Closes: https://github.com/dart-lang/sdk/issues/38158
This CL implements unaligned access for float/double on arm32, but does
not expose it in an API. Rather it only uses these loads/stores inside
the getters and setters of packed structs.
Bug: https://github.com/dart-lang/sdk/issues/45009
Besides unaligned access for float/double, this CL is mostly a CFE
change. The only VM change is reading the packing in the
`_FfiStructLayout` annotation.
The implementation of using the packing in the VM, and analyzer changes
have landed separately in earlier CLs.
tools/test.py ffi ffi_2
TEST=tests/ffi(_2)/(.*)by_value_(*.)_test.dart
Change-Id: Ic3106ecc626d2e30674a49bf03f65ae12d2b3502
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/+/186143
Reviewed-by: Aske Simon Christensen <askesc@google.com>
This is a follow-up to https://dart-review.googlesource.com/c/sdk/+/186680.
After that change, tree shaker started reusing Reference to a Field
when it is replaced with a getter. As a side-effect, this makes
FieldInitializer.field to crash as Reference is now pointing to Procedure,
not a Field.
The fix is to carefully use 'node.fieldReference.asMember' instead of
'node.field' and then retrieve original Field node if it was replaced
with a getter.
TEST=pkg/vm/testcases/transformations/type_flow/transformer/regress_45324_2.dart
Fixes https://github.com/dart-lang/sdk/issues/45324
Change-Id: Ic34e8b9933b00997cd350a4ad93f798c86ac60ad
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/191323
Auto-Submit: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Evaluated constant instances in kernel are represented as a list of
field->value pairs. Tree shaker should not remove fields used in
constants as it may potentially affect identity of constant instances
(non-identical constants may become identical).
Previously, fields used in constants were only marked as written,
so they could still be removed as write-only fields.
Furthermore, if such field was also used as an interface target
it could be converted to an abstract getter, causing the conflict
during kernel AST serialization.
This fix marks fields used in constants as also read, effectively
preventing their tree shaking.
Flutter gallery AOT snapshot size is not affected.
TEST=pkg/vm/testcases/transformations/type_flow/transformer/regress_45324.dart
Fixes https://github.com/dart-lang/sdk/issues/45324
Change-Id: Id6d35604757ad6ba1e23f7227aa9b769448c2688
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/191280
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
The CL is a step towards have a more restricted and wellstructured
handled of references and canonical names.
The CL moves Reference to canonical_name.dart and makes
CanonicalName.reference private, and replaces CanonicalName.getReference
with a 'reference' getter.
It also removes NamedNode.canonicalName, Field.getterCanonicalName and
Field.setterCanonicalName so that these can only be accessed through the
corresponding reference. This is to reduce the reliance on the
canonical names which, ideally, should only be part of serialization and
deserialization.
TEST=existing
Change-Id: I955fb7d52d4e112d8741f7c12dcf38b74ae0c91a
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/190442
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
This CL only changes dart:ffi API, CFE, and analyzer. No VM changes
were needed because the dimensions of inline arrays can be flattened
before passing them to the VM. The multi-dimensionality does not
impact the ABI.
Closes: https://github.com/dart-lang/sdk/issues/45023
TEST=pkg/analyzer/test/src/diagnostics/size_annotation_dimensions_test.dart
TEST=pkg/front_end/testcases/nnbd/ffi_struct_inline_array_multi_dimensional.dart
TEST=tests/ffi/function_structs_by_value_generated_test.dart
TEST=tests/ffi/inline_array_multi_dimensional_test.dart
Change-Id: Ica2c01fccbea7e513879365b34086d8968b54c5b
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/+/188286
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Aske Simon Christensen <askesc@google.com>
This removes the @fields and @=fields canonical name
encodings that allowed for encoding of conflicting members
and didn't support field<->getter/setter conversion
between dills or between outline and full dill.
