1 .\" Automatically generated by Pod::Man v1.37, Pod::Parser v1.35
4 .\" ========================================================================
5 .de Sh \" Subsection heading
13 .de Sp \" Vertical space (when we can't use .PP)
17 .de Vb \" Begin verbatim text
22 .de Ve \" End verbatim text
26 .\" Set up some character translations and predefined strings. \*(-- will
27 .\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left
28 .\" double quote, and \*(R" will give a right double quote. | will give a
29 .\" real vertical bar. \*(C+ will give a nicer C++. Capital omega is used to
30 .\" do unbreakable dashes and therefore won't be available. \*(C` and \*(C'
31 .\" expand to `' in nroff, nothing in troff, for use with C<>.
33 .ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p'
37 . if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch
38 . if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch
51 .\" If the F register is turned on, we'll generate index entries on stderr for
52 .\" titles (.TH), headers (.SH), subsections (.Sh), items (.Ip), and index
53 .\" entries marked with X<> in POD. Of course, you'll have to process the
54 .\" output yourself in some meaningful fashion.
57 . tm Index:\\$1\t\\n%\t"\\$2"
63 .\" For nroff, turn off justification. Always turn off hyphenation; it makes
64 .\" way too many mistakes in technical documents.
68 .\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
69 .\" Fear. Run. Save yourself. No user-serviceable parts.
70 . \" fudge factors for nroff and troff
79 . ds #H ((1u-(\\\\n(.fu%2u))*.13m)
85 . \" simple accents for nroff and troff
95 . ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
96 . ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
97 . ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
98 . ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
99 . ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
100 . ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
102 . \" troff and (daisy-wheel) nroff accents
103 .ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
104 .ds 8 \h'\*(#H'\(*b\h'-\*(#H'
105 .ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
106 .ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
107 .ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
108 .ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
109 .ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
110 .ds ae a\h'-(\w'a'u*4/10)'e
111 .ds Ae A\h'-(\w'A'u*4/10)'E
112 . \" corrections for vroff
113 .if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
114 .if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
115 . \" for low resolution devices (crt and lpr)
116 .if \n(.H>23 .if \n(.V>19 \
129 .\" ========================================================================
131 .IX Title ""<STANDARD INPUT>" 1"
132 .TH "<STANDARD INPUT>" 1 "2007-11-23" "perl v5.8.8" "User Contributed Perl Documentation"
134 libev \- a high performance full\-featured event loop written in C
136 .IX Header "SYNOPSIS"
141 .IX Header "DESCRIPTION"
142 Libev is an event loop: you register interest in certain events (such as a
143 file descriptor being readable or a timeout occuring), and it will manage
144 these event sources and provide your program with events.
146 To do this, it must take more or less complete control over your process
147 (or thread) by executing the \fIevent loop\fR handler, and will then
148 communicate events via a callback mechanism.
150 You register interest in certain events by registering so-called \fIevent
151 watchers\fR, which are relatively small C structures you initialise with the
152 details of the event, and then hand it over to libev by \fIstarting\fR the
155 .IX Header "FEATURES"
156 Libev supports select, poll, the linux-specific epoll and the bsd-specific
157 kqueue mechanisms for file descriptor events, relative timers, absolute
158 timers with customised rescheduling, signal events, process status change
159 events (related to \s-1SIGCHLD\s0), and event watchers dealing with the event
160 loop mechanism itself (idle, prepare and check watchers). It also is quite
161 fast (see this benchmark comparing
162 it to libevent for example).
164 .IX Header "CONVENTIONS"
165 Libev is very configurable. In this manual the default configuration
166 will be described, which supports multiple event loops. For more info
167 about various configuration options please have a look at the file
168 \&\fI\s-1README\s0.embed\fR in the libev distribution. If libev was configured without
169 support for multiple event loops, then all functions taking an initial
170 argument of name \f(CW\*(C`loop\*(C'\fR (which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR)
171 will not have this argument.
172 .SH "TIME REPRESENTATION"
173 .IX Header "TIME REPRESENTATION"
174 Libev represents time as a single floating point number, representing the
175 (fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near
176 the beginning of 1970, details are complicated, don't ask). This type is
177 called \f(CW\*(C`ev_tstamp\*(C'\fR, which is what you should use too. It usually aliases
178 to the double type in C.
179 .SH "GLOBAL FUNCTIONS"
180 .IX Header "GLOBAL FUNCTIONS"
181 These functions can be called anytime, even before initialising the
183 .IP "ev_tstamp ev_time ()" 4
184 .IX Item "ev_tstamp ev_time ()"
185 Returns the current time as libev would use it. Please note that the
186 \&\f(CW\*(C`ev_now\*(C'\fR function is usually faster and also often returns the timestamp
187 you actually want to know.
188 .IP "int ev_version_major ()" 4
189 .IX Item "int ev_version_major ()"
191 .IP "int ev_version_minor ()" 4
192 .IX Item "int ev_version_minor ()"
194 You can find out the major and minor version numbers of the library
195 you linked against by calling the functions \f(CW\*(C`ev_version_major\*(C'\fR and
196 \&\f(CW\*(C`ev_version_minor\*(C'\fR. If you want, you can compare against the global
197 symbols \f(CW\*(C`EV_VERSION_MAJOR\*(C'\fR and \f(CW\*(C`EV_VERSION_MINOR\*(C'\fR, which specify the
198 version of the library your program was compiled against.
