As part of making the compiler as well as hot-reload independent of
a particular isolate we move this boolean to `IsolateGroup`.
For hot-reloading isolate groups there are more changes needed that
come in the future. This CL is preparation for making the compiler
independent of current isolate.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Refactoring of existing code.
Change-Id: I434cf1180bc963462e2901dfa83e915df9dc1712
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182561
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
As part of making the compiler independent of the current isolate (since
JITed code will be shared across all isolate) we make the compiler not
depend on global field state, with one exception: we preserve an existing
optimization that utilizes knowledge whether a global final field was
already initialized (if --enable-isolate-groups is turned off) - we do
so by asking the isolate group for it's own isolate member.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Refactoring of existing code.
Change-Id: I2f08d854af6102e05e5a1df8d5b66b514d6f3ce4
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182540
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
When using reachability fence as an optimization barrier (e.g. to
prevent dead code elimination or allocation sinking) it is helpful
to reduce fence's impact on optimizations themselves.
TEST=ci
Change-Id: I4581c73e4704e94870a9942e05e16ab38ca2d864
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182625
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
As part of making the compiler independent of a particular isolate
(since all isolates in a group share the same JITed code) we move the
boxed field list to IsolateGroup.
Any isolate that hits the lazy compile stub will then drain the fields
to be deoptimized and performs that work.
A follow-up investigation can decide whether it's beneficial (or
hurtful) for single vs multi isolate scenarios to deoptimize the field
immediately in BG compiler (and interrupt mutator). It is possibly
disadvantagues for single isolate scenarios to do so, since it causes
more mutator interruptions.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Refactoring of existing code.
Change-Id: I64b42ad99056b43a7264bcc30280b9ba39389a2e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182505
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
This is follow up to 6681bce565
the intent was always to perform a sign-extending 4 bytes load.
TEST=ci
Cq-Include-Trybots: luci.dart.try:vm-kernel-linux-release-simarm64-try,vm-kernel-precomp-linux-release-simarm64-try
Change-Id: Ie2b733c4de9bd564cb666d54c9e7b483dbfbeabb
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182261
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
When handling hot reload via library tag handler's Kernel tag
VM must guarantee that external typed data it receives from the
embedder stays alive for as long as VM has any structures
referencing this kernel binary. This is achieved by attaching
original typed data to KernelProgramInfo objects containing
the views into it.
Unfortunately kernel::Reader APIs were somewhat unsafe: allowing
to construct reader object from Program's raw buffer and forget to
set the link betwen the reader and original typed data.
This created a situation where reloading using a multicomponent
Kernel binary would result in KernelProgramInfo objects without
link to the original typed data, which in turn leads to premature
finalization of external typed data and subsequent crashes
when trying to use delete kernel binary.
This CL reworks kernel::Reader API (by introducing kernel::ProgramBinary
wrapper) to make it safer and make sure that connection is preserved.
TEST=vm/cc/IsolateReload_RegressB179030011
Bug: b/179030011
Change-Id: I05f8c31c3cb8e67de6e94a20d9501a5f476b7e27
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182280
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Option --write_code_comments_as_synthetic_source_to=file tells AOT
compiler to create a file which contains code comments for all code
objects written into an ELF snapshot and then emit a DWARF line number
program into .debug_line section which attributes code to code comments
in a way similar to how our own disassembler does it.
This enables tools like objdump to display our code comments as part of
the disassembly.
This CL also tweaks ifdefs in such a way that IL printer and code comments
facilities is now included into PRODUCT gen_snapshot binary.
TEST=manually run product build with --print-flow-graph and --write_code_comments_as_synthetic_source_to
Change-Id: Id6741013d43e1733b4ddeb34891a4d2fc06b9313
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/181380
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
For microbenchmarking it is good to be able to force VM to drop
all stack overflow checks in a given function to reduce overhead of the
measuring loop to the minimum.
This pragma might have other applications, e.g. to allow writing tight
loops in core libraries.
TEST=CheckStackOverflowElimination_NoInterruptsPragma
Change-Id: I07c7fec5c49a893bbb0e4f557d89f022b90be141
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/181400
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
When --dwarf-stack-trace option is used, we don't use nor write inlined
functions information to the snapshot. However, we were retaining
Function objects corresponding to inlined functions.
On a large app, AOT-compiled in PRODUCT mode with
--dwarf-stack-traces on arm64:
Function objects in AOT snapshot -19.4%
Heap size used by objects in snapshot -4.45%
Total AOT snapshot size -2.35%
Flutter gallery AOT snapshot size
(release mode with --dwarf-stack-traces):
arm64 -1.1% (gzip -1.3%, brotli -1.4%)
arm -0.95% (gzip -1.2%, brotli -1.4%)
TEST=ci
Change-Id: Id2d7a2d6b6bff1a5929be97477b0e17a9afae475
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/181700
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
This statistics is printed with --print_snapshot_sizes_verbose.
