ViewVC Help
View File | Revision Log | Show Annotations | Download File
Revision: 1.437
Committed: Sun Oct 11 15:55:48 2015 UTC (8 years, 9 months ago) by root
Branch: MAIN
CVS Tags: EV-rel-4_22, rel-4_22
Changes since 1.436: +1 -1 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 =encoding utf-8
3 =head1 NAME
5 libev - a high performance full-featured event loop written in C
7 =head1 SYNOPSIS
9 #include <ev.h>
13 // a single header file is required
14 #include <ev.h>
16 #include <stdio.h> // for puts
18 // every watcher type has its own typedef'd struct
19 // with the name ev_TYPE
20 ev_io stdin_watcher;
21 ev_timer timeout_watcher;
23 // all watcher callbacks have a similar signature
24 // this callback is called when data is readable on stdin
25 static void
26 stdin_cb (EV_P_ ev_io *w, int revents)
27 {
28 puts ("stdin ready");
29 // for one-shot events, one must manually stop the watcher
30 // with its corresponding stop function.
31 ev_io_stop (EV_A_ w);
33 // this causes all nested ev_run's to stop iterating
34 ev_break (EV_A_ EVBREAK_ALL);
35 }
37 // another callback, this time for a time-out
38 static void
39 timeout_cb (EV_P_ ev_timer *w, int revents)
40 {
41 puts ("timeout");
42 // this causes the innermost ev_run to stop iterating
43 ev_break (EV_A_ EVBREAK_ONE);
44 }
46 int
47 main (void)
48 {
49 // use the default event loop unless you have special needs
50 struct ev_loop *loop = EV_DEFAULT;
52 // initialise an io watcher, then start it
53 // this one will watch for stdin to become readable
54 ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
55 ev_io_start (loop, &stdin_watcher);
57 // initialise a timer watcher, then start it
58 // simple non-repeating 5.5 second timeout
59 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
60 ev_timer_start (loop, &timeout_watcher);
62 // now wait for events to arrive
63 ev_run (loop, 0);
65 // break was called, so exit
66 return 0;
67 }
71 This document documents the libev software package.
73 The newest version of this document is also available as an html-formatted
74 web page you might find easier to navigate when reading it for the first
75 time: L<>.
77 While this document tries to be as complete as possible in documenting
78 libev, its usage and the rationale behind its design, it is not a tutorial
79 on event-based programming, nor will it introduce event-based programming
80 with libev.
82 Familiarity with event based programming techniques in general is assumed
83 throughout this document.
87 This manual tries to be very detailed, but unfortunately, this also makes
88 it very long. If you just want to know the basics of libev, I suggest
89 reading L</ANATOMY OF A WATCHER>, then the L</EXAMPLE PROGRAM> above and
90 look up the missing functions in L</GLOBAL FUNCTIONS> and the C<ev_io> and
91 C<ev_timer> sections in L</WATCHER TYPES>.
93 =head1 ABOUT LIBEV
95 Libev is an event loop: you register interest in certain events (such as a
96 file descriptor being readable or a timeout occurring), and it will manage
97 these event sources and provide your program with events.
99 To do this, it must take more or less complete control over your process
100 (or thread) by executing the I<event loop> handler, and will then
101 communicate events via a callback mechanism.
103 You register interest in certain events by registering so-called I<event
104 watchers>, which are relatively small C structures you initialise with the
105 details of the event, and then hand it over to libev by I<starting> the
106 watcher.
108 =head2 FEATURES
110 Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
111 BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
112 for file descriptor events (C<ev_io>), the Linux C<inotify> interface
113 (for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
114 inter-thread wakeup (C<ev_async>)/signal handling (C<ev_signal>)) relative
115 timers (C<ev_timer>), absolute timers with customised rescheduling
116 (C<ev_periodic>), synchronous signals (C<ev_signal>), process status
117 change events (C<ev_child>), and event watchers dealing with the event
118 loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
119 C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
120 limited support for fork events (C<ev_fork>).
122 It also is quite fast (see this
123 L<benchmark|> comparing it to libevent
124 for example).
126 =head2 CONVENTIONS
128 Libev is very configurable. In this manual the default (and most common)
129 configuration will be described, which supports multiple event loops. For
130 more info about various configuration options please have a look at
131 B<EMBED> section in this manual. If libev was configured without support
132 for multiple event loops, then all functions taking an initial argument of
133 name C<loop> (which is always of type C<struct ev_loop *>) will not have
134 this argument.
138 Libev represents time as a single floating point number, representing
139 the (fractional) number of seconds since the (POSIX) epoch (in practice
140 somewhere near the beginning of 1970, details are complicated, don't
141 ask). This type is called C<ev_tstamp>, which is what you should use
142 too. It usually aliases to the C<double> type in C. When you need to do
143 any calculations on it, you should treat it as some floating point value.
145 Unlike the name component C<stamp> might indicate, it is also used for
146 time differences (e.g. delays) throughout libev.
150 Libev knows three classes of errors: operating system errors, usage errors
151 and internal errors (bugs).
153 When libev catches an operating system error it cannot handle (for example
154 a system call indicating a condition libev cannot fix), it calls the callback
155 set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
156 abort. The default is to print a diagnostic message and to call C<abort
157 ()>.
159 When libev detects a usage error such as a negative timer interval, then
160 it will print a diagnostic message and abort (via the C<assert> mechanism,
161 so C<NDEBUG> will disable this checking): these are programming errors in
162 the libev caller and need to be fixed there.
164 Libev also has a few internal error-checking C<assert>ions, and also has
165 extensive consistency checking code. These do not trigger under normal
166 circumstances, as they indicate either a bug in libev or worse.
171 These functions can be called anytime, even before initialising the
172 library in any way.
174 =over 4
176 =item ev_tstamp ev_time ()
178 Returns the current time as libev would use it. Please note that the
179 C<ev_now> function is usually faster and also often returns the timestamp
180 you actually want to know. Also interesting is the combination of
181 C<ev_now_update> and C<ev_now>.
183 =item ev_sleep (ev_tstamp interval)
185 Sleep for the given interval: The current thread will be blocked
186 until either it is interrupted or the given time interval has
187 passed (approximately - it might return a bit earlier even if not
188 interrupted). Returns immediately if C<< interval <= 0 >>.
190 Basically this is a sub-second-resolution C<sleep ()>.
192 The range of the C<interval> is limited - libev only guarantees to work
193 with sleep times of up to one day (C<< interval <= 86400 >>).
195 =item int ev_version_major ()
197 =item int ev_version_minor ()
199 You can find out the major and minor ABI version numbers of the library
200 you linked against by calling the functions C<ev_version_major> and
201 C<ev_version_minor>. If you want, you can compare against the global
202 symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
203 version of the library your program was compiled against.
205 These version numbers refer to the ABI version of the library, not the
206 release version.
208 Usually, it's a good idea to terminate if the major versions mismatch,
209 as this indicates an incompatible change. Minor versions are usually
210 compatible to older versions, so a larger minor version alone is usually
211 not a problem.
213 Example: Make sure we haven't accidentally been linked against the wrong
214 version (note, however, that this will not detect other ABI mismatches,
215 such as LFS or reentrancy).
217 assert (("libev version mismatch",
218 ev_version_major () == EV_VERSION_MAJOR
219 && ev_version_minor () >= EV_VERSION_MINOR));
221 =item unsigned int ev_supported_backends ()
223 Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
224 value) compiled into this binary of libev (independent of their
225 availability on the system you are running on). See C<ev_default_loop> for
226 a description of the set values.
228 Example: make sure we have the epoll method, because yeah this is cool and
229 a must have and can we have a torrent of it please!!!11
231 assert (("sorry, no epoll, no sex",
232 ev_supported_backends () & EVBACKEND_EPOLL));
234 =item unsigned int ev_recommended_backends ()
236 Return the set of all backends compiled into this binary of libev and
237 also recommended for this platform, meaning it will work for most file
238 descriptor types. This set is often smaller than the one returned by
239 C<ev_supported_backends>, as for example kqueue is broken on most BSDs
240 and will not be auto-detected unless you explicitly request it (assuming
241 you know what you are doing). This is the set of backends that libev will
242 probe for if you specify no backends explicitly.
244 =item unsigned int ev_embeddable_backends ()
246 Returns the set of backends that are embeddable in other event loops. This
247 value is platform-specific but can include backends not available on the
248 current system. To find which embeddable backends might be supported on
249 the current system, you would need to look at C<ev_embeddable_backends ()
250 & ev_supported_backends ()>, likewise for recommended ones.
252 See the description of C<ev_embed> watchers for more info.
254 =item ev_set_allocator (void *(*cb)(void *ptr, long size) throw ())
256 Sets the allocation function to use (the prototype is similar - the
257 semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
258 used to allocate and free memory (no surprises here). If it returns zero
259 when memory needs to be allocated (C<size != 0>), the library might abort
260 or take some potentially destructive action.
262 Since some systems (at least OpenBSD and Darwin) fail to implement
263 correct C<realloc> semantics, libev will use a wrapper around the system
264 C<realloc> and C<free> functions by default.
266 You could override this function in high-availability programs to, say,
267 free some memory if it cannot allocate memory, to use a special allocator,
268 or even to sleep a while and retry until some memory is available.
270 Example: Replace the libev allocator with one that waits a bit and then
271 retries (example requires a standards-compliant C<realloc>).
273 static void *
274 persistent_realloc (void *ptr, size_t size)
275 {
276 for (;;)
277 {
278 void *newptr = realloc (ptr, size);
280 if (newptr)
281 return newptr;
283 sleep (60);
284 }
285 }
287 ...
288 ev_set_allocator (persistent_realloc);
290 =item ev_set_syserr_cb (void (*cb)(const char *msg) throw ())
292 Set the callback function to call on a retryable system call error (such
293 as failed select, poll, epoll_wait). The message is a printable string
294 indicating the system call or subsystem causing the problem. If this
295 callback is set, then libev will expect it to remedy the situation, no
296 matter what, when it returns. That is, libev will generally retry the
297 requested operation, or, if the condition doesn't go away, do bad stuff
298 (such as abort).
300 Example: This is basically the same thing that libev does internally, too.
302 static void
303 fatal_error (const char *msg)
304 {
305 perror (msg);
306 abort ();
307 }
309 ...
310 ev_set_syserr_cb (fatal_error);
312 =item ev_feed_signal (int signum)
314 This function can be used to "simulate" a signal receive. It is completely
315 safe to call this function at any time, from any context, including signal
316 handlers or random threads.
318 Its main use is to customise signal handling in your process, especially
319 in the presence of threads. For example, you could block signals
320 by default in all threads (and specifying C<EVFLAG_NOSIGMASK> when
321 creating any loops), and in one thread, use C<sigwait> or any other
322 mechanism to wait for signals, then "deliver" them to libev by calling
323 C<ev_feed_signal>.
325 =back
329 An event loop is described by a C<struct ev_loop *> (the C<struct> is
330 I<not> optional in this case unless libev 3 compatibility is disabled, as
331 libev 3 had an C<ev_loop> function colliding with the struct name).
333 The library knows two types of such loops, the I<default> loop, which
334 supports child process events, and dynamically created event loops which
335 do not.
337 =over 4
339 =item struct ev_loop *ev_default_loop (unsigned int flags)
341 This returns the "default" event loop object, which is what you should
342 normally use when you just need "the event loop". Event loop objects and
343 the C<flags> parameter are described in more detail in the entry for
344 C<ev_loop_new>.
346 If the default loop is already initialised then this function simply
347 returns it (and ignores the flags. If that is troubling you, check
348 C<ev_backend ()> afterwards). Otherwise it will create it with the given
349 flags, which should almost always be C<0>, unless the caller is also the
350 one calling C<ev_run> or otherwise qualifies as "the main program".
352 If you don't know what event loop to use, use the one returned from this
353 function (or via the C<EV_DEFAULT> macro).
355 Note that this function is I<not> thread-safe, so if you want to use it
356 from multiple threads, you have to employ some kind of mutex (note also
357 that this case is unlikely, as loops cannot be shared easily between
358 threads anyway).
360 The default loop is the only loop that can handle C<ev_child> watchers,
361 and to do this, it always registers a handler for C<SIGCHLD>. If this is
362 a problem for your application you can either create a dynamic loop with
363 C<ev_loop_new> which doesn't do that, or you can simply overwrite the
364 C<SIGCHLD> signal handler I<after> calling C<ev_default_init>.
366 Example: This is the most typical usage.
368 if (!ev_default_loop (0))
369 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
371 Example: Restrict libev to the select and poll backends, and do not allow
372 environment settings to be taken into account:
376 =item struct ev_loop *ev_loop_new (unsigned int flags)
378 This will create and initialise a new event loop object. If the loop
379 could not be initialised, returns false.
381 This function is thread-safe, and one common way to use libev with
382 threads is indeed to create one loop per thread, and using the default
383 loop in the "main" or "initial" thread.
385 The flags argument can be used to specify special behaviour or specific
386 backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
388 The following flags are supported:
390 =over 4
392 =item C<EVFLAG_AUTO>
394 The default flags value. Use this if you have no clue (it's the right
395 thing, believe me).
397 =item C<EVFLAG_NOENV>
399 If this flag bit is or'ed into the flag value (or the program runs setuid
400 or setgid) then libev will I<not> look at the environment variable
401 C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
402 override the flags completely if it is found in the environment. This is
403 useful to try out specific backends to test their performance, to work
404 around bugs, or to make libev threadsafe (accessing environment variables
405 cannot be done in a threadsafe way, but usually it works if no other
406 thread modifies them).
410 Instead of calling C<ev_loop_fork> manually after a fork, you can also
411 make libev check for a fork in each iteration by enabling this flag.
413 This works by calling C<getpid ()> on every iteration of the loop,
414 and thus this might slow down your event loop if you do a lot of loop
415 iterations and little real work, but is usually not noticeable (on my
416 GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
417 without a system call and thus I<very> fast, but my GNU/Linux system also has
418 C<pthread_atfork> which is even faster).
420 The big advantage of this flag is that you can forget about fork (and
421 forget about forgetting to tell libev about forking, although you still
422 have to ignore C<SIGPIPE>) when you use this flag.
424 This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
425 environment variable.
429 When this flag is specified, then libev will not attempt to use the
430 I<inotify> API for its C<ev_stat> watchers. Apart from debugging and
431 testing, this flag can be useful to conserve inotify file descriptors, as
432 otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
436 When this flag is specified, then libev will attempt to use the
437 I<signalfd> API for its C<ev_signal> (and C<ev_child>) watchers. This API
438 delivers signals synchronously, which makes it both faster and might make
439 it possible to get the queued signal data. It can also simplify signal
440 handling with threads, as long as you properly block signals in your
441 threads that are not interested in handling them.
443 Signalfd will not be used by default as this changes your signal mask, and
444 there are a lot of shoddy libraries and programs (glib's threadpool for
445 example) that can't properly initialise their signal masks.
449 When this flag is specified, then libev will avoid to modify the signal
450 mask. Specifically, this means you have to make sure signals are unblocked
451 when you want to receive them.
453 This behaviour is useful when you want to do your own signal handling, or
454 want to handle signals only in specific threads and want to avoid libev
455 unblocking the signals.
457 It's also required by POSIX in a threaded program, as libev calls
458 C<sigprocmask>, whose behaviour is officially unspecified.
460 This flag's behaviour will become the default in future versions of libev.
462 =item C<EVBACKEND_SELECT> (value 1, portable select backend)
464 This is your standard select(2) backend. Not I<completely> standard, as
465 libev tries to roll its own fd_set with no limits on the number of fds,
466 but if that fails, expect a fairly low limit on the number of fds when
467 using this backend. It doesn't scale too well (O(highest_fd)), but its
468 usually the fastest backend for a low number of (low-numbered :) fds.
470 To get good performance out of this backend you need a high amount of
471 parallelism (most of the file descriptors should be busy). If you are
472 writing a server, you should C<accept ()> in a loop to accept as many
473 connections as possible during one iteration. You might also want to have
474 a look at C<ev_set_io_collect_interval ()> to increase the amount of
475 readiness notifications you get per iteration.
