ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.html
Revision: 1.22
Committed: Mon Nov 12 09:02:16 2007 UTC (16 years, 6 months ago) by root
Content type: text/html
Branch: MAIN
Changes since 1.21: +5 -4 lines
Log Message:
*** empty log message ***

File Contents

# User Rev Content
1 root 1.1 <?xml version="1.0" encoding="UTF-8"?>
2     <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
3     <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
4     <head>
5     <title>libev</title>
6     <meta name="description" content="Pod documentation for libev" />
7     <meta name="inputfile" content="&lt;standard input&gt;" />
8     <meta name="outputfile" content="&lt;standard output&gt;" />
9 root 1.22 <meta name="created" content="Mon Nov 12 10:02:16 2007" />
10 root 1.1 <meta name="generator" content="Pod::Xhtml 1.57" />
11     <link rel="stylesheet" href="http://res.tst.eu/pod.css"/></head>
12     <body>
13     <div class="pod">
14     <!-- INDEX START -->
15     <h3 id="TOP">Index</h3>
16    
17     <ul><li><a href="#NAME">NAME</a></li>
18     <li><a href="#SYNOPSIS">SYNOPSIS</a></li>
19     <li><a href="#DESCRIPTION">DESCRIPTION</a></li>
20     <li><a href="#FEATURES">FEATURES</a></li>
21     <li><a href="#CONVENTIONS">CONVENTIONS</a></li>
22 root 1.18 <li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li>
23     <li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li>
24 root 1.1 <li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li>
25     <li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a>
26     <ul><li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
27     </ul>
28     </li>
29     <li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
30 root 1.11 <ul><li><a href="#code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</a></li>
31 root 1.10 <li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</a></li>
32 root 1.14 <li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron</a></li>
33 root 1.10 <li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</a></li>
34     <li><a href="#code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</a></li>
35     <li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</a></li>
36 root 1.17 <li><a href="#code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop</a></li>
37 root 1.1 </ul>
38     </li>
39     <li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
40     <li><a href="#AUTHOR">AUTHOR</a>
41     </li>
42     </ul><hr />
43     <!-- INDEX END -->
44    
45     <h1 id="NAME">NAME</h1><p><a href="#TOP" class="toplink">Top</a></p>
46     <div id="NAME_CONTENT">
47     <p>libev - a high performance full-featured event loop written in C</p>
48    
49     </div>
50     <h1 id="SYNOPSIS">SYNOPSIS</h1><p><a href="#TOP" class="toplink">Top</a></p>
51     <div id="SYNOPSIS_CONTENT">
52     <pre> #include &lt;ev.h&gt;
53    
54     </pre>
55    
56     </div>
57     <h1 id="DESCRIPTION">DESCRIPTION</h1><p><a href="#TOP" class="toplink">Top</a></p>
58     <div id="DESCRIPTION_CONTENT">
59     <p>Libev is an event loop: you register interest in certain events (such as a
60     file descriptor being readable or a timeout occuring), and it will manage
61 root 1.4 these event sources and provide your program with events.</p>
62 root 1.1 <p>To do this, it must take more or less complete control over your process
63     (or thread) by executing the <i>event loop</i> handler, and will then
64     communicate events via a callback mechanism.</p>
65     <p>You register interest in certain events by registering so-called <i>event
66     watchers</i>, which are relatively small C structures you initialise with the
67     details of the event, and then hand it over to libev by <i>starting</i> the
68     watcher.</p>
69    
70     </div>
71     <h1 id="FEATURES">FEATURES</h1><p><a href="#TOP" class="toplink">Top</a></p>
72     <div id="FEATURES_CONTENT">
73     <p>Libev supports select, poll, the linux-specific epoll and the bsd-specific
74     kqueue mechanisms for file descriptor events, relative timers, absolute
75     timers with customised rescheduling, signal events, process status change
76     events (related to SIGCHLD), and event watchers dealing with the event
77 root 1.5 loop mechanism itself (idle, prepare and check watchers). It also is quite
78 root 1.7 fast (see this <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing
79     it to libevent for example).</p>
80 root 1.1
81     </div>
82     <h1 id="CONVENTIONS">CONVENTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
83     <div id="CONVENTIONS_CONTENT">
84     <p>Libev is very configurable. In this manual the default configuration
85     will be described, which supports multiple event loops. For more info
86 root 1.7 about various configuration options please have a look at the file
87 root 1.1 <cite>README.embed</cite> in the libev distribution. If libev was configured without
88     support for multiple event loops, then all functions taking an initial
89     argument of name <code>loop</code> (which is always of type <code>struct ev_loop *</code>)
90     will not have this argument.</p>
91    
92     </div>
93 root 1.18 <h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
