Implement incremental marking in Oilpan |
||||||||||||||||||||||||||||||||||||||
Issue description
Showing comments 85 - 184
of 184
Older ›
,
Apr 4 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/0ce3313cc995916da0b54adf440f4ec468298124 commit 0ce3313cc995916da0b54adf440f4ec468298124 Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed Apr 04 07:25:10 2018 [oilpan] Enable heap verification for incremental marking Enables heap verification when running in the incremental marking configuration. No-try: true Bug: chromium:757440 Change-Id: I4bc75ac0779021e34d0e637192157695567e825c Reviewed-on: https://chromium-review.googlesource.com/993056 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#547998} [modify] https://crrev.com/0ce3313cc995916da0b54adf440f4ec468298124/third_party/WebKit/Source/platform/heap/BUILD.gn [modify] https://crrev.com/0ce3313cc995916da0b54adf440f4ec468298124/third_party/WebKit/Source/platform/heap/HeapPage.cpp
,
Apr 4 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/fdd33088beafe06beebd337e58e1fdc77f527c2f commit fdd33088beafe06beebd337e58e1fdc77f527c2f Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed Apr 04 13:25:01 2018 [oilpan] Faster write barrier guard Add a general counter for threads that are currently in incremental marking. This guard can be checked as first reason for bailing out in any write barrier as it only requires a single load and does not need to go through TLS or page lookups. Bug: chromium:757440 Change-Id: Ida63e560bb52ccd66da739139c0510229b05a550 Reviewed-on: https://chromium-review.googlesource.com/995272 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#548033} [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/Heap.cpp [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/Heap.h [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/HeapAllocator.h [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/HeapPage.cpp [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/HeapPage.h [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/IncrementalMarkingTest.cpp [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/Member.h [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/ThreadState.cpp [modify] https://crrev.com/fdd33088beafe06beebd337e58e1fdc77f527c2f/third_party/WebKit/Source/platform/heap/ThreadState.h
,
Apr 4 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/df7973c46ecc157e63f710177eac3cdb02bc089c commit df7973c46ecc157e63f710177eac3cdb02bc089c Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed Apr 04 20:42:32 2018 [oilpan] Avoid individual write barriers when swapping backing stores The backing store write barrier is enough. Bug: chromium:757440 Change-Id: Id7aa827b2777b4060d4a5c6e2ba0849095f50abd Reviewed-on: https://chromium-review.googlesource.com/995477 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#548184} [modify] https://crrev.com/df7973c46ecc157e63f710177eac3cdb02bc089c/third_party/WebKit/Source/platform/heap/IncrementalMarkingTest.cpp [modify] https://crrev.com/df7973c46ecc157e63f710177eac3cdb02bc089c/third_party/WebKit/Source/platform/wtf/HashTable.h [modify] https://crrev.com/df7973c46ecc157e63f710177eac3cdb02bc089c/third_party/WebKit/Source/platform/wtf/Vector.h
,
Apr 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/1e1a0e3002ef6767749cd38dbd2f90caca0c3978 commit 1e1a0e3002ef6767749cd38dbd2f90caca0c3978 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 05 13:09:53 2018 [oilpan] Incremental marking: DCHECK for deleted values Since df7973c46ecc157e63f710177eac3cdb02bc089c there is no way to trigger individual write barriers on deleted values, because: a. write barriers are emitted for complete backing stores and their processing already takes care of deleted values b. for moving and copying the reinsertion takes care of omitting deleted values Bug: chromium:757440 Change-Id: I43fb8f43936d471e7ad6d1fb7f0da6c8bd626e52 Reviewed-on: https://chromium-review.googlesource.com/992932 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#548399} [modify] https://crrev.com/1e1a0e3002ef6767749cd38dbd2f90caca0c3978/third_party/WebKit/Source/platform/heap/HeapAllocator.h [modify] https://crrev.com/1e1a0e3002ef6767749cd38dbd2f90caca0c3978/third_party/WebKit/Source/platform/heap/IncrementalMarkingTest.cpp [modify] https://crrev.com/1e1a0e3002ef6767749cd38dbd2f90caca0c3978/third_party/WebKit/Source/platform/heap/Visitor.h [modify] https://crrev.com/1e1a0e3002ef6767749cd38dbd2f90caca0c3978/third_party/WebKit/Source/platform/wtf/HashTraits.h [modify] https://crrev.com/1e1a0e3002ef6767749cd38dbd2f90caca0c3978/third_party/WebKit/Source/platform/wtf/VectorTraits.h
,
Apr 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/f13cd25fdd73ee6d0d87ea4e9e7d0565c0d7bdc8 commit f13cd25fdd73ee6d0d87ea4e9e7d0565c0d7bdc8 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 05 14:59:17 2018 [oilpan] Fix object start bitmap verification Verification requires iterability of pages implying that we cannot have a linear allocation area. The area was however potentially set up during compaction. Bug: chromium:757440 Change-Id: I9b12ecac032893aaa6d6f022149cec0a6130ebf9 Reviewed-on: https://chromium-review.googlesource.com/997672 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#548421} [modify] https://crrev.com/f13cd25fdd73ee6d0d87ea4e9e7d0565c0d7bdc8/third_party/WebKit/Source/platform/heap/HeapPage.cpp [modify] https://crrev.com/f13cd25fdd73ee6d0d87ea4e9e7d0565c0d7bdc8/third_party/WebKit/Source/platform/heap/HeapPage.h
,
Apr 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/a0c28b8a366c961f29e53565c068ef9ea37a0882 commit a0c28b8a366c961f29e53565c068ef9ea37a0882 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 05 22:07:03 2018 [wrapper-tracing] Fix write barrier test The test was manually hooking up a wrapper in a non-supported way. Switch to using the providing bindings methods. This was flushed out when incremental marking started scanning roots right after the test and encountered a broken leftover object. Bug: chromium:757440 Change-Id: I0636c48a22d88f2d85ded1ebab2732d3ef6de077 Reviewed-on: https://chromium-review.googlesource.com/997795 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#548581} [modify] https://crrev.com/a0c28b8a366c961f29e53565c068ef9ea37a0882/third_party/WebKit/Source/bindings/core/v8/ScriptWrappableMarkingVisitorTest.cpp [modify] https://crrev.com/a0c28b8a366c961f29e53565c068ef9ea37a0882/third_party/WebKit/Source/platform/bindings/V8DOMWrapper.h
,
Apr 6 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/73e7ae2870242ecb195b14318700e23e77df0945 commit 73e7ae2870242ecb195b14318700e23e77df0945 Author: Michael Lippautz <mlippautz@chromium.org> Date: Fri Apr 06 18:00:02 2018 [oilpan] Merge post marking processing into weak processing Post marking processing assumed that the backing store for weak HashTable's remains unchanged during marking which is not correct in the presence of incremental marking. Since post marking was just marking the header of the backing store this can be merged into the weak processing steps. This CL - Remove Heap::post_marking_worklist_ completely, avoid any memory consumption it might require. - Merges WeakProcessing and PostMarkingProcessing as the post marking steps were only used for weak HashTables. Bug: chromium:828537 , chromium:757440 Change-Id: I0d0b7552786659becf039419bff3701e7bb72aee Reviewed-on: https://chromium-review.googlesource.com/999485 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#548854} [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/Heap.cpp [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/Heap.h [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/HeapAllocator.h [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/IncrementalMarkingTest.cpp [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/MarkingVerifier.h [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/MarkingVisitor.cpp [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/MarkingVisitor.h [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/ThreadState.cpp [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/ThreadState.h [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/heap/Visitor.h [modify] https://crrev.com/73e7ae2870242ecb195b14318700e23e77df0945/third_party/WebKit/Source/platform/wtf/HashTable.h
