v8::TracedReference and their underlying traced handles are not
actually phantom references but have been historically represented as
such. This patch splits off their backing implementaton into
`TracedHandles`.
As a drive-by, re-organize the free list to rely on indices instead of
pointers, saving a full word for each node. In additon, the setup also
allows for freeing fully empty node blocks after GCs.
Allocation/free of nodes is also organized in a way that avoids
jumping between handles/blocks/nodes for accounting purposes.
Removing CRTP in GlobalHandle internals is left for a follow up.
Bug: v8:13372
Change-Id: Ib55aa9ba71202d5918b486c161fe6f747d31c719
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3966952
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83872}
This CL makes the object keeping stack information thread-local, moving
it from Heap to ThreadLocalTop. In this way, stack scanning will work
correctly when switching between threads, e.g., using v8::Locker.
It also introduces a mechanism for setting a stack marker, to be used
for scanning only the part of stack between its start and the marker
(instead of the current stack top).
Bug: v8:13257
Change-Id: I01091f5f49d9a8143d50aeef53789a98bdb29048
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3960991
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@{#83848}
Turbofan generates bitcast nodes like BitcastWordToTaggedSigned to
ensure the value types of definitions and uses are matched. These
nodes can be elided after MachineGraphVerifier verifying the graph.
This can avoid generating redundant instructions:
Before:
xorl r15,r15
cmpl [rdx+0xb],r15
After:
cmpl [rdx+0xb],0x0
Change-Id: I84bc1b05d77ed9487001e34a93dfe14e45a7a678
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3837161
Commit-Queue: Hao A Xu <hao.a.xu@intel.com>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83811}
Drive-by fix: reorder members of RandomAccessStackDominatorNode to
save 8 bytes on the total size of Block.
Bug: v8:12783
Change-Id: I4923490b0d2f4de22ea001eeba44c950c6451633
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3893853
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Darius Mercadier <dmercadier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83679}
A BoundedSize is just a regular size_t when the sandbox is disabled.
However, when the sandbox is enabled, a BoundedLength is guaranteed to
be in the range [0, kMaxSafeBufferSizeForSandbox]. This is (currently)
achieved by storing the length shifted to the left, then right-shifting
it when loading it. This guarantees that the top bits are zero.
BoundedSizes are used to ensure safe access to variable-sized buffers,
in particular ArrayBuffers and their views, located inside the sandbox.
If a full size_t is used to represent their size, it may allow an
attacker to "reach out of" the sandbox address space by setting the
length to a very large value. A BoundedSize prevents this.
Bug: chromium:1360375
Change-Id: I0579693db528af96c41eeaa64bd3ed71266aacd9
Cq-Include-Trybots: luci.v8.try.triggered:v8_linux64_no_sandbox_dbg_ng_triggered
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3876823
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Samuel Groß <saelo@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83631}
This is needed to simplify concurrent sweeping for MinorMC.
Also: move evacuation verifiers to a separate file so that they can be
used from heap.cc as well.
Bug: v8:12612
Change-Id: I2a738a31e83a357f4fdded8a30ccb2ff6ba70553
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3932720
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83557}
The CL sligthly generalizes SlotSet by parameterizing it with slot size.
SlotSet is abstracted into BasicSlotSet, which is moved to heap::base::.
V8 GC related parts stay in slot-set.h
Bug: chromium:1029379
Change-Id: I093332b77682d2b31e61a91d4b0110fa95b5c908
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3695595
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83553}
Detach the marking state from the collectors and move them to heap.
Also update users to access via heap and reduce dependencies on the
collectors.
This is a prerequisite for moving sweeper to the heap, which is needed
for concurrent sweeping in MinorMC.
Bug: v8:12612
Change-Id: Ia0bb2b7566b24eeb0d75c911edbfd626f07dad0f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3925548
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83542}
Implementation of property sequences for regular expressions is unused
(likely since switching to icu).
Bug: v8:11935
Change-Id: Ic4cf6219de8d6eb99464292a20f637e1fd423341
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3920135
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Commit-Queue: Jakob Linke <jgruber@chromium.org>
Auto-Submit: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83451}
- Use non-idle tasks to be able to process finalizers on time.