200 Usually, it's a good idea to terminate if the major versions mismatch,
201 as this indicates an incompatible change. Minor versions are usually
202 compatible to older versions, so a larger minor version alone is usually
204 .IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4
205 .IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))"
206 Sets the allocation function to use (the prototype is similar to the
207 realloc C function, the semantics are identical). It is used to allocate
208 and free memory (no surprises here). If it returns zero when memory
209 needs to be allocated, the library might abort or take some potentially
210 destructive action. The default is your system realloc function.
212 You could override this function in high-availability programs to, say,
213 free some memory if it cannot allocate memory, to use a special allocator,
214 or even to sleep a while and retry until some memory is available.
215 .IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4
216 .IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));"
217 Set the callback function to call on a retryable syscall error (such
218 as failed select, poll, epoll_wait). The message is a printable string
219 indicating the system call or subsystem causing the problem. If this
220 callback is set, then libev will expect it to remedy the sitution, no
221 matter what, when it returns. That is, libev will generally retry the
222 requested operation, or, if the condition doesn't go away, do bad stuff
224 .SH "FUNCTIONS CONTROLLING THE EVENT LOOP"
225 .IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP"
226 An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two
227 types of such loops, the \fIdefault\fR loop, which supports signals and child
228 events, and dynamically created loops which do not.
230 If you use threads, a common model is to run the default event loop
231 in your main thread (or in a separate thread) and for each thread you
232 create, you also create another event loop. Libev itself does no locking
233 whatsoever, so if you mix calls to the same event loop in different
234 threads, make sure you lock (this is usually a bad idea, though, even if
235 done correctly, because it's hideous and inefficient).
236 .IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4
237 .IX Item "struct ev_loop *ev_default_loop (unsigned int flags)"
238 This will initialise the default event loop if it hasn't been initialised
239 yet and return it. If the default loop could not be initialised, returns
240 false. If it already was initialised it simply returns it (and ignores the
243 If you don't know what event loop to use, use the one returned from this
246 The flags argument can be used to specify special behaviour or specific
247 backends to use, and is usually specified as 0 (or \s-1EVFLAG_AUTO\s0).
249 It supports the following flags:
251 .ie n .IP """EVFLAG_AUTO""" 4
252 .el .IP "\f(CWEVFLAG_AUTO\fR" 4
253 .IX Item "EVFLAG_AUTO"
254 The default flags value. Use this if you have no clue (it's the right
256 .ie n .IP """EVFLAG_NOENV""" 4
257 .el .IP "\f(CWEVFLAG_NOENV\fR" 4
258 .IX Item "EVFLAG_NOENV"
259 If this flag bit is ored into the flag value (or the program runs setuid
260 or setgid) then libev will \fInot\fR look at the environment variable
261 \&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will
262 override the flags completely if it is found in the environment. This is
263 useful to try out specific backends to test their performance, or to work
265 .ie n .IP """EVMETHOD_SELECT"" (value 1, portable select backend)" 4
266 .el .IP "\f(CWEVMETHOD_SELECT\fR (value 1, portable select backend)" 4
267 .IX Item "EVMETHOD_SELECT (value 1, portable select backend)"
268 This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as
269 libev tries to roll its own fd_set with no limits on the number of fds,
270 but if that fails, expect a fairly low limit on the number of fds when
271 using this backend. It doesn't scale too well (O(highest_fd)), but its usually
272 the fastest backend for a low number of fds.
273 .ie n .IP """EVMETHOD_POLL"" (value 2, poll backend, available everywhere except on windows)" 4
274 .el .IP "\f(CWEVMETHOD_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4
275 .IX Item "EVMETHOD_POLL (value 2, poll backend, available everywhere except on windows)"
276 And this is your standard \fIpoll\fR\|(2) backend. It's more complicated than
277 select, but handles sparse fds better and has no artificial limit on the
278 number of fds you can use (except it will slow down considerably with a
279 lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
280 .ie n .IP """EVMETHOD_EPOLL"" (value 4, Linux)" 4
281 .el .IP "\f(CWEVMETHOD_EPOLL\fR (value 4, Linux)" 4
282 .IX Item "EVMETHOD_EPOLL (value 4, Linux)"
283 For few fds, this backend is a bit little slower than poll and select,
284 but it scales phenomenally better. While poll and select usually scale like
285 O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
286 either O(1) or O(active_fds).
288 While stopping and starting an I/O watcher in the same iteration will
289 result in some caching, there is still a syscall per such incident
290 (because the fd could point to a different file description now), so its
291 best to avoid that. Also, \fIdup()\fRed file descriptors might not work very
292 well if you register events for both fds.
