Commit Graph

31 Commits

Author SHA1 Message Date
Vyacheslav Egorov bf36861cb3 [vm] Fix race in NativeMessageHandler
Introduce `NativeMessageHandler::Cleanup` which
waits for all pending `NativeMessageHandler`
deletions to complete.

Fixes https://github.com/dart-lang/sdk/issues/56744

TEST=hard to distill this specific race into a test

Change-Id: Id6196f955ff7a874c07d3a11f766d74accbde96c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/385780
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
2024-09-18 21:07:59 +00:00
Vyacheslav Egorov e2efffa005 [vm] Simplify implementation of native ports
This reland commit 5a32d8bc7c with a fix
for thread leak (Issue #56717): when `ThreadPool` is shutting down
asynchronously the last worker should detach itself to prevent
leaking associated low-level data structures, because no thread will
join it.

A hang in service isolate shutdown (caused by an existing bug) was fixed by commit 157a0dc7f9.

This CL turns native ports into a thin abstraction over underlying
thread pool instead of building them as full fledged MessageHandler.

This allows to easily implement a variation of native ports which can
handle messages concurrently with the given degree of concurrency.
This type of port can be used to greatly simplify implementation of
IOService - which previously had to do its own concurrency management
on top of "single threaded" native ports. This capability is exposed
as `Dart_NewConcurrentNativePort` API.

The new implementation is in general much cleaner then the old one
with one exception: `Dart_CloseNativePort` API has unfortunate design
where underlying message handler is destroyed asynchronously and
`Dart_CloseNativePort` returns immediately without waiting for pending
tasks to complete. Implementing this on top of `ThreadPool` requires
some changes to thread pool implementation.

Issue https://github.com/dart-lang/sdk/issues/55844

Closes https://github.com/dart-lang/sdk/issues/56717

TEST=ci

Change-Id: Ic68bfb60757685afd75c80a70cdec66cc13c149b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/385000
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
2024-09-12 15:16:12 +00:00
Slava Egorov 120dac365d Revert "[vm] Simplify implementation of native ports"
This reverts commit 5a32d8bc7c.

Reason for revert: DartIsolateTest.CanCreateServiceIsolate flutter engine unit test hangs on Windows.

Original change's description:
> [vm] Simplify implementation of native ports
>
> This CL turns native ports into a thin abstraction over underlying
> thread pool instead of building them as full fledged MessageHandler.
>
> This allows to easily implement a variation of native ports which can
> handle messages concurrently with the given degree of concurrency.
> This type of port can be used to greatly simplify implementation of
> IOService - which previously had to do its own concurrency management
> on top of "single threaded" native ports. This capability is exposed
> as `Dart_NewConcurrentNativePort` API.
>
> The new implementation is in general much cleaner then the old one
> with one exception: `Dart_CloseNativePort` API has unfortunate design
> where underlying message handler is destroyed asynchronously and
> `Dart_CloseNativePort` returns immediately without waiting for pending
> tasks to complete. Implementing this on top of `ThreadPool` requires
> some changes to thread pool implementation.
>
> Issue https://github.com/dart-lang/sdk/issues/55844
>
> TEST=ci
>
> Change-Id: I062040ff233e93962ae93684e9b044d8facdaffc
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/382163
> Commit-Queue: Slava Egorov <vegorov@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>

Change-Id: I7de27793a54072e974bf1a9f17a07c12159a202d
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/384481
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
2024-09-12 08:25:27 +00:00
Vyacheslav Egorov 5a32d8bc7c [vm] Simplify implementation of native ports
This CL turns native ports into a thin abstraction over underlying
thread pool instead of building them as full fledged MessageHandler.

This allows to easily implement a variation of native ports which can
handle messages concurrently with the given degree of concurrency.
This type of port can be used to greatly simplify implementation of
IOService - which previously had to do its own concurrency management
on top of "single threaded" native ports. This capability is exposed
as `Dart_NewConcurrentNativePort` API.

The new implementation is in general much cleaner then the old one
with one exception: `Dart_CloseNativePort` API has unfortunate design
where underlying message handler is destroyed asynchronously and
`Dart_CloseNativePort` returns immediately without waiting for pending
tasks to complete. Implementing this on top of `ThreadPool` requires
some changes to thread pool implementation.

Issue https://github.com/dart-lang/sdk/issues/55844

TEST=ci

Change-Id: I062040ff233e93962ae93684e9b044d8facdaffc
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/382163
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
2024-09-11 13:38:24 +00:00
Martin Kustermann 06b2967188 [vm] Disentagle PortMap/MessageHandler from port status tracking
The main purpose of the low-level [PortMap] is to coordinate between
ports being opened & closed and concurrent message senders. That's the
only thing it should do.

