summaryrefslogtreecommitdiffstats
path: root/runtime/gc/heap.h
Commit message (Collapse)AuthorAgeFilesLines
* Implement TLAB fast paths in artAllocObjectFromCode.Hiroshi Yamauchi2014-08-061-1/+1
| | | | | | | | | | | | | | | | | | | GSS/TLAB GC speedup on N4 (ms): MemAllocTest 2963 -> 2792 BinaryTrees 2205 -> 2113 Also, measured wth -XX:IgnoreMaxFootprint to invoke GC less often (only when the bump pointer space is filled rather than based on the target utilization): MemAllocTest 2707 -> 2590 BinaryTrees 2023 -> 1906 TODO: implement fast paths for array allocations. Bug: 9986565 Change-Id: I73ff6327b229704f8ae5924ae9b747443c229841
* Fix memory leak in RemoveRememberedSet.Mathieu Chartier2014-07-241-0/+1
| | | | | | | RemoveRememberedSet now deletes the remembered set. Bug: 16532086 Change-Id: I01092931cc20cd0688dd42eed3dde9ad140889b2
* Recycle mem-maps for collector transitions.Mathieu Chartier2014-07-221-4/+9
| | | | | | | | | | | | | | | | | | | | We now create spaces when we need them for collector transitions or homogeneous compaction by recycling mem maps. Change the bump pointer space size to be as large as the heap capacity instead of 1/2 heap capacity like it used to be. For GSS, bump pointer spaces are set to 32MB currently. Changed GSS to have main space == non moving space since we don't need to copy from the main space. Fixes GC stress tests 074, 096. Fixed test 080 oom throw with -Xmx2m for GC stress test, this was broken since it was allocating a 4 MB array before starting the OOM process. Bug: 14059466 Bug: 16406852 Change-Id: I62877cfa24ec944a6f34ffac30334f454a8002fd
* Disable adding main and non moving spaces to immune region in GSSMathieu Chartier2014-07-161-1/+2
| | | | | | | | | | | | | | | | | | Disabled adding the main and non moving space to the immune region. This will enable us to recycle bump pointer spaces for malloc space -> malloc space compaction as well as collector transitions. Also added logic for falling back to the non moving space, we may copy objects there. Refactored mod union table logic into MarkReachableObjects. No measurable performance benefit or regression. Bug: 14059466 Bug: 16291259 Change-Id: If663d9fdbde943b988173b7f6ac844e5f78a0327
* ART: Compacting ROS/DlMalloc spaces with semispace copy GCZuo Wang2014-07-111-4/+61
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Current semispace copy GC is mainly associated with bump pointer spaces. Though it squeezes fragmentation most aggressively, an extra copy is required to re-establish the data in the ROS/DlMalloc space to allow CMS GCs to happen afterwards. As semispace copy GC is still stop-the-world, this not only introduces unnecessary overheads but also longer response time. Response time indicates the time duration between the start of transition request and the start of transition animation, which may impact the user experience. Using semispace copy GC to compact the data in a ROS space to another ROS(or DlMalloc space to another DlMalloc) space solves this problem. Although it squeezes less fragmentation, CMS GCs can run immediately after the compaction. We apply this algorithm in two cases: 1) Right before throwing an OOM if -XX:EnableHSpaceCompactForOOM is passed in as true. 2) When app is switched to background if the -XX:BackgroundGC option has value HSpaceCompact. For case 1), OOMs are significantly delayed in the harmony GC stress test, with compaction ratio up to 0.87. For case 2), compaction ratio around 0.5 is observed in both built-in SMS and browser. Similar results have been obtained on other apps as well. Change-Id: Iad9eabc6d046659fda3535ae20f21bc31f89ded3 Signed-off-by: Wang, Zuo <zuo.wang@intel.com> Signed-off-by: Chang, Yang <yang.chang@intel.com> Signed-off-by: Lei Li <lei.l.li@intel.com> Signed-off-by: Lin Zang <lin.zang@intel.com>
* Change default heap maximum size to be 256m.Mathieu Chartier2014-07-101-1/+1
| | | | | | Useful for command line benchmarks. Change-Id: Ie525863cd8eff93c64ce76639b1108fbdad91633
* Improve the OOME fragmentation message.Hiroshi Yamauchi2014-07-101-2/+1
| | | | Change-Id: I390d3622f8d572ec7e34ea6dff9e1e0936e81ac1
* Add notion of released vs empty pages to ROSAlloc.Mathieu Chartier2014-07-091-2/+2