293 .ie n .IP """EVMETHOD_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4
294 .el .IP "\f(CWEVMETHOD_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4
295 .IX Item "EVMETHOD_KQUEUE (value 8, most BSD clones)"
296 Kqueue deserves special mention, as at the time of this writing, it
297 was broken on all BSDs except NetBSD (usually it doesn't work with
298 anything but sockets and pipes, except on Darwin, where of course its
299 completely useless). For this reason its not being \*(L"autodetected\*(R" unless
300 you explicitly specify the flags (i.e. you don't use \s-1EVFLAG_AUTO\s0).
302 It scales in the same way as the epoll backend, but the interface to the
303 kernel is more efficient (which says nothing about its actual speed, of
304 course). While starting and stopping an I/O watcher does not cause an
305 extra syscall as with epoll, it still adds up to four event changes per
306 incident, so its best to avoid that.
307 .ie n .IP """EVMETHOD_DEVPOLL"" (value 16, Solaris 8)" 4
308 .el .IP "\f(CWEVMETHOD_DEVPOLL\fR (value 16, Solaris 8)" 4
309 .IX Item "EVMETHOD_DEVPOLL (value 16, Solaris 8)"
310 This is not implemented yet (and might never be).
311 .ie n .IP """EVMETHOD_PORT"" (value 32, Solaris 10)" 4
312 .el .IP "\f(CWEVMETHOD_PORT\fR (value 32, Solaris 10)" 4
313 .IX Item "EVMETHOD_PORT (value 32, Solaris 10)"
314 This uses the Solaris 10 port mechanism. As with everything on Solaris,
315 it's really slow, but it still scales very well (O(active_fds)).
316 .ie n .IP """EVMETHOD_ALL""" 4
317 .el .IP "\f(CWEVMETHOD_ALL\fR" 4
318 .IX Item "EVMETHOD_ALL"
319 Try all backends (even potentially broken ones that wouldn't be tried
320 with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as
321 \&\f(CW\*(C`EVMETHOD_ALL & ~EVMETHOD_KQUEUE\*(C'\fR.
325 If one or more of these are ored into the flags value, then only these
326 backends will be tried (in the reverse order as given here). If none are
327 specified, most compiled-in backend will be tried, usually in reverse
328 order of their flag values :)
330 .IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4
331 .IX Item "struct ev_loop *ev_loop_new (unsigned int flags)"
332 Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is
333 always distinct from the default loop. Unlike the default loop, it cannot
334 handle signal and child watchers, and attempts to do so will be greeted by
335 undefined behaviour (or a failed assertion if assertions are enabled).
336 .IP "ev_default_destroy ()" 4
337 .IX Item "ev_default_destroy ()"
338 Destroys the default loop again (frees all memory and kernel state
339 etc.). This stops all registered event watchers (by not touching them in
340 any way whatsoever, although you cannot rely on this :).
341 .IP "ev_loop_destroy (loop)" 4
342 .IX Item "ev_loop_destroy (loop)"
343 Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an
344 earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR.
345 .IP "ev_default_fork ()" 4
346 .IX Item "ev_default_fork ()"
347 This function reinitialises the kernel state for backends that have
348 one. Despite the name, you can call it anytime, but it makes most sense
349 after forking, in either the parent or child process (or both, but that
350 again makes little sense).
352 You \fImust\fR call this function in the child process after forking if and
353 only if you want to use the event library in both processes. If you just
354 fork+exec, you don't have to call it.
356 The function itself is quite fast and it's usually not a problem to call
357 it just in case after a fork. To make this easy, the function will fit in
358 quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR:
361 \& pthread_atfork (0, 0, ev_default_fork);
363 .IP "ev_loop_fork (loop)" 4
364 .IX Item "ev_loop_fork (loop)"
365 Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by
366 \&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop
367 after fork, and how you do this is entirely your own problem.
368 .IP "unsigned int ev_method (loop)" 4
369 .IX Item "unsigned int ev_method (loop)"
370 Returns one of the \f(CW\*(C`EVMETHOD_*\*(C'\fR flags indicating the event backend in
372 .IP "ev_tstamp ev_now (loop)" 4
373 .IX Item "ev_tstamp ev_now (loop)"
374 Returns the current \*(L"event loop time\*(R", which is the time the event loop
375 got events and started processing them. This timestamp does not change
376 as long as callbacks are being processed, and this is also the base time
377 used for relative timers. You can treat it as the timestamp of the event
378 occuring (or more correctly, the mainloop finding out about it).
379 .IP "ev_loop (loop, int flags)" 4
380 .IX Item "ev_loop (loop, int flags)"
381 Finally, this is it, the event handler. This function usually is called
382 after you initialised all your watchers and you want to start handling
385 If the flags argument is specified as 0, it will not return until either
386 no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called.
388 A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle
389 those events and any outstanding ones, but will not block your process in
390 case there are no events and will return after one iteration of the loop.
392 A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if
393 neccessary) and will handle those and any outstanding ones. It will block
394 your process until at least one new event arrives, and will return after
395 one iteration of the loop.
397 This flags value could be used to implement alternative looping
398 constructs, but the \f(CW\*(C`prepare\*(C'\fR and \f(CW\*(C`check\*(C'\fR watchers provide a better and
399 more generic mechanism.
401 Here are the gory details of what ev_loop does:
404 \& 1. If there are no active watchers (reference count is zero), return.
405 \& 2. Queue and immediately call all prepare watchers.