94     <div id="TIME_REPRESENTATION_CONTENT">
95 root 1.2 <p>Libev represents time as a single floating point number, representing the
96     (fractional) number of seconds since the (POSIX) epoch (somewhere near
97     the beginning of 1970, details are complicated, don't ask). This type is
98 root 1.1 called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
99     to the double type in C.</p>
100 root 1.18
101     </div>
102     <h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
103     <div id="GLOBAL_FUNCTIONS_CONTENT">
104 root 1.21 <p>These functions can be called anytime, even before initialising the
105     library in any way.</p>
106 root 1.1 <dl>
107     <dt>ev_tstamp ev_time ()</dt>
108     <dd>
109     <p>Returns the current time as libev would use it.</p>
110     </dd>
111     <dt>int ev_version_major ()</dt>
112     <dt>int ev_version_minor ()</dt>
113     <dd>
114     <p>You can find out the major and minor version numbers of the library
115     you linked against by calling the functions <code>ev_version_major</code> and
116     <code>ev_version_minor</code>. If you want, you can compare against the global
117     symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
118     version of the library your program was compiled against.</p>
119 root 1.9 <p>Usually, it's a good idea to terminate if the major versions mismatch,
120 root 1.1 as this indicates an incompatible change. Minor versions are usually
121     compatible to older versions, so a larger minor version alone is usually
122     not a problem.</p>
123     </dd>
124     <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt>
125     <dd>
126     <p>Sets the allocation function to use (the prototype is similar to the
127 root 1.7 realloc C function, the semantics are identical). It is used to allocate
128     and free memory (no surprises here). If it returns zero when memory
129     needs to be allocated, the library might abort or take some potentially
130     destructive action. The default is your system realloc function.</p>
131 root 1.1 <p>You could override this function in high-availability programs to, say,
132     free some memory if it cannot allocate memory, to use a special allocator,
133     or even to sleep a while and retry until some memory is available.</p>
134     </dd>
135     <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
136     <dd>
137     <p>Set the callback function to call on a retryable syscall error (such
138     as failed select, poll, epoll_wait). The message is a printable string
139     indicating the system call or subsystem causing the problem. If this
140     callback is set, then libev will expect it to remedy the sitution, no
141 root 1.7 matter what, when it returns. That is, libev will generally retry the
142 root 1.1 requested operation, or, if the condition doesn't go away, do bad stuff
143     (such as abort).</p>
144     </dd>
145     </dl>
146    
147     </div>
148     <h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1><p><a href="#TOP" class="toplink">Top</a></p>
149     <div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
150     <p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
151     types of such loops, the <i>default</i> loop, which supports signals and child
152     events, and dynamically created loops which do not.</p>
153     <p>If you use threads, a common model is to run the default event loop
154 root 1.18 in your main thread (or in a separate thread) and for each thread you
155 root 1.7 create, you also create another event loop. Libev itself does no locking
156     whatsoever, so if you mix calls to the same event loop in different
157     threads, make sure you lock (this is usually a bad idea, though, even if
158 root 1.9 done correctly, because it's hideous and inefficient).</p>
159 root 1.1 <dl>
160     <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
161     <dd>
162     <p>This will initialise the default event loop if it hasn't been initialised
163     yet and return it. If the default loop could not be initialised, returns
164     false. If it already was initialised it simply returns it (and ignores the
165     flags).</p>
166     <p>If you don't know what event loop to use, use the one returned from this
167     function.</p>
168     <p>The flags argument can be used to specify special behaviour or specific
169 root 1.8 backends to use, and is usually specified as 0 (or EVFLAG_AUTO).</p>
170 root 1.1 <p>It supports the following flags:</p>
171     <p>
172     <dl>
173 root 1.10 <dt><code>EVFLAG_AUTO</code></dt>
174 root 1.1 <dd>
175 root 1.9 <p>The default flags value. Use this if you have no clue (it's the right
176 root 1.1 thing, believe me).</p>
177     </dd>
178 root 1.10 <dt><code>EVFLAG_NOENV</code></dt>
179 root 1.1 <dd>
180 root 1.8 <p>If this flag bit is ored into the flag value (or the program runs setuid
181     or setgid) then libev will <i>not</i> look at the environment variable
182     <code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
183     override the flags completely if it is found in the environment. This is
184     useful to try out specific backends to test their performance, or to work
185     around bugs.</p>
186 root 1.1 </dd>