Each isolate owns [ReceivePort]s. Only the isolate mutator can create
ports, delete them or change their "keeps-isolate-alive" state. Right
now it requires going via [PortMap] (which acquires lock) and
[MessageHandler] (which acquires lock) to change the
"keeps-isolate-alive" state of a port.

We'll move information whether a dart [ReceivePort] is closed and
whether it keeps the isolate alive into the [ReceivePort] object itself.

=> Changing the "keeps-isolate-alive" state of the port no longer
requires any locks. We could even avoid the runtime call itself in a
future CL.

Isolates are kept alive if there's any open receive ports (that have not
been marked as "does not keep isolate alive"). This is a property of an
isolate not of the message handler. For native message handlers we do
have a 1<->1 correspondence between port and handler (i.e. there's no
"number of open ports" tracking needed).

=> We'll move the logic of counting open receive ports and ports that
keep the isolate alive to the [Isolate].
=> We'll also remove locking around incrementing/decrementing or
accessing the counts.
=> The [IsolateMessageHandler] will ask the [Isolate] whether there's
any open ports for determining whether to shut down.
=> For native ports, the `Dart_NewNativePort()` & `Dart_CloseNativePort()`
functions will manage the lifetime (as their name also suggests).

Overall this makes the [Isolate] responsible for creation of dart
[ReceivePort]s and tracking whether the isolate should be kept alive:
  * Isolate::CreateReceivePort()
  * Isolate::SetReceivePortKeepAliveState()
  * Isolate::CloseReceivePort()

TEST=ci

Change-Id: I847ae357c26254d3810cc277962e05deca18a1de
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/317960
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
2023-08-08 10:57:47 +00:00
Ryan Macnak f9a6a5bdd2 [vm] Update NULL to nullptr in runtime/vm.
TEST=build
Change-Id: I2834ef7cf7cb7c8770f8167a2438cbedcee5c623
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/292063
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
2023-04-10 18:15:12 +00:00
Ryan Macnak bbefc05748 [vm] Rewrite isolate message serialization.
TEST=ci
Bug: https://github.com/dart-lang/sdk/issues/25559
Bug: https://github.com/dart-lang/sdk/issues/27285
Bug: https://github.com/flutter/flutter/issues/84691
Change-Id: I04608c5d337144195748d5f3509a29eb3116585d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/207865
Reviewed-by: Siva Annamalai <asiva@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
2021-07-29 16:55:55 +00:00
Alexander Aprelev d33e51df3c [build] Support --clang when building dart sdk on Windows.
It relies on flutter copy of clang distribution, same one that is used to build flutter/engine.

It addressed several deprecated warnings from clang compiler for functions like strdup, unlink, etc.

It allows few warnings still since they are triggered in third_party sources.

Change-Id: Ieb13792c011438d46dbbc0fa030e1b5e4ea14315
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/142704
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
2020-06-25 03:36:41 +00:00
Matthew Dempsky a53c12d07a [vm] Use std::unique_ptr with Message
Message is a C++ type with a simple ownership model appropriate for
std::unique_ptr. This CL applies the following changes:

1. All uses of "new Message(...)" are replaced with
"Message::New(...)", which is effectively
"std::make_unique<Message>(...)". (The latter was only added in C++14,
but Dart still compiles in C++11 mode.)

2. All owning Message* are replaced with std::unique_ptr<Message>. The
notable exception is MessageQueue, which still uses raw Message*
internally to simplify the linked list handling.

3. All "delete message;" statements are removed.

4. Uses of "NULL" replaced with "nullptr" as necessary.

Change-Id: I05b5804289f2a225bfa05d3c1631129358fed373
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/101222
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Matthew Dempsky <mdempsky@google.com>
2019-05-06 21:01:39 +00:00
Zachary Anderson 6cd8a79078 VM: Re-format to use at most one newline between functions
R=asiva@google.com

Review-Url: https://codereview.chromium.org/2974233002 .
2017-07-13 08:08:37 -07:00
Vyacheslav Egorov 9bae50bf47 VM: [Kernel] Revert changes to the native ports that introduced peers.
The problem with these changes was that closed ports still continue to receive
messages and as a result we start using dead peer objects.

Fixing that would require more intrusive changes into message handler implementation - so instead we are reverting the changes and restoring manual PORT -> PEER mapping.

