archive/jdk
1
0
Fork 0
mirror of https://github.com/openjdk/jdk.git synced 2025-09-21 19:44:41 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

6668573: CMS: reference processing crash if ParallelCMSThreads > ParallelGCThreads

Browse source 
Use _max_num_q = max(discovery_degree, processing_degree), and let balance_queues() redistribute from discovery_degree to processing_degree of queues. This should also allow a more dynamic and flexible parallelism policy in the future.

Reviewed-by: jmasa, johnc
This commit is contained in:
Y. Srinivas Ramakrishna 2011-03-17 10:32:46 -07:00
parent 7293fabc43
commit b4b287e7ac
13 changed files with 148 additions and 153 deletions
Download patch file Download diff file Expand all files Collapse all files

18
hotspot/src/share/vm/gc_implementation/parallelScavenge/psScavenge.cpp
View file

@ -1,5 +1,5 @@
/*
* Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2002, 2011, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
@ -796,13 +796,15 @@ void PSScavenge::initialize() {
// Initialize ref handling object for scavenging.
MemRegion mr = young_gen->reserved();
_ref_processor = ReferenceProcessor::create_ref_processor(
mr, // span
true, // atomic_discovery
true, // mt_discovery
NULL, // is_alive_non_header
ParallelGCThreads,
ParallelRefProcEnabled);
_ref_processor =
new ReferenceProcessor(mr, // span
ParallelRefProcEnabled && (ParallelGCThreads > 1), // mt processing
(int) ParallelGCThreads, // mt processing degree
true, // mt discovery
(int) ParallelGCThreads, // mt discovery degree
true, // atomic_discovery
NULL, // header provides liveness info
false); // next field updates do not need write barrier
// Cache the cardtable
BarrierSet* bs = Universe::heap()->barrier_set();