187 root 1.10 <dt><code>EVMETHOD_SELECT</code> (portable select backend)</dt>
188     <dt><code>EVMETHOD_POLL</code> (poll backend, available everywhere except on windows)</dt>
189     <dt><code>EVMETHOD_EPOLL</code> (linux only)</dt>
190     <dt><code>EVMETHOD_KQUEUE</code> (some bsds only)</dt>
191     <dt><code>EVMETHOD_DEVPOLL</code> (solaris 8 only)</dt>
192     <dt><code>EVMETHOD_PORT</code> (solaris 10 only)</dt>
193 root 1.1 <dd>
194     <p>If one or more of these are ored into the flags value, then only these
195     backends will be tried (in the reverse order as given here). If one are
196     specified, any backend will do.</p>
197     </dd>
198     </dl>
199     </p>
200     </dd>
201     <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
202     <dd>
203     <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
204     always distinct from the default loop. Unlike the default loop, it cannot
205     handle signal and child watchers, and attempts to do so will be greeted by
206     undefined behaviour (or a failed assertion if assertions are enabled).</p>
207     </dd>
208     <dt>ev_default_destroy ()</dt>
209     <dd>
210     <p>Destroys the default loop again (frees all memory and kernel state
211     etc.). This stops all registered event watchers (by not touching them in
212 root 1.9 any way whatsoever, although you cannot rely on this :).</p>
213 root 1.1 </dd>
214     <dt>ev_loop_destroy (loop)</dt>
215     <dd>
216     <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
217     earlier call to <code>ev_loop_new</code>.</p>
218     </dd>
219     <dt>ev_default_fork ()</dt>
220     <dd>
221     <p>This function reinitialises the kernel state for backends that have
222     one. Despite the name, you can call it anytime, but it makes most sense
223     after forking, in either the parent or child process (or both, but that
224     again makes little sense).</p>
225     <p>You <i>must</i> call this function after forking if and only if you want to
226     use the event library in both processes. If you just fork+exec, you don't
227     have to call it.</p>
228 root 1.9 <p>The function itself is quite fast and it's usually not a problem to call
229 root 1.1 it just in case after a fork. To make this easy, the function will fit in
230     quite nicely into a call to <code>pthread_atfork</code>:</p>
231     <pre> pthread_atfork (0, 0, ev_default_fork);
232    
233     </pre>
234     </dd>
235     <dt>ev_loop_fork (loop)</dt>
236     <dd>
237     <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
238     <code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
239     after fork, and how you do this is entirely your own problem.</p>
240     </dd>
241     <dt>unsigned int ev_method (loop)</dt>
242     <dd>
243     <p>Returns one of the <code>EVMETHOD_*</code> flags indicating the event backend in
244     use.</p>
245     </dd>
246 root 1.9 <dt>ev_tstamp ev_now (loop)</dt>
247 root 1.1 <dd>
248     <p>Returns the current &quot;event loop time&quot;, which is the time the event loop
249     got events and started processing them. This timestamp does not change
250     as long as callbacks are being processed, and this is also the base time
251     used for relative timers. You can treat it as the timestamp of the event
252     occuring (or more correctly, the mainloop finding out about it).</p>
253     </dd>
254     <dt>ev_loop (loop, int flags)</dt>
255     <dd>
256     <p>Finally, this is it, the event handler. This function usually is called
257     after you initialised all your watchers and you want to start handling
258     events.</p>
259     <p>If the flags argument is specified as 0, it will not return until either
260     no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
261     <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
262     those events and any outstanding ones, but will not block your process in
263 root 1.9 case there are no events and will return after one iteration of the loop.</p>
264 root 1.1 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
265     neccessary) and will handle those and any outstanding ones. It will block
266 root 1.9 your process until at least one new event arrives, and will return after
267     one iteration of the loop.</p>
268 root 1.1 <p>This flags value could be used to implement alternative looping
269     constructs, but the <code>prepare</code> and <code>check</code> watchers provide a better and
270     more generic mechanism.</p>
271     </dd>
272     <dt>ev_unloop (loop, how)</dt>
273     <dd>
274 root 1.9 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
275     has processed all outstanding events). The <code>how</code> argument must be either
276     <code>EVUNLOOP_ONCE</code>, which will make the innermost <code>ev_loop</code> call return, or
277     <code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
278 root 1.1 </dd>
279     <dt>ev_ref (loop)</dt>
280     <dt>ev_unref (loop)</dt>
281     <dd>
282 root 1.9 <p>Ref/unref can be used to add or remove a reference count on the event
283     loop: Every watcher keeps one reference, and as long as the reference
284     count is nonzero, <code>ev_loop</code> will not return on its own. If you have