477 This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
478 C<writefds> set (and to work around Microsoft Windows bugs, also onto the
479 C<exceptfds> set on that platform).
481 =item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
483 And this is your standard poll(2) backend. It's more complicated
484 than select, but handles sparse fds better and has no artificial
485 limit on the number of fds you can use (except it will slow down
486 considerably with a lot of inactive fds). It scales similarly to select,
487 i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
488 performance tips.
490 This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
493 =item C<EVBACKEND_EPOLL> (value 4, Linux)
495 Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
496 kernels).
498 For few fds, this backend is a bit little slower than poll and select, but
499 it scales phenomenally better. While poll and select usually scale like
500 O(total_fds) where total_fds is the total number of fds (or the highest
501 fd), epoll scales either O(1) or O(active_fds).
503 The epoll mechanism deserves honorable mention as the most misdesigned
504 of the more advanced event mechanisms: mere annoyances include silently
505 dropping file descriptors, requiring a system call per change per file
506 descriptor (and unnecessary guessing of parameters), problems with dup,
507 returning before the timeout value, resulting in additional iterations
508 (and only giving 5ms accuracy while select on the same platform gives
509 0.1ms) and so on. The biggest issue is fork races, however - if a program
510 forks then I<both> parent and child process have to recreate the epoll
511 set, which can take considerable time (one syscall per file descriptor)
512 and is of course hard to detect.
514 Epoll is also notoriously buggy - embedding epoll fds I<should> work,
515 but of course I<doesn't>, and epoll just loves to report events for
516 totally I<different> file descriptors (even already closed ones, so
517 one cannot even remove them from the set) than registered in the set
518 (especially on SMP systems). Libev tries to counter these spurious
519 notifications by employing an additional generation counter and comparing
520 that against the events to filter out spurious ones, recreating the set
521 when required. Epoll also erroneously rounds down timeouts, but gives you
522 no way to know when and by how much, so sometimes you have to busy-wait
523 because epoll returns immediately despite a nonzero timeout. And last
524 not least, it also refuses to work with some file descriptors which work
525 perfectly fine with C<select> (files, many character devices...).
527 Epoll is truly the train wreck among event poll mechanisms, a frankenpoll,
528 cobbled together in a hurry, no thought to design or interaction with
529 others. Oh, the pain, will it ever stop...
531 While stopping, setting and starting an I/O watcher in the same iteration
532 will result in some caching, there is still a system call per such
533 incident (because the same I<file descriptor> could point to a different
534 I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
535 file descriptors might not work very well if you register events for both
536 file descriptors.
538 Best performance from this backend is achieved by not unregistering all
539 watchers for a file descriptor until it has been closed, if possible,
540 i.e. keep at least one watcher active per fd at all times. Stopping and
541 starting a watcher (without re-setting it) also usually doesn't cause
542 extra overhead. A fork can both result in spurious notifications as well
543 as in libev having to destroy and recreate the epoll object, which can
544 take considerable time and thus should be avoided.
546 All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
547 faster than epoll for maybe up to a hundred file descriptors, depending on
548 the usage. So sad.
550 While nominally embeddable in other event loops, this feature is broken in
551 all kernel versions tested so far.
553 This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
556 =item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
558 Kqueue deserves special mention, as at the time of this writing, it
559 was broken on all BSDs except NetBSD (usually it doesn't work reliably
560 with anything but sockets and pipes, except on Darwin, where of course
561 it's completely useless). Unlike epoll, however, whose brokenness
562 is by design, these kqueue bugs can (and eventually will) be fixed
563 without API changes to existing programs. For this reason it's not being
564 "auto-detected" unless you explicitly specify it in the flags (i.e. using
565 C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
566 system like NetBSD.
568 You still can embed kqueue into a normal poll or select backend and use it
569 only for sockets (after having made sure that sockets work with kqueue on
570 the target platform). See C<ev_embed> watchers for more info.
572 It scales in the same way as the epoll backend, but the interface to the
573 kernel is more efficient (which says nothing about its actual speed, of
574 course). While stopping, setting and starting an I/O watcher does never
575 cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
576 two event changes per incident. Support for C<fork ()> is very bad (you
577 might have to leak fd's on fork, but it's more sane than epoll) and it
578 drops fds silently in similarly hard-to-detect cases.
580 This backend usually performs well under most conditions.
582 While nominally embeddable in other event loops, this doesn't work
583 everywhere, so you might need to test for this. And since it is broken
584 almost everywhere, you should only use it when you have a lot of sockets
585 (for which it usually works), by embedding it into another event loop
586 (e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
587 also broken on OS X)) and, did I mention it, using it only for sockets.
589 This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
590 C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
591 C<NOTE_EOF>.
593 =item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
595 This is not implemented yet (and might never be, unless you send me an
596 implementation). According to reports, C</dev/poll> only supports sockets
597 and is not embeddable, which would limit the usefulness of this backend
598 immensely.
600 =item C<EVBACKEND_PORT> (value 32, Solaris 10)
602 This uses the Solaris 10 event port mechanism. As with everything on Solaris,
603 it's really slow, but it still scales very well (O(active_fds)).
605 While this backend scales well, it requires one system call per active
606 file descriptor per loop iteration. For small and medium numbers of file
607 descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
608 might perform better.
610 On the positive side, this backend actually performed fully to
611 specification in all tests and is fully embeddable, which is a rare feat
612 among the OS-specific backends (I vastly prefer correctness over speed
613 hacks).
615 On the negative side, the interface is I<bizarre> - so bizarre that
616 even sun itself gets it wrong in their code examples: The event polling
617 function sometimes returns events to the caller even though an error
618 occurred, but with no indication whether it has done so or not (yes, it's
619 even documented that way) - deadly for edge-triggered interfaces where you
620 absolutely have to know whether an event occurred or not because you have
621 to re-arm the watcher.
623 Fortunately libev seems to be able to work around these idiocies.
625 This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
628 =item C<EVBACKEND_ALL>
630 Try all backends (even potentially broken ones that wouldn't be tried
631 with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
634 It is definitely not recommended to use this flag, use whatever
635 C<ev_recommended_backends ()> returns, or simply do not specify a backend
636 at all.
640 Not a backend at all, but a mask to select all backend bits from a
641 C<flags> value, in case you want to mask out any backends from a flags
642 value (e.g. when modifying the C<LIBEV_FLAGS> environment variable).
644 =back
646 If one or more of the backend flags are or'ed into the flags value,
647 then only these backends will be tried (in the reverse order as listed
648 here). If none are specified, all backends in C<ev_recommended_backends
649 ()> will be tried.
651 Example: Try to create a event loop that uses epoll and nothing else.
653 struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
654 if (!epoller)
655 fatal ("no epoll found here, maybe it hides under your chair");
657 Example: Use whatever libev has to offer, but make sure that kqueue is
658 used if available.
660 struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE);
662 =item ev_loop_destroy (loop)
664 Destroys an event loop object (frees all memory and kernel state
665 etc.). None of the active event watchers will be stopped in the normal
666 sense, so e.g. C<ev_is_active> might still return true. It is your
667 responsibility to either stop all watchers cleanly yourself I<before>
668 calling this function, or cope with the fact afterwards (which is usually
669 the easiest thing, you can just ignore the watchers and/or C<free ()> them
670 for example).
672 Note that certain global state, such as signal state (and installed signal
673 handlers), will not be freed by this function, and related watchers (such
674 as signal and child watchers) would need to be stopped manually.
676 This function is normally used on loop objects allocated by
677 C<ev_loop_new>, but it can also be used on the default loop returned by
678 C<ev_default_loop>, in which case it is not thread-safe.
680 Note that it is not advisable to call this function on the default loop
681 except in the rare occasion where you really need to free its resources.
682 If you need dynamically allocated loops it is better to use C<ev_loop_new>
683 and C<ev_loop_destroy>.
685 =item ev_loop_fork (loop)
687 This function sets a flag that causes subsequent C<ev_run> iterations
688 to reinitialise the kernel state for backends that have one. Despite
689 the name, you can call it anytime you are allowed to start or stop
690 watchers (except inside an C<ev_prepare> callback), but it makes most
691 sense after forking, in the child process. You I<must> call it (or use
692 C<EVFLAG_FORKCHECK>) in the child before resuming or calling C<ev_run>.
694 In addition, if you want to reuse a loop (via this function or
695 C<EVFLAG_FORKCHECK>), you I<also> have to ignore C<SIGPIPE>.
697 Again, you I<have> to call it on I<any> loop that you want to re-use after
698 a fork, I<even if you do not plan to use the loop in the parent>. This is
699 because some kernel interfaces *cough* I<kqueue> *cough* do funny things
700 during fork.
702 On the other hand, you only need to call this function in the child
703 process if and only if you want to use the event loop in the child. If
704 you just fork+exec or create a new loop in the child, you don't have to
705 call it at all (in fact, C<epoll> is so badly broken that it makes a
706 difference, but libev will usually detect this case on its own and do a
707 costly reset of the backend).
709 The function itself is quite fast and it's usually not a problem to call
710 it just in case after a fork.
712 Example: Automate calling C<ev_loop_fork> on the default loop when
713 using pthreads.
715 static void
716 post_fork_child (void)
717 {
718 ev_loop_fork (EV_DEFAULT);
719 }
721 ...
722 pthread_atfork (0, 0, post_fork_child);
724 =item int ev_is_default_loop (loop)
726 Returns true when the given loop is, in fact, the default loop, and false
727 otherwise.
729 =item unsigned int ev_iteration (loop)
731 Returns the current iteration count for the event loop, which is identical
732 to the number of times libev did poll for new events. It starts at C<0>
733 and happily wraps around with enough iterations.
735 This value can sometimes be useful as a generation counter of sorts (it
736 "ticks" the number of loop iterations), as it roughly corresponds with
737 C<ev_prepare> and C<ev_check> calls - and is incremented between the
738 prepare and check phases.
740 =item unsigned int ev_depth (loop)
742 Returns the number of times C<ev_run> was entered minus the number of
743 times C<ev_run> was exited normally, in other words, the recursion depth.
745 Outside C<ev_run>, this number is zero. In a callback, this number is
746 C<1>, unless C<ev_run> was invoked recursively (or from another thread),
747 in which case it is higher.
749 Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread,
750 throwing an exception etc.), doesn't count as "exit" - consider this
751 as a hint to avoid such ungentleman-like behaviour unless it's really
752 convenient, in which case it is fully supported.
754 =item unsigned int ev_backend (loop)
756 Returns one of the C<EVBACKEND_*> flags indicating the event backend in
757 use.
759 =item ev_tstamp ev_now (loop)
761 Returns the current "event loop time", which is the time the event loop
762 received events and started processing them. This timestamp does not
763 change as long as callbacks are being processed, and this is also the base
764 time used for relative timers. You can treat it as the timestamp of the
765 event occurring (or more correctly, libev finding out about it).
767 =item ev_now_update (loop)
769 Establishes the current time by querying the kernel, updating the time
770 returned by C<ev_now ()> in the progress. This is a costly operation and
771 is usually done automatically within C<ev_run ()>.
773 This function is rarely useful, but when some event callback runs for a
774 very long time without entering the event loop, updating libev's idea of
775 the current time is a good idea.
777 See also L</The special problem of time updates> in the C<ev_timer> section.
779 =item ev_suspend (loop)
781 =item ev_resume (loop)
783 These two functions suspend and resume an event loop, for use when the
784 loop is not used for a while and timeouts should not be processed.
786 A typical use case would be an interactive program such as a game: When
787 the user presses C<^Z> to suspend the game and resumes it an hour later it
788 would be best to handle timeouts as if no time had actually passed while
789 the program was suspended. This can be achieved by calling C<ev_suspend>
790 in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
791 C<ev_resume> directly afterwards to resume timer processing.
793 Effectively, all C<ev_timer> watchers will be delayed by the time spend
794 between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
795 will be rescheduled (that is, they will lose any events that would have
796 occurred while suspended).
798 After calling C<ev_suspend> you B<must not> call I<any> function on the
799 given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
800 without a previous call to C<ev_suspend>.
802 Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
803 event loop time (see C<ev_now_update>).
805 =item bool ev_run (loop, int flags)
807 Finally, this is it, the event handler. This function usually is called
808 after you have initialised all your watchers and you want to start
809 handling events. It will ask the operating system for any new events, call
810 the watcher callbacks, and then repeat the whole process indefinitely: This
811 is why event loops are called I<loops>.
813 If the flags argument is specified as C<0>, it will keep handling events
814 until either no event watchers are active anymore or C<ev_break> was
815 called.
817 The return value is false if there are no more active watchers (which
818 usually means "all jobs done" or "deadlock"), and true in all other cases
819 (which usually means " you should call C<ev_run> again").
821 Please note that an explicit C<ev_break> is usually better than
822 relying on all watchers to be stopped when deciding when a program has
823 finished (especially in interactive programs), but having a program
824 that automatically loops as long as it has to and no longer by virtue
825 of relying on its watchers stopping correctly, that is truly a thing of
826 beauty.
828 This function is I<mostly> exception-safe - you can break out of a
829 C<ev_run> call by calling C<longjmp> in a callback, throwing a C++
830 exception and so on. This does not decrement the C<ev_depth> value, nor
831 will it clear any outstanding C<EVBREAK_ONE> breaks.
833 A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
834 those events and any already outstanding ones, but will not wait and
835 block your process in case there are no events and will return after one
836 iteration of the loop. This is sometimes useful to poll and handle new
837 events while doing lengthy calculations, to keep the program responsive.
839 A flags value of C<EVRUN_ONCE> will look for new events (waiting if
840 necessary) and will handle those and any already outstanding ones. It
841 will block your process until at least one new event arrives (which could
842 be an event internal to libev itself, so there is no guarantee that a
843 user-registered callback will be called), and will return after one
844 iteration of the loop.
846 This is useful if you are waiting for some external event in conjunction
847 with something not expressible using other libev watchers (i.e. "roll your
848 own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
849 usually a better approach for this kind of thing.
851 Here are the gory details of what C<ev_run> does (this is for your
852 understanding, not a guarantee that things will work exactly like this in
853 future versions):
855 - Increment loop depth.
856 - Reset the ev_break status.
857 - Before the first iteration, call any pending watchers.
858 LOOP:
859 - If EVFLAG_FORKCHECK was used, check for a fork.
860 - If a fork was detected (by any means), queue and call all fork watchers.
861 - Queue and call all prepare watchers.
862 - If ev_break was called, goto FINISH.
863 - If we have been forked, detach and recreate the kernel state
864 as to not disturb the other process.
865 - Update the kernel state with all outstanding changes.
866 - Update the "event loop time" (ev_now ()).
867 - Calculate for how long to sleep or block, if at all
868 (active idle watchers, EVRUN_NOWAIT or not having
869 any active watchers at all will result in not sleeping).
870 - Sleep if the I/O and timer collect interval say so.
871 - Increment loop iteration counter.
872 - Block the process, waiting for any events.
873 - Queue all outstanding I/O (fd) events.
874 - Update the "event loop time" (ev_now ()), and do time jump adjustments.
875 - Queue all expired timers.
876 - Queue all expired periodics.
877 - Queue all idle watchers with priority higher than that of pending events.
878 - Queue all check watchers.
879 - Call all queued watchers in reverse order (i.e. check watchers first).
880 Signals and child watchers are implemented as I/O watchers, and will
881 be handled here by queueing them when their watcher gets executed.
882 - If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
883 were used, or there are no active watchers, goto FINISH, otherwise
884 continue with step LOOP.
886 - Reset the ev_break status iff it was EVBREAK_ONE.
887 - Decrement the loop depth.
888 - Return.
890 Example: Queue some jobs and then loop until no events are outstanding
891 anymore.
893 ... queue jobs here, make sure they register event watchers as long
894 ... as they still have work to do (even an idle watcher will do..)
895 ev_run (my_loop, 0);
896 ... jobs done or somebody called break. yeah!
