Develop Reference

Deadlock when updating a UI control from a worker thread

To simplify the explanation of the strange behavior I am experiencing, I have this simple class named Log which fires 1 log events every 1000msec.
public static class Log
{
public delegate void LogDel(string msg);
public static event LogDel logEvent;
public static void StartMessageGeneration ()
{
for (int i = 0; i < 1000; i++)
{
logEvent.Invoke(i.ToString());
Task.Delay(1000);
}
}
}
I have the Form class below which is subscribed to the log events of the Log class so it can handle them and display in a simple text box.
Once a log message arrives, it is added to a list. Every 500msec, a timer object access that list so its content can be displayed in a text box.
public partial class Form1 : Form
{
private SynchronizationContext context;
private System.Threading.Timer guiTimer = null;
private readonly object syncLock = new object();
private List<string> listOfMessages = new List<string>();
public Form1()
{
InitializeComponent();
context = SynchronizationContext.Current;
guiTimer = new System.Threading.Timer(TimerProcessor, this, 0, 500);
Log.logEvent += Log_logEvent;
}
private void Log_logEvent(string msg)
{
lock (syncLock)
listOfMessages.Add(msg);
}
private void TimerProcessor(object obj)
{
Form1 myForm = obj as Form1;
lock (myForm.syncLock)
{
if (myForm.listOfMessages.Count == 0)
return;
myForm.context.Send(new SendOrPostCallback(delegate
{
foreach (string item in myForm.listOfMessages)
myForm.textBox1.AppendText(item + "\n");
}), null);
listOfMessages.Clear();
}
}
private void button1_Click(object sender, EventArgs e)
{
Log.StartMessageGeneration();
}
}
The problem I see is that sometimes, there is a dead lock (application stuck). Seems that the 2 locks (1st one for adding to the list and the 2nd one for "retrieving" from the list) are somehow blocking each others.
Hints:
1) reducing the rate of sending the messages from 1 sec to 200msec seems to help (not sure why)
2) Somehow something happens when returning to the GUI thread (using the synchronization context) and accessing the GUI control. If I don't return to the GUI thread, the 2 locks are working fine together...
Thanks everyone!