285     a watcher you never unregister that should not keep <code>ev_loop</code> from
286     returning, ev_unref() after starting, and ev_ref() before stopping it. For
287     example, libev itself uses this for its internal signal pipe: It is not
288     visible to the libev user and should not keep <code>ev_loop</code> from exiting if
289     no event watchers registered by it are active. It is also an excellent
290     way to do this for generic recurring timers or from within third-party
291     libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
292 root 1.1 </dd>
293     </dl>
294    
295     </div>
296     <h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1><p><a href="#TOP" class="toplink">Top</a></p>
297     <div id="ANATOMY_OF_A_WATCHER_CONTENT">
298     <p>A watcher is a structure that you create and register to record your
299     interest in some event. For instance, if you want to wait for STDIN to
300 root 1.10 become readable, you would create an <code>ev_io</code> watcher for that:</p>
301 root 1.1 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
302     {
303     ev_io_stop (w);
304     ev_unloop (loop, EVUNLOOP_ALL);
305     }
306    
307     struct ev_loop *loop = ev_default_loop (0);
308     struct ev_io stdin_watcher;
309     ev_init (&amp;stdin_watcher, my_cb);
310     ev_io_set (&amp;stdin_watcher, STDIN_FILENO, EV_READ);
311     ev_io_start (loop, &amp;stdin_watcher);
312     ev_loop (loop, 0);
313    
314     </pre>
315     <p>As you can see, you are responsible for allocating the memory for your
316     watcher structures (and it is usually a bad idea to do this on the stack,
317     although this can sometimes be quite valid).</p>
318     <p>Each watcher structure must be initialised by a call to <code>ev_init
319     (watcher *, callback)</code>, which expects a callback to be provided. This
320     callback gets invoked each time the event occurs (or, in the case of io
321     watchers, each time the event loop detects that the file descriptor given
322     is readable and/or writable).</p>
323     <p>Each watcher type has its own <code>ev_&lt;type&gt;_set (watcher *, ...)</code> macro
324     with arguments specific to this watcher type. There is also a macro
325     to combine initialisation and setting in one call: <code>ev_&lt;type&gt;_init
326     (watcher *, callback, ...)</code>.</p>
327     <p>To make the watcher actually watch out for events, you have to start it
328     with a watcher-specific start function (<code>ev_&lt;type&gt;_start (loop, watcher
329     *)</code>), and you can stop watching for events at any time by calling the
330     corresponding stop function (<code>ev_&lt;type&gt;_stop (loop, watcher *)</code>.</p>
331     <p>As long as your watcher is active (has been started but not stopped) you
332     must not touch the values stored in it. Most specifically you must never
333     reinitialise it or call its set method.</p>
334 root 1.14 <p>You can check whether an event is active by calling the <code>ev_is_active
335 root 1.4 (watcher *)</code> macro. To see whether an event is outstanding (but the
336 root 1.14 callback for it has not been called yet) you can use the <code>ev_is_pending
337 root 1.1 (watcher *)</code> macro.</p>
338     <p>Each and every callback receives the event loop pointer as first, the
339     registered watcher structure as second, and a bitset of received events as
340     third argument.</p>
341 root 1.14 <p>The received events usually include a single bit per event type received
342 root 1.1 (you can receive multiple events at the same time). The possible bit masks
343     are:</p>
344     <dl>
345 root 1.10 <dt><code>EV_READ</code></dt>
346     <dt><code>EV_WRITE</code></dt>
347 root 1.1 <dd>
348 root 1.10 <p>The file descriptor in the <code>ev_io</code> watcher has become readable and/or
349 root 1.1 writable.</p>
350     </dd>
351 root 1.10 <dt><code>EV_TIMEOUT</code></dt>
352 root 1.1 <dd>
353 root 1.10 <p>The <code>ev_timer</code> watcher has timed out.</p>
354 root 1.1 </dd>
355 root 1.10 <dt><code>EV_PERIODIC</code></dt>
356 root 1.1 <dd>
357 root 1.10 <p>The <code>ev_periodic</code> watcher has timed out.</p>
358 root 1.1 </dd>
359 root 1.10 <dt><code>EV_SIGNAL</code></dt>
360 root 1.1 <dd>
361 root 1.10 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
362 root 1.1 </dd>
363 root 1.10 <dt><code>EV_CHILD</code></dt>
364 root 1.1 <dd>
365 root 1.10 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
366 root 1.1 </dd>
367 root 1.10 <dt><code>EV_IDLE</code></dt>
368 root 1.1 <dd>
369 root 1.10 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
370 root 1.1 </dd>
371 root 1.10 <dt><code>EV_PREPARE</code></dt>
372     <dt><code>EV_CHECK</code></dt>
373 root 1.1 <dd>
374 root 1.10 <p>All <code>ev_prepare</code> watchers are invoked just <i>before</i> <code>ev_loop</code> starts
375     to gather new events, and all <code>ev_check</code> watchers are invoked just after
376 root 1.1 <code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
377     received events. Callbacks of both watcher types can start and stop as
378     many watchers as they want, and all of them will be taken into account