TEST=existing tests+add aot expectation tests
Change-Id: I119b0c95f90e456356146cdc2d9241de4c1b4fff
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/186680
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
The CFEs FormattedMessage always had two getters to get the text inside
one that would supposedly give an ansi formated version of the message
and one that would supposedly give a plaintext formated version of the
message. They both returned the same string, though, which would either
be with ansi escape codes or plain text depending on the environment at
compile time.
This CL fixes that by having both messages, and letting the reporting
(i.e. whenever the message is read) decide which to use. That way we
can - for instance - report errors with color if the terminal supports
it correctly when reusing a dill (and reissuing problems, but where the
terminal support changes) and if printing the problem to an html <pre>
field (like observatory does).
It also cleans up two different implementations of whether we think
the terminal supports colors or not, by deleting one of them.
This is the second try. Patchset #1 is the original.
Patchset #2(and possibly beyond) is the changes.
TEST=Existing test suites.
Change-Id: I8e483049ce81ce1bd8e5396b588a31e0ad3a8630
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/187402
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Adds support for single dimension inline arrays in structs. Multi-
dimensional arrays will be supported in a future CL.
This CL adds:
- CFE static error checks for inline arrays.
- CFE transformations for inline arrays.
- VM consumption of inline array fields for NativeType.
- Test generator support for inline arrays + generated tests.
Previous CLs added support for inline arrays in:
- analyzer https://dart-review.googlesource.com/c/sdk/+/183684
- updated in this CL to new API.
- ABI calculation https://dart-review.googlesource.com/c/sdk/+/183682
Closes: https://github.com/dart-lang/sdk/issues/35763
Open issue: https://github.com/dart-lang/sdk/issues/45101
CFE transformations are tested with expectation files:
TEST=pkg/front_end/testcases/(.*)/ffi_struct_inline_array.dart
Trampolines and CArray API are tested with end-to-end Dart tests:
TEST=tests/ffi(_2)/(.*)by_value(.*)test.dart
TEST=tests/ffi(_2)/inline_array_test.dart
Compile-time errors (both CFE and analyzer) are tested in:
TEST=tests/ffi(_2)/vmspecific_static_checks_test.dart
Change-Id: I014c0e4153f1b885638adce80de6ab3cac8e6bb2
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/+/183640
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Clement Skau <cskau@google.com>
This reverts commit f63f7736c5.
Reason for revert: This somehow breaks the Windows bots. `python tools/test.py -n dartk-win-release-x64 standalone_2/io/process_shell_test`.
Original change's description:
> [cfe] Actually have both ansi and plain text formatted messages
>
> The CFEs FormattedMessage always had two getters to get the text inside
> one that would supposedly give an ansi formated version of the message
> and one that would supposedly give a plaintext formated version of the
> message. They both returned the same string, though, which would either
> be with ansi escape codes or plain text depending on the environment at
> compile time.
>
> This CL fixes that by having both messages, and letting the reporting
> (i.e. whenever the message is read) decide which to use. That way we
> can - for instance - report errors with color if the terminal supports
> it correctly when reusing a dill (and reissuing problems, but where the
> terminal support changes) and if printing the problem to an html <pre>
> field (like observatory does (1)).
>
> It also cleans up two different implementations of whether we think
> the terminal supports colors or not, by deleting one of them.
>
> (1) At least sometimes. It works - I think - only for 'evaluateInFrame',
> but that's another story (and will be fixed in a follow-up CL).
>
> TEST=Existing test suites.
>
> Change-Id: Iedaedd9a5c41458d40c23ed4b706324c004ae943
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/186291
> Commit-Queue: Jens Johansen <jensj@google.com>
> Reviewed-by: Johnni Winther <johnniwinther@google.com>
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
Change-Id: I0b53f943a61f76705badfead30d9e1ee35baff57
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/186941
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
The CFEs FormattedMessage always had two getters to get the text inside
one that would supposedly give an ansi formated version of the message
and one that would supposedly give a plaintext formated version of the
message. They both returned the same string, though, which would either
be with ansi escape codes or plain text depending on the environment at
compile time.
