gecko-dev/dom/serviceworkers/ServiceWorkerOp.h
Andrew Sutherland b21cc21e8a Bug 1775784 - Improve worker shutdown ordering. r=dom-worker-reviewers,smaug
This patch simplifies and cleans up the RemoteWorker state machine by:
- (In a previous patch in the stack) making PRemoteWorker refcounted instead of
  ManualAlloc allows us to eliminate use of ThreadSafeWeakPtr and some legwork
  that was attempting to ensure we only drop references to RemoteWorkerChild on
  the "Worker Launcher" thread.
  - Although we are not allowed to call IPC methods on a different thread, we
    can absolutely add and remove refcounts on other threads as needed.
  - This allowed removal of the `SelfHolder` class as well as a number of calls
    to NS_ProxyRelease.
  - This allowed removal of the `{get,m}RemoteWorkerControllerWeakRef` hack.
- Adding 2 optional lambda callbacks to the WorkerPrivate constructor that will
  be notified on the parent thread in order to simplify tracking the state of
  the worker:
  - CancellationCallback: Invoked when the parent thread tells the worker to
    cancel itself.  This can be the automatic result of a script load error,
    a call to `self.close()`, or an explicit call to Cancel() by the parent.
    - Note that cancellation was always being initiated by the parent thread,
      even in the case of `self.close()`!
  - TerminationCallback: Invoked when the worker thread has finished and the
    WorkerPrivate's self-ref is about to be dropped.  This indicates that the
    worker transitioned to Killing and fully completed that process.
- Changing the RemoteWorkerChild to always perform state transitions on the
  owning parent thread (although the state may be checked from other threads,
  more in the code comments).
  - The CancellationCallback mechanism replaces a complicated use of
    WorkerRefs.  This allowed removal of ReleaseWorkerRunnable.
  - In general this reduces the number of permutations of thread interactions
    we have to worry about.
- Changing the RemoteWorkerChild to only report that the worker has finished
  shutting down as a result of the TerminationCallback rather than when the
  worker has transitioned to Cancelling.
- Changing the states to better represent the worker's state in terms of
  WorkerStatus terminology.

Some related cleanups:
- Use of mechanisms provided by IPC:
  - RemoteWorkerChild had an mOwningEventTarget with a GetOwningEventTarget
    getter that we were able to replace with the IPC-maintained
    GetActorEventTarget().
  - We replaced the `mIPCActive` with use of IPC's CanSend().
- MOZ_LOG logging was added under "RemoteWorkerChild".

Differential Revision: https://phabricator.services.mozilla.com/D164644
2023-03-14 23:57:57 +00:00

