<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>Control.Concurrent</title><link href="ocean.css" rel="stylesheet" type="text/css" title="Ocean" /><script src="haddock-util.js" type="text/javascript"></script><script type="text/javascript">//<![CDATA[
window.onload = function () {pageLoad();setSynopsis("mini_Control-Concurrent.html");};
//]]>
</script></head><body><div id="package-header"><ul class="links" id="page-menu"><li><a href="index.html">Contents</a></li><li><a href="doc-index.html">Index</a></li></ul><p class="caption">base-4.5.1.0: Basic libraries</p></div><div id="content"><div id="module-header"><table class="info"><tr><th>Portability</th><td>non-portable (concurrency)</td></tr><tr><th>Stability</th><td>experimental</td></tr><tr><th>Maintainer</th><td>libraries@haskell.org</td></tr><tr><th>Safe Haskell</th><td>Trustworthy</td></tr></table><p class="caption">Control.Concurrent</p></div><div id="table-of-contents"><p class="caption">Contents</p><ul><li><a href="#g:1">Concurrent Haskell
</a></li><li><a href="#g:2">Basic concurrency operations
</a><ul><li><a href="#g:3">Threads with affinity
</a></li></ul></li><li><a href="#g:4">Scheduling
</a><ul><li><a href="#g:5">Blocking
</a></li><li><a href="#g:6">Waiting
</a></li></ul></li><li><a href="#g:7">Communication abstractions
</a></li><li><a href="#g:8">Merging of streams
</a></li><li><a href="#g:9">Bound Threads
</a></li><li><a href="#g:10">GHC's implementation of concurrency
</a><ul><li><a href="#g:11">Haskell threads and Operating System threads
</a></li><li><a href="#g:12">Terminating the program
</a></li><li><a href="#g:13">Pre-emption
</a></li></ul></li><li><a href="#g:14">Deprecated functions
</a></li></ul></div><div id="description"><p class="caption">Description</p><div class="doc"><p>A common interface to a collection of useful concurrency
abstractions.
</p></div></div><div id="synopsis"><p id="control.syn" class="caption expander" onclick="toggleSection('syn')">Synopsis</p><ul id="section.syn" class="hide" onclick="toggleSection('syn')"><li class="src short"><span class="keyword">data</span> <a href="#t:ThreadId">ThreadId</a> </li><li class="src short"><a href="#v:myThreadId">myThreadId</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li><li class="src short"><a href="#v:forkIO">forkIO</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li><li class="src short"><a href="#v:forkIOWithUnmask">forkIOWithUnmask</a> :: ((<span class="keyword">forall</span> a. <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a) -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a>) -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li><li class="src short"><a href="#v:killThread">killThread</a> :: <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></li><li class="src short"><a href="#v:throwTo">throwTo</a> :: <a href="Control-Exception-Base.html#t:Exception">Exception</a> e => <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a> -> e -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></li><li class="src short"><a href="#v:forkOn">forkOn</a> :: <a href="Data-Int.html#t:Int">Int</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li><li class="src short"><a href="#v:forkOnWithUnmask">forkOnWithUnmask</a> :: <a href="Data-Int.html#t:Int">Int</a> -> ((<span class="keyword">forall</span> a. <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a) -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a>) -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li><li class="src short"><a href="#v:getNumCapabilities">getNumCapabilities</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="Data-Int.html#t:Int">Int</a></li><li class="src short"><a href="#v:threadCapability">threadCapability</a> :: <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a> -> <a href="System-IO.html#t:IO">IO</a> (<a href="Data-Int.html#t:Int">Int</a>, <a href="Data-Bool.html#t:Bool">Bool</a>)</li><li class="src short"><a href="#v:yield">yield</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></li><li class="src short"><a href="#v:threadDelay">threadDelay</a> :: <a href="Data-Int.html#t:Int">Int</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></li><li class="src short"><a href="#v:threadWaitRead">threadWaitRead</a> :: <a href="System-Posix-Types.html#t:Fd">Fd</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></li><li class="src short"><a href="#v:threadWaitWrite">threadWaitWrite</a> :: <a href="System-Posix-Types.html#t:Fd">Fd</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></li><li class="src short">module <a href="Control-Concurrent-MVar.html">Control.Concurrent.MVar</a></li><li class="src short">module <a href="Control-Concurrent-Chan.html">Control.Concurrent.Chan</a></li><li class="src short">module <a href="Control-Concurrent-QSem.html">Control.Concurrent.QSem</a></li><li class="src short">module <a href="Control-Concurrent-QSemN.html">Control.Concurrent.QSemN</a></li><li class="src short">module <a href="Control-Concurrent-SampleVar.html">Control.Concurrent.SampleVar</a></li><li class="src