406 \& 3. If we have been forked, recreate the kernel state.
407 \& 4. Update the kernel state with all outstanding changes.
408 \& 5. Update the "event loop time".
409 \& 6. Calculate for how long to block.
410 \& 7. Block the process, waiting for events.
411 \& 8. Update the "event loop time" and do time jump handling.
412 \& 9. Queue all outstanding timers.
413 \& 10. Queue all outstanding periodics.
414 \& 11. If no events are pending now, queue all idle watchers.
415 \& 12. Queue all check watchers.
416 \& 13. Call all queued watchers in reverse order (i.e. check watchers first).
417 \& 14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
418 \& was used, return, otherwise continue with step #1.
420 .IP "ev_unloop (loop, how)" 4
421 .IX Item "ev_unloop (loop, how)"
422 Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it
423 has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either
424 \&\f(CW\*(C`EVUNLOOP_ONE\*(C'\fR, which will make the innermost \f(CW\*(C`ev_loop\*(C'\fR call return, or
425 \&\f(CW\*(C`EVUNLOOP_ALL\*(C'\fR, which will make all nested \f(CW\*(C`ev_loop\*(C'\fR calls return.
426 .IP "ev_ref (loop)" 4
427 .IX Item "ev_ref (loop)"
429 .IP "ev_unref (loop)" 4
430 .IX Item "ev_unref (loop)"
432 Ref/unref can be used to add or remove a reference count on the event
433 loop: Every watcher keeps one reference, and as long as the reference
434 count is nonzero, \f(CW\*(C`ev_loop\*(C'\fR will not return on its own. If you have
435 a watcher you never unregister that should not keep \f(CW\*(C`ev_loop\*(C'\fR from
436 returning, \fIev_unref()\fR after starting, and \fIev_ref()\fR before stopping it. For
437 example, libev itself uses this for its internal signal pipe: It is not
438 visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if
439 no event watchers registered by it are active. It is also an excellent
440 way to do this for generic recurring timers or from within third-party
441 libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR.
442 .SH "ANATOMY OF A WATCHER"
443 .IX Header "ANATOMY OF A WATCHER"
444 A watcher is a structure that you create and register to record your
445 interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to
446 become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that:
449 \& static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
452 \& ev_unloop (loop, EVUNLOOP_ALL);
457 \& struct ev_loop *loop = ev_default_loop (0);
458 \& struct ev_io stdin_watcher;
459 \& ev_init (&stdin_watcher, my_cb);
460 \& ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
461 \& ev_io_start (loop, &stdin_watcher);
462 \& ev_loop (loop, 0);
465 As you can see, you are responsible for allocating the memory for your
466 watcher structures (and it is usually a bad idea to do this on the stack,
467 although this can sometimes be quite valid).
469 Each watcher structure must be initialised by a call to \f(CW\*(C`ev_init
470 (watcher *, callback)\*(C'\fR, which expects a callback to be provided. This
471 callback gets invoked each time the event occurs (or, in the case of io
472 watchers, each time the event loop detects that the file descriptor given
473 is readable and/or writable).
475 Each watcher type has its own \f(CW\*(C`ev_<type>_set (watcher *, ...)\*(C'\fR macro
476 with arguments specific to this watcher type. There is also a macro
477 to combine initialisation and setting in one call: \f(CW\*(C`ev_<type>_init
478 (watcher *, callback, ...)\*(C'\fR.
480 To make the watcher actually watch out for events, you have to start it
481 with a watcher-specific start function (\f(CW\*(C`ev_<type>_start (loop, watcher
482 *)\*(C'\fR), and you can stop watching for events at any time by calling the
483 corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR.
485 As long as your watcher is active (has been started but not stopped) you
486 must not touch the values stored in it. Most specifically you must never
487 reinitialise it or call its set method.
489 You can check whether an event is active by calling the \f(CW\*(C`ev_is_active
490 (watcher *)\*(C'\fR macro. To see whether an event is outstanding (but the
491 callback for it has not been called yet) you can use the \f(CW\*(C`ev_is_pending
492 (watcher *)\*(C'\fR macro.
494 Each and every callback receives the event loop pointer as first, the
495 registered watcher structure as second, and a bitset of received events as
498 The received events usually include a single bit per event type received
499 (you can receive multiple events at the same time). The possible bit masks
501 .ie n .IP """EV_READ""" 4
502 .el .IP "\f(CWEV_READ\fR" 4
505 .ie n .IP """EV_WRITE""" 4
506 .el .IP "\f(CWEV_WRITE\fR" 4
509 The file descriptor in the \f(CW\*(C`ev_io\*(C'\fR watcher has become readable and/or
511 .ie n .IP """EV_TIMEOUT""" 4
512 .el .IP "\f(CWEV_TIMEOUT\fR" 4
513 .IX Item "EV_TIMEOUT"
514 The \f(CW\*(C`ev_timer\*(C'\fR watcher has timed out.
515 .ie n .IP """EV_PERIODIC""" 4
516 .el .IP "\f(CWEV_PERIODIC\fR" 4
517 .IX Item "EV_PERIODIC"
