Why will the synchronous continuations no longer be executed synchronously? - c#

In the following article, Stephen Toub describes how to build a ManualResetEvent using the TPL. In particular, he states that in the following method:
public void Set()
{
var tcs = m_tcs;
Task.Factory.StartNew(
s => ((TaskCompletionSource<bool>)s).TrySetResult(true),
tcs,
CancellationToken.None,
TaskCreationOptions.PreferFairness,
TaskScheduler.Default);
tcs.Task.Wait();
}
the tcs.Task.Wait() will "block until the task being completed is actually finished (not including the task’s synchronous continuations, just the task itself)". Why won't we block on any synchronous continuations? That is, what will prevent any synchronous continuations on any Tasks awaiting tcs.Task's completion from executing synchronously?

He means that if you do is this way:
public void Set()
{
m_tsc.TrySetResult(true);
}
Then all synchronous continuations of TaskCompletionSource Task will run synchronously on this same thread, blocking Set method for the duration.
However if you do it the way described in your question, then all those continuations will still run, but on thread pool thread, because that's where you execute TrySetResult. As soon as TaskCompletionSource task completes - your tcs.Task.Wait() will return, before synchronous continuations will run. So in effect, Set method will not block waiting for them to complete.

This is because the work in ContinueWith is a different task. All ContinueWith variants say:
Returns
A new continuation Task.
and in the Remarks we can read:
Remarks
The returned Task will not be scheduled for execution until the current task has completed.
BTW. These days ContinueWith is discouraged and await is preferred in almost all cases.

the tcs.Task.Wait() will "block until the task being completed is actually finished (not including the task’s synchronous continuations, just the task itself)". Why won't we block on any synchronous continuations? That is, what will prevent any synchronous continuations on any Tasks awaiting tcs.Task's completion from executing synchronously?
Your question is about Tasks in general, not about the TaskCompletionSource.Task in particular. The answer is that when an incomplete Task is waited, the Wait is performed by a ManualResetEventSlim, which is Set by a task continuation, which is prioritized in front of all other continuations that might be attached previously. Here is a part of the private SpinThenBlockingWait method, from the Task.cs source code:
private bool SpinThenBlockingWait(int millisecondsTimeout,
CancellationToken cancellationToken)
{
bool infiniteWait = millisecondsTimeout == Timeout.Infinite;
uint startTimeTicks = infiniteWait ? 0 : (uint)Environment.TickCount;
bool returnValue = SpinWait(millisecondsTimeout);
if (!returnValue)
{
var mres = new SetOnInvokeMres();
try
{
AddCompletionAction(mres, addBeforeOthers: true);
if (infiniteWait)
{
bool notifyWhenUnblocked = ThreadPool.NotifyThreadBlocked();
try
{
returnValue = mres.Wait(Timeout.Infinite, cancellationToken);
//...
This method is invoked internally by the Wait, when the task is not complete at the moment, and the current thread must be blocked. The SetOnInvokeMres is a small class that inherits from the ManualResetEventSlim:
private sealed class SetOnInvokeMres : ManualResetEventSlim, ITaskCompletionAction
{
internal SetOnInvokeMres() : base(false, 0) { }
public void Invoke(Task completingTask) { Set(); }
public bool InvokeMayRunArbitraryCode => false;
}
The interesting point is the addBeforeOthers: true argument. This ensures that the ManualResetEventSlim will be signaled before invoking any continuation that is associated with user-supplied code. The Task Parallel Library exposes no public API that allows to attach a continuation with addBeforeOthers: true. Actually this parameter seems to exist exclusively for supporting the signaling of ManualResetEventSlims with priority.