,
Apr 7 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/321dd6ae9b61d9edcd761c654aeb0cff910cb27b commit 321dd6ae9b61d9edcd761c654aeb0cff910cb27b Author: Michael Lippautz <mlippautz@chromium.org> Date: Sat Apr 07 10:46:49 2018 [oilpan] Incremental marking: Transition into full garbage collection Avoid doing two garbage collections back to back. Bug: chromium:757440 Change-Id: I89933906fe055ed368d9d454ce69e4a0aa37ae6f Reviewed-on: https://chromium-review.googlesource.com/993133 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#549055} [modify] https://crrev.com/321dd6ae9b61d9edcd761c654aeb0cff910cb27b/third_party/WebKit/Source/platform/heap/ThreadState.cpp
,
Apr 7 2018
,
Apr 9 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/ff278139f629a991fe28c949cb210ef19b9d2997 commit ff278139f629a991fe28c949cb210ef19b9d2997 Author: Michael Lippautz <mlippautz@chromium.org> Date: Mon Apr 09 11:11:24 2018 [oilpan] Incremental marking: Force another GC for termination Bug: chromium:757440 Change-Id: I7b417e79ba205dc15a83f7ab768f1f47c1840221 Reviewed-on: https://chromium-review.googlesource.com/1002553 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#549149} [modify] https://crrev.com/ff278139f629a991fe28c949cb210ef19b9d2997/third_party/blink/renderer/platform/heap/thread_state.cc
,
Apr 10 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/2da3fcd0c9a8a6ec5504ddbf2293d5b9b8d6122c commit 2da3fcd0c9a8a6ec5504ddbf2293d5b9b8d6122c Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Apr 10 09:16:52 2018 [oilpan] Filter deleted weakcell slots during incremental marking Slots may be deleted after recording a callback for them. Filter out those slots accordingly during processing. Bug: chromium:757440 Change-Id: I23c17c2f79af41a8539b8d4734ec677fb0b99c0b Reviewed-on: https://chromium-review.googlesource.com/1000865 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#549465} [modify] https://crrev.com/2da3fcd0c9a8a6ec5504ddbf2293d5b9b8d6122c/third_party/blink/renderer/platform/heap/heap.h
,
Apr 10 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/b1708929d4af0ee675f4cdcc3782182b3651eddf commit b1708929d4af0ee675f4cdcc3782182b3651eddf Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Apr 10 11:09:16 2018 [oilpan] Reset allocation point before CompleteSweep Some phases require iterability of a page. This property is already established before LazySweep during OutOfLineAllocate but was missing in the case where we completed sweeping for *all* arenas. Bug: chromium:757440 Change-Id: I103b55a0d9dba11574ef56fae38db482f7fdad6a Reviewed-on: https://chromium-review.googlesource.com/1004581 Reviewed-by: Keishi Hattori <keishi@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#549482} [modify] https://crrev.com/b1708929d4af0ee675f4cdcc3782182b3651eddf/third_party/blink/renderer/platform/heap/heap_page.cc [modify] https://crrev.com/b1708929d4af0ee675f4cdcc3782182b3651eddf/third_party/blink/renderer/platform/heap/heap_page.h
,
Apr 11 2018
,
Apr 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e commit 4c835f2ce638c5a6804d11b5ae401e5ac2aa570e Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 12 19:27:06 2018 [oilpan] Create single entry for write barrier Similar to ScriptWrappableMarkingVisitor, thread all write barrier calls through MarkingVisitor::WriteBarrier. Drive-by: - Add missing write barrier to CSSRule. Bug: chromium:757440 Change-Id: Ibc9885ac0724ec157df9c7c314be2aa14555691b Reviewed-on: https://chromium-review.googlesource.com/1010204 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#550304} [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/core/css/css_rule.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/core/dom/node.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/heap.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/heap.h [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/heap_allocator.h [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/incremental_marking_test.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/marking_visitor.h [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/member.h
,
Apr 13 2018
,
Apr 17 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e commit 4c835f2ce638c5a6804d11b5ae401e5ac2aa570e Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 12 19:27:06 2018 [oilpan] Create single entry for write barrier Similar to ScriptWrappableMarkingVisitor, thread all write barrier calls through MarkingVisitor::WriteBarrier. Drive-by: - Add missing write barrier to CSSRule. Bug: chromium:757440 Change-Id: Ibc9885ac0724ec157df9c7c314be2aa14555691b Reviewed-on: https://chromium-review.googlesource.com/1010204 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#550304} [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/core/css/css_rule.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/core/dom/node.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/heap.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/heap.h [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/heap_allocator.h [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/incremental_marking_test.cc [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/marking_visitor.h [modify] https://crrev.com/4c835f2ce638c5a6804d11b5ae401e5ac2aa570e/third_party/blink/renderer/platform/heap/member.h
,
Apr 26 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/d48e4a5027d99df10556c28382948c31f419e815 commit d48e4a5027d99df10556c28382948c31f419e815 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 26 06:59:13 2018 [oilpan] Introduce incremental marking stress flag This frees up the regular HeapIncrementalMarking flag for implementing an initial incremental marking strategy. Bug: chromium:757440 Change-Id: Iee924ef718f450b4d1e865ef6880242f501376fb Reviewed-on: https://chromium-review.googlesource.com/1028273 Reviewed-by: John Budorick <jbudorick@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#553935} [modify] https://crrev.com/d48e4a5027d99df10556c28382948c31f419e815/testing/buildbot/chromium.fyi.json [modify] https://crrev.com/d48e4a5027d99df10556c28382948c31f419e815/testing/buildbot/test_suite_exceptions.pyl [modify] https://crrev.com/d48e4a5027d99df10556c28382948c31f419e815/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/d48e4a5027d99df10556c28382948c31f419e815/third_party/blink/renderer/platform/heap/thread_state.h [modify] https://crrev.com/d48e4a5027d99df10556c28382948c31f419e815/third_party/blink/renderer/platform/runtime_enabled_features.json5
,