This CL fixes that by having both messages, and letting the reporting
(i.e. whenever the message is read) decide which to use. That way we
can - for instance - report errors with color if the terminal supports
it correctly when reusing a dill (and reissuing problems, but where the
terminal support changes) and if printing the problem to an html <pre>
field (like observatory does (1)).
It also cleans up two different implementations of whether we think
the terminal supports colors or not, by deleting one of them.
(1) At least sometimes. It works - I think - only for 'evaluateInFrame',
but that's another story (and will be fixed in a follow-up CL).
TEST=Existing test suites.
Change-Id: Iedaedd9a5c41458d40c23ed4b706324c004ae943
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/186291
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This reverts commit 3892e95547.
Reason for revert: Breaks Flutter web
Original change's description:
> [cfe] Encode field references as @getters and @setters
>
> This removes the @fields and @=fields canonical name
> encodings that allowed for encoding of conflicting members
> and didn't support field<->getter/setter conversion
> between dills or between outline and full dill.
>
> TEST=existing tests
>
> Change-Id: Id15e58ad4d1847d2c98a688705e5945196146c6d
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/184783
> Commit-Queue: Johnni Winther <johnniwinther@google.com>
> Reviewed-by: Jens Johansen <jensj@google.com>
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
Change-Id: I744e284b16e097fa0833c5bdf1bc7653f13bdf63
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/186147
Reviewed-by: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
This removes the @fields and @=fields canonical name
encodings that allowed for encoding of conflicting members
and didn't support field<->getter/setter conversion
between dills or between outline and full dill.
TEST=existing tests
Change-Id: Id15e58ad4d1847d2c98a688705e5945196146c6d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/184783
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Before this CL the FFI transformation was always run (for target VM),
so if for instance running dart2js (from source), e.g. like
`out/ReleaseX64/dart -DDFE_VERBOSE=true pkg/compiler/bin/dart2js.dart --help`
one could see how it spends something like 120-140 ms, which is
something along the lines of 2.5% of the entire request processing
(i.e. compile and serialization).
This CL first checks if dart:ffi is imported at all. If it's not the
transformation is skipped.
When compiling dart2js (where, at least currently, dart:ffi is not used)
this skips the transformation, thus saving the 120-140 ms (or ~2.5%).
All measurements on my machine. Your mileage may vary.
TEST=Existing tests.
Change-Id: Ia5d3d7b989f549ca5da257660c204e8c59f15d4b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/185543
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
This CL completes the migration of the first wave of
interdependent libraries in package:kernel, including ast.dart.
In order to ensure non-nullability on AST properties, the Transformer
has been split in 2 variants: Transformer which doesn't support
removal of nodes and RemovingTransformer which supports removal where
allowed by the context using 'removal sentinels'.
Start reviewing Transformer and RemovingTransformer in visitors.dart
since many of the changes are caused by the changes here.
Included in the migration are the mixin_deduplication.dart and
unreachable_code_elimination.dart since these needed porting to
the RemovingTransformer which was aided by opting in the libraries
which only depended on ast.dart.
TEST=existing
Change-Id: I9e63b985bd24896c25edd4ee51e37770187bcc17
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/184786
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
The front end previously enforced a stricter-than-spec requirement on
conflicting imports on its own code. The check was included of the
kernel snapshot and therefore always enforced, even in published sdks.
The extra check was removed a month ago and now tools/sdks/ have been
updated to use a later version of the sdk, so the unneeded hide
combinators can now be removed from the source code.
Closes#44667
TEST=existing
Change-Id: I1d1053b1ef9a40b6a918eef515a02d7b404906c9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/185084
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
Removes the entry-point on the `#sizeOf` field of `Struct` subtypes,
because we are no longer using it in the runtime.
Does not remove the entry-point from the constructor yet, because FFI
trampolines can instantiate these objects. Updated the TODOs to reflect
this.