short"><a href="#v:mergeIO">mergeIO</a> :: [a] -> [a] -> <a href="System-IO.html#t:IO">IO</a> [a]</li><li class="src short"><a href="#v:nmergeIO">nmergeIO</a> :: [[a]] -> <a href="System-IO.html#t:IO">IO</a> [a]</li><li class="src short"><a href="#v:rtsSupportsBoundThreads">rtsSupportsBoundThreads</a> :: <a href="Data-Bool.html#t:Bool">Bool</a></li><li class="src short"><a href="#v:forkOS">forkOS</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li><li class="src short"><a href="#v:isCurrentThreadBound">isCurrentThreadBound</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="Data-Bool.html#t:Bool">Bool</a></li><li class="src short"><a href="#v:runInBoundThread">runInBoundThread</a> :: <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a</li><li class="src short"><a href="#v:runInUnboundThread">runInUnboundThread</a> :: <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a</li><li class="src short"><a href="#v:forkIOUnmasked">forkIOUnmasked</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></li></ul></div><div id="interface"><h1 id="g:1">Concurrent Haskell
</h1><div class="doc"><p>The concurrency extension for Haskell is described in the paper
<em>Concurrent Haskell</em>
<a href="http://www.haskell.org/ghc/docs/papers/concurrent-haskell.ps.gz">http://www.haskell.org/ghc/docs/papers/concurrent-haskell.ps.gz</a>.
</p><p>Concurrency is "lightweight", which means that both thread creation
and context switching overheads are extremely low. Scheduling of
Haskell threads is done internally in the Haskell runtime system, and
doesn't make use of any operating system-supplied thread packages.
</p><p>However, if you want to interact with a foreign library that expects your
program to use the operating system-supplied thread package, you can do so
by using <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> instead of <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>.
</p><p>Haskell threads can communicate via <code><a href="Control-Concurrent-MVar.html#t:MVar">MVar</a></code>s, a kind of synchronised
mutable variable (see <a href="Control-Concurrent-MVar.html">Control.Concurrent.MVar</a>). Several common
concurrency abstractions can be built from <code><a href="Control-Concurrent-MVar.html#t:MVar">MVar</a></code>s, and these are
provided by the <a href="Control-Concurrent.html">Control.Concurrent</a> library.
In GHC, threads may also communicate via exceptions.
</p></div><h1 id="g:2">Basic concurrency operations
</h1><div class="top"><p class="src"><span class="keyword">data</span> <a name="t:ThreadId" class="def">ThreadId</a> </p><div class="doc"><p>A <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> is an abstract type representing a handle to a thread.
<code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> is an instance of <code><a href="Data-Eq.html#t:Eq">Eq</a></code>, <code><a href="Data-Ord.html#t:Ord">Ord</a></code> and <code><a href="Text-Show.html#t:Show">Show</a></code>, where
the <code><a href="Data-Ord.html#t:Ord">Ord</a></code> instance implements an arbitrary total ordering over
<code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code>s. The <code><a href="Text-Show.html#t:Show">Show</a></code> instance lets you convert an arbitrary-valued
<code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> to string form; showing a <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> value is occasionally
useful when debugging or diagnosing the behaviour of a concurrent
program.
</p><p><em>Note</em>: in GHC, if you have a <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code>, you essentially have
a pointer to the thread itself. This means the thread itself can't be
garbage collected until you drop the <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code>.
This misfeature will hopefully be corrected at a later date.
</p><p><em>Note</em>: Hugs does not provide any operations on other threads;
it defines <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> as a synonym for ().
</p></div><div class="subs instances"><p id="control.i:ThreadId" class="caption collapser" onclick="toggleSection('i:ThreadId')">Instances</p><div id="section.i:ThreadId" class="show"><table><tr><td class="src"><a href="Data-Eq.html#t:Eq">Eq</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></td><td class="doc empty"> </td></tr><tr><td class="src"><a href="Data-Ord.html#t:Ord">Ord</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></td><td class="doc empty"> </td></tr><tr><td class="src"><a href="Text-Show.html#t:Show">Show</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></td><td class="doc empty"> </td></tr><tr><td class="src"><a href="Data-Typeable-Internal.html#t:Typeable">Typeable</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></td><td class="doc empty"> </td></tr></table></div></div></div><div class="top"><p class="src"><a name="v:myThreadId" class="def">myThreadId</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>Returns the <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> of the calling thread (GHC only).
