The main changes of this CL are:
It should no longer be assumed that an empty ArrayBuffer has a nullptr
backing store. This is in preparation for the move to caged pointers,
which cannot represent nullptr, and will instead likely provide a
EmptyBackingStore constant pointing inside the virtual memory cage. For
that reason, a new JSArrayBuffer::IsEmpty() helper is introduced, which
should be used instead of checking against nullptr.
CodeStubAssembler::GetTypedArrayBuffer now checks for on-heap
TypedArrays instead of comparing the backing store pointer to nullptr.
This is consistent with the implementation in JSTypedArray::GetBuffer.
v8::ArrayBufferView::CopyContents now uses JSTypedArray::DataPtr instead
of relying on nullptr backing stores to handle on-heap TypedArrays.
The serializer and deserializer now check for IsEmpty() and use the
kEmptyBackingStoreRefSentinel value to serialize empty backing stores.
Empty ArrayBuffers allocated for on-heap TypedArrays now have a
byte_length of zero. This allows removing the allocation_length() (and
allocation_buffer()) methods, which were only (incorrectly, as they
don't account for GSABs) used for memory measurements.
Bug: chromium:1218005
Change-Id: Ib889ccf855f68525f7a614f3963e46ea56865fa3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3297709
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: Samuel Groß <saelo@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78069}
Rename flags to align them with other flags that are named in an
enabled way and drop the "never" prefix.
Drive-by: Refactor compaction entry point.
Bug: v8:12251
Change-Id: If2b189152f3cd22038b87fe3cc2ba0db4953ae23
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3270534
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77813}
The tests are modeled after another patch that includes
v8::CFunctions into Node.js's builtin snapshot.
Refs: https://github.com/nodejs/node/pull/40649
Change-Id: I5a91682f7944ef06a0d3caf7333b09f974bcd64b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3251138
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Commit-Queue: Joyee Cheung <joyee@igalia.com>
Cr-Commit-Position: refs/heads/main@{#77726}
- Introduce v8::ScriptCompiler::CompileFunction
- Deprecate v8::ScriptCompiler::CompileFunctionInContext
- Add v8::Function::GetUnboundScript
- Add v8::Script::GetResourceName
The ScriptOrModule out-parameter is only used by NodeJS since we don't
allow arbitrary objects has host-defined options and they need a way to
keep the options alive.
This CL deprecates the out-parameter and adds helper methods to
address the most common use-cases.
The final fix still requires more fundamental changes on how host-defined
options are handled.
Bug: chromium:1244145
Change-Id: Id29de53521ad626c41391b8300146ee37a1b8a51
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3245117
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Auto-Submit: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77564}
For the upcoming host_defined_options fixes we will have to explicitly
pass the host-defined options to Invoke so we will be able to install
it in the script context in the future.
Bug: chromium:1244145
Change-Id: I690cc774d6a17278db4381aba8c3408e979606c1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3222765
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77524}
This is a reland of 929b83fb7b
This version of the CL also fixes initialization of the
marking_barrier_ in the LocalHeap constructor.
This CL also got rebased on Victor's CL in https://crrev.com/c/3229361.
It added a code_space_allocator_ in LocalHeap which needs to be
initialized a bit later on the main thread as well.
Original change's description:
> [heap] Attach to shared isolate after setting up main thread
>
> Attach to the shared isolate after the main thread was set up. Otherwise
> it could happen that a shared GC initiated from another isolate might
> see no threads are running and performs the safepoint operation in the
> middle of isolate deserialization.
>
> We use DisallowSafepoints to check that the isolate doesn't join a
> global safepoint before deserialization is complete. DisallowSafepoints
> used to prevent only invocations of Safepoint() but was updated to
> also prevent Park() and Unpark() invocations. Each state change could
> cause the thread to reach a safepoint, which would allow a shared GC
> to run.
>
> We now also DCHECK that every isolate has at least one local heap and
> that shared collections aren't started before deserialization is
> complete.
>
> Bug: v8:11708
> Change-Id: Iba3fb59dd951d5ee4fc9934158062287302fc279
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3221157
> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Reviewed-by: Shu-yu Guo <syg@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#77424}
Bug: v8:11708
Change-Id: I7d44e4a5f76cc09092c2444cede10e9331222c1d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3229361
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77448}
This reverts commit 929b83fb7b.
Reason for revert: https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux64%20UBSan/18725/overview
Original change's description:
> [heap] Attach to shared isolate after setting up main thread
>
> Attach to the shared isolate after the main thread was set up. Otherwise
> it could happen that a shared GC initiated from another isolate might
> see no threads are running and performs the safepoint operation in the
> middle of isolate deserialization.
>
> We use DisallowSafepoints to check that the isolate doesn't join a
> global safepoint before deserialization is complete. DisallowSafepoints
> used to prevent only invocations of Safepoint() but was updated to
> also prevent Park() and Unpark() invocations. Each state change could
> cause the thread to reach a safepoint, which would allow a shared GC
> to run.
>
> We now also DCHECK that every isolate has at least one local heap and
> that shared collections aren't started before deserialization is
> complete.
>
> Bug: v8:11708
> Change-Id: Iba3fb59dd951d5ee4fc9934158062287302fc279
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3221157
> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Reviewed-by: Shu-yu Guo <syg@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#77424}
Bug: v8:11708
Change-Id: I0633150b6b40b297a335a39bf1a087ca93592e04
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3225937
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Owners-Override: Nico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Nico Hartmann <nicohartmann@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77425}
Attach to the shared isolate after the main thread was set up. Otherwise
it could happen that a shared GC initiated from another isolate might
see no threads are running and performs the safepoint operation in the
middle of isolate deserialization.
We use DisallowSafepoints to check that the isolate doesn't join a
global safepoint before deserialization is complete. DisallowSafepoints
used to prevent only invocations of Safepoint() but was updated to
also prevent Park() and Unpark() invocations. Each state change could
cause the thread to reach a safepoint, which would allow a shared GC
to run.
We now also DCHECK that every isolate has at least one local heap and
that shared collections aren't started before deserialization is
complete.
Bug: v8:11708
Change-Id: Iba3fb59dd951d5ee4fc9934158062287302fc279
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3221157
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77424}
Tip of tree puts both internalized and in-place-internalizable strings
into the shared heap object cache. But only internalized strings need
to go in there, since we can't have duplicates of those. It's fine to
allocate in-place-internalizable strings in the shared heap each time
a new Isolate is initialized, it'll be deduplicated if it's
internalized eventually.
Bug: chromium:1258918, v8:12007
Change-Id: I0e46b73a5ac3be83d0eaa31915a3a24f47a8c2bd
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3219690
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77388}
This CL adds a new snapshot to hold objects that are in the shared heap
or may need to be in the shared heap depending on runtime flags.
Currently this is to support --shared-string-table, which puts all
in-place-internalizable strings, internalized strings, and the
string table into the shared heap.
The shared heap snapshot is never deserialized into client Isolates.