200 lines
6.6 KiB
C++

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#ifndef mozilla_dom_serviceworkerop_h__
#define mozilla_dom_serviceworkerop_h__
#include <functional>
#include "mozilla/dom/ServiceWorkerOpPromise.h"
#include "nsISupportsImpl.h"
#include "ServiceWorkerEvents.h"
#include "ServiceWorkerOpPromise.h"
#include "mozilla/Attributes.h"
#include "mozilla/Maybe.h"
#include "mozilla/RefPtr.h"
#include "mozilla/TimeStamp.h"
#include "mozilla/dom/PromiseNativeHandler.h"
#include "mozilla/dom/RemoteWorkerChild.h"
#include "mozilla/dom/ServiceWorkerOpArgs.h"
#include "mozilla/dom/WorkerRunnable.h"
namespace mozilla::dom {
class FetchEventOpProxyChild;
class ServiceWorkerOp : public RemoteWorkerChild::Op {
public:
// `aCallback` will be called when the operation completes or is canceled.
static already_AddRefed<ServiceWorkerOp> Create(
ServiceWorkerOpArgs&& aArgs,
std::function<void(const ServiceWorkerOpResult&)>&& aCallback);
ServiceWorkerOp(
ServiceWorkerOpArgs&& aArgs,
std::function<void(const ServiceWorkerOpResult&)>&& aCallback);
ServiceWorkerOp(const ServiceWorkerOp&) = delete;
ServiceWorkerOp& operator=(const ServiceWorkerOp&) = delete;
ServiceWorkerOp(ServiceWorkerOp&&) = default;
ServiceWorkerOp& operator=(ServiceWorkerOp&&) = default;
// Returns `true` if the operation has started and `false` otherwise.
bool MaybeStart(RemoteWorkerChild* aOwner,
RemoteWorkerChild::State& aState) final;
void StartOnMainThread(RefPtr<RemoteWorkerChild>& aOwner) final;
void Cancel() final;
protected:
~ServiceWorkerOp();
bool Started() const;
bool IsTerminationOp() const;
// Override to provide a runnable that's not a `ServiceWorkerOpRunnable.`
virtual RefPtr<WorkerRunnable> GetRunnable(WorkerPrivate* aWorkerPrivate);
// Overridden by ServiceWorkerOp subclasses, it should return true when
// the ServiceWorkerOp was executed successfully (and false if it did fail).
// Content throwing an exception during event dispatch is still considered
// success.
virtual bool Exec(JSContext* aCx, WorkerPrivate* aWorkerPrivate) = 0;
// Override to reject any additional MozPromises that subclasses may contain.
virtual void RejectAll(nsresult aStatus);
ServiceWorkerOpArgs mArgs;
// Subclasses must settle this promise when appropriate.
MozPromiseHolder<ServiceWorkerOpPromise> mPromiseHolder;
private:
class ServiceWorkerOpRunnable;
bool mStarted = false;
};
class ExtendableEventOp : public ServiceWorkerOp,
public ExtendableEventCallback {
using ServiceWorkerOp::ServiceWorkerOp;
protected:
~ExtendableEventOp() = default;
void FinishedWithResult(ExtendableEventResult aResult) override;
};
class FetchEventOp final : public ExtendableEventOp,
public PromiseNativeHandler {
using ExtendableEventOp::ExtendableEventOp;
public:
NS_DECL_THREADSAFE_ISUPPORTS
/**
* This must be called once and only once before the first call to
* `MaybeStart()`; `aActor` will be used for `AsyncLog()` and
* `ReportCanceled().`
*/
void SetActor(RefPtr<FetchEventOpProxyChild> aActor);
void RevokeActor(FetchEventOpProxyChild* aActor);
// This must be called at most once before the first call to `MaybeStart().`
RefPtr<FetchEventRespondWithPromise> GetRespondWithPromise();
// This must be called when `FetchEvent::RespondWith()` is called.
void RespondWithCalledAt(const nsCString& aRespondWithScriptSpec,
uint32_t aRespondWithLineNumber,
uint32_t aRespondWithColumnNumber);
void ReportCanceled(const nsCString& aPreventDefaultScriptSpec,
uint32_t aPreventDefaultLineNumber,
uint32_t aPreventDefaultColumnNumber);
private:
class AutoCancel;
~FetchEventOp();
bool Exec(JSContext* aCx, WorkerPrivate* aWorkerPrivate) override;
void RejectAll(nsresult aStatus) override;
void FinishedWithResult(ExtendableEventResult aResult) override;
/**
* `{Resolved,Reject}Callback()` are use to handle the
* `FetchEvent::RespondWith()` promise.
*/
void ResolvedCallback(JSContext* aCx, JS::Handle<JS::Value> aValue,
ErrorResult& aRv) override;
void RejectedCallback(JSContext* aCx, JS::Handle<JS::Value> aValue,
ErrorResult& aRv) override;
void MaybeFinished();
// Requires mRespondWithClosure to be non-empty.
void AsyncLog(const nsCString& aMessageName, nsTArray<nsString> aParams);
void AsyncLog(const nsCString& aScriptSpec, uint32_t aLineNumber,
uint32_t aColumnNumber, const nsCString& aMessageName,
nsTArray<nsString> aParams);
void GetRequestURL(nsAString& aOutRequestURL);
// A failure code means that the dispatch failed.
nsresult DispatchFetchEvent(JSContext* aCx, WorkerPrivate* aWorkerPrivate);
// Worker Launcher thread only. Used for `AsyncLog().`
RefPtr<FetchEventOpProxyChild> mActor;
/**
* Created on the Worker Launcher thread and settled on the worker thread.
* If this isn't settled before `mPromiseHolder` (which it should be),
* `FetchEventOpChild` will cancel the intercepted network request.
*/
MozPromiseHolder<FetchEventRespondWithPromise> mRespondWithPromiseHolder;
// Worker thread only.
Maybe<ExtendableEventResult> mResult;
bool mPostDispatchChecksDone = false;
// Worker thread only; set when `FetchEvent::RespondWith()` is called.
Maybe<FetchEventRespondWithClosure> mRespondWithClosure;
// Must be set to `nullptr` on the worker thread because `Promise`'s
// destructor must be called on the worker thread.
RefPtr<Promise> mHandled;
// Must be set to `nullptr` on the worker thread because `Promise`'s
// destructor must be called on the worker thread.
RefPtr<Promise> mPreloadResponse;
// Holds the callback that resolves mPreloadResponse.
MozPromiseRequestHolder<FetchEventPreloadResponseAvailablePromise>
mPreloadResponseAvailablePromiseRequestHolder;
MozPromiseRequestHolder<FetchEventPreloadResponseTimingPromise>
mPreloadResponseTimingPromiseRequestHolder;
MozPromiseRequestHolder<FetchEventPreloadResponseEndPromise>
mPreloadResponseEndPromiseRequestHolder;
TimeStamp mFetchHandlerStart;
TimeStamp mFetchHandlerFinish;
};
} // namespace mozilla::dom
#endif // mozilla_dom_serviceworkerop_h__