Bug: https://github.com/dart-lang/sdk/issues/38721
TEST=tests/ffi(_2)/* in precompiled mode.
Change-Id: If3889e782b8fe34ef1c34cf8af83da041b4d2ef5
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/185540
Reviewed-by: Aske Simon Christensen <askesc@google.com>
Commit-Queue: Daco Harkes <dacoharkes@google.com>
As part of implementing hot-reload for isolate groups, we have to ensure
reloads on a group are performed using the same state of incremental
compiler.
This means we cannot use a specific isolate's information (such as it's
main port) for reloading purposes. Instead we use the unqiue isolate
group id when communicating with the kernel service.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Existing test coverage, future test when hot-reload is fully
implemented with isolate groups.
Change-Id: Ifab39cd2ba689c08507dfab4091cd26951ed54e5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/185320
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
This CL changes the semantics of
`Pointer<T extends NativeType>.elementAt` and
`sizeOf<T extends NativeType>` to use the compile-time `T` rather than
the runtime `T`.
Issue: https://github.com/dart-lang/sdk/issues/38721
TEST=tests/ffi/data_test.dart
TEST=tests/ffi/sizeof_test.dart
TEST=tests/ffi/structs_test.dart
TEST=tests/ffi/vmspecific_static_checks_test.dart
Change-Id: Ifb25a4bd66d50a385d3db6dec9213b96dff21722
Cq-Include-Trybots: luci.dart.try:vm-precomp-ffi-qemu-linux-release-arm-try,analyzer-analysis-server-linux-try,analyzer-linux-release-try,analyzer-nnbd-linux-release-try,front-end-linux-release-x64-try,front-end-nnbd-linux-release-x64-try,benchmark-linux-try,dart-sdk-linux-try,pkg-linux-release-try,vm-ffi-android-release-arm-try,vm-ffi-android-release-arm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/178200
Reviewed-by: Aske Simon Christensen <askesc@google.com>
This CL:
* Makes the standard filesystem more synchronous by default.
This makes reading files for normal compiles faster.
* Adds a more asynchronous version of `exists` and `readAsBytes`.
This potentially allows for reading files faster when in a context
where the files can actually be read in parallel. The client have
to make the choise though.
* Skips `Uri.base.resolveUri` when the uri has a scheme. This makes it
slightly faster in my tests, but does not remove any `.` and `..`.
Having those when having a scheme seems sort of weird though.
If this turns out to be a problem we can add it back.
* Re-orders what the standard filesystem checks in `exists` to assume files
which seems more likely to happen. This - for files - means less checks
and should make checking for existance faster. (This was noticed by Siggi).
Using the same measurement procedure as in
https://dart-review.googlesource.com/c/sdk/+/85442
timing how long fasta spends on *reading* files when compiling dart2js
(when compiling via the VM) I get a good speedup.
