Currently getting the following 2 errors:
```
error: reference to 'MarkingType' is ambiguous
error: reference to 'SweepingType' is ambiguous
```
Change-Id: Ia50d1b5ea8af0fcc85acb9c0dc5cfae1956cec62
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3540624
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Commit-Queue: Milad Farazmand <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#79556}
Access to the object start bitmap is only safe during marking until
sweeping is started as the concurrent sweeper may clear and rebuild
the bitmap at any time during sweeping.
Adds a DCHECK and an additional test for a previously broken
pre-finalizer scenario.
Bug: chromium:1307471
Change-Id: If67ade43f7cdad6de4720c0efeac11bfe8c22b3c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3535782
Reviewed-by: Nikolaos Papaspyrou <nikolaos@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79550}
This CL moves the call to GCTracer::StopCycle for the full GC from
Heap::CompleteSweeping full, which is called to force sweeping to
finish, to GCTracer::NotifySweepingComplete, which is called as soon
as sweeping finishes --- and symmetrically to a new method
GCTracer::NotifyCppGCCompleted, which is called as soon as sweeping
of the managed C++ heap finishes. In this way, a full GC cycle is
reported as soon as sweeping is finished both for the V8 and the C++
managed heap.
The changes introduced in this CL are essentially a partial revert of
https://crrev.com/c/3456563, fixed in such a way that when the full
GC cycle is reported, the current tracer event will be the correct
one corresponding to that cycle.
Bug: v8:12503
Bug: chromium:1154636
Change-Id: Icea07cf35a9565994e798b0500e9da72cd95f9ac
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3497318
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79470}
Instead of using the high water mark for determining this metric, we use
a bitset for all active/used system pages on a V8 heap page. Each time
when allocating a LAB on a page, we add the pages of that memory range
to that bitset. During sweeping we rebuild that bitset from scratch and
replace it with the old one in case free pages are discarded by the GC.
We DCHECK here that the sweeper only ever removes pages. This has the
nice benefit of ensuring that we don't miss any allocations (like we
do now for concurrent allocations).
CommittedPhysicalMemory for a page is then calculated by counting the
set bits in the bitset and multiplying it with the system page size.
This should be simpler to verify and track the "real" effective size
more precisely.
One case where we are partially less precise than the current
implementation is for LABs. In order to reduce complexity we now treat
all pages of a LAB allocation as active immediately. In the current
implementation we tried to only account the actual used part of the LAB
when changing the LAB later. This is more complex to track correctly
but also doesn't account the currently used LAB in effective size.
Change-Id: Ia83df9ad5fbb852f0717c4c396b5074604bd21e9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3497363
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79428}
The only user of OverrideEmbedderStackStateScope is Blink where it is
used to override stack state of top-level tasks. Adjust the behavior
here to allow using this scope broadly while still supporting explicit
garbage collection calls.
Bug: chromium:1300492
Change-Id: I78c418c5f08991bf6857147cd4a537246bfcc556
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3497744
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79420}
Before a young GC, the scavenger finalizes a pending full GC sweeping,
in case there are no more running sweeping jobs, to avoid unnecessarily
pausing sweeping and then resuming it.
This CL moves this sweeping finalization from
ScavengerCollector::CollectGarbage to Heap::CompleteSweepingYoung, so
that it is also performed for the minor mark-compactor and that sweeping
is correctly attributed to the previous full GC cycle (instead of the
beginning young cycle). Furthermore, it also finalizes CppGC sweeping
if there are no more running sweeping jobs.
Bug: chromium:1154636
Change-Id: Ic9ba4803f49db32c0a539f080329f012859bc8bc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3508011
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79407}
This renaming was missed in
https://chromium-review.googlesource.com/c/v8/v8/+/3468577, because the
code adding these #ifdef blocks landed in-between the renaming change
being uploaded and landed.
Bug: chromium:1298417
Change-Id: I8c2e951099dafcce7e19a59e40e36f308e2ed867
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3498349
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Benoit Lize <lizeb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79345}
The following histograms have been deprecated, as they are superseded
by V8.GC.Cycle.*.Young or not needed anymore (next to each, the
corresponding isolate counter):
- V8.GCScavenger (gc_scavenger)
- V8.GCScavengerBackground (gc_scavenger_background)
- V8.GCScavengeReason (scavenge_reason)
- V8.GCScavengerForeground (gc_scavenger_foreground)
- V8.GCBackgroundScavenger (background_scavenger)
- V8.GCMarkCompactor (gc_mark_compactor)
This CL removes the corresponding instrumentation in the code and the
isolate counters.