Related

Why would One choose SemaphoreSlim Wait over WaitAsync [duplicate]

I'm trying to find out what is the difference between the SemaphoreSlim use of Wait and WaitAsync, used in this kind of context:
private SemaphoreSlim semaphore = new SemaphoreSlim(1);
public async Task<string> Get()
{
// What's the difference between using Wait and WaitAsync here?
this.semaphore.Wait(); // await this.semaphore.WaitAsync()
string result;
try {
result = this.GetStringAsync();
}
finally {
this.semaphore.Release();
}
return result;
}
If you have async method - you want to avoid any blocking calls if possible. SemaphoreSlim.Wait() is a blocking call. So what will happen if you use Wait() and semaphore is not available at the moment? It will block the caller, which is very unexpected thing for async methods:
// this will _block_ despite calling async method and using await
// until semaphore is available
var myTask = Get();
var myString = await Get(); // will block also
If you use WaitAsync - it will not block the caller if semaphore is not available at the moment.
var myTask = Get();
// can continue with other things, even if semaphore is not available
Also you should beware to use regular locking mechanisms together with async\await. After doing this:
result = await this.GetStringAsync();
You may be on another thread after await, which means when you try to release the lock you acquired - it might fail, because you are trying to release it not from the same thread you acquired it. Note this is NOT the case for semaphore, because it does not have thread affinity (unlike other such constructs like Monitor.Enter, ReaderWriterLock and so on).
The difference is that Wait blocks the current thread until semaphore is released, while WaitAsync does not.

Async method blocking on unawaited task

In my current project, I have a piece of code that, after simplifying it down to where I'm having issues, looks something like this:
private async Task RunAsync(CancellationToken cancel)
{
bool finished = false;
while (!cancel.IsCancellationRequested && !finished)
finished = await FakeTask();
}
private Task<bool> FakeTask()
{
return Task.FromResult(false);
}
If I use this code without awaiting, I end up blocking anyway:
// example 1
var task = RunAsync(cancel); // Code blocks here...
... // Other code that could run while RunAsync is doing its thing, but is forced to wait
await task;
// example 2
var task = RunAsync(cancelSource.Token); // Code blocks here...
cancelSource.Cancel(); // Never called
In the actual project, I'm not actually using FakeTask, and there usually will be some Task.Delay I'm awaiting in there, so the code most of the time doesn't actually block, or only for a limited amount of iterations.
In unit testing, however, I'm using a mock object that does pretty much do what FakeTask does, so when I want to see if RunAsync responds to its CancellationToken getting cancelled the way I expect it to, I'm stuck.
I have found I can fix this issue by adding for example await Task.Delay(1) at the top of RunAsync, to force it to truly run asynchronous, but this feels a bit hacky. Are there better alternatives?
You have an incorrect mental picture of what await does. The meaning of await is:
Check to see if the awaitable object is complete. If it is, fetch its result and continue executing the coroutine.
If it is not complete, sign up the remainder of the current method as the continuation of the awaitable and suspend the coroutine by returning control to the caller. (Note that this makes it a semicoroutine.)
In your program, the "fake" awaitable is always complete, so there is never a suspension of the coroutine.
Are there better alternatives?
If your control flow logic requires you to suspend the coroutine then use Task.Yield.
Task.FromResult actually runs synchronously, as would await Task.Delay(0). If you want to actually simulate asynchronous code, call Task.Yield(). That creates an awaitable task that asynchronously yields back to the current context when awaited.
As #SLaks said, your code will run synchronously. One thing is running async code, and another thing is running parallel code.
If you need to run your code in parallel you can use Task.Run.
class Program
{
static async Task Main(string[] args)
{
var tcs = new CancellationTokenSource();
var task = Task.Run(() => RunAsync("1", tcs.Token));
var task2 = Task.Run(() => RunAsync("2", tcs.Token));
await Task.Delay(1000);
tcs.Cancel();
Console.ReadLine();
}
private static async Task RunAsync(string source, CancellationToken cancel)
{
bool finished = false;
while (!cancel.IsCancellationRequested && !finished)
finished = await FakeTask(source);
}
private static Task<bool> FakeTask(string source)
{
Console.WriteLine(source);
return Task.FromResult(false);
}
}
C#'s async methods execute synchronously up to the point where they have to wait for a result.
In your example there is no such point where the method has to wait for a result, so the loop keeps running forever and thereby blocking the caller.
Inserting an await Task.Yield() to simulate some real async work should help.

Does a method has to be async when invoked inside an async method?

