This change adds initial support for working with isolates in the Dart
Runtime Service and its backends.
The new `IsolateManager` base class tracks the set of active isolates
and their lifecycle events. The `VmIsolateManager` extends this class,
adding support specific to interacting with isolates within the Dart VM.
TEST=vm/cc/DartAPI_InvokeVMServiceMethod*_Exp
Change-Id: I3dfa298722c40dbdfdd58105cc78f31d058dd7a2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/486560
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
Reviewed-by: Jessy Yameogo <yjessy@google.com>
This change includes an initial implementation of the new VM service
implementation based on `package:dart_runtime_service`, along with the
necessary plumbing to start it in place of the legacy VM service
implementation.
The entrypoint for the new VM service implementation is located in
dart_runtime_service_vm/bin/vm_service_entrypoint.dart, which is
compiled into AppJIT and AOT snapshots when the
`--include-experimental-vm-service` flag is provided to `build.py`. To run
the VM with the new VM service implementation, the
`--experimental-vm-service` flag must be provided.
Currently, the experimental VM service implementation supports:
- User specified ports
- Authentication code flags
- Enabling the HTTP server via SIGQUIT
- Some service protocol RPCs that don't require an isolate ID (e.g.,
`getVM`)
See go/dart-runtime-services-unification for more details.
TEST=Manual
CoreLibraryReviewExempt: dart:_vmservice is private
Change-Id: I4a58cd1fa0a386313baa3d5c5345720231279123
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/484820
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Alexander Thomas <athom@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
This lets bloaty attribute the data symbol to its source file, allowing the binary size visualization to nest it appropriately.
TEST=runtime/tools/binary_size
Change-Id: I683ad6729d2fe2e1730dd3be70c6cce15bdf0003
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/486524
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
This allows a great deal of linker GC resulting in a much smaller binary. The benefit multiplies for Flutter development because the Flutter tool downloads many variant copies of gen_snapshot.
out/ReleaseX64/exe.stripped/gen_snapshot -1.55M (-23.4%)
TEST=ci
Change-Id: I9f6e650f78ea52c9c6953320a669b71019403236
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/482965
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Adds a new code path for NativeCallable.isolateLocal invocations. If
the current thread is not entered into any isolate, but owns the
target isolate, then it enters the target isolate, invokes, then exits
the isolate.
Fixes: https://github.com/dart-lang/sdk/issues/61623
TEST=tests/ffi/function_callbacks_isolate_ownership_test.dart
Change-Id: I401f185fadf7d2a55190dafd15387e1c418c67c9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/452380
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Liam Appelbe <liama@google.com>
- copy sanitizer runtime when using ubsan;
- avoid use of winnt.h CONTAINING_RECORD since it triggers ubsan "member access within null-pointer of type" error, use our copy which uses `offsetof`;
- have default virtual destructor in `ValueObject` to avoid ubsan complains about "insufficient space for an object of type 'dart:ValueObject'" at NoTemporaryAllocator use/declaration site;
- have virtual destructor in ZoneAllocated to avoid ubsan complains about "not having enough space to allocate object" at new RegExpEmpty() instantiation site;
- avoid using crashpad with ubsan as it causes dartvm to exit with error code 3;
- switch to windows, mac-friendly `[[gnu::no_sanitize(check)]]` from `__GNUC__` and `__has_feature` checks.
TEST=ci
Bug: https://github.com/dart-lang/sdk/issues/62267
Change-Id: I8b922a8da329af276d4cefaa88fb841cc0457124
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/469840
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Aprelev <aam@google.com>
For structs, we were already loading the size of the struct at
runtime for the size of the memcopy. This CL makes that work for
the `Array`s as well.