- Only process finalizers while concurrent marking is still running.
Bug: v8:13294
Change-Id: I1a2812c3fc350ea679c4c916c230cf736f2aa3ea
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3904648
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83367}
- Add Turboshaft to the wasm pipeline (behind a flag).
- Add a few operators.
- Implement SimplifyLoopsPhase, which ensures each loop has at most
two inputs.
- Remove the unneeded effect argument from
{FlagContinuation::FromTrap}.
Bug: v8:12783
Change-Id: I03a3f8cf3af40fc75bf57cfbad973b754b13dd8c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3899126
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83365}
This moves the existing PGO code to a separate cc file with a separate
header. As the implementation will be further extended in follow-up CLs,
it's better to have it separated.
R=jkummerow@chromium.org
Bug: v8:13209
Change-Id: I7b7b5bf9c8d3d542dae734f3874499dccee152a9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3899321
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83250}
This CL fixes redesigns the current API, which does not correctly
manage lifetimes of the shared object conveyors.
See design doc at
https://docs.google.com/document/d/1TV6agY9dafVJFvdPrUAGbEvos8wL2WDnsmf84n3OJVU/edit?usp=sharing
This CL also removes the incorrect behavior of serializing all shared
strings by sharing instead of copying. Shared strings may be sent to
another process, which should still work.
Bug: v8:12547
Change-Id: I7413abd2d871fd3d52c9b433445cfa1d03e4a732
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3868713
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83044}
The src/wasm/stacks.h file became progressively bigger, create a
cc file for it and move some function definitions there.
R=clemensb@chromium.org
Bug: v8:12191
Change-Id: I6880a4e11884856bd2f4d777cb745a3c06c71841
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3874936
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83039}
This CL adds src/base/container-utils.h wich contains a few utilities
to make working with containers easier by providing a few additional
functions (e.g. contains, all_equal, ...) that are not (yet) shipped
with C++ standard containers.
Change-Id: I365b88c4286bf58bcac32c7bb89a5b0a98fc3509
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3865966
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82907}
- Remove TruncateInt64ToInt32 instead of translating to Turboshaft, since it has no effect. Removing it simplifies pattern-matching in optimizations.
- Change how exception handling is done in Turboshaft: The exception value is obtained as the result of `CatchExceptionOp` instead of a special projection. This simplifies projections.
- Add `TupleOp` as the counterpart to `ProjectionOp`, which is useful
for lowerings of operations that have multiple outputs.
- Split BinopOp into WordBinopOp and FloatBinopOp because they have quite different semantics and many kinds only exist for one of them.
- rename IntegerUnary to WordUnary and other occurences of
Integer/Integral
- rename ChangeOp::Kind::kUnsignedFloatTruncate` to `kJSFloatTruncate`
because it actually has JS wrap-around semantics.
- move/add representation DCHECKs to operation constructors.
- add some convinience helpers to `AssemblerInterface`.
- Add a mechanism to check which operations are supported by the machine.
Drive-by fix: Abort current block in OptimizationPhase::VisitBlock if
we lower to a block-terminator.
Bug: v8:12783
Change-Id: Ib738accccd22fb1606d9dab86f57ac1e739fcec2
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3857449
Auto-Submit: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Darius Mercadier <dmercadier@chromium.org>
Commit-Queue: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82857}
Methods are now defined in heap-verifier.h in the HeapVerifier class.
Bug: v8:11708
Change-Id: I13e7f1760598f3659ad6aa31082840caf2e44038
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3857558
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82810}
Shared values need to be kept alive while being conveyed across threads (i.e. by
postMessage). Currently they are meant to be conveyed through the
serializer/deserializer by the embedder via API. This both clunky and
the embedder has no good choice for what to do, because the most natural
choice is v8::Global and it is not designed to be threadsafe.
This CL removes the API and transparently handles conveying shared
values by using a wrapper around PersistentHandles called
SharedValueConveyors. Any isolate can own the conveyor provided it
outlives the receipt of the message by the receiving isolate. For
simpler lifetime management, the shared isolate currently owns all
conveyors.