898 =item ev_break (loop, how)
900 Can be used to make a call to C<ev_run> return early (but only after it
901 has processed all outstanding events). The C<how> argument must be either
902 C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
903 C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
905 This "break state" will be cleared on the next call to C<ev_run>.
907 It is safe to call C<ev_break> from outside any C<ev_run> calls, too, in
908 which case it will have no effect.
910 =item ev_ref (loop)
912 =item ev_unref (loop)
914 Ref/unref can be used to add or remove a reference count on the event
915 loop: Every watcher keeps one reference, and as long as the reference
916 count is nonzero, C<ev_run> will not return on its own.
918 This is useful when you have a watcher that you never intend to
919 unregister, but that nevertheless should not keep C<ev_run> from
920 returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
921 before stopping it.
923 As an example, libev itself uses this for its internal signal pipe: It
924 is not visible to the libev user and should not keep C<ev_run> from
925 exiting if no event watchers registered by it are active. It is also an
926 excellent way to do this for generic recurring timers or from within
927 third-party libraries. Just remember to I<unref after start> and I<ref
928 before stop> (but only if the watcher wasn't active before, or was active
929 before, respectively. Note also that libev might stop watchers itself
930 (e.g. non-repeating timers) in which case you have to C<ev_ref>
931 in the callback).
933 Example: Create a signal watcher, but keep it from keeping C<ev_run>
934 running when nothing else is active.
936 ev_signal exitsig;
937 ev_signal_init (&exitsig, sig_cb, SIGINT);
938 ev_signal_start (loop, &exitsig);
939 ev_unref (loop);
941 Example: For some weird reason, unregister the above signal handler again.
943 ev_ref (loop);
944 ev_signal_stop (loop, &exitsig);
946 =item ev_set_io_collect_interval (loop, ev_tstamp interval)
948 =item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
950 These advanced functions influence the time that libev will spend waiting
951 for events. Both time intervals are by default C<0>, meaning that libev
952 will try to invoke timer/periodic callbacks and I/O callbacks with minimum
953 latency.
955 Setting these to a higher value (the C<interval> I<must> be >= C<0>)
956 allows libev to delay invocation of I/O and timer/periodic callbacks
957 to increase efficiency of loop iterations (or to increase power-saving
958 opportunities).
960 The idea is that sometimes your program runs just fast enough to handle
961 one (or very few) event(s) per loop iteration. While this makes the
962 program responsive, it also wastes a lot of CPU time to poll for new
963 events, especially with backends like C<select ()> which have a high
964 overhead for the actual polling but can deliver many events at once.
966 By setting a higher I<io collect interval> you allow libev to spend more
967 time collecting I/O events, so you can handle more events per iteration,
968 at the cost of increasing latency. Timeouts (both C<ev_periodic> and
969 C<ev_timer>) will not be affected. Setting this to a non-null value will
970 introduce an additional C<ev_sleep ()> call into most loop iterations. The
971 sleep time ensures that libev will not poll for I/O events more often then
972 once per this interval, on average (as long as the host time resolution is
973 good enough).
975 Likewise, by setting a higher I<timeout collect interval> you allow libev
976 to spend more time collecting timeouts, at the expense of increased
977 latency/jitter/inexactness (the watcher callback will be called
978 later). C<ev_io> watchers will not be affected. Setting this to a non-null
979 value will not introduce any overhead in libev.
981 Many (busy) programs can usually benefit by setting the I/O collect
982 interval to a value near C<0.1> or so, which is often enough for
983 interactive servers (of course not for games), likewise for timeouts. It
984 usually doesn't make much sense to set it to a lower value than C<0.01>,
985 as this approaches the timing granularity of most systems. Note that if
986 you do transactions with the outside world and you can't increase the
987 parallelity, then this setting will limit your transaction rate (if you
988 need to poll once per transaction and the I/O collect interval is 0.01,
989 then you can't do more than 100 transactions per second).
991 Setting the I<timeout collect interval> can improve the opportunity for
992 saving power, as the program will "bundle" timer callback invocations that
993 are "near" in time together, by delaying some, thus reducing the number of
994 times the process sleeps and wakes up again. Another useful technique to
995 reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
996 they fire on, say, one-second boundaries only.
998 Example: we only need 0.1s timeout granularity, and we wish not to poll
999 more often than 100 times per second:
1001 ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
1002 ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
1004 =item ev_invoke_pending (loop)
1006 This call will simply invoke all pending watchers while resetting their
1007 pending state. Normally, C<ev_run> does this automatically when required,
1008 but when overriding the invoke callback this call comes handy. This
1009 function can be invoked from a watcher - this can be useful for example
1010 when you want to do some lengthy calculation and want to pass further
1011 event handling to another thread (you still have to make sure only one
1012 thread executes within C<ev_invoke_pending> or C<ev_run> of course).
1014 =item int ev_pending_count (loop)
1016 Returns the number of pending watchers - zero indicates that no watchers
1017 are pending.
1019 =item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
1021 This overrides the invoke pending functionality of the loop: Instead of
1022 invoking all pending watchers when there are any, C<ev_run> will call
1023 this callback instead. This is useful, for example, when you want to
1024 invoke the actual watchers inside another context (another thread etc.).
1026 If you want to reset the callback, use C<ev_invoke_pending> as new
1027 callback.
1029 =item ev_set_loop_release_cb (loop, void (*release)(EV_P) throw (), void (*acquire)(EV_P) throw ())
1031 Sometimes you want to share the same loop between multiple threads. This
1032 can be done relatively simply by putting mutex_lock/unlock calls around
1033 each call to a libev function.
1035 However, C<ev_run> can run an indefinite time, so it is not feasible
1036 to wait for it to return. One way around this is to wake up the event
1037 loop via C<ev_break> and C<ev_async_send>, another way is to set these
1038 I<release> and I<acquire> callbacks on the loop.
1040 When set, then C<release> will be called just before the thread is
1041 suspended waiting for new events, and C<acquire> is called just
1042 afterwards.
1044 Ideally, C<release> will just call your mutex_unlock function, and
1045 C<acquire> will just call the mutex_lock function again.
1047 While event loop modifications are allowed between invocations of
1048 C<release> and C<acquire> (that's their only purpose after all), no
1049 modifications done will affect the event loop, i.e. adding watchers will
1050 have no effect on the set of file descriptors being watched, or the time
1051 waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
1052 to take note of any changes you made.
1054 In theory, threads executing C<ev_run> will be async-cancel safe between
1055 invocations of C<release> and C<acquire>.
1057 See also the locking example in the C<THREADS> section later in this
1058 document.
1060 =item ev_set_userdata (loop, void *data)
1062 =item void *ev_userdata (loop)
1064 Set and retrieve a single C<void *> associated with a loop. When
1065 C<ev_set_userdata> has never been called, then C<ev_userdata> returns
1066 C<0>.
1068 These two functions can be used to associate arbitrary data with a loop,
1069 and are intended solely for the C<invoke_pending_cb>, C<release> and
1070 C<acquire> callbacks described above, but of course can be (ab-)used for
1071 any other purpose as well.
1073 =item ev_verify (loop)
1075 This function only does something when C<EV_VERIFY> support has been
1076 compiled in, which is the default for non-minimal builds. It tries to go
1077 through all internal structures and checks them for validity. If anything
1078 is found to be inconsistent, it will print an error message to standard
1079 error and call C<abort ()>.
1081 This can be used to catch bugs inside libev itself: under normal
1082 circumstances, this function will never abort as of course libev keeps its
1083 data structures consistent.
1085 =back
1090 In the following description, uppercase C<TYPE> in names stands for the
1091 watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
1092 watchers and C<ev_io_start> for I/O watchers.
1094 A watcher is an opaque structure that you allocate and register to record
1095 your interest in some event. To make a concrete example, imagine you want
1096 to wait for STDIN to become readable, you would create an C<ev_io> watcher
1097 for that:
1099 static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
1100 {
1101 ev_io_stop (w);
1102 ev_break (loop, EVBREAK_ALL);
1103 }
1105 struct ev_loop *loop = ev_default_loop (0);
1107 ev_io stdin_watcher;
1109 ev_init (&stdin_watcher, my_cb);
1110 ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
1111 ev_io_start (loop, &stdin_watcher);
1113 ev_run (loop, 0);
1115 As you can see, you are responsible for allocating the memory for your
1116 watcher structures (and it is I<usually> a bad idea to do this on the
1117 stack).
1119 Each watcher has an associated watcher structure (called C<struct ev_TYPE>
1120 or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
1122 Each watcher structure must be initialised by a call to C<ev_init (watcher
1123 *, callback)>, which expects a callback to be provided. This callback is
1124 invoked each time the event occurs (or, in the case of I/O watchers, each
1125 time the event loop detects that the file descriptor given is readable
1126 and/or writable).
1128 Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
1129 macro to configure it, with arguments specific to the watcher type. There
1130 is also a macro to combine initialisation and setting in one call: C<<
1131 ev_TYPE_init (watcher *, callback, ...) >>.
1133 To make the watcher actually watch out for events, you have to start it
1134 with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
1135 *) >>), and you can stop watching for events at any time by calling the
1136 corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
1138 As long as your watcher is active (has been started but not stopped) you
1139 must not touch the values stored in it. Most specifically you must never
1140 reinitialise it or call its C<ev_TYPE_set> macro.
1142 Each and every callback receives the event loop pointer as first, the
1143 registered watcher structure as second, and a bitset of received events as
1144 third argument.
1146 The received events usually include a single bit per event type received
1147 (you can receive multiple events at the same time). The possible bit masks
1148 are:
1150 =over 4
1152 =item C<EV_READ>
1154 =item C<EV_WRITE>
1156 The file descriptor in the C<ev_io> watcher has become readable and/or
1157 writable.
1159 =item C<EV_TIMER>
1161 The C<ev_timer> watcher has timed out.
1163 =item C<EV_PERIODIC>
1165 The C<ev_periodic> watcher has timed out.
1167 =item C<EV_SIGNAL>
1169 The signal specified in the C<ev_signal> watcher has been received by a thread.
1171 =item C<EV_CHILD>
1173 The pid specified in the C<ev_child> watcher has received a status change.
1175 =item C<EV_STAT>
1177 The path specified in the C<ev_stat> watcher changed its attributes somehow.
1179 =item C<EV_IDLE>
1181 The C<ev_idle> watcher has determined that you have nothing better to do.
1183 =item C<EV_PREPARE>
1185 =item C<EV_CHECK>
1187 All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts to
1188 gather new events, and all C<ev_check> watchers are queued (not invoked)
1189 just after C<ev_run> has gathered them, but before it queues any callbacks
1190 for any received events. That means C<ev_prepare> watchers are the last
1191 watchers invoked before the event loop sleeps or polls for new events, and
1192 C<ev_check> watchers will be invoked before any other watchers of the same
1193 or lower priority within an event loop iteration.
1195 Callbacks of both watcher types can start and stop as many watchers as
1196 they want, and all of them will be taken into account (for example, a
1197 C<ev_prepare> watcher might start an idle watcher to keep C<ev_run> from
1198 blocking).
1200 =item C<EV_EMBED>
1202 The embedded event loop specified in the C<ev_embed> watcher needs attention.
1204 =item C<EV_FORK>
1206 The event loop has been resumed in the child process after fork (see
1207 C<ev_fork>).
1209 =item C<EV_CLEANUP>
1211 The event loop is about to be destroyed (see C<ev_cleanup>).
1213 =item C<EV_ASYNC>
1215 The given async watcher has been asynchronously notified (see C<ev_async>).
1217 =item C<EV_CUSTOM>
1219 Not ever sent (or otherwise used) by libev itself, but can be freely used
1220 by libev users to signal watchers (e.g. via C<ev_feed_event>).
1222 =item C<EV_ERROR>
1224 An unspecified error has occurred, the watcher has been stopped. This might
1225 happen because the watcher could not be properly started because libev
1226 ran out of memory, a file descriptor was found to be closed or any other
1227 problem. Libev considers these application bugs.
1229 You best act on it by reporting the problem and somehow coping with the
1230 watcher being stopped. Note that well-written programs should not receive
1231 an error ever, so when your watcher receives it, this usually indicates a
1232 bug in your program.
1234 Libev will usually signal a few "dummy" events together with an error, for
1235 example it might indicate that a fd is readable or writable, and if your
1236 callbacks is well-written it can just attempt the operation and cope with
1237 the error from read() or write(). This will not work in multi-threaded
1238 programs, though, as the fd could already be closed and reused for another
1239 thing, so beware.
1241 =back
1245 =over 4
1247 =item C<ev_init> (ev_TYPE *watcher, callback)
1249 This macro initialises the generic portion of a watcher. The contents
1250 of the watcher object can be arbitrary (so C<malloc> will do). Only
1251 the generic parts of the watcher are initialised, you I<need> to call
1252 the type-specific C<ev_TYPE_set> macro afterwards to initialise the
1253 type-specific parts. For each type there is also a C<ev_TYPE_init> macro
1254 which rolls both calls into one.
1256 You can reinitialise a watcher at any time as long as it has been stopped
1257 (or never started) and there are no pending events outstanding.
1259 The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
1260 int revents)>.
1262 Example: Initialise an C<ev_io> watcher in two steps.
1264 ev_io w;
1265 ev_init (&w, my_cb);
1266 ev_io_set (&w, STDIN_FILENO, EV_READ);
1268 =item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
1270 This macro initialises the type-specific parts of a watcher. You need to
1271 call C<ev_init> at least once before you call this macro, but you can
1272 call C<ev_TYPE_set> any number of times. You must not, however, call this
1273 macro on a watcher that is active (it can be pending, however, which is a
1274 difference to the C<ev_init> macro).
1276 Although some watcher types do not have type-specific arguments
1277 (e.g. C<ev_prepare>) you still need to call its C<set> macro.
1279 See C<ev_init>, above, for an example.
1281 =item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
1283 This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
1284 calls into a single call. This is the most convenient method to initialise
1285 a watcher. The same limitations apply, of course.
1287 Example: Initialise and set an C<ev_io> watcher in one step.
1289 ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
1291 =item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
1293 Starts (activates) the given watcher. Only active watchers will receive
1294 events. If the watcher is already active nothing will happen.
1296 Example: Start the C<ev_io> watcher that is being abused as example in this
1297 whole section.
1299 ev_io_start (EV_DEFAULT_UC, &w);
1301 =item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
1303 Stops the given watcher if active, and clears the pending status (whether
1304 the watcher was active or not).
1306 It is possible that stopped watchers are pending - for example,
1307 non-repeating timers are being stopped when they become pending - but
1308 calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
1309 pending. If you want to free or reuse the memory used by the watcher it is
1310 therefore a good idea to always call its C<ev_TYPE_stop> function.
1312 =item bool ev_is_active (ev_TYPE *watcher)
1314 Returns a true value iff the watcher is active (i.e. it has been started
1315 and not yet been stopped). As long as a watcher is active you must not modify
1316 it.
1318 =item bool ev_is_pending (ev_TYPE *watcher)
1320 Returns a true value iff the watcher is pending, (i.e. it has outstanding
1321 events but its callback has not yet been invoked). As long as a watcher
1322 is pending (but not active) you must not call an init function on it (but
1323 C<ev_TYPE_set> is safe), you must not change its priority, and you must
1324 make sure the watcher is available to libev (e.g. you cannot C<free ()>
1325 it).
1327 =item callback ev_cb (ev_TYPE *watcher)
1329 Returns the callback currently set on the watcher.
1331 =item ev_set_cb (ev_TYPE *watcher, callback)
1333 Change the callback. You can change the callback at virtually any time
1334 (modulo threads).
1336 =item ev_set_priority (ev_TYPE *watcher, int priority)
1338 =item int ev_priority (ev_TYPE *watcher)
1340 Set and query the priority of the watcher. The priority is a small
1341 integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
1342 (default: C<-2>). Pending watchers with higher priority will be invoked
1343 before watchers with lower priority, but priority will not keep watchers
1344 from being executed (except for C<ev_idle> watchers).
1346 If you need to suppress invocation when higher priority events are pending
1347 you need to look at C<ev_idle> watchers, which provide this functionality.