Apr 26 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/6c2e15ce3aaf1b7eeb067fca773632c1a11aeff1 commit 6c2e15ce3aaf1b7eeb067fca773632c1a11aeff1 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 26 09:44:23 2018 [oilpan] Fix incremental marking tests Fix compilaton broken by stricter compile-time checks. No-try: true Tbr: haraken@chromium.org Bug: chromium:757440 Change-Id: Ib4444900e0fc68ab915683ec68d598aaf73a27fd Reviewed-on: https://chromium-review.googlesource.com/1029862 Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#553966} [modify] https://crrev.com/6c2e15ce3aaf1b7eeb067fca773632c1a11aeff1/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
Apr 26 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/062062c330db34f6d2f53e86242bb2c689712d86 commit 062062c330db34f6d2f53e86242bb2c689712d86 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 26 11:03:28 2018 [oilpan] Adjust layout test expectations for incremental marking The stressbot now runs with HeapIncrementalMarkingStress No-try: true Tbr: haraken@chromium.org Bug: chromium:757440 Change-Id: Ieb520b74417a635ff91ea9db2be7bcc6a875e19b Reviewed-on: https://chromium-review.googlesource.com/1030090 Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#553976} [rename] https://crrev.com/062062c330db34f6d2f53e86242bb2c689712d86/third_party/WebKit/LayoutTests/FlagExpectations/enable-blink-features=HeapIncrementalMarkingStress
,
Apr 26 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/544122c6009ae3e632e9346223716df3e83ee68d commit 544122c6009ae3e632e9346223716df3e83ee68d Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Apr 26 15:02:17 2018 [oilpan] Fix incremental marking write barrier test WriteBarrierDuringMixinConstruction assumed a specific layout of the underlying marking work lists. This is brittle as, depending on other build flags, containers be resized and rehashed differently, resulting in different objects on the work lists. Rework the test to only check a simple condition: whether the object in question can be found on the |not_fully_constructed_worklist|. Tbr: haraken@chromium.org No-try: true Bug: chromium:757440 Change-Id: I1377e2e727ad8844b0aefa88fb0084492f4bb526 Reviewed-on: https://chromium-review.googlesource.com/1030494 Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#554030} [modify] https://crrev.com/544122c6009ae3e632e9346223716df3e83ee68d/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
Apr 27 2018
,
May 2 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/d8a31d453aaa429829d6ba183f94b5ba016768bd commit d8a31d453aaa429829d6ba183f94b5ba016768bd Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed May 02 17:23:13 2018 [oilpan] Enable incremental marking buildflag Includes incremental garbage collection infrastructure in regular builds. Does not enable incremental marking at runtime! This CL may cause throughput regressions and is an attempt to collect a broad range of potential performance issues. Bug: chromium:757440 Change-Id: Ie7b75f8d96f9230400beed1a3258e45e0ca742ed Reviewed-on: https://chromium-review.googlesource.com/980753 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#555449} [modify] https://crrev.com/d8a31d453aaa429829d6ba183f94b5ba016768bd/third_party/blink/renderer/platform/heap/BUILD.gn
,
May 2 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/b1940e0ce70651ca8cb36787ed0a96ace34eace3 commit b1940e0ce70651ca8cb36787ed0a96ace34eace3 Author: Ned Nguyen <nednguyen@google.com> Date: Wed May 02 20:49:12 2018 Revert "[oilpan] Enable incremental marking buildflag" This reverts commit d8a31d453aaa429829d6ba183f94b5ba016768bd. Reason for revert: suspect breaking blink_heap_unittests BUG: chromium:839114 Original change's description: > [oilpan] Enable incremental marking buildflag > > Includes incremental garbage collection infrastructure in regular > builds. Does not enable incremental marking at runtime! > > This CL may cause throughput regressions and is an attempt > to collect a broad range of potential performance issues. > > Bug: chromium:757440 > Change-Id: Ie7b75f8d96f9230400beed1a3258e45e0ca742ed > Reviewed-on: https://chromium-review.googlesource.com/980753 > Commit-Queue: Michael Lippautz <mlippautz@chromium.org> > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Cr-Commit-Position: refs/heads/master@{#555449} TBR=haraken@chromium.org,keishi@chromium.org,hpayer@chromium.org,mlippautz@chromium.org Change-Id: I1d1cd812dcf24a9a87ad6100658e4c18979f29bd No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: chromium:757440 Reviewed-on: https://chromium-review.googlesource.com/1040688 Reviewed-by: Ned Nguyen <nednguyen@google.com> Commit-Queue: Ned Nguyen <nednguyen@google.com> Cr-Commit-Position: refs/heads/master@{#555504} [modify] https://crrev.com/b1940e0ce70651ca8cb36787ed0a96ace34eace3/third_party/blink/renderer/platform/heap/BUILD.gn
,
May 2 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/d1a23ee75d8ef824d53df708ea09017ebafa00e4 commit d1a23ee75d8ef824d53df708ea09017ebafa00e4 Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed May 02 21:11:20 2018 [oilpan] Fix estimated marking time computation Compute estimated marking time in PostSweep were both, marked bytes and marking time, refer to the same garbage collection cycle. Bug: chromium:830196 , chromium:757440 Change-Id: I80ea0c028ba7668f05b94b1acbe9b0c38afa801a Reviewed-on: https://chromium-review.googlesource.com/1039566 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#555531} [modify] https://crrev.com/d1a23ee75d8ef824d53df708ea09017ebafa00e4/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/d1a23ee75d8ef824d53df708ea09017ebafa00e4/third_party/blink/renderer/platform/heap/thread_state.h
,
May 3 2018
,
May 3 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/6524551518762533169e817cf3efe47eb4cbbd1a commit 6524551518762533169e817cf3efe47eb4cbbd1a Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu May 03 09:51:34 2018 [oilpan] Fix CFI in incremental marking tests Bug: chromium:839114 , chromium:757440 Change-Id: I9784179f8af19739a8c4398e3deea04611e12b09 Reviewed-on: https://chromium-review.googlesource.com/1041825 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#555684} [modify] https://crrev.com/6524551518762533169e817cf3efe47eb4cbbd1a/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
May 3 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/bd9ac13acc9d99a4f192b2734613545c187a319c commit bd9ac13acc9d99a4f192b2734613545c187a319c Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu May 03 12:12:31 2018 [oilpan] Enable incremental marking buildflag Includes incremental garbage collection infrastructure in regular builds. Does not enable incremental marking at runtime! This CL may cause throughput regressions and is an attempt to collect a broad range of potential performance issues. Bug: chromium:757440 Change-Id: If4d91be2eb325417736a38e27c481ab23ce7a0e9 Reviewed-on: https://chromium-review.googlesource.com/1041986 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#555698} [modify] https://crrev.com/bd9ac13acc9d99a4f192b2734613545c187a319c/third_party/blink/renderer/platform/heap/BUILD.gn
,
May 4 2018
,