Lets say I have a method defined as follows:
public async Task CreateUser()
{
await GetUserDetails();
GetUserOrder();
}
private void GetUserDetails() {
private void GetUserOrder() {
Does the method GetUserDetails(); and GetUserOrder() have to be async as well to avoid UI blocking ?
I cannot await the GetUserDetails() method since it is not async. How can I achieve this in c# ? I want to ensure all these methods are invoked step by step.
The relevant question is in a comment:
How can I ensure all my methods are invoked completely sequentially?
The fact that you're asking the question indicates that you don't understand what "await" is. Await is the sequencing operation on a asynchronous workflows. An await means this workflow will not proceed until the awaited task is complete. It's an asynchronous wait, hence the name await.
Consider this question: in a synchronous workflow, what is the sequencing operation?
No, really, give it some thought.
.
.
.
It is ;. When you say
fResult = foo();
bResult = bar();
qResult = qux();
that means that foo has to finish completely before bar can begin. That is not true for asynchronous workflows. If we have
fTask = fooAsync();
bTask = barAsync();
qTask = quxAsync();
Then the asynchronous operations can complete in any order. If we say
await fooAsync();
await barAsync();
await quxAsync();
Then barAsync will not start until fooAsync's task completes. await sequences the asynchronous workflow. The difference is that the thread can continue to do other unrelated work while asynchronously waiting for foo to complete, which is not true in a synchronous workflow; in a synchronous workflow the thread is already busy computing the foo result, so it can't do other work.
yes if you want to wait than you have to write await for that methods also. because after first await your code agian will be synchronous ..and if UI thread than it will run on it.
1.you code will be , so by this code you code become asynchronous for GetUserORder also. you just have to wrap method in Task construct and return
public async Task CreateUser()
{
await GetUserDetails();
await Task.Factory.SartNew(()=> GetUserOrder());
}
2.or you can do this also
public async Task CreateUser()
{
await Task.Factory.SartNew(()=>{
GetUserDetails();
GetUserOrder(); });
}
3.or you can do like this also, in below code will not wait for getuserorder method and excute await one method
public async Task CreateUser()
{
Task.Factory.SartNew(()=> GetUserOrder()).ContinueWith((t)=> Console.WriteLine("Completed");
await GetUserDetails();
}
4.or last one variation, here you start GetUserOrder first and dont wait for it than you call GetUserDetails in async fashion , and if you want to work on GetUserOrder method want to wait just use Wait method.
public async Task CreateUser()
{
var task = Task.Factory.SartNew(()=> GetUserOrder());
await GetUserDetails();
if(!task.IsCompleted)
task.Wait();
}
in your case you can go for 3 and if you want to wait go for 4th one.
As you asked me in comment what is difference between Task.Run and statnew method -: for that you can check this SO question : Regarding usage of Task.Start() , Task.Run() and Task.Factory.StartNew()
You should put await only in front of async methods. To run a synchronous one that you don't want to wait, you can use a new thread from the tread pool:
new Thread(() => DoSomething()).Start();
or
Task.Factory.SartNew(()=> DoSomething());
Here is the help page: https://msdn.microsoft.com/en-us/library/dd321439(v=vs.110).aspx
Otherwise, your call to GetUserDetails will have to finish before you execute the next line.

How to wait for Task to start?