Bug: chromium:1154636
Bug: chromium:1299555
Change-Id: I62d28ff60ef47a058fe148c7855af8e2c1cc0aed
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487548
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79308}
This CL refactors Heap::GCTypeTimer and Heap::GCTypePriorityTimer
and moves them to a GCTracer::RecordGCPhasesInfo class. This is
a necessary change for deprecating counters that are used for
old style GC metrics, like gc_scavenger. When all such counters
are deprecated, GCTracer::RecordGCPhasesInfo will no longer be
necessary and will be removed.
Bug: chromium:1154636
Change-Id: I04504a0f6c7a0955f4300a1c94c969aaeb23b77f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3486556
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79257}
Custom callbacks assume that untraced pointers always point to valid,
not freed objects. They must make sure that upon callback completion no
UntracedMembers point to an unreachable object. This may not hold true
if a custom callback for an old object operates with a reference to a
young object that was freed on a minor collection cycle. To maintain
the mentioned invariant, the CL calls custom callbacks for old objects
on every minor collection cycle.
The alternative options could be:
1) Replacing all UntracedMembers with WeakMembers, since WeakMember
supports tracing and the barrier.
2) Emitting the generational barrier for UntracedMember + tracing
UntracedMember on minor collection cycles.
The first option requires changing multiple use sites and can bring some
performance regression. The second option requires changing the GC logic
and the semantics of UntracedMember.
Bug: chromium:1029379
Change-Id: I9bb89e4787daf05990feed374dceca940be7be63
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3472499
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79221}
This CL refactors all remembered set logic from heap-base and
explicit-management to a new class OldToNewRememberedSet.
Bug: chromium:1029379
Change-Id: Id032b9dcc01af6f9bb9e546ed9bc6324da6d9b66
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3472498
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79212}
The test case was fixed in c7fbac6a72
This reverts commit 34c0f0fced.
Bug: chromium:1278780
Change-Id: If04e41a7fc1f0c744fe785a834880e598f482ef8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3401592
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79201}
Weak containers are retraced if they are found through the stack using
the conservative scanner, possibly resulting in a race with the
concurrent marker.
Bug: v8:12648
Change-Id: I0936a2953e3e2151cea4191f335a091b0e334e6b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3474678
Auto-Submit: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79184}
Previously, V8_OS_MACOSX was, somewhat confusingly, also used for iOS.
With this CL, V8_OS_DARWIN will be set on both macOS and iOS,
V8_OS_MACOS only on macOS, and V8_OS_IOS only on iOS.
This CL also renames V8_TARGET_OS_MACOSX to V8_TARGET_OS_MACOS and
renames platform-xnu.cc to platform-darwin.cc.
Change-Id: I4bcafc7c337586662114144f6c7ccf47d978da1f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3468577
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Samuel Groß <saelo@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79167}
Guard pages are 4k areas at the beginning and end of each oilpan page
(128kiB) which are meant to be inaccessible. However on ARM64 macOS, the
OS page size is 16kiB, meaning that these are not inaccessible. But we
do pay for these, as they are part of the first and last OS
page. Meaning that we effectively waste 2 * 4kiB = 6.25% of each Oilpan
page.
Since these are not serving their purpose, disable them on this
platform. Another fix could be to make the guard page 16kiB, but given
that the entire oilpan page is 128kiB, this may have adverse effects on
e.g. fragmentation.
Note that this doesn't regress security, as the regions were never
protected to begin with on this platform.
Bug: chromium:1298417
Change-Id: Iad5d05670962780e6d1eeab2bb8a331deb7aa1f3
Cq-Include-Trybots: luci.v8.try:v8_linux_arm64_rel_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3471558
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Benoit Lize <lizeb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79151}
This CL simplifies the reporting of full GC cycles and the connection
between the GC of the managed C++ heap and the managed Javascript heap.
It moves the call to GCTracer::RecordFullCycleToRecorder to be part of
GCTracer::StopCycle.
Bug: v8:12503
Bug: chromium:1154636
Change-Id: I332dbcd81d2e5bdda83f3353c6526fc18e23ebd5
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3456563
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79075}
The generational barrier for source objects records the entire source
object to be processed later during remembered set visitation. It's
planned to be used for Blink backing stores when an inlined object (or a
range thereof) is added (HeapAllocator::NotifyNewObject(s)).