May 7 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/faf462c0c6690b6433f20b0b7f28090d5277f0c5 commit faf462c0c6690b6433f20b0b7f28090d5277f0c5 Author: Michael Lippautz <mlippautz@chromium.org> Date: Mon May 07 15:05:11 2018 [oilpan] Move address cache to roots marking Address cache is only actively used during roots marking. For incremental marking we pass over roots multiple times, each time potentially requires clearing the cache. Drive-by: - Add unit tests - Rename to AddressCache - Encapsulate handling fully in AddressCache - Remove outdated caller Bug: chromium:757440 , chromium:830196 Change-Id: I51780e0bbbb8566ad0c83df98229e15133b42825 Reviewed-on: https://chromium-review.googlesource.com/1044214 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#556439} [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/BUILD.gn [add] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/address_cache.cc [add] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/address_cache.h [add] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/address_cache_test.cc [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/heap.cc [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/heap.h [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/heap_page.cc [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/heap_page.h [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/heap_test.cc [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/page_memory.h [modify] https://crrev.com/faf462c0c6690b6433f20b0b7f28090d5277f0c5/third_party/blink/renderer/platform/heap/thread_state.cc
,
May 7 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/e2cf7974aa9444877e0c9f71cff2b946310b83a7 commit e2cf7974aa9444877e0c9f71cff2b946310b83a7 Author: Michael Lippautz <mlippautz@chromium.org> Date: Mon May 07 15:35:56 2018 Revert "[oilpan] Enable incremental marking buildflag" This reverts commit bd9ac13acc9d99a4f192b2734613545c187a319c. Reason for revert: Unfortunately there's another V8 CL close which massively affects speedometer. We need to reland to get a clear picture. Original change's description: > [oilpan] Enable incremental marking buildflag > > Includes incremental garbage collection infrastructure in regular > builds. Does not enable incremental marking at runtime! > > This CL may cause throughput regressions and is an attempt > to collect a broad range of potential performance issues. > > Bug: chromium:757440 > Change-Id: If4d91be2eb325417736a38e27c481ab23ce7a0e9 > Reviewed-on: https://chromium-review.googlesource.com/1041986 > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Commit-Queue: Michael Lippautz <mlippautz@chromium.org> > Cr-Commit-Position: refs/heads/master@{#555698} TBR=haraken@chromium.org,keishi@chromium.org,hpayer@chromium.org,mlippautz@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: chromium:757440 Change-Id: I87dd96d9b9cd142503ff22534b8207da27b8c352 Reviewed-on: https://chromium-review.googlesource.com/1047265 Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#556450} [modify] https://crrev.com/e2cf7974aa9444877e0c9f71cff2b946310b83a7/third_party/blink/renderer/platform/heap/BUILD.gn
,
May 10 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/838fb3bd2db6a2869a0efc6e7e3154f954f74b70 commit 838fb3bd2db6a2869a0efc6e7e3154f954f74b70 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu May 10 17:51:18 2018 [oilpan] Enable incremental marking buildflag Includes incremental garbage collection infrastructure in regular builds. Does not enable incremental marking at runtime! Sheriffs: - This CL may cause throughput regressions and is an attempt to collect a broad range of potential performance issues. - Binary size regressions of up to 100kb are expected. The infra added here is the minimal base for any future incremental and concurrent marking changes. Long term we will save a bit of memory by removing trace methods. Bug: chromium:757440 Change-Id: I61f4dca6cb1933a9f10416bc6641aae3e8794455 Reviewed-on: https://chromium-review.googlesource.com/1052108 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#557576} [modify] https://crrev.com/838fb3bd2db6a2869a0efc6e7e3154f954f74b70/third_party/blink/renderer/platform/heap/BUILD.gn
,
May 14 2018
,
May 14 2018
,
May 14 2018
,
May 14 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/cc7faaf76c126b4a2900d7b56bda3d06d5cc50ed commit cc7faaf76c126b4a2900d7b56bda3d06d5cc50ed Author: Michael Lippautz <mlippautz@chromium.org> Date: Mon May 14 13:56:28 2018 [oilpan] Cleanup incremental marking tests Remove template parameter for observed objects, making it impossible to cast an marking worklist item to that type. This was a source of UBSAN errors in the tests. Bug: chromium:757440 Change-Id: I35843b1bcfd81c404e840790bb00a7911e2bc3aa Reviewed-on: https://chromium-review.googlesource.com/1057234 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#558289} [modify] https://crrev.com/cc7faaf76c126b4a2900d7b56bda3d06d5cc50ed/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
May 14 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/c6ff060cc760c7cb8a44be7314ce5e96757213c6 commit c6ff060cc760c7cb8a44be7314ce5e96757213c6 Author: Michael Lippautz <mlippautz@chromium.org> Date: Mon May 14 23:23:14 2018 [oilpan] Cleanup tracing scopes - Use ThreadHeapStatsCollector scopes instead of manual scopes - Allow adding parameters - Allow enabling by default - Fix telemetry unit tests Bug: chromium:840789, chromium:757440 Change-Id: If291b08d989d11233a31a4fbfc28db0b2f594e2b Reviewed-on: https://chromium-review.googlesource.com/1057626 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#558522} [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/third_party/blink/renderer/platform/heap/heap_page.cc [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/third_party/blink/renderer/platform/heap/heap_stats_collector.cc [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/third_party/blink/renderer/platform/heap/heap_stats_collector.h [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/third_party/blink/renderer/platform/heap/heap_stats_collector_test.cc [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/tools/perf/contrib/oilpan/oilpan_gc_times.py [modify] https://crrev.com/c6ff060cc760c7cb8a44be7314ce5e96757213c6/tools/perf/contrib/oilpan/oilpan_gc_times_unittest.py
,
May 15 2018
,
May 15 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/984762d2100f6aff3dacb8507595b813e68e96a5 commit 984762d2100f6aff3dacb8507595b813e68e96a5 Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue May 15 15:33:20 2018 [oilpan,wtf] Rework HashTable write barriers Instead of emitting write barriers for single writes in HashTable, conservatively emit a barrier for the backing store during rehashing. This moves us in the tradeoff space: - Removes 40KiB of binary bloat - Avoids emitting barriers when re-inserting single elements which improves throughput for the case where incremental marking is off - Emits a single conservative barrier at the end of RehashTo. This potentially means that we (re-)scan more objects when incremental marking is on Bug: chromium:839514 , chromium:757440 Change-Id: Ib5495c9e2210836424d7b26f2138d852f8dafec7 Reviewed-on: https://chromium-review.googlesource.com/1055400 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#558713} [modify] https://crrev.com/984762d2100f6aff3dacb8507595b813e68e96a5/third_party/blink/renderer/platform/heap/heap_allocator.h [modify] https://crrev.com/984762d2100f6aff3dacb8507595b813e68e96a5/third_party/blink/renderer/platform/heap/marking_visitor.h [modify] https://crrev.com/984762d2100f6aff3dacb8507595b813e68e96a5/third_party/blink/renderer/platform/wtf/allocator/partition_allocator.h [modify] https://crrev.com/984762d2100f6aff3dacb8507595b813e68e96a5/third_party/blink/renderer/platform/wtf/hash_table.h
,
May 19 2018
,
May 21 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/15c2fb6c240000
,
May 21 2018
📍 Job complete. See results below. https://pinpoint-dot-chromeperf.appspot.com/job/15c2fb6c240000
,
May 21 2018
,
May 23 2018
,
