log

age author description
Fri, 28 Oct 2016 11:30:24 +0200 roland Condition code not set after CAS on aarch64
Fri, 28 Oct 2016 12:04:19 +0200 shade Chunked array processing should first push the continuation.
Tue, 25 Oct 2016 17:53:35 +0200 shade Native Brooks ptr accesses can be constant-folded.
Tue, 25 Oct 2016 17:11:49 +0200 rkennke Insert load-load fence in obj-eq-barrier, to prevent brooks ptr loads from floating above comparison.
Fri, 21 Oct 2016 18:40:55 +0200 rkennke Reuse C2 write barrier stub in interpreter and C1.
Fri, 21 Oct 2016 18:40:17 +0200 rkennke Implement/improve Shenandoah store checks.
Fri, 21 Oct 2016 12:29:51 -0400 chf Merge
Fri, 21 Oct 2016 10:38:49 -0400 chf Better logging
Thu, 20 Oct 2016 11:22:48 +0200 rkennke Rewrite barriers for locks to invoke barriers only when mark word is actually touched.
Tue, 18 Oct 2016 09:31:37 -0400 zgu Parallelize CLDG and CodeCache root scanning and evacuation
Mon, 17 Oct 2016 17:05:10 +0200 shade Prune is_in call from in_cset_fast_test.
Fri, 14 Oct 2016 13:45:56 +0000 rkennke Use non-checking version of mov for 0xdeaddead to avoid assert.
Tue, 04 Oct 2016 16:24:00 -0400 zgu Fixed Shenandoah support in jvmci
Thu, 13 Oct 2016 12:31:58 +0200 rkennke Added includes to fix aarch64 build.
Wed, 12 Oct 2016 10:19:53 +0200 shade Improve UseLargePages support.
Mon, 10 Oct 2016 16:18:43 +0200 roland C2's barrier verification pass doesn't cover compressed oops correctly
Mon, 10 Oct 2016 15:51:20 +0200 shade Parallel AlwaysPreTouch: do heap pre-touch operation in parallel.
Mon, 10 Oct 2016 09:53:56 +0200 rkennke Fix types of oops/oopDesc*/HeapWord* in new asserts to make compiler happy.
Fri, 07 Oct 2016 21:31:46 +0200 shade Refactor BrooksPointer utility class: reflow API, add more asserts, etc.
Fri, 07 Oct 2016 19:07:55 +0200 shade TestWriteBarrierClearControl fails in release build: needs -XX:+UnlockDiagnosticVMOptions.
Fri, 07 Oct 2016 17:49:55 +0200 rkennke More asserts to find where from-space oops come from in biased locking.
Fri, 07 Oct 2016 11:37:56 +0200 roland implicit null checks broken with compressed oops
Fri, 07 Oct 2016 09:57:04 +0200 roland ShenandoahBarrierNode::needs_barrier() support for compressed oops
Fri, 07 Oct 2016 16:15:16 +0200 rkennke Added assert in biased locking to see where to-space objects might come from.
Fri, 07 Oct 2016 11:15:01 +0200 rkennke Added asserts all over synchronizer code to check that target objects are in to-space.
Thu, 06 Oct 2016 18:43:37 +0200 rkennke Use unsafe_equals() instead of == for comparing oops.
Thu, 06 Oct 2016 18:02:45 +0200 rkennke Additional assert and debug output.
Thu, 06 Oct 2016 11:39:41 +0200 rkennke Use MacroAssembler cas-obj code for C1 too. Use movl for compressed oops. Use sete instead of setz to set result reg.
Wed, 05 Oct 2016 10:38:06 +0200 rkennke Remove unnecessary and obsolete cmd line arg from test.
Wed, 05 Oct 2016 09:41:55 +0200 rkennke Rewrite Shenandoah logging to use unified logging framework.
Tue, 04 Oct 2016 10:53:58 +0200 rkennke Cleanup Shenandoah arguments.
Fri, 30 Sep 2016 16:35:24 +0200 shade Throw proper OOMEs instead of crashing on internal asserts.
Fri, 30 Sep 2016 11:30:47 +0200 roland Clearing control during final graph reshape causes memory barrier to loose dependency on null check
Tue, 20 Sep 2016 18:26:48 +0200 roland c2: logic that removes write barrier projection is not accurate
Thu, 29 Sep 2016 10:33:49 +0200 shade Move shenandoah oop CAS to MacroAssembler, cleanup, document, and optimize it a little.
Wed, 28 Sep 2016 19:57:19 +0200 shade Minor cleanup on ShenandoahHeapRegion's LiveData counter.
Wed, 28 Sep 2016 19:42:20 +0200 shade Missing memory tags for marking bitmaps.
Tue, 27 Sep 2016 15:09:28 +0200 rkennke Handle single-threaded GC correctly.
Thu, 22 Sep 2016 10:25:26 +0200 rkennke Fix: Shenandoah crashes with -XX:ObjectAlignmentInBytes=16.
Wed, 21 Sep 2016 17:05:31 +0200 rkennke Add ShenandoahGC into GC flag collision test.
Wed, 21 Sep 2016 16:43:17 +0200 rkennke Added regression test for humongous off-by-one error.
Wed, 21 Sep 2016 16:40:58 +0200 rkennke Added read barrier to C1's vectorizedMismatch() intrinsic.
Wed, 21 Sep 2016 15:44:54 +0200 rkennke Pin regions that contain JNI critical regions, instead of bail-and-retry protocol.
Wed, 21 Sep 2016 10:29:37 +0200 roland c2: don't clear write barrier control in final graph reshape if CastPP input was processed before
Tue, 20 Sep 2016 15:22:14 +0200 rkennke Exit gracefully when trying Shenandoah on unsupported platforms.
Mon, 19 Sep 2016 11:14:10 +0200 rkennke Proper GC mode in hs_err for Shenandoah.
Thu, 15 Sep 2016 10:07:15 +0200 roland loop optimizations for shenandoah read barriers
Fri, 16 Sep 2016 18:42:00 +0200 rkennke Fix Shenandoah's heap validation.
Wed, 14 Sep 2016 16:45:39 +0200 roland When a write barrier's projection is about to be removed, all users of the projection must first update their memory edge to the memory input of the write barrier
Wed, 14 Sep 2016 16:46:52 +0200 roland move loop invariant shenandoah write barriers before loops
Tue, 13 Sep 2016 14:15:11 +0200 roland memory edge of write barriers shouldn't be unconditionally cleared after optimizations
Wed, 14 Sep 2016 10:22:37 +0200 rkennke Merge
Thu, 08 Sep 2016 21:11:53 +0000 lana Added tag jdk-9+135 for changeset 3b1c4562953djdk-9+136
Fri, 02 Sep 2016 02:41:12 +0000 lana Mergejdk-9+135
Wed, 31 Aug 2016 09:22:53 -0700 amurillo Merge
Fri, 26 Aug 2016 10:02:47 -0700 amurillo Merge
Wed, 24 Aug 2016 19:54:03 -0400 dholmes 8157904: Atomic::cmpxchg for jbyte is missing a fence on initial failure
Wed, 24 Aug 2016 20:38:21 +0200 rehn 8164208: Update tests with redefine classes UL options and tags?
Tue, 23 Aug 2016 21:49:33 -0400 ctornqvi 8163146: Remove os::check_heap on Windows
Tue, 23 Aug 2016 22:51:24 +0200 egahlin Merge