TEST=pkg/vm/testcases/transformations/ffi/regress_62087.dart
TEST=tests/ffi/native_assets/asset_absolute_test.dart
Closes: https://github.com/dart-lang/sdk/issues/62087
Change-Id: I15a5796d713879a613971e080d18101111704b9b
Cq-Include-Trybots: dart/try:vm-aot-linux-debug-arm64-try,vm-aot-linux-debug-x64-try,vm-aot-linux-debug-x64c-try,vm-aot-mac-debug-arm64-try,vm-aot-mac-debug-x64-try,vm-aot-obfuscate-linux-release-x64-try,vm-aot-optimization-level-linux-release-x64-try,vm-aot-win-debug-arm64-try,vm-aot-win-debug-x64-try,vm-aot-win-debug-x64c-try,vm-asan-linux-release-arm64-try,vm-asan-linux-release-x64-try,vm-asan-mac-release-arm64-try,vm-asan-win-release-x64-try,vm-dyn-linux-debug-x64-try,vm-dyn-mac-debug-arm64-try,vm-ffi-dyn-mac-debug-simarm64_arm64-try,vm-msan-linux-release-arm64-try,vm-msan-linux-release-x64-try,vm-tsan-linux-release-arm64-try,vm-tsan-linux-release-x64-try,vm-tsan-mac-release-arm64-try,vm-ubsan-linux-release-arm64-try,vm-ubsan-linux-release-x64-try,vm-ubsan-mac-release-arm64-try,vm-ubsan-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/464706
Reviewed-by: Tess Strickland <sstrickl@google.com>
Reviewed-by: Michael Goderbauer <goderbauer@google.com>
This format specifier does not perform a wide char to utf8 conversion
but rather a locale dependent wide char to multibyte conversion
(essentially wcstombs), which is not what this function is expecting
to do (it is expected to return utf8 encoded error message).
Issue https://github.com/dart-lang/sdk/issues/62252
TEST=not tested, speculative fix
Change-Id: Icd5ddbf9969dbd4acd2f7f48626ce0bd6a6a6964
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/468800
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
The code for inheriting stdio handles was not taking into account that
some of the stdio handles might be the same causing them to be
duplicated in the list of inherited handles. However passing duplicate
handles in PROC_THREAD_ATTRIBUTE_HANDLE_LIST causes CreateProcess to
return E_INVALIDARG.
This CL fixes the code to avoid duplicate handles.
We also fix printing of errors, which was broken when Process::Exec
was used without Dart scope causing it to print (null) instead of
actual error message.
Fixes https://github.com/dart-lang/sdk/issues/61981
TEST=vm/dart/regress_61981
Fixed: 61981
Change-Id: I93a6b56476f5020de9ebf894614cee51586b6751
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/464382
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
This allows analyze_snapshot to analyze Mach-O snapshots as well, and
also removes the restriction on only building analyze_snapshot for
Linux and Android.
TEST=vm/dart/analyze_snapshot_binary_test
Change-Id: I82797db2f20a4a80cdc52922aeebc789f45bf9bb
Cq-Include-Trybots: luci.dart.try:vm-aot-mac-release-arm64-try,vm-aot-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/461482
Reviewed-by: Slava Egorov <vegorov@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
In order to appropriately generate .dSYMs for a snapshot and allow
the strip tool to be used on the snapshot afterwards, add a mode that
not only outputs the snapshot (sans DWARF information) but also
an associated relocatable object file that contains the program's DWARF
information. That allows dsymutil to retrieve the DWARF information
from the relocatable object when run on the snapshot prior to stripping.
To specify that the relocatable object file should be output, use
the new --macho-object command line argument to gen_snapshot
to specify where the object file should be written.
This CL also adds an additional command line argument,
--macho-reduce-padding, which reduces the alignment used for segments
and the text/const sections in Mach-O outputs from 16KB to 64 bytes.
The larger padding is needed for some uses, like non-native loading
of Mach-O objects, but can be elided for other uses like Flutter builds.