May 24 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/4024aad376eea8ef01d693a89dab09585b2a4f9e commit 4024aad376eea8ef01d693a89dab09585b2a4f9e Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu May 24 08:33:41 2018 [oilpan] Split write barrier in fast and slow parts Split write barrier into fast and slow parts: - The fast part only checks whether any Oilpan heap is currently marking. This check is only approximate as it does not consider the current ThreadState. In general, we expect this check to be enough to allow bailing out of the barrier. - The slow version checks whether the current ThreadState is marking and whether the values actually require a write barrier. This way we emit only a short instruction sequence for the fast cases and avoid poluting the regular instruction sequences. Verified locally on the microbenchmark blink_perf.parser query-selector-deep which showed a 42% regression. Scores (higher is better): - ToT: 8932 - Without barrier: 15188 - With this CL: 13352 Bug: chromium:844576 , chromium:757440 Change-Id: Ie8ebbf95fef0ff59ad8f1a111dd5daecfabc4109 Reviewed-on: https://chromium-review.googlesource.com/1071272 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#561424} [modify] https://crrev.com/4024aad376eea8ef01d693a89dab09585b2a4f9e/third_party/blink/renderer/platform/heap/marking_visitor.cc [modify] https://crrev.com/4024aad376eea8ef01d693a89dab09585b2a4f9e/third_party/blink/renderer/platform/heap/marking_visitor.h [modify] https://crrev.com/4024aad376eea8ef01d693a89dab09585b2a4f9e/third_party/blink/renderer/platform/heap/member.h [modify] https://crrev.com/4024aad376eea8ef01d693a89dab09585b2a4f9e/third_party/blink/renderer/platform/heap/thread_state.h
,
May 24 2018
,
May 28 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/943547ec9faefe687f67244d36ae1b59e0e79697 commit 943547ec9faefe687f67244d36ae1b59e0e79697 Author: Keishi Hattori <keishi@chromium.org> Date: Mon May 28 07:09:27 2018 Oilpan: Implement IncrementalIdleGC scheduler behind HeapIncrementalMarking flag Schedules an IncrementalIdleGC instead of an IdleGC when the HeapIncrementalMarking runtime flag is enabled. Added GCReason kIncrementalIdleGC so we can distinguish them in UMA. IncrementalMarking tasks will be posted and executed as normal tasks, but if the reason is kTesting (in stress mode or unit test), they will be executed from RunScheduledGC as before. ThreadState::RunIncrementalMarking*Task() has been added to support tasks canceled after posting. Bug: 757440 Change-Id: Id305620facb0e7aa26561e9b6ebd3ef4db8b8f34 Reviewed-on: https://chromium-review.googlesource.com/1065719 Reviewed-by: Alexander Timin <altimin@chromium.org> Reviewed-by: Brian White <bcwhite@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Hajime Hoshi <hajimehoshi@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Commit-Queue: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#562198} [modify] https://crrev.com/943547ec9faefe687f67244d36ae1b59e0e79697/third_party/blink/renderer/platform/heap/blink_gc.h [modify] https://crrev.com/943547ec9faefe687f67244d36ae1b59e0e79697/third_party/blink/renderer/platform/heap/incremental_marking_test.cc [modify] https://crrev.com/943547ec9faefe687f67244d36ae1b59e0e79697/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/943547ec9faefe687f67244d36ae1b59e0e79697/third_party/blink/renderer/platform/heap/thread_state.h [modify] https://crrev.com/943547ec9faefe687f67244d36ae1b59e0e79697/tools/metrics/histograms/enums.xml [modify] https://crrev.com/943547ec9faefe687f67244d36ae1b59e0e79697/tools/metrics/histograms/histograms.xml
,
May 29 2018
,
May 29 2018
,
May 30 2018
,
May 30 2018
,
May 30 2018
,
May 30 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3 commit ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3 Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed May 30 10:41:42 2018 Add chrome://flags entry for Blink incremental marking - Adds chrome://flags entry - Adds a FeatureFlag - Make existing Blink flag dependent on the Chrome flag Bug: chromium:757440 , chromium:847735 Change-Id: Iccb63fc31b330a6bb69e95d5c0d7599cce10e2ee Reviewed-on: https://chromium-review.googlesource.com/1077870 Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Kinuko Yasuda <kinuko@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#562792} [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/chrome/browser/about_flags.cc [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/chrome/browser/flag_descriptions.cc [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/chrome/browser/flag_descriptions.h [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/content/child/runtime_features.cc [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/content/public/common/content_features.cc [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/content/public/common/content_features.h [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/third_party/blink/public/platform/web_runtime_features.h [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/third_party/blink/renderer/platform/exported/web_runtime_features.cc [modify] https://crrev.com/ebbc5a8d60fa8e1d9cbe9e4176c9206e03b26eb3/tools/metrics/histograms/enums.xml
,
May 31 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/50ff9ebe68493b7cdb512bb688b1db9f6e69e27b commit 50ff9ebe68493b7cdb512bb688b1db9f6e69e27b Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu May 31 00:09:10 2018 [oilpan] Factor out common parts of the atomic phase In a follow up, we will make CollectGarbage the bottleneck for the GC atomic phase. Bug: chromium:757440 Change-Id: Icf27cfde4a61375eda52207e967c84c4ed9c7c31 Reviewed-on: https://chromium-review.googlesource.com/1078214 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#563077} [modify] https://crrev.com/50ff9ebe68493b7cdb512bb688b1db9f6e69e27b/third_party/blink/renderer/platform/heap/heap_stats_collector.h [modify] https://crrev.com/50ff9ebe68493b7cdb512bb688b1db9f6e69e27b/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/50ff9ebe68493b7cdb512bb688b1db9f6e69e27b/third_party/blink/renderer/platform/heap/thread_state.h
,
May 31 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/14dd8136240000
,
May 31 2018
📍 Job complete. See results below. https://pinpoint-dot-chromeperf.appspot.com/job/14dd8136240000
,
May 31 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/11f5f156240000
,
May 31 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/14b7c57a240000
,
May 31 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/12dd8136240000
,
May 31 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/14d8df56240000
,
May 31 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/14a6c7a6240000
,
May 31 2018
😿 Pinpoint job stopped with an error. https://pinpoint-dot-chromeperf.appspot.com/job/11f5f156240000 Buildbucket says the build completed successfully, but Pinpoint can't find the isolate hash.
,
May 31 2018
😿 Pinpoint job stopped with an error. https://pinpoint-dot-chromeperf.appspot.com/job/14b7c57a240000 Buildbucket says the build completed successfully, but Pinpoint can't find the isolate hash.
,
May 31 2018
😿 Pinpoint job stopped with an error. https://pinpoint-dot-chromeperf.appspot.com/job/12dd8136240000 Buildbucket says the build completed successfully, but Pinpoint can't find the isolate hash.
,
May 31 2018
😿 Pinpoint job stopped with an error. https://pinpoint-dot-chromeperf.appspot.com/job/14a6c7a6240000 Buildbucket says the build completed successfully, but Pinpoint can't find the isolate hash.
