HashTable<..., GrowableArrayStorageTraits> passes the zone to placement-new of GrowableArrayStorageTraits::Array. Because this class did not extend ZoneAllocated, the default placement-new interpreted the zone as the allocated address and clobbered the beginning of the zone.
TEST=build
Change-Id: I95805e00b3012f5282f11cfb6da1567f00d56c46
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208982
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
- Delete the port set during cleanup while holding the lock
- Check for port set being valid in every function that accesses
the port set to avoid racy access while the VM is shutting
down (e.g: Dart_InvokeVMServiceMethod)
TEST=ci
Change-Id: Ib5c3a6a36260820c040304d03fdc49558dae611b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208983
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Siva Annamalai <asiva@google.com>
This brings the native port deserializer into agreement with the native port serializer and the interpretation of TypedData::length_.
TEST=ci
Change-Id: I52eb28af7292fc15b0ad4779dc4d4e723f6e128b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208283
Reviewed-by: Siva Annamalai <asiva@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
This reverts commit 1796160606.
Reason for revert: Still seeing failures.
Original change's description:
> Reland "[ package:dds ] Add support for caching CPU samples based on UserTag"
>
> This reverts commit 4160747ef6.
>
> It seems like 9397b8ff05 should have fixed the original failures we were seeing and the failing bots in question seem to be passing when running try jobs.
>
> TBR=asiva@ (discussed offline)
>
> TEST=Existing service tests + get_cached_cpu_samples_test
> Change-Id: I1c50c0e79375df819a0bd68e68ac28c3064874c8
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208441
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Ben Konyi <bkonyi@google.com>
TBR=bkonyi@google.com,asiva@google.com
Change-Id: Ic522ab61b430673533c94ecd0c91bee569a0add6
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208960
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
Combines one and two byte string serialization into a unified
string serialization cluster that's built on CanonicalSet
serialization. This approach is already used when serializing
to RO data, but we can't use RO data in compressed mode.
Change-Id: I28b2473efbc855c0d30fa9ec39fa9c6d42dc2b67
TEST=CI
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208022
Commit-Queue: Liam Appelbe <liama@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
This reverts commit 4160747ef6.
It seems like 9397b8ff05 should have fixed the original failures we were seeing and the failing bots in question seem to be passing when running try jobs.
TBR=asiva@ (discussed offline)
TEST=Existing service tests + get_cached_cpu_samples_test
Change-Id: I1c50c0e79375df819a0bd68e68ac28c3064874c8
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208441
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
All handles passed to IL instructions should be ZoneHandles.
Recently added code for generating constructor tear-offs has been
passing Handle instead of ZoneHandle by mistake.
This change fixes the bug and adds assertions to AllocateObjectInstr so
it would check handle type earlier and on all platforms.
TEST=co19/LanguageFeatures/Constructor-tear-offs in debug/ia32 mode
Issue https://github.com/dart-lang/sdk/issues/46231
Change-Id: I99e6889e59bf40f4788a7f3c6d37dad89aa91ab0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208686
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Both SampleBlock::ReserveAllocationSample and
SampleBlock::ReserveCPUSample can return nullptr, so that needs to be
appropriately handled in SampleBlock::ReserveSampleAndLink. The only
caller of the latter checks for nullptr, so pass though nullptr results
from the former methods.
TEST=Fixes current service tests failures on some trybots,
see test results on CL.
Cq-Include-Trybots: luci.dart.try:vm-kernel-linux-debug-ia32-try,app-kernel-linux-debug-x64-try
Change-Id: Ib70cc6726862bc5fe3bfb0c06ee2aae84ec026c3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208321
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Clement Skau <cskau@google.com>
Adds GetNativeField(o,i) to access NativeFieldWrapperClass{1-4}
native fields directly in Dart instead of through VM API.
This will allow us to pass `this` as a raw pointer instead of as
a Dart handle which needs to be unwrapped.
This will in turn allow ffi native calls for instance methods.
TEST=Adds vm/cc/DartAPI_NativeFieldAccess
Bug: https://github.com/dart-lang/sdk/issues/43889
Change-Id: I8006c735569fcbe5e49915a4c7e72c4ca85b356e
Cq-Do-Not-Cancel-Tryjobs: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/205482
Commit-Queue: Clement Skau <cskau@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
DDS can be configured to listen for CPU sample events and cache samples
that were collected while certain UserTags are active. These cached
samples are stored in a ring buffer and are stored until the isolate
shuts down.
TEST=pkg/dds/test/get_cached_cpu_samples_test.dart
Change-Id: Ib20770f59f1672c703413486f87795b3bb23f676
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/207206
Commit-Queue: Ben Konyi <bkonyi@google.com>
Reviewed-by: Kenzie Schmoll <kenzieschmoll@google.com>
Also make NBitMask generic, replace uses of SignedNBitMask
with generic NBitMask, and remove SignedNBitMask.