</p></div></div><div class="top"><p class="src"><a name="v:forkIO" class="def">forkIO</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>Sparks off a new thread to run the <code><a href="System-IO.html#t:IO">IO</a></code> computation passed as the
first argument, and returns the <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code> of the newly created
thread.
</p><p>The new thread will be a lightweight thread; if you want to use a foreign
library that uses thread-local storage, use <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> instead.
</p><p>GHC note: the new thread inherits the <em>masked</em> state of the parent
(see <code><a href="Control-Exception.html#v:mask">mask</a></code>).
</p><p>The newly created thread has an exception handler that discards the
exceptions <code><a href="Control-Exception-Base.html#t:BlockedIndefinitelyOnMVar">BlockedIndefinitelyOnMVar</a></code>, <code><a href="Control-Exception-Base.html#t:BlockedIndefinitelyOnSTM">BlockedIndefinitelyOnSTM</a></code>, and
<code><a href="Control-Exception-Base.html#v:ThreadKilled">ThreadKilled</a></code>, and passes all other exceptions to the uncaught
exception handler.
</p></div></div><div class="top"><p class="src"><a name="v:forkIOWithUnmask" class="def">forkIOWithUnmask</a> :: ((<span class="keyword">forall</span> a. <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a) -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a>) -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>Like <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>, but the child thread is passed a function that can
be used to unmask asynchronous exceptions. This function is
typically used in the following way
</p><pre> ... mask_ $ forkIOWithUnmask $ \unmask ->
catch (unmask ...) handler
</pre><p>so that the exception handler in the child thread is established
with asynchronous exceptions masked, meanwhile the main body of
the child thread is executed in the unmasked state.
</p><p>Note that the unmask function passed to the child thread should
only be used in that thread; the behaviour is undefined if it is
invoked in a different thread.
</p></div></div><div class="top"><p class="src"><a name="v:killThread" class="def">killThread</a> :: <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></p><div class="doc"><p><code><a href="Control-Concurrent.html#v:killThread">killThread</a></code> raises the <code><a href="Control-Exception-Base.html#v:ThreadKilled">ThreadKilled</a></code> exception in the given
thread (GHC only).
</p><pre> killThread tid = throwTo tid ThreadKilled
</pre></div></div><div class="top"><p class="src"><a name="v:throwTo" class="def">throwTo</a> :: <a href="Control-Exception-Base.html#t:Exception">Exception</a> e => <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a> -> e -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></p><div class="doc"><p><code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> raises an arbitrary exception in the target thread (GHC only).
</p><p><code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> does not return until the exception has been raised in the
target thread.
The calling thread can thus be certain that the target
thread has received the exception. This is a useful property to know
when dealing with race conditions: eg. if there are two threads that
can kill each other, it is guaranteed that only one of the threads
will get to kill the other.
</p><p>Whatever work the target thread was doing when the exception was
raised is not lost: the computation is suspended until required by
another thread.
</p><p>If the target thread is currently making a foreign call, then the
exception will not be raised (and hence <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> will not return)
until the call has completed. This is the case regardless of whether
the call is inside a <code><a href="Control-Exception-Base.html#v:mask">mask</a></code> or not. However, in GHC a foreign call
can be annotated as <code>interruptible</code>, in which case a <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> will
cause the RTS to attempt to cause the call to return; see the GHC
documentation for more details.
</p><p>Important note: the behaviour of <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> differs from that described in
the paper "Asynchronous exceptions in Haskell"
(<a href="http://research.microsoft.com/~simonpj/Papers/asynch-exns.htm">http://research.microsoft.com/~simonpj/Papers/asynch-exns.htm</a>).
In the paper, <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> is non-blocking; but the library implementation adopts
a more synchronous design in which <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> does not return until the exception
is received by the target thread. The trade-off is discussed in Section 9 of the paper.
Like any blocking operation, <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> is therefore interruptible (see Section 5.3 of
the paper). Unlike other interruptible operations, however, <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code>
is <em>always</em> interruptible, even if it does not actually block.
</p><p>There is no guarantee that the exception will be delivered promptly,
although the runtime will endeavour to ensure that arbitrary
delays don't occur. In GHC, an exception can only be raised when a
thread reaches a <em>safe point</em>, where a safe point is where memory
allocation occurs. Some loops do not perform any memory allocation
inside the loop and therefore cannot be interrupted by a <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code>.