| | | | | | | | | | | | | | | | | A notion of released vs empty pages helps get a more accurate view of how much memory was released during heap trimming. Otherwise we get that the same pages possibly get madvised multiple times without getting dirtied. Also enabled heap trimming of rosalloc spaces even when we care about jank. This is safe to do since the trimming process only acquires locks for short periods of time. Dalvik PSS reduces from ~52M to ~50M after boot on N4. Bug: 9969166 Change-Id: I4012e0a2554f413d18efe1a0371fe18d1edabaa9
* Add easy way to ensure the next allocation does GC.Mathieu Chartier2014-07-091-0/+20
| | | | | | | | | Added a class called ScopedHeapFill which changes the bytes allocated counter to be equal to the growth limit. This causes the next allocation to do a GC and possibly generate an OOM error. This is useful for tests which need GC to happen at specific point. Change-Id: Ibd8f3d5928b58534c5165ba7c296980002aa2c28
* Don't allow negative GetFreeMemory.Mathieu Chartier2014-06-271-1/+4
| | | | | | | | | | | | The max allowed footprint is only updated after the GC. But we can still allocate even if bytes_allocated > max_allowed_footprint_. This means that we used to be able to get a negative value if bytes_allocated > max_allowed_footprint_. External bug: https://code.google.com/p/android/issues/detail?id=72221 Change-Id: I4ef9a534e29211786e82cdcb2582c11ab37a348a
* Shared single GC iteration accounting for all GCs.Mathieu Chartier2014-06-201-1/+12
| | | | | | | | | | | | Previously, each garbage collector had data that was only used during collection. Since only one collector can be running at any given time, we can make this data be shared between all collectors. This reduces memory usage since we don't need to have redundant information for each GC types. Also reduced how much code is required to sweep spaces. Bug: 9969166 Change-Id: I31caf0ee4d572f75e0c66863fe7db12c08ae08e7
* Add mark compact collector.Mathieu Chartier2014-06-191-1/+4
| | | | | | | | | | | | | | | | | | | | | | The mark compact collector is a 4 phase collection, doing a normal full mark_sweep, calculating forwarding addresses of objects in the from space, updating references of objects in the from space, and moving the objects in the from space. Support is diabled by default since it needs to have non movable classes and field arrays. Performance numbers is around 50% as fast. The main advantage that this has over semispace is that the worst case memory usage is 50% since we only need one space isntead of two. TODO: Make field arrays and classes movable. This causes complication since Object::VisitReferences relies on these, so if we update the fields of an object but another future object uses this object to figure out what fields are reference fields it doesn't work. Bug: 14059466 Change-Id: I661ed3b71ad4dde124ef80312c95696b4a5665a1
* Fix GetFreeMemory to use fooprint limit instead of max memory.Mathieu Chartier2014-06-091-1/+1
| | | | | | | | Based on definitions in: http://developer.android.com/reference/java/lang/Runtime.html Bug: 15507122 Change-Id: I02f34682d7ac2d379a07631b5207b6cfb224da6b
* Change GetFreeMemory to use the growth limit.Mathieu Chartier2014-05-291-1/+1
| | | | | | | | | This makes more sense since it is what the allocator uses. Also fixed somewhere where we were not properly passing in whether or not it was a large object allocation. Bug: 15327879 Change-Id: Ieab7af5427f5cdc2760390186b67e2c96d4bafa7
* Add reserve area to allocation stacks.Mathieu Chartier2014-05-231-0/+4
| | | | | | | | | | | This fixes an issue with heap verification which was caused when the allocation stack overflowed. This resulted in heap verification failures since we were storing the newly allocated object in a handle scope without having it be live either in the live bitmap or allocation stack. We now push the object in the reserve area before we do a GC due to allocation stack overflow. Change-Id: I83b42c4b3250d7eaab1b49e53066e21c8656a740
* Merge "Begin migration of art::Atomic to std::atomic."Ian Rogers2014-05-211-2/+2
|\
| * Begin migration of art::Atomic to std::atomic.Ian Rogers2014-05-201-2/+2
| | | | | | | | Change-Id: I4858d9cbed95e5ca560956b9dabd976cebe68333
* | Change zygote_creation_lock_ to be member instead of static.Mathieu Chartier2014-05-211-0/+3
| | | | | | | | | | | | | | | | Static variables aren't thread safe and could cause the zygote to be created twice. Bug: 15133494 Change-Id: I65c8f089bed8de93f895b62b3dcff4c936931860