This means when V8 is started without a shared Isolate, the shared heap
snapshot is deserialized into all Isolates.
Bug: v8:12007
Change-Id: I7eeab73080cda2e8250a5a49747f25b2440a349d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3173905
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77309}
.. and refactor js-regexp.h.
- Hide the generic DataAt/SetDataAt accessors and replace them by
dedicated accessors. Use the common lower_case naming scheme for
these.
- Shuffle around definitions in js-regexp.h s.t. they are in a
meaningful order.
- Dedupe the source/flags accessors - these fields are stored both
on the instance and on the data array. We keep only accessors for
the instance. Previously, these were disambiguated through naming
oddities (e.g. Pattern() returned data->source).
Change-Id: I3d53c8b095f0d59621ff779608438f7fa5e8c92a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3193534
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77138}
This means we don't need to copy over properties, and accessors stay in
place similar to when we deserialize a custom snapshot.
This slightly changes the semantics of Context::New, so let's see
whether someone depends on this behaviour. We may need to revert if so
(hopefully until we can update the embedder).
Bug: v8:12113
Change-Id: I8325480a00bab5b2bb6ea42274e295b0d4dfc85c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3162143
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76928}
This reverts commit 810d34dfe6.
Reason for revert: The stricter host checks prevent
certain security issues. We will have to live with regressions
until we have a more flexible caching solution in place.
Original change's description:
> [codegen] Disable host-defined options checks in cache
>
> We see too many regressions for now in M94 (~10% more misses in
> some cases).
>
> This CL reverts the logic to the state before landing
> https://crrev.com/c/3069152 without having to revert the several
> refactoring CLs that landed on top of it.
>
> Bug: v8:10284, chromium:1238312, chromium:1237242
> Change-Id: I57e66b9e0d58c36d2f1563b07720e3729c88ec94
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3103006
> Commit-Queue: Camillo Bruni <cbruni@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#76362}
Bug: v8:10284, chromium:1238312, chromium:1237242
Change-Id: I4c662dd0ac16a4406f06fb2a62b9e4e65fa428ce
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3114057
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76504}
This is a reland of d1b27019d3
Fixes include:
Adding missing file to bazel build
Forward-declaring classing before friend-classing them to fix win/gcc
Add missing v8-isolate.h include for vtune builds
Original change's description:
> [include] Split out v8.h
>
> This moves every single class/function out of include/v8.h into a
> separate header in include/, which v8.h then includes so that
> externally nothing appears to have changed.
>
> Every include of v8.h from inside v8 has been changed to a more
> fine-grained include.
>
> Previously inline functions defined at the bottom of v8.h would call
> private non-inline functions in the V8 class. Since that class is now
> in v8-initialization.h and is rarely included (as that would create
> dependency cycles), this is not possible and so those methods have been
> moved out of the V8 class into the namespace v8::api_internal.
>
> None of the previous files in include/ now #include v8.h, which means
> if embedders were relying on this transitive dependency then it will
> give compile failures.
>
> v8-inspector.h does depend on v8-scripts.h for the time being to ensure
> that Chrome continue to compile but that change will be reverted once
> those transitive #includes in chrome are changed to include it directly.
>
> Full design:
> https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing
>
> Bug: v8:11965
> Change-Id: I53b84b29581632710edc80eb11f819c2097a2877
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Reviewed-by: Camillo Bruni <cbruni@chromium.org>
> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Commit-Queue: Dan Elphick <delphick@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#76424}
Cq-Include-Trybots: luci.v8.try:v8_linux_vtunejit
Bug: v8:11965
Change-Id: I99f5d3a73bf8fe25b650adfaf9567dc4e44a09e6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3113629
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76460}
Re-use the same check we already have in place for the
compilation cache for when we use CodeSerializer::Deserialize.
- Move HasOrigin to SharedFunctionInfo::HasMatchingOrigin
- HasMatchingOrigin no longer allocates
- Pass ScriptDetails in more places
Bug: v8:10284
Change-Id: I6e074bd1e7db9a35fdf7123d04a65841d9813e02
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3090968
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76451}
This reverts commit d1b27019d3.
Reason for revert: Broke vtune build, tsan build and possibly others
Original change's description:
> [include] Split out v8.h
>
> This moves every single class/function out of include/v8.h into a
> separate header in include/, which v8.h then includes so that
> externally nothing appears to have changed.
>
> Every include of v8.h from inside v8 has been changed to a more
> fine-grained include.
>
> Previously inline functions defined at the bottom of v8.h would call
> private non-inline functions in the V8 class. Since that class is now
> in v8-initialization.h and is rarely included (as that would create
> dependency cycles), this is not possible and so those methods have been
> moved out of the V8 class into the namespace v8::api_internal.
>
> None of the previous files in include/ now #include v8.h, which means
> if embedders were relying on this transitive dependency then it will
> give compile failures.
>
> v8-inspector.h does depend on v8-scripts.h for the time being to ensure
> that Chrome continue to compile but that change will be reverted once
> those transitive #includes in chrome are changed to include it directly.
>
> Full design:
> https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing
>
> Bug: v8:11965
> Change-Id: I53b84b29581632710edc80eb11f819c2097a2877
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Reviewed-by: Camillo Bruni <cbruni@chromium.org>
> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Commit-Queue: Dan Elphick <delphick@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#76424}
Bug: v8:11965
Change-Id: Id57313ae992e720c8b19abc975cd69729e1344aa
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3113627
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Owners-Override: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76428}
This moves every single class/function out of include/v8.h into a
separate header in include/, which v8.h then includes so that
externally nothing appears to have changed.
Every include of v8.h from inside v8 has been changed to a more
fine-grained include.
Previously inline functions defined at the bottom of v8.h would call
private non-inline functions in the V8 class. Since that class is now
in v8-initialization.h and is rarely included (as that would create
dependency cycles), this is not possible and so those methods have been
moved out of the V8 class into the namespace v8::api_internal.
None of the previous files in include/ now #include v8.h, which means
if embedders were relying on this transitive dependency then it will
give compile failures.
v8-inspector.h does depend on v8-scripts.h for the time being to ensure
that Chrome continue to compile but that change will be reverted once
those transitive #includes in chrome are changed to include it directly.
Full design:
https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing
Bug: v8:11965
Change-Id: I53b84b29581632710edc80eb11f819c2097a2877
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76424}
We see too many regressions for now in M94 (~10% more misses in
some cases).
This CL reverts the logic to the state before landing
https://crrev.com/c/3069152 without having to revert the several
refactoring CLs that landed on top of it.
Bug: v8:10284, chromium:1238312, chromium:1237242
Change-Id: I57e66b9e0d58c36d2f1563b07720e3729c88ec94
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3103006
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76362}
Make off-thread deserialization play well with the Isolate compilation
cache, by moving the Finish call into GetSharedFunctionInfoForScript.
