This CL removes static fields for storing the `@Native`'s function
addresses. Instead, the function addresses are stored in the object
pool for all archs except for ia32. ia32 has no AOT and no AppJit
snapshots, so the addresses are directly embedded in the code.
This CL removes the closure wrapping for `@Native`s. Instead of
`pointer.asFunctionInternal()()` where `asFunction` returns a closure
which contains the trampoline, the function is compiled to a body
which contains the trampoline `Native()`. This is possible for
`@Native`s because the dylib and symbol names are known statically.
Doing the compilation in kernel_to_il instead of a CFE transform
enables supporting static linking later. (The alternative would have
been to transform in the cfe to a `@pragma('vm:cachable-idempotent')`
instead of constructing the IL in kernel_to_il.
To enable running resolution in ia32 in kernel_to_il.cc, the
resolution function has been made available via
`runtime/lib/ffi_dynamic_library.h`.
Because the new calls are simply static calls, the TFA can figure
out const arguments flowing to these calls. This leads to constant
locations in the parameters to FfiCalls. So, this CL also introduces
logic to move constants into `NativeLocation`s.
TEST=runtime/vm/compiler/backend/il_test.cc
TEST=tests/ffi/function_*_native_(leaf_)test.dart
TEST=pkg/vm/testcases/transformations/ffi/ffinative_compound_return.dart
Closes: https://github.com/dart-lang/sdk/issues/47625
Closes: https://github.com/dart-lang/sdk/issues/51618
Change-Id: Ic5d017005dedcedea40c455c4d24dbe774f91603
CoreLibraryReviewExempt: Internal FFI implementation changes
Cq-Include-Trybots: luci.dart.try:vm-aot-android-release-arm64c-try,vm-aot-android-release-arm_x64-try,vm-aot-linux-debug-x64-try,vm-aot-linux-debug-x64c-try,vm-aot-mac-release-arm64-try,vm-aot-mac-release-x64-try,vm-aot-obfuscate-linux-release-x64-try,vm-aot-win-debug-arm64-try,vm-aot-win-debug-x64c-try,vm-aot-win-release-x64-try,vm-appjit-linux-debug-x64-try,vm-asan-linux-release-x64-try,vm-checked-mac-release-arm64-try,vm-eager-optimization-linux-release-ia32-try,vm-eager-optimization-linux-release-x64-try,vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64c-try,vm-ffi-qemu-linux-release-arm-try,vm-ffi-qemu-linux-release-riscv64-try,vm-fuchsia-release-x64-try,vm-kernel-linux-debug-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-linux-debug-ia32-try,vm-linux-debug-x64-try,vm-linux-debug-x64c-try,vm-mac-debug-arm64-try,vm-mac-debug-x64-try,vm-msan-linux-release-x64-try,vm-reload-linux-debug-x64-try,vm-reload-rollback-linux-debug-x64-try,vm-ubsan-linux-release-x64-try,vm-win-debug-arm64-try,vm-win-debug-x64-try,vm-win-debug-x64c-try,vm-win-release-ia32-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/284300
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This fixes `identical` checks when the input doesn't need case mapping.
This change was originally made in [1], but I'm trying to split it into
smaller CLs as it currently has a lot of conflicts with the main branch.
[1]: https://dart-review.googlesource.com/c/sdk/+/316628
Change-Id: I88da52a3a73c9d587acefe2b14fd39edaf01c966
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/332200
Reviewed-by: Aske Simon Christensen <askesc@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Refactor libraries so that JSCM will only use `JSStringImpl` class for
strings.
The goal is to disentangle the native string classes and `JSStringImpl`
and start testing `JSStringImpl` in isolation.
Changes:
- `dart:_string` is no longer available in JSCM.
- Make `int.toString` external to allow patching it differently in JSCM
and normal modes.
`toString` implementations are in `boxed_int_to_string.dart` patch
files.
- `int.parse` now uses JS `parseInt`. However `parseInt` is not
compatible with Dart's `int.parse` so this will cause some more test
failures in JSCM for now.
- Any dependencies to `dart:_string` from JSCM `dart:convert` are
removed. The library implementation now uses JS `TextDecoder` for
UTF-8 decoding.
Note: `TextDecoder` is not available on d8, so text decoding tests
will fail on d8.
JSON encoding and decoding in `dart:convert` will be updated in a
follow-up CL.