Note that the measurements are in microseconds:
Before this CL (run like `out/ReleaseX64/dart pkg/compiler/bin/dart2js.dart 2>&1 | grep "Read file (total)" | tail -n 1` 10 times):
Read file (total): 94678
Read file (total): 94232
Read file (total): 91726
Read file (total): 77472
Read file (total): 95582
Read file (total): 88113
Read file (total): 89245
Read file (total): 91575
Read file (total): 96291
Read file (total): 95730
With this CL (run like `out/ReleaseX64/dart pkg/compiler/bin/dart2js.dart 2>&1 | grep "Read file (total)" | tail -n 1` 10 times):
Read file (total): 68379
Read file (total): 69320
Read file (total): 72930
Read file (total): 69692
Read file (total): 68685
Read file (total): 73548
Read file (total): 64649
Read file (total): 71951
Read file (total): 73486
Read file (total): 70621
Difference at 95.0% confidence
-21138.3 +/- 4193
-23.111% +/- 4.5843%
(Student's t, pooled s = 4462.56)
Furthermore, using an internal benchmark (discussed in an email thread)
I get these numbers (note that the measurements are in milliseconds):
before:
4453
4249
4187
4190
4216
after:
2310
2379
2449
2467
2407
Difference at 95.0% confidence
-1856.6 +/- 131.443
-43.5924% +/- 3.08625%
(Student's t, pooled s = 90.1257)
Using the asynchronous versions of exists and readAsBytes and checking for
existance and reading in parallel on the same internal benchmark it gets to
1137
1185
1067
1089
1189
For completion, comparing to using "raw" `File` (again using the internal benchmark):
sequential standard
2377
2419
2485
2373
2431
sequential io_sync
2279
2182
2237
2312
2324
io_sync faster:
-150.2 +/- 76.3122
-6.21432% +/- 3.15731%
(Student's t, pooled s = 52.3245)
sequential standardasync
4865
4915
4753
5185
4760
sequential io_async
4746
4834
4989
4840
4891
No difference proven at 95.0% confidence
parallel standard
2526
2563
2652
2616
2685
parallel io_sync
2335
2184
2244
2341
2401
io_async faster
Difference at 95.0% confidence
-307.4 +/- 111.037
-11.785% +/- 4.25691%
(Student's t, pooled s = 76.1341)
parallel standardasync
1137
1185
1067
1089
1189
parallel io_async
1105
1088
1145
1048
1113
No difference proven at 95.0% confidence
TEST=Existing tests.
Change-Id: I8ba56ab0768df8672bcdb693782d3f1eec86b683
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/185101
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Gary Roumanis <grouma@google.com>
package:vm "clones" an old incremental compiler when creating a new one
for a new isolate. It does that by calling .computeDelta on the first
available compiler. If that compiler is already compiling, though, things
can go bad.
The risk of that happening was made bigger by awaiting in the standard
filesystem instead of using the sync filesystem operations, but it can
happen either way.
If it has any practial importance (outside of tests) is an open quesiton
but this should nevertheless fix the issue by only letting the
incremental compiler allow one compile at the time (creating an implicit
queue when more is tried).
TEST=Existing test suites + added test.
Change-Id: I694e360aa4bce604db41908730c66ef3b96e6d8b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/184788
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
This CL changes `@pragma('vm:ffi:struct-fields', [...])` to
`@pragma('vm:ffi:struct-fields', _FfiStructLayout([...]))` which makes
it easier to add more data in subsequent CLs.
Extends `FindPragma` to allow returning multiple matched pragma's, so
that we can filter them. (In this case to avoid matching user-defined
pragma's that do not have an instance of the private class.)
Separated out from https://dart-review.googlesource.com/c/sdk/+/183640
because of the extra constant in existing expectation files.
Bug: https://github.com/dart-lang/sdk/issues/35763
Bug: https://github.com/dart-lang/sdk/issues/38158
TEST=tests/ffi(_2)/*_by_value_*_test.dart
Change-Id: Idef9f82e9b53c2a32dffabcec19669eae550fe2f
Cq-Include-Trybots: luci.dart.try:front-end-nnbd-mac-release-x64-try,front-end-linux-release-x64-try,vm-precomp-ffi-qemu-linux-release-arm-try,vm-kernel-nnbd-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/184181
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Clement Skau <cskau@google.com>
This is in preparation to migrate package:kernel to null safety.
For the visitor interfaces to support non-nullable return types, the
implementations must avoid using `null` as return value in its base case.
TEST=Refactoring
Change-Id: Ie8fa5d41b99850d9e4abb59634c72920c64128d9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/183691
Reviewed-by: Jens Johansen <jensj@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
This is in preparation to migrate package:kernel to null safety.
For the visitor interfaces to support non-nullable return types, the
implementations must avoid using `null` as return value in its base case.
TEST=Refactoring
Change-Id: Ie5e4153f8d3779d94957bb13b3d2d2a942040ff2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/179760
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
This CL changes the semantics of `Pointer<T extends Struct>.ref` to use
the compile-time `T` rather than the runtime `T`.