Bug: v8:12547
Change-Id: I8f71b2faa0f8a1973f8b97ffccf4f5ad230f4e16
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3857453
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82795}
- Move StringForwardingTable implementation to own compilation unit.
- Refactoring preparing for layout change (Introduce explicit record
class to make transition from contiguous Tagged_t fields to a
heterogeneous record layout easier).
- Replace RootVisitor pattern for transitioning/cleanup during GC with
callback.
- Minor cleanups.
Bug: v8:12957
Change-Id: Iae343393f470130eac0c54148a1303b67fb95aa4
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3845635
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82730}
That entails that V8 built with bazel does not support
snapshot compression.
NOTE: Snapshot compression has always been supported, but it has never
been enabled. We do not even have a flag for it.
Change-Id: I4b67db618be2c62e6a6f34e71eb0dac2ec26fa36
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3834031
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82537}
- Move InnerPointerToCodeCache to using internal callbacks.
- Refactor internal and external callbacks to use a unified interface.
Bug: v8:13184
Change-Id: If0006d324b0433f5d6bbf00b6d0fc1a2589227bc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3834583
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82515}
StringConstantXXX were introduced when we switched to concurrent
compilation, as a way to build strings in Turbofan in a background
thread, without having to actually allocate them on the main heap
from the background. See https://crrev.com/c/1221807.
Now that we have local heaps, we can actually allocate strings from
the background, making StringConstantXXX useless.
Moreover, we would fold constant string concatenations into
ConsString, which sounds a bit dubious for performance. Now, small
constant string concatenations will be folded into SeqStrings, while
larger ones will remain ConsString, just to avoid the quadratic
worst-case.
Change-Id: I0479d16aa5691c9d774187c4cc0d03ff4fe2b4f9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3811291
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Darius Mercadier <dmercadier@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82381}
This is a reland of commit 6953b5550e
The reland fixes tests that retrieved the stack start from a
non-inlined frame's fp. This does not work in certain configurations
as the resulting marker is too low to consider the first local
variables in subsequent calls.
The fix uses an inline frame address for the tests to get an upper
bound of stack addresses to consider.
Original change's description:
> [handles] Remove precise on-stack representation of global handles
>
> Since https://crrev.com/c/3806439 on-stack traced handles are marked
> conservatively when being used in combination with CppHeap.
>
> This change removes the precise on-stack representation of the
> internal traced nodes as they nodes would anyways be marked
> conservatively. The effects are:
> - cheaper representation (just a single node space);
> - uniform handling: no checks to distinguish on-stack vs on-heap;
> - no brittleness around cleaning on-stack handles when the event loop
> is empty;
>
> Change-Id: Id859623bfed77a66bdd064ea8065536264515eae
> Bug: v8:13141
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3812039
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#82306}
Bug: v8:13141
Change-Id: I53ece36220e99d02be6df18f83c18450e5d5037b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3820585
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82342}
This reverts commit 6953b5550e.
Reason for revert: Failing on CI: https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux%20-%20full%20debug/22876/overview
Original change's description:
> [handles] Remove precise on-stack representation of global handles
>
> Since https://crrev.com/c/3806439 on-stack traced handles are marked
> conservatively when being used in combination with CppHeap.
>
> This change removes the precise on-stack representation of the
> internal traced nodes as they nodes would anyways be marked
> conservatively. The effects are:
> - cheaper representation (just a single node space);
> - uniform handling: no checks to distinguish on-stack vs on-heap;
> - no brittleness around cleaning on-stack handles when the event loop
> is empty;
>
> Change-Id: Id859623bfed77a66bdd064ea8065536264515eae
> Bug: v8:13141
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3812039
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#82306}
Bug: v8:13141
Change-Id: I5ebb2ed9eeec1394c2dd48504e91f6ff789b0711
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3820584
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#82313}
Since https://crrev.com/c/3806439 on-stack traced handles are marked
conservatively when being used in combination with CppHeap.
