stan/jdk-24
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
46,241 Commits 2 Branches 0 Tags
Commit Graph

747 Commits

Author SHA1 Message Date
Sangheon Kim
cf5132e887 8078555: GC: implement ranges (optionally constraints) for those flags that have them missing 
Add ranges and constraint functions for GC flags.

Reviewed-by: jmasa, kbarrett, gziemski, ddmitriev, drwhite
 2015-08-31 11:51:29 -07:00
Kim Barrett
15196341a5 8131330: G1CollectedHeap::verify_dirty_young_list fails with assert 
Use assembly loop to avoid compiler optimization into memset

Reviewed-by: ecaspole, tschatzl
 2015-08-31 13:06:01 -04:00
Kim Barrett
7706e36194 8072817: CardTableExtension kind() should be BarrierSet::CardTableExtension 
Use BarrierSet::CardTableForRS where needed, and update concrete bs tags.

Reviewed-by: jwilhelm, jmasa
 2015-08-18 17:48:35 -04:00
Erik Helin
6239649ec3 8134509: G1ParCopyClosure does not need a ReferenceProcessor 
Reviewed-by: stefank, mgerdin
 2015-08-28 09:05:46 +02:00
Erik Helin
ec7659c972 8134504: Remove usage of EvacuationInfo from G1CollectorPolicy 
Reviewed-by: mgerdin, jwilhelm
 2015-08-27 16:28:05 +02:00
Jiangli Zhou
9ee2b211f9 8131734: assert(!is_null(v)) failed: narrow klass value can never be zero with -Xshared:auto 
Handle shared string mapping failures.

Co-authored-by: Tom Benson <tom.benson@oracle.com>
Reviewed-by: tschatzl, kbarrett, ddmitriev, hseigel
 2015-08-25 17:22:56 -04:00
Thomas Schatzl
494872e535 8133456: HeapRegionManager::shrink_by() iterates suboptimally across regions 
Reviewed-by: brutisso, tbenson
 2015-08-24 16:32:50 +02:00
Thomas Schatzl
472a3333a8 8067336: Allow that PLAB allocations at the end of regions are flexible 
PLAB allocations may return a buffer that is between minimum size (current allocation) and the desired size. This allows removes a large amount of fragmentation at the end of regions.

Reviewed-by: tbenson, mgerdin
 2015-08-24 16:27:41 +02:00
Thomas Schatzl
c91c77abd2 8067339: PLAB reallocation might result in failure to allocate object in that recently allocated PLAB 
Properly size the PLAB to make sure that the object to allocate always has enough space in it.

Reviewed-by: mgerdin, brutisso
 2015-08-20 15:17:43 +02:00
Thomas Schatzl
041c76d48b 8073146: G1 merges thread local age tables too early with global age table 
Move merging of age and survivor table to the end of GC.

Reviewed-by: mgerdin, david
 2015-08-20 15:17:43 +02:00
Thomas Schatzl
6cca258181 8040162: Avoid reallocating PLABs between GC phases in G1 
Allocate G1ParScanThreadState at the start of GC and only deallocate at the end of GC. This automatically keeps the current PLAB intact without flushing it (and loosing lots of memory) for every reallocation.

Reviewed-by: david, mgerdin
 2015-08-20 15:17:43 +02:00
Thomas Schatzl
4a849fb21c 8133530: Add JFR event for evacuation statistics 
Introduce two new JFR events for young/old generation allocation statistics based on previous changes.

Reviewed-by: ehelin, mgerdin
 2015-08-20 15:17:41 +02:00
Thomas Schatzl
0ade675b8b 8073013: Add detailed information about PLAB memory usage 
Track and provide information about direct allocation, region end waste and failure waste for G1.

Reviewed-by: jmasa, david, ehelin, mgerdin
 2015-08-19 13:59:39 +02:00
Thomas Schatzl
43a6ce61de 8133470: Uses of Atomic methods in plab.hpp should be moved to .inline.hpp file 
Fix include file dependencies according to general guidelines for plab.hpp

Reviewed-by: ehelin, mgerdin
 2015-08-19 13:50:50 +02:00
Thomas Schatzl
8c2da662ff 8003237: G1: Reduce unnecessary (and failing) allocation attempts when handling an evacuation failure 
Remember for every generation whether the memory for that generation has already been exhausted. If so, do not try to get a new region for these generations. Further, if the last generation is full, early exit from copy_to_survivor_space.