TEST=No changes to functionality for existing callers, so existing tests.
Cq-Include-Trybots: luci.dart.try:vm-kernel-linux-release-simarm-try,vm-kernel-linux-release-simarm64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-linux-release-simarm64-try
Change-Id: Ic33785e50be98123c75782c6e62ef32ec45f4191
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208181
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Split off https://dart-review.googlesource.com/c/sdk/+/203765 to make
that diff smaller.
Serializing the index incurs a large size penalty to snapshots. Rather
than serializing, we reconstruct the index based on the hashes that we
do serialize for all canonical objects. (Reconstructing the index is a
relatively cheap operation, recomputing the hashes is not.)
Note that to_snapshot is currently dead code, because we never serialize
non-const maps and sets in clustered snapshots.
Of course we do exercise the layout of the maps and sets itself.
Bug: https://github.com/dart-lang/sdk/issues/45908
TEST=The whole test suite including all map and set tests.
Change-Id: If54117497029e4da212c702a76c82de6d1531092
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-debug-simarm_x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208086
Reviewed-by: Tess Strickland <sstrickl@google.com>
Commit-Queue: Daco Harkes <dacoharkes@google.com>
For members of program structure that we cache in the ObjectStore we try
to provide only read-only access in the public API. The lazy population
is encapculated in the ObjectStore class.
Use the same mechanism we use for core/isolate/... library members also
for the `dart:_internal` members.
Furthermore cache the Symbol._name field in the ObjectStore to avoid
expensive lookup as well as avoid an extra TLS lookup.
TEST=Existing test suite.
Change-Id: I299faceb39f20576e23244c71c58d698f80f80c7
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208180
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
We want to reuse CanonicalizeHash for const maps and sets. See design in
go/dart-vm-const-maps.
Before, we only used CanonicalizeHash on a per-class basis. However,
when we start using it for const maps and sets we don't want hash
collisions between const instances with identical layout but a different
class id. So we include the class id into the hash rather than the size.
Bug: https://github.com/dart-lang/sdk/issues/45908
TEST=vm/cc/CanonicalizeHash_Const_Instances
TEST=A ton of bots for existing tests.
Change-Id: Idc0eed49fbae8f76dcfdda21dd9fe2674dc4a240
Cq-Include-Trybots: luci.dart.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-canary-linux-debug-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-checked-linux-release-x64-try,vm-kernel-linux-debug-x64c-try,vm-kernel-linux-debug-x64-try,vm-kernel-linux-debug-simarm64c-try,vm-kernel-nnbd-linux-release-simarm-try,vm-kernel-optcounter-threshold-linux-release-x64-try,vm-kernel-precomp-android-release-arm_x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64c-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-precomp-ffi-qemu-linux-release-arm-try,vm-kernel-precomp-linux-debug-simarm_x64-try,vm-kernel-precomp-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208085
Reviewed-by: Tess Strickland <sstrickl@google.com>
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Allows for function objects to be displayed in an understandable fashion
in developer tooling (e.g., a DevTools program structure inspector).
Also adds 'implicit' property to @Function which specifies whether or not a function was defined implicitly (e.g., getter/setter).
TEST=service/get_object_rpc_test.dart
Change-Id: I2da1aa15c58635046b905e205bfd29d654bbde2f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/203200
Reviewed-by: Siva Annamalai <asiva@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
Allows for tools like the Dart test runner to hide the main isolate by
marking it as a system isolate, only showing the isolates users are
interested in.
TEST=pkg/vm_service/mark_main_isolate_as_system_isolate_test.dart
Change-Id: I24d0f20614e89076a05499c206d576c355489a13
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/207203
Reviewed-by: Jacob Richman <jacobr@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
Previously, prologue builder omitted loading of delayed type arguments
from closure object in case of inling. This would effectively switch
from using delayed type arguments of an instantiated generic closure to
type arguments passed on the stack (null/all-dynamic if none), which is
not correct.
The fix is to load delayed type arguments from closure object even
in case of inlining.
TEST=tests/language/regress/regress46550_test.dart
Fixes https://github.com/dart-lang/sdk/issues/46550
Change-Id: Ibe50b0bc2ebda1d3638c4940fe9a26cd5ba7ce32
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/207704
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
that we do not have any state mismatch when the Dart VM is initialzed
and cleaned up multiple times.
TEST=existing unit test in the engine which is failing in a flaky manner
Change-Id: I073ca2b53a8c3d386be46b5222d547a29513714a
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/207642
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Siva Annamalai <asiva@google.com>
Ben Konyi