TEST=vm/dart/use_dwarf_stack_traces_flag_test
vm/dart/use_macho_reduce_padding_flag_test
Change-Id: I2bf4bacb70c41299b8b6fdb7635c2374acf7a07d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/457420
Reviewed-by: Slava Egorov <vegorov@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Basic support for Unix domain sockets on Windows within
the limitations of the operating system itself: no
support for abstract addresses, datagram or
packet sockets and ancillary data.
This also fixes File::GetType on Windows to correctly
identify reparse points representing AF_UNIX sockets
as such, rather then identify them as links.
Finally, it is worth noting that there is an existing
discrepancy between POSIX OSes and Windows: on POSIX OSes
File(sockPath).existsSync() returns true, but on Windows the
same code returns false because Unix domain sockets are not
considered regular files by stat.
This CL does a bunch of refactoring around RawAddr class which
surves as a wrapper around various structures in sockaddr_* family.
Distinguishing anonymous AF_UNIX address from abstract AF_UNIX
address requires passing around the length of the address
structure. Thus we incorporate this information into
RawAddr. This will also make possible to properly support
full-range of abstract AF_UNIX addresses in the future because
supporting these requires properly handling embedded NUL bytes.
(See https://github.com/dart-lang/sdk/issues/46158).
Fixes https://github.com/dart-lang/sdk/issues/41161
TEST=standalone/io/unix_socket
Cq-Include-Trybots: luci.dart.try:vm-fuchsia-release-x64-try
Change-Id: I016cb33ebdd62f0cac1ae97d822105366a6a6964
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/457720
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
Instead of requiring embedder to initialize dart:io by bits and pieces
by invoking various hooks and setting fields provide a structured C++
API for preparing dart:io for execution.
TEST=ci
CoreLibraryReviewExempt: VM specific changes only
Change-Id: I66e95c0a451d384b4f7582d1df059e506a6a6964
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/454062
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
This CL tries to address version skew issues which occur in internal
monorepo which we created by checking in *.pbzero.h files into
runtime/vm/protos/*. These files refer to protozero internal headers,
which in the internal monorepo come from a shared Perfetto dependency
which does not necessarily match the version we pinned in our DEPS and
used to generated these files.
This version skew caused problems for the Perfetto team, see
https://github.com/google/perfetto/pull/3195
To resolve this change layout of our repo to allow us to depend on the
shared Perfetto entirely - including its version of *.pbzero.h headers.
We change include paths in the VM source to
#include "third_party/perfetto/protos/..."
which match location of these headers in the internal monorepo.
To make the code compile in the normal SDK checkout we shift location of
Perfetto and checked in code:
* Perfetto source moves from third_party/perfetto to
third_party/perfetto/src.
* Generated code moves from runtime/vm/protos into
third_party/perfetto/protos.
Note: experiments show that *.pbzero.h files tree-shake very well and
you pay only for things that you use. That means depending on *.pbzero.h
files generated from full message definitions rather than our manually
tree-shaken versions does not actually increase the size of the VM
binaries. We could switch our external build to do the same but this
would make protoc a build time dependency, which seems excessive.
TEST=ci
Change-Id: I724b9d9304be83302a5d83c2337154b36a6a6964
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/454260
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Ben Konyi <bkonyi@google.com>
Fixes bug in commit 25df2b3f11.
The code for converting FSEventStream notification into Dart format
was checking file existence using relative path instead of absolute
path. This caused it to incorrectly classify rename events which
were creations as deletions.
This was not caught by testing because this was not well covered
by the test and the test closest to checking this was even
disabled on Mac. CL updates updates the test to cover this better.
Fixes https://github.com/dart-lang/sdk/issues/61693
TEST=standalone/io/file_system_watcher_test
Change-Id: I2a7c984dfcec0be74c974f13b6ead7d86a6a6964
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/453880
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Watcher used pipes to fit into the "socket"-like implementation used by
other OSes. However it turns out that pipe buffer on Mac OS X is not
necessarily large enough to fit the whole FSEvent structure meaning that
we can't expect to be able to read FSEvents atomically from a pipe.