- Compiler (translator, constant generator, code generator etc.) is
updated to allocate the `JSStringImpl`s in JSCM.
Initially this will make some JSCM test fail as `int` parsing is not
quite right, those will be fixed in follow-up CLs.
Co-authored-by: Joshua Litt <joshualitt@google.com>
Change-Id: I366e06f44cdc369d28fe47b24015234260304399
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/332680
Reviewed-by: Aske Simon Christensen <askesc@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
This reverts commit 29979649bd.
Reason for revert: makes `promiseToFuture` not exported from `dart:html`, see b/295129286
Original change's description:
> [analyzer] simplify the libraries.dart file
>
> Cleanup and re-landing of https://dart-review.googlesource.com/c/sdk/+/317803.
>
> Change-Id: I9fd679fdb111895ce662ecbe6d0fc5844909c3bc
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/332441
> Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
> Commit-Queue: Devon Carew <devoncarew@google.com>
Change-Id: Id93b4a1d4bdea706ae10cfb089319514a0238a7d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/332683
Reviewed-by: Alexander Thomas <athom@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Ilya Yanok <yanok@google.com>
This uses the allowed_experiments.json to opt package:web and
dart:js_interop in to the inline-class experiment (the extension
types feature) for early access and development.
Change-Id: Ieb828b654f2320b147a3c48bbb4a142da1c18380
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/332421
Reviewed-by: Srujan Gaddam <srujzs@google.com>
Reviewed-by: William Hesse <whesse@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Flow analysis has improved and issue #40041 is resolved. These
`FutureOr<T>` conditionals are promoting to `Future<T>` or `T` and no
longer need an explicit cast.
R=lrn@google.com
Change-Id: I925db6d7f52aad54c88434ff02e1d0ae4e72278b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/331206
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Stephen Adams <sra@google.com>
Reviewed-by: Sigmund Cherem <sigmund@google.com>
Auto-Submit: Nate Bosch <nbosch@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
Change-Id: I70c593ca2b7f9ba6b7cc94886f622f4af5dd20fd
CoreLibraryReviewExempt: documentation update
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330620
Reviewed-by: Brian Quinlan <bquinlan@google.com>
Commit-Queue: Alexander Thomas <athom@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Intercept all `Utf8Decoder.convert` calls to check the input type and
call the right UTF8 decoder function.
This currently only optimizes `U8List` parsing, and only the fastest
path where the input is an ASCII string. Follow-up CLs will optimize JS
typed array parsing by calling browser's UTF8 decoder and the slow paths
of non-ASCII and malformed buffers.
Results from an internal benchmark extracted from a real use case:
Before:
Parse(RunTime): 330,500 us.
After:
Parse(RunTime): 90,310 us. (-72%)
dart2js -O4 results of the same benchmark:
Parse(RunTime): 234,250 us.
Change-Id: I470697a9eb516e1e031e9b865ddb29e8f25569bf
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/331187
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Jackson Gardner <jacksongardner@google.com>
This reverts commit b657773d58.
Reason for revert: blocking Dart SDK -> Engine roll (https://github.com/flutter/flutter/issues/137054)
Original change's description:
> [typed_data] Deprecate UnmodifiableUint8ListView and friends
>
> This is the first of several steps to remove the unmodifiable views for typed data classes. The end goal is that dart2js has only one class implementing `Uint8List` so that `Uint8List` accesses can always be compiled down to JavaScript code that directly uses indexed property accesses (`a[i]`).
>
> This first step deprecates the unmodifiable view classes to help prevent their use in new code, and adds `asUnmodifiableView()` methods as a replacement for the small number of places that use the classes.
>
> The next steps (see #53785) are to remove uses of the unmodifiable view classes from the SDK. Once this is complete the classes themselves can be removed.
>
> TEST=ci
>
> Issue: #53218
> Issue: #53785
>
> Change-Id: I04d4feb0d9f1619e6eee65236e559f5e6adf2661
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/321922
> Reviewed-by: Nicholas Shahan <nshahan@google.com>
> Reviewed-by: Lasse Nielsen <lrn@google.com>
> Commit-Queue: Stephen Adams <sra@google.com>
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>
> Reviewed-by: Ömer Ağacan <omersa@google.com>
Issue: #53218
Issue: #53785
Change-Id: I0bb042269f9ff8e5cd69619cf97b60c79ea98cbf
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/331680
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Derek Xu <derekx@google.com>
This is the first of several steps to remove the unmodifiable views for typed data classes. The end goal is that dart2js has only one class implementing `Uint8List` so that `Uint8List` accesses can always be compiled down to JavaScript code that directly uses indexed property accesses (`a[i]`).