379 root 1.10 (for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
380 root 1.1 <code>ev_loop</code> from blocking).</p>
381     </dd>
382 root 1.10 <dt><code>EV_ERROR</code></dt>
383 root 1.1 <dd>
384     <p>An unspecified error has occured, the watcher has been stopped. This might
385     happen because the watcher could not be properly started because libev
386     ran out of memory, a file descriptor was found to be closed or any other
387     problem. You best act on it by reporting the problem and somehow coping
388     with the watcher being stopped.</p>
389     <p>Libev will usually signal a few &quot;dummy&quot; events together with an error,
390     for example it might indicate that a fd is readable or writable, and if
391     your callbacks is well-written it can just attempt the operation and cope
392     with the error from read() or write(). This will not work in multithreaded
393     programs, though, so beware.</p>
394     </dd>
395     </dl>
396    
397     </div>
398     <h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
399     <div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
400     <p>Each watcher has, by default, a member <code>void *data</code> that you can change
401 root 1.14 and read at any time, libev will completely ignore it. This can be used
402 root 1.1 to associate arbitrary data with your watcher. If you need more data and
403     don't want to allocate memory and store a pointer to it in that data
404     member, you can also &quot;subclass&quot; the watcher type and provide your own
405     data:</p>
406     <pre> struct my_io
407     {
408     struct ev_io io;
409     int otherfd;
410     void *somedata;
411     struct whatever *mostinteresting;
412     }
413    
414     </pre>
415     <p>And since your callback will be called with a pointer to the watcher, you
416     can cast it back to your own type:</p>
417     <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
418     {
419     struct my_io *w = (struct my_io *)w_;
420     ...
421     }
422    
423     </pre>
424     <p>More interesting and less C-conformant ways of catsing your callback type
425     have been omitted....</p>
426    
427    
428    
429    
430    
431     </div>
432     <h1 id="WATCHER_TYPES">WATCHER TYPES</h1><p><a href="#TOP" class="toplink">Top</a></p>
433     <div id="WATCHER_TYPES_CONTENT">
434     <p>This section describes each watcher in detail, but will not repeat
435     information given in the last section.</p>
436    
437     </div>
438 root 1.11 <h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</h2>
439     <div id="code_ev_io_code_is_this_file_descrip-2">
440 root 1.4 <p>I/O watchers check whether a file descriptor is readable or writable
441 root 1.1 in each iteration of the event loop (This behaviour is called
442     level-triggering because you keep receiving events as long as the
443 root 1.14 condition persists. Remember you can stop the watcher if you don't want to
444 root 1.1 act on the event and neither want to receive future events).</p>
445 root 1.8 <p>In general you can register as many read and/or write event watchers oer
446     fd as you want (as long as you don't confuse yourself). Setting all file
447     descriptors to non-blocking mode is also usually a good idea (but not
448     required if you know what you are doing).</p>
449     <p>You have to be careful with dup'ed file descriptors, though. Some backends
450     (the linux epoll backend is a notable example) cannot handle dup'ed file
451     descriptors correctly if you register interest in two or more fds pointing
452     to the same file/socket etc. description.</p>
453     <p>If you must do this, then force the use of a known-to-be-good backend
454     (at the time of this writing, this includes only EVMETHOD_SELECT and
455     EVMETHOD_POLL).</p>
456 root 1.1 <dl>
457     <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
458     <dt>ev_io_set (ev_io *, int fd, int events)</dt>
459     <dd>
460 root 1.10 <p>Configures an <code>ev_io</code> watcher. The fd is the file descriptor to rceeive
461 root 1.1 events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_READ |
462     EV_WRITE</code> to receive the given events.</p>
463     </dd>
464     </dl>
465    
466     </div>
467 root 1.10 <h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</h2>
468     <div id="code_ev_timer_code_relative_and_opti-2">
469 root 1.1 <p>Timer watchers are simple relative timers that generate an event after a
470     given time, and optionally repeating in regular intervals after that.</p>
471     <p>The timers are based on real time, that is, if you register an event that
472     times out after an hour and youreset your system clock to last years
473     time, it will still time out after (roughly) and hour. &quot;Roughly&quot; because
474     detecting time jumps is hard, and soem inaccuracies are unavoidable (the
475     monotonic clock option helps a lot here).</p>
476 root 1.9 <p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
477     time. This is usually the right thing as this timestamp refers to the time
478     of the event triggering whatever timeout you are modifying/starting. If
479     you suspect event processing to be delayed and you *need* to base the timeout