This change removes the precise on-stack representation of the
internal traced nodes as they nodes would anyways be marked
conservatively. The effects are:
- cheaper representation (just a single node space);
- uniform handling: no checks to distinguish on-stack vs on-heap;
- no brittleness around cleaning on-stack handles when the event loop
is empty;
Change-Id: Id859623bfed77a66bdd064ea8065536264515eae
Bug: v8:13141
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3812039
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82306}
This is a strictly code moving change. We move the diffing algorithm
out of liveedit.cc into its own file.
We want to replace the current diffing algorithm and to stay safe we
will ship old and new algorithm side-by-side for a release for easy
revertability. Given that liveedit.cc is already large enough we
extract the diffing algo into a separate file.
R=kimanh@chromium.org
Bug: chromium:1205288
Change-Id: If5ebb6c2dff2f00387c9e2ab87e4bb61d1f1484a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3802687
Reviewed-by: Kim-Anh Tran <kimanh@chromium.org>
Commit-Queue: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/main@{#82124}
This is a reland of commit e895b7af73
The unit test has been updated to work correctly when
--stress-incremental-marking is enabled.
Original change's description:
> Background merging of deserialized scripts
>
> Recently, https://crrev.com/c/v8/v8/+/3681880 added new API functions
> with which an embedder could request that V8 merge newly deserialized
> script data into an existing Script from the Isolate's compilation
> cache. This change implements those new functions. This functionality is
> still disabled by default due to the flag
> merge_background_deserialized_script_with_compilation_cache.
>
> The goal of this new functionality is to reduce memory usage when
> multiple frames load the same script with a long delay between (long
> enough for the script to have been evicted from Blink's in-memory cache
> and for the top-level SharedFunctionInfo to be flushed). In that case,
> there are two Script objects for the same script: one which was found in
> the Isolate compilation cache (the "old" script), and one which was
> recently deserialized (the "new" script). The new script's object graph
> is essentially standalone: it may point to internalized strings and
> readonly objects such as the empty feedback metadata, but otherwise
> it is unconnected to the rest of the heap. The merging logic takes any
> useful data from the new script's object graph and attaches it into the
> old script's object graph, so that the new Script object and any other
> duplicated objects can be discarded. More specifically:
>
> 1. If the new Script has a SharedFunctionInfo for a particular function
> literal, and the old Script does not, then the old Script is updated
> to refer to the new SharedFunctionInfo.
> 2. If the new Script has a compiled SharedFunctionInfo for a particular
> function literal, and the old Script has an uncompiled
> SharedFunctionInfo, then the old SharedFunctionInfo is updated to
> point to the function_data and feedback_metadata from the new
> SharedFunctionInfo.
> 3. If any used object from the new object graph points to a
> SharedFunctionInfo, where the old object graph contains a matching
> SharedFunctionInfo for the same function literal, then that pointer
> is updated to point to the old SharedFunctionInfo.
>
> The document at [0] includes diagrams showing an example merge on a very
> small script.
>
> Steps 1 and 2 above are pretty simple, but step 3 requires walking a
> possibly large set of objects, so this new API lets the embedder run
> step 3 from a background thread. Steps 1 and 2 are performed later, on
> the main thread.
>
> The next important question is: in what ways can the old script's object
> graph be modified during the background execution of step 3, or during
> the time after step 3 but before steps 1 and 2?
>
> A. SharedFunctionInfos can go from compiled to uncompiled due to
> flushing. This is okay; the worst outcome is that the function would
> need to be compiled again later. Such a risk is already present,
> since V8 doesn't keep IsCompiledScopes for every compiled function in
> a background-deserialized script.
> B. SharedFunctionInfos can go from uncompiled to compiled due to lazy
> compilation. This is also okay; the merge completion logic on the
> main thread will just keep this lazily compiled data rather than
> inserting compiled data from the newly deserialized object graph.
> C. SharedFunctionInfos can be cleared from the Script's weak array if
> they are no longer referenced. This is mostly okay, because any
> SharedFunctionInfo that is needed by the background merge is strongly
> referenced and therefore can't be cleared. The only problem arises if
> the top-level SharedFunctionInfo gets cleared, so the merge task must
> deliberately keep a reference to that one.