It is useful in order to figure out how much memory is consumed by
program structure objects without running an application.
TEST=manual testing with --print_snapshot_sizes_verbose
Change-Id: I6eeae7b303ecaf44806deb0f532058492c1d5808
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/181444
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Inlining will evaluate metadata, which can involve creating constants if the metadata wasn't previously evaluated. We currently can't handle creating new constants in the compiler because we lack a mechanism to trigger the field guard deopts from threads other than the main thread. Compare the general inlining case, which aborts if the inlining candidate wasn't previously compiled (on the main thread).
TEST=ci
Bug: https://github.com/dart-lang/sdk/issues/44714
Change-Id: Ib06d5697be6151f373f3242345be49d5c6ec61be
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/180084
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This CL removes asserts that prevented background compiler from doing class finalization.
It adds program_lock synchronization for when object properties are updated during class finalization.
It also encloses code instructions installation into safepoint scopes to ensure that no code is executed during installation.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=existing test suite
Change-Id: If297e44f51b187242eca2cdcfff066c8f6386d97
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/178163
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Currently we have things called XPtr which are not what you get from ptr().
Old world:
handle->raw() returns RawObject* (tagged)
raw_obj->ptr() returns RawObject* (untagged)
After 6fe15f6df9:
handle->raw() returns ObjectPtr
obj_ptr->ptr() returns ObjectLayout*
New world:
handle->ptr() returns ObjectPtr
obj_ptr->untag() returns UntaggedObject*
TEST=ci
Change-Id: I6c7f34014cf20737607caaf84979838300d12df2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/149367
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
When a function type does not describe the signature of a function (e.g. when it just describes a function type literal), the names of the type parameters are not relevant and can be replaced by canonical names allowing proper canonicalization of type literals.
TEST=fixes upcoming language/nnbd/const/type_literals_test
Change-Id: I62924ef619ae3f5a2fe5964f45c8d60990f3aa69
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/179060
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
As per planned breaking change to let platforms decide how they and what throw for late initiaization errors,
we no longer need a public `LateInitializationError` class. It's confusing to have one if some platforms throw
something else instead.
Removes the public abstract class. The dart:_internal implementation class `LateError` no longer implements it.
This is the only implementation of the public interface, and the class which platforms either throw directly,
or through front-end lowering of the feature.
Remove mentions in tests. All tests now just expect `Error`, some platform specific tests might test the message.
TEST=rewrote tests referring to LateInitializationError.
Change-Id: I54344a67f89ce101ed770412db134e12354cdcc4
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/174928
Commit-Queue: Lasse R.H. Nielsen <lrn@google.com>
Reviewed-by: Nate Bosch <nbosch@google.com>
This is a step toward deferring literals, to effectively split binary size for libraries that are more data than code, such as localizations. Bare instructions mode still has all constants in the root snapshot because it has one global ObjectPool.
dart2js+analyzer product X64 non-bare:
app.so 10072728 -> 8479824 (-15.8%)
app.so-2.part.so 14074512 -> 15016976 (6.69%)
app.so-3.part.so 9225936 -> 9888984 (7.18%)
TEST=ci
Bug: https://github.com/dart-lang/sdk/issues/41974
Change-Id: If61f6156e3fc2dd539331335b3bf0458102b7c75
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/168991
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
For normal Dart code a usage of a constant (e.g. from a static const
field) is "inlined": It is represented as a `ConstantExpression` AST
node with a reference to the "evaluated" `Constant` node.
=> Static const fields are never accessed by generated code.
We would therefore not need to represent static const fields in
Kernel/VM anymore. Though there are a few exceptions where the VM's
runtime code needs access to them, which is why they are kept.
The VM runtime code doesn't need an isolate-specific slot for such
static const fields (it's a waste).
=> This CL changes the runtime code that accesses such fields to operate
on the `IsolateGroup`s initial field table instead.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Pure refactoring - relying on existing test coverage.
Change-Id: I5994811a1c1016b2f523b318b698c9e052a98c29
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/177709
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
As part of making the compiler and other subsystems independent
of `Isolate` we have to move various state from `Isolate` to
`IsolateGroup` (or to another place).
The access of `ObjectStore::closure_functions()` was spread out over our
codebase, including from `Isolate`. This CL moves it to a centralized
place - ClosureFunctionsCache.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Pure refactoring - relying on existing test coverage.
Change-Id: I3b282623f07614cdc2c2a863bd5b9d6c7748d36f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/177130
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
As part of making the compiler and other subsystems independent
of `Isolate` we have to move various state from `Isolate` to
`IsolateGroup`.
The class_table and object_store were already moved to `IsolateGroup`.
This CL only replaces usages of `Isolate::{object_store,class_table}`
with the equivalent in `IsolateGroup`.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Pure refactoring - relying on existing test coverage.
