johnc [Mon, 01 Jul 2013 09:30:23 -0700] rev 18496
8017070: G1: assert(_card_counts[card_num] <= G1ConcRSHotCardLimit) failed
Summary: The assert is invalid when a card is being refined by two different threads and its count crosses the hot threshold - the refinement count will be updated once by each thread triggering the assert. Remove the assert and update the count using a bounded expression.
Reviewed-by: jmasa, tamao, brutisso
brutisso [Sun, 30 Jun 2013 21:42:07 +0200] rev 18495
8014022: G1: Non Java threads should lock the shared SATB queue lock without safepoint checks.
Reviewed-by: tschatzl, brutisso, jmasa, ysr
Contributed-by: per.liden@oracle.com
ehelin [Fri, 28 Jun 2013 18:28:17 +0200] rev 18494
Merge
brutisso [Thu, 27 Jun 2013 09:59:59 +0200] rev 18493
8017483: G1 tests fail with native OOME on Solaris x86 after HeapBaseMinAddress has been increased
Summary: Set HeapBaseMinAddress as default rather than ergo
Reviewed-by: stefank, jmasa, kvn
coleenp [Tue, 02 Jul 2013 16:54:24 +0200] rev 18492
Merge
coleenp [Tue, 02 Jul 2013 08:42:37 -0400] rev 18491
8015391: NPG: With -XX:+UseCompressedKlassPointers OOME due to exhausted metadata space could occur when metaspace is almost empty
Summary: Allocate medium chunks for class metaspace when class loader has lots of classes
Reviewed-by: mgerdin, jmasa
sspitsyn [Mon, 01 Jul 2013 14:13:12 -0700] rev 18490
8009204: [dtrace] signatures returned by Java 7 jstack() are corrupted on Solaris
Summary: The fix is basically a backport of JDK-7019165 (pstack issue) to jhelper.d.
Reviewed-by: coleenp, sspitsyn
Contributed-by: tomas.hurka@oracle.com