This situation happens because Mac OS X inherits BSD behavior of
limiting total number of kernel memory reserved for pipe buffers. Once
kern.ipc.maxpipekva limit is crossed pipe buffers no longer grow to 64KB
and remain 512 bytes large (which is absolute minimum allowed by POSIX
which requires writes smaller than PIPE_BUF to be atomic).
So our code needs to be prepared that only half of FSEvent structure
(which is 1032 bytes large) will fit into pipe buffer.
We could fix this by reading each FSEvent in chunks but this seems
ridiculously ineffecient. Instead we rewrite the code to pass file
system events via SendPort instead.
Fixes https://github.com/dart-lang/sdk/issues/61551
TEST=runtime/tests/vm/dart/regress_61551_test.dart
CoreLibraryReviewExempt: VM only changes.
Change-Id: I6a6a696499044832cfb61998bc6a519bbaadf8bc
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/451260
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Existing implementation is an entangled mess which consists of shared
code residing in the base class which in random places invokes a number
of undocumented poorly named methods overloaded in OS specific
subclasses. Some of these methods mutate static state. There are no
clear lifetime guarantees for different parts of the system (including
comments saying that some values might or might not be valid at certain
points).
The rewrite aims to clean most of this up - sharing everything that can
be shared and moving OS specific logic to clearly documented methods.
Furthermore, we change the code to ensure proper lifetime guarantees -
so we no longer find ourself in situations where we don't know whether
pathId is valid or not.
This refactoring by itself fixes a number of issues, most specifically a
bug where watcher would stop receiving events on Windows because
DirectoryWatchHandle ends up allocated at precisely the same address as
a previous destroyed one - which confuses Dart side to think that newly
created handle is the same as the old one (due to a race between event
handler thread and Dart thread).
We fix Windows lifetime issue by a) not keeping pathId based mapping in
the watcher anymore and b) keeping DirectoryWatchHandler alive until it
is stoped by the Dart side - this is achieved by retaining it after it
is created and releasing it once path is unwatched. This way Dart side
is always sure that pathId values are valid until they are explicitly
released via _unwatchPath - which makes code very uniform.
To make sure that native objects created by _watchPath are released when
surrounding isolate exists abruptly (e.g. via Isolate.exit - without
letting Dart code to shutdown and call _unwatchPath naturally) we attach
NativeFinalizer to them. This fixes the existing leak of file watchers
on Mac OS X - as Node objects it created were not freed if surrounding
isolate exited. Note that inotify descriptors did not leak in the same
way because they were wrapped into sockets.
Finally, this refactoring also make sure that the last subscriber
cancelling subscription on filesystem event stream will get a proper
cancellation future back and can wait for the watcher to shutdown.
Previously implementation used broadcast streams which simply return an
already completed future when subscriber cancels. New implementation
uses Stream.multi instead which gives a better result. Now doing
watch().listen().cancel() returns a future which will only complete once
watcher is fully disposed (e.g. inotify descriptor is closed). Bad
behavior was revealed by analysing standalone/regress_52715 - which
revealed that repeatedly watching and cancelling might flakely cause us
to hit fd limit depending on whether eventhandler thread can keep up
closing file descriptors created by the main thread or not.
Fixes https://github.com/dart-lang/sdk/issues/61378
TEST=standalone/{regress_61378,file_system_watcher_isolate_exit_leak}
CoreLibraryReviewExempt: VM only changes.
Change-Id: I6a6a69642b1f2673f2be78434bc64270846ad8c5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/450921
Reviewed-by: Lasse Nielsen <lrn@google.com>
This library no longer has old school natives and uses FFI natives
instead which are configured in bootstrap instead.
TEST=ci
Change-Id: I6a6a6964e1ee692153dfeb8dbaf692b35851e893
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/452141
Reviewed-by: Alexander Aprelev <aam@google.com>
Ben Konyi