,
May 31 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/32787e23a00ccb7f62a9162b87afbaead91521de commit 32787e23a00ccb7f62a9162b87afbaead91521de Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu May 31 13:30:02 2018 [oilpan] Call into CollectGarbage from incremental finalization This avoids having multiple code paths that can finalize garbage collection by calling the regular bottleneck from the incremental finalization. The only adjustment needed is for MarkPhasePrologue which in this case already executed a minimal version upon starting incremental marking. Bug: chromium:757440 , chromium:830196 Change-Id: I9380a8c867e4b32aaf3c73eecae8391eda8e2853 Reviewed-on: https://chromium-review.googlesource.com/1080347 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#563209} [modify] https://crrev.com/32787e23a00ccb7f62a9162b87afbaead91521de/third_party/blink/renderer/platform/heap/heap_test.cc [modify] https://crrev.com/32787e23a00ccb7f62a9162b87afbaead91521de/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/32787e23a00ccb7f62a9162b87afbaead91521de/third_party/blink/renderer/platform/heap/thread_state.h
,
Jun 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/fdd98637aba24a7fcaabea785e85252419e4d7a3 commit fdd98637aba24a7fcaabea785e85252419e4d7a3 Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Jun 05 02:40:05 2018 [oilpan] Update collection rate histograms for all GC reasons Bug: chromium:840789, chromium:757440 Change-Id: I9841524427dcb072ece75888f928b3e09bbf368c Reviewed-on: https://chromium-review.googlesource.com/1086430 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#564353} [modify] https://crrev.com/fdd98637aba24a7fcaabea785e85252419e4d7a3/third_party/blink/renderer/platform/heap/thread_state.cc
,
Jun 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/3860f3e4d199e934052b3486a226dde44fb22e9a commit 3860f3e4d199e934052b3486a226dde44fb22e9a Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Jun 05 05:11:12 2018 [oilpan] Use atomic phase marking time for BlinkGC.CollectGarbage The histogram will be deprecated soon. Up until then it should use the atomic phase marking time though and not the sum of all marking time. No-try: true Bug: chromium:840789, chromium:757440 Change-Id: Ic425ba4538ec7d32b4055f9f5c905d387711a637 Reviewed-on: https://chromium-review.googlesource.com/1086559 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#564362} [modify] https://crrev.com/3860f3e4d199e934052b3486a226dde44fb22e9a/third_party/blink/renderer/platform/heap/thread_state.cc
,
Jun 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/d58b508a28d4b5beb82e6931a2b78080900d32fc commit d58b508a28d4b5beb82e6931a2b78080900d32fc Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Jun 05 05:21:12 2018 [oilpan] Update GC reason in stats collection This reason is reported to UMA. No-try: true Bug: chromium:840789, chromium:757440 Change-Id: Ie2000aa96b34e13e8280b08c7e7c7931f3efe216 Reviewed-on: https://chromium-review.googlesource.com/1086428 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#564364} [modify] https://crrev.com/d58b508a28d4b5beb82e6931a2b78080900d32fc/third_party/blink/renderer/platform/heap/heap_stats_collector.cc [modify] https://crrev.com/d58b508a28d4b5beb82e6931a2b78080900d32fc/third_party/blink/renderer/platform/heap/heap_stats_collector.h [modify] https://crrev.com/d58b508a28d4b5beb82e6931a2b78080900d32fc/third_party/blink/renderer/platform/heap/heap_stats_collector_test.cc [modify] https://crrev.com/d58b508a28d4b5beb82e6931a2b78080900d32fc/third_party/blink/renderer/platform/heap/thread_state.cc
,
Jun 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/09f81d9f83fc84bf05fabc870a57d3451dc88590 commit 09f81d9f83fc84bf05fabc870a57d3451dc88590 Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Jun 05 08:45:00 2018 [oilpan] Add new BlinkGC UMA counters Histgram CL: https://chromium-review.googlesource.com/c/chromium/src/+/1086432 Bug: chromium:840789, chromium:757440 Change-Id: Ie2b3510ac774ba8ac66ee9f11dc8aa8dd22ac677 Reviewed-on: https://chromium-review.googlesource.com/1086809 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#564411} [modify] https://crrev.com/09f81d9f83fc84bf05fabc870a57d3451dc88590/third_party/blink/renderer/platform/heap/thread_state.cc
,
Jun 5 2018
,
Jun 7 2018
,
Jun 7 2018
,
Jun 7 2018
,
Jun 7 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/2a198bd391edea387150f5040c4bdf78e683faa6 commit 2a198bd391edea387150f5040c4bdf78e683faa6 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Jun 07 00:36:04 2018 histograms: Update BlinkGC histograms - Fix labels of GCReason suffixes - Deprecate BlinkGC.CollectGarbage in favor of more exact buckets. See histograms.xml - Introduce BlinkGC.AtomicPhaseMarking for the atomic marking phase - Apply suffixes to interesting buckets to get exact pause times per GC reason No-try: true Bug: chromium:840789, chromium:757440 , chromium:847735 Change-Id: I611fea498c925c70286e44d10a885aa769552bcb Reviewed-on: https://chromium-review.googlesource.com/1086432 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Mark Pearson <mpearson@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#565123} [modify] https://crrev.com/2a198bd391edea387150f5040c4bdf78e683faa6/tools/metrics/histograms/histograms.xml
,
Jun 13 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/0db490b65e9f828bf6dca20edc26c343431e9510 commit 0db490b65e9f828bf6dca20edc26c343431e9510 Author: Keishi Hattori <keishi@chromium.org> Date: Wed Jun 13 07:07:51 2018 Oilpan: Introduce IncrementalV8FollowupGC Add new IncrementalV8FollowupGC. It is a incremental marking enabled version of PreciseGC. Bug: 757440 Change-Id: I1b480eac30936a8ad9b56a88b2191b0d417a61e7 Reviewed-on: https://chromium-review.googlesource.com/1096798 Commit-Queue: Keishi Hattori <keishi@chromium.org> Reviewed-by: Brian White <bcwhite@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#566741} [modify] https://crrev.com/0db490b65e9f828bf6dca20edc26c343431e9510/third_party/blink/renderer/platform/heap/blink_gc.h [modify] https://crrev.com/0db490b65e9f828bf6dca20edc26c343431e9510/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/0db490b65e9f828bf6dca20edc26c343431e9510/third_party/blink/renderer/platform/heap/thread_state.h [modify] https://crrev.com/0db490b65e9f828bf6dca20edc26c343431e9510/third_party/blink/renderer/platform/heap/thread_state_scheduling_test.cc [modify] https://crrev.com/0db490b65e9f828bf6dca20edc26c343431e9510/tools/metrics/histograms/enums.xml [modify] https://crrev.com/0db490b65e9f828bf6dca20edc26c343431e9510/tools/metrics/histograms/histograms.xml
,