This means that
a) The isolate cache is checked before the Finish, allowing it to be
hit, and
b) Results of off-thread deserializations are written into the Isolate
cache.
Bug: chromium:1075999
Change-Id: I535935180bbe77f3e718253830e649bd62857634
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3094006
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76341}
The previous CL https://crrev.com/c/3069152 only did a pointer equality
check for host defined options. This broke code caching for chrome.
This CL extends the check to use a shallow strict equals check on the
host defined options elements.
Bug: v8:10284, chromium:1237242
Change-Id: Ie0ab17a5f5abe024061b6c3d3d68367d9e92b78b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3081607
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76169}
To consume a code cache off-thread
1. The embedder creates a CachedData object wrapping the data blob.
2. The embedder calls ScriptCompiler::StartConsumingCodeCache with the
CachedData, and receives a ScriptCompiler::CodeCacheConsumeTask
which takes ownership of the CachedData.
3. The embedder calls ScriptCompiler::CodeCacheConsumeTask::Run
on a different thread.
4. Once this completes, the embedded passes the completed task as an
optional argument into Source constructor, and calls Compile as
before.
This is roughly similar to how streaming compilation works, with the
QoL improvement that Source owns the CodeCacheConsumeTask and therefore
we can reuse the same Compile method and do the off-thread finalization
behind the scenes inside Compile.
On the v8::internal side, ScriptCompiler::CodeCacheConsumeTask wraps a
v8::internal::BackgroundDeserializeTask, which has a Run and a Finish
method. The Run creates a LocalIsolate (again, similar to
BackgroundCompileTask), calls some helpers on CodeSerializer, and stores
the pre-finalization result in a OffThreadDeserializeData structure.
This stores Persistent Handles to the off-thread initialized SFI and
a vector of Scripts needing fixing up, and it owns the PersistentHandles
object which owns those Handles. Finally, the Finish method consumes
this OffThreadDeserializeData structure, fixes up Scripts, moves the
SFI Handle into the caller HandleScope, and that's it.
Since we don't yet have the source at off-thread deserialization time,
the various code cache sanity checks are done without the source hash
when deserializing, and the Finish method re-does them now that the
source is available.
Bug: chromium:1075999
Change-Id: If1faf35ba3ef840fa4e735581d0b29c96c1d5fc8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3067322
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76155}
Optimizing compilation can no longer collect source positions on demand
since it may now run concurrently without serialization.
Instead, we now collect full source positions when any component that
needs them is enabled (profiler, debugger).
Bug: v8:7790,v8:12030
Change-Id: I6a2a82eb2b0d3e92121e101b4d9bf330c1f6c065
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3067226
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Mythri Alle <mythria@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76114}
Follow-up CLs will use the ScriptDetails object for code cache lookups
instead of only the ScriptOriginOptions.
Bug: v8:10284
Change-Id: Idc83e6e79cfca283369a9b5ceab8bc53dae5f2dd
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3069149
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76073}
The adding of base:: was mostly prepared using git grep and sed:
git grep -l <pattern> | grep -v base/vector.h | \
xargs sed -i 's/\b<pattern>\b/base::<pattern>/
with lots of manual clean-ups due to the resulting
v8::internal::base::Vectors.
#includes were fixed using:
git grep -l "src/utils/vector.h" | \
axargs sed -i 's!src/utils/vector.h!src/base/vector.h!'
Bug: v8:11879
Change-Id: I3e6d622987fee4478089c40539724c19735bd625
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2968412
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#75243}
Remove the requirement to pass the native context into the script cache,
simple to be able to access the empty function. Instead, for script
cache keys, use Smi::zero() in the 'owner function' slot.
This allows CompileUnboundScript to be called outside of a Context
scope.
Change-Id: I9b4fe6dd43f14944728664f7203b748ced750e76
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2794440
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73737}
This reverts commit 3a6f75ac94.
Reason for revert: performance regressions https://bugs.chromium.org/p/chromium/issues/detail?id=1163063
Original change's description:
> [objects] Remove MakeExternal case for uncached internal strings
>
> Concurrently accessing internal external uncached strings is not
> thread-safe. We are removing a case where we can make such a string
> through MakeExternal.
>
> Bug: v8:7790
> Change-Id: I958062c15cf40ccc330600bb572de98620866e54
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2565511
> Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org>
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#71573}
TBR=leszeks@chromium.org,solanes@chromium.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: v8:7790
Change-Id: I5dcc734869c3c921eacd89426309141127a85f47
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2633547
Reviewed-by: Santiago Aboy Solanes <solanes@chromium.org>
Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72123}
Remove the ambient dependency on the currently entered isolate, let the
embedder pass it in explicitly.
Bug: v8:11287
Change-Id: I03690390a308a59e2c6ea5c6ae268780d836b717
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2608209
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72105}
This CL changes SharedFunctionInfo::GetBytecodeArray to a function
template, which is specialized for Isolate and LocalIsolate arguments.
This allows main thread only uses to avoid taking a lock.
Bug: v8:7790, chromium:1154603
Change-Id: I3462c4e36b66073e09393c01c765dd8a018a98f0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2595307
Commit-Queue: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71833}
Concurrently accessing internal external uncached strings is not
thread-safe. We are removing a case where we can make such a string
through MakeExternal.
Bug: v8:7790
Change-Id: I958062c15cf40ccc330600bb572de98620866e54
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2565511
Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71573}
Scopes in V8 are used to guarantee one or more properties during its
lifetimes. If a scope is not named e.g MyClassScope(args) instead of
MyClassScope scope(args) it will get created and automatically destroyed
and therefore, being useless as a scope. This CL would produce a
compiling warning when that happens to ward off this developer error.
Follow-up to ccrev.com/2552415 in which it was introduced and
implemented for Guard classes.
Change-Id: Ifa0fb89cc3d9bdcdee0fd8150a2618af5ef45cbf
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2555001
Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71425}
Because of LocalHeap safepoints, our existing assert scopes don't
necessarily maintain the same guarantees as desired. In particular,
DisallowHeapAllocation no longer guarantees that objects don't move.
This patch transitions DisallowHeapAllocation to
DisallowGarbageCollection, to ensure that code using this scope is
also protected against safepoints.
Change-Id: I0411425884f6849982611205fb17bb072881c722
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2540547
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71319}
* Replace deprecated Factory::NewFunction* calls with JSFunctionBuilder.
* Drive-by: rename Factory::NewFunctionForTest to ..ForTesting (this is
the correct suffix recognized by our tooling to ensure it's only
called from tests).
Tbr: clemensb@chromium.org
Bug: v8:8888
Change-Id: I110063803e5b467bd91b75fe8fea2ca4174f2bcc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2529129
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71101}
This relands commit 3f4e9bbe43.
which was a reland of c4a062a958
which was a reland of 28a30c578c
which was a reland of 5d7a29c90e
The change had an issue that embedders implementing heap tracing (e.g.