* | Improve heap verification for invalid roots.Mathieu Chartier2014-05-201-3/+2
|/ | | | | | | | | | | The new root verification prints the root type and owner thread id as well as the type of the object. Also a bit of work for planned multi-threaded verification. Bug: 14289301 Change-Id: Ia73c517dc11ec6dd82f3d945604ee3836b3db536
* Now we have a proper C++ library, use std::unique_ptr.Ian Rogers2014-05-191-8/+8
| | | | | | | Also remove the Android.libcxx.mk and other bits of stlport compatibility mechanics. Change-Id: Icdf7188ba3c79cdf5617672c1cfd0a68ae596a61
* Follow up to "Add ISA directory to image and odex pathnames."Brian Carlstrom2014-05-191-1/+1
| | | | Change-Id: I7f08cc3052fbed93a56ccf1ab7675ae8bc129da9
* Add Handle/HandleScope and delete SirtRef.Mathieu Chartier2014-05-131-2/+9
| | | | | | | | | | | | | | | | Delete SirtRef and replaced it with Handle. Handles are value types which wrap around StackReference*. Renamed StackIndirectReferenceTable to HandleScope. Added a scoped handle wrapper which wraps around an Object** and restores it in its destructor. Renamed Handle::get -> Get. Bug: 8473721 Change-Id: Idbfebd4f35af629f0f43931b7c5184b334822c7a
* Add concurrent reference processing.Mathieu Chartier2014-05-071-29/+11
| | | | | | | | | | | | | | | | | Concurrent reference processing currently works by going into native code from java.lang.ref.Reference.get(). From there, we have a fast path if the references aren't being processed which returns the referent without needing to access any locks. In the slow path we block until reference processing is complete. It may be possible to improve the slow path if the referent is blackened. TODO: Investigate doing the fast path in java code by using racy reads of a static volatile boolean. This will work as long as there are no suspend points inbetween the boolean read and referent read. Bug: 14381653 Change-Id: I1546b55be4691fe4ff4aa6d857b234cce7187d87
* Add RecordFree to the GarbageCollector interfaceMathieu Chartier2014-05-051-1/+1
| | | | | | | RecordFree now calls the Heap::RecordFree as well as updates the garbage collector's internal bytes freed accounting. Change-Id: I8cb03748b0768e3c8c50ea709572960e6e4ad219
* Merge "Add finalizer references from the entrypoints."Mathieu Chartier2014-05-021-1/+1
|\
| * Add finalizer references from the entrypoints.Mathieu Chartier2014-05-011-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We now have an invariant where we never allocate finalizable objects with the Initialized or Resolved entrypoints. This speeds up allocation by only doing the check in the slow path. Before: MemAllocTest: 3625, 3707, 3641 EvaluateAndApplyChanges: 3448, 3421, 3413 After: MemAllocTest: 3164, 3109, 3135 EvaluateAndApplyChanges: 3272, 3299, 3353 Bug: 14078487 Change-Id: I2b0534af3e7c75ea5e5257cf3647744f7abfb74e
* | Add GC cause to WaitForGcToComplete message.Mathieu Chartier2014-05-021-2/+3
|/ | | | Change-Id: I8fe107d90a84de065c407b8d29fd106267ac440d
* Merge "Clean up Add/Remove space."Mathieu Chartier2014-05-011-4/+5
|\
| * Clean up Add/Remove space.Mathieu Chartier2014-04-301-4/+5
| | | | | | | | | | | | | | Deleted the set_as_default parameter and added a new function SetSpaceAsDefault instead. Change-Id: Ic4c359854d08e64ac0d0df92f0105447adb9df36
* | Use instruction specific dalvik cache dirs.Narayan Kamath2014-05-011-0/+2
|/ | | | | | | | | | | | - All oat & art files are now placed under /data/dalvik-cache/<isa>/. - GetDalvikCacheOrDie now requires a mandatory subdirectory argument, and is implicitly rooted under /data/. - Added helper methods to convert InstructionSet enums into strings and vice versa. (cherry picked from commit 2974bc3d8a5d161d449dd66826d668d87bdc3cbe) Change-Id: Ic7986938e6a7091a2af675ebafec768f7b5fb8cd
* Force inlining on trivial accessors.Ian Rogers2014-04-291-10/+1
| | | | | | | | | Make volatility for GetFieldObject a template parameter. Move some trivial mirror::String routines to a -inl.h. Bug: 14285442 Change-Id: Ie23b11d4f18cb15a62c3bbb42837a8aaf6b68f92