This first step deprecates the unmodifiable view classes to help prevent their use in new code, and adds `asUnmodifiableView()` methods as a replacement for the small number of places that use the classes.
The next steps (see #53785) are to remove uses of the unmodifiable view classes from the SDK. Once this is complete the classes themselves can be removed.
TEST=ci
Issue: #53218
Issue: #53785
Change-Id: I04d4feb0d9f1619e6eee65236e559f5e6adf2661
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/321922
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Stephen Adams <sra@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
CoreLibraryReviewExempt: Comment only.
Change-Id: Ibaec3812e430ba0c8644315fba062bca62f3c173
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/331241
Auto-Submit: Lasse Nielsen <lrn@google.com>
Commit-Queue: Lasse Nielsen <lrn@google.com>
Reviewed-by: Alexander Thomas <athom@google.com>
- In development mode (DDC) the extra null safety errors will be thrown.
- Remove extra code paths that called unsound helpers.
- Fix expectations in weak_null_safety_errors_test.dart.
Change-Id: I107c602b0ae38b13038e501564cba9b8cfc58e70
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/329568
Reviewed-by: Mayank Patke <fishythefish@google.com>
Commit-Queue: Nicholas Shahan <nshahan@google.com>
The example was originally written without a `main`, and with an example
using `Duration.operator *`. An update added a call to the example
method from a new `main`, removed the usage of `Duration.operator *`,
but did not remove the prose reference.
- Remove the sentence mentioning that the duration can be const or
computed.
- Remove the intermediate `scheduleTimeout` method, move the `Timer`
construction to `main`. This removes any consideration of `Duration`
vs `int milliseconds` which had been added in an attempt make the docs
more focused on `Timer` than `Duration`.
R=lrn@google.com
CoreLibraryReviewExempt:Documentation change.
Change-Id: I26bb4e12910eb50cf5975b8a6573701a5a862070
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330980
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Nate Bosch <nbosch@google.com>
Also adds `JS_FALSE` to replace uses of `JS_GET_FLAG(false)`. See the
doc comment for details.
Change-Id: I33f89f158c1955a19fa299f22c50d51e36ede3d8
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330171
Commit-Queue: Mayank Patke <fishythefish@google.com>
Reviewed-by: Stephen Adams <sra@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Nate Biggs <natebiggs@google.com>
Now that we have our own `dart:convert` implementations in dart2wasm
targets (1) and the internal types are now internally public (2), this
simplifies some of the string implementation code by using the
implementation classes directly.
Note: stringref patch can be ignored, it's currently unused and I'm only
updating it to make it compile. It will be rewritten based on
`JSStringImpl` in [3].
To keep the CLs as small as possible, this CL does not do any of the
optimizations we can do now, such as optimizing
`OneByteString._setRange` to Wasm `array.copy` instructions. These will
be done in follow-up CLs.
[1]: https://dart-review.googlesource.com/c/sdk/+/330781
[2]: https://dart-review.googlesource.com/c/sdk/+/330783
[3]: https://dart-review.googlesource.com/c/sdk/+/316628
Change-Id: I92918d58a565ad32f5a221430e02d81562f03b00
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/331040
Reviewed-by: Jackson Gardner <jacksongardner@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
This is a part of [1] and a continuation of [2].
- Make typed data implementation file a library, to allow importing it
in `dart:convert` patch.
- Make SIMD implementation file a library, to allow importing it from
the typed data library.
- Make typed data implementation classes internally public (available in
other `dart:...` libraries).
- Relax `mayDefineRestrictedType` to allow implementing restricted types
anywhere in the standard library implementation.
[1]: https://dart-review.googlesource.com/c/sdk/+/330121
[2]: https://dart-review.googlesource.com/c/sdk/+/330781
Change-Id: I70bc6869a20fd2479bee081cfceef27d6de19974
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330783
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Jackson Gardner <jacksongardner@google.com>
This copies VM's convert implementation in dart2wasm's default and JSCM
targets, to be able to optimize the implementation based on these
targets' typed data and string implementation classes.