Unified Heap with Blink) could be passed an uninitialized pointer if
marking happened during deserialization of an object containing such a
pointer. Because of the 0xdeadbed0 uninitialized filler value, these
embedders would then receive the value 0xdeadbed0deadbed0 as the
'pointer', and crash on dereference.
There is, however, special handling already for null pointers in heap
tracing, also for dealing with not-yet initialized values. So, we can
make the uninitialized Smi filler be 0x00000000, and that will make such
embedded fields have a nullptr representation, making them follow the
normal uninitialized value bailouts.
In addition, it relands the following dependent changes, which are
relanding unchanged and are followup performance improvements.
Relanding them in the same change should allow for cleaner reverts
should they be needed.
This relands commit 76ad3ab597
[identity-map] Change resize heuristic
This relands commit 77cc96aa48
[identity-map] Cache the calculated Hash
This relands commit bee5b996aa
[serializer] Remove Deserializer::Initialize
This relands commit c8f73f2266
[serializer] Cache instance type in PostProcessNewObject
This relands commit 4e7c99abda
[identity-map] Remove double-lookups in IdentityMap
Original change's description:
> Reland^3 "[serializer] Allocate during deserialization"
>
> This is a reland of c4a062a958
> which was a reland of 28a30c578c
> which was a reland of 5d7a29c90e
>
> Fixes TSAN errors from non-atomic writes in the deserializer. Now all
> writes are (relaxed) atomic.
>
> Original change's description:
> > Reland^2 "[serializer] Allocate during deserialization"
> >
> > This is a reland of 28a30c578c
> > which was a reland of 5d7a29c90e
> >
> > The crashes were from calling RegisterDeserializerFinished on a null
> > Isolate pointer, for a deserializer that was never initialised
> > (specifically, ReadOnlyDeserializer when ROHeap is shared).
> >
> > Original change's description:
> > > Reland "[serializer] Allocate during deserialization"
> > >
> > > This is a reland of 5d7a29c90e
> > >
> > > This reland shuffles around the order of checks in Heap::AllocateRawWith
> > > to not check the new space addresses until it's known that this is a new
> > > space allocation. This fixes an UBSan failure during read-only space
> > > deserialization, which happens before the new space is initialized.
> > >
> > > It also fixes some issues discovered by --stress-snapshot, around
> > > serializing ThinStrings (which are now elided as part of serialization),
> > > handle counts (I bumped the maximum handle count in that check), and
> > > clearing map transitions (the map backpointer field needed a Smi
> > > uninitialized value check).
> > >
> > > Original change's description:
> > > > [serializer] Allocate during deserialization
> > > >
> > > > This patch removes the concept of reservations and a specialized
> > > > deserializer allocator, and instead makes the deserializer allocate
> > > > directly with the Heap's Allocate method.
> > > >
> > > > The major consequence of this is that the GC can now run during
> > > > deserialization, which means that:
> > > >
> > > > a) Deserialized objects are visible to the GC, and
> > > > b) Objects that the deserializer/deserialized objects point to can
> > > > move.
> > > >
> > > > Point a) is mostly not a problem due to previous work in making
> > > > deserialized objects "GC valid", i.e. making sure that they have a valid
> > > > size before any subsequent allocation/safepoint. We now additionally
> > > > have to initialize the allocated space with a valid tagged value -- this
> > > > is a magic Smi value to keep "uninitialized" checks simple.
> > > >
> > > > Point b) is solved by Handlifying the deserializer. This involves
> > > > changing any vectors of objects into vectors of Handles, and any object
> > > > keyed map into an IdentityMap (we can't use Handles as keys because
> > > > the object's address is no longer a stable hash).
> > > >
> > > > Back-references can no longer be direct chunk offsets, so instead the
> > > > deserializer stores a Handle to each deserialized object, and the
> > > > backreference is an index into this handle array. This encoding could
> > > > be optimized in the future with e.g. a second pass over the serialized
> > > > array which emits a different bytecode for objects that are and aren't
> > > > back-referenced.
> > > >
> > > > Additionally, the slot-walk over objects to initialize them can no
> > > > longer use absolute slot offsets, as again an object may move and its
> > > > slot address would become invalid. Now, slots are walked as relative
> > > > offsets to a Handle to the object, or as absolute slots for the case of
> > > > root pointers. A concept of "slot accessor" is introduced to share the
> > > > code between these two modes, and writing the slot (including write
> > > > barriers) is abstracted into this accessor.
> > > >
> > > > Finally, the Code body walk is modified to deserialize all objects
> > > > referred to by RelocInfos before doing the RelocInfo walk itself. This
> > > > is because RelocInfoIterator uses raw pointers, so we cannot allocate
> > > > during a RelocInfo walk.
> > > >
> > > > As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> > > > size rather than byte size -- the size is expected to be tagged-aligned
> > > > anyway, so now we get an extra few bits in the size encoding.
> > > >
> > > > Bug: chromium:1075999
> > > > Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> > > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> > > > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > > > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > > > Cr-Commit-Position: refs/heads/master@{#70229}
Bug: chromium:1075999
Change-Id: Ib514a4ef16bd02bfb60d046ecbf8fae1ead64a98
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2452689
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70366}
This reverts commit 3f4e9bbe43, along
with the following dependent changes (reverted to make this a clean revert):
76ad3ab597 [identity-map] Change resize heuristic
77cc96aa48 [identity-map] Cache the calculated Hash
bee5b996aa [serializer] Remove Deserializer::Initialize
c8f73f2266 [serializer] Cache instance type in PostProcessNewObject
4e7c99abda [identity-map] Remove double-lookups in IdentityMap
Reason for revert: major crash spike on Canary (https://crbug.com/1135027)
Original change's description:
> Reland^3 "[serializer] Allocate during deserialization"
>
> This is a reland of c4a062a958
> which was a reland of 28a30c578c
> which was a reland of 5d7a29c90e
>
> Fixes TSAN errors from non-atomic writes in the deserializer. Now all
> writes are (relaxed) atomic.
>
> Original change's description:
> > Reland^2 "[serializer] Allocate during deserialization"
> >
> > This is a reland of 28a30c578c
> > which was a reland of 5d7a29c90e
> >
> > The crashes were from calling RegisterDeserializerFinished on a null
> > Isolate pointer, for a deserializer that was never initialised
> > (specifically, ReadOnlyDeserializer when ROHeap is shared).
> >
> > Original change's description:
> > > Reland "[serializer] Allocate during deserialization"
> > >
> > > This is a reland of 5d7a29c90e
> > >
> > > This reland shuffles around the order of checks in Heap::AllocateRawWith
> > > to not check the new space addresses until it's known that this is a new
> > > space allocation. This fixes an UBSan failure during read-only space
> > > deserialization, which happens before the new space is initialized.
> > >
> > > It also fixes some issues discovered by --stress-snapshot, around
> > > serializing ThinStrings (which are now elided as part of serialization),
> > > handle counts (I bumped the maximum handle count in that check), and
> > > clearing map transitions (the map backpointer field needed a Smi
> > > uninitialized value check).