480     ion the current time, use something like this to adjust for this:</p>
481     <pre> ev_timer_set (&amp;timer, after + ev_now () - ev_time (), 0.);
482    
483     </pre>
484 root 1.1 <dl>
485     <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
486     <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
487     <dd>
488     <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
489     <code>0.</code>, then it will automatically be stopped. If it is positive, then the
490     timer will automatically be configured to trigger again <code>repeat</code> seconds
491     later, again, and again, until stopped manually.</p>
492     <p>The timer itself will do a best-effort at avoiding drift, that is, if you
493     configure a timer to trigger every 10 seconds, then it will trigger at
494     exactly 10 second intervals. If, however, your program cannot keep up with
495     the timer (ecause it takes longer than those 10 seconds to do stuff) the
496     timer will not fire more than once per event loop iteration.</p>
497     </dd>
498     <dt>ev_timer_again (loop)</dt>
499     <dd>
500     <p>This will act as if the timer timed out and restart it again if it is
501     repeating. The exact semantics are:</p>
502     <p>If the timer is started but nonrepeating, stop it.</p>
503     <p>If the timer is repeating, either start it if necessary (with the repeat
504     value), or reset the running timer to the repeat value.</p>
505     <p>This sounds a bit complicated, but here is a useful and typical
506     example: Imagine you have a tcp connection and you want a so-called idle
507     timeout, that is, you want to be called when there have been, say, 60
508     seconds of inactivity on the socket. The easiest way to do this is to
509 root 1.10 configure an <code>ev_timer</code> with after=repeat=60 and calling ev_timer_again each
510 root 1.1 time you successfully read or write some data. If you go into an idle
511     state where you do not expect data to travel on the socket, you can stop
512     the timer, and again will automatically restart it if need be.</p>
513     </dd>
514     </dl>
515    
516     </div>
517 root 1.14 <h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron</h2>
518 root 1.10 <div id="code_ev_periodic_code_to_cron_or_not-2">
519 root 1.1 <p>Periodic watchers are also timers of a kind, but they are very versatile
520     (and unfortunately a bit complex).</p>
521 root 1.10 <p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
522 root 1.1 but on wallclock time (absolute time). You can tell a periodic watcher
523     to trigger &quot;at&quot; some specific point in time. For example, if you tell a
524     periodic watcher to trigger in 10 seconds (by specifiying e.g. c&lt;ev_now ()
525     + 10.&gt;) and then reset your system clock to the last year, then it will
526 root 1.10 take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
527 root 1.1 roughly 10 seconds later and of course not if you reset your system time
528     again).</p>
529     <p>They can also be used to implement vastly more complex timers, such as
530     triggering an event on eahc midnight, local time.</p>
531     <dl>
532     <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
533     <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
534     <dd>
535     <p>Lots of arguments, lets sort it out... There are basically three modes of
536     operation, and we will explain them from simplest to complex:</p>
537    
538    
539    
540    
541     <p>
542     <dl>
543     <dt>* absolute timer (interval = reschedule_cb = 0)</dt>
544     <dd>
545     <p>In this configuration the watcher triggers an event at the wallclock time
546     <code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
547     that is, if it is to be run at January 1st 2011 then it will run when the
548     system time reaches or surpasses this time.</p>
549     </dd>
550     <dt>* non-repeating interval timer (interval &gt; 0, reschedule_cb = 0)</dt>
551     <dd>
552     <p>In this mode the watcher will always be scheduled to time out at the next
553     <code>at + N * interval</code> time (for some integer N) and then repeat, regardless
554     of any time jumps.</p>
555     <p>This can be used to create timers that do not drift with respect to system
556     time:</p>
557     <pre> ev_periodic_set (&amp;periodic, 0., 3600., 0);
558    
559     </pre>
560     <p>This doesn't mean there will always be 3600 seconds in between triggers,
561     but only that the the callback will be called when the system time shows a
562 root 1.12 full hour (UTC), or more correctly, when the system time is evenly divisible
563 root 1.1 by 3600.</p>
564     <p>Another way to think about it (for the mathematically inclined) is that
565 root 1.10 <code>ev_periodic</code> will try to run the callback in this mode at the next possible
566 root 1.1 time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
567     </dd>
568     <dt>* manual reschedule mode (reschedule_cb = callback)</dt>
569     <dd>
570     <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
571     ignored. Instead, each time the periodic watcher gets scheduled, the