BUG=
R=erikcorry@google.com

Review-Url: https://codereview.chromium.org/2666063002 .
2017-01-31 10:30:15 +01:00
Vyacheslav Egorov 43a0500e37 VM: [Kernel] Fix bootstraping when Kernel isolate is used.
We must bootstrap from Kernel instead of Source when running with --dfe

BUG=
R=asiva@google.com, kustermann@google.com

Review-Url: https://codereview.chromium.org/2651633002 .
2017-01-30 20:52:59 +01:00
Zachary Anderson a1bcf051d8 clang-format runtime/vm
R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org/2481873005 .
2016-11-08 13:54:47 -08:00
Todd Turnidge f8ff7bd5c4 Delete NativeMessageHandler once a native port closes.
Closes #27002

R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org/2217693002 .
2016-08-04 15:34:24 -07:00
John McCutchan 757b7d094c Revert "Fix NativeMessageHandler leak"
This reverts commit 425efb9808.

BUG=

Review URL: https://codereview.chromium.org/2207663002 .
2016-08-02 11:33:12 -07:00
John McCutchan 425efb9808 Fix NativeMessageHandler leak
Fixes #27002

R=zra@google.com

Review URL: https://codereview.chromium.org/2200903002 .
2016-08-02 11:08:18 -07:00
Zachary Anderson fa8a0613a3 Fold ApiObjectConverter use into ApiMessageReader
TODO from change on Friday.

R=turnidge@google.com

Review URL: https://codereview.chromium.org/1523013002 .
2015-12-14 11:12:25 -08:00
Zachary Anderson a0970835f0 Adds a special case for sending an int over a port with the native API.
This improves dart <=> C++ IPC round-trip-times for Mojo by 10-20%.

BUG=

Review URL: https://codereview.chromium.org/1499853004 .
2015-12-11 16:07:16 -08:00
Todd Turnidge 0d501ad53d VM restart + shutdown fixes
This change add the ability to restart the vm through the service
protocol.  All isolates are killed, and then the main isolate is
restarted cooperatively by the embedder.

This change also fixes the message handler to prevent it from
accidentally ignoring vm shutdown messages.

Previously, we would stop handling messages whenever we hit an error
(such as a compile error or an unhandled exception).  This would leave
shutdown requests sitting the oob queue, neglected.

We now process *all* oob requests, up to the first shutdown request.
When we hit a shutdown request, we clear the oob queue and process no
more messages.

To make all of this work, we had to change the return value of
HandleMessage from bool to a new enum type, allowing the message
handler to distinguish *normal* error cases from the more rarified
shutdown and restart cases.

R=johnmccutchan@google.com, zra@google.com

Review URL: https://codereview.chromium.org/1371193005 .
2015-10-06 11:27:26 -07:00
Siva Annamalai 810823f8d5 Use the zone in ApiNativeScope for allocating objects in the ApiMessageReader instead of passing in an allocator.
BUG=
R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org//1319583003 .
2015-08-25 14:48:41 -07:00
iposva@google.com 14ecdbd66a - Separate the thread implementation used in bin/ and vm/
to allow us to make VM specific changes.

R=asiva@google.com

Review URL: https://codereview.chromium.org//463993002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39172 260f80e4-7a28-3924-810f-c04153c831b5
2014-08-12 23:19:53 +00:00
sgjesse@google.com 2f27215203 Remove the reply port form the native isolate handler
The handler function for a native isolate no longer gets an explicit
reply port. Instead the reply port must be sent as part of the message
if required.

The Dart_CObject structure now exposes the send ports in a message to
a native isolate.

R=asiva@google.com, floitsch@google.com, whesse@google.com
BUG=

Review URL: https://codereview.chromium.org//43483004

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29418 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-29 08:45:32 +00:00
tball@google.com 2cd8436c5a Changed StackZone and ApiZone to be containers for Zone.
BUG=

Review URL: https://codereview.chromium.org//11028145

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@13707 260f80e4-7a28-3924-810f-c04153c831b5
2012-10-16 22:05:40 +00:00
turnidge@google.com bd995a345e Avoid trusting the length encoded in the Snapshot if there is an
external length available.  We now pass a length with all messages in
the vm and verify that there is no mismatch with the length from the
Snapshot.

Fixed a bug in the use of ApiMessageReader.  We were always manually
adding Snapshot::kHeaderSize to the data, but neglecting to subtract
kHeaderSize from the message length.