</p><p>If the target of <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> is the calling thread, then the behaviour
is the same as <code><a href="Control-Exception.html#v:throwIO">throwIO</a></code>, except that the exception
is thrown as an asynchronous exception. This means that if there is
an enclosing pure computation, which would be the case if the current
IO operation is inside <code><a href="System-IO-Unsafe.html#v:unsafePerformIO">unsafePerformIO</a></code> or <code><a href="System-IO-Unsafe.html#v:unsafeInterleaveIO">unsafeInterleaveIO</a></code>, that
computation is not permanently replaced by the exception, but is
suspended as if it had received an asynchronous exception.
</p><p>Note that if <code><a href="Control-Concurrent.html#v:throwTo">throwTo</a></code> is called with the current thread as the
target, the exception will be thrown even if the thread is currently
inside <code><a href="Control-Exception-Base.html#v:mask">mask</a></code> or <code><a href="Control-Exception-Base.html#v:uninterruptibleMask">uninterruptibleMask</a></code>.
</p></div></div><h2 id="g:3">Threads with affinity
</h2><div class="top"><p class="src"><a name="v:forkOn" class="def">forkOn</a> :: <a href="Data-Int.html#t:Int">Int</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>Like <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>, but lets you specify on which processor the thread
should run. Unlike a <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> thread, a thread created by <code><a href="Control-Concurrent.html#v:forkOn">forkOn</a></code>
will stay on the same processor for its entire lifetime (<code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>
threads can migrate between processors according to the scheduling
policy). <code><a href="Control-Concurrent.html#v:forkOn">forkOn</a></code> is useful for overriding the scheduling policy when
you know in advance how best to distribute the threads.
</p><p>The <code><a href="Data-Int.html#t:Int">Int</a></code> argument specifies a <em>capability number</em> (see
<code><a href="Control-Concurrent.html#v:getNumCapabilities">getNumCapabilities</a></code>). Typically capabilities correspond to physical
processors, but the exact behaviour is implementation-dependent. The
value passed to <code><a href="Control-Concurrent.html#v:forkOn">forkOn</a></code> is interpreted modulo the total number of
capabilities as returned by <code><a href="Control-Concurrent.html#v:getNumCapabilities">getNumCapabilities</a></code>.
</p><p>GHC note: the number of capabilities is specified by the <code>+RTS -N</code>
option when the program is started. Capabilities can be fixed to
actual processor cores with <code>+RTS -qa</code> if the underlying operating
system supports that, although in practice this is usually unnecessary
(and may actually degrade perforamnce in some cases - experimentation
is recommended).
</p></div></div><div class="top"><p class="src"><a name="v:forkOnWithUnmask" class="def">forkOnWithUnmask</a> :: <a href="Data-Int.html#t:Int">Int</a> -> ((<span class="keyword">forall</span> a. <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a) -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a>) -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>Like <code><a href="Control-Concurrent.html#v:forkIOWithUnmask">forkIOWithUnmask</a></code>, but the child thread is pinned to the
given CPU, as with <code><a href="Control-Concurrent.html#v:forkOn">forkOn</a></code>.
</p></div></div><div class="top"><p class="src"><a name="v:getNumCapabilities" class="def">getNumCapabilities</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="Data-Int.html#t:Int">Int</a></p><div class="doc"><p>Returns the number of Haskell threads that can run truly
simultaneously (on separate physical processors) at any given time.
The number passed to <code><a href="Control-Concurrent.html#v:forkOn">forkOn</a></code> is interpreted modulo this
value.
</p><p>An implementation in which Haskell threads are mapped directly to
OS threads might return the number of physical processor cores in
the machine, and <code><a href="Control-Concurrent.html#v:forkOn">forkOn</a></code> would be implemented using the OS's
affinity facilities. An implementation that schedules Haskell
threads onto a smaller number of OS threads (like GHC) would return
the number of such OS threads that can be running simultaneously.
</p><p>GHC notes: this returns the number passed as the argument to the
<code>+RTS -N</code> flag. In current implementations, the value is fixed
when the program starts and never changes, but it is possible that
in the future the number of capabilities might vary at runtime.