572     reschedule callback will be called with the watcher as first, and the
573     current time as second argument.</p>
574 root 1.21 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
575     ever, or make any event loop modifications</i>. If you need to stop it,
576     return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
577     starting a prepare watcher).</p>
578 root 1.13 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
579     ev_tstamp now)</code>, e.g.:</p>
580 root 1.1 <pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
581     {
582     return now + 60.;
583     }
584    
585     </pre>
586     <p>It must return the next time to trigger, based on the passed time value
587     (that is, the lowest time value larger than to the second argument). It
588     will usually be called just before the callback will be triggered, but
589     might be called at other times, too.</p>
590 root 1.21 <p>NOTE: <i>This callback must always return a time that is later than the
591 root 1.22 passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
592 root 1.1 <p>This can be used to create very complex timers, such as a timer that
593     triggers on each midnight, local time. To do this, you would calculate the
594 root 1.22 next midnight after <code>now</code> and return the timestamp value for this. How
595     you do this is, again, up to you (but it is not trivial, which is the main
596     reason I omitted it as an example).</p>
597 root 1.1 </dd>
598     </dl>
599     </p>
600     </dd>
601     <dt>ev_periodic_again (loop, ev_periodic *)</dt>
602     <dd>
603     <p>Simply stops and restarts the periodic watcher again. This is only useful
604     when you changed some parameters or the reschedule callback would return
605     a different time than the last time it was called (e.g. in a crond like
606     program when the crontabs have changed).</p>
607     </dd>
608     </dl>
609    
610     </div>
611 root 1.10 <h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</h2>
612     <div id="code_ev_signal_code_signal_me_when_a-2">
613 root 1.1 <p>Signal watchers will trigger an event when the process receives a specific
614     signal one or more times. Even though signals are very asynchronous, libev
615 root 1.9 will try it's best to deliver signals synchronously, i.e. as part of the
616 root 1.1 normal event processing, like any other event.</p>
617 root 1.14 <p>You can configure as many watchers as you like per signal. Only when the
618 root 1.1 first watcher gets started will libev actually register a signal watcher
619     with the kernel (thus it coexists with your own signal handlers as long
620     as you don't register any with libev). Similarly, when the last signal
621     watcher for a signal is stopped libev will reset the signal handler to
622     SIG_DFL (regardless of what it was set to before).</p>
623     <dl>
624     <dt>ev_signal_init (ev_signal *, callback, int signum)</dt>
625     <dt>ev_signal_set (ev_signal *, int signum)</dt>
626     <dd>
627     <p>Configures the watcher to trigger on the given signal number (usually one
628     of the <code>SIGxxx</code> constants).</p>
629     </dd>
630     </dl>
631    
632     </div>
633 root 1.10 <h2 id="code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</h2>
634     <div id="code_ev_child_code_wait_for_pid_stat-2">
635 root 1.1 <p>Child watchers trigger when your process receives a SIGCHLD in response to
636     some child status changes (most typically when a child of yours dies).</p>
637     <dl>
638     <dt>ev_child_init (ev_child *, callback, int pid)</dt>
639     <dt>ev_child_set (ev_child *, int pid)</dt>
640     <dd>
641     <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
642     <i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
643     at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
644 root 1.14 the status word (use the macros from <code>sys/wait.h</code> and see your systems
645     <code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
646     process causing the status change.</p>
647 root 1.1 </dd>
648     </dl>
649    
650     </div>
651 root 1.10 <h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</h2>
652     <div id="code_ev_idle_code_when_you_ve_got_no-2">
653 root 1.14 <p>Idle watchers trigger events when there are no other events are pending
654     (prepare, check and other idle watchers do not count). That is, as long
655     as your process is busy handling sockets or timeouts (or even signals,
656     imagine) it will not be triggered. But when your process is idle all idle
657     watchers are being called again and again, once per event loop iteration -
658     until stopped, that is, or your process receives more events and becomes
659     busy.</p>
660 root 1.1 <p>The most noteworthy effect is that as long as any idle watchers are
661     active, the process will not block when waiting for new events.</p>
662     <p>Apart from keeping your process non-blocking (which is a useful
663     effect on its own sometimes), idle watchers are a good place to do
664     &quot;pseudo-background processing&quot;, or delay processing stuff to after the
665     event loop has handled all outstanding events.</p>
666     <dl>