1349 You I<must not> change the priority of a watcher as long as it is active or
1350 pending.
1352 Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
1353 fine, as long as you do not mind that the priority value you query might
1354 or might not have been clamped to the valid range.
1356 The default priority used by watchers when no priority has been set is
1357 always C<0>, which is supposed to not be too high and not be too low :).
1359 See L</WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
1360 priorities.
1362 =item ev_invoke (loop, ev_TYPE *watcher, int revents)
1364 Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
1365 C<loop> nor C<revents> need to be valid as long as the watcher callback
1366 can deal with that fact, as both are simply passed through to the
1367 callback.
1369 =item int ev_clear_pending (loop, ev_TYPE *watcher)
1371 If the watcher is pending, this function clears its pending status and
1372 returns its C<revents> bitset (as if its callback was invoked). If the
1373 watcher isn't pending it does nothing and returns C<0>.
1375 Sometimes it can be useful to "poll" a watcher instead of waiting for its
1376 callback to be invoked, which can be accomplished with this function.
1378 =item ev_feed_event (loop, ev_TYPE *watcher, int revents)
1380 Feeds the given event set into the event loop, as if the specified event
1381 had happened for the specified watcher (which must be a pointer to an
1382 initialised but not necessarily started event watcher). Obviously you must
1383 not free the watcher as long as it has pending events.
1385 Stopping the watcher, letting libev invoke it, or calling
1386 C<ev_clear_pending> will clear the pending event, even if the watcher was
1387 not started in the first place.
1389 See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
1390 functions that do not need a watcher.
1392 =back
1397 =head2 WATCHER STATES
1399 There are various watcher states mentioned throughout this manual -
1400 active, pending and so on. In this section these states and the rules to
1401 transition between them will be described in more detail - and while these
1402 rules might look complicated, they usually do "the right thing".
1404 =over 4
1406 =item initialised
1408 Before a watcher can be registered with the event loop it has to be
1409 initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
1410 C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
1412 In this state it is simply some block of memory that is suitable for
1413 use in an event loop. It can be moved around, freed, reused etc. at
1414 will - as long as you either keep the memory contents intact, or call
1415 C<ev_TYPE_init> again.
1417 =item started/running/active
1419 Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
1420 property of the event loop, and is actively waiting for events. While in
1421 this state it cannot be accessed (except in a few documented ways), moved,
1422 freed or anything else - the only legal thing is to keep a pointer to it,
1423 and call libev functions on it that are documented to work on active watchers.
1425 =item pending
1427 If a watcher is active and libev determines that an event it is interested
1428 in has occurred (such as a timer expiring), it will become pending. It will
1429 stay in this pending state until either it is stopped or its callback is
1430 about to be invoked, so it is not normally pending inside the watcher
1431 callback.
1433 The watcher might or might not be active while it is pending (for example,
1434 an expired non-repeating timer can be pending but no longer active). If it
1435 is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
1436 but it is still property of the event loop at this time, so cannot be
1437 moved, freed or reused. And if it is active the rules described in the
1438 previous item still apply.
1440 It is also possible to feed an event on a watcher that is not active (e.g.
1441 via C<ev_feed_event>), in which case it becomes pending without being
1442 active.
1444 =item stopped
1446 A watcher can be stopped implicitly by libev (in which case it might still
1447 be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
1448 latter will clear any pending state the watcher might be in, regardless
1449 of whether it was active or not, so stopping a watcher explicitly before
1450 freeing it is often a good idea.
1452 While stopped (and not pending) the watcher is essentially in the
1453 initialised state, that is, it can be reused, moved, modified in any way
1454 you wish (but when you trash the memory block, you need to C<ev_TYPE_init>
1455 it again).
1457 =back
1461 Many event loops support I<watcher priorities>, which are usually small
1462 integers that influence the ordering of event callback invocation
1463 between watchers in some way, all else being equal.
1465 In libev, Watcher priorities can be set using C<ev_set_priority>. See its
1466 description for the more technical details such as the actual priority
1467 range.
1469 There are two common ways how these these priorities are being interpreted
1470 by event loops:
1472 In the more common lock-out model, higher priorities "lock out" invocation
1473 of lower priority watchers, which means as long as higher priority
1474 watchers receive events, lower priority watchers are not being invoked.
1476 The less common only-for-ordering model uses priorities solely to order
1477 callback invocation within a single event loop iteration: Higher priority
1478 watchers are invoked before lower priority ones, but they all get invoked
1479 before polling for new events.
1481 Libev uses the second (only-for-ordering) model for all its watchers
1482 except for idle watchers (which use the lock-out model).
1484 The rationale behind this is that implementing the lock-out model for
1485 watchers is not well supported by most kernel interfaces, and most event
1486 libraries will just poll for the same events again and again as long as
1487 their callbacks have not been executed, which is very inefficient in the
1488 common case of one high-priority watcher locking out a mass of lower
1489 priority ones.
1491 Static (ordering) priorities are most useful when you have two or more
1492 watchers handling the same resource: a typical usage example is having an
1493 C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
1494 timeouts. Under load, data might be received while the program handles
1495 other jobs, but since timers normally get invoked first, the timeout
1496 handler will be executed before checking for data. In that case, giving
1497 the timer a lower priority than the I/O watcher ensures that I/O will be
1498 handled first even under adverse conditions (which is usually, but not
1499 always, what you want).
1501 Since idle watchers use the "lock-out" model, meaning that idle watchers
1502 will only be executed when no same or higher priority watchers have
1503 received events, they can be used to implement the "lock-out" model when
1504 required.
1506 For example, to emulate how many other event libraries handle priorities,
1507 you can associate an C<ev_idle> watcher to each such watcher, and in
1508 the normal watcher callback, you just start the idle watcher. The real
1509 processing is done in the idle watcher callback. This causes libev to
1510 continuously poll and process kernel event data for the watcher, but when
1511 the lock-out case is known to be rare (which in turn is rare :), this is
1512 workable.
1514 Usually, however, the lock-out model implemented that way will perform
1515 miserably under the type of load it was designed to handle. In that case,
1516 it might be preferable to stop the real watcher before starting the
1517 idle watcher, so the kernel will not have to process the event in case
1518 the actual processing will be delayed for considerable time.
1520 Here is an example of an I/O watcher that should run at a strictly lower
1521 priority than the default, and which should only process data when no
1522 other events are pending:
1524 ev_idle idle; // actual processing watcher
1525 ev_io io; // actual event watcher
1527 static void
1528 io_cb (EV_P_ ev_io *w, int revents)
1529 {
1530 // stop the I/O watcher, we received the event, but
1531 // are not yet ready to handle it.
1532 ev_io_stop (EV_A_ w);
1534 // start the idle watcher to handle the actual event.
1535 // it will not be executed as long as other watchers
1536 // with the default priority are receiving events.
1537 ev_idle_start (EV_A_ &idle);
1538 }
1540 static void
1541 idle_cb (EV_P_ ev_idle *w, int revents)
1542 {
1543 // actual processing
1544 read (STDIN_FILENO, ...);
1546 // have to start the I/O watcher again, as
1547 // we have handled the event
1548 ev_io_start (EV_P_ &io);
1549 }
1551 // initialisation
1552 ev_idle_init (&idle, idle_cb);
1553 ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
1554 ev_io_start (EV_DEFAULT_ &io);
1556 In the "real" world, it might also be beneficial to start a timer, so that
1557 low-priority connections can not be locked out forever under load. This
1558 enables your program to keep a lower latency for important connections
1559 during short periods of high load, while not completely locking out less
1560 important ones.
1563 =head1 WATCHER TYPES
1565 This section describes each watcher in detail, but will not repeat
1566 information given in the last section. Any initialisation/set macros,
1567 functions and members specific to the watcher type are explained.
1569 Members are additionally marked with either I<[read-only]>, meaning that,
1570 while the watcher is active, you can look at the member and expect some
1571 sensible content, but you must not modify it (you can modify it while the
1572 watcher is stopped to your hearts content), or I<[read-write]>, which
1573 means you can expect it to have some sensible content while the watcher
1574 is active, but you can also modify it. Modifying it may not do something
1575 sensible or take immediate effect (or do anything at all), but libev will
1576 not crash or malfunction in any way.
1579 =head2 C<ev_io> - is this file descriptor readable or writable?
1581 I/O watchers check whether a file descriptor is readable or writable
1582 in each iteration of the event loop, or, more precisely, when reading
1583 would not block the process and writing would at least be able to write
1584 some data. This behaviour is called level-triggering because you keep
1585 receiving events as long as the condition persists. Remember you can stop
1586 the watcher if you don't want to act on the event and neither want to
1587 receive future events.
1589 In general you can register as many read and/or write event watchers per
1590 fd as you want (as long as you don't confuse yourself). Setting all file
1591 descriptors to non-blocking mode is also usually a good idea (but not
1592 required if you know what you are doing).
1594 Another thing you have to watch out for is that it is quite easy to
1595 receive "spurious" readiness notifications, that is, your callback might
1596 be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1597 because there is no data. It is very easy to get into this situation even
1598 with a relatively standard program structure. Thus it is best to always
1599 use non-blocking I/O: An extra C<read>(2) returning C<EAGAIN> is far
1600 preferable to a program hanging until some data arrives.
1602 If you cannot run the fd in non-blocking mode (for example you should
1603 not play around with an Xlib connection), then you have to separately
1604 re-test whether a file descriptor is really ready with a known-to-be good
1605 interface such as poll (fortunately in the case of Xlib, it already does
1606 this on its own, so its quite safe to use). Some people additionally
1607 use C<SIGALRM> and an interval timer, just to be sure you won't block
1608 indefinitely.
1610 But really, best use non-blocking mode.
1612 =head3 The special problem of disappearing file descriptors
1614 Some backends (e.g. kqueue, epoll) need to be told about closing a file
1615 descriptor (either due to calling C<close> explicitly or any other means,
1616 such as C<dup2>). The reason is that you register interest in some file
1617 descriptor, but when it goes away, the operating system will silently drop
1618 this interest. If another file descriptor with the same number then is
1619 registered with libev, there is no efficient way to see that this is, in
1620 fact, a different file descriptor.
1622 To avoid having to explicitly tell libev about such cases, libev follows
1623 the following policy: Each time C<ev_io_set> is being called, libev
1624 will assume that this is potentially a new file descriptor, otherwise
1625 it is assumed that the file descriptor stays the same. That means that
1626 you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
1627 descriptor even if the file descriptor number itself did not change.
1629 This is how one would do it normally anyway, the important point is that
1630 the libev application should not optimise around libev but should leave
1631 optimisations to libev.
1633 =head3 The special problem of dup'ed file descriptors
1635 Some backends (e.g. epoll), cannot register events for file descriptors,
1636 but only events for the underlying file descriptions. That means when you
1637 have C<dup ()>'ed file descriptors or weirder constellations, and register
1638 events for them, only one file descriptor might actually receive events.
1640 There is no workaround possible except not registering events
1641 for potentially C<dup ()>'ed file descriptors, or to resort to
1644 =head3 The special problem of files
1646 Many people try to use C<select> (or libev) on file descriptors
1647 representing files, and expect it to become ready when their program
1648 doesn't block on disk accesses (which can take a long time on their own).
1650 However, this cannot ever work in the "expected" way - you get a readiness
1651 notification as soon as the kernel knows whether and how much data is
1652 there, and in the case of open files, that's always the case, so you
1653 always get a readiness notification instantly, and your read (or possibly
1654 write) will still block on the disk I/O.
1656 Another way to view it is that in the case of sockets, pipes, character
1657 devices and so on, there is another party (the sender) that delivers data
1658 on its own, but in the case of files, there is no such thing: the disk
1659 will not send data on its own, simply because it doesn't know what you
1660 wish to read - you would first have to request some data.
1662 Since files are typically not-so-well supported by advanced notification
1663 mechanism, libev tries hard to emulate POSIX behaviour with respect
1664 to files, even though you should not use it. The reason for this is
1665 convenience: sometimes you want to watch STDIN or STDOUT, which is
1666 usually a tty, often a pipe, but also sometimes files or special devices
1667 (for example, C<epoll> on Linux works with F</dev/random> but not with
1668 F</dev/urandom>), and even though the file might better be served with
1669 asynchronous I/O instead of with non-blocking I/O, it is still useful when
1670 it "just works" instead of freezing.
1672 So avoid file descriptors pointing to files when you know it (e.g. use
1673 libeio), but use them when it is convenient, e.g. for STDIN/STDOUT, or
1674 when you rarely read from a file instead of from a socket, and want to
1675 reuse the same code path.
1677 =head3 The special problem of fork
1679 Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
1680 useless behaviour. Libev fully supports fork, but needs to be told about
1681 it in the child if you want to continue to use it in the child.
1683 To support fork in your child processes, you have to call C<ev_loop_fork
1684 ()> after a fork in the child, enable C<EVFLAG_FORKCHECK>, or resort to
1687 =head3 The special problem of SIGPIPE
1689 While not really specific to libev, it is easy to forget about C<SIGPIPE>:
1690 when writing to a pipe whose other end has been closed, your program gets
1691 sent a SIGPIPE, which, by default, aborts your program. For most programs
1692 this is sensible behaviour, for daemons, this is usually undesirable.
1694 So when you encounter spurious, unexplained daemon exits, make sure you
1695 ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
1696 somewhere, as that would have given you a big clue).
1698 =head3 The special problem of accept()ing when you can't
1700 Many implementations of the POSIX C<accept> function (for example,
1701 found in post-2004 Linux) have the peculiar behaviour of not removing a
1702 connection from the pending queue in all error cases.
1704 For example, larger servers often run out of file descriptors (because
1705 of resource limits), causing C<accept> to fail with C<ENFILE> but not
1706 rejecting the connection, leading to libev signalling readiness on
1707 the next iteration again (the connection still exists after all), and
1708 typically causing the program to loop at 100% CPU usage.
1710 Unfortunately, the set of errors that cause this issue differs between
1711 operating systems, there is usually little the app can do to remedy the
1712 situation, and no known thread-safe method of removing the connection to
1713 cope with overload is known (to me).
1715 One of the easiest ways to handle this situation is to just ignore it
1716 - when the program encounters an overload, it will just loop until the
1717 situation is over. While this is a form of busy waiting, no OS offers an
1718 event-based way to handle this situation, so it's the best one can do.
1720 A better way to handle the situation is to log any errors other than
1721 C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
1722 messages, and continue as usual, which at least gives the user an idea of
1723 what could be wrong ("raise the ulimit!"). For extra points one could stop
1724 the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
1725 usage.
1727 If your program is single-threaded, then you could also keep a dummy file
1728 descriptor for overload situations (e.g. by opening F</dev/null>), and
1729 when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
1730 close that fd, and create a new dummy fd. This will gracefully refuse
1731 clients under typical overload conditions.
1733 The last way to handle it is to simply log the error and C<exit>, as
1734 is often done with C<malloc> failures, but this results in an easy
1735 opportunity for a DoS attack.
1737 =head3 Watcher-Specific Functions
1739 =over 4
1741 =item ev_io_init (ev_io *, callback, int fd, int events)
1743 =item ev_io_set (ev_io *, int fd, int events)
1745 Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1746 receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
1747 C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
1749 =item int fd [read-only]
1751 The file descriptor being watched.
1753 =item int events [read-only]
1755 The events being watched.
1757 =back
1759 =head3 Examples
1761 Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1762 readable, but only once. Since it is likely line-buffered, you could
1763 attempt to read a whole line in the callback.
1765 static void
1766 stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1767 {
1768 ev_io_stop (loop, w);
1769 .. read from stdin here (or from w->fd) and handle any I/O errors
1770 }
1772 ...
1773 struct ev_loop *loop = ev_default_init (0);
1774 ev_io stdin_readable;
1775 ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1776 ev_io_start (loop, &stdin_readable);
1777 ev_run (loop, 0);
1780 =head2 C<ev_timer> - relative and optionally repeating timeouts
1782 Timer watchers are simple relative timers that generate an event after a
1783 given time, and optionally repeating in regular intervals after that.