Jun 15 2018
,
Jun 15 2018
,
Jun 15 2018
,
Jun 15 2018
,
Jun 15 2018
,
Jun 15 2018
,
Jun 15 2018
,
Jul 3
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/96eb702eb19acebd1f792436749e3021e0b8f54c commit 96eb702eb19acebd1f792436749e3021e0b8f54c Author: Keishi Hattori <keishi@chromium.org> Date: Tue Jul 03 14:45:47 2018 Oilpan: Update incremental marking step duration dynamically Calls UpdateIncrementalMarkingStepDuration() every time we allocate a new page. UpdateIncrementalMarkingStepDuration() will check if the estimated time left is increasing (i.e. marking speed is not keeping up with allocations). If so it will double the step size. Also increases default duration from 1ms to 2ms. Bug: 757440 Change-Id: Icf262a4b851fd730698e6b09384c5c181c671de2 Reviewed-on: https://chromium-review.googlesource.com/1105695 Commit-Queue: Keishi Hattori <keishi@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#572208} [modify] https://crrev.com/96eb702eb19acebd1f792436749e3021e0b8f54c/third_party/blink/renderer/platform/heap/heap_stats_collector.h [modify] https://crrev.com/96eb702eb19acebd1f792436749e3021e0b8f54c/third_party/blink/renderer/platform/heap/thread_state.cc [modify] https://crrev.com/96eb702eb19acebd1f792436749e3021e0b8f54c/third_party/blink/renderer/platform/heap/thread_state.h
,
Jul 4
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/4e34ccd5643183a799de377d9eaba175835ea659 commit 4e34ccd5643183a799de377d9eaba175835ea659 Author: Keishi Hattori <keishi@chromium.org> Date: Wed Jul 04 02:30:10 2018 Oilpan: Add reason arg to IncrementalMarkingStart trace event Add reason arg to IncrementalMarkingStart trace event so we can see what trigger the incremental marking GC. Bug: 757440 Change-Id: Ibecb88c157dec70e44287a146f93cdf7e5ef6211 Reviewed-on: https://chromium-review.googlesource.com/1124342 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#572452} [modify] https://crrev.com/4e34ccd5643183a799de377d9eaba175835ea659/third_party/blink/renderer/platform/heap/thread_state.cc
,
Jul 4
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/c225e90eeb77271d63d7e6299c1ff9948839891f commit c225e90eeb77271d63d7e6299c1ff9948839891f Author: Keishi Hattori <keishi@chromium.org> Date: Wed Jul 04 13:29:48 2018 Oilpan: Enable IncrementalV8FollowupGC behind the HeapIncrementalMarking flag Schedules an kIncrementalV8FollowupGC instead of an PreciseGC when the HeapIncrementalMarking runtime flag is enabled. Bug: 757440 Change-Id: I1d05afcc69508592d1e5f411dea1a3afa950baca Reviewed-on: https://chromium-review.googlesource.com/1095200 Commit-Queue: Keishi Hattori <keishi@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#572549} [modify] https://crrev.com/c225e90eeb77271d63d7e6299c1ff9948839891f/third_party/blink/renderer/platform/heap/thread_state.cc
,
Aug 1
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/d92ab8be365f7c52d1ead03f2d2f3a75510957f3 commit d92ab8be365f7c52d1ead03f2d2f3a75510957f3 Author: Keishi Hattori <keishi@chromium.org> Date: Wed Aug 01 10:08:20 2018 Oilpan: Enable incremental marking by default PSA: https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/9YCeDToqDU8 Bug: 757440 Change-Id: Id13ec3a8aa0c6a134a83acffd9811e2eee99fd2b Reviewed-on: https://chromium-review.googlesource.com/1125892 Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#579746} [modify] https://crrev.com/d92ab8be365f7c52d1ead03f2d2f3a75510957f3/third_party/blink/renderer/platform/runtime_enabled_features.json5
,
Aug 3
,
Aug 3
,
Aug 4
,
Aug 6
,
Aug 6
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/632f6b70acb4754d8fd6fc674b15cad43cf43078 commit 632f6b70acb4754d8fd6fc674b15cad43cf43078 Author: Keishi Hattori <keishi@chromium.org> Date: Mon Aug 06 12:06:13 2018 Revert "Oilpan: Enable incremental marking by default" This reverts commit d92ab8be365f7c52d1ead03f2d2f3a75510957f3. Reason for revert: causing chromium.linux bots to fail Original change's description: > Oilpan: Enable incremental marking by default > > PSA: https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/9YCeDToqDU8 > > Bug: 757440 > Change-Id: Id13ec3a8aa0c6a134a83acffd9811e2eee99fd2b > Reviewed-on: https://chromium-review.googlesource.com/1125892 > Reviewed-by: Michael Lippautz <mlippautz@chromium.org> > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Commit-Queue: Keishi Hattori <keishi@chromium.org> > Cr-Commit-Position: refs/heads/master@{#579746} TBR=haraken@chromium.org,keishi@chromium.org,mlippautz@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: 757440 Change-Id: I997d18b640ce20d9db5ba7ca8ce5928f4dfe71c4 Reviewed-on: https://chromium-review.googlesource.com/1163421 Reviewed-by: Keishi Hattori <keishi@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Commit-Queue: Keishi Hattori <keishi@chromium.org> Cr-Commit-Position: refs/heads/master@{#580849} [modify] https://crrev.com/632f6b70acb4754d8fd6fc674b15cad43cf43078/third_party/blink/renderer/platform/runtime_enabled_features.json5
,
Aug 7
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/1638e067d7dccd748c024dcdf763a63a943c9656 commit 1638e067d7dccd748c024dcdf763a63a943c9656 Author: Michael Lippautz <mlippautz@chromium.org> Date: Tue Aug 07 15:58:10 2018 [heap] Do not revive WeakMember during incremental marking Overwriting deleted values with cleared values potentially revives buckets in collections. For atomic garbage collections this is not a problem as those values are filered out in the visitors already. For incremental garbage collections the deleted values may be introduced by erasing a value from a container after it has been visited by the incremental marker. Bug: chromium:870196 , chromium:757440 Change-Id: I1d628f00c2e2ea6e36e880d50a397dcbcc8566d1 Reviewed-on: https://chromium-review.googlesource.com/1164950 Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#581238} [modify] https://crrev.com/1638e067d7dccd748c024dcdf763a63a943c9656/third_party/blink/renderer/platform/heap/heap.h [modify] https://crrev.com/1638e067d7dccd748c024dcdf763a63a943c9656/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
Aug 7