667     <dt>ev_idle_init (ev_signal *, callback)</dt>
668     <dd>
669     <p>Initialises and configures the idle watcher - it has no parameters of any
670     kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
671     believe me.</p>
672     </dd>
673     </dl>
674    
675     </div>
676 root 1.17 <h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop</h2>
677 root 1.16 <div id="code_ev_prepare_code_and_code_ev_che-2">
678 root 1.14 <p>Prepare and check watchers are usually (but not always) used in tandem:
679     Prepare watchers get invoked before the process blocks and check watchers
680     afterwards.</p>
681 root 1.1 <p>Their main purpose is to integrate other event mechanisms into libev. This
682     could be used, for example, to track variable changes, implement your own
683     watchers, integrate net-snmp or a coroutine library and lots more.</p>
684     <p>This is done by examining in each prepare call which file descriptors need
685 root 1.14 to be watched by the other library, registering <code>ev_io</code> watchers for
686     them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
687     provide just this functionality). Then, in the check watcher you check for
688     any events that occured (by checking the pending status of all watchers
689     and stopping them) and call back into the library. The I/O and timer
690     callbacks will never actually be called (but must be valid neverthelles,
691     because you never know, you know?).</p>
692     <p>As another example, the Perl Coro module uses these hooks to integrate
693 root 1.1 coroutines into libev programs, by yielding to other active coroutines
694     during each prepare and only letting the process block if no coroutines
695 root 1.14 are ready to run (its actually more complicated, it only runs coroutines
696     with priority higher than the event loop and one lower priority once,
697     using idle watchers to keep the event loop from blocking if lower-priority
698     coroutines exist, thus mapping low-priority coroutines to idle/background
699     tasks).</p>
700 root 1.1 <dl>
701     <dt>ev_prepare_init (ev_prepare *, callback)</dt>
702     <dt>ev_check_init (ev_check *, callback)</dt>
703     <dd>
704     <p>Initialises and configures the prepare or check watcher - they have no
705     parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
706 root 1.14 macros, but using them is utterly, utterly and completely pointless.</p>
707 root 1.1 </dd>
708     </dl>
709    
710     </div>
711     <h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
712     <div id="OTHER_FUNCTIONS_CONTENT">
713 root 1.14 <p>There are some other functions of possible interest. Described. Here. Now.</p>
714 root 1.1 <dl>
715     <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
716     <dd>
717     <p>This function combines a simple timer and an I/O watcher, calls your
718     callback on whichever event happens first and automatically stop both
719     watchers. This is useful if you want to wait for a single event on an fd
720     or timeout without havign to allocate/configure/start/stop/free one or
721     more watchers yourself.</p>
722 root 1.14 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
723     is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
724     <code>events</code> set will be craeted and started.</p>
725 root 1.1 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
726 root 1.14 started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
727     repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
728     dubious value.</p>
729     <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
730     passed an events set like normal event callbacks (with a combination of
731     <code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
732     value passed to <code>ev_once</code>:</p>
733 root 1.1 <pre> static void stdin_ready (int revents, void *arg)
734     {
735     if (revents &amp; EV_TIMEOUT)
736 root 1.14 /* doh, nothing entered */;
737 root 1.1 else if (revents &amp; EV_READ)
738 root 1.14 /* stdin might have data for us, joy! */;
739 root 1.1 }
740    
741 root 1.14 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
742 root 1.1
743     </pre>
744     </dd>
745     <dt>ev_feed_event (loop, watcher, int events)</dt>
746     <dd>
747     <p>Feeds the given event set into the event loop, as if the specified event
748 root 1.14 had happened for the specified watcher (which must be a pointer to an
749     initialised but not necessarily started event watcher).</p>
750 root 1.1 </dd>
751     <dt>ev_feed_fd_event (loop, int fd, int revents)</dt>
752     <dd>
753 root 1.14 <p>Feed an event on the given fd, as if a file descriptor backend detected
754     the given events it.</p>
755 root 1.1 </dd>
756     <dt>ev_feed_signal_event (loop, int signum)</dt>
757     <dd>
758     <p>Feed an event as if the given signal occured (loop must be the default loop!).</p>
759     </dd>
760     </dl>
761    
762     </div>
763     <h1 id="AUTHOR">AUTHOR</h1><p><a href="#TOP" class="toplink">Top</a></p>
764     <div id="AUTHOR_CONTENT">
765     <p>Marc Lehmann &lt;libev@schmorp.de&gt;.</p>
766    
767     </div>
768     </div></body>
769     </html>