GitLab

Revert of Simple BlinkGC heap compaction. (patchset #22 id:420001 of https://codereview.chromium.org/2531973002/ )

Reason for revert:
Speculative revert for a layout test breakage. https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11/builds/13912

Original issue's description:
> Simple BlinkGC heap compaction.
>
> This implements heap compaction for the Blink GC infrastructure
> (Oilpan), compacting the arenas of the BlinkGC heap which are most
> susceptible to becoming fragmented during actual use.
>
> Fragmentation is a real problem and a growing one while browsing anything
> but static pages: the amount of unused, but allocated, memory is
> fluctuating higher over time.
>
> To avoid leaving increasing amounts of unused holes in our heaps,
> heap compaction will periodically squeeze out the unused portions,
> packing together the live objects. The heap pages that are then
> left as unused, are subsequently released and returned to the OS.
>
> Due to a fortunate property of Blink heap collection types, providing
> such compaction is within relatively easy reach. Experiments show that
> the arenas which hold such collection objects ("backing stores") are
> the ones that develop fragmentation the most & persistently. While not
> a complete heap compactor of all Blink GC arenas, it addresses the
> fragmentation problem where it is most pressing. More can be done, later.
>
> Explainer / design document:
>
>  https://docs.google.com/document/d/1k-vivOinomDXnScw8Ew5zpsYCXiYqj76OCOYZSvHkaU
>
> R=haraken
> BUG=672030
>
> Review-Url: https://codereview.chromium.org/2531973002

TBR=oilpan-reviews@chromium.org,haraken@chromium.org,sigbjornf@opera.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=672030

Review-Url: https://codereview.chromium.org/2570483002
Cr-Commit-Position: refs/heads/master@{#437834}

Skip presubmit warning when the introduced useless forward
declaration is made in third_party (with the exception of
blink). Take the opportunity to fix one test and style.

BUG=662195
TEST=PRESUBMIT_test.py ForwardDeclarationTest

Review-Url: https://codereview.chromium.org/2568473002
Cr-Commit-Position: refs/heads/master@{#437833}

Cr-Commit-Position: refs/heads/master@{#437832}

This patch adds a FlakyTest annotation for testAddInvalidAddressAndCancel

BUG=673136
TBR=dfalcantara

Review-Url: https://codereview.chromium.org/2566923003
Cr-Commit-Position: refs/heads/master@{#437831}

If the crash still occurs and the CHECKs pass, we know the null
active version is due to SWRegistration::DeleteVersion.

BUG=655910

Review-Url: https://codereview.chromium.org/2570453002
Cr-Commit-Position: refs/heads/master@{#437830}

This implements heap compaction for the Blink GC infrastructure
(Oilpan), compacting the arenas of the BlinkGC heap which are most
susceptible to becoming fragmented during actual use.

Fragmentation is a real problem and a growing one while browsing anything
but static pages: the amount of unused, but allocated, memory is
fluctuating higher over time.

To avoid leaving increasing amounts of unused holes in our heaps,
heap compaction will periodically squeeze out the unused portions,
packing together the live objects. The heap pages that are then
left as unused, are subsequently released and returned to the OS.

Due to a fortunate property of Blink heap collection types, providing
such compaction is within relatively easy reach. Experiments show that
the arenas which hold such collection objects ("backing stores") are
the ones that develop fragmentation the most & persistently. While not
a complete heap compactor of all Blink GC arenas, it addresses the
fragmentation problem where it is most pressing. More can be done, later.

Explainer / design document:

 https://docs.google.com/document/d/1k-vivOinomDXnScw8Ew5zpsYCXiYqj76OCOYZSvHkaU

R=haraken
BUG=672030

Review-Url: https://codereview.chromium.org/2531973002
Cr-Commit-Position: refs/heads/master@{#437829}

This patch is an initial try to setup call flows and produces fragment
tree for inline layout, a step towards to NG inline layout to produce
LayoutObject tree.

Most of the necessary logic are still missing. For sizing the block
container, the logic should be shared with NGBlockLayoutAlgorithm using
composition. This will be filled in future patches.

For more steps of inline layout (e.g., line breaking, bidi reordering),
making the fragment tree meaningful, copying the fragment tree to
LayoutObject tree, etc. will be in future patches too.