</p></div></div><div class="top"><p class="src"><a name="v:threadCapability" class="def">threadCapability</a> :: <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a> -> <a href="System-IO.html#t:IO">IO</a> (<a href="Data-Int.html#t:Int">Int</a>, <a href="Data-Bool.html#t:Bool">Bool</a>)</p><div class="doc"><p>returns the number of the capability on which the thread is currently
running, and a boolean indicating whether the thread is locked to
that capability or not. A thread is locked to a capability if it
was created with <code>forkOn</code>.
</p></div></div><h1 id="g:4">Scheduling
</h1><div class="doc"><p>Scheduling may be either pre-emptive or co-operative,
depending on the implementation of Concurrent Haskell (see below
for information related to specific compilers). In a co-operative
system, context switches only occur when you use one of the
primitives defined in this module. This means that programs such
as:
</p><pre> main = forkIO (write 'a') >> write 'b'
where write c = putChar c >> write c
</pre><p>will print either <code>aaaaaaaaaaaaaa...</code> or <code>bbbbbbbbbbbb...</code>,
instead of some random interleaving of <code>a</code>s and <code>b</code>s. In
practice, cooperative multitasking is sufficient for writing
simple graphical user interfaces.
</p></div><div class="top"><p class="src"><a name="v:yield" class="def">yield</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></p><div class="doc"><p>The <code><a href="Control-Concurrent.html#v:yield">yield</a></code> action allows (forces, in a co-operative multitasking
implementation) a context-switch to any other currently runnable
threads (if any), and is occasionally useful when implementing
concurrency abstractions.
</p></div></div><h2 id="g:5">Blocking
</h2><div class="doc"><p>Different Haskell implementations have different characteristics with
regard to which operations block <em>all</em> threads.
</p><p>Using GHC without the <code>-threaded</code> option, all foreign calls will block
all other Haskell threads in the system, although I/O operations will
not. With the <code>-threaded</code> option, only foreign calls with the <code>unsafe</code>
attribute will block all other threads.
</p><p>Using Hugs, all I/O operations and foreign calls will block all other
Haskell threads.
</p></div><h2 id="g:6">Waiting
</h2><div class="top"><p class="src"><a name="v:threadDelay" class="def">threadDelay</a> :: <a href="Data-Int.html#t:Int">Int</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></p><div class="doc"><p>Suspends the current thread for a given number of microseconds
(GHC only).
</p><p>There is no guarantee that the thread will be rescheduled promptly
when the delay has expired, but the thread will never continue to
run <em>earlier</em> than specified.
</p></div></div><div class="top"><p class="src"><a name="v:threadWaitRead" class="def">threadWaitRead</a> :: <a href="System-Posix-Types.html#t:Fd">Fd</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></p><div class="doc"><p>Block the current thread until data is available to read on the
given file descriptor (GHC only).
</p><p>This will throw an <code><a href="System-IO-Error.html#t:IOError">IOError</a></code> if the file descriptor was closed
while this thread was blocked. To safely close a file descriptor
that has been used with <code><a href="Control-Concurrent.html#v:threadWaitRead">threadWaitRead</a></code>, use
<code><a href="GHC-Conc.html#v:closeFdWith">closeFdWith</a></code>.
</p></div></div><div class="top"><p class="src"><a name="v:threadWaitWrite" class="def">threadWaitWrite</a> :: <a href="System-Posix-Types.html#t:Fd">Fd</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a></p><div class="doc"><p>Block the current thread until data can be written to the
given file descriptor (GHC only).
</p><p>This will throw an <code><a href="System-IO-Error.html#t:IOError">IOError</a></code> if the file descriptor was closed
while this thread was blocked. To safely close a file descriptor
that has been used with <code><a href="Control-Concurrent.html#v:threadWaitWrite">threadWaitWrite</a></code>, use
<code><a href="GHC-Conc.html#v:closeFdWith">closeFdWith</a></code>.
</p></div></div><h1 id="g:7">Communication abstractions
</h1><div class="top"><p class="src">module <a href="Control-Concurrent-MVar.html">Control.Concurrent.MVar</a></p></div><div class="top"><p class="src">module <a href="Control-Concurrent-Chan.html">Control.Concurrent.Chan</a></p></div><div class="top"><p class="src">module <a href="Control-Concurrent-QSem.html">Control.Concurrent.QSem</a></p></div><div class="top"><p class="src">module <a href="Control-Concurrent-QSemN.html">Control.Concurrent.QSemN</a></p></div><div class="top"><p class="src">module <a href="Control-Concurrent-SampleVar.html">Control.Concurrent.SampleVar</a></p></div><h1 id="g:8">Merging of streams
</h1><div class="top"><p class="src"><a name="v:mergeIO" class="def">mergeIO</a> :: [a] -> [a] -> <a href="System-IO.html#t:IO">IO</a> [a]</p></div><div class="top"><p class="src"><a name="v:nmergeIO" class="def">nmergeIO</a> :: [[a]] -> <a href="System-IO.html#t:IO">IO</a> [a]</p></div><div class="doc"><p>The <code><a href="Control-Concurrent.html#v:mergeIO">mergeIO</a></code> and <code><a href="Control-Concurrent.html#v:nmergeIO">nmergeIO</a></code> functions fork one thread for each
input list that concurrently evaluates that list; the results are
merged into a single output list.