Change-Id: I34a0682d715b054d6c5faff077a513980f59a348
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/177126
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Changes fields to be either mutable or immutable by construction.
This ensure that we don't create setter references for fields that
cannot be assigned to and is a prerequisite for replacing @fields/
@fields= canonical names with @getters/@setters.
TEST=existing expectation tests and verification
Change-Id: I70b9a504ee6f221b7c334ac02620feb0d5f7ae01
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/176665
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Dmitry Stefantsov <dmitryas@google.com>
This CL adds support for nested structs in FFI calls, callbacks, and
memory loads and stores through the Struct classes itself.
Nesting empty structs and nesting a structs in themselves (directly or
indirectly) is reported as error.
This feature is almost fully implemented in the CFE transformation.
Because structs depend on the sizes of their nested structs, the structs
now need to be processed in topological order.
Field access to nested structs branches at runtime on making a derived
Pointer if the backing memory of the outer struct was a Pointer or
making a TypedDataView if the backing memory of the outer struct was
a TypedData.
Assigning to a nested struct is a byte for byte copy from the source.
The only changes in the VM are contained in the native calling
convention calculation which now recursively needs to reason about
fundamental types instead of just 1 struct deep.
Because of the amount of corner cases in the calling conventions that
need to be covered, the tests are generated, rather than hand-written.
ABIs tested on CQ: x64 (Linux, MacOS, Windows), ia32 (Linux, Windows),
arm (Android softFP, Linux hardFP), arm64 Android.
ABIs tested locally through Flutter: arm64 iOS.
ABIs not tested: ia32 Android (emulator), x64 iOS (simulator), arm iOS.
TEST=runtime/bin/ffi_test/ffi_test_functions_generated.cc
TEST=runtime/bin/ffi_test/ffi_test_functions.cc
TEST=tests/{ffi,ffi_2}/function_structs_by_value_generated_test.dart
TEST=tests/{ffi,ffi_2}/function_callbacks_structs_by_value_generated_tes
TEST=tests/{ffi,ffi_2}/function_callbacks_structs_by_value_test.dart
TEST=tests/{ffi,ffi_2}/vmspecific_static_checks_test.dart
Closes https://github.com/dart-lang/sdk/issues/37271.
Contains a temporary workaround for
https://github.com/dart-lang/sdk/issues/44454.
Change-Id: I5e5d10e09e5c3fc209f5f7e997efe17bd362214d
Cq-Include-Trybots: luci.dart.try:dart-sdk-linux-try,dart-sdk-mac-try,dart-sdk-win-try,vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-mac-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-kernel-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-ia32-try,vm-kernel-nnbd-mac-release-x64-try,vm-kernel-nnbd-win-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-debug-simarm_x64-try,vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-precomp-win-release-x64-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-precomp-ffi-qemu-linux-release-arm-try,vm-kernel-precomp-obfuscate-linux-release-x64-try,vm-kernel-msan-linux-release-x64-try,vm-kernel-precomp-msan-linux-release-x64-try,vm-kernel-precomp-android-release-arm_x64-try,analyzer-analysis-server-linux-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/169221
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Dmitry Stefantsov <dmitryas@google.com>
When creating new instructions that inherit a token position that
represents a source location from another instruction, the inheriting
instruction must also have the same inlining ID in order for the source
position represented by the token position to be looked up in the
correct script.
Force this by wrapping both in a single InstructionSource struct which
is taken by instructions which take token positions instead of just a
token position. That way, it's more work to manually transfer the token
position separately from the inlining ID of the instruction than doing
the right thing of transfering both at once.
To ensure this information is kept consistent, we pass InstructionSource
structs through the FlowGraphCompiler all the way down to the
CodeSourceMapBuilder.
This CL also makes the following changes:
* Cache the upper bound of source positions in scripts and use it to add
a check for if a given real token position is valid for the script
without iterating over the line starts data for each token position.
* Start inlining intervals appropriately when adding descriptor and
null check information to the code source map.
Code size changes are minimal on Flutter gallery in release mode
(<0.05% decrease).
TEST=Existing tests on trybots, with manual checking with
--check-token-positions that previous errors are now removed.
Bug: https://github.com/dart-lang/sdk/issues/44436
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-x64-try,vm-kernel-linux-debug-x64-try,vm-kernel-linux-release-x64-try,vm-kernel-nnbd-linux-release-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-linux-product-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-debug-simarm_x64-try,vm-kernel-precomp-linux-release-simarm64-try,vm-kernel-linux-release-simarm64-try,vm-kernel-linux-release-simarm-try,vm-kernel-linux-release-ia32-try
Change-Id: I23ced262cb4e9fe9d81356f409e7e8d220d63ee0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/173967
Reviewed-by: Régis Crelier <regis@google.com>
Daco Harkes