> > >
> > > Original change's description:
> > > > [serializer] Allocate during deserialization
> > > >
> > > > This patch removes the concept of reservations and a specialized
> > > > deserializer allocator, and instead makes the deserializer allocate
> > > > directly with the Heap's Allocate method.
> > > >
> > > > The major consequence of this is that the GC can now run during
> > > > deserialization, which means that:
> > > >
> > > > a) Deserialized objects are visible to the GC, and
> > > > b) Objects that the deserializer/deserialized objects point to can
> > > > move.
> > > >
> > > > Point a) is mostly not a problem due to previous work in making
> > > > deserialized objects "GC valid", i.e. making sure that they have a valid
> > > > size before any subsequent allocation/safepoint. We now additionally
> > > > have to initialize the allocated space with a valid tagged value -- this
> > > > is a magic Smi value to keep "uninitialized" checks simple.
> > > >
> > > > Point b) is solved by Handlifying the deserializer. This involves
> > > > changing any vectors of objects into vectors of Handles, and any object
> > > > keyed map into an IdentityMap (we can't use Handles as keys because
> > > > the object's address is no longer a stable hash).
> > > >
> > > > Back-references can no longer be direct chunk offsets, so instead the
> > > > deserializer stores a Handle to each deserialized object, and the
> > > > backreference is an index into this handle array. This encoding could
> > > > be optimized in the future with e.g. a second pass over the serialized
> > > > array which emits a different bytecode for objects that are and aren't
> > > > back-referenced.
> > > >
> > > > Additionally, the slot-walk over objects to initialize them can no
> > > > longer use absolute slot offsets, as again an object may move and its
> > > > slot address would become invalid. Now, slots are walked as relative
> > > > offsets to a Handle to the object, or as absolute slots for the case of
> > > > root pointers. A concept of "slot accessor" is introduced to share the
> > > > code between these two modes, and writing the slot (including write
> > > > barriers) is abstracted into this accessor.
> > > >
> > > > Finally, the Code body walk is modified to deserialize all objects
> > > > referred to by RelocInfos before doing the RelocInfo walk itself. This
> > > > is because RelocInfoIterator uses raw pointers, so we cannot allocate
> > > > during a RelocInfo walk.
> > > >
> > > > As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> > > > size rather than byte size -- the size is expected to be tagged-aligned
> > > > anyway, so now we get an extra few bits in the size encoding.
> > > >
> > > > Bug: chromium:1075999
> > > > Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> > > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> > > > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > > > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > > > Cr-Commit-Position: refs/heads/master@{#70229}
> > >
> > > Bug: chromium:1075999
> > > Change-Id: Ibc77cc48b3440b4a28b09746cfc47e50c340ce54
> > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440828
> > > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > > Auto-Submit: Leszek Swirski <leszeks@chromium.org>
> > > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > > Cr-Commit-Position: refs/heads/master@{#70267}
> >
> > Tbr: jgruber@chromium.org,ulan@chromium.org
> > Bug: chromium:1075999
> > Change-Id: Iaa8dc54895866ada0e34a7c9e8fff9ae1cb13f2d
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2444991
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70279}
>
> Tbr: jgruber@chromium.org,ulan@chromium.org
> Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel_ng,v8_linux64_tsan_no_cm_rel_ng,v8_linux64_tsan_isolates_rel_ng
> Bug: chromium:1075999
> Change-Id: I0b9b11644aebc4cc8b07c62a0f765b24e4d73d89
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2445872
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Auto-Submit: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70288}
TBR=ulan@chromium.org,jgruber@chromium.org,leszeks@chromium.org,dinfuehr@chromium.org
Bug: chromium:1075999, chromium:1135027
Change-Id: I5d0d9e49c0302d94ff7291834f5f18e7a0839eb7
Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel_ng,v8_linux64_tsan_no_cm_rel_ng,v8_linux64_tsan_isolates_rel_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2451030
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70328}
Remove the separate Initialize method from Deserializer, opting instead
to pass around SnapshotData where appropriate and pass the isolate
directly into the Deserializer's constructor.
Change-Id: I0092fadd9c81f14b2ce75145fd81af37c3947c65
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2448466
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70310}
This is a reland of c4a062a958
which was a reland of 28a30c578c
which was a reland of 5d7a29c90e
Fixes TSAN errors from non-atomic writes in the deserializer. Now all
writes are (relaxed) atomic.
Original change's description:
> Reland^2 "[serializer] Allocate during deserialization"
>
> This is a reland of 28a30c578c
> which was a reland of 5d7a29c90e
>
> The crashes were from calling RegisterDeserializerFinished on a null
> Isolate pointer, for a deserializer that was never initialised
> (specifically, ReadOnlyDeserializer when ROHeap is shared).
>
> Original change's description:
> > Reland "[serializer] Allocate during deserialization"
> >
> > This is a reland of 5d7a29c90e
> >
> > This reland shuffles around the order of checks in Heap::AllocateRawWith
> > to not check the new space addresses until it's known that this is a new
> > space allocation. This fixes an UBSan failure during read-only space
> > deserialization, which happens before the new space is initialized.
> >
> > It also fixes some issues discovered by --stress-snapshot, around
> > serializing ThinStrings (which are now elided as part of serialization),
> > handle counts (I bumped the maximum handle count in that check), and
> > clearing map transitions (the map backpointer field needed a Smi
> > uninitialized value check).
> >
> > Original change's description:
> > > [serializer] Allocate during deserialization
> > >
> > > This patch removes the concept of reservations and a specialized
> > > deserializer allocator, and instead makes the deserializer allocate
> > > directly with the Heap's Allocate method.
> > >
> > > The major consequence of this is that the GC can now run during
> > > deserialization, which means that:
> > >
> > > a) Deserialized objects are visible to the GC, and
> > > b) Objects that the deserializer/deserialized objects point to can
> > > move.
> > >
> > > Point a) is mostly not a problem due to previous work in making
> > > deserialized objects "GC valid", i.e. making sure that they have a valid
> > > size before any subsequent allocation/safepoint. We now additionally
> > > have to initialize the allocated space with a valid tagged value -- this
> > > is a magic Smi value to keep "uninitialized" checks simple.
> > >
> > > Point b) is solved by Handlifying the deserializer. This involves
> > > changing any vectors of objects into vectors of Handles, and any object
> > > keyed map into an IdentityMap (we can't use Handles as keys because
> > > the object's address is no longer a stable hash).
> > >
> > > Back-references can no longer be direct chunk offsets, so instead the
> > > deserializer stores a Handle to each deserialized object, and the
> > > backreference is an index into this handle array. This encoding could
> > > be optimized in the future with e.g. a second pass over the serialized
> > > array which emits a different bytecode for objects that are and aren't
> > > back-referenced.