NGInlineLayoutTest is changed to test the result after layout completes,
as NGInlineBox collects inline text during the layout in this patch.

BUG=636993

Review-Url: https://codereview.chromium.org/2551793002
Cr-Commit-Position: refs/heads/master@{#437828}

Memory-Infra on Windows doesn't use memory usage reported by the OS (proportional set size, private dirty, etc.) yet.

This CL added functionality to calculate proportional set size for a process.

BUG=623499

Review-Url: https://codereview.chromium.org/2549803003
Cr-Commit-Position: refs/heads/master@{#437827}

The JoystickScrollProvider Java class used to emulate a View
scroll events and call ContentViewCore.scrollBy(), now it
emulates the mouse wheel events directly, dropping the calculation
of the current event position (it is set to (0,0)).

This CL also removes JoystickScrollProvider dependency on
ContentViewCore, it talks to its own native counterpart.

BUG=620929, 650351

Review-Url: https://codereview.chromium.org/2548363007
Cr-Commit-Position: refs/heads/master@{#437826}

This CL adds a Java-side interface to the Site Engagement Service, as
well as a basic integration test for the interface. This allows Java
code (particularly tests) to programatically adjust and retrieve the
score.

BUG=671100

Review-Url: https://codereview.chromium.org/2553013002
Cr-Commit-Position: refs/heads/master@{#437825}

https://skia.googlesource.com/skia.git/+log/2790c52e36dd..bbab477b49e2

$ git log 2790c52e3..bbab477b4 --date=short --no-merges --format='%ad %ae %s'
2016-12-08 jvanverth Add documentation for Viewer

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
TBR=msarett@google.com

Review-Url: https://codereview.chromium.org/2570433003
Cr-Commit-Position: refs/heads/master@{#437824}

Also add a hook for content::RenderThreadObserver implementations, such
as ChromeRenderThreadObserver, to register their Mojo interfaces.

Tested manually, by creating a new Chrome extension a la
https://developer.chrome.com/extensions/getstarted
and replacing its popup.js script with:

document.addEventListener('DOMContentLoaded', function() {
  document.getElementById('status').textContent =
    'inIncognitoContext: ' + chrome.extension.inIncognitoContext;
});

BUG=577685

Review-Url: https://codereview.chromium.org/2531133002
Cr-Commit-Position: refs/heads/master@{#437823}

https://crrev.com/2558743002 fixed a bug that loading metric doesn't
work for webview traces. This patch re-enables page_cycler_v2 and
loading.mobile benchmarks on webview bots.

BUG=651188
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq

Review-Url: https://codereview.chromium.org/2564773002
Cr-Commit-Position: refs/heads/master@{#437822}

Now that ContextLifecycleObserver clears its observing context after calling
contextDestroyed. So there is no need to explicitly call clearContext().

BUG=610176

Review-Url: https://codereview.chromium.org/2552703006
Cr-Commit-Position: refs/heads/master@{#437821}

This is a rename requested by https://codereview.chromium.org/2554813002/.

Previously ActiveDOMObject was used to implement hasPendingActivity() of DOM wrappers,
where the name "ActiveDOMObject" made sense. Then we moved hasPendingActivity() to
ActiveScriptWrappable. Now ActiveDOMObject is a class that just provides resume/suspend.
It would make more sense to rename it to SuspendableObject.

BUG=610176
TBR=jam@chromium.org

Review-Url: https://codereview.chromium.org/2567913002
Cr-Commit-Position: refs/heads/master@{#437820}

BUG=671459
TBR=avi@chromium.org

Review-Url: https://codereview.chromium.org/2565273002
Cr-Commit-Position: refs/heads/master@{#437819}

BUG=NONE

Committed: https://crrev.com/98f0f335c82dbbd932d3d5bad9fbb22585dafb65
Review-Url: https://codereview.chromium.org/2556803002
Cr-Original-Commit-Position: refs/heads/master@{#437292}
Cr-Commit-Position: refs/heads/master@{#437818}

Cr-Commit-Position: refs/heads/master@{#437817}

Remove ArcBridgeService::Get().
Now all usage was gone. It's time to delete the function.

BUG=657687
BUG=b/31079732
TEST=Ran bots.