An example optimization that specializes UTF-8 decoder to `_U8List` can
be seen in [1], which improves a real-world benchmark (extracted from
internal) from 220ms to 90ms. (-59%)
Another optimization is we will be using the browser's UTF8 and JSON
decoders in JSCM convert patch.
VM's convert patch is moved from vm_shared to vm, as it's no longer used
outside of VM.
[1]: https://dart-review.googlesource.com/c/sdk/+/330121
Tested: existing tests.
Change-Id: I981070615a106e4f356ed8b292a29ec950bd4d97
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330781
Reviewed-by: Jackson Gardner <jacksongardner@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Bug:https://github.com/dart-lang/sdk/issues/53161
Change-Id: I3f13af3cb852b3656341922b9656ec91fc413eed
Tested: documentation + unit test only
CoreLibraryReviewExempt: Only adds documentation and adds a unit test to verify existing behavior
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/323426
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Brian Quinlan <bquinlan@google.com>
Reviewed-by: Sigmund Cherem <sigmund@google.com>
Cleanup for https://github.com/dart-lang/dartdoc/issues/3531
This makes these comment references more idiomatic, I think
more readable, and supports dropping the leading `new ` syntax
in doc comments.
Change-Id: Id832ad14d9ea08fe03fe3125065755f49b1b93ed
CoreLibraryReviewExempt: doc comment only change
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330683
Reviewed-by: Devon Carew <devoncarew@google.com>
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Samuel Rawlins <srawlins@google.com>
This is a reland of commit 06d7a2352e
This version fixes an issue when a phi node has multiple inputs with
different unboxed integer representations. The original CL made a
change where only the representations were considered, not the range
of values for the Phi calculated by range analysis. The reland goes
back to the old behavior for this case.
Also fixes the new tests on 32-bit architectures.
Original change's description:
> [vm/compiler] Change MemoryCopy to also take untagged addresses.
>
> This CL adds the ability to pass the payload address of the source
> and destination directly to the MemoryCopy instruction as an untagged
> value.
>
> The new translation of the _TypedListBase._memMoveN methods use the new
> MemoryCopy constructor, retrieving the untagged value of the data field
> of both the source and destination. This way, if inlining exposes the
> allocation of the object from which the data field is being retrieved,
> then allocation sinking can remove the intermediate allocation if there
> are no escaping uses of the object.
>
> Since Pointer.asTypedList allocates such ExternalTypedData objects,
> this CL makes that method inlined if at all possible, which removes
> the intermediate allocation if the only use of the TypedData object
> is to call setRange for memory copying purposes.
>
> This CL also separates unboxed native slots into two groups: those
> that contain untagged addresses and those that do not. The former
> group now have the kUntagged representation, which mimics the old
> use of LoadUntagged for the PointerBase data field and also ensures
> that any arithmetic operations on untagged addresses must first be
> explicitly converted to an unboxed integer and then explicitly converted
> back to untagged before being stored in a slot that contains untagged
> addresses.
>
> When a unboxed native slot that contains untagged addresses is defined,
> the definition also includes a boolean which represents whether
> addresses that may be moved by the GC can be stored in this slot or not.
> The redundancy eliminator uses this to decide whether it is safe to
> eliminate a duplicate load, replace a load with the value originally
> stored in the slot, or lift a load out of a loop.
>
> In particular, the PointerBase data field may contain GC-moveable
> addresses, but only for internal TypedData objects and views, not
> for external TypedData objects or Pointers. To allow load optimizations
> involving the latter, the LoadField and StoreField instructions now
> take boolean flags for whether loads or stores from the slot are
> guaranteed to not be GC-moveable, to override the information from
> the slot argument.