* Enable concurrent sweeping for non-concurrent GC.Mathieu Chartier2014-04-291-11/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Refactored the GarbageCollector to let all of the phases be run by the collector's RunPhases virtual method. This lets the GC decide which phases should be concurrent and reduces how much baked in GC logic resides in GarbageCollector. Enabled concurrent sweeping in the semi space and non concurrent mark sweep GCs. Changed the semi-space collector to have a swap semi spaces boolean which can be changed with a setter. Fixed tests to pass with GSS collector, there was an error related to the large object space limit. Before (EvaluateAndApplyChanges): GSS paused GC time 7.81s/7.81s, score: 3920 After (EvaluateAndApplyChanges): GSS paused GC time 6.94s/7.71s, score: 3900 Benchmark score doesn't go up since the GC happens in the allocating thread. There is a slight reduction in pause times experienced by other threads (0.8s total). Added options for pre sweeping GC heap verification and pre sweeping rosalloc verification. Bug: 14226004 Bug: 14250892 Bug: 14386356 Change-Id: Ib557d0590c1ed82a639d0f0281ba67cf8cae938c
* Enable reading page map without lock in RosAlloc::BulkFreeMathieu Chartier2014-04-251-4/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Enabling this flag greatly reduces how much time was spent in the GC. It was not done previously since it was regressing MemAllocTest. With these RosAlloc changes, the benchmark score no longer regresses after we enable the flag. Changed Run::AllocSlot to only have one mode of allocation. The new mode is finding the first free bit in the bitmap. This was previously the slow path but is now the fast path. Some optimizations which enabled this include always having the alloc bitmap bits which correspond to invalid slots be set to 1. This prevents us from needing a bound check since we will never end up allocating there. Changed revoking thread local buffer to point to an invalid run. The invalid run is just a run which always has all the allocation bits set to 1. When a thread attempts to do a thread local allocation from here it will always fail and go slow path. This eliminates the need for a null check for revoked runs. Changed zeroing of memory to happen during free, AllocPages should always return zeroed memory. Added prefetching which happens when we allocate a run. Some refactoring to reduce duplicated code. Ergonomics changes: Changed kStickyGcThroughputAdjustment to 1.0, this helps reduce GC time. Measurements (3 samples per benchmark): Before: MemAllocTest scores: 3463, 3445, 3431 EvaluateAndApplyChanges score | total GC time Iter 1: 3485, 23.602436s Iter 2: 3434, 22.499882s Iter 3: 3483, 23.253274s After: MemAllocTest scores: 3495, 3417, 3409 EvaluateAndApplyChanges score | total GC time: Iter 1: 3375, 17.463462s Iter 2: 3358, 16.185188s Iter 3: 3367, 15.822312s Bug: 8788501 Bug: 11790317 Bug: 9986565 Change-Id: Ifd273a054824028dabed27c07c081dde1816f93c
* Replace ObjectSet with LargeObjectBitmap.Mathieu Chartier2014-04-171-2/+2
| | | | | | | | | | | | | | | | | | | | | | | Speeds up large object marking since large objects no longer required a lock. Changed the GCs to use the heap bitmap for marking objects which aren't in the fast path. This eliminates the need for a MarkLargeObject function. Maps before (10 GC iterations): Mean partial time: 180ms Mean sticky time: 151ms Maps after: Mean partial time: 161ms Mean sticky time: 101ms Note: the GC durations are long due to recent ergonomic changes and because the fast bulk free hasn't yet been enabled. Over 50% of the GC time is spent in RosAllocSpace::FreeList. Bug: 13571028 Change-Id: Id8f94718aeaa13052672ccbae1e8edf77d653f62
* Decrease target utilization for foreground apps.Mathieu Chartier2014-04-161-3/+10
| | | | | | | | | | GC time in FormulaEvaluationActions.EvaluateAndApplyChanges goes from 26.1s to 23.2s. Benchmark score goes down ~50 in FormulaEvaluationActions.EvaluateAndApplyChanges, and up ~50 in GenericCalcActions.MemAllocTest. Bug: 8788501 Change-Id: I412af1205f8b67e70a12237c990231ea62167bc0
* Refactor space bitmap to support different alignments.Mathieu Chartier2014-04-141-1/+3
| | | | | | | | | | | Required for: Using space bitmaps instead of std::set in mod union table + remembered set. Using a bitmap instead of set for large object marking. Bug: 13571028 Change-Id: Id024e9563d4ca4278f79607cdb2f81895121b113