This enables tree shaking of subtypes of Struct and optimizing `.ref`.
Bug: https://github.com/dart-lang/sdk/issues/38721
TEST=tests/ffi/vmspecific_static_checks_test.dart
TEST=tests/ffi/*struct*test_.dart
Change-Id: Ie19bc3259d1cb721d0ce56d68e82d09dc3a4ad0e
Cq-Include-Trybots: luci.dart.try:analyzer-linux-release-try,analyzer-nnbd-linux-release-try,app-kernel-linux-debug-x64-try,dart-sdk-linux-try,front-end-nnbd-linux-release-x64-try,pkg-linux-debug-try,vm-kernel-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-x64-try,vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/180190
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Aske Simon Christensen <askesc@google.com>
This rewrites `Pointer<Struct>.ref` and `Pointer<Struct>[]` calls in
the CFE to skip the runtime entry when the type argument is constant.
The runtime entry is still used when the type argument is generic.
Forcing the type argument to be constant and removing the runtime entry
will be done in follow up CLs.
Bug: https://github.com/dart-lang/sdk/issues/38721
Bug: https://github.com/dart-lang/sdk/issues/44621
Removing the runtime entry speeds up `.ref` significantly.
before: FfiStruct.FieldLoadStore(RunTime): 18868.140186915887 us.
after: FfiStruct.FieldLoadStore(RunTime): 270.5877976190476 us.
Measurements from Linux x64 in JIT mode.
Closes: https://github.com/dart-lang/sdk/issues/38648
TEST=tests/ffi/structs_test.dart
Change-Id: I82abd930b5a9c5c78a8999c2bc49802d67d37534
Cq-Include-Trybots: luci.dart.try:analyzer-linux-release-try,analyzer-nnbd-linux-release-try,app-kernel-linux-debug-x64-try,dart-sdk-linux-try,front-end-nnbd-linux-release-x64-try,pkg-linux-debug-try,vm-kernel-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/182265
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Clement Skau <cskau@google.com>
The analyzer and CFE now report a warning on calls to
`Pointer<T extends NativeType>.elementAt()` and
`sizeOf<T extends NativeType>()` where `T` is a generic.
Adapted from https://dart-review.googlesource.com/c/sdk/+/178200 to
only deprecate but not error out on generic calls.
Does not roll forward `package:ffi` to a version with the `Allocator`
but keeps the generic `sizeOf` invocation. This causes extra warnings
in the pkg/front_end testcases.
Issue: https://github.com/dart-lang/sdk/issues/38721
TEST=tests/ffi/data_test.dart
TEST=tests/ffi/sizeof_test.dart
TEST=tests/ffi/structs_test.dart
TEST=tests/ffi/vmspecific_static_checks_test.dart
Change-Id: I8f41c4dc04fc44e7e6c540ba87a3f41604130fe9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/180560
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This CL changes the semantics of `Pointer<T extends Struct>.ref` to use
the compile-time `T` rather than the runtime `T`.
This enables tree shaking of subtypes of Struct and optimizing `.ref`.
Issue: https://github.com/dart-lang/sdk/issues/38721
TEST=tests/ffi/vmspecific_static_checks_test.dart
TEST=tests/ffi/*struct*test_.dart
Change-Id: I3f5b08c08ec0799ef8aab3c4177e2ac70d25501c
Cq-Include-Trybots: luci.dart.try:vm-precomp-ffi-qemu-linux-release-arm-try,analyzer-analysis-server-linux-try,analyzer-linux-release-try,analyzer-nnbd-linux-release-try,front-end-linux-release-x64-try,front-end-nnbd-linux-release-x64-try,benchmark-linux-try,dart-sdk-linux-try,pkg-linux-release-try,vm-ffi-android-release-arm-try,vm-ffi-android-release-arm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/177862
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Aske Simon Christensen <askesc@google.com>
Daco Harkes