Reviewed-by: jmasa, brutisso
 2015-08-19 13:47:40 +02:00
Erik Helin
9bc2e29405 8133825: Remove the class G1CollectorPolicyExt 
Reviewed-by: jwilhelm, ecaspole
 2015-08-20 10:27:04 +02:00
Eric Caspole
3c41a37bd8 8133349: CMS: Assert failed: Ctl pt invariant 
_survivor_chunk_array should be sized based on PLAB::min_size() not ThreadLocalAllocBuffer::min_size()

Reviewed-by: tschatzl, jwilhelm
 2015-08-18 14:04:24 -04:00
David Lindholm
4445cbb273 8133553: Running with -XX:+UseParallelGC -XX:OldSize=30k crashes jvm 
Reviewed-by: jwilhelm, sjohanss
 2015-08-17 10:43:11 +02:00
Stefan Karlsson
0e252b2a93 8087324: Use semaphores when starting and stopping GC task threads 
Reviewed-by: jmasa, sjohanss
 2015-06-29 11:11:12 +02:00
Stefan Karlsson
e25bcfd3d3 8087323: Unify and split the work gang classes 
Reviewed-by: jmasa, sjohanss
 2015-06-29 11:09:39 +02:00
Axel Siebenborn
d32b06427d 8133121: Move implementation of process_grey_object to concurrentMark.inline.hpp 
Move implementation of process_grey_object to inline.hpp

Reviewed-by: kbarrett, simonis
 2015-08-13 09:32:01 +02:00
David Lindholm
b206c669f7 8133193: Memory leak in G1 because G1RootProcessor doesn't have desctructor 
Reviewed-by: mgerdin, tschatzl
 2015-08-12 16:32:29 +02:00
Kim Barrett
a4537791a8 Merge 2015-08-08 00:27:39 +02:00
Thomas Schatzl
c6cfb3a7fd 8133042: Remove some direct accesses of G1Allocator to G1CollectedHeap members 
Remove direct access to private members of G1CollectedHeap from G1Allocator and G1ArchiveAllocator. Allows removal of friend delcarations.

Reviewed-by: kbarrett, jmasa, jwilhelm
 2015-08-07 15:37:07 +02:00
Eric Caspole
bcca305c14 8078904: CMS: Assert failed: Ctl pt invariant 
Survivor rescan array now baed on MinTLABSize so all entries will fit

Reviewed-by: tschatzl, kbarrett
 2015-08-07 09:15:42 -04:00
Thomas Schatzl
c2d0a8dbd0 8133047: Rename G1ParScanThreadState::_queue_num to _worker_id 
Reviewed-by: jmasa, jwilhelm
 2015-08-06 15:52:33 +02:00
Thomas Schatzl
183850cec4 8133043: Clean up code related to termination stats printing 
Reformat termination stats related code to make it look more similar to existing code.

Reviewed-by: jwilhelm, jmasa
 2015-08-06 15:49:52 +02:00
Thomas Schatzl
6e3ffb1aae 8073052: Rename and clean up the allocation manager hierarchy in g1Allocator.?pp 
Fix naming of G1ParGCAllocator, decrease dependencies between G1CollectedHeap, G1Allocator and G1AllocRegion, additional documentation.

Reviewed-by: mgerdin, jmasa, kbarrett
 2015-08-06 15:49:50 +02:00
Sangheon Kim
a13fe250b0 8130459: Add additional validation after heap creation 
Some PLAB related flags' validation can happen after heap creation. Introduced additional stage of validation check.

Reviewed-by: coleenp, dholmes, kbarrett, gziemski, ecaspole, ddmitriev
 2015-07-27 13:56:26 -07:00
Kim Barrett
093ee9b659 8079082: VerifyNoCSetOopsClosure is derived twice from Closure 
Change closure to a function object and update iteration

Reviewed-by: mgerdin, ecaspole
 2015-07-27 14:05:55 -04:00
Eric Caspole
cf7ea2f64e 8132148: G1 hs_err region dump legend out of sync with region values 
Use E for Eden, S for Survivor, O for Old

Reviewed-by: tschatzl, tbenson
 2015-07-24 10:32:00 -04:00
Walter Florian Gugenberger
ca0fd4b1fe 8004687: G1: Parallelize object self-forwarding and scanning during an evacuation failure 
Use the regular task queue during evacuation failure and allow per-thread preserved header queues to remove the global lock during evacuation failure.