</p><p>Note: Hugs does not provide these functions, since they require
preemptive multitasking.
</p></div><h1 id="g:9">Bound Threads
</h1><div class="doc"><p><a name="boundthreads"></a>
</p><p>Support for multiple operating system threads and bound threads as described
below is currently only available in the GHC runtime system if you use the
<em>-threaded</em> option when linking.
</p><p>Other Haskell systems do not currently support multiple operating system threads.
</p><p>A bound thread is a haskell thread that is <em>bound</em> to an operating system
thread. While the bound thread is still scheduled by the Haskell run-time
system, the operating system thread takes care of all the foreign calls made
by the bound thread.
</p><p>To a foreign library, the bound thread will look exactly like an ordinary
operating system thread created using OS functions like <code>pthread_create</code>
or <code>CreateThread</code>.
</p><p>Bound threads can be created using the <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> function below. All foreign
exported functions are run in a bound thread (bound to the OS thread that
called the function). Also, the <code>main</code> action of every Haskell program is
run in a bound thread.
</p><p>Why do we need this? Because if a foreign library is called from a thread
created using <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>, it won't have access to any <em>thread-local state</em> -
state variables that have specific values for each OS thread
(see POSIX's <code>pthread_key_create</code> or Win32's <code>TlsAlloc</code>). Therefore, some
libraries (OpenGL, for example) will not work from a thread created using
<code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>. They work fine in threads created using <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> or when called
from <code>main</code> or from a <code>foreign export</code>.
</p><p>In terms of performance, <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> (aka bound) threads are much more
expensive than <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> (aka unbound) threads, because a <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code>
thread is tied to a particular OS thread, whereas a <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> thread
can be run by any OS thread. Context-switching between a <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code>
thread and a <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> thread is many times more expensive than between
two <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> threads.
</p><p>Note in particular that the main program thread (the thread running
<code>Main.main</code>) is always a bound thread, so for good concurrency
performance you should ensure that the main thread is not doing
repeated communication with other threads in the system. Typically
this means forking subthreads to do the work using <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>, and
waiting for the results in the main thread.
</p></div><div class="top"><p class="src"><a name="v:rtsSupportsBoundThreads" class="def">rtsSupportsBoundThreads</a> :: <a href="Data-Bool.html#t:Bool">Bool</a></p><div class="doc"><p><code><a href="Data-Bool.html#v:True">True</a></code> if bound threads are supported.
If <code>rtsSupportsBoundThreads</code> is <code><a href="Data-Bool.html#v:False">False</a></code>, <code><a href="Control-Concurrent.html#v:isCurrentThreadBound">isCurrentThreadBound</a></code>
will always return <code><a href="Data-Bool.html#v:False">False</a></code> and both <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> and <code><a href="Control-Concurrent.html#v:runInBoundThread">runInBoundThread</a></code> will
fail.
</p></div></div><div class="top"><p class="src"><a name="v:forkOS" class="def">forkOS</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>Like <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>, this sparks off a new thread to run the <code><a href="System-IO.html#t:IO">IO</a></code>
computation passed as the first argument, and returns the <code><a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></code>
of the newly created thread.
</p><p>However, <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> creates a <em>bound</em> thread, which is necessary if you
need to call foreign (non-Haskell) libraries that make use of
thread-local state, such as OpenGL (see <a href="Control-Concurrent.html#boundthreads">Control.Concurrent</a>).
</p><p>Using <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> instead of <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> makes no difference at all to the
scheduling behaviour of the Haskell runtime system. It is a common
misconception that you need to use <code><a href="Control-Concurrent.html#v:forkOS">forkOS</a></code> instead of <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> to
avoid blocking all the Haskell threads when making a foreign call;
this isn't the case. To allow foreign calls to be made without
blocking all the Haskell threads (with GHC), it is only necessary to
use the <code>-threaded</code> option when linking your program, and to make sure
the foreign import is not marked <code>unsafe</code>.