Added FullSnapshotWriter and MessageWriter classes.
Review URL: https://chromiumcodereview.appspot.com//10829444

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@11413 260f80e4-7a28-3924-810f-c04153c831b5
2012-08-27 20:25:44 +00:00
turnidge@google.com 2803d78dfd Change the zone allocation api.
Instead of passing a size in bytes to the allocation function, we now
have a templatized Alloc function:

  zone->Alloc<Type>(len)

This is better for security, as we can check for integer overflow in
the size computation before performing the allocation.  Before, we
often failed to check this.
Review URL: https://chromiumcodereview.appspot.com//10836061

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@10254 260f80e4-7a28-3924-810f-c04153c831b5
2012-08-03 21:51:44 +00:00
turnidge@google.com 9d72035ce5 Use the ThreadPool for all isolates and native ports. Previously,
each isolate or native port had a dedicated thread.

Refactored the MessageHandler api...

- Added a Run function to allow a MessageHandler to run on a
  ThreadPool.  These functions take a start and end callback to allow for
  isolate initialization and shutdown.

- Made the queue private to the MessageHandler and moved all message
  processing code inside the MessageHandler (got rid of all of the
  different flavors of RunLoop).  This helps remove some code
  duplication and hides the details of how messages are handled.

- Moved all locking and notification out of MessageQueue and moved it
  up to MessageHandler.  Moved OOB support out of MessageQueue and up
  to MessageHandler.  These changes make the MessageQueue much
  simpler.

- Refactored native port and isolate MessageHandlers to share more code.

- Improved --trace_isolates output.

- Added tests for MessageHandler.

Refactored lib/isolate code...

- Use the new MessageHandler::Run api.

- Got rid of the LongJump stuff in RunIsolate.  No longer needed.

- Use the new StartIsolateScope/SwitchIsolateScope to make the code
  less verbose and less error-prone.

- Store top-level isolate errors in the sticky_error.

Added StartIsolateScope/SwitchIsolateScope classes.
Review URL: https://chromiumcodereview.appspot.com//9924015

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@6762 260f80e4-7a28-3924-810f-c04153c831b5
2012-04-19 19:47:27 +00:00
turnidge@google.com e2a88a357a Revert my last change. Odd test failures that I will investigate tomorrow.
Review URL: https://chromiumcodereview.appspot.com//9570051

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@4846 260f80e4-7a28-3924-810f-c04153c831b5
2012-03-02 00:38:18 +00:00
turnidge@google.com cdf40e7b15 Make the message queue private in the message handler.
Move locking up from the message queue to the message handler.
Review URL: https://chromiumcodereview.appspot.com//9570046

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@4842 260f80e4-7a28-3924-810f-c04153c831b5
2012-03-01 23:57:49 +00:00
sgjesse@google.com 3481c5661a Decode the Dart message into a Dart_CMessage structure before calling the native port callback
The native port callback is now passed the message as a decodes
Dart_CMessage structure. The Dart_CMessage structure is allocated
in a zone and the callback receiving it should expect the
lifetime to be controlled by the caller.

Added support for zones which do not require a current
isolate. Changed the GrowableArray to support allocating in
aprovided zone instead of the zone for the current isolate.

R=turnidge@google.com, asiva@google.com

BUG=
TEST=

Review URL: https://chromiumcodereview.appspot.com//9325022

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@4068 260f80e4-7a28-3924-810f-c04153c831b5
2012-02-09 08:47:19 +00:00
sgjesse@google.com 42b70dcddb Change the thread interface in runtime/platform and use it starting all threads
The platform thread interface (dart::thread) is now refactored to an
all static interface as suggested by iposva@ and asiva@. Use this
interface for running all threads in the VM.

R=ager@google.com, iposva@google.com

BUG=
TEST=

Review URL: https://chromiumcodereview.appspot.com//9141005

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@3830 260f80e4-7a28-3924-810f-c04153c831b5
2012-02-02 08:59:26 +00:00
turnidge@google.com 1f5364cd3b Add support for native ports in the vm.
Dart_NewNativePort creates a port associated with a C handler
function.  When messages come in on this port, they are forwarded to
the C function for processing.

To support this, refactored PortMap so that it operates on a new
MessageHandler type instead of directly on Isolates.

For now, native ports have a dedicated single thread.  Eventually we
will back native ports (and possibly Isolates as well) by a shared
thread pool.
Review URL: https://chromiumcodereview.appspot.com//9169063

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@3804 260f80e4-7a28-3924-810f-c04153c831b5
2012-02-01 18:53:40 +00:00