518 The \f(CW\*(C`ev_periodic\*(C'\fR watcher has timed out.
519 .ie n .IP """EV_SIGNAL""" 4
520 .el .IP "\f(CWEV_SIGNAL\fR" 4
522 The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread.
523 .ie n .IP """EV_CHILD""" 4
524 .el .IP "\f(CWEV_CHILD\fR" 4
526 The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change.
527 .ie n .IP """EV_IDLE""" 4
528 .el .IP "\f(CWEV_IDLE\fR" 4
530 The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do.
531 .ie n .IP """EV_PREPARE""" 4
532 .el .IP "\f(CWEV_PREPARE\fR" 4
533 .IX Item "EV_PREPARE"
535 .ie n .IP """EV_CHECK""" 4
536 .el .IP "\f(CWEV_CHECK\fR" 4
539 All \f(CW\*(C`ev_prepare\*(C'\fR watchers are invoked just \fIbefore\fR \f(CW\*(C`ev_loop\*(C'\fR starts
540 to gather new events, and all \f(CW\*(C`ev_check\*(C'\fR watchers are invoked just after
541 \&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any
542 received events. Callbacks of both watcher types can start and stop as
543 many watchers as they want, and all of them will be taken into account
544 (for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep
545 \&\f(CW\*(C`ev_loop\*(C'\fR from blocking).
546 .ie n .IP """EV_ERROR""" 4
547 .el .IP "\f(CWEV_ERROR\fR" 4
549 An unspecified error has occured, the watcher has been stopped. This might
550 happen because the watcher could not be properly started because libev
551 ran out of memory, a file descriptor was found to be closed or any other
552 problem. You best act on it by reporting the problem and somehow coping
553 with the watcher being stopped.
555 Libev will usually signal a few \*(L"dummy\*(R" events together with an error,
556 for example it might indicate that a fd is readable or writable, and if
557 your callbacks is well-written it can just attempt the operation and cope
558 with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded
559 programs, though, so beware.
560 .Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0"
561 .IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER"
562 Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change
563 and read at any time, libev will completely ignore it. This can be used
564 to associate arbitrary data with your watcher. If you need more data and
565 don't want to allocate memory and store a pointer to it in that data
566 member, you can also \*(L"subclass\*(R" the watcher type and provide your own
575 \& struct whatever *mostinteresting;
579 And since your callback will be called with a pointer to the watcher, you
580 can cast it back to your own type:
583 \& static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
585 \& struct my_io *w = (struct my_io *)w_;
590 More interesting and less C\-conformant ways of catsing your callback type
591 have been omitted....
593 .IX Header "WATCHER TYPES"
594 This section describes each watcher in detail, but will not repeat
595 information given in the last section.
596 .ie n .Sh """ev_io"" \- is this file descriptor readable or writable"
597 .el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable"
598 .IX Subsection "ev_io - is this file descriptor readable or writable"
599 I/O watchers check whether a file descriptor is readable or writable
600 in each iteration of the event loop (This behaviour is called
601 level-triggering because you keep receiving events as long as the
602 condition persists. Remember you can stop the watcher if you don't want to
603 act on the event and neither want to receive future events).
605 In general you can register as many read and/or write event watchers per
606 fd as you want (as long as you don't confuse yourself). Setting all file
607 descriptors to non-blocking mode is also usually a good idea (but not
608 required if you know what you are doing).
610 You have to be careful with dup'ed file descriptors, though. Some backends
611 (the linux epoll backend is a notable example) cannot handle dup'ed file
612 descriptors correctly if you register interest in two or more fds pointing
613 to the same underlying file/socket etc. description (that is, they share
614 the same underlying \*(L"file open\*(R").
616 If you must do this, then force the use of a known-to-be-good backend
617 (at the time of this writing, this includes only \s-1EVMETHOD_SELECT\s0 and
618 \&\s-1EVMETHOD_POLL\s0).
619 .IP "ev_io_init (ev_io *, callback, int fd, int events)" 4
620 .IX Item "ev_io_init (ev_io *, callback, int fd, int events)"
622 .IP "ev_io_set (ev_io *, int fd, int events)" 4
623 .IX Item "ev_io_set (ev_io *, int fd, int events)"
625 Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The fd is the file descriptor to rceeive
626 events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or \f(CW\*(C`EV_READ |
627 EV_WRITE\*(C'\fR to receive the given events.
628 .ie n .Sh """ev_timer"" \- relative and optionally recurring timeouts"
629 .el .Sh "\f(CWev_timer\fP \- relative and optionally recurring timeouts"
630 .IX Subsection "ev_timer - relative and optionally recurring timeouts"
631 Timer watchers are simple relative timers that generate an event after a
632 given time, and optionally repeating in regular intervals after that.
634 The timers are based on real time, that is, if you register an event that
635 times out after an hour and you reset your system clock to last years
636 time, it will still time out after (roughly) and hour. \*(L"Roughly\*(R" because
637 detecting time jumps is hard, and some inaccuracies are unavoidable (the
638 monotonic clock option helps a lot here).