> > >
> > > Additionally, the slot-walk over objects to initialize them can no
> > > longer use absolute slot offsets, as again an object may move and its
> > > slot address would become invalid. Now, slots are walked as relative
> > > offsets to a Handle to the object, or as absolute slots for the case of
> > > root pointers. A concept of "slot accessor" is introduced to share the
> > > code between these two modes, and writing the slot (including write
> > > barriers) is abstracted into this accessor.
> > >
> > > Finally, the Code body walk is modified to deserialize all objects
> > > referred to by RelocInfos before doing the RelocInfo walk itself. This
> > > is because RelocInfoIterator uses raw pointers, so we cannot allocate
> > > during a RelocInfo walk.
> > >
> > > As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> > > size rather than byte size -- the size is expected to be tagged-aligned
> > > anyway, so now we get an extra few bits in the size encoding.
> > >
> > > Bug: chromium:1075999
> > > Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> > > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > > Cr-Commit-Position: refs/heads/master@{#70229}
> >
> > Bug: chromium:1075999
> > Change-Id: Ibc77cc48b3440b4a28b09746cfc47e50c340ce54
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440828
> > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > Auto-Submit: Leszek Swirski <leszeks@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70267}
>
> Tbr: jgruber@chromium.org,ulan@chromium.org
> Bug: chromium:1075999
> Change-Id: Iaa8dc54895866ada0e34a7c9e8fff9ae1cb13f2d
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2444991
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70279}
Tbr: jgruber@chromium.org,ulan@chromium.org
Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel_ng,v8_linux64_tsan_no_cm_rel_ng,v8_linux64_tsan_isolates_rel_ng
Bug: chromium:1075999
Change-Id: I0b9b11644aebc4cc8b07c62a0f765b24e4d73d89
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2445872
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70288}
This reverts commit c4a062a958.
Reason for revert: TSan issues: https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20TSAN/33504
Original change's description:
> Reland^2 "[serializer] Allocate during deserialization"
>
> This is a reland of 28a30c578c
> which was a reland of 5d7a29c90e
>
> The crashes were from calling RegisterDeserializerFinished on a null
> Isolate pointer, for a deserializer that was never initialised
> (specifically, ReadOnlyDeserializer when ROHeap is shared).
>
> Original change's description:
> > Reland "[serializer] Allocate during deserialization"
> >
> > This is a reland of 5d7a29c90e
> >
> > This reland shuffles around the order of checks in Heap::AllocateRawWith
> > to not check the new space addresses until it's known that this is a new
> > space allocation. This fixes an UBSan failure during read-only space
> > deserialization, which happens before the new space is initialized.
> >
> > It also fixes some issues discovered by --stress-snapshot, around
> > serializing ThinStrings (which are now elided as part of serialization),
> > handle counts (I bumped the maximum handle count in that check), and
> > clearing map transitions (the map backpointer field needed a Smi
> > uninitialized value check).
> >
> > Original change's description:
> > > [serializer] Allocate during deserialization
> > >
> > > This patch removes the concept of reservations and a specialized
> > > deserializer allocator, and instead makes the deserializer allocate
> > > directly with the Heap's Allocate method.
> > >
> > > The major consequence of this is that the GC can now run during
> > > deserialization, which means that:
> > >
> > > a) Deserialized objects are visible to the GC, and
> > > b) Objects that the deserializer/deserialized objects point to can
> > > move.
> > >
> > > Point a) is mostly not a problem due to previous work in making
> > > deserialized objects "GC valid", i.e. making sure that they have a valid
> > > size before any subsequent allocation/safepoint. We now additionally
> > > have to initialize the allocated space with a valid tagged value -- this
> > > is a magic Smi value to keep "uninitialized" checks simple.
> > >
> > > Point b) is solved by Handlifying the deserializer. This involves
> > > changing any vectors of objects into vectors of Handles, and any object
> > > keyed map into an IdentityMap (we can't use Handles as keys because
> > > the object's address is no longer a stable hash).
> > >
> > > Back-references can no longer be direct chunk offsets, so instead the
> > > deserializer stores a Handle to each deserialized object, and the
> > > backreference is an index into this handle array. This encoding could
> > > be optimized in the future with e.g. a second pass over the serialized
> > > array which emits a different bytecode for objects that are and aren't
> > > back-referenced.
> > >
> > > Additionally, the slot-walk over objects to initialize them can no
> > > longer use absolute slot offsets, as again an object may move and its
> > > slot address would become invalid. Now, slots are walked as relative
> > > offsets to a Handle to the object, or as absolute slots for the case of
> > > root pointers. A concept of "slot accessor" is introduced to share the
> > > code between these two modes, and writing the slot (including write
> > > barriers) is abstracted into this accessor.
> > >
> > > Finally, the Code body walk is modified to deserialize all objects
> > > referred to by RelocInfos before doing the RelocInfo walk itself. This
> > > is because RelocInfoIterator uses raw pointers, so we cannot allocate
> > > during a RelocInfo walk.
> > >
> > > As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> > > size rather than byte size -- the size is expected to be tagged-aligned
> > > anyway, so now we get an extra few bits in the size encoding.
> > >
> > > Bug: chromium:1075999
> > > Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> > > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > > Cr-Commit-Position: refs/heads/master@{#70229}
> >
> > Bug: chromium:1075999
> > Change-Id: Ibc77cc48b3440b4a28b09746cfc47e50c340ce54
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440828
> > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > Auto-Submit: Leszek Swirski <leszeks@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70267}
>
> Tbr: jgruber@chromium.org,ulan@chromium.org
> Bug: chromium:1075999
> Change-Id: Iaa8dc54895866ada0e34a7c9e8fff9ae1cb13f2d
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2444991
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70279}
TBR=ulan@chromium.org,jgruber@chromium.org,leszeks@chromium.org
Change-Id: Ib2f01db4cd9b55639d6a4af971bda865edb45e84
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:1075999
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2445250
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70280}
This is a reland of 28a30c578c
which was a reland of 5d7a29c90e
The crashes were from calling RegisterDeserializerFinished on a null
Isolate pointer, for a deserializer that was never initialised
(specifically, ReadOnlyDeserializer when ROHeap is shared).
Original change's description:
> Reland "[serializer] Allocate during deserialization"
>
> This is a reland of 5d7a29c90e
>
> This reland shuffles around the order of checks in Heap::AllocateRawWith
> to not check the new space addresses until it's known that this is a new
> space allocation. This fixes an UBSan failure during read-only space
> deserialization, which happens before the new space is initialized.
>
> It also fixes some issues discovered by --stress-snapshot, around
> serializing ThinStrings (which are now elided as part of serialization),
> handle counts (I bumped the maximum handle count in that check), and
> clearing map transitions (the map backpointer field needed a Smi
> uninitialized value check).
>
> Original change's description:
> > [serializer] Allocate during deserialization
> >
> > This patch removes the concept of reservations and a specialized
> > deserializer allocator, and instead makes the deserializer allocate
> > directly with the Heap's Allocate method.