1785 The timers are based on real time, that is, if you register an event that
1786 times out after an hour and you reset your system clock to January last
1787 year, it will still time out after (roughly) one hour. "Roughly" because
1788 detecting time jumps is hard, and some inaccuracies are unavoidable (the
1789 monotonic clock option helps a lot here).
1791 The callback is guaranteed to be invoked only I<after> its timeout has
1792 passed (not I<at>, so on systems with very low-resolution clocks this
1793 might introduce a small delay, see "the special problem of being too
1794 early", below). If multiple timers become ready during the same loop
1795 iteration then the ones with earlier time-out values are invoked before
1796 ones of the same priority with later time-out values (but this is no
1797 longer true when a callback calls C<ev_run> recursively).
1799 =head3 Be smart about timeouts
1801 Many real-world problems involve some kind of timeout, usually for error
1802 recovery. A typical example is an HTTP request - if the other side hangs,
1803 you want to raise some error after a while.
1805 What follows are some ways to handle this problem, from obvious and
1806 inefficient to smart and efficient.
1808 In the following, a 60 second activity timeout is assumed - a timeout that
1809 gets reset to 60 seconds each time there is activity (e.g. each time some
1810 data or other life sign was received).
1812 =over 4
1814 =item 1. Use a timer and stop, reinitialise and start it on activity.
1816 This is the most obvious, but not the most simple way: In the beginning,
1817 start the watcher:
1819 ev_timer_init (timer, callback, 60., 0.);
1820 ev_timer_start (loop, timer);
1822 Then, each time there is some activity, C<ev_timer_stop> it, initialise it
1823 and start it again:
1825 ev_timer_stop (loop, timer);
1826 ev_timer_set (timer, 60., 0.);
1827 ev_timer_start (loop, timer);
1829 This is relatively simple to implement, but means that each time there is
1830 some activity, libev will first have to remove the timer from its internal
1831 data structure and then add it again. Libev tries to be fast, but it's
1832 still not a constant-time operation.
1834 =item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
1836 This is the easiest way, and involves using C<ev_timer_again> instead of
1837 C<ev_timer_start>.
1839 To implement this, configure an C<ev_timer> with a C<repeat> value
1840 of C<60> and then call C<ev_timer_again> at start and each time you
1841 successfully read or write some data. If you go into an idle state where
1842 you do not expect data to travel on the socket, you can C<ev_timer_stop>
1843 the timer, and C<ev_timer_again> will automatically restart it if need be.
1845 That means you can ignore both the C<ev_timer_start> function and the
1846 C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
1847 member and C<ev_timer_again>.
1849 At start:
1851 ev_init (timer, callback);
1852 timer->repeat = 60.;
1853 ev_timer_again (loop, timer);
1855 Each time there is some activity:
1857 ev_timer_again (loop, timer);
1859 It is even possible to change the time-out on the fly, regardless of
1860 whether the watcher is active or not:
1862 timer->repeat = 30.;
1863 ev_timer_again (loop, timer);
1865 This is slightly more efficient then stopping/starting the timer each time
1866 you want to modify its timeout value, as libev does not have to completely
1867 remove and re-insert the timer from/into its internal data structure.
1869 It is, however, even simpler than the "obvious" way to do it.
1871 =item 3. Let the timer time out, but then re-arm it as required.
1873 This method is more tricky, but usually most efficient: Most timeouts are
1874 relatively long compared to the intervals between other activity - in
1875 our example, within 60 seconds, there are usually many I/O events with
1876 associated activity resets.
1878 In this case, it would be more efficient to leave the C<ev_timer> alone,
1879 but remember the time of last activity, and check for a real timeout only
1880 within the callback:
1882 ev_tstamp timeout = 60.;
1883 ev_tstamp last_activity; // time of last activity
1884 ev_timer timer;
1886 static void
1887 callback (EV_P_ ev_timer *w, int revents)
1888 {
1889 // calculate when the timeout would happen
1890 ev_tstamp after = last_activity - ev_now (EV_A) + timeout;
1892 // if negative, it means we the timeout already occurred
1893 if (after < 0.)
1894 {
1895 // timeout occurred, take action
1896 }
1897 else
1898 {
1899 // callback was invoked, but there was some recent
1900 // activity. simply restart the timer to time out
1901 // after "after" seconds, which is the earliest time
1902 // the timeout can occur.
1903 ev_timer_set (w, after, 0.);
1904 ev_timer_start (EV_A_ w);
1905 }
1906 }
1908 To summarise the callback: first calculate in how many seconds the
1909 timeout will occur (by calculating the absolute time when it would occur,
1910 C<last_activity + timeout>, and subtracting the current time, C<ev_now
1911 (EV_A)> from that).
1913 If this value is negative, then we are already past the timeout, i.e. we
1914 timed out, and need to do whatever is needed in this case.
1916 Otherwise, we now the earliest time at which the timeout would trigger,
1917 and simply start the timer with this timeout value.
1919 In other words, each time the callback is invoked it will check whether
1920 the timeout occurred. If not, it will simply reschedule itself to check
1921 again at the earliest time it could time out. Rinse. Repeat.
1923 This scheme causes more callback invocations (about one every 60 seconds
1924 minus half the average time between activity), but virtually no calls to
1925 libev to change the timeout.
1927 To start the machinery, simply initialise the watcher and set
1928 C<last_activity> to the current time (meaning there was some activity just
1929 now), then call the callback, which will "do the right thing" and start
1930 the timer:
1932 last_activity = ev_now (EV_A);
1933 ev_init (&timer, callback);
1934 callback (EV_A_ &timer, 0);
1936 When there is some activity, simply store the current time in
1937 C<last_activity>, no libev calls at all:
1939 if (activity detected)
1940 last_activity = ev_now (EV_A);
1942 When your timeout value changes, then the timeout can be changed by simply
1943 providing a new value, stopping the timer and calling the callback, which
1944 will again do the right thing (for example, time out immediately :).
1946 timeout = new_value;
1947 ev_timer_stop (EV_A_ &timer);
1948 callback (EV_A_ &timer, 0);
1950 This technique is slightly more complex, but in most cases where the
1951 time-out is unlikely to be triggered, much more efficient.
1953 =item 4. Wee, just use a double-linked list for your timeouts.
1955 If there is not one request, but many thousands (millions...), all
1956 employing some kind of timeout with the same timeout value, then one can
1957 do even better:
1959 When starting the timeout, calculate the timeout value and put the timeout
1960 at the I<end> of the list.
1962 Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
1963 the list is expected to fire (for example, using the technique #3).
1965 When there is some activity, remove the timer from the list, recalculate
1966 the timeout, append it to the end of the list again, and make sure to
1967 update the C<ev_timer> if it was taken from the beginning of the list.
1969 This way, one can manage an unlimited number of timeouts in O(1) time for
1970 starting, stopping and updating the timers, at the expense of a major
1971 complication, and having to use a constant timeout. The constant timeout
1972 ensures that the list stays sorted.
1974 =back
1976 So which method the best?
1978 Method #2 is a simple no-brain-required solution that is adequate in most
1979 situations. Method #3 requires a bit more thinking, but handles many cases
1980 better, and isn't very complicated either. In most case, choosing either
1981 one is fine, with #3 being better in typical situations.
1983 Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
1984 rather complicated, but extremely efficient, something that really pays
1985 off after the first million or so of active timers, i.e. it's usually
1986 overkill :)
1988 =head3 The special problem of being too early
1990 If you ask a timer to call your callback after three seconds, then
1991 you expect it to be invoked after three seconds - but of course, this
1992 cannot be guaranteed to infinite precision. Less obviously, it cannot be
1993 guaranteed to any precision by libev - imagine somebody suspending the
1994 process with a STOP signal for a few hours for example.
1996 So, libev tries to invoke your callback as soon as possible I<after> the
1997 delay has occurred, but cannot guarantee this.
1999 A less obvious failure mode is calling your callback too early: many event
2000 loops compare timestamps with a "elapsed delay >= requested delay", but
2001 this can cause your callback to be invoked much earlier than you would
2002 expect.
2004 To see why, imagine a system with a clock that only offers full second
2005 resolution (think windows if you can't come up with a broken enough OS
2006 yourself). If you schedule a one-second timer at the time 500.9, then the
2007 event loop will schedule your timeout to elapse at a system time of 500
2008 (500.9 truncated to the resolution) + 1, or 501.
2010 If an event library looks at the timeout 0.1s later, it will see "501 >=
2011 501" and invoke the callback 0.1s after it was started, even though a
2012 one-second delay was requested - this is being "too early", despite best
2013 intentions.
2015 This is the reason why libev will never invoke the callback if the elapsed
2016 delay equals the requested delay, but only when the elapsed delay is
2017 larger than the requested delay. In the example above, libev would only invoke
2018 the callback at system time 502, or 1.1s after the timer was started.
2020 So, while libev cannot guarantee that your callback will be invoked
2021 exactly when requested, it I<can> and I<does> guarantee that the requested
2022 delay has actually elapsed, or in other words, it always errs on the "too
2023 late" side of things.
2025 =head3 The special problem of time updates
2027 Establishing the current time is a costly operation (it usually takes
2028 at least one system call): EV therefore updates its idea of the current
2029 time only before and after C<ev_run> collects new events, which causes a
2030 growing difference between C<ev_now ()> and C<ev_time ()> when handling
2031 lots of events in one iteration.
2033 The relative timeouts are calculated relative to the C<ev_now ()>
2034 time. This is usually the right thing as this timestamp refers to the time
2035 of the event triggering whatever timeout you are modifying/starting. If
2036 you suspect event processing to be delayed and you I<need> to base the
2037 timeout on the current time, use something like the following to adjust
2038 for it:
2040 ev_timer_set (&timer, after + (ev_time () - ev_now ()), 0.);
2042 If the event loop is suspended for a long time, you can also force an
2043 update of the time returned by C<ev_now ()> by calling C<ev_now_update
2044 ()>, although that will push the event time of all outstanding events
2045 further into the future.
2047 =head3 The special problem of unsynchronised clocks
2049 Modern systems have a variety of clocks - libev itself uses the normal
2050 "wall clock" clock and, if available, the monotonic clock (to avoid time
2051 jumps).
2053 Neither of these clocks is synchronised with each other or any other clock
2054 on the system, so C<ev_time ()> might return a considerably different time
2055 than C<gettimeofday ()> or C<time ()>. On a GNU/Linux system, for example,
2056 a call to C<gettimeofday> might return a second count that is one higher
2057 than a directly following call to C<time>.
2059 The moral of this is to only compare libev-related timestamps with
2060 C<ev_time ()> and C<ev_now ()>, at least if you want better precision than
2061 a second or so.
2063 One more problem arises due to this lack of synchronisation: if libev uses
2064 the system monotonic clock and you compare timestamps from C<ev_time>
2065 or C<ev_now> from when you started your timer and when your callback is
2066 invoked, you will find that sometimes the callback is a bit "early".
2068 This is because C<ev_timer>s work in real time, not wall clock time, so
2069 libev makes sure your callback is not invoked before the delay happened,
2070 I<measured according to the real time>, not the system clock.
2072 If your timeouts are based on a physical timescale (e.g. "time out this
2073 connection after 100 seconds") then this shouldn't bother you as it is
2074 exactly the right behaviour.
2076 If you want to compare wall clock/system timestamps to your timers, then
2077 you need to use C<ev_periodic>s, as these are based on the wall clock
2078 time, where your comparisons will always generate correct results.
2080 =head3 The special problems of suspended animation
2082 When you leave the server world it is quite customary to hit machines that
2083 can suspend/hibernate - what happens to the clocks during such a suspend?
2085 Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
2086 all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
2087 to run until the system is suspended, but they will not advance while the
2088 system is suspended. That means, on resume, it will be as if the program
2089 was frozen for a few seconds, but the suspend time will not be counted
2090 towards C<ev_timer> when a monotonic clock source is used. The real time
2091 clock advanced as expected, but if it is used as sole clocksource, then a
2092 long suspend would be detected as a time jump by libev, and timers would
2093 be adjusted accordingly.
2095 I would not be surprised to see different behaviour in different between
2096 operating systems, OS versions or even different hardware.
2098 The other form of suspend (job control, or sending a SIGSTOP) will see a
2099 time jump in the monotonic clocks and the realtime clock. If the program
2100 is suspended for a very long time, and monotonic clock sources are in use,
2101 then you can expect C<ev_timer>s to expire as the full suspension time
2102 will be counted towards the timers. When no monotonic clock source is in
2103 use, then libev will again assume a timejump and adjust accordingly.
2105 It might be beneficial for this latter case to call C<ev_suspend>
2106 and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
2107 deterministic behaviour in this case (you can do nothing against
2108 C<SIGSTOP>).
2110 =head3 Watcher-Specific Functions and Data Members
2112 =over 4
2114 =item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
2116 =item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
2118 Configure the timer to trigger after C<after> seconds. If C<repeat>
2119 is C<0.>, then it will automatically be stopped once the timeout is
2120 reached. If it is positive, then the timer will automatically be
2121 configured to trigger again C<repeat> seconds later, again, and again,
2122 until stopped manually.
2124 The timer itself will do a best-effort at avoiding drift, that is, if
2125 you configure a timer to trigger every 10 seconds, then it will normally
2126 trigger at exactly 10 second intervals. If, however, your program cannot
2127 keep up with the timer (because it takes longer than those 10 seconds to
2128 do stuff) the timer will not fire more than once per event loop iteration.
2130 =item ev_timer_again (loop, ev_timer *)
2132 This will act as if the timer timed out, and restarts it again if it is
2133 repeating. It basically works like calling C<ev_timer_stop>, updating the
2134 timeout to the C<repeat> value and calling C<ev_timer_start>.
2136 The exact semantics are as in the following rules, all of which will be
2137 applied to the watcher:
2139 =over 4
2141 =item If the timer is pending, the pending status is always cleared.
2143 =item If the timer is started but non-repeating, stop it (as if it timed
2144 out, without invoking it).
2146 =item If the timer is repeating, make the C<repeat> value the new timeout
2147 and start the timer, if necessary.
2149 =back
2151 This sounds a bit complicated, see L</Be smart about timeouts>, above, for a
2152 usage example.
2154 =item ev_tstamp ev_timer_remaining (loop, ev_timer *)
2156 Returns the remaining time until a timer fires. If the timer is active,
2157 then this time is relative to the current event loop time, otherwise it's
2158 the timeout value currently configured.
2160 That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
2161 C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
2162 will return C<4>. When the timer expires and is restarted, it will return
2163 roughly C<7> (likely slightly less as callback invocation takes some time,
2164 too), and so on.
2166 =item ev_tstamp repeat [read-write]
2168 The current C<repeat> value. Will be used each time the watcher times out
2169 or C<ev_timer_again> is called, and determines the next timeout (if any),
2170 which is also when any modifications are taken into account.
2172 =back
2174 =head3 Examples
2176 Example: Create a timer that fires after 60 seconds.
2178 static void
2179 one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
2180 {
2181 .. one minute over, w is actually stopped right here
2182 }
2184 ev_timer mytimer;
2185 ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
2186 ev_timer_start (loop, &mytimer);
2188 Example: Create a timeout timer that times out after 10 seconds of
2189 inactivity.
2191 static void
2192 timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
2193 {
2194 .. ten seconds without any activity
2195 }
2197 ev_timer mytimer;
2198 ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
2199 ev_timer_again (&mytimer); /* start timer */
2200 ev_run (loop, 0);
2202 // and in some piece of code that gets executed on any "activity":
2203 // reset the timeout to start ticking again at 10 seconds
2204 ev_timer_again (&mytimer);
2207 =head2 C<ev_periodic> - to cron or not to cron?
2209 Periodic watchers are also timers of a kind, but they are very versatile
2210 (and unfortunately a bit complex).
2212 Unlike C<ev_timer>, periodic watchers are not based on real time (or
2213 relative time, the physical time that passes) but on wall clock time
2214 (absolute time, the thing you can read on your calender or clock). The
2215 difference is that wall clock time can run faster or slower than real
2216 time, and time jumps are not uncommon (e.g. when you adjust your
2217 wrist-watch).