</p></div></div><div class="top"><p class="src"><a name="v:isCurrentThreadBound" class="def">isCurrentThreadBound</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="Data-Bool.html#t:Bool">Bool</a></p><div class="doc"><p>Returns <code><a href="Data-Bool.html#v:True">True</a></code> if the calling thread is <em>bound</em>, that is, if it is
safe to use foreign libraries that rely on thread-local state from the
calling thread.
</p></div></div><div class="top"><p class="src"><a name="v:runInBoundThread" class="def">runInBoundThread</a> :: <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a</p><div class="doc"><p>Run the <code><a href="System-IO.html#t:IO">IO</a></code> computation passed as the first argument. If the calling thread
is not <em>bound</em>, a bound thread is created temporarily. <code>runInBoundThread</code>
doesn't finish until the <code><a href="System-IO.html#t:IO">IO</a></code> computation finishes.
</p><p>You can wrap a series of foreign function calls that rely on thread-local state
with <code>runInBoundThread</code> so that you can use them without knowing whether the
current thread is <em>bound</em>.
</p></div></div><div class="top"><p class="src"><a name="v:runInUnboundThread" class="def">runInUnboundThread</a> :: <a href="System-IO.html#t:IO">IO</a> a -> <a href="System-IO.html#t:IO">IO</a> a</p><div class="doc"><p>Run the <code><a href="System-IO.html#t:IO">IO</a></code> computation passed as the first argument. If the calling thread
is <em>bound</em>, an unbound thread is created temporarily using <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code>.
<code>runInBoundThread</code> doesn't finish until the <code><a href="System-IO.html#t:IO">IO</a></code> computation finishes.
</p><p>Use this function <em>only</em> in the rare case that you have actually observed a
performance loss due to the use of bound threads. A program that
doesn't need it's main thread to be bound and makes <em>heavy</em> use of concurrency
(e.g. a web server), might want to wrap it's <code>main</code> action in
<code>runInUnboundThread</code>.
</p><p>Note that exceptions which are thrown to the current thread are thrown in turn
to the thread that is executing the given computation. This ensures there's
always a way of killing the forked thread.
</p></div></div><h1 id="g:10">GHC's implementation of concurrency
</h1><div class="doc"><p>This section describes features specific to GHC's
implementation of Concurrent Haskell.
</p></div><h2 id="g:11">Haskell threads and Operating System threads
</h2><div class="doc"><p><a name="osthreads"></a> In GHC, threads created by <code><a href="Control-Concurrent.html#v:forkIO">forkIO</a></code> are lightweight threads, and
are managed entirely by the GHC runtime. Typically Haskell
threads are an order of magnitude or two more efficient (in
terms of both time and space) than operating system threads.
</p><p>The downside of having lightweight threads is that only one can
run at a time, so if one thread blocks in a foreign call, for
example, the other threads cannot continue. The GHC runtime
works around this by making use of full OS threads where
necessary. When the program is built with the <code>-threaded</code>
option (to link against the multithreaded version of the
runtime), a thread making a <code>safe</code> foreign call will not block
the other threads in the system; another OS thread will take
over running Haskell threads until the original call returns.
The runtime maintains a pool of these <em>worker</em> threads so that
multiple Haskell threads can be involved in external calls
simultaneously.
</p><p>The <a href="System-IO.html">System.IO</a> library manages multiplexing in its own way. On
Windows systems it uses <code>safe</code> foreign calls to ensure that
threads doing I/O operations don't block the whole runtime,
whereas on Unix systems all the currently blocked I/O requests
are managed by a single thread (the <em>IO manager thread</em>) using
a mechanism such as <code>epoll</code> or <code>kqueue</code>, depending on what is
provided by the host operating system.
</p><p>The runtime will run a Haskell thread using any of the available
worker OS threads. If you need control over which particular OS
thread is used to run a given Haskell thread, perhaps because
you need to call a foreign library that uses OS-thread-local
state, then you need bound threads (see <a href="Control-Concurrent.html#boundthreads">Control.Concurrent</a>).
</p><p>If you don't use the <code>-threaded</code> option, then the runtime does
not make use of multiple OS threads. Foreign calls will block
all other running Haskell threads until the call returns. The
<a href="System-IO.html">System.IO</a> library still does multiplexing, so there can be multiple
threads doing I/O, and this is handled internally by the runtime using
<code>select</code>.