There's a few problems with your code, and a few... silly things.
First, your Log.StartMessageGeneration doesn't actually produce a log message every second, because you're not awaiting the task returned by Task.Delay - you're basically just creating a thousand timers very quickly (and pointlessly). The log generation is limited only by the Invoke. Using Thread.Sleep is a blocking alternative to Task.Delay if you don't want to use Tasks, await etc. Of course, therein lies your biggest problem - StartMessageGeneration is not asynchronous with respect to the UI thread!
Second, there's little point in using System.Threading.Timer on your form. Instead, just use the windows forms timer - it's entirely on the UI thread so there's no need for marshalling your code back to the UI thread. Since your TimerProcessor doesn't do any CPU work and it only blocks for a very short time, it's the more straight-forward solution.
If you decide to keep using System.Threading.Timer anyway, there's no point in manually dealing with synchronization contexts - just use BeginInvoke on the form; the same way, there's no point in passing the form as an argument to the method, since the method isn't static. this is your form. You can actually see this is the case since you omitted myForm in listOfMessages.Clear() - the two instances are the same, myForm is superfluous.
A simple pause in the debugger will easily tell you where the program is hung - learn to use the debugger well, and it will save you a lot of time. But let's just look at this logically. StartMessageGeneration runs on the UI thread, while System.Threading.Timer uses a thread-pool thread. When the timer locks syncLock, StartMessageGeneration can't enter the same lock, of course - that's fine. But then you Send to the UI thread, and... the UI thread can't do anything, since it's blocked by StartMessageGeneration, which never gives the UI an opportunity to do anything. And StartMessageGeneration can't proceed, because it's waiting on the lock. The only case where this "works" is when StartMessageGeneration runs fast enough to complete before your timer fires (thus freeing the UI thread to do its work) - which is very much possible due to your incorrect use of Task.Delay.
Now let's look on your "hints" with all we know. 1) is simply your bias in measurements. Since you never wait on the Task.Delay in any way, changing the interval does absolutely nothing (with a tiny change in case the delay is zero). 2) of course - that's where your deadlock is. Two pieces of code that depend on a shared resource, while they both require to take posession of another resource. It's a very typical case of a deadlock. Thread 1 is waiting for A to release B, and thread 2 is waiting for B to release A (in this case, A being syncLock and B being the UI thread). When you remove the Send (or replace it with Post), thread 1 no longer has to wait on B, and the deadlock disappears.
There's other things that make writing code like this simpler. There's little point in declaring your own delegate when you can just use Action<string>, for example; using await helps quite a bit when dealing with mixed UI/non-UI code, as well as managing any kind of asynchronous code. You don't need to use event where a simple function will suffice - you can just pass that delegate to a function that needs it if that makes sense, and it may make perfect sense not to allow multiple event handlers to be called. If you decide to keep with the event, at least make sure it conforms to the EventHandler delegate.
To show how your code can be rewritten to be a bit more up-to-date and actually work:
void Main()
{
Application.Run(new LogForm());
}
public static class Log
{
public static async Task GenerateMessagesAsync(Action<string> logEvent,
CancellationToken cancel)
{
for (int i = 0; i < 1000; i++)
{
cancel.ThrowIfCancellationRequested();
logEvent(i.ToString());
await Task.Delay(1000, cancel);
}
}
}
public partial class LogForm : Form
{
private readonly List<string> messages;
private readonly Button btnStart;
private readonly Button btnStop;
private readonly TextBox tbxLog;
private readonly System.Windows.Forms.Timer timer;
public LogForm()
{
messages = new List<string>();
btnStart = new Button { Text = "Start" };
btnStart.Click += btnStart_Click;
Controls.Add(btnStart);
btnStop =
new Button { Text = "Stop", Location = new Point(80, 0), Enabled = false };
Controls.Add(btnStop);
tbxLog = new TextBox { Height = 200, Multiline = true, Dock = DockStyle.Bottom };
Controls.Add(tbxLog);
timer = new System.Windows.Forms.Timer { Interval = 500 };
timer.Tick += TimerProcessor;
timer.Start();
}
private void TimerProcessor(object sender, EventArgs e)
{
foreach (var message in messages)
{
tbxLog.AppendText(message + Environment.NewLine);
}
messages.Clear();
}
private async void btnStart_Click(object sender, EventArgs e)
{
btnStart.Enabled = false;
var cts = new CancellationTokenSource();
EventHandler stopAction = (_, __) => cts.Cancel();
btnStop.Click += stopAction;
btnStop.Enabled = true;
try
{
await Log.GenerateMessagesAsync(message => messages.Add(message), cts.Token);
}
catch (TaskCanceledException)
{
messages.Add("Cancelled.");
}
finally
{
btnStart.Enabled = true;
btnStop.Click -= stopAction;
btnStop.Enabled = false;
}
}
protected override void Dispose(bool disposing)
{
if (disposing)
{
timer.Dispose();
btnStart.Dispose();
btnStop.Dispose();
tbxLog.Dispose();
}
base.Dispose(disposing);
}
}

SynchronizationContext.Send is run synchronously. When you call it, you actually block the UI thread until the operation is complete. But if UI thread is already in lock state, then it just make sense that you are in deadlock.
You can use SynchronizationContext.Post to avoid this.
I just answer on your question, but the truth is that your code need a "little" refactoring..