640 The relative timeouts are calculated relative to the \f(CW\*(C`ev_now ()\*(C'\fR
641 time. This is usually the right thing as this timestamp refers to the time
642 of the event triggering whatever timeout you are modifying/starting. If
643 you suspect event processing to be delayed and you \fIneed\fR to base the timeout
644 on the current time, use something like this to adjust for this:
647 \& ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
650 The callback is guarenteed to be invoked only when its timeout has passed,
651 but if multiple timers become ready during the same loop iteration then
652 order of execution is undefined.
653 .IP "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 4
654 .IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)"
656 .IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4
657 .IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)"
659 Configure the timer to trigger after \f(CW\*(C`after\*(C'\fR seconds. If \f(CW\*(C`repeat\*(C'\fR is
660 \&\f(CW0.\fR, then it will automatically be stopped. If it is positive, then the
661 timer will automatically be configured to trigger again \f(CW\*(C`repeat\*(C'\fR seconds
662 later, again, and again, until stopped manually.
664 The timer itself will do a best-effort at avoiding drift, that is, if you
665 configure a timer to trigger every 10 seconds, then it will trigger at
666 exactly 10 second intervals. If, however, your program cannot keep up with
667 the timer (because it takes longer than those 10 seconds to do stuff) the
668 timer will not fire more than once per event loop iteration.
669 .IP "ev_timer_again (loop)" 4
670 .IX Item "ev_timer_again (loop)"
671 This will act as if the timer timed out and restart it again if it is
672 repeating. The exact semantics are:
674 If the timer is started but nonrepeating, stop it.
676 If the timer is repeating, either start it if necessary (with the repeat
677 value), or reset the running timer to the repeat value.
679 This sounds a bit complicated, but here is a useful and typical
680 example: Imagine you have a tcp connection and you want a so-called idle
681 timeout, that is, you want to be called when there have been, say, 60
682 seconds of inactivity on the socket. The easiest way to do this is to
683 configure an \f(CW\*(C`ev_timer\*(C'\fR with after=repeat=60 and calling ev_timer_again each
684 time you successfully read or write some data. If you go into an idle
685 state where you do not expect data to travel on the socket, you can stop
686 the timer, and again will automatically restart it if need be.
687 .ie n .Sh """ev_periodic"" \- to cron or not to cron"
688 .el .Sh "\f(CWev_periodic\fP \- to cron or not to cron"
689 .IX Subsection "ev_periodic - to cron or not to cron"
690 Periodic watchers are also timers of a kind, but they are very versatile
691 (and unfortunately a bit complex).
693 Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time)
694 but on wallclock time (absolute time). You can tell a periodic watcher
695 to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a
696 periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()
697 + 10.>) and then reset your system clock to the last year, then it will
698 take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger
699 roughly 10 seconds later and of course not if you reset your system time
702 They can also be used to implement vastly more complex timers, such as
703 triggering an event on eahc midnight, local time.
705 As with timers, the callback is guarenteed to be invoked only when the
706 time (\f(CW\*(C`at\*(C'\fR) has been passed, but if multiple periodic timers become ready
707 during the same loop iteration then order of execution is undefined.
708 .IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4
709 .IX Item "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)"
711 .IP "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 4
712 .IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)"
714 Lots of arguments, lets sort it out... There are basically three modes of
715 operation, and we will explain them from simplest to complex:
717 .IP "* absolute timer (interval = reschedule_cb = 0)" 4
718 .IX Item "absolute timer (interval = reschedule_cb = 0)"
719 In this configuration the watcher triggers an event at the wallclock time
720 \&\f(CW\*(C`at\*(C'\fR and doesn't repeat. It will not adjust when a time jump occurs,
721 that is, if it is to be run at January 1st 2011 then it will run when the
722 system time reaches or surpasses this time.
723 .IP "* non-repeating interval timer (interval > 0, reschedule_cb = 0)" 4
724 .IX Item "non-repeating interval timer (interval > 0, reschedule_cb = 0)"
725 In this mode the watcher will always be scheduled to time out at the next
726 \&\f(CW\*(C`at + N * interval\*(C'\fR time (for some integer N) and then repeat, regardless
729 This can be used to create timers that do not drift with respect to system
733 \& ev_periodic_set (&periodic, 0., 3600., 0);
736 This doesn't mean there will always be 3600 seconds in between triggers,
737 but only that the the callback will be called when the system time shows a
738 full hour (\s-1UTC\s0), or more correctly, when the system time is evenly divisible
741 Another way to think about it (for the mathematically inclined) is that
742 \&\f(CW\*(C`ev_periodic\*(C'\fR will try to run the callback in this mode at the next possible
743 time where \f(CW\*(C`time = at (mod interval)\*(C'\fR, regardless of any time jumps.
744 .IP "* manual reschedule mode (reschedule_cb = callback)" 4
745 .IX Item "manual reschedule mode (reschedule_cb = callback)"
746 In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`at\*(C'\fR are both being
747 ignored. Instead, each time the periodic watcher gets scheduled, the
748 reschedule callback will be called with the watcher as first, and the
749 current time as second argument.
751 \&\s-1NOTE:\s0 \fIThis callback \s-1MUST\s0 \s-1NOT\s0 stop or destroy any periodic watcher,
752 ever, or make any event loop modifications\fR. If you need to stop it,
753 return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop it afterwards (e.g. by
754 starting a prepare watcher).