</p></div><h2 id="g:12">Terminating the program
</h2><div class="doc"><p>In a standalone GHC program, only the main thread is
required to terminate in order for the process to terminate.
Thus all other forked threads will simply terminate at the same
time as the main thread (the terminology for this kind of
behaviour is "daemonic threads").
</p><p>If you want the program to wait for child threads to
finish before exiting, you need to program this yourself. A
simple mechanism is to have each child thread write to an
<code><a href="Control-Concurrent-MVar.html#t:MVar">MVar</a></code> when it completes, and have the main
thread wait on all the <code><a href="Control-Concurrent-MVar.html#t:MVar">MVar</a></code>s before
exiting:
</p><pre> myForkIO :: IO () -> IO (MVar ())
myForkIO io = do
mvar <- newEmptyMVar
forkIO (io `finally` putMVar mvar ())
return mvar
</pre><p>Note that we use <code><a href="Control-Exception-Base.html#v:finally">finally</a></code> from the
<a href="Control-Exception.html">Control.Exception</a> module to make sure that the
<code><a href="Control-Concurrent-MVar.html#t:MVar">MVar</a></code> is written to even if the thread dies or
is killed for some reason.
</p><p>A better method is to keep a global list of all child
threads which we should wait for at the end of the program:
</p><pre> children :: MVar [MVar ()]
children = unsafePerformIO (newMVar [])
waitForChildren :: IO ()
waitForChildren = do
cs <- takeMVar children
case cs of
[] -> return ()
m:ms -> do
putMVar children ms
takeMVar m
waitForChildren
forkChild :: IO () -> IO ThreadId
forkChild io = do
mvar <- newEmptyMVar
childs <- takeMVar children
putMVar children (mvar:childs)
forkIO (io `finally` putMVar mvar ())
main =
later waitForChildren $
...
</pre><p>The main thread principle also applies to calls to Haskell from
outside, using <code>foreign export</code>. When the <code>foreign export</code>ed
function is invoked, it starts a new main thread, and it returns
when this main thread terminates. If the call causes new
threads to be forked, they may remain in the system after the
<code>foreign export</code>ed function has returned.
</p></div><h2 id="g:13">Pre-emption
</h2><div class="doc"><p>GHC implements pre-emptive multitasking: the execution of
threads are interleaved in a random fashion. More specifically,
a thread may be pre-empted whenever it allocates some memory,
which unfortunately means that tight loops which do no
allocation tend to lock out other threads (this only seems to
happen with pathological benchmark-style code, however).
</p><p>The rescheduling timer runs on a 20ms granularity by
default, but this may be altered using the
<code>-i<n></code> RTS option. After a rescheduling
"tick" the running thread is pre-empted as soon as
possible.
</p><p>One final note: the
<code>aaaa</code> <code>bbbb</code> example may not
work too well on GHC (see Scheduling, above), due
to the locking on a <code><a href="System-IO.html#t:Handle">Handle</a></code>. Only one thread
may hold the lock on a <code><a href="System-IO.html#t:Handle">Handle</a></code> at any one
time, so if a reschedule happens while a thread is holding the
lock, the other thread won't be able to run. The upshot is that
the switch from <code>aaaa</code> to
<code>bbbbb</code> happens infrequently. It can be
improved by lowering the reschedule tick period. We also have a
patch that causes a reschedule whenever a thread waiting on a
lock is woken up, but haven't found it to be useful for anything
other than this example :-)
</p></div><h1 id="g:14">Deprecated functions
</h1><div class="top"><p class="src"><a name="v:forkIOUnmasked" class="def">forkIOUnmasked</a> :: <a href="System-IO.html#t:IO">IO</a> <a href="../ghc-prim-0.2.0.0/GHC-Tuple.html#t:-40--41-">()</a> -> <a href="System-IO.html#t:IO">IO</a> <a href="Control-Concurrent.html#t:ThreadId">ThreadId</a></p><div class="doc"><p>This function is deprecated; use <code>forkIOWIthUnmask</code> instead
</p></div></div></div></div><div id="footer"><p>Produced by <a href="http://www.haskell.org/haddock/">Haddock</a> version 2.11.0</p></div></body></html>
distrib
>
Mageia
>
5
>
x86_64
>
media
>
core-release
>
by-pkgid
>
956c458aa5fe9afc4d2c00cb7b491287
>
files
>
2220