>
> Notable benchmark changes on x64 (similar for other archs unless noted):
>
> JIT:
> * FfiMemory.PointerPointer: 250.7%
> * FfiStructCopy.Copy1Bytes: -26.73% (only x64)
> * FfiStructCopy.Copy32Bytes: -25.18% (only x64)
> * MemoryCopy.64.setRange.Pointer.Uint8: 19.36%
> * MemoryCopy.64.setRange.Pointer.Double: 18.96%
> * MemoryCopy.8.setRange.Pointer.Double: 17.59%
> * MemoryCopy.8.setRange.Pointer.Uint8: 19.46%
>
> AOT:
> * FfiMemory.PointerPointer: 323.5%
> * FfiStruct.FieldLoadStore: 483.3%
> * FileIO_readwrite_64kb: 15.39%
> * FileIO_readwrite_512kb (Intel Xeon): 46.22%
> * MemoryCopy.512.setRange.Pointer.Uint8: 35.20%
> * MemoryCopy.64.setRange.Pointer.Uint8: 55.40%
> * MemoryCopy.512.setRange.Pointer.Double: 29.45%
> * MemoryCopy.64.setRange.Pointer.Double: 60.37%
> * MemoryCopy.8.setRange.Pointer.Double: 59.54%
> * MemoryCopy.8.setRange.Pointer.Uint8: 55.40%
> * FfiStructCopy.Copy32Bytes: 398.3%
> * FfiStructCopy.Copy1Bytes: 1233%
>
> TEST=vm/dart/address_local_pointer, vm/dart/pointer_as_typed_list
>
> Issue: https://github.com/dart-lang/sdk/issues/42072
> Fixes: https://github.com/dart-lang/sdk/issues/53124
>
> Cq-Include-Trybots: luci.dart.try:vm-ffi-qemu-linux-release-arm-try,vm-eager-optimization-linux-release-x64-try,vm-linux-release-x64-try,vm-linux-debug-x64-try,vm-aot-linux-release-x64-try,vm-aot-linux-debug-x64-try
> Change-Id: I563e0bfac5b1ac6cf1111649934067c12891b631
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/324820
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Commit-Queue: Tess Strickland <sstrickl@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>
TEST=vm/dart/address_local_pointer, vm/dart/pointer_as_typed_list
Issue: https://github.com/dart-lang/sdk/issues/42072
Fixes: https://github.com/dart-lang/sdk/issues/53124
Change-Id: Iabb0e910f12636d0ff51e711c8c9c98ad40e5811
Cq-Include-Trybots: luci.dart.try:vm-ffi-qemu-linux-release-arm-try,vm-eager-optimization-linux-release-x64-try,vm-linux-release-x64-try,vm-linux-debug-x64-try,vm-aot-linux-release-x64-try,vm-aot-linux-debug-x64-try,vm-aot-linux-release-simarm_x64-try,vm-aot-linux-debug-simarm_x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330600
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
https://dart-review.googlesource.com/c/sdk/+/328280
1) Combine _setKeepIsolateAlive/_getKeepIsolateAlive into
abstract bool _keepIsolateAlive.
2) Refactor close to delegate to _close, rather than calling
super.close(). Makes _isClosed validation clearer.
Change-Id: I61a97f5ba35c3ff00ab4270b9f7a763968e703fc
TEST=CI (no behaviour changes)
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330580
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Liam Appelbe <liama@google.com>
This reverts commit 06d7a2352e.
Reason for revert: everything crashes on vm-aot-linux-debug-simarm_x64
Original change's description:
> [vm/compiler] Change MemoryCopy to also take untagged addresses.
>
> This CL adds the ability to pass the payload address of the source
> and destination directly to the MemoryCopy instruction as an untagged
> value.
>
> The new translation of the _TypedListBase._memMoveN methods use the new
> MemoryCopy constructor, retrieving the untagged value of the data field
> of both the source and destination. This way, if inlining exposes the
> allocation of the object from which the data field is being retrieved,
> then allocation sinking can remove the intermediate allocation if there
> are no escaping uses of the object.
>
> Since Pointer.asTypedList allocates such ExternalTypedData objects,
> this CL makes that method inlined if at all possible, which removes
> the intermediate allocation if the only use of the TypedData object
> is to call setRange for memory copying purposes.
>
> This CL also separates unboxed native slots into two groups: those
> that contain untagged addresses and those that do not. The former
> group now have the kUntagged representation, which mimics the old
> use of LoadUntagged for the PointerBase data field and also ensures
> that any arithmetic operations on untagged addresses must first be
> explicitly converted to an unboxed integer and then explicitly converted
> back to untagged before being stored in a slot that contains untagged
> addresses.
>
> When a unboxed native slot that contains untagged addresses is defined,
> the definition also includes a boolean which represents whether
> addresses that may be moved by the GC can be stored in this slot or not.
> The redundancy eliminator uses this to decide whether it is safe to
> eliminate a duplicate load, replace a load with the value originally
> stored in the slot, or lift a load out of a loop.