Calling a method after set amount of time and/or aborting thread issues

So I've got an application that employs a filesystemWatcher and triggers an event just fine. The FSW will trigger a bunch of times pretty close together. I want to create a function that triggers say an hour after the last time the FSW was triggered.
I first tried using a backgroundworker: (All code is shortened for clarity)
namespace Devo
{
public partial class Form1 : Form
{
BackgroundWorker bw = new BackgroundWorker();
private void fileSystemWatcher_Created(object sender, FileSystemEventArgs e)
{
if (bw.IsBusy)
{
bw.CancelAsync(); //this is to, in a way, reset the timer for the delayed method.
}
//do a lot of stuff
bw.RunWorkerAsync();
}
private void backgroundWorker_DoWork(object sender, DoWorkEventArgs e)
{
Stopwatch sw = new Stopwatch();
sw.Start();
while(sw.ElapsedMilliseconds < 180000)
{
if (bw.CancellationPending == true)
{
sw.Stop();
sw.Reset();
e.Cancel = true;
return;
}
}
sw.Stop();
sw.Reset();
DelayedMethod();
}
}
}
This didn't work as the second time bw.RunWorkerAsync() was called it was apparently busy, even though the call to bw.CancelAsync().
My next attempt involved a regular thread as I read somewhere on SO (can't find the link now) that one could not "restart" a backgroundWorker as I am trying to do.
The thread attemp is nearly identical but I thought I'd try in since there might be some constraints within the backgroundWorker that is not present in a regular thread. I thought.
namespace Devo
{
public partial class Form1 : Form
{
Thread PWC_counter_thread = new Thread(PWC_Counter);
private void fileSystemWatcher_Created(object sender, FileSystemEventArgs e)
{
if (PWC_counter_thread.IsAlive)
PWC_counter_thread.Abort();
//do a lot of stuff
PWC_counter_thread.Start();
}
static void PWC_Counter()
{
Thread.Sleep(180000);
DelayedMethod();
}
}
}
But this gave me the same error. On the second call to PWC_counter_thread.Start() is was busy.
I'm assuming that a race condition is not present as the second thread waits for, in this example, 3 minutes, and the initial FSW method takes a good full second to execute, therefore I believe that the call to .Abort() and .CancelAsync() both are done before their respective methods are completed.
Now for the questions:
Is it possible to restart a thread in the fashion I am trying? If so, what am I doing wrong?
Should I delay my method call in another way? If so, tips?
EDIT/UPDATE/SOLUTION
I never got starting and stopping a thread to work as I wanted so I found another solution to my situation.
The situation was that I had a second thread that worked as a sort of timer where it would call a method after a set amount of time. My first thread did some work and upon finishing it would start the second thread. If the first thread got fired up again before the timer-thread had finished it was supposed to kill the thread and restart it.
This proved, for me, to be difficult to get the way I wanted. So I instead took another approach towards my wanted end result. Instead of restarting the thread I simply restarted the stopwatch that my second thread was using as a counter. This gave me the result I wanted. It's probably bad practice but it works.

In your BackgroundWorker example you probably have an issue with racing. CancelAsync(), as its name implies, is an asynchronious call, meaning that BackgroundWorker does not stop working immediately and it might still work when try to restart it. To avoid that, you should subscribe to RunWorkerCompleted event and wait for it to fire before calling bw.RunWorkerAsync(); again. For example:
public Form1()
{
bw = new BackgroundWorker();
bw.RunWorkerCompleted += OnCompleted;
}
private BackgroundWorker bw;
private ManualResetEvent completed = new ManualResetEvent(false);
private void OnCompleted(object sender, RunWorkerCompletedEventArgs e)
{
completed.Set();
}
private void fileSystemWatcher_Created(object sender, FileSystemEventArgs e)
{
if (bw.IsBusy)
{
bw.CancelAsync();
completed.WaitOne();
}
//do a lot of stuff
completed.Reset();
bw.RunWorkerAsync();
}
You have multiple issues with your Thread-based example.
You should never call Thread.Abort(). Instead, you should implement a cancellation mechanism, similar to that of BackgroundWorker. Make a bool field (_isCancelled or something) and check it periodically in thread delegate.
You can not reuse a Thread object. You should always create a new one.

You would be best off encapsulating this in a class, and use a System.Threading.Timer to detect the inactivity.
Here's an example I put together. The idea is that you create an InactivityDetector with the appropriate inactivity threshold (an hour in your case) and a callback method that will be called when that period of inactivity is exceeded.
You have to call InactivityDetector.RegisterActivity() whenever activity is detected (e.g. in your case a file creation is detected).
Once the inactivity callback has been issued, it will not be called again until RegisterActivity() has been called again (this prevents multiple callbacks for the same period of extended inactivity).
Your code would pass DelayedMethod for the inactivity Action delegate.
Note that the callback is on a separate thread!
(Also note that I didn't put in any parameter validation, to keep the code shorter.)
using System;
using System.Threading;
namespace ConsoleApp1
{
sealed class Program
{
void test()
{
using (var inactivityDetector = new InactivityDetector(TimeSpan.FromSeconds(2), inactivityDetected))
{
for (int loop = 0; loop < 3; ++loop)
{
Console.WriteLine("Keeping busy once a second for 5 seconds.");
for (int i = 0; i < 5; ++i)
{
Thread.Sleep(1000);
Console.WriteLine("Registering activity");
inactivityDetector.RegisterActivity();
}
Console.WriteLine("Entering 3 second inactivity");
Thread.Sleep(3000);
inactivityDetector.RegisterActivity();
}
}
}
static void inactivityDetected()
{
Console.WriteLine("Inactivity detected.");
}
static void Main(string[] args)
{
new Program().test();
}
}
public sealed class InactivityDetector: IDisposable
{
public InactivityDetector(TimeSpan inactivityThreshold, Action onInactivity)
{
_inactivityThreshold = inactivityThreshold;
_onInactivity = onInactivity;
_timer = new Timer(timerCallback, null, (int)inactivityThreshold.TotalMilliseconds, -1);
}
public void RegisterActivity()
{
_timer.Change(-1, -1);
_timer.Change((int)_inactivityThreshold.TotalMilliseconds, -1);
}
private void timerCallback(object state)
{
_timer.Change(-1, -1);
_onInactivity();
}
public void Dispose()
{
_timer.Dispose();
}
private readonly TimeSpan _inactivityThreshold;
private readonly Action _onInactivity;
private readonly Timer _timer;
}
}