An alternative approach would be to eagerly process the inlined objects
using a custom callback. However, this requires changing Visitors to
bring slots into the context. This approach should better work for
scenarios where small ranges or single elements are added, to avoid
processing potentially large backing stores. The followup CL implements
this idea.
Bug: chromium:1029379
Change-Id: Iacb59e4b10a66354526ed293d7f43f14d8761a8f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3460402
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79073}
This CL fixes a bug in the tracing of full GC cycles that was introduced
by https://crrev.com/3432211. In doing so, it refactors the tracing of
cycles by introducing an explicit state in GC tracing events, which
follows the phase within the GC cycle as perceived by the tracer. Two
new methods, (Start|Stop)AtomicPause are introduced; together with
(Start|Stop)Cycle they mark the state transitions. The existing methods
(Start|Stop)ObservablePause are now disentangled from cycles and state
transitions.
Bug: v8:12503
Bug: chromium:1154636
Change-Id: Ie4b863bc27f81dd6858103a8988874d89e6e8517
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3440663
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79043}
The methods in explicit_management.h should be called via the public
variants in the subtle namespace. Calling the variants in the internal
namespace directly skips asserts and required size coversions. Doing so
may cause misuse of the api that may break GC inernals
Change-Id: I58a0f324ca1ee0839bb85eb9b53ce57785dc7b91
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3435187
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78920}
This CL adds invalidations for slots that reside in promptly freed or
shrunk storage.
Bug: chromium:1029379
Change-Id: I05e0ede55c202c952b26f452053b8777d1a2ffae
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3431488
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78912}
Conceptually, a full GC cycle completes when the sweeping phase is
finished. As sweeping is performed concurrently, this happens after
Heap::CollectGarbage has returned and, at the latest, before the next
full GC cycle begins. However, an arbitrary number of young GC cycles
may happen in the meantime. Tracing information for the sweeping phase
must be added to the corresponding full GC cycle event. Until now, this
was not done correctly: this information was added to the GCTracer's
current event and could thus be attributed to a subsequent young or full
GC cycle.
This CL introduces methods GCTracer::(Start|Stop)Cycle to delimit a
cycle (still allowing for full GC cycles to be interrupted by young GC
cycles). These methods are different from (Start|Stop)ObservablePause,
which delimit the observable pause of each GC. The events of "pending"
full GC cycles are kept until they are properly amended and reported,
when the sweeping phase is finished.
This is a reland of 4ad20bff97
which was reviewed here: https://crrev.com/3404733
Bug: v8:12503
Bug: chromium:1154636
Change-Id: Icc315b53cff1f3b19b8efe49db34340a5608bcd2
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3432211
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78911}
This reverts commit 4ad20bff97.
Reason for revert: New test seems to be failing on TSAN/incremental marking stress (https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux64%20TSAN%20-%20stress-incremental-marking/6346/overview)
Original change's description:
> heap: Fix the tracing of GC cycles
>
> Conceptually, a full GC cycle completes when the sweeping phase is
> finished. As sweeping is performed concurrently, this happens after
> Heap::CollectGarbage has returned and, at the latest, before the next
> full GC cycle begins. However, an arbitrary number of young GC cycles
> may happen in the meantime. Tracing information for the sweeping phase
> must be added to the corresponding full GC cycle event. Until now, this
> was not done correctly: this information was added to the GCTracer's
> current event and could thus be attributed to a subsequent young or full
> GC cycle.
>
> This CL introduces methods GCTracer::(Start|Stop)Cycle to delimit a
> cycle (still allowing for full GC cycles to be interrupted by young GC
> cycles). These methods are different from (Start|Stop)ObservablePause,
> which delimit the observable pause of each GC. The events of "pending"
> full GC cycles are kept until they are properly amended and reported,
> when the sweeping phase is finished.
>
> Bug: chromium:1154636
> Change-Id: I2fbc65d4807c78656d4abc8c451043f6f86211b1
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3404733
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Reviewed-by: Omer Katz <omerkatz@chromium.org>
> Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#78905}
Bug: chromium:1154636
Change-Id: Id6688cfe982f9d8159c66d715b7079782a371bed
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3431489
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Owners-Override: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78908}
Conceptually, a full GC cycle completes when the sweeping phase is
finished. As sweeping is performed concurrently, this happens after
Heap::CollectGarbage has returned and, at the latest, before the next
full GC cycle begins. However, an arbitrary number of young GC cycles
may happen in the meantime. Tracing information for the sweeping phase
must be added to the corresponding full GC cycle event. Until now, this
was not done correctly: this information was added to the GCTracer's
current event and could thus be attributed to a subsequent young or full
GC cycle.
