z+0x153
V  [jvm.dll+0x2b51dc]  GangWorker::loop+0x2cc;;  ?loop@GangWorker@@MAEXXZ+0x2cc
V  [jvm.dll+0x2d9e7c]  java_start+0xcc;;  ?java_start@@YGIPAVThread@@@Z+0xcc
C  [msvcr100.dll+0x5c556]
C  [msvcr100.dll+0x5c600]
C  [kernel32.dll+0x4d3c9]
C  [ntdll.dll+0x41603]
C  [ntdll.dll+0x415d6]



Comments
This is a sort-of duplicate of JDK-8069367, not JDK-8075215. Unlike suggested before, this is not about references in the SATB buffers, but on the local mark stack. Since JDK-8069367 is going to handle this, this is a "better" duplicate.
09-04-2015

Looks similar to JDK-8075215 where a reference to a freed region/eagerly reclaimed humongous object is found in the SATB buffer. Too strict assertion results in this benign assert.
09-04-2015