How to invoke a UI method from another thread

Playing round with Timers.
Context: a winforms with two labels.
I would like to see how System.Timers.Timer works so I've not used the Forms timer.
I understand that the form and myTimer will now be running in different threads.
Is there an easy way to represent the elapsed time on lblValue in the following form?
I've looked here on MSDN but is there an easier way !
Here's the winforms code:
using System.Timers;
namespace Ariport_Parking
{
public partial class AirportParking : Form
{
//instance variables of the form
System.Timers.Timer myTimer;
int ElapsedCounter = 0;
int MaxTime = 5000;
int elapsedTime = 0;
static int tickLength = 100;
public AirportParking()
{
InitializeComponent();
keepingTime();
lblValue.Text = "hello";
}
//method for keeping time
public void keepingTime() {
myTimer = new System.Timers.Timer(tickLength);
myTimer.Elapsed += new ElapsedEventHandler(myTimer_Elapsed);
myTimer.AutoReset = true;
myTimer.Enabled = true;
myTimer.Start();
}
void myTimer_Elapsed(Object myObject,EventArgs myEventArgs){
myTimer.Stop();
ElapsedCounter += 1;
elapsedTime += tickLength;
if (elapsedTime < MaxTime)
{
this.lblElapsedTime.Text = elapsedTime.ToString();
if (ElapsedCounter % 2 == 0)
this.lblValue.Text = "hello world";
else
this.lblValue.Text = "hello";
myTimer.Start();
}
else
{ myTimer.Start(); }
}
}
}

I guess your code is just a test so I won't discuss about what you do with your timer. The problem here is how to do something with an user interface control inside your timer callback.
Most of Control's methods and properties can be accessed only from the UI thread (in reality they can be accessed only from the thread where you created them but this is another story). This is because each thread has to have its own message loop (GetMessage() filters out messages by thread) then to do something with a Control you have to dispatch a message from your thread to the main thread. In .NET it is easy because every Control inherits a couple of methods for this purpose: Invoke/BeginInvoke/EndInvoke. To know if executing thread must call those methods you have the property InvokeRequired. Just change your code with this to make it works:
if (elapsedTime < MaxTime)
{
this.BeginInvoke(new MethodInvoker(delegate
{
this.lblElapsedTime.Text = elapsedTime.ToString();
if (ElapsedCounter % 2 == 0)
this.lblValue.Text = "hello world";
else
this.lblValue.Text = "hello";
}));
}
Please check MSDN for the list of methods you can call from any thread, just as reference you can always call Invalidate, BeginInvoke, EndInvoke, Invoke methods and to read InvokeRequired property. In general this is a common usage pattern (assuming this is an object derived from Control):
void DoStuff() {
// Has been called from a "wrong" thread?
if (InvokeRequired) {
// Dispatch to correct thread, use BeginInvoke if you don't need
// caller thread until operation completes
Invoke(new MethodInvoker(DoStuff));
} else {
// Do things
}
}
Note that current thread will block until UI thread completed method execution. This may be an issue if thread's timing is important (do not forget that UI thread may be busy or hung for a little). If you don't need method's return value you may simply replace Invoke with BeginInvoke, for WinForms you don't even need subsequent call to EndInvoke:
void DoStuff() {
if (InvokeRequired) {
BeginInvoke(new MethodInvoker(DoStuff));
} else {
// Do things
}
}
If you need return value then you have to deal with usual IAsyncResult interface.
How it works?
A GUI Windows application is based on the window procedure with its message loops. If you write an application in plain C you have something like this:
MSG message;
while (GetMessage(&message, NULL, 0, 0))
{
TranslateMessage(&message);
DispatchMessage(&message);
}
With these few lines of code your application wait for a message and then delivers the message to the window procedure. The window procedure is a big switch/case statement where you check the messages (WM_) you know and you process them somehow (you paint the window for WM_PAINT, you quit your application for WM_QUIT and so on).
Now imagine you have a working thread, how can you call your main thread? Simplest way is using this underlying structure to do the trick. I oversimplify the task but these are the steps:
Create a (thread-safe) queue of functions to invoke (some examples here on SO).
Post a custom message to the window procedure. If you make this queue a priority queue then you can even decide priority for these calls (for example a progress notification from a working thread may have a lower priority than an alarm notification).
In the window procedure (inside your switch/case statement) you understand that message then you can peek the function to call from the queue and to invoke it.
Both WPF and WinForms use this method to deliver (dispatch) a message from a thread to the UI thread. Take a look to this article on MSDN for more details about multiple threads and user interface, WinForms hides a lot of these details and you do not have to take care of them but you may take a look to understand how it works under the hood.