>
> In particular, the PointerBase data field may contain GC-moveable
> addresses, but only for internal TypedData objects and views, not
> for external TypedData objects or Pointers. To allow load optimizations
> involving the latter, the LoadField and StoreField instructions now
> take boolean flags for whether loads or stores from the slot are
> guaranteed to not be GC-moveable, to override the information from
> the slot argument.
>
> Notable benchmark changes on x64 (similar for other archs unless noted):
>
> JIT:
> * FfiMemory.PointerPointer: 250.7%
> * FfiStructCopy.Copy1Bytes: -26.73% (only x64)
> * FfiStructCopy.Copy32Bytes: -25.18% (only x64)
> * MemoryCopy.64.setRange.Pointer.Uint8: 19.36%
> * MemoryCopy.64.setRange.Pointer.Double: 18.96%
> * MemoryCopy.8.setRange.Pointer.Double: 17.59%
> * MemoryCopy.8.setRange.Pointer.Uint8: 19.46%
>
> AOT:
> * FfiMemory.PointerPointer: 323.5%
> * FfiStruct.FieldLoadStore: 483.3%
> * FileIO_readwrite_64kb: 15.39%
> * FileIO_readwrite_512kb (Intel Xeon): 46.22%
> * MemoryCopy.512.setRange.Pointer.Uint8: 35.20%
> * MemoryCopy.64.setRange.Pointer.Uint8: 55.40%
> * MemoryCopy.512.setRange.Pointer.Double: 29.45%
> * MemoryCopy.64.setRange.Pointer.Double: 60.37%
> * MemoryCopy.8.setRange.Pointer.Double: 59.54%
> * MemoryCopy.8.setRange.Pointer.Uint8: 55.40%
> * FfiStructCopy.Copy32Bytes: 398.3%
> * FfiStructCopy.Copy1Bytes: 1233%
>
> TEST=vm/dart/address_local_pointer, vm/dart/pointer_as_typed_list
>
> Issue: https://github.com/dart-lang/sdk/issues/42072
> Fixes: https://github.com/dart-lang/sdk/issues/53124
>
> Cq-Include-Trybots: luci.dart.try:vm-ffi-qemu-linux-release-arm-try,vm-eager-optimization-linux-release-x64-try,vm-linux-release-x64-try,vm-linux-debug-x64-try,vm-aot-linux-release-x64-try,vm-aot-linux-debug-x64-try
> Change-Id: I563e0bfac5b1ac6cf1111649934067c12891b631
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/324820
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Commit-Queue: Tess Strickland <sstrickl@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>
Issue: https://github.com/dart-lang/sdk/issues/42072
Change-Id: I7c31434e01108487de69a32154bbefd1538c6f0f
Cq-Include-Trybots: luci.dart.try:vm-ffi-qemu-linux-release-arm-try,vm-eager-optimization-linux-release-x64-try,vm-linux-release-x64-try,vm-linux-debug-x64-try,vm-aot-linux-release-x64-try,vm-aot-linux-debug-x64-try
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/330523
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Alexander Aprelev <aam@google.com>
Auto-Submit: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
This CL adds the ability to pass the payload address of the source
and destination directly to the MemoryCopy instruction as an untagged
value.
The new translation of the _TypedListBase._memMoveN methods use the new
MemoryCopy constructor, retrieving the untagged value of the data field
of both the source and destination. This way, if inlining exposes the
allocation of the object from which the data field is being retrieved,
then allocation sinking can remove the intermediate allocation if there
are no escaping uses of the object.
Since Pointer.asTypedList allocates such ExternalTypedData objects,
this CL makes that method inlined if at all possible, which removes
the intermediate allocation if the only use of the TypedData object
is to call setRange for memory copying purposes.
This CL also separates unboxed native slots into two groups: those
that contain untagged addresses and those that do not. The former
group now have the kUntagged representation, which mimics the old
use of LoadUntagged for the PointerBase data field and also ensures
that any arithmetic operations on untagged addresses must first be
explicitly converted to an unboxed integer and then explicitly converted
back to untagged before being stored in a slot that contains untagged
addresses.
When a unboxed native slot that contains untagged addresses is defined,
the definition also includes a boolean which represents whether
addresses that may be moved by the GC can be stored in this slot or not.