> D. SharedFunctionInfos can be created if they are needed due to lazy
> compilation of a parent function. This change is somewhat troublesome
> because it invalidates the background thread's work and requires a
> re-traversal on the main thread to update any pointers that should
> point to this lazily compiled SharedFunctionInfo.
>
> At a high level, this change implements three previously unimplemented
> functions in BackgroundDeserializeTask (in compiler.cc) and updates one:
>
> - BackgroundDeserializeTask::SourceTextAvailable, run on the main
> thread, checks whether there is a matching Script in the Isolate
> compilation cache which doesn't already have a top-level
> SharedFunctionInfo. If so, it saves that Script in a persistent
> handle.
> - BackgroundDeserializeTask::ShouldMergeWithExistingScript checks
> whether the persistent handle from the first step exists (a fast
> operation which can be called from any thread).
> - BackgroundDeserializeTask::MergeWithExistingScript, run on a
> background thread, performs step 3 of the merge described above and
> generates lists of persistent data describing how the main thread can
> complete the merge.
> - BackgroundDeserializeTask::Finish is updated to perform the merge
> steps 1 and 2 listed above, as well as a possible re-traversal of the
> graph if required due to newly created SharedFunctionInfos in the old
> Script.
>
> The merge logic has nothing to do with deserialization, and indeed I
> hope to reuse it for background compilation tasks as well, so it is all
> contained within a new class BackgroundMergeTask (in compiler.h,cc). It
> uses a second class, ForwardPointersVisitor (in compiler.cc) to perform
> the object visitation that updates pointers to SharedFunctionInfos.
>
> [0] https://docs.google.com/document/d/1UksB5Vm7TT1-f3S9W1dK_rP9jKn_ly0WVm_UDPpWuBw/edit
>
> Bug: v8:12808
> Change-Id: Id405869e9d5b106ca7afd9c4b08cb5813e6852c6
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3739232
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Cr-Commit-Position: refs/heads/main@{#81941}
Bug: v8:12808
Change-Id: Id2036dfa4eba8670cac899773d7a906825fa2c50
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3787266
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#82045}
The implicit copy constructor triggers a deprecation warning if the
struct contains a deprecated field. We can fix this by explicitly
declaring the copy and move constructors and assignment operators
with the deprecation warning disabled.
This CL also adds a test to check that we can indeed call the
constructors and assignment operators, which did not work before.
R=leszeks@chromium.org
Bug: v8:13092
Change-Id: Ia63ff9375de13fc6e5b5a8d59d827a742c99fb39
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3785145
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#81962}
This reverts commit e895b7af73.
Reason for revert: TSAN failures: https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux64%20TSAN%20-%20stress-incremental-marking/8468/overview
Original change's description:
> Background merging of deserialized scripts
>
> Recently, https://crrev.com/c/v8/v8/+/3681880 added new API functions
> with which an embedder could request that V8 merge newly deserialized
> script data into an existing Script from the Isolate's compilation
> cache. This change implements those new functions. This functionality is
> still disabled by default due to the flag
> merge_background_deserialized_script_with_compilation_cache.
>
> The goal of this new functionality is to reduce memory usage when
> multiple frames load the same script with a long delay between (long
> enough for the script to have been evicted from Blink's in-memory cache
> and for the top-level SharedFunctionInfo to be flushed). In that case,
> there are two Script objects for the same script: one which was found in
> the Isolate compilation cache (the "old" script), and one which was
> recently deserialized (the "new" script). The new script's object graph
> is essentially standalone: it may point to internalized strings and
> readonly objects such as the empty feedback metadata, but otherwise
> it is unconnected to the rest of the heap. The merging logic takes any
> useful data from the new script's object graph and attaches it into the
> old script's object graph, so that the new Script object and any other
> duplicated objects can be discarded. More specifically:
>
> 1. If the new Script has a SharedFunctionInfo for a particular function
> literal, and the old Script does not, then the old Script is updated
> to refer to the new SharedFunctionInfo.