Reviewed-by: mgerdin, jmasa
 2015-07-23 11:14:47 +02:00
Erik Helin
57553520be 8131319: Move G1Allocator::_summary_bytes_used back to G1CollectedHeap 
G1Allocator::_summary_bytes_used fits better in G1CollectedHeap since we added the archive allocator. Move it back to the original location.

Reviewed-by: mgerdin, tbenson
 2015-07-23 11:14:24 +02:00
Kim Barrett
717679c169 8130931: Refactor CardTableModRefBS[ForCTRS] 
Remove friends and push CTRS-specific code down from base to derived

Reviewed-by: tschatzl, mgerdin
 2015-07-22 00:37:01 -04:00
Thomas Schatzl
70eeaa4735 8131166: Remove additional whitespace in G1Allocator 
Reviewed-by: jmasa, tbenson
 2015-07-16 11:54:37 +02:00
Eric Caspole
34bcc977bc 8129961: SIGSEGV when copying to survivor space 
Remove "include_young" parameter from GenCollectedHeap::no_allocs_since_save_marks() since all existing uses pass true to always rescan young gen.

Reviewed-by: jmasa, kbarrett
 2015-07-13 11:49:23 -04:00
Erik Helin
3a37b5d81c 8142402: G1 should not redirty cards in free regions 
Reviewed-by: mgerdin, tschatzl
 2015-04-29 13:17:46 +02:00
Erik Helin
62bdcb0abf 8144145: G1GCPhaseTimes should allow externally accounted time 
Reviewed-by: ehelin, tschatzl
 2015-11-29 10:00:31 +01:00
Erik Helin
66f8f80768 8144077: Add getter for G1CollectorPolicy::_collectionSetChooser 
Reviewed-by: mgerdin, tschatzl
 2015-11-29 10:00:23 +01:00
Jon Masamitsu
48ed80d136 7012980: PSOldGen is increased if there is no space in Metaspace 
Reviewed-by: tschatzl, tbenson
 2015-04-15 11:30:36 -07:00
Jon Masamitsu
bd52f0a2ef Merge 2015-07-07 18:16:06 +00:00
Sangheon Kim
3e36930486 8079555: REDO - Determining the desired PLAB size adjusts to the the number of threads at the wrong place 
Calculate the desired PLAB value for a single thread and then return desired PLAB size according to the current number of threads when needed

Reviewed-by: jmasa, tschatzl
 2015-07-07 06:37:10 -07:00
Thomas Schatzl
c8c4924aa3 8129558: Coalesce dead objects during removal of self-forwarded pointers 
To improve performance of self-forwarding fixup during evacuation failure, coalesce the work done for dead objects.

Reviewed-by: mgerdin, tbenson, tonyp
 2015-07-07 14:20:08 +02:00
Stefan Johansson
c85e4ae56f 8078901: Add trace event for G1 MMU information 
Reviewed-by: ehelin, mlarsson
 2015-04-30 13:33:23 +02:00
Poonam Bajaj
b34dec66c6 Merge 2015-06-25 18:14:54 +00:00
Thomas Schatzl
e5a5f9beab Merge 2015-06-25 15:43:16 +00:00
Poonam Bajaj
c7612a1138 8129108: nmethod related crash in CMS 
Add SO_AllCodeCache to root scanning options when not unloading classes with a CMS collection cycle

Reviewed-by: mgerdin, jwilhelm
 2015-06-25 06:52:05 -07:00
Bengt Rutisson
e7a2bfa767 Merge 2015-06-25 10:52:37 +00:00
Thomas Schatzl
2431e87477 8129573: CollectedHeap::fill_with_objects() needs to use multiple arrays in 32 bit mode too 
In JDK-8042668 we introduced a custom fill threshold for G1. This leads to CollectedHeap::fill_with_objects create too large objects in G1 when using it in 32 bit mode, as the code to create multiple filler objects is IFDEF'ed out on 32 bit. Enable this code on 32 bit too.

Reviewed-by: tonyp, mgerdin, tbenson
 2015-06-25 09:06:35 +02:00
Bengt Rutisson
5f4a67087f 8129626: G1: set_in_progress() and clear_started() needs a barrier on non-TSO platforms 
Also reviewed by vitalyd@gmail.com

Reviewed-by: pliden, bpittore, bdelsart
 2015-06-25 08:15:07 +02:00