This CL introduces methods GCTracer::(Start|Stop)Cycle to delimit a
cycle (still allowing for full GC cycles to be interrupted by young GC
cycles). These methods are different from (Start|Stop)ObservablePause,
which delimit the observable pause of each GC. The events of "pending"
full GC cycles are kept until they are properly amended and reported,
when the sweeping phase is finished.
Bug: chromium:1154636
Change-Id: I2fbc65d4807c78656d4abc8c451043f6f86211b1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3404733
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78905}
The CL does following:
1) Makes sure young generation works and tests pass;
2) Provides CollectGarbageInYoungGenerationForTesting() that is needed
to support remaining tests in Blink;
3) Moved cppgc_enable_young_generation GN flag to v8.gni to refer to it
from Blink;
4) Bails out from marking TracedReferences in UnifiedHeapMarkingState;
5) Disables (temporarily) prompt freeing for young generation;
6) Fixes remembered set visitation for nullptr|kSentinel slots.
Bug: chromium:1029379
Change-Id: I5165fa22c8a0eaa708ef7a35a9978cb12e1cb13e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3429202
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Auto-Submit: Anton Bikineev <bikineev@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78901}
Split method into ShouldRecordRelocSlot and ProcessRelocInfo.
ProcessRelocInfo can then be reused in the write barrier and in the
future for the OLD_TO_SHARED remembered set. SlotTypeForRelocInfoMode
got moved into ProcessRelocInfo.
In addition rename and document SlotTypes. This CL does not change
behavior.
Bug: v8:11708
Change-Id: Iff712e7e6f5d3a4da64510b67b604b9e04998361
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3400968
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78701}
This is a reland of 142dd775b4
Original change's description:
> cppgc-js,heap: Implement snapshots for embedder fields
>
> https://crrev.com/c/3293410 added concurrent processing of C++ objects
> found through V8 embedder fields. The CL missed that those embedder
> fields are not read atomically from JS objects. The problem is that
> embedder fields are only aligned to kTaggedSize on builds with pointer
> compression and are as such mis-aligned for atomic ops. This is not a
> problem for on-heap values as the upper 32bits are anyways computed
> from the cage. Is is a problem for generic C++ values though, as they
> are used with Oilpan.
>
> This CL adds the standard marker snapshot protocol for embedder fields.
>
> Marker:
> 1. Snapshot embedder fields
> 2. Try to mark host object
> 3. On success: process snapshot
>
> Main thread:
> 1. On setting embedder fields mark the object black first
> 2. Emit a write barrier for the embedder fields
>
> This will get simpler with the heap sandbox that uses a separate table
> for embedder fields. Once the sandbox is the default configuration, we
> can use it as dependency for the concurrent fast path.
>
> Bug: chromium:1285706
> Change-Id: I6b975ea561be08cda840ef0dd27a11627de93900
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3380983
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#78604}
Bug: chromium:1285706
Change-Id: I024e50fc0757fbcd13cb9ffde027dff55f99d25c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3386600
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78631}
This reverts commit 142dd775b4.
Reason for revert: TSAN breaks: https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux64%20TSAN%20-%20stress-incremental-marking/6113/overview
Original change's description:
> cppgc-js,heap: Implement snapshots for embedder fields
>
> https://crrev.com/c/3293410 added concurrent processing of C++ objects
> found through V8 embedder fields. The CL missed that those embedder
> fields are not read atomically from JS objects. The problem is that
> embedder fields are only aligned to kTaggedSize on builds with pointer
> compression and are as such mis-aligned for atomic ops. This is not a
> problem for on-heap values as the upper 32bits are anyways computed
> from the cage. Is is a problem for generic C++ values though, as they
> are used with Oilpan.
>
> This CL adds the standard marker snapshot protocol for embedder fields.
>
> Marker:
> 1. Snapshot embedder fields
> 2. Try to mark host object
> 3. On success: process snapshot
>
> Main thread:
> 1. On setting embedder fields mark the object black first
> 2. Emit a write barrier for the embedder fields
>
> This will get simpler with the heap sandbox that uses a separate table
> for embedder fields. Once the sandbox is the default configuration, we
> can use it as dependency for the concurrent fast path.