2219 You can tell a periodic watcher to trigger after some specific point
2220 in time: for example, if you tell a periodic watcher to trigger "in 10
2221 seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
2222 not a delay) and then reset your system clock to January of the previous
2223 year, then it will take a year or more to trigger the event (unlike an
2224 C<ev_timer>, which would still trigger roughly 10 seconds after starting
2225 it, as it uses a relative timeout).
2227 C<ev_periodic> watchers can also be used to implement vastly more complex
2228 timers, such as triggering an event on each "midnight, local time", or
2229 other complicated rules. This cannot be done with C<ev_timer> watchers, as
2230 those cannot react to time jumps.
2232 As with timers, the callback is guaranteed to be invoked only when the
2233 point in time where it is supposed to trigger has passed. If multiple
2234 timers become ready during the same loop iteration then the ones with
2235 earlier time-out values are invoked before ones with later time-out values
2236 (but this is no longer true when a callback calls C<ev_run> recursively).
2238 =head3 Watcher-Specific Functions and Data Members
2240 =over 4
2242 =item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
2244 =item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
2246 Lots of arguments, let's sort it out... There are basically three modes of
2247 operation, and we will explain them from simplest to most complex:
2249 =over 4
2251 =item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
2253 In this configuration the watcher triggers an event after the wall clock
2254 time C<offset> has passed. It will not repeat and will not adjust when a
2255 time jump occurs, that is, if it is to be run at January 1st 2011 then it
2256 will be stopped and invoked when the system clock reaches or surpasses
2257 this point in time.
2259 =item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
2261 In this mode the watcher will always be scheduled to time out at the next
2262 C<offset + N * interval> time (for some integer N, which can also be
2263 negative) and then repeat, regardless of any time jumps. The C<offset>
2264 argument is merely an offset into the C<interval> periods.
2266 This can be used to create timers that do not drift with respect to the
2267 system clock, for example, here is an C<ev_periodic> that triggers each
2268 hour, on the hour (with respect to UTC):
2270 ev_periodic_set (&periodic, 0., 3600., 0);
2272 This doesn't mean there will always be 3600 seconds in between triggers,
2273 but only that the callback will be called when the system time shows a
2274 full hour (UTC), or more correctly, when the system time is evenly divisible
2275 by 3600.
2277 Another way to think about it (for the mathematically inclined) is that
2278 C<ev_periodic> will try to run the callback in this mode at the next possible
2279 time where C<time = offset (mod interval)>, regardless of any time jumps.
2281 The C<interval> I<MUST> be positive, and for numerical stability, the
2282 interval value should be higher than C<1/8192> (which is around 100
2283 microseconds) and C<offset> should be higher than C<0> and should have
2284 at most a similar magnitude as the current time (say, within a factor of
2285 ten). Typical values for offset are, in fact, C<0> or something between
2286 C<0> and C<interval>, which is also the recommended range.
2288 Note also that there is an upper limit to how often a timer can fire (CPU
2289 speed for example), so if C<interval> is very small then timing stability
2290 will of course deteriorate. Libev itself tries to be exact to be about one
2291 millisecond (if the OS supports it and the machine is fast enough).
2293 =item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
2295 In this mode the values for C<interval> and C<offset> are both being
2296 ignored. Instead, each time the periodic watcher gets scheduled, the
2297 reschedule callback will be called with the watcher as first, and the
2298 current time as second argument.
2300 NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
2301 or make ANY other event loop modifications whatsoever, unless explicitly
2302 allowed by documentation here>.
2304 If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
2305 it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
2306 only event loop modification you are allowed to do).
2308 The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
2309 *w, ev_tstamp now)>, e.g.:
2311 static ev_tstamp
2312 my_rescheduler (ev_periodic *w, ev_tstamp now)
2313 {
2314 return now + 60.;
2315 }
2317 It must return the next time to trigger, based on the passed time value
2318 (that is, the lowest time value larger than to the second argument). It
2319 will usually be called just before the callback will be triggered, but
2320 might be called at other times, too.
2322 NOTE: I<< This callback must always return a time that is higher than or
2323 equal to the passed C<now> value >>.
2325 This can be used to create very complex timers, such as a timer that
2326 triggers on "next midnight, local time". To do this, you would calculate the
2327 next midnight after C<now> and return the timestamp value for this. How
2328 you do this is, again, up to you (but it is not trivial, which is the main
2329 reason I omitted it as an example).
2331 =back
2333 =item ev_periodic_again (loop, ev_periodic *)
2335 Simply stops and restarts the periodic watcher again. This is only useful
2336 when you changed some parameters or the reschedule callback would return
2337 a different time than the last time it was called (e.g. in a crond like
2338 program when the crontabs have changed).
2340 =item ev_tstamp ev_periodic_at (ev_periodic *)
2342 When active, returns the absolute time that the watcher is supposed
2343 to trigger next. This is not the same as the C<offset> argument to
2344 C<ev_periodic_set>, but indeed works even in interval and manual
2345 rescheduling modes.
2347 =item ev_tstamp offset [read-write]
2349 When repeating, this contains the offset value, otherwise this is the
2350 absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
2351 although libev might modify this value for better numerical stability).
2353 Can be modified any time, but changes only take effect when the periodic
2354 timer fires or C<ev_periodic_again> is being called.
2356 =item ev_tstamp interval [read-write]
2358 The current interval value. Can be modified any time, but changes only
2359 take effect when the periodic timer fires or C<ev_periodic_again> is being
2360 called.
2362 =item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
2364 The current reschedule callback, or C<0>, if this functionality is
2365 switched off. Can be changed any time, but changes only take effect when
2366 the periodic timer fires or C<ev_periodic_again> is being called.
2368 =back
2370 =head3 Examples
2372 Example: Call a callback every hour, or, more precisely, whenever the
2373 system time is divisible by 3600. The callback invocation times have
2374 potentially a lot of jitter, but good long-term stability.
2376 static void
2377 clock_cb (struct ev_loop *loop, ev_periodic *w, int revents)
2378 {
2379 ... its now a full hour (UTC, or TAI or whatever your clock follows)
2380 }
2382 ev_periodic hourly_tick;
2383 ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
2384 ev_periodic_start (loop, &hourly_tick);
2386 Example: The same as above, but use a reschedule callback to do it:
2388 #include <math.h>
2390 static ev_tstamp
2391 my_scheduler_cb (ev_periodic *w, ev_tstamp now)
2392 {
2393 return now + (3600. - fmod (now, 3600.));
2394 }
2396 ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
2398 Example: Call a callback every hour, starting now:
2400 ev_periodic hourly_tick;
2401 ev_periodic_init (&hourly_tick, clock_cb,
2402 fmod (ev_now (loop), 3600.), 3600., 0);
2403 ev_periodic_start (loop, &hourly_tick);
2406 =head2 C<ev_signal> - signal me when a signal gets signalled!
2408 Signal watchers will trigger an event when the process receives a specific
2409 signal one or more times. Even though signals are very asynchronous, libev
2410 will try its best to deliver signals synchronously, i.e. as part of the
2411 normal event processing, like any other event.
2413 If you want signals to be delivered truly asynchronously, just use
2414 C<sigaction> as you would do without libev and forget about sharing
2415 the signal. You can even use C<ev_async> from a signal handler to
2416 synchronously wake up an event loop.
2418 You can configure as many watchers as you like for the same signal, but
2419 only within the same loop, i.e. you can watch for C<SIGINT> in your
2420 default loop and for C<SIGIO> in another loop, but you cannot watch for
2421 C<SIGINT> in both the default loop and another loop at the same time. At
2422 the moment, C<SIGCHLD> is permanently tied to the default loop.
2424 Only after the first watcher for a signal is started will libev actually
2425 register something with the kernel. It thus coexists with your own signal
2426 handlers as long as you don't register any with libev for the same signal.
2428 If possible and supported, libev will install its handlers with
2429 C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
2430 not be unduly interrupted. If you have a problem with system calls getting
2431 interrupted by signals you can block all signals in an C<ev_check> watcher
2432 and unblock them in an C<ev_prepare> watcher.
2434 =head3 The special problem of inheritance over fork/execve/pthread_create
2436 Both the signal mask (C<sigprocmask>) and the signal disposition
2437 (C<sigaction>) are unspecified after starting a signal watcher (and after
2438 stopping it again), that is, libev might or might not block the signal,
2439 and might or might not set or restore the installed signal handler (but
2442 While this does not matter for the signal disposition (libev never
2443 sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
2444 C<execve>), this matters for the signal mask: many programs do not expect
2445 certain signals to be blocked.
2447 This means that before calling C<exec> (from the child) you should reset
2448 the signal mask to whatever "default" you expect (all clear is a good
2449 choice usually).
2451 The simplest way to ensure that the signal mask is reset in the child is
2452 to install a fork handler with C<pthread_atfork> that resets it. That will
2453 catch fork calls done by libraries (such as the libc) as well.
2455 In current versions of libev, the signal will not be blocked indefinitely
2456 unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
2457 the window of opportunity for problems, it will not go away, as libev
2458 I<has> to modify the signal mask, at least temporarily.
2460 So I can't stress this enough: I<If you do not reset your signal mask when
2461 you expect it to be empty, you have a race condition in your code>. This
2462 is not a libev-specific thing, this is true for most event libraries.
2464 =head3 The special problem of threads signal handling
2466 POSIX threads has problematic signal handling semantics, specifically,
2467 a lot of functionality (sigfd, sigwait etc.) only really works if all
2468 threads in a process block signals, which is hard to achieve.
2470 When you want to use sigwait (or mix libev signal handling with your own
2471 for the same signals), you can tackle this problem by globally blocking
2472 all signals before creating any threads (or creating them with a fully set
2473 sigprocmask) and also specifying the C<EVFLAG_NOSIGMASK> when creating
2474 loops. Then designate one thread as "signal receiver thread" which handles
2475 these signals. You can pass on any signals that libev might be interested
2476 in by calling C<ev_feed_signal>.
2478 =head3 Watcher-Specific Functions and Data Members
2480 =over 4
2482 =item ev_signal_init (ev_signal *, callback, int signum)
2484 =item ev_signal_set (ev_signal *, int signum)
2486 Configures the watcher to trigger on the given signal number (usually one
2487 of the C<SIGxxx> constants).
2489 =item int signum [read-only]
2491 The signal the watcher watches out for.
2493 =back
2495 =head3 Examples
2497 Example: Try to exit cleanly on SIGINT.
2499 static void
2500 sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
2501 {
2502 ev_break (loop, EVBREAK_ALL);
2503 }
2505 ev_signal signal_watcher;
2506 ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
2507 ev_signal_start (loop, &signal_watcher);
2510 =head2 C<ev_child> - watch out for process status changes
2512 Child watchers trigger when your process receives a SIGCHLD in response to
2513 some child status changes (most typically when a child of yours dies or
2514 exits). It is permissible to install a child watcher I<after> the child
2515 has been forked (which implies it might have already exited), as long
2516 as the event loop isn't entered (or is continued from a watcher), i.e.,
2517 forking and then immediately registering a watcher for the child is fine,
2518 but forking and registering a watcher a few event loop iterations later or
2519 in the next callback invocation is not.
2521 Only the default event loop is capable of handling signals, and therefore
2522 you can only register child watchers in the default event loop.
2524 Due to some design glitches inside libev, child watchers will always be
2525 handled at maximum priority (their priority is set to C<EV_MAXPRI> by
2526 libev)
2528 =head3 Process Interaction
2530 Libev grabs C<SIGCHLD> as soon as the default event loop is
2531 initialised. This is necessary to guarantee proper behaviour even if the
2532 first child watcher is started after the child exits. The occurrence
2533 of C<SIGCHLD> is recorded asynchronously, but child reaping is done
2534 synchronously as part of the event loop processing. Libev always reaps all
2535 children, even ones not watched.
2537 =head3 Overriding the Built-In Processing
2539 Libev offers no special support for overriding the built-in child
2540 processing, but if your application collides with libev's default child
2541 handler, you can override it easily by installing your own handler for
2542 C<SIGCHLD> after initialising the default loop, and making sure the
2543 default loop never gets destroyed. You are encouraged, however, to use an
2544 event-based approach to child reaping and thus use libev's support for
2545 that, so other libev users can use C<ev_child> watchers freely.
2547 =head3 Stopping the Child Watcher
2549 Currently, the child watcher never gets stopped, even when the
2550 child terminates, so normally one needs to stop the watcher in the
2551 callback. Future versions of libev might stop the watcher automatically
2552 when a child exit is detected (calling C<ev_child_stop> twice is not a
2553 problem).
2555 =head3 Watcher-Specific Functions and Data Members
2557 =over 4
2559 =item ev_child_init (ev_child *, callback, int pid, int trace)
2561 =item ev_child_set (ev_child *, int pid, int trace)
2563 Configures the watcher to wait for status changes of process C<pid> (or
2564 I<any> process if C<pid> is specified as C<0>). The callback can look
2565 at the C<rstatus> member of the C<ev_child> watcher structure to see
2566 the status word (use the macros from C<sys/wait.h> and see your systems
2567 C<waitpid> documentation). The C<rpid> member contains the pid of the
2568 process causing the status change. C<trace> must be either C<0> (only
2569 activate the watcher when the process terminates) or C<1> (additionally
2570 activate the watcher when the process is stopped or continued).
2572 =item int pid [read-only]
2574 The process id this watcher watches out for, or C<0>, meaning any process id.
2576 =item int rpid [read-write]
2578 The process id that detected a status change.
2580 =item int rstatus [read-write]
2582 The process exit/trace status caused by C<rpid> (see your systems
2583 C<waitpid> and C<sys/wait.h> documentation for details).
2585 =back
2587 =head3 Examples
2589 Example: C<fork()> a new process and install a child handler to wait for
2590 its completion.
2592 ev_child cw;
2594 static void
2595 child_cb (EV_P_ ev_child *w, int revents)
2596 {
2597 ev_child_stop (EV_A_ w);
2598 printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
2599 }
2601 pid_t pid = fork ();
2603 if (pid < 0)
2604 // error
2605 else if (pid == 0)
2606 {
2607 // the forked child executes here
2608 exit (1);
2609 }
2610 else
2611 {
2612 ev_child_init (&cw, child_cb, pid, 0);
2613 ev_child_start (EV_DEFAULT_ &cw);
2614 }
2617 =head2 C<ev_stat> - did the file attributes just change?
2619 This watches a file system path for attribute changes. That is, it calls
2620 C<stat> on that path in regular intervals (or when the OS says it changed)
2621 and sees if it changed compared to the last time, invoking the callback
2622 if it did. Starting the watcher C<stat>'s the file, so only changes that
2623 happen after the watcher has been started will be reported.
2625 The path does not need to exist: changing from "path exists" to "path does
2626 not exist" is a status change like any other. The condition "path does not
2627 exist" (or more correctly "path cannot be stat'ed") is signified by the
2628 C<st_nlink> field being zero (which is otherwise always forced to be at
2629 least one) and all the other fields of the stat buffer having unspecified
2630 contents.
2632 The path I<must not> end in a slash or contain special components such as
2633 C<.> or C<..>. The path I<should> be absolute: If it is relative and
2634 your working directory changes, then the behaviour is undefined.
2636 Since there is no portable change notification interface available, the
2637 portable implementation simply calls C<stat(2)> regularly on the path
2638 to see if it changed somehow. You can specify a recommended polling
2639 interval for this case. If you specify a polling interval of C<0> (highly
2640 recommended!) then a I<suitable, unspecified default> value will be used
2641 (which you can expect to be around five seconds, although this might
2642 change dynamically). Libev will also impose a minimum interval which is
2643 currently around C<0.1>, but that's usually overkill.
2645 This watcher type is not meant for massive numbers of stat watchers,
2646 as even with OS-supported change notifications, this can be
2647 resource-intensive.
2649 At the time of this writing, the only OS-specific interface implemented
2650 is the Linux inotify interface (implementing kqueue support is left as an
2651 exercise for the reader. Note, however, that the author sees no way of
2652 implementing C<ev_stat> semantics with kqueue, except as a hint).