> 2. If the new Script has a compiled SharedFunctionInfo for a particular
> function literal, and the old Script has an uncompiled
> SharedFunctionInfo, then the old SharedFunctionInfo is updated to
> point to the function_data and feedback_metadata from the new
> SharedFunctionInfo.
> 3. If any used object from the new object graph points to a
> SharedFunctionInfo, where the old object graph contains a matching
> SharedFunctionInfo for the same function literal, then that pointer
> is updated to point to the old SharedFunctionInfo.
>
> The document at [0] includes diagrams showing an example merge on a very
> small script.
>
> Steps 1 and 2 above are pretty simple, but step 3 requires walking a
> possibly large set of objects, so this new API lets the embedder run
> step 3 from a background thread. Steps 1 and 2 are performed later, on
> the main thread.
>
> The next important question is: in what ways can the old script's object
> graph be modified during the background execution of step 3, or during
> the time after step 3 but before steps 1 and 2?
>
> A. SharedFunctionInfos can go from compiled to uncompiled due to
> flushing. This is okay; the worst outcome is that the function would
> need to be compiled again later. Such a risk is already present,
> since V8 doesn't keep IsCompiledScopes for every compiled function in
> a background-deserialized script.
> B. SharedFunctionInfos can go from uncompiled to compiled due to lazy
> compilation. This is also okay; the merge completion logic on the
> main thread will just keep this lazily compiled data rather than
> inserting compiled data from the newly deserialized object graph.
> C. SharedFunctionInfos can be cleared from the Script's weak array if
> they are no longer referenced. This is mostly okay, because any
> SharedFunctionInfo that is needed by the background merge is strongly
> referenced and therefore can't be cleared. The only problem arises if
> the top-level SharedFunctionInfo gets cleared, so the merge task must
> deliberately keep a reference to that one.
> D. SharedFunctionInfos can be created if they are needed due to lazy
> compilation of a parent function. This change is somewhat troublesome
> because it invalidates the background thread's work and requires a
> re-traversal on the main thread to update any pointers that should
> point to this lazily compiled SharedFunctionInfo.
>
> At a high level, this change implements three previously unimplemented
> functions in BackgroundDeserializeTask (in compiler.cc) and updates one:
>
> - BackgroundDeserializeTask::SourceTextAvailable, run on the main
> thread, checks whether there is a matching Script in the Isolate
> compilation cache which doesn't already have a top-level
> SharedFunctionInfo. If so, it saves that Script in a persistent
> handle.
> - BackgroundDeserializeTask::ShouldMergeWithExistingScript checks
> whether the persistent handle from the first step exists (a fast
> operation which can be called from any thread).
> - BackgroundDeserializeTask::MergeWithExistingScript, run on a
> background thread, performs step 3 of the merge described above and
> generates lists of persistent data describing how the main thread can
> complete the merge.
> - BackgroundDeserializeTask::Finish is updated to perform the merge
> steps 1 and 2 listed above, as well as a possible re-traversal of the
> graph if required due to newly created SharedFunctionInfos in the old
> Script.
>
> The merge logic has nothing to do with deserialization, and indeed I
> hope to reuse it for background compilation tasks as well, so it is all
> contained within a new class BackgroundMergeTask (in compiler.h,cc). It
> uses a second class, ForwardPointersVisitor (in compiler.cc) to perform
> the object visitation that updates pointers to SharedFunctionInfos.
>
> [0] https://docs.google.com/document/d/1UksB5Vm7TT1-f3S9W1dK_rP9jKn_ly0WVm_UDPpWuBw/edit
>
> Bug: v8:12808
> Change-Id: Id405869e9d5b106ca7afd9c4b08cb5813e6852c6
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3739232
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Cr-Commit-Position: refs/heads/main@{#81941}
Bug: v8:12808
Change-Id: I82a080e6287828445293cb6b4b94a5e8f15eb8f3
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3787213
Auto-Submit: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Owners-Override: Deepti Gandluri <gdeepti@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#81943}
