Previously, the VM used a linear array to cache previous instantiations
of a type arguments object. Now once the cache hits a certain number
of occupied entries, the VM changes to using a hash-based approach.
The InstantiateTypeArguments stubs have not yet been updated to
traverse the hash-based cache, so once the cache has grown too large,
all attempts at instantiations, even those that are in the cache, go to
the runtime. Thus, until the stubs are updated, this is only an
improvement if the cost of traversing the linear cache dominates the
cost of making a runtime call. Our benchmarks see a ~40% performance
regression for hash-based caches of size 100 but a ~400% performance
improvement for hash-based caches of size 1000. Thus, we currently
split the difference and set the maximum size of linear caches to 500.
TEST=vm/cc/TypeArguments_Cache_ManyInstantiations
Bug: https://github.com/dart-lang/sdk/issues/48344
Change-Id: I7f1376943523bb5bcd8b175cfb1936779ea73d60
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-dwarf-linux-product-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-nnbd-mac-release-arm64-try,vm-kernel-precomp-nnbd-linux-release-simarm_x64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-nnbd-linux-release-simarm64-try,vm-kernel-precomp-nnbd-linux-debug-simriscv64-try,vm-kernel-precomp-tsan-linux-release-x64-try,vm-kernel-tsan-linux-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/265325
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
All `Pointer`s will now have the `Never` type argument at runtime.
TEST=tests/ffi(_2)/*
Bug: https://github.com/dart-lang/sdk/issues/49935
Change-Id: I83c8bba9461c2cab22992ba2e3cf42b7b5f43c36
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-win-debug-x64-try,vm-kernel-precomp-win-debug-x64c-try,vm-kernel-mac-debug-x64-try,dart-sdk-mac-arm64-try,vm-precomp-ffi-qemu-linux-release-arm-try,vm-ffi-android-debug-arm64c-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/272201
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
This CL introduces new embedding APIs for supporting heap sample
profiling. A registered sampling callback is invoked approximately every
N bytes based on an exponential distribution, providing information
about the isolate group the allocation occurred in, the user visible
name of the allocated object type, a weak persistent handle to the
allocated object, and the size of the allocation.
Sampling is triggered using artificial TLAB boundaries to cause
allocations to be sampled to take the allocation slow path where the
registered callback can be invoked with the allocation information.
Only new space allocations are currently traced, with old space
allocation support to be added in a future CL.
TEST=Dart_HeapSampling
Change-Id: I22bcdeec6e823bc1ab44898d4c596fbed7169fa1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/264520
Commit-Queue: Ben Konyi <bkonyi@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
To prepare for the eventual removal of Observatory, we plan on disabling
Observatory by default while providing an escape hatch to manually serve
the tool for some period of time before completely removing Observatory
from the SDK. This change adds flags that can be used to configure
whether or not Observatory is served.
Currently, '--serve-observatory' is the default behavior, but will be
changed to '--no-serve-observatory' once tooling is ready to support the
escape hatch behavior.
Part of https://github.com/dart-lang/sdk/issues/50233
TEST=run_test.dart
Change-Id: Ib6d1e1587d9fbd3c61d4a4c75d90635052835844
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/267720
Reviewed-by: Siva Annamalai <asiva@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
Some large applications are bumping up against the 16-bit limit.
The object header changes from hash:32,cid:16,size:8,gc:8 to hash:32,cid:20,size:4,gc:8.
Because the CID field is no longer a power of two, it takes two instructions to load instead of one, increasing code size. The alternative change to cid:32,hash:24,gc:8 was considered and rejected because the reduction in hash bits very noticeably affects performance in applications with large data sets.
TEST=ci
Bug: b/255934984
Change-Id: I5fa36ba48a6852359d994393cf80c6a761c5d84c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/269120
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Adds a stack map to the catch block entry pc.