>
> Bug: chromium:1285706
> Change-Id: I6b975ea561be08cda840ef0dd27a11627de93900
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3380983
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#78604}
Bug: chromium:1285706
Change-Id: If1976c0356f450fc068aa4dcc39fb9a0d5417a40
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3386598
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Owners-Override: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#78605}
https://crrev.com/c/3293410 added concurrent processing of C++ objects
found through V8 embedder fields. The CL missed that those embedder
fields are not read atomically from JS objects. The problem is that
embedder fields are only aligned to kTaggedSize on builds with pointer
compression and are as such mis-aligned for atomic ops. This is not a
problem for on-heap values as the upper 32bits are anyways computed
from the cage. Is is a problem for generic C++ values though, as they
are used with Oilpan.
This CL adds the standard marker snapshot protocol for embedder fields.
Marker:
1. Snapshot embedder fields
2. Try to mark host object
3. On success: process snapshot
Main thread:
1. On setting embedder fields mark the object black first
2. Emit a write barrier for the embedder fields
This will get simpler with the heap sandbox that uses a separate table
for embedder fields. Once the sandbox is the default configuration, we
can use it as dependency for the concurrent fast path.
Bug: chromium:1285706
Change-Id: I6b975ea561be08cda840ef0dd27a11627de93900
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3380983
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78604}
This is a reland of 86038ecfdc
Compared to the previous CL this one is adding a TSAN suppression
for GlobalSafepoint::EnterSafepointScope. local_heaps_mutex_ of client
isolates may be locked in any order. This would be detected by TSAN as a
potential race. Add some additional DCHECKs to compensate for that
missing test coverage.
As a cleanup this CL also removes the unused methods ContainsLocalHeap()
and ContainsAnyLocalHeap() from LocalHeap.
Original change's description:
> [heap] Optimize time to reach global safepoint
>
> Initial support for global safepoints kept it simple by entering a
> safepoint for each of them one after another. This means
> time-to-global-safepoint is the sum of all time-to-safepoint operations.
> We can improve this slightly by splitting up the safepoint iteration
> into two operations:
>
> 1) Initiate safepoint lock (locks local_heaps_mutex_, arms the barrier
> and sets SafepointRequested flag for all client threads)
> 2) Block until all runnning client threads reach a safepoint
>
> We now perform operation 1) for all clients first and only then start
> with operation 2).
>
> Bug: v8:11708
> Change-Id: Iaafd3c6d70bcf7026f722633e9250b04148b3da6
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3310910
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#78308}
Bug: v8:11708, v8:12492
Change-Id: I7087ba23c08f2d4edb9b632eef3c218fc76342e7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3328786
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78583}
This CL doesn't change behavior, only refactors MemoryAllocator:
* De-templatify class, MemoryAllocator is used on slow path and doesn't
really need templates for performance.
* Rename FreeMode names
* Move methods into private section of class
Change-Id: I7894fba956dcd7aa78ad0284d0924662fef4acae
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3379812
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78580}
Included in this CL:
(*) Introduce CppMarkingState that V8 should use to push references to
Oilpan. CppMarkingState allocates its own Worklist::Locals to
support concurrent updates from V8.
(*) Split Oilpan MarkingWorklist object to form a base class used by
CppMarkingState.
(*) Remove MarkerFactory and split marking initialization. Marking
worklists should already be initialized when V8 initializes
visitors. For incremental marking, this requires splitting
marking initialization and marking start.
(*) Drive-by: Mark JSObject::IsApiWrapper and
JSObject::IsDroppableApiWrapper as const.
Bug: v8:12407
Change-Id: I35cc816343da86f69a68306204675720e9b3913f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3293410
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78446}
This CL renames a number of things related to the V8 sandbox.
Mainly, what used to be under V8_HEAP_SANDBOX is now under
V8_SANDBOXED_EXTERNAL_POINTERS, while the previous V8 VirtualMemoryCage
is now simply the V8 Sandbox:
V8_VIRTUAL_MEMORY_CAGE => V8_SANDBOX
V8_HEAP_SANDBOX => V8_SANDBOXED_EXTERNAL_POINTERS
V8_CAGED_POINTERS => V8_SANDBOXED_POINTERS
V8VirtualMemoryCage => Sandbox
CagedPointer => SandboxedPointer
fake cage => partially reserved sandbox
src/security => src/sandbox
This naming scheme should simplify things: the sandbox is now the large
region of virtual address space inside which V8 mainly operates and
which should be considered untrusted. Mechanisms like sandboxed pointers
are then used to attempt to prevent escapes from the sandbox (i.e.
corruption of memory outside of it). Furthermore, the new naming scheme
avoids the confusion with the various other "cages" in V8, in
particular, the VirtualMemoryCage class, by dropping that name entirely.