Personally when I work in an application that works with threads out of the UI one, I usually write this little snippet:
private void InvokeUI(Action a)
{
this.BeginInvoke(new MethodInvoker(a));
}
When I do an async call in a different thread I can always callback using:
InvokeUI(() => {
Label1.Text = "Super Cool";
});
Simple and clean.

As asked, here is my answer that checks for cross thread calls, synchronises variable updates, doesen't stop and start the timer and doesn't use the timer for counting elapsed time.
EDIT fixed BeginInvoke call. I've done the cross thread invoke using a generic Action, This allows the sender and eventargs to be passed. If these are unused (as they are here) it is more efficient to use MethodInvoker but I suspect the handling would need to be moved into a parameterless method.
public partial class AirportParking : Form
{
private Timer myTimer = new Timer(100);
private int elapsedCounter = 0;
private readonly DateTime startTime = DateTime.Now;
private const string EvenText = "hello";
private const string OddText = "hello world";
public AirportParking()
{
lblValue.Text = EvenText;
myTimer.Elapsed += MyTimerElapsed;
myTimer.AutoReset = true;
myTimer.Enabled = true;
myTimer.Start();
}
private void MyTimerElapsed(object sender,EventArgs myEventArgs)
{
If (lblValue.InvokeRequired)
{
var self = new Action<object, EventArgs>(MyTimerElapsed);
this.BeginInvoke(self, new [] {sender, myEventArgs});
return;
}
lock (this)
{
lblElapsedTime.Text = DateTime.Now.SubTract(startTime).ToString();
elapesedCounter++;
if(elapsedCounter % 2 == 0)
{
lblValue.Text = EvenText;
}
else
{
lblValue.Text = OddText;
}
}
}
}

First, in Windows Forms (and most frameworks), a control can only be accessed (unless documented as "thread safe") by the UI thread.
So this.lblElapsedTime.Text = ... in your callback is plain wrong. Take a look at Control.BeginInvoke.
Second, You should use System.DateTime and System.TimeSpan for your time computations.
Untested:
DateTime startTime = DateTime.Now;
void myTimer_Elapsed(...) {
TimeSpan elapsed = DateTime.Now - startTime;
this.lblElapsedTime.BeginInvoke(delegate() {
this.lblElapsedTime.Text = elapsed.ToString();
});
}

Ended up using the following. It's a combination of the suggestions given:
using System.Timers;
namespace Ariport_Parking
{
public partial class AirportParking : Form
{
//>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
//instance variables of the form
System.Timers.Timer myTimer;
private const string EvenText = "hello";
private const string OddText = "hello world";
static int tickLength = 100;
static int elapsedCounter;
private int MaxTime = 5000;
private TimeSpan elapsedTime;
private readonly DateTime startTime = DateTime.Now;
//<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
public AirportParking()
{
InitializeComponent();
lblValue.Text = EvenText;
keepingTime();
}
//method for keeping time
public void keepingTime() {
using (System.Timers.Timer myTimer = new System.Timers.Timer(tickLength))
{
myTimer.Elapsed += new ElapsedEventHandler(myTimer_Elapsed);
myTimer.AutoReset = true;
myTimer.Enabled = true;
myTimer.Start();
}
}
private void myTimer_Elapsed(Object myObject,EventArgs myEventArgs){
elapsedCounter++;
elapsedTime = DateTime.Now.Subtract(startTime);
if (elapsedTime.TotalMilliseconds < MaxTime)
{
this.BeginInvoke(new MethodInvoker(delegate
{
this.lblElapsedTime.Text = elapsedTime.ToString();
if (elapsedCounter % 2 == 0)
this.lblValue.Text = EvenText;
else
this.lblValue.Text = OddText;
}));
}
else {myTimer.Stop();}
}
}
}