The alternative would be to prevent yielding to a safepoint between these two point. For throws coming from Dart or the embedding API, this is possible by unwinding the transitions before the first point and doing only a non-yielding VM->Generated between the two points. For FFI, this does not work because the final state is Native instead of Generated, and VM->Native is a blocking transition.
TEST=ci
Bug: https://github.com/dart-lang/sdk/issues/48668
Change-Id: I8164fddae018ed95d9147fafa65785b7638e5301
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/269746
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Apparently the web implementation of RegExp has drifted from the native implementation in a very subtle way. To prevent further drift, this CL refactors named-captures_test into two tests, one which fails on all web backends currently(named_captures_2_test), and one that passes on all backends(named-captures_test, named unchanged to preserve history).
Change-Id: Ie94185948873fe3e592c10a45c6127dc6d0a4238
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/267282
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Stephen Adams <sra@google.com>
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Joshua Litt <joshualitt@google.com>
_InternalLinkedHashMap => _Map
_InternalImmutableLinkedHashMap => _ConstMap
_InternalLinkedHashSet => _Set
_InternalImmutableLinkedHashSet => _ConstSet
This makes things nicer to read in places that display implementation names, such as stack traces, debuggers, profilers and inspectors.
TEST=ci
Change-Id: Iec851c80ea2086cbe79934565dbf35f04809a836
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/266303
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Chloe Stefantsova <cstefantsova@google.com>
Language spec does not require that evaluations of the same function
literal should create distinct objects. Remove the parts in `Function`
documentation to reflect that.
Also fixes formatting of markdown.
Invalid test file (with Dart 2 and 3 versions) removed: the tests assume
function literals won't be lifted to top-level.
Change-Id: Ib7a9464ad992cf461e77ef2d8ef336c7b0f4875a
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/269721
Reviewed-by: Lasse Nielsen <lrn@google.com>
Reviewed-by: Aske Simon Christensen <askesc@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Erik Ernst <eernst@google.com>
I checked if the retaining paths displayed on Observatory currently made sense, and they do. So, I think that the `getRetainingPath` responses currently being produced are correct, but they are not described well in the spec.
TEST=N/A
Change-Id: I6809ea9ea20574fd18bf5ee3d7d2c190aad62e7d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/268820
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Derek Xu <derekx@google.com>
Before, we only stripped `org-dartlang-sdk:///` as a prefix when it was
followed by `sdk/...`, which happens for code compiled with the Dart
SDK. However, the same prefix can show up in Flutter code (e.g.,
`org-dartlang-sdk:///third_party/dart/sdk/...`). Thus, except for one
case, just strip the prefix at all times when `--resolve-dwarf-paths`
is set, leaving a relative path.
The one case is `org-dartlang-sdk:///flutter/`, which appears in the
resolved version of `dart:ui` (`org-dartlang-sdk:///flutter/lib/ui`),
where the `flutter` isn't part of the actual filesystem path. In this
case, we strip off the `flutter/` as well.
TEST=vm/dart{,_2}/use_resolve_dwarf_paths_flag
Bug: https://github.com/dart-lang/sdk/issues/44325
Change-Id: Ia9abca877e41657089a438d4723ff08a2e16fe69
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-dwarf-linux-product-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-nnbd-mac-release-arm64-try,vm-kernel-precomp-nnbd-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/268762
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Per b/253491057, there are invalid keys (i.e. keys with value 0) being
passed around on occasion. These keys are resulting in flaky test.
This change adds additional error logging to help determine when the
zero-valued keys are originating from.
This change also removes the zx_assert to avoid crashing when a
zero-value key is encountered, and instead an error is logged.
TEST=built for fuchsia and tested on nest hub device.
Bug: b/253491057
Change-Id: Ic94a3f526318ef2873d3af7a3ceab064b67859d5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/268603
Reviewed-by: Zach Anderson <zra@google.com>
Commit-Queue: Brian Quinlan <bquinlan@google.com>
Reviewed-by: Brian Quinlan <bquinlan@google.com>
Tess Strickland