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/460c4724a3221dfb0b438fa619b41ec4022d2d38 commit 460c4724a3221dfb0b438fa619b41ec4022d2d38 Author: Fredrik Hubinette <hubbe@chromium.org> Date: Tue Aug 07 18:04:18 2018 Revert "[heap] Do not revive WeakMember during incremental marking" This reverts commit 1638e067d7dccd748c024dcdf763a63a943c9656. Reason for revert: New tests are flaky and causing problems when submitting code. Original change's description: > [heap] Do not revive WeakMember during incremental marking > > Overwriting deleted values with cleared values potentially revives > buckets in collections. For atomic garbage collections this is not a > problem as those values are filered out in the visitors already. For > incremental garbage collections the deleted values may be introduced by > erasing a value from a container after it has been visited by the > incremental marker. > > Bug: chromium:870196 , chromium:757440 > Change-Id: I1d628f00c2e2ea6e36e880d50a397dcbcc8566d1 > Reviewed-on: https://chromium-review.googlesource.com/1164950 > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Reviewed-by: Keishi Hattori <keishi@chromium.org> > Commit-Queue: Michael Lippautz <mlippautz@chromium.org> > Cr-Commit-Position: refs/heads/master@{#581238} TBR=haraken@chromium.org,keishi@chromium.org,mlippautz@chromium.org Change-Id: I632b5594d137acd37b4639ebcd46f8d61638dbc0 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: chromium:870196 , chromium:757440 Reviewed-on: https://chromium-review.googlesource.com/1165328 Reviewed-by: Fredrik Hubinette <hubbe@chromium.org> Commit-Queue: Fredrik Hubinette <hubbe@chromium.org> Cr-Commit-Position: refs/heads/master@{#581277} [modify] https://crrev.com/460c4724a3221dfb0b438fa619b41ec4022d2d38/third_party/blink/renderer/platform/heap/heap.h [modify] https://crrev.com/460c4724a3221dfb0b438fa619b41ec4022d2d38/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
Aug 8
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/3972c07ca0793f1f0a8d84ed79b43bd1faf42ea1 commit 3972c07ca0793f1f0a8d84ed79b43bd1faf42ea1 Author: Michael Lippautz <mlippautz@chromium.org> Date: Wed Aug 08 07:33:18 2018 Reland "[heap] Do not revive WeakMember during incremental marking" Overwriting deleted values with cleared values potentially revives buckets in collections. For atomic garbage collections this is not a problem as those values are filered out in the visitors already. For incremental garbage collections the deleted values may be introduced by erasing a value from a container after it has been visited by the incremental marker. This reverts commit 460c4724a3221dfb0b438fa619b41ec4022d2d38. Bug: chromium:870196 , chromium:757440 Change-Id: I2d4c173a2dc94b3b8615699f81b85e1b61c4c705 Reviewed-on: https://chromium-review.googlesource.com/1165556 Reviewed-by: Kentaro Hara <haraken@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#581494} [modify] https://crrev.com/3972c07ca0793f1f0a8d84ed79b43bd1faf42ea1/third_party/blink/renderer/platform/heap/heap.h [modify] https://crrev.com/3972c07ca0793f1f0a8d84ed79b43bd1faf42ea1/third_party/blink/renderer/platform/heap/incremental_marking_test.cc
,
Aug 9
,
Aug 9
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/f3e02276ecbbaf87c8c5143a6361582faff3ca33 commit f3e02276ecbbaf87c8c5143a6361582faff3ca33 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Aug 09 11:30:03 2018 [heap] Incremental marking: Do not promptly free HashTable backings Previously we checked whether a backing could be freed during incremental marking using the provided Traits and kWeakHandlingFlag. It turns out that WeakMember may also be part of otherwise strongly treated in-place constructed objects, e.g., std::pair<WeakMember<T>,U>. Completely turn off promptly freeing for HashTable backings during incremental marking. Bug: chromium:870306 , chromium:757440 Change-Id: I94aacad0dbba707a4e0fb8c8119591fd413f9314 Reviewed-on: https://chromium-review.googlesource.com/1165353 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#581865} [modify] https://crrev.com/f3e02276ecbbaf87c8c5143a6361582faff3ca33/third_party/blink/renderer/platform/heap/heap_allocator.cc [modify] https://crrev.com/f3e02276ecbbaf87c8c5143a6361582faff3ca33/third_party/blink/renderer/platform/heap/heap_allocator.h [modify] https://crrev.com/f3e02276ecbbaf87c8c5143a6361582faff3ca33/third_party/blink/renderer/platform/heap/incremental_marking_test.cc [modify] https://crrev.com/f3e02276ecbbaf87c8c5143a6361582faff3ca33/third_party/blink/renderer/platform/wtf/allocator/partition_allocator.cc [modify] https://crrev.com/f3e02276ecbbaf87c8c5143a6361582faff3ca33/third_party/blink/renderer/platform/wtf/allocator/partition_allocator.h [modify] https://crrev.com/f3e02276ecbbaf87c8c5143a6361582faff3ca33/third_party/blink/renderer/platform/wtf/hash_table.h
,
Aug 10
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/f789a9453217467376d32c4d20b67501e48f3129 commit f789a9453217467376d32c4d20b67501e48f3129 Author: Michael Lippautz <mlippautz@chromium.org> Date: Fri Aug 10 11:29:10 2018 Oilpan: Enable incremental marking per default PSA: https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/9YCeDToqDU8 Bug: chromium:757440 Change-Id: Ie53f3a5f68ace78bc47828cfb441d5d0a53fd9aa Reviewed-on: https://chromium-review.googlesource.com/1170692 Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Keishi Hattori <keishi@chromium.org> Reviewed-by: Hannes Payer <hpayer@chromium.org> Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#582117} [modify] https://crrev.com/f789a9453217467376d32c4d20b67501e48f3129/third_party/blink/renderer/platform/runtime_enabled_features.json5
,
Oct 18
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/cdf14f5ad9f7f77d94b841416590a8e2d50c2f80 commit cdf14f5ad9f7f77d94b841416590a8e2d50c2f80 Author: Michael Lippautz <mlippautz@chromium.org> Date: Thu Oct 18 08:10:38 2018 [heap] Remove Blink incremental marking flag from chrome://flags There should be no reason to override this from the UI anymore. Also fix flag toggle. The toggle was only working from 'default false' -> 'enabled' but could not be used to explicitly disable the feature. Tbr: jochen@chromium.org Bug: 757440 Change-Id: I732c508503692437f9780d745f43d9bc4f09a611 Reviewed-on: https://chromium-review.googlesource.com/c/1286857 Commit-Queue: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#600687} [modify] https://crrev.com/cdf14f5ad9f7f77d94b841416590a8e2d50c2f80/chrome/browser/about_flags.cc [modify] https://crrev.com/cdf14f5ad9f7f77d94b841416590a8e2d50c2f80/chrome/browser/flag_descriptions.cc [modify] https://crrev.com/cdf14f5ad9f7f77d94b841416590a8e2d50c2f80/chrome/browser/flag_descriptions.h [modify] https://crrev.com/cdf14f5ad9f7f77d94b841416590a8e2d50c2f80/content/child/runtime_features.cc [modify] https://crrev.com/cdf14f5ad9f7f77d94b841416590a8e2d50c2f80/content/public/common/content_features.cc
,
Nov 30
This launched in M70. Wohoo! Any bugfixes or further improvements should get their own bugs.
,
Nov 30
Showing comments 85 - 184
of 184
Older ›
|
||||||||||||||||||||||||||||||||||||||
►
Sign in to add a comment |
||||||||||||||||||||||||||||||||||||||