> >
> > The major consequence of this is that the GC can now run during
> > deserialization, which means that:
> >
> > a) Deserialized objects are visible to the GC, and
> > b) Objects that the deserializer/deserialized objects point to can
> > move.
> >
> > Point a) is mostly not a problem due to previous work in making
> > deserialized objects "GC valid", i.e. making sure that they have a valid
> > size before any subsequent allocation/safepoint. We now additionally
> > have to initialize the allocated space with a valid tagged value -- this
> > is a magic Smi value to keep "uninitialized" checks simple.
> >
> > Point b) is solved by Handlifying the deserializer. This involves
> > changing any vectors of objects into vectors of Handles, and any object
> > keyed map into an IdentityMap (we can't use Handles as keys because
> > the object's address is no longer a stable hash).
> >
> > Back-references can no longer be direct chunk offsets, so instead the
> > deserializer stores a Handle to each deserialized object, and the
> > backreference is an index into this handle array. This encoding could
> > be optimized in the future with e.g. a second pass over the serialized
> > array which emits a different bytecode for objects that are and aren't
> > back-referenced.
> >
> > Additionally, the slot-walk over objects to initialize them can no
> > longer use absolute slot offsets, as again an object may move and its
> > slot address would become invalid. Now, slots are walked as relative
> > offsets to a Handle to the object, or as absolute slots for the case of
> > root pointers. A concept of "slot accessor" is introduced to share the
> > code between these two modes, and writing the slot (including write
> > barriers) is abstracted into this accessor.
> >
> > Finally, the Code body walk is modified to deserialize all objects
> > referred to by RelocInfos before doing the RelocInfo walk itself. This
> > is because RelocInfoIterator uses raw pointers, so we cannot allocate
> > during a RelocInfo walk.
> >
> > As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> > size rather than byte size -- the size is expected to be tagged-aligned
> > anyway, so now we get an extra few bits in the size encoding.
> >
> > Bug: chromium:1075999
> > Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70229}
>
> Bug: chromium:1075999
> Change-Id: Ibc77cc48b3440b4a28b09746cfc47e50c340ce54
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440828
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Auto-Submit: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70267}
Tbr: jgruber@chromium.org,ulan@chromium.org
Bug: chromium:1075999
Change-Id: Iaa8dc54895866ada0e34a7c9e8fff9ae1cb13f2d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2444991
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70279}
This reverts commit 28a30c578c.
Reason for revert: Broke Test262 https://ci.chromium.org/p/v8/builders/ci/V8%20Linux%20-%20shared/38638?
Original change's description:
> Reland "[serializer] Allocate during deserialization"
>
> This is a reland of 5d7a29c90e
>
> This reland shuffles around the order of checks in Heap::AllocateRawWith
> to not check the new space addresses until it's known that this is a new
> space allocation. This fixes an UBSan failure during read-only space
> deserialization, which happens before the new space is initialized.
>
> It also fixes some issues discovered by --stress-snapshot, around
> serializing ThinStrings (which are now elided as part of serialization),
> handle counts (I bumped the maximum handle count in that check), and
> clearing map transitions (the map backpointer field needed a Smi
> uninitialized value check).
>
> Original change's description:
> > [serializer] Allocate during deserialization
> >
> > This patch removes the concept of reservations and a specialized
> > deserializer allocator, and instead makes the deserializer allocate
> > directly with the Heap's Allocate method.
> >
> > The major consequence of this is that the GC can now run during
> > deserialization, which means that:
> >
> > a) Deserialized objects are visible to the GC, and
> > b) Objects that the deserializer/deserialized objects point to can
> > move.
> >
> > Point a) is mostly not a problem due to previous work in making
> > deserialized objects "GC valid", i.e. making sure that they have a valid
> > size before any subsequent allocation/safepoint. We now additionally
> > have to initialize the allocated space with a valid tagged value -- this
> > is a magic Smi value to keep "uninitialized" checks simple.
> >
> > Point b) is solved by Handlifying the deserializer. This involves
> > changing any vectors of objects into vectors of Handles, and any object
> > keyed map into an IdentityMap (we can't use Handles as keys because
> > the object's address is no longer a stable hash).
> >
> > Back-references can no longer be direct chunk offsets, so instead the
> > deserializer stores a Handle to each deserialized object, and the
> > backreference is an index into this handle array. This encoding could
> > be optimized in the future with e.g. a second pass over the serialized
> > array which emits a different bytecode for objects that are and aren't
> > back-referenced.
> >
> > Additionally, the slot-walk over objects to initialize them can no
> > longer use absolute slot offsets, as again an object may move and its
> > slot address would become invalid. Now, slots are walked as relative
> > offsets to a Handle to the object, or as absolute slots for the case of
> > root pointers. A concept of "slot accessor" is introduced to share the
> > code between these two modes, and writing the slot (including write
> > barriers) is abstracted into this accessor.
> >
> > Finally, the Code body walk is modified to deserialize all objects
> > referred to by RelocInfos before doing the RelocInfo walk itself. This
> > is because RelocInfoIterator uses raw pointers, so we cannot allocate
> > during a RelocInfo walk.
> >
> > As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> > size rather than byte size -- the size is expected to be tagged-aligned
> > anyway, so now we get an extra few bits in the size encoding.
> >
> > Bug: chromium:1075999
> > Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> > Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> > Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70229}
>
> Bug: chromium:1075999
> Change-Id: Ibc77cc48b3440b4a28b09746cfc47e50c340ce54
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440828
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Auto-Submit: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70267}
TBR=ulan@chromium.org,jgruber@chromium.org,leszeks@chromium.org
Change-Id: Ieed68332ef6a7ad36db061e3f48be0f28673d7a2
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:1075999
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2441608
Reviewed-by: Zhi An Ng <zhin@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70268}
This is a reland of 5d7a29c90e
This reland shuffles around the order of checks in Heap::AllocateRawWith
to not check the new space addresses until it's known that this is a new
space allocation. This fixes an UBSan failure during read-only space
deserialization, which happens before the new space is initialized.
It also fixes some issues discovered by --stress-snapshot, around
serializing ThinStrings (which are now elided as part of serialization),
handle counts (I bumped the maximum handle count in that check), and
clearing map transitions (the map backpointer field needed a Smi
uninitialized value check).
Original change's description:
> [serializer] Allocate during deserialization
>
> This patch removes the concept of reservations and a specialized
> deserializer allocator, and instead makes the deserializer allocate
> directly with the Heap's Allocate method.
>
> The major consequence of this is that the GC can now run during
> deserialization, which means that:
>
> a) Deserialized objects are visible to the GC, and
> b) Objects that the deserializer/deserialized objects point to can
> move.
>
> Point a) is mostly not a problem due to previous work in making
> deserialized objects "GC valid", i.e. making sure that they have a valid
> size before any subsequent allocation/safepoint. We now additionally
> have to initialize the allocated space with a valid tagged value -- this
> is a magic Smi value to keep "uninitialized" checks simple.
