RFR (XS): JDK-8023145: G1: G1CollectedHeap::mark_strong_code_roots() needs to handle ParallelGCThreads=0

[Bengt Rutisson]

Hi everyone,

Could I have a couple of reviews for this small change?
http://cr.openjdk.java.net/~brutisso/8023145/webrev.00/

 From the bug report:

With the fix for JDK-7145569 the method 
G1CollectedHeap::mark_strong_code_roots() was introduced. However, this 
method assumes that the work gang has been set up. If we are running 
with ParallelGCThreads=0 we don't set up the work gang and thus get a 
null pointer:

This simple JTreg test illustrates the issue:


/**
  * @test Test8000311
  * @key gc
  * @bug 8000311
  * @summary G1: ParallelGCThreads==0 broken
  * @run main/othervm -XX:+UseG1GC -XX:ParallelGCThreads=0 
-XX:+ResizePLAB -XX:+ExplicitGCInvokesConcurrent Test8000311
  * @author filipp.zhinkin at oracle.com
  */

import java.util.*;

public class Test8000311 {
   public static void main(String args[]) {
     for(int i = 0; i<100; i++) {
       byte[] garbage = new byte[1000];
       System.gc();
     }
   }
}

It crashes with a null pointer:

# SIGSEGV (0xb) at pc=0xffffffff7d3fadf4, pid=22622, tid=6

si_addr=0x0000000000000000

Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, 
C=native code)
V [libjvm.so+0x7fadf4] void 
G1CollectedHeap::mark_strong_code_roots(unsigned)+0x44;; 
__1cPG1CollectedHeapWmark_strong_code_roots6MI_v_+0x44
V [libjvm.so+0x7f5f80] void 
G1CollectedHeap::g1_process_strong_roots(bool,SharedHeap::ScanningOption,OopClosure*,OopsInHeapRegionClosure*,G1KlassScanClosure*,int)+0x5e0;; 
__1cPG1CollectedHeapXg1_process_strong_roots6MbnKSharedHeapOScanningOption_pnKOopClosure_pnXOopsInHeapRegionClosure_pnSG1KlassScanClosure_i_v_+0x5e0
V [libjvm.so+0x7fccdc] void G1ParTask::work(unsigned)+0x67c;; 
__1cJG1ParTaskEwork6MI_v_+0x67c
V [libjvm.so+0x7f7500] void 
G1CollectedHeap::evacuate_collection_set(EvacuationInfo&)+0x418;; 
__1cPG1CollectedHeapXevacuate_collection_set6MrnOEvacuationInfo__v_+0x418
V [libjvm.so+0x7f1a84] bool 
G1CollectedHeap::do_collection_pause_at_safepoint(double)+0x824;; 
__1cPG1CollectedHeapbGdo_collection_pause_at_safepoint6Md_b_+0x824
V [libjvm.so+0x1241a7c] void VM_G1IncCollectionPause::doit()+0x2e4;; 
__1cXVM_G1IncCollectionPauseEdoit6M_v_+0x2e4
V [libjvm.so+0x123de38] void VM_Operation::evaluate()+0xf8;; 
__1cMVM_OperationIevaluate6M_v_+0xf8
V [libjvm.so+0x1239dcc] void 
VMThread::evaluate_operation(VM_Operation*)+0x254;; 
__1cIVMThreadSevaluate_operation6MpnMVM_Operation__v_+0x254
V [libjvm.so+0x123a83c] void VMThread::loop()+0x654;; 
__1cIVMThreadEloop6M_v_+0x654
V [libjvm.so+0x1239834] void VMThread::run()+0xe4;; 
__1cIVMThreadDrun6M_v_+0xe4
V [libjvm.so+0xef8af0] java_start+0x258;; java_start+0x258

With the proposed fix the test passes.

Thanks,
Bengt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20130816/187e4338/attachment.htm>