The redundancy eliminator uses this to decide whether it is safe to
eliminate a duplicate load, replace a load with the value originally
stored in the slot, or lift a load out of a loop.
In particular, the PointerBase data field may contain GC-moveable
addresses, but only for internal TypedData objects and views, not
for external TypedData objects or Pointers. To allow load optimizations
involving the latter, the LoadField and StoreField instructions now
take boolean flags for whether loads or stores from the slot are
guaranteed to not be GC-moveable, to override the information from
the slot argument.
Notable benchmark changes on x64 (similar for other archs unless noted):
JIT:
* FfiMemory.PointerPointer: 250.7%
* FfiStructCopy.Copy1Bytes: -26.73% (only x64)
* FfiStructCopy.Copy32Bytes: -25.18% (only x64)
* MemoryCopy.64.setRange.Pointer.Uint8: 19.36%
* MemoryCopy.64.setRange.Pointer.Double: 18.96%
* MemoryCopy.8.setRange.Pointer.Double: 17.59%
* MemoryCopy.8.setRange.Pointer.Uint8: 19.46%
AOT:
* FfiMemory.PointerPointer: 323.5%
* FfiStruct.FieldLoadStore: 483.3%
* FileIO_readwrite_64kb: 15.39%
* FileIO_readwrite_512kb (Intel Xeon): 46.22%
* MemoryCopy.512.setRange.Pointer.Uint8: 35.20%
* MemoryCopy.64.setRange.Pointer.Uint8: 55.40%
* MemoryCopy.512.setRange.Pointer.Double: 29.45%
* MemoryCopy.64.setRange.Pointer.Double: 60.37%
* MemoryCopy.8.setRange.Pointer.Double: 59.54%
* MemoryCopy.8.setRange.Pointer.Uint8: 55.40%
* FfiStructCopy.Copy32Bytes: 398.3%
* FfiStructCopy.Copy1Bytes: 1233%
TEST=vm/dart/address_local_pointer, vm/dart/pointer_as_typed_list
Issue: https://github.com/dart-lang/sdk/issues/42072
Fixes: https://github.com/dart-lang/sdk/issues/53124
Cq-Include-Trybots: luci.dart.try:vm-ffi-qemu-linux-release-arm-try,vm-eager-optimization-linux-release-x64-try,vm-linux-release-x64-try,vm-linux-debug-x64-try,vm-aot-linux-release-x64-try,vm-aot-linux-debug-x64-try
Change-Id: I563e0bfac5b1ac6cf1111649934067c12891b631
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/324820
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Use URLSearchParams to escape query parameters.
- For large query parameters (>100KB), where escaping the parameters causes jank, this can be 2x-5x faster, reducing ~100ms pauses to nearer frame rate.
- For small query parameters (<100B) it can be slightly (10-20%) slower, but still well below 1 millisecond.
TEST=ci
Issue: #53712
Change-Id: I045bac7a067a658a58aaac4266409d526ccda774
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/329822
Commit-Queue: Stephen Adams <sra@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Lasse Nielsen <lrn@google.com>
The type normalization rules specify that `FutureOr<T>?` is normalized
to `FutureOr<T>` when `T` is nullable. However, it's more practical
for subtype testing if the declared nullability on the runtime
representation of the `FutureOr` type reflects the true nullability
(nullable if the `FutureOr` is declared nullable or its type argument
is nullable), rather than being normalized as per the spec.
This changes the static and dynamic normalization rules in dart2wasm
thus and compensates by computing the proper spec normalization when
the type is converted to a string.
The added test exposed a number of bugs in DDC and the VM:
https://github.com/dart-lang/sdk/issues/53175https://github.com/dart-lang/sdk/issues/53737https://github.com/dart-lang/sdk/issues/53738
Change-Id: I0ad0a09fe935ccbd3eb65e6958c958d29e0bb088
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/320821
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Aske Simon Christensen <askesc@google.com>
This CL passes the offset and uri of the file (here called a script uri
as opposed to a library uri, the two will be different if we're in a
part) when doing expression compilation.
This CL only passes the data, but doesn't actually use it.
Future CL(s) will use this data to calculate the static type of
available variables which is needed for an upcomming feature.
TEST=Existing tests.
CoreLibraryReviewExempt: Not changing SDK APIs.
Change-Id: I67ead461ab4bb9341424e693946f3e4afe35ce92
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/329322
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