Single threaded timer

I wanted a timer with the following properties:
No matter how many times start is called, only one call back thread is ever running
The time spent in the call back function was ignored with regards to the interval. E.g if the interval is 100ms and the call back takes 4000ms to execute, the callback is called at 100ms, 4100ms etc.
I couldn't see anything available so wrote the following code. Is there a better way to do this?
/**
* Will ensure that only one thread is ever in the callback
*/
public class SingleThreadedTimer : Timer
{
protected static readonly object InstanceLock = new object();
//used to check whether timer has been disposed while in call back
protected bool running = false;
virtual new public void Start()
{
lock (InstanceLock)
{
this.AutoReset = false;
this.Elapsed -= new ElapsedEventHandler(SingleThreadedTimer_Elapsed);
this.Elapsed += new ElapsedEventHandler(SingleThreadedTimer_Elapsed);
this.running = true;
base.Start();
}
}
virtual public void SingleThreadedTimer_Elapsed(object sender, ElapsedEventArgs e)
{
lock (InstanceLock)
{
DoSomethingCool();
//check if stopped while we were waiting for the lock,
//we don't want to restart if this is the case..
if (running)
{
this.Start();
}
}
}
virtual new public void Stop()
{
lock (InstanceLock)
{
running = false;
base.Stop();
}
}
}

Here's a quick example I just knocked up;
using System.Threading;
//...
public class TimerExample
{
private System.Threading.Timer m_objTimer;
private bool m_blnStarted;
private readonly int m_intTickMs = 1000;
private object m_objLockObject = new object();
public TimerExample()
{
//Create your timer object, but don't start anything yet
m_objTimer = new System.Threading.Timer(callback, m_objTimer, Timeout.Infinite, Timeout.Infinite);
}
public void Start()
{
if (!m_blnStarted)
{
lock (m_objLockObject)
{
if (!m_blnStarted) //double check after lock to be thread safe
{
m_blnStarted = true;
//Make it start in 'm_intTickMs' milliseconds,
//but don't auto callback when it's done (Timeout.Infinite)
m_objTimer.Change(m_intTickMs, Timeout.Infinite);
}
}
}
}
public void Stop()
{
lock (m_objLockObject)
{
m_blnStarted = false;
}
}
private void callback(object state)
{
System.Diagnostics.Debug.WriteLine("callback invoked");
//TODO: your code here
Thread.Sleep(4000);
//When your code has finished running, wait 'm_intTickMs' milliseconds
//and call the callback method again,
//but don't auto callback (Timeout.Infinite)
m_objTimer.Change(m_intTickMs, Timeout.Infinite);
}
}

The .NET Framework provides four timers. Two of these are general-purpose multithreaded
timers:
System.Threading.Timer
System.Timers.Timer
The other two are special-purpose single-threaded timers:
System.Windows.Forms.Timer (Windows Forms timer)
System.Windows.Threading.DispatcherTimer (WPF timer)
The last 2 are designed to eliminate thread-safety issues for WPF and Windows Forms applications.
For example, using WebBrowser inside a timer to capture screenshots from webpage needs to be single-threaded and gives an error at runtime if it is on another thread.
The single-thread timers have the following benefits
You can forget about thread safety.
A fresh Tick will never fire until the previous Tick has finished
processing.
You can update user interface elements and controls directly from
Tick event handling code, without calling Control.BeginInvoke or
Dispatcher.BeginIn voke.
and main disadvantage to note
One thread serves all timers—as well as the processing UI events.
Which means that the Tick event handler must execute quickly,
otherwise the user interface becomes unresponsive.
source: most are scraps from C# in a Nutshell book -> Chapter 22 -> Advanced threading -> Timers -> Single-Threaded Timers

For anyone who needs a single thread timer and wants the timer start to tick after task done.
System.Timers.Timer could do the trick without locking or [ThreadStatic]
System.Timers.Timer tmr;
void InitTimer(){
tmr = new System.Timers.Timer();
tmr.Interval = 300;
tmr.AutoReset = false;
tmr.Elapsed += OnElapsed;
}
void OnElapsed( object sender, System.Timers.ElapsedEventArgs e )
{
backgroundWorking();
// let timer start ticking
tmr.Enabled = true;
}
Credit to Alan N
source https://www.codeproject.com/Answers/405715/System-Timers-Timer-single-threaded-usage#answer2
Edit: spacing