Review-Url: https://codereview.chromium.org/2558313002
Cr-Commit-Position: refs/heads/master@{#437816}

https://chromium.googlesource.com/angle/angle.git/+log/e735511..d023939

BUG=chromium:671021
TBR=jmadill@chromium.org
TEST=bots
NOTRY=true
NOPRESUBMIT=true

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.android:android_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2567923002
Cr-Commit-Position: refs/heads/master@{#437815}

This CL has no behavior changes.

BUG=668300

Review-Url: https://codereview.chromium.org/2555743005
Cr-Commit-Position: refs/heads/master@{#437814}

Sort fields in ComputedStyleBase in decreasing order of size. This is
a micro optimization that improves field alignment and minimizes wasted
memory space.

From Intel on 'Coding for Performance':
https://software.intel.com/en-us/articles/coding-for-performance-data-alignment-and-structures
"One can try to minimize this memory wastage by ordering the structure
elements such that the widest (largest) element comes first, followed by
the second widest, and so on."

Another good explanation article:
https://msdn.microsoft.com/en-us/library/ewwyfdbe.aspx

BUG=628043

Review-Url: https://codereview.chromium.org/2558733002
Cr-Commit-Position: refs/heads/master@{#437813}

Revert of gpu: Move ShaderDiskCache into //gpu/ipc/host component. (patchset #5 id:80001 of https://codereview.chromium.org/2555693005/ )

Reason for revert:
Makes StoragePartitionShaderClearTest.ClearShaderCache fail on Linux Tests bot: https://build.chromium.org/p/chromium.linux/builders/Linux%20Tests%20%28dbg%29%281%29

Original issue's description:
> gpu: Move ShaderDiskCache into //gpu/ipc/host component.
>
> BUG=643746, 604369
> CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
>
> Review-Url: https://codereview.chromium.org/2555693005

TBR=piman@chromium.org,rch@chromium.org,rsleevi@chromium.org,ben@chromium.org,sadrul@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=643746, 604369

Review-Url: https://codereview.chromium.org/2565243002
Cr-Commit-Position: refs/heads/master@{#437812}

This CL gets rid of using ArcBridgeService::Get() in
arc_robot_auto_code_fetcher_browsertest, by getting rid of
depending of ArcAuthService.

Test for AttemptUserExit is moved to arc_session_manager_unittest.
For that purpose, fixed username_hash for ARC Kiosk fake user.

BUG=657687
BUG=b/31079732
TEST=Ran bots.

Review-Url: https://codereview.chromium.org/2553193002
Cr-Commit-Position: refs/heads/master@{#437811}

As a follow-up for [1], this CL fixes the flakiness of
timeline-receive-response-event.html by dumping only the first
ResourceReceivedData event for each request.

1: https://crrev.com/48f3e047d17e5f78543bb0bec1c421ba3a69cb79

BUG=666217

Review-Url: https://codereview.chromium.org/2563613004
Cr-Commit-Position: refs/heads/master@{#437810}

TBR=mkwst@chromium.org
NOTRY=true
BUG=610835

Review-Url: https://codereview.chromium.org/2567933002
Cr-Commit-Position: refs/heads/master@{#437809}

BUG=672516

Review-Url: https://codereview.chromium.org/2562063002
Cr-Commit-Position: refs/heads/master@{#437808}

Created by adding the following to testing/buildbot/manage.py in
def process_file(..):

   if 'views_unittests' in seen and 'native_theme_unittests' not in seen:
     data = [d for d in gtest_tests if d['test'] == 'views_unittests'][0].copy()
     data['test'] = 'native_theme_unittests'
     gtest_tests.append(data)

Then running manage.py --write. It told me to put an entry in
gn_isolate_map.pyl as well.

BUG=398316

Review-Url: https://codereview.chromium.org/2560633004
Cr-Commit-Position: refs/heads/master@{#437807}

Review-Url: https://codereview.chromium.org/2560423002
Cr-Commit-Position: refs/heads/master@{#437806}

It's only one test now (zero on Android, so it breaks on Android FYI)
and will be merged into service_manager_unittests.