* Clean-up heap spaces.Mathieu Chartier2014-04-111-18/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | We now use the CMS collector instead of the semispace collector when the phone is booting. We still perform compaction during the zygote space creation. This reduces time spent in GC by ~2s during boot and doesn't affect zygote space size. Changed the space creation logic to create the temp space when a background transition occurs. Added a flag to each space which is true if you are allowed to move objects that are within this space. Removed SwapSemiSpaces call from the semi space collector, it is now the job of the caller to do this with threads suspended. This simplifies the logic in the zygote compaction / heap transition code since these do not copy from one semispace to another. Added Space::Clear to RosAllocSpace and DlMallocSpace. This greatly simplifies the code used for collector transitions. Time spent in GC creating zygote space: Before: 3.4s, After: 1.28s No change in zygote space size. Bug: 13878055 Change-Id: I700348ab7d5bf3aa537c0cd70c0fed09aa4b0623
* Improve invalid root dumping.Mathieu Chartier2014-04-041-0/+2
| | | | | | The invalid root dumping now attempts to print the root type. Change-Id: Ie821296d569f34909ba6e2705f5c347cd2143a3a
* Fix Object::Clone()'s pre-fence barrier.Hiroshi Yamauchi2014-04-031-7/+8
| | | | | | | | | | | | | | Pass in a pre-fence barrier object that sets in the array length instead of setting it after returning from AllocObject(). Fix another potential bug due to the wrong default pre-fence barrier parameter value. Since this appears error-prone, removed the default parameter value and make it an explicit parameter. Fix another potential moving GC bug due to a lack of a SirtRef. Bug: 13097759 Change-Id: I466aa0e50f9e1a5dbf20be5a195edee619c7514e
* Merge "Refactor some GC code."Mathieu Chartier2014-03-281-0/+3
|\
| * Refactor some GC code.Mathieu Chartier2014-03-281-0/+3
| | | | | | | | | | | | | | | | Reduced amount of code in mark sweep / semi space by moving common logic to garbage_collector.cc. Cleaned up mod union tables and deleted an unused implementation. Change-Id: I4bcc6ba41afd96d230cfbaf4d6636f37c52e37ea
* | Merge "An empty collector skeleton for a read barrier-based collector."Hiroshi Yamauchi2014-03-281-2/+5
|\ \ | |/ |/|
| * An empty collector skeleton for a read barrier-based collector.Hiroshi Yamauchi2014-03-271-2/+5
| | | | | | | | | | | | Bug: 12687968 Change-Id: Ic2a3a7b9943ca64e7f60f4d6ed552a316ea4a6f3
* | Change sticky GC ergonomics to use GC throughput.Mathieu Chartier2014-03-271-1/+4
|/ | | | | | | | | | | | | | | | | The old sticky ergonomics used partial/full GC when the bytes until the footprint limit was < min free. This was suboptimal. The new sticky GC ergonomics do partial/full GC when the throughput of the current sticky GC iteration is <= mean throughput of the partial/full GC. Total GC time on FormulaEvaluationActions.EvaluateAndApplyChanges. Before: 26.4s After: 24.8s No benchmark score change measured. Bug: 8788501 Change-Id: I90000305e93fd492a8ef5a06ec9620d830eaf90d
* Revoke rosalloc thread-local buffers at the checkpoint.Hiroshi Yamauchi2014-03-211-0/+2
| | | | | | | | | | | | | | In the mark sweep collector, rosalloc thread-local buffers were revoked during the pause. Now, they are revoked at the thread checkpoint, as opposed to during the pause, which appears to help reduce the pause time. In Ritz MemAllocTest, the average sticky pause time went down ~20% (925 us -> 724 us). Bug: 13394464 Bug: 9986565 Change-Id: I104992a11b46d59264c0b9aa2db82b1ccf2826bc
* Merge "Refactor the garbage collector driver (GarbageCollector::Run)."Hiroshi Yamauchi2014-03-211-4/+6
|\
| * Refactor the garbage collector driver (GarbageCollector::Run).Hiroshi Yamauchi2014-03-201-4/+6
| | | | | | | | | | | | Bug: 12687968 Change-Id: Ifc9ee86249f7938f51495ea1498cf0f7853a27e8
* | Merge "Fix RecordFree to take signed parameters."Mathieu Chartier2014-03-201-1/+1
|\ \
| * | Fix RecordFree to take signed parameters.Mathieu Chartier2014-03-201-1/+1
| |/ | | | | | | | | | | | | | | | | | | RecordFree can get negative bytes allocated when background compaction foreground transitions occur. This caused a DCHECK to fail on debug builds. Also did some refactoring in PreProcessReferences. Bug: 13568814 Change-Id: I57543f1c78544a94f1d241459698b736dba8cfa8
* | Add Heap::RunningOnValgrind and call it from the spaces.Mathieu Chartier2014-03-201-0/+3
|/ | | | | | Makes it easier to disable valgrind support. Change-Id: I1bde792f1b76a2dd968fa03c6142e92fcc3670b0
generated by cgit v1.1 at 2025-02-09 19:05:01 +0000