Look at the [ThreadStatic] attribute and the .Net 4.0 ThreadLocal generic type. This will probably quickly give you a way to code this without messing with thread locking etc.
You could have a stack inside your time class, and you could implement a Monitor() method that returns a IDisposable, so you can use the timer like so:
using (_threadTimer.Monitor())
{
// do stuff
}
Have the timer-monitor pop the the interval timestamp off the stack during Dispose().
Manually coding all the locking and thread recognition is an option as has been mentioned. However, locking will influence the time used, most likely more than having to initialize an instance per thread using ThreadLocal
If you're interested, I might knock up an example later

Here is a simple PeriodicNonOverlappingTimer class, that provides just the requested features, and nothing more than that. This timer cannot be started and stopped on demand, and neither can have its interval changed. It just invokes the specified action periodically in a non overlapping manner, until the timer is disposed.
/// <summary>
/// Invokes an action on the ThreadPool at specified intervals, ensuring
/// that the invocations will not overlap, until the timer is disposed.
/// </summary>
public class PeriodicNonOverlappingTimer : IDisposable, IAsyncDisposable
{
private readonly System.Threading.Timer _timer;
public PeriodicNonOverlappingTimer(Action periodicAction,
TimeSpan dueTime, TimeSpan period)
{
// Arguments validation omitted
_timer = new(_ =>
{
var stopwatch = Stopwatch.StartNew();
periodicAction();
var nextDueTime = period - stopwatch.Elapsed;
if (nextDueTime < TimeSpan.Zero) nextDueTime = TimeSpan.Zero;
try { _timer.Change(nextDueTime, Timeout.InfiniteTimeSpan); }
catch (ObjectDisposedException) { } // Ignore this exception
});
_timer.Change(dueTime, Timeout.InfiniteTimeSpan);
}
public void Dispose() => _timer.DisposeAsync().AsTask().Wait();
public ValueTask DisposeAsync() => _timer.DisposeAsync();
}
Usage example. Shows how to create a non-overlapping timer that starts immediately, with a period of 10 seconds.
var timer = new PeriodicNonOverlappingTimer(() =>
{
DoSomethingCool();
}, TimeSpan.Zero, TimeSpan.FromSeconds(10));
//...
timer.Dispose(); // Stop the timer once and for all
In case the DoSomethingCool fails, the exception will be thrown on the ThreadPool, causing the process to crash. So you may want to add a try/catch block, and handle all the exceptions that may occur.
The Dispose is a potentially blocking method. If the periodicAction is currently running, the Dispose will block until the last invocation is completed.
If you don't want to wait for this to happen, you can do this instead:
_ = timer.DisposeAsync(); // Stop the timer without waiting it to finish

Non-reentrant C# timer

I'm trying to invoke a method f() every t time, but if the previous invocation of f() has not finished yet, wait until it's finished.
I've read a bit about the available timers but couldn't find any good way of doing what I want, save for manually writing it all. Any help about how to achieve this will be appreciated, though I fear I might not be able to find a simple solution using timers.
To clarify, if t is one second, and f() runs the arbitrary durations I've written below, then:
Step Operation Time taken
1 wait 1s
2 f() 0.6s
3 wait 0.4s (because f already took 0.6 seconds)
4 f() 10s
5 wait 0s (we're late)
6 f() 0.3s
7 wait 0.7s (we can disregard the debt from step 4)
Notice that the nature of this timer is that f() will not need to be safe regarding re-entrance, and a thread pool of size 1 is enough here.

Use a System.Threading.Timer. Initialize it with a period of Timeout.Infinite so it acts like a one-shot timer. When f() completes, call its Change() method to recharge it again.

You could just use a 'global' level var (or more likely, a public property in the same class as f()) which returns true if f() is already running.
So if f() was in a class named TimedEvent, the first thing f() would do is set Running true
That way your timer fires every second, then launches the timed event if it isnt already running
if (!timedEvent.Running) timedEvent.f()
You commented that f() wouldnt repeat immediately if it took longer than the timer interval. Thats a fair point. I would probably include logic like that inside f() so that Running stays true. So it would look something like this:
public void f(int t) // t is interval in seconds
{
this.running = true;
Stopwatch stopWatch = new Stopwatch();
stopWatch.Start();
do
{
stopwatch.Reset();
// Do work here
} while (stopWatch.Elapsed.Seconds > t); // repeat if f() took longer than t
this.running = false;
}

You can use a non-restarting timer, then manually restart the timer after the method finishes.
Note that this will result in timing that is somewhat different from what you're asking for. (There will always be a gap of t time between invocations)
You could solve that by setting the interval to lastTick + t - Now, and running the method immediately if that's <= 0.
Beware of race conditions if you need to stop the timer.