How can I wait for the task to start. The following code fails:
var asyncmethod = ...a Task<TReturn>, with .Start() called on it...;
int waitcounter = 0;
while (!asyncmethod.Wait(1000))
{
waitcounter++;
Log("waiting very long...");
}
ret = asyncmethod.Result;
The asyncmethod.Wait(1000) waits 1 seconds as expected, but the Task is in the state WaitingToRun and will never start running when Wait()ing. On the other hand, when .Result is called, it will start running. How to get it to run without calling .Result?
the Task is in the state WaitingToRun and will never start running when Wait()ing
When a task is in the WaitingToRun state, that means it is ready to start running and is just waiting for its scheduling context to be available, so it can be scheduled and run (as I describe on my blog).
Since the task is still in this state after Wait(1000) finishes, then presumably the task is waiting for the scheduling context that is used by the calling thread, and thus cannot be scheduled until that thread is free.
Task.Result can trigger task inlining and execute task, but apparently Wait() cannot.
Both .Result and .Wait() will permit the task to be inlined, but of course .Wait(x) cannot because it has to honor the timeout.
However, neither .Result nor .Wait() will guarantee inlining - and it's important to keep that in mind when writing reliable code.
the code shouldn't break, regardless of if the task is scheduled on the "current" or a separate thread.
That's an extremely difficult requirement to satisfy. Are you sure you need that?
The easiest solution would be to wait asynchronously:
Task<T> asyncmethod = ...;
int waitcounter = 0;
while (await Task.WhenAny(Task.Delay(1000), asyncmethod) != asyncmethod)
{
waitcounter++;
Log("waiting very long...");
}
ret = await asyncmethod;
Just wait for the task to be completed using:
asyncmethod.Start();
asyncmethod.Wait(); // not needed in most cases
// but if used, the task is completed at this point.
var ret = asyncmethod.Result; // automatically waits for the task to be completed
but basically, the waiting is not neccesary, unless you have a reason for this. From the Task<TResult>.Result-docs:
The get accessor for this property ensures that the asynchronous
operation is complete before returning. Once the result of the
computation is available, it is stored and will be returned
immediately on later calls to Result. (from msdn)
Not really sure why you're doing this, but this can be achieved without blocking the calling thread using Task.IsCompleted and Task.Delay:
public async Task FooAsync()
{
var waitCounter = -1;
var task = Task.Run(() => { });
do
{
waitCounter++;
await Task.Delay(1000);
}
while (!task.IsCompleted)
}
This snippet will call Log a single time if the Task takes more than 1000ms to complete.
private async static void StartTask()
{
Task<object> asyncmethod = ... ;
LogDurationTooLong(asyncmethod, 1000);
var result = await asyncmethod;
}
/// <summary>
/// Logs if a task takes too long to complete.
/// </summary>
/// <param name="asyncmethod">The task to reference.</param>
/// <param name="duration">The duration after which a log entry is made.</param>
private async static void LogDurationTooLong(Task asyncmethod, int duration)
{
Task completedTask = await Task.WhenAny(Task.Delay(duration), asyncmethod);
if (completedTask != asyncmethod)
{
Log("waiting very long...");
}
}

Task inlining and Task.Wait

I just realized that when i start a task from within a task and call Task.Wait the new task will not be inlined, while calling Task.Result will always inline the task.
As we wrap our tasks with a RAII pattern (implemented in ExecuteWithCancel), inlining will reuse allocated resources and is preferable.
But we sometime want to wait a certain time and cancel the task after that.
The waiting code looks like this:
using (var cts = new CancellationTokenSource())
{
// Task scheduler decides whether to execute synchronous or asynchronous
var task = new Task<TResult>(() => ExecuteWithCancel<TResult>(cts.Token, nameOfTaskPerformer, arguments), cts.Token)
if (timeout==TimeSpan.Zero || task.Wait(timeout)) // this creates an all or nothing timeout
return task.Result;
cts.Cancel();
throw new TimeoutException("");
}
When timeout is TimeSpan.Zero the Task is inlined, otherwise it always uses another thread.
Is there an easy way to redesign this code to use inlining and waiting/timeout?
Pretty sure that's not possible. Suppose you are running the following code on thread A:
var task = Task.Factory.StartNew(() => Thread.Sleep(Timeout.Infinite));
task.Wait(5000);
If the task is inlined, thread A will block indefinitely - how will it wake up after the timeout?
Looking at the reference source (Task.cs) we can see exactly that:
internal bool InternalWait(int millisecondsTimeout, CancellationToken cancellationToken)
{
...
// we will attempt inline execution only if an infinite wait was requested
// Inline execution doesn't make sense for finite timeouts and if a cancellation token was specified
// because we don't know how long the task delegate will take.
if (millisecondsTimeout == Timeout.Infinite && !cancellationToken.CanBeCanceled &&
WrappedTryRunInline() && IsCompleted)
{
returnValue = true;
}
else
{
returnValue = CompletedEvent.Wait(millisecondsTimeout, cancellationToken);
}
As per your question, in order to benefit from inlining with finite timeouts, you'd have to implement the timeout logic inside the task itself, perhaps something like:
ExecuteWithCancel<TResult>(cts.Token, TimeSpan timeout, nameOfTaskPerformer, arguments)
And then use a regular Wait() (or Result).

Categories

Resources