Future sandbox features are developed under their own V8_SANDBOX_X flag,
and will, once final, be merged into V8_SANDBOX. Current future features
are sandboxed external pointers (using the external pointer table), and
sandboxed pointers (pointers guaranteed to point into the sandbox, e.g.
because they are encoded as offsets). This CL then also introduces a new
build flag, v8_enable_sandbox_future, which enables all future features.
Bug: v8:10391
Change-Id: I5174ea8f5ab40fb96a04af10853da735ad775c96
Cq-Include-Trybots: luci.v8.try:v8_linux64_heap_sandbox_dbg_ng,v8_linux_arm64_sim_heap_sandbox_dbg_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3322981
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Samuel Groß <saelo@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78384}
Introduce a ReusableUnoptimizedCompileState class, passed to ParseInfo,
which stores a couple of pointers and most importantly the Zone and
AstValueFactory of the parse. This allows the Zone and AstValueFactory
to be reused across multiple parses, rather than re-initialising
per-Parse.
With this, we can amend the LazyCompileDispatcher to initialise one
LocalIsolate, Zone and AstValueFactory per background thread loop,
rather than one per compile task, which allows us to reduce per-task
costs and re-use the AstValueFactory's string table and previous String
internalizations.
Change-Id: Ia0e29c4e31fbe29af57674ebb10916865d38b2ce
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3313106
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78289}
Checks whether a Persistent is used from the creation thread on slow
path allocations. In practice, these currently happen every 256
Persistent allocations. This is a best effort check that may help to
flush out issues that are missed with DCHECK builds.
Bug: chromium:1276570
Change-Id: Ia868ca436341b1b5ef427d5b3ec04926c1394e41
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3318658
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78276}
This interface is meant to eventually replace the existing
v8::PageAllocator interface. Beyond general refactoring of the
PageAllocator APIs, the new interface now supports the concept of
(contiguous) address space reservations, which previously had to be
implemented through page allocations. These reservations now make better
use of provided OS primitives on Fuchsia (VMARs) and Windows
(placeholder mappings) and can be used to back many of the cages and
virtual memory regions that V8 creates.
The new interface is not yet stable and may change at any time without
deprecating the old version first.
Bug: chromium:1218005
Change-Id: I295253c42e04cf311393c5dab9f8c06bd7451ce3
Cq-Include-Trybots: luci.v8.try:v8_linux64_heap_sandbox_dbg_ng,v8_linux_arm64_sim_heap_sandbox_dbg_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3301475
Commit-Queue: Samuel Groß <saelo@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78235}
This allows the embedder to determine whether some function has been
called from a destructor.
See discussion in
https://crrev.com/c/3302810
Bug: chromium:1273928
Change-Id: Icb5d98eff777574488a7d6de5e693c502c2fb53e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3303793
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78113}
... and thus avoid the need for special handling of objects located
in external code space.
This will also allow making HeapObject::IsBlah() checks faster when
external code space is enabled.
Bug: v8:11880
Change-Id: I12d07c05451ff198f0a6182d9b5849f76015e7fb
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3300140
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78104}
Unify parse post-processing between main-thread and background-thread
parsing, now that we have LocalIsolate and can Internalize on background
threads.
As part of this, simplify the LocalIsolate parking pattern to explicitly
park during ParseOnBackground, rather than being implicitly parked when
ParseOnBackground is called. This reduces the amound of scoping needed
in the BackgroundCompileTask::Run method.
Change-Id: Ifdb128b763129bda78bd1bae89dac1c62f872350
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3277876
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77872}
Adds support for aborting compaction when finalizing with stack:
- never_compact_with_stack: All pages are aborted;
- never_compact_code_space_with_stack: Only code space pages are
aborted;
This flags allow simulating a worst case where a stack cannot
be considered precise, or evacuation candiate is refered to from
a stack slot that V8 has no info for.
Bug: v8:12251
Change-Id: Ice24ac87a985b8ecf7b5cbb5c106ad4a3ae1944b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3173682
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77792}
Explicit write barriers for internal fields are deprecated as they are
automatically emitted by V8.
Depends on:
https://crrev.com/c/3263924
Bug: v8:12356
Change-Id: I171ba5b42a6570ce52e2e2ea1b7c1029d5a8a3a7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3263888
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77772}