2654 =head3 ABI Issues (Largefile Support)
2656 Libev by default (unless the user overrides this) uses the default
2657 compilation environment, which means that on systems with large file
2658 support disabled by default, you get the 32 bit version of the stat
2659 structure. When using the library from programs that change the ABI to
2660 use 64 bit file offsets the programs will fail. In that case you have to
2661 compile libev with the same flags to get binary compatibility. This is
2662 obviously the case with any flags that change the ABI, but the problem is
2663 most noticeably displayed with ev_stat and large file support.
2665 The solution for this is to lobby your distribution maker to make large
2666 file interfaces available by default (as e.g. FreeBSD does) and not
2667 optional. Libev cannot simply switch on large file support because it has
2668 to exchange stat structures with application programs compiled using the
2669 default compilation environment.
2671 =head3 Inotify and Kqueue
2673 When C<inotify (7)> support has been compiled into libev and present at
2674 runtime, it will be used to speed up change detection where possible. The
2675 inotify descriptor will be created lazily when the first C<ev_stat>
2676 watcher is being started.
2678 Inotify presence does not change the semantics of C<ev_stat> watchers
2679 except that changes might be detected earlier, and in some cases, to avoid
2680 making regular C<stat> calls. Even in the presence of inotify support
2681 there are many cases where libev has to resort to regular C<stat> polling,
2682 but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
2683 many bugs), the path exists (i.e. stat succeeds), and the path resides on
2684 a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
2685 xfs are fully working) libev usually gets away without polling.
2687 There is no support for kqueue, as apparently it cannot be used to
2688 implement this functionality, due to the requirement of having a file
2689 descriptor open on the object at all times, and detecting renames, unlinks
2690 etc. is difficult.
2692 =head3 C<stat ()> is a synchronous operation
2694 Libev doesn't normally do any kind of I/O itself, and so is not blocking
2695 the process. The exception are C<ev_stat> watchers - those call C<stat
2696 ()>, which is a synchronous operation.
2698 For local paths, this usually doesn't matter: unless the system is very
2699 busy or the intervals between stat's are large, a stat call will be fast,
2700 as the path data is usually in memory already (except when starting the
2701 watcher).
2703 For networked file systems, calling C<stat ()> can block an indefinite
2704 time due to network issues, and even under good conditions, a stat call
2705 often takes multiple milliseconds.
2707 Therefore, it is best to avoid using C<ev_stat> watchers on networked
2708 paths, although this is fully supported by libev.
2710 =head3 The special problem of stat time resolution
2712 The C<stat ()> system call only supports full-second resolution portably,
2713 and even on systems where the resolution is higher, most file systems
2714 still only support whole seconds.
2716 That means that, if the time is the only thing that changes, you can
2717 easily miss updates: on the first update, C<ev_stat> detects a change and
2718 calls your callback, which does something. When there is another update
2719 within the same second, C<ev_stat> will be unable to detect unless the
2720 stat data does change in other ways (e.g. file size).
2722 The solution to this is to delay acting on a change for slightly more
2723 than a second (or till slightly after the next full second boundary), using
2724 a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
2725 ev_timer_again (loop, w)>).
2727 The C<.02> offset is added to work around small timing inconsistencies
2728 of some operating systems (where the second counter of the current time
2729 might be be delayed. One such system is the Linux kernel, where a call to
2730 C<gettimeofday> might return a timestamp with a full second later than
2731 a subsequent C<time> call - if the equivalent of C<time ()> is used to
2732 update file times then there will be a small window where the kernel uses
2733 the previous second to update file times but libev might already execute
2734 the timer callback).
2736 =head3 Watcher-Specific Functions and Data Members
2738 =over 4
2740 =item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
2742 =item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
2744 Configures the watcher to wait for status changes of the given
2745 C<path>. The C<interval> is a hint on how quickly a change is expected to
2746 be detected and should normally be specified as C<0> to let libev choose
2747 a suitable value. The memory pointed to by C<path> must point to the same
2748 path for as long as the watcher is active.
2750 The callback will receive an C<EV_STAT> event when a change was detected,
2751 relative to the attributes at the time the watcher was started (or the
2752 last change was detected).
2754 =item ev_stat_stat (loop, ev_stat *)
2756 Updates the stat buffer immediately with new values. If you change the
2757 watched path in your callback, you could call this function to avoid
2758 detecting this change (while introducing a race condition if you are not
2759 the only one changing the path). Can also be useful simply to find out the
2760 new values.
2762 =item ev_statdata attr [read-only]
2764 The most-recently detected attributes of the file. Although the type is
2765 C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
2766 suitable for your system, but you can only rely on the POSIX-standardised
2767 members to be present. If the C<st_nlink> member is C<0>, then there was
2768 some error while C<stat>ing the file.
2770 =item ev_statdata prev [read-only]
2772 The previous attributes of the file. The callback gets invoked whenever
2773 C<prev> != C<attr>, or, more precisely, one or more of these members
2774 differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
2775 C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
2777 =item ev_tstamp interval [read-only]
2779 The specified interval.
2781 =item const char *path [read-only]
2783 The file system path that is being watched.
2785 =back
2787 =head3 Examples
2789 Example: Watch C</etc/passwd> for attribute changes.
2791 static void
2792 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
2793 {
2794 /* /etc/passwd changed in some way */
2795 if (w->attr.st_nlink)
2796 {
2797 printf ("passwd current size %ld\n", (long)w->attr.st_size);
2798 printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
2799 printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
2800 }
2801 else
2802 /* you shalt not abuse printf for puts */
2803 puts ("wow, /etc/passwd is not there, expect problems. "
2804 "if this is windows, they already arrived\n");
2805 }
2807 ...
2808 ev_stat passwd;
2810 ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
2811 ev_stat_start (loop, &passwd);
2813 Example: Like above, but additionally use a one-second delay so we do not
2814 miss updates (however, frequent updates will delay processing, too, so
2815 one might do the work both on C<ev_stat> callback invocation I<and> on
2816 C<ev_timer> callback invocation).
2818 static ev_stat passwd;
2819 static ev_timer timer;
2821 static void
2822 timer_cb (EV_P_ ev_timer *w, int revents)
2823 {
2824 ev_timer_stop (EV_A_ w);
2826 /* now it's one second after the most recent passwd change */
2827 }
2829 static void
2830 stat_cb (EV_P_ ev_stat *w, int revents)
2831 {
2832 /* reset the one-second timer */
2833 ev_timer_again (EV_A_ &timer);
2834 }
2836 ...
2837 ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
2838 ev_stat_start (loop, &passwd);
2839 ev_timer_init (&timer, timer_cb, 0., 1.02);
2842 =head2 C<ev_idle> - when you've got nothing better to do...
2844 Idle watchers trigger events when no other events of the same or higher
2845 priority are pending (prepare, check and other idle watchers do not count
2846 as receiving "events").
2848 That is, as long as your process is busy handling sockets or timeouts
2849 (or even signals, imagine) of the same or higher priority it will not be
2850 triggered. But when your process is idle (or only lower-priority watchers
2851 are pending), the idle watchers are being called once per event loop
2852 iteration - until stopped, that is, or your process receives more events
2853 and becomes busy again with higher priority stuff.
2855 The most noteworthy effect is that as long as any idle watchers are
2856 active, the process will not block when waiting for new events.
2858 Apart from keeping your process non-blocking (which is a useful
2859 effect on its own sometimes), idle watchers are a good place to do
2860 "pseudo-background processing", or delay processing stuff to after the
2861 event loop has handled all outstanding events.
2863 =head3 Abusing an C<ev_idle> watcher for its side-effect
2865 As long as there is at least one active idle watcher, libev will never
2866 sleep unnecessarily. Or in other words, it will loop as fast as possible.
2867 For this to work, the idle watcher doesn't need to be invoked at all - the
2868 lowest priority will do.
2870 This mode of operation can be useful together with an C<ev_check> watcher,
2871 to do something on each event loop iteration - for example to balance load
2872 between different connections.
2874 See L</Abusing an ev_check watcher for its side-effect> for a longer
2875 example.
2877 =head3 Watcher-Specific Functions and Data Members
2879 =over 4
2881 =item ev_idle_init (ev_idle *, callback)
2883 Initialises and configures the idle watcher - it has no parameters of any
2884 kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
2885 believe me.
2887 =back
2889 =head3 Examples
2891 Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
2892 callback, free it. Also, use no error checking, as usual.
2894 static void
2895 idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
2896 {
2897 // stop the watcher
2898 ev_idle_stop (loop, w);
2900 // now we can free it
2901 free (w);
2903 // now do something you wanted to do when the program has
2904 // no longer anything immediate to do.
2905 }
2907 ev_idle *idle_watcher = malloc (sizeof (ev_idle));
2908 ev_idle_init (idle_watcher, idle_cb);
2909 ev_idle_start (loop, idle_watcher);
2912 =head2 C<ev_prepare> and C<ev_check> - customise your event loop!
2914 Prepare and check watchers are often (but not always) used in pairs:
2915 prepare watchers get invoked before the process blocks and check watchers
2916 afterwards.
2918 You I<must not> call C<ev_run> (or similar functions that enter the
2919 current event loop) or C<ev_loop_fork> from either C<ev_prepare> or
2920 C<ev_check> watchers. Other loops than the current one are fine,
2921 however. The rationale behind this is that you do not need to check
2922 for recursion in those watchers, i.e. the sequence will always be
2923 C<ev_prepare>, blocking, C<ev_check> so if you have one watcher of each
2924 kind they will always be called in pairs bracketing the blocking call.
2926 Their main purpose is to integrate other event mechanisms into libev and
2927 their use is somewhat advanced. They could be used, for example, to track
2928 variable changes, implement your own watchers, integrate net-snmp or a
2929 coroutine library and lots more. They are also occasionally useful if
2930 you cache some data and want to flush it before blocking (for example,
2931 in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
2932 watcher).
2934 This is done by examining in each prepare call which file descriptors
2935 need to be watched by the other library, registering C<ev_io> watchers
2936 for them and starting an C<ev_timer> watcher for any timeouts (many
2937 libraries provide exactly this functionality). Then, in the check watcher,
2938 you check for any events that occurred (by checking the pending status
2939 of all watchers and stopping them) and call back into the library. The
2940 I/O and timer callbacks will never actually be called (but must be valid
2941 nevertheless, because you never know, you know?).
2943 As another example, the Perl Coro module uses these hooks to integrate
2944 coroutines into libev programs, by yielding to other active coroutines
2945 during each prepare and only letting the process block if no coroutines
2946 are ready to run (it's actually more complicated: it only runs coroutines
2947 with priority higher than or equal to the event loop and one coroutine
2948 of lower priority, but only once, using idle watchers to keep the event
2949 loop from blocking if lower-priority coroutines are active, thus mapping
2950 low-priority coroutines to idle/background tasks).
2952 When used for this purpose, it is recommended to give C<ev_check> watchers
2953 highest (C<EV_MAXPRI>) priority, to ensure that they are being run before
2954 any other watchers after the poll (this doesn't matter for C<ev_prepare>
2955 watchers).
2957 Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
2958 activate ("feed") events into libev. While libev fully supports this, they
2959 might get executed before other C<ev_check> watchers did their job. As
2960 C<ev_check> watchers are often used to embed other (non-libev) event
2961 loops those other event loops might be in an unusable state until their
2962 C<ev_check> watcher ran (always remind yourself to coexist peacefully with
2963 others).
2965 =head3 Abusing an C<ev_check> watcher for its side-effect
2967 C<ev_check> (and less often also C<ev_prepare>) watchers can also be
2968 useful because they are called once per event loop iteration. For
2969 example, if you want to handle a large number of connections fairly, you
2970 normally only do a bit of work for each active connection, and if there
2971 is more work to do, you wait for the next event loop iteration, so other
2972 connections have a chance of making progress.
2974 Using an C<ev_check> watcher is almost enough: it will be called on the
2975 next event loop iteration. However, that isn't as soon as possible -
2976 without external events, your C<ev_check> watcher will not be invoked.
2978 This is where C<ev_idle> watchers come in handy - all you need is a
2979 single global idle watcher that is active as long as you have one active
2980 C<ev_check> watcher. The C<ev_idle> watcher makes sure the event loop
2981 will not sleep, and the C<ev_check> watcher makes sure a callback gets
2982 invoked. Neither watcher alone can do that.
2984 =head3 Watcher-Specific Functions and Data Members
2986 =over 4
2988 =item ev_prepare_init (ev_prepare *, callback)
2990 =item ev_check_init (ev_check *, callback)
2992 Initialises and configures the prepare or check watcher - they have no
2993 parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
2994 macros, but using them is utterly, utterly, utterly and completely
2995 pointless.
2997 =back
2999 =head3 Examples
3001 There are a number of principal ways to embed other event loops or modules
3002 into libev. Here are some ideas on how to include libadns into libev
3003 (there is a Perl module named C<EV::ADNS> that does this, which you could
3004 use as a working example. Another Perl module named C<EV::Glib> embeds a
3005 Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
3006 Glib event loop).
3008 Method 1: Add IO watchers and a timeout watcher in a prepare handler,
3009 and in a check watcher, destroy them and call into libadns. What follows
3010 is pseudo-code only of course. This requires you to either use a low
3011 priority for the check watcher or use C<ev_clear_pending> explicitly, as
3012 the callbacks for the IO/timeout watchers might not have been called yet.
3014 static ev_io iow [nfd];
3015 static ev_timer tw;
3017 static void
3018 io_cb (struct ev_loop *loop, ev_io *w, int revents)
3019 {
3020 }
3022 // create io watchers for each fd and a timer before blocking
3023 static void
3024 adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
3025 {
3026 int timeout = 3600000;
3027 struct pollfd fds [nfd];
3028 // actual code will need to loop here and realloc etc.
3029 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
3031 /* the callback is illegal, but won't be called as we stop during check */
3032 ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
3033 ev_timer_start (loop, &tw);
3035 // create one ev_io per pollfd
3036 for (int i = 0; i < nfd; ++i)
3037 {
3038 ev_io_init (iow + i, io_cb, fds [i].fd,
3039 ((fds [i].events & POLLIN ? EV_READ : 0)
3040 | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
3042 fds [i].revents = 0;
3043 ev_io_start (loop, iow + i);
3044 }
3045 }
3047 // stop all watchers after blocking
3048 static void
3049 adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
3050 {
3051 ev_timer_stop (loop, &tw);
3053 for (int i = 0; i < nfd; ++i)
3054 {
3055 // set the relevant poll flags
3056 // could also call adns_processreadable etc. here
3057 struct pollfd *fd = fds + i;
3058 int revents = ev_clear_pending (iow + i);
3059 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
3060 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
3062 // now stop the watcher
3063 ev_io_stop (loop, iow + i);
3064 }
3066 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
3067 }
3069 Method 2: This would be just like method 1, but you run C<adns_afterpoll>
3070 in the prepare watcher and would dispose of the check watcher.
3072 Method 3: If the module to be embedded supports explicit event
3073 notification (libadns does), you can also make use of the actual watcher
3074 callbacks, and only destroy/create the watchers in the prepare watcher.
3076 static void
3077 timer_cb (EV_P_ ev_timer *w, int revents)
3078 {
3079 adns_state ads = (adns_state)w->data;
3080 update_now (EV_A);
3082 adns_processtimeouts (ads, &tv_now);
3083 }
3085 static void
3086 io_cb (EV_P_ ev_io *w, int revents)
3087 {
3088 adns_state ads = (adns_state)w->data;
3089 update_now (EV_A);
3091 if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
3092 if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
3093 }
3095 // do not ever call adns_afterpoll
3097 Method 4: Do not use a prepare or check watcher because the module you
3098 want to embed is not flexible enough to support it. Instead, you can
3099 override their poll function. The drawback with this solution is that the
3100 main loop is now no longer controllable by EV. The C<Glib::EV> module uses
3101 this approach, effectively embedding EV as a client into the horrible
3102 libglib event loop.