756 Its prototype is \f(CW\*(C`ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
757 ev_tstamp now)\*(C'\fR, e.g.:
760 \& static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
766 It must return the next time to trigger, based on the passed time value
767 (that is, the lowest time value larger than to the second argument). It
768 will usually be called just before the callback will be triggered, but
769 might be called at other times, too.
771 \&\s-1NOTE:\s0 \fIThis callback must always return a time that is later than the
772 passed \f(CI\*(C`now\*(C'\fI value\fR. Not even \f(CW\*(C`now\*(C'\fR itself will do, it \fImust\fR be larger.
774 This can be used to create very complex timers, such as a timer that
775 triggers on each midnight, local time. To do this, you would calculate the
776 next midnight after \f(CW\*(C`now\*(C'\fR and return the timestamp value for this. How
777 you do this is, again, up to you (but it is not trivial, which is the main
778 reason I omitted it as an example).
782 .IP "ev_periodic_again (loop, ev_periodic *)" 4
783 .IX Item "ev_periodic_again (loop, ev_periodic *)"
784 Simply stops and restarts the periodic watcher again. This is only useful
785 when you changed some parameters or the reschedule callback would return
786 a different time than the last time it was called (e.g. in a crond like
787 program when the crontabs have changed).
788 .ie n .Sh """ev_signal"" \- signal me when a signal gets signalled"
789 .el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled"
790 .IX Subsection "ev_signal - signal me when a signal gets signalled"
791 Signal watchers will trigger an event when the process receives a specific
792 signal one or more times. Even though signals are very asynchronous, libev
793 will try it's best to deliver signals synchronously, i.e. as part of the
794 normal event processing, like any other event.
796 You can configure as many watchers as you like per signal. Only when the
797 first watcher gets started will libev actually register a signal watcher
798 with the kernel (thus it coexists with your own signal handlers as long
799 as you don't register any with libev). Similarly, when the last signal
800 watcher for a signal is stopped libev will reset the signal handler to
801 \&\s-1SIG_DFL\s0 (regardless of what it was set to before).
802 .IP "ev_signal_init (ev_signal *, callback, int signum)" 4
803 .IX Item "ev_signal_init (ev_signal *, callback, int signum)"
805 .IP "ev_signal_set (ev_signal *, int signum)" 4
806 .IX Item "ev_signal_set (ev_signal *, int signum)"
808 Configures the watcher to trigger on the given signal number (usually one
809 of the \f(CW\*(C`SIGxxx\*(C'\fR constants).
810 .ie n .Sh """ev_child"" \- wait for pid status changes"
811 .el .Sh "\f(CWev_child\fP \- wait for pid status changes"
812 .IX Subsection "ev_child - wait for pid status changes"
813 Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to
814 some child status changes (most typically when a child of yours dies).
815 .IP "ev_child_init (ev_child *, callback, int pid)" 4
816 .IX Item "ev_child_init (ev_child *, callback, int pid)"
818 .IP "ev_child_set (ev_child *, int pid)" 4
819 .IX Item "ev_child_set (ev_child *, int pid)"
821 Configures the watcher to wait for status changes of process \f(CW\*(C`pid\*(C'\fR (or
822 \&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look
823 at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see
824 the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems
825 \&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the
826 process causing the status change.
827 .ie n .Sh """ev_idle"" \- when you've got nothing better to do"
828 .el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do"
829 .IX Subsection "ev_idle - when you've got nothing better to do"
830 Idle watchers trigger events when there are no other events are pending
831 (prepare, check and other idle watchers do not count). That is, as long
832 as your process is busy handling sockets or timeouts (or even signals,
833 imagine) it will not be triggered. But when your process is idle all idle
834 watchers are being called again and again, once per event loop iteration \-
835 until stopped, that is, or your process receives more events and becomes
838 The most noteworthy effect is that as long as any idle watchers are
839 active, the process will not block when waiting for new events.
841 Apart from keeping your process non-blocking (which is a useful
842 effect on its own sometimes), idle watchers are a good place to do
843 \&\*(L"pseudo\-background processing\*(R", or delay processing stuff to after the
844 event loop has handled all outstanding events.
845 .IP "ev_idle_init (ev_signal *, callback)" 4
846 .IX Item "ev_idle_init (ev_signal *, callback)"
847 Initialises and configures the idle watcher \- it has no parameters of any
848 kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless,
850 .ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop"
851 .el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop"
852 .IX Subsection "ev_prepare and ev_check - customise your event loop"
853 Prepare and check watchers are usually (but not always) used in tandem:
854 prepare watchers get invoked before the process blocks and check watchers
857 Their main purpose is to integrate other event mechanisms into libev. This
858 could be used, for example, to track variable changes, implement your own
859 watchers, integrate net-snmp or a coroutine library and lots more.
861 This is done by examining in each prepare call which file descriptors need
862 to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for
863 them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries
864 provide just this functionality). Then, in the check watcher you check for
865 any events that occured (by checking the pending status of all watchers
866 and stopping them) and call back into the library. The I/O and timer
867 callbacks will never actually be called (but must be valid nevertheless,
868 because you never know, you know?).