>
> Point b) is solved by Handlifying the deserializer. This involves
> changing any vectors of objects into vectors of Handles, and any object
> keyed map into an IdentityMap (we can't use Handles as keys because
> the object's address is no longer a stable hash).
>
> Back-references can no longer be direct chunk offsets, so instead the
> deserializer stores a Handle to each deserialized object, and the
> backreference is an index into this handle array. This encoding could
> be optimized in the future with e.g. a second pass over the serialized
> array which emits a different bytecode for objects that are and aren't
> back-referenced.
>
> Additionally, the slot-walk over objects to initialize them can no
> longer use absolute slot offsets, as again an object may move and its
> slot address would become invalid. Now, slots are walked as relative
> offsets to a Handle to the object, or as absolute slots for the case of
> root pointers. A concept of "slot accessor" is introduced to share the
> code between these two modes, and writing the slot (including write
> barriers) is abstracted into this accessor.
>
> Finally, the Code body walk is modified to deserialize all objects
> referred to by RelocInfos before doing the RelocInfo walk itself. This
> is because RelocInfoIterator uses raw pointers, so we cannot allocate
> during a RelocInfo walk.
>
> As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> size rather than byte size -- the size is expected to be tagged-aligned
> anyway, so now we get an extra few bits in the size encoding.
>
> Bug: chromium:1075999
> Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70229}
Bug: chromium:1075999
Change-Id: Ibc77cc48b3440b4a28b09746cfc47e50c340ce54
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440828
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70267}
This reverts commit 5d7a29c90e.
Reason for revert: UBSan -- https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20UBSan/13100
Original change's description:
> [serializer] Allocate during deserialization
>
> This patch removes the concept of reservations and a specialized
> deserializer allocator, and instead makes the deserializer allocate
> directly with the Heap's Allocate method.
>
> The major consequence of this is that the GC can now run during
> deserialization, which means that:
>
> a) Deserialized objects are visible to the GC, and
> b) Objects that the deserializer/deserialized objects point to can
> move.
>
> Point a) is mostly not a problem due to previous work in making
> deserialized objects "GC valid", i.e. making sure that they have a valid
> size before any subsequent allocation/safepoint. We now additionally
> have to initialize the allocated space with a valid tagged value -- this
> is a magic Smi value to keep "uninitialized" checks simple.
>
> Point b) is solved by Handlifying the deserializer. This involves
> changing any vectors of objects into vectors of Handles, and any object
> keyed map into an IdentityMap (we can't use Handles as keys because
> the object's address is no longer a stable hash).
>
> Back-references can no longer be direct chunk offsets, so instead the
> deserializer stores a Handle to each deserialized object, and the
> backreference is an index into this handle array. This encoding could
> be optimized in the future with e.g. a second pass over the serialized
> array which emits a different bytecode for objects that are and aren't
> back-referenced.
>
> Additionally, the slot-walk over objects to initialize them can no
> longer use absolute slot offsets, as again an object may move and its
> slot address would become invalid. Now, slots are walked as relative
> offsets to a Handle to the object, or as absolute slots for the case of
> root pointers. A concept of "slot accessor" is introduced to share the
> code between these two modes, and writing the slot (including write
> barriers) is abstracted into this accessor.
>
> Finally, the Code body walk is modified to deserialize all objects
> referred to by RelocInfos before doing the RelocInfo walk itself. This
> is because RelocInfoIterator uses raw pointers, so we cannot allocate
> during a RelocInfo walk.
>
> As a drive-by, the VariableRawData bytecode is tweaked to use tagged
> size rather than byte size -- the size is expected to be tagged-aligned
> anyway, so now we get an extra few bits in the size encoding.
>
> Bug: chromium:1075999
> Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70229}
TBR=ulan@chromium.org,jgruber@chromium.org,leszeks@chromium.org
Change-Id: I2bd792a24861e8f54897e51522769b50f8f814e2
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:1075999
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440827
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70231}
This patch removes the concept of reservations and a specialized
deserializer allocator, and instead makes the deserializer allocate
directly with the Heap's Allocate method.
The major consequence of this is that the GC can now run during
deserialization, which means that:
a) Deserialized objects are visible to the GC, and
b) Objects that the deserializer/deserialized objects point to can
move.
Point a) is mostly not a problem due to previous work in making
deserialized objects "GC valid", i.e. making sure that they have a valid
size before any subsequent allocation/safepoint. We now additionally
have to initialize the allocated space with a valid tagged value -- this
is a magic Smi value to keep "uninitialized" checks simple.
Point b) is solved by Handlifying the deserializer. This involves
changing any vectors of objects into vectors of Handles, and any object
keyed map into an IdentityMap (we can't use Handles as keys because
the object's address is no longer a stable hash).
Back-references can no longer be direct chunk offsets, so instead the
deserializer stores a Handle to each deserialized object, and the
backreference is an index into this handle array. This encoding could
be optimized in the future with e.g. a second pass over the serialized
array which emits a different bytecode for objects that are and aren't
back-referenced.
Additionally, the slot-walk over objects to initialize them can no
longer use absolute slot offsets, as again an object may move and its
slot address would become invalid. Now, slots are walked as relative
offsets to a Handle to the object, or as absolute slots for the case of
root pointers. A concept of "slot accessor" is introduced to share the
code between these two modes, and writing the slot (including write
barriers) is abstracted into this accessor.
Finally, the Code body walk is modified to deserialize all objects
referred to by RelocInfos before doing the RelocInfo walk itself. This
is because RelocInfoIterator uses raw pointers, so we cannot allocate
during a RelocInfo walk.
As a drive-by, the VariableRawData bytecode is tweaked to use tagged
size rather than byte size -- the size is expected to be tagged-aligned
anyway, so now we get an extra few bits in the size encoding.
Bug: chromium:1075999
Change-Id: I672c42f553f2669888cc5e35d692c1b8ece1845e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2404451
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70229}
This change moves external pointers into a separate table and turns
external pointers in heap objects into indices into that table.
This CL implements one of two possible ownership models for the table
entries. With this one, every heap object owns its table entries, and
they are allocated when the owning object is allocated. As such, setting
external pointer fields does not require allocation of table entries. On
the other hand, table indices cannot be shared between multiple objects.
This CL does not yet implement freeing of external pointer table
entires. This will later happen by a table garbage collector.
Bug: v8:10391
Change-Id: I4d37785295c25a7d1dcbc9871dd5887b9d788a4f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2235700
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Samuel Groß <saelo@google.com>
Cr-Commit-Position: refs/heads/master@{#70204}
Samuel Groß
Michael Lippautz
Maya Lekova
Joyee Cheung
Ng Zhi An
Camillo Bruni
Dominik Inführ
Nico Hartmann
Shu-yu Guo
Jakob Gruber
Toon Verwaest
Dan Elphick
Leszek Swirski
Wenyu Zhao
Gus Caplan
Santiago Aboy Solanes
Ben Noordhuis
Adam Klein
Clemens Backes
Samuel Groß