BUG=None
R=dpranke@chromium.org

Review-Url: https://codereview.chromium.org/2563253003
Cr-Commit-Position: refs/heads/master@{#437805}

We moved to checking need-for-repaint and invalidating based
on anchor node in http://crrev.com/2401363003 but we
incorrectly left null-checking based on non-anchor node.

BUG=663319

Review-Url: https://codereview.chromium.org/2565103002
Cr-Commit-Position: refs/heads/master@{#437804}

Because refX and refY are in the coordinate space of each marker instance,
they are included in markerTransformation() instead of being part of the
marker viewport. As a result, only a viewport size is maintained in
LayoutSVGResourceMarker. This patch makes this distinction a little
clearer by refactoring m_viewport to m_viewportSize, as well as adding
some comments.

No new tests; no change in behavior.

Review-Url: https://codereview.chromium.org/2567853002
Cr-Commit-Position: refs/heads/master@{#437803}

https://chromium.googlesource.com/chromium/src/+/c896e79e5ba34

BUG=548616
TBR=robhogan@gmail.com

Review-Url: https://codereview.chromium.org/2563323002 .
Cr-Commit-Position: refs/heads/master@{#437802}

ReadingListStoreDelegate has virtual methods, so it should have a virtual
destructor.

BUG=673169

Review-Url: https://codereview.chromium.org/2565213002
Cr-Commit-Position: refs/heads/master@{#437801}

ReadingListStore inherits from multiple classes, including from
ReadingListModelStorage that does not define a virtual destructor.

Since ReadingListModelImpl has a scoped_ptr to ReadingListModelStorage,
it calls the base destructor with a pointer that does not correspond
to memory returned by "new" (it points to somewhere inside the memory
block).

The issue can be fixed by defining a virtual destructor on the base
class ReadingListModelStorage (as all class with virtual method should
do). Also mark the class a non-copyable since copying would cause slicing.

BUG=673169

Review-Url: https://codereview.chromium.org/2563303002
Cr-Commit-Position: refs/heads/master@{#437800}

Revert of Enable RL by default (patchset #6 id:100001 of https://codereview.chromium.org/2562643003/ )

Reason for revert:
Speculative revert as I think this CL broke the downstream autoroller.

It is a bit hard to debug as the autoroller was unable to land some CL as the downstream tree was closed, but the last successful roll would have happened at 39e086ae [1], and the first failing roll is at a229e828 [2].

This CL is the only one having an impact on the iOS build, so I think it is causing the failures:

$ git log --oneline 39e086ae...a229e828
a229e828 Strength-reduce the "scale-factor changed" condition in LayoutSVGRoot
d1693364 Update V8 to version 5.7.202.
7cfbed3d Enable RL by default
314c528d Revert of Improve EAT_STREAM_PARAMETERS for Windows x86 (patchset #10 id:240001 of https://codereview.chromium.org/2559323007/ )

[1]: https://uberchromegw.corp.google.com/i/internal.bling.main/builders/trunk-autoroller/builds/6861
[2]: https://uberchromegw.corp.google.com/i/internal.bling.main/builders/trunk-autoroller/builds/6862

Original issue's description:
> Enable RL by default
>
> BUG=577659
>
> Review-Url: https://codereview.chromium.org/2562643003

TBR=noyau@chromium.org,mardini@chromium.org,olivierrobin@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=577659
NOTRY=true

Review-Url: https://codereview.chromium.org/2564293002
Cr-Commit-Position: refs/heads/master@{#437799}

Computing the offset for a list marker after the rest of the objects on the line
it is on have been laid out means it will avoid floats it ought not to.

Instead, compute the offset when laying out the marker and cache it for use later.

BUG=548616
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2553793003
Cr-Commit-Position: refs/heads/master@{#437798}

BUG=671181

Review-Url: https://codereview.chromium.org/2559223002
Cr-Commit-Position: refs/heads/master@{#437797}

BUG=672516

Review-Url: https://codereview.chromium.org/2567483003
Cr-Commit-Position: refs/heads/master@{#437796}

Without this check we'd crash when creating the watcher and
asking for the server file descriptor from a null server instance.

TBR=sky@chromium.org
BUG=
TEST=chrome --enable-wayland-server (without XDG_RUNTIME_DIR set)

Review-Url: https://codereview.chromium.org/2569633002
Cr-Commit-Position: refs/heads/master@{#437795}