870 As another example, the Perl Coro module uses these hooks to integrate
871 coroutines into libev programs, by yielding to other active coroutines
872 during each prepare and only letting the process block if no coroutines
873 are ready to run (it's actually more complicated: it only runs coroutines
874 with priority higher than or equal to the event loop and one coroutine
875 of lower priority, but only once, using idle watchers to keep the event
876 loop from blocking if lower-priority coroutines are active, thus mapping
877 low-priority coroutines to idle/background tasks).
878 .IP "ev_prepare_init (ev_prepare *, callback)" 4
879 .IX Item "ev_prepare_init (ev_prepare *, callback)"
881 .IP "ev_check_init (ev_check *, callback)" 4
882 .IX Item "ev_check_init (ev_check *, callback)"
884 Initialises and configures the prepare or check watcher \- they have no
885 parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR
886 macros, but using them is utterly, utterly and completely pointless.
887 .SH "OTHER FUNCTIONS"
888 .IX Header "OTHER FUNCTIONS"
889 There are some other functions of possible interest. Described. Here. Now.
890 .IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4
891 .IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)"
892 This function combines a simple timer and an I/O watcher, calls your
893 callback on whichever event happens first and automatically stop both
894 watchers. This is useful if you want to wait for a single event on an fd
895 or timeout without having to allocate/configure/start/stop/free one or
896 more watchers yourself.
898 If \f(CW\*(C`fd\*(C'\fR is less than 0, then no I/O watcher will be started and events
899 is being ignored. Otherwise, an \f(CW\*(C`ev_io\*(C'\fR watcher for the given \f(CW\*(C`fd\*(C'\fR and
900 \&\f(CW\*(C`events\*(C'\fR set will be craeted and started.
902 If \f(CW\*(C`timeout\*(C'\fR is less than 0, then no timeout watcher will be
903 started. Otherwise an \f(CW\*(C`ev_timer\*(C'\fR watcher with after = \f(CW\*(C`timeout\*(C'\fR (and
904 repeat = 0) will be started. While \f(CW0\fR is a valid timeout, it is of
907 The callback has the type \f(CW\*(C`void (*cb)(int revents, void *arg)\*(C'\fR and gets
908 passed an \f(CW\*(C`revents\*(C'\fR set like normal event callbacks (a combination of
909 \&\f(CW\*(C`EV_ERROR\*(C'\fR, \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or \f(CW\*(C`EV_TIMEOUT\*(C'\fR) and the \f(CW\*(C`arg\*(C'\fR
910 value passed to \f(CW\*(C`ev_once\*(C'\fR:
913 \& static void stdin_ready (int revents, void *arg)
915 \& if (revents & EV_TIMEOUT)
916 \& /* doh, nothing entered */;
917 \& else if (revents & EV_READ)
918 \& /* stdin might have data for us, joy! */;
923 \& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
925 .IP "ev_feed_event (loop, watcher, int events)" 4
926 .IX Item "ev_feed_event (loop, watcher, int events)"
927 Feeds the given event set into the event loop, as if the specified event
928 had happened for the specified watcher (which must be a pointer to an
929 initialised but not necessarily started event watcher).
930 .IP "ev_feed_fd_event (loop, int fd, int revents)" 4
931 .IX Item "ev_feed_fd_event (loop, int fd, int revents)"
932 Feed an event on the given fd, as if a file descriptor backend detected
934 .IP "ev_feed_signal_event (loop, int signum)" 4
935 .IX Item "ev_feed_signal_event (loop, int signum)"
936 Feed an event as if the given signal occured (loop must be the default loop!).
937 .SH "LIBEVENT EMULATION"
938 .IX Header "LIBEVENT EMULATION"
939 Libev offers a compatibility emulation layer for libevent. It cannot
940 emulate the internals of libevent, so here are some usage hints:
941 .IP "* Use it by including <event.h>, as usual." 4
942 .IX Item "Use it by including <event.h>, as usual."
944 .IP "* The following members are fully supported: ev_base, ev_callback, ev_arg, ev_fd, ev_res, ev_events." 4
945 .IX Item "The following members are fully supported: ev_base, ev_callback, ev_arg, ev_fd, ev_res, ev_events."
946 .IP "* Avoid using ev_flags and the EVLIST_*\-macros, while it is maintained by libev, it does not work exactly the same way as in libevent (consider it a private \s-1API\s0)." 4
947 .IX Item "Avoid using ev_flags and the EVLIST_*-macros, while it is maintained by libev, it does not work exactly the same way as in libevent (consider it a private API)."
948 .IP "* Priorities are not currently supported. Initialising priorities will fail and all watchers will have the same priority, even though there is an ev_pri field." 4
949 .IX Item "Priorities are not currently supported. Initialising priorities will fail and all watchers will have the same priority, even though there is an ev_pri field."
950 .IP "* Other members are not supported." 4
951 .IX Item "Other members are not supported."
952 .IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4
953 .IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library."
956 .IX Header " SUPPORT"
960 Marc Lehmann <libev@schmorp.de>.
projects / software / libev.git / blob