You cannot get a timer to call you at exactly scheduled intervals. All timers do is call you back no sooner than the requested time.
Some timers are better than others (e.g. Windows.Forms.Timer is very erratic and unreliable compared to System.Threading.Timer)
To stop your timer being called re-entrantly, one approach is to Stop the timer while your method is running. (Depending on the type of timer you use, you either stop it and start it again when your handler exits, or with some timers you can request a single callback rather than repeating callbacks, so each execution of your handler simply enqueues the next call).
To keep the timing relatively even between these calls you can record the time since your handler last executed and use that to calculate the delay until the next event is required. e.g. If you want to be called once per second and your timer completed provcessing at 1.02s, then you can set up the next timer callback at a duration of 0.98s to accomodate the fact that you've already "used up" part of the next second during your processing.

A straightforward solution:
private class Worker : IDisposable
{
private readonly TimeSpan _interval;
private WorkerContext _workerContext;
private sealed class WorkerContext
{
private readonly ManualResetEvent _evExit;
private readonly Thread _thread;
private readonly TimeSpan _interval;
public WorkerContext(ParameterizedThreadStart threadProc, TimeSpan interval)
{
_evExit = new ManualResetEvent(false);
_thread = new Thread(threadProc);
_interval = interval;
}
public ManualResetEvent ExitEvent
{
get { return _evExit; }
}
public TimeSpan Interval
{
get { return _interval; }
}
public void Run()
{
_thread.Start(this);
}
public void Stop()
{
_evExit.Set();
}
public void StopAndWait()
{
_evExit.Set();
_thread.Join();
}
}
~Worker()
{
Stop();
}
public Worker(TimeSpan interval)
{
_interval = interval;
}
public TimeSpan Interval
{
get { return _interval; }
}
private void DoWork()
{
/* do your work here */
}
public void Start()
{
var context = new WorkerContext(WorkThreadProc, _interval);
if(Interlocked.CompareExchange<WorkerContext>(ref _workerContext, context, null) == null)
{
context.Run();
}
else
{
context.ExitEvent.Close();
throw new InvalidOperationException("Working alredy.");
}
}
public void Stop()
{
var context = Interlocked.Exchange<WorkerContext>(ref _workerContext, null);
if(context != null)
{
context.Stop();
}
}
private void WorkThreadProc(object p)
{
var context = (WorkerContext)p;
// you can use whatever time-measurement mechanism you want
var sw = new System.Diagnostics.Stopwatch();
int sleep = (int)context.Interval.TotalMilliseconds;
while(true)
{
if(context.ExitEvent.WaitOne(sleep)) break;
sw.Reset();
sw.Start();
DoWork();
sw.Stop();
var time = sw.Elapsed;
if(time < _interval)
sleep = (int)(_interval - time).TotalMilliseconds;
else
sleep = 0;
}
context.ExitEvent.Close();
}
public void Dispose()
{
Stop();
GC.SuppressFinalize(this);
}
}

How about using delegates to method f(), queuing them to a stack, and popping the stack as each delegate completes? You still need the timer, of course.

A simple thread is the easiest way to achieve this. Your still not going to be certain that your called 'precisely' when you want, but it should be close.... Also you can decide if you want to skip calls that should happen or attempt to catch back up... Here is simple helper routine for creating the thread.
public static Thread StartTimer(TimeSpan interval, Func<bool> operation)
{
Thread t = new Thread(new ThreadStart(
delegate()
{
DateTime when = DateTime.Now;
TimeSpan wait = interval;
while (true)
{
Thread.Sleep(wait);
if (!operation())
return;
DateTime dt = DateTime.Now;
when += interval;
while (when < dt)
when += interval;
wait = when - dt;
}
}
));
t.IsBackground = true;
t.Start();
return t;
}

For the benefit of people who land here searching for "re-entrancy": (I know this may be too late for the original question)
If one is not averse to using open source libraries that already provide for such functionality, I have successfully achieved this through an implementation using Quartz.NET
When you create a job and attach a trigger, you can specify what should be done if a previous trigger has not completed executing it's job