3104 static gint
3105 event_poll_func (GPollFD *fds, guint nfds, gint timeout)
3106 {
3107 int got_events = 0;
3109 for (n = 0; n < nfds; ++n)
3110 // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
3112 if (timeout >= 0)
3113 // create/start timer
3115 // poll
3116 ev_run (EV_A_ 0);
3118 // stop timer again
3119 if (timeout >= 0)
3120 ev_timer_stop (EV_A_ &to);
3122 // stop io watchers again - their callbacks should have set
3123 for (n = 0; n < nfds; ++n)
3124 ev_io_stop (EV_A_ iow [n]);
3126 return got_events;
3127 }
3130 =head2 C<ev_embed> - when one backend isn't enough...
3132 This is a rather advanced watcher type that lets you embed one event loop
3133 into another (currently only C<ev_io> events are supported in the embedded
3134 loop, other types of watchers might be handled in a delayed or incorrect
3135 fashion and must not be used).
3137 There are primarily two reasons you would want that: work around bugs and
3138 prioritise I/O.
3140 As an example for a bug workaround, the kqueue backend might only support
3141 sockets on some platform, so it is unusable as generic backend, but you
3142 still want to make use of it because you have many sockets and it scales
3143 so nicely. In this case, you would create a kqueue-based loop and embed
3144 it into your default loop (which might use e.g. poll). Overall operation
3145 will be a bit slower because first libev has to call C<poll> and then
3146 C<kevent>, but at least you can use both mechanisms for what they are
3147 best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
3149 As for prioritising I/O: under rare circumstances you have the case where
3150 some fds have to be watched and handled very quickly (with low latency),
3151 and even priorities and idle watchers might have too much overhead. In
3152 this case you would put all the high priority stuff in one loop and all
3153 the rest in a second one, and embed the second one in the first.
3155 As long as the watcher is active, the callback will be invoked every
3156 time there might be events pending in the embedded loop. The callback
3157 must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
3158 sweep and invoke their callbacks (the callback doesn't need to invoke the
3159 C<ev_embed_sweep> function directly, it could also start an idle watcher
3160 to give the embedded loop strictly lower priority for example).
3162 You can also set the callback to C<0>, in which case the embed watcher
3163 will automatically execute the embedded loop sweep whenever necessary.
3165 Fork detection will be handled transparently while the C<ev_embed> watcher
3166 is active, i.e., the embedded loop will automatically be forked when the
3167 embedding loop forks. In other cases, the user is responsible for calling
3168 C<ev_loop_fork> on the embedded loop.
3170 Unfortunately, not all backends are embeddable: only the ones returned by
3171 C<ev_embeddable_backends> are, which, unfortunately, does not include any
3172 portable one.
3174 So when you want to use this feature you will always have to be prepared
3175 that you cannot get an embeddable loop. The recommended way to get around
3176 this is to have a separate variables for your embeddable loop, try to
3177 create it, and if that fails, use the normal loop for everything.
3179 =head3 C<ev_embed> and fork
3181 While the C<ev_embed> watcher is running, forks in the embedding loop will
3182 automatically be applied to the embedded loop as well, so no special
3183 fork handling is required in that case. When the watcher is not running,
3184 however, it is still the task of the libev user to call C<ev_loop_fork ()>
3185 as applicable.
3187 =head3 Watcher-Specific Functions and Data Members
3189 =over 4
3191 =item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
3193 =item ev_embed_set (ev_embed *, struct ev_loop *embedded_loop)
3195 Configures the watcher to embed the given loop, which must be
3196 embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
3197 invoked automatically, otherwise it is the responsibility of the callback
3198 to invoke it (it will continue to be called until the sweep has been done,
3199 if you do not want that, you need to temporarily stop the embed watcher).
3201 =item ev_embed_sweep (loop, ev_embed *)
3203 Make a single, non-blocking sweep over the embedded loop. This works
3204 similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
3205 appropriate way for embedded loops.
3207 =item struct ev_loop *other [read-only]
3209 The embedded event loop.
3211 =back
3213 =head3 Examples
3215 Example: Try to get an embeddable event loop and embed it into the default
3216 event loop. If that is not possible, use the default loop. The default
3217 loop is stored in C<loop_hi>, while the embeddable loop is stored in
3218 C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
3219 used).
3221 struct ev_loop *loop_hi = ev_default_init (0);
3222 struct ev_loop *loop_lo = 0;
3223 ev_embed embed;
3225 // see if there is a chance of getting one that works
3226 // (remember that a flags value of 0 means autodetection)
3227 loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
3228 ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
3229 : 0;
3231 // if we got one, then embed it, otherwise default to loop_hi
3232 if (loop_lo)
3233 {
3234 ev_embed_init (&embed, 0, loop_lo);
3235 ev_embed_start (loop_hi, &embed);
3236 }
3237 else
3238 loop_lo = loop_hi;
3240 Example: Check if kqueue is available but not recommended and create
3241 a kqueue backend for use with sockets (which usually work with any
3242 kqueue implementation). Store the kqueue/socket-only event loop in
3243 C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
3245 struct ev_loop *loop = ev_default_init (0);
3246 struct ev_loop *loop_socket = 0;
3247 ev_embed embed;
3249 if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
3250 if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
3251 {
3252 ev_embed_init (&embed, 0, loop_socket);
3253 ev_embed_start (loop, &embed);
3254 }
3256 if (!loop_socket)
3257 loop_socket = loop;
3259 // now use loop_socket for all sockets, and loop for everything else
3262 =head2 C<ev_fork> - the audacity to resume the event loop after a fork
3264 Fork watchers are called when a C<fork ()> was detected (usually because
3265 whoever is a good citizen cared to tell libev about it by calling
3266 C<ev_loop_fork>). The invocation is done before the event loop blocks next
3267 and before C<ev_check> watchers are being called, and only in the child
3268 after the fork. If whoever good citizen calling C<ev_default_fork> cheats
3269 and calls it in the wrong process, the fork handlers will be invoked, too,
3270 of course.
3272 =head3 The special problem of life after fork - how is it possible?
3274 Most uses of C<fork ()> consist of forking, then some simple calls to set
3275 up/change the process environment, followed by a call to C<exec()>. This
3276 sequence should be handled by libev without any problems.
3278 This changes when the application actually wants to do event handling
3279 in the child, or both parent in child, in effect "continuing" after the
3280 fork.
3282 The default mode of operation (for libev, with application help to detect
3283 forks) is to duplicate all the state in the child, as would be expected
3284 when I<either> the parent I<or> the child process continues.
3286 When both processes want to continue using libev, then this is usually the
3287 wrong result. In that case, usually one process (typically the parent) is
3288 supposed to continue with all watchers in place as before, while the other
3289 process typically wants to start fresh, i.e. without any active watchers.
3291 The cleanest and most efficient way to achieve that with libev is to
3292 simply create a new event loop, which of course will be "empty", and
3293 use that for new watchers. This has the advantage of not touching more
3294 memory than necessary, and thus avoiding the copy-on-write, and the
3295 disadvantage of having to use multiple event loops (which do not support
3296 signal watchers).
3298 When this is not possible, or you want to use the default loop for
3299 other reasons, then in the process that wants to start "fresh", call
3300 C<ev_loop_destroy (EV_DEFAULT)> followed by C<ev_default_loop (...)>.
3301 Destroying the default loop will "orphan" (not stop) all registered
3302 watchers, so you have to be careful not to execute code that modifies
3303 those watchers. Note also that in that case, you have to re-register any
3304 signal watchers.
3306 =head3 Watcher-Specific Functions and Data Members
3308 =over 4
3310 =item ev_fork_init (ev_fork *, callback)
3312 Initialises and configures the fork watcher - it has no parameters of any
3313 kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
3314 really.
3316 =back
3319 =head2 C<ev_cleanup> - even the best things end
3321 Cleanup watchers are called just before the event loop is being destroyed
3322 by a call to C<ev_loop_destroy>.
3324 While there is no guarantee that the event loop gets destroyed, cleanup
3325 watchers provide a convenient method to install cleanup hooks for your
3326 program, worker threads and so on - you just to make sure to destroy the
3327 loop when you want them to be invoked.
3329 Cleanup watchers are invoked in the same way as any other watcher. Unlike
3330 all other watchers, they do not keep a reference to the event loop (which
3331 makes a lot of sense if you think about it). Like all other watchers, you
3332 can call libev functions in the callback, except C<ev_cleanup_start>.
3334 =head3 Watcher-Specific Functions and Data Members
3336 =over 4
3338 =item ev_cleanup_init (ev_cleanup *, callback)
3340 Initialises and configures the cleanup watcher - it has no parameters of
3341 any kind. There is a C<ev_cleanup_set> macro, but using it is utterly
3342 pointless, I assure you.
3344 =back
3346 Example: Register an atexit handler to destroy the default loop, so any
3347 cleanup functions are called.
3349 static void
3350 program_exits (void)
3351 {
3352 ev_loop_destroy (EV_DEFAULT_UC);
3353 }
3355 ...
3356 atexit (program_exits);
3359 =head2 C<ev_async> - how to wake up an event loop
3361 In general, you cannot use an C<ev_loop> from multiple threads or other
3362 asynchronous sources such as signal handlers (as opposed to multiple event
3363 loops - those are of course safe to use in different threads).
3365 Sometimes, however, you need to wake up an event loop you do not control,
3366 for example because it belongs to another thread. This is what C<ev_async>
3367 watchers do: as long as the C<ev_async> watcher is active, you can signal
3368 it by calling C<ev_async_send>, which is thread- and signal safe.
3370 This functionality is very similar to C<ev_signal> watchers, as signals,
3371 too, are asynchronous in nature, and signals, too, will be compressed
3372 (i.e. the number of callback invocations may be less than the number of
3373 C<ev_async_send> calls). In fact, you could use signal watchers as a kind
3374 of "global async watchers" by using a watcher on an otherwise unused
3375 signal, and C<ev_feed_signal> to signal this watcher from another thread,
3376 even without knowing which loop owns the signal.
3378 =head3 Queueing
3380 C<ev_async> does not support queueing of data in any way. The reason
3381 is that the author does not know of a simple (or any) algorithm for a
3382 multiple-writer-single-reader queue that works in all cases and doesn't
3383 need elaborate support such as pthreads or unportable memory access
3384 semantics.
3386 That means that if you want to queue data, you have to provide your own
3387 queue. But at least I can tell you how to implement locking around your
3388 queue:
3390 =over 4
3392 =item queueing from a signal handler context
3394 To implement race-free queueing, you simply add to the queue in the signal
3395 handler but you block the signal handler in the watcher callback. Here is
3396 an example that does that for some fictitious SIGUSR1 handler:
3398 static ev_async mysig;
3400 static void
3401 sigusr1_handler (void)
3402 {
3403 sometype data;
3405 // no locking etc.
3406 queue_put (data);
3407 ev_async_send (EV_DEFAULT_ &mysig);
3408 }
3410 static void
3411 mysig_cb (EV_P_ ev_async *w, int revents)
3412 {
3413 sometype data;
3414 sigset_t block, prev;
3416 sigemptyset (&block);
3417 sigaddset (&block, SIGUSR1);
3418 sigprocmask (SIG_BLOCK, &block, &prev);
3420 while (queue_get (&data))
3421 process (data);
3423 if (sigismember (&prev, SIGUSR1)
3424 sigprocmask (SIG_UNBLOCK, &block, 0);
3425 }
3427 (Note: pthreads in theory requires you to use C<pthread_setmask>
3428 instead of C<sigprocmask> when you use threads, but libev doesn't do it
3429 either...).
3431 =item queueing from a thread context
3433 The strategy for threads is different, as you cannot (easily) block
3434 threads but you can easily preempt them, so to queue safely you need to
3435 employ a traditional mutex lock, such as in this pthread example:
3437 static ev_async mysig;
3438 static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
3440 static void
3441 otherthread (void)
3442 {
3443 // only need to lock the actual queueing operation
3444 pthread_mutex_lock (&mymutex);
3445 queue_put (data);
3446 pthread_mutex_unlock (&mymutex);
3448 ev_async_send (EV_DEFAULT_ &mysig);
3449 }
3451 static void
3452 mysig_cb (EV_P_ ev_async *w, int revents)
3453 {
3454 pthread_mutex_lock (&mymutex);
3456 while (queue_get (&data))
3457 process (data);
3459 pthread_mutex_unlock (&mymutex);
3460 }
3462 =back
3465 =head3 Watcher-Specific Functions and Data Members
3467 =over 4
3469 =item ev_async_init (ev_async *, callback)
3471 Initialises and configures the async watcher - it has no parameters of any
3472 kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
3473 trust me.
3475 =item ev_async_send (loop, ev_async *)
3477 Sends/signals/activates the given C<ev_async> watcher, that is, feeds
3478 an C<EV_ASYNC> event on the watcher into the event loop, and instantly
3479 returns.
3481 Unlike C<ev_feed_event>, this call is safe to do from other threads,
3482 signal or similar contexts (see the discussion of C<EV_ATOMIC_T> in the
3483 embedding section below on what exactly this means).
3485 Note that, as with other watchers in libev, multiple events might get
3486 compressed into a single callback invocation (another way to look at
3487 this is that C<ev_async> watchers are level-triggered: they are set on
3488 C<ev_async_send>, reset when the event loop detects that).
3490 This call incurs the overhead of at most one extra system call per event
3491 loop iteration, if the event loop is blocked, and no syscall at all if
3492 the event loop (or your program) is processing events. That means that
3493 repeated calls are basically free (there is no need to avoid calls for
3494 performance reasons) and that the overhead becomes smaller (typically
3495 zero) under load.
3497 =item bool = ev_async_pending (ev_async *)
3499 Returns a non-zero value when C<ev_async_send> has been called on the
3500 watcher but the event has not yet been processed (or even noted) by the
3501 event loop.
3503 C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
3504 the loop iterates next and checks for the watcher to have become active,
3505 it will reset the flag again. C<ev_async_pending> can be used to very
3506 quickly check whether invoking the loop might be a good idea.
3508 Not that this does I<not> check whether the watcher itself is pending,
3509 only whether it has been requested to make this watcher pending: there
3510 is a time window between the event loop checking and resetting the async
3511 notification, and the callback being invoked.
3513 =back
3518 There are some other functions of possible interest. Described. Here. Now.
3520 =over 4
3522 =item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
3524 This function combines a simple timer and an I/O watcher, calls your
3525 callback on whichever event happens first and automatically stops both
3526 watchers. This is useful if you want to wait for a single event on an fd
3527 or timeout without having to allocate/configure/start/stop/free one or
3528 more watchers yourself.
3530 If C<fd> is less than 0, then no I/O watcher will be started and the
3531 C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
3532 the given C<fd> and C<events> set will be created and started.
3534 If C<timeout> is less than 0, then no timeout watcher will be
3535 started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
3536 repeat = 0) will be started. C<0> is a valid timeout.
3538 The callback has the type C<void (*cb)(int revents, void *arg)> and is
3539 passed an C<revents> set like normal event callbacks (a combination of
3540 C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
3541 value passed to C<ev_once>. Note that it is possible to receive I<both>
3542 a timeout and an io event at the same time - you probably should give io
3543 events precedence.
3545 Example: wait up to ten seconds for data to appear on STDIN_FILENO.
3547 static void stdin_ready (int revents, void *arg)
3548 {
3549 if (revents & EV_READ)
3550 /* stdin might have data for us, joy! */;
3551 else if (revents & EV_TIMER)
3552 /* doh, nothing entered */;
3553 }
3555 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
3557 =item ev_feed_fd_event (loop, int fd, int revents)
3559 Feed an event on the given fd, as if a file descriptor backend detected
3560 the given events.
3562 =item ev_feed_signal_event (loop, int signum)
3564 Feed an event as if the given signal occurred. See also C<ev_feed_signal>,
3565 which is async-safe.
3567 =back
3572 This section explains some common idioms that are not immediately
3573 obvious. Note that examples are sprinkled over the whole manual, and this
3574 section only contains stuff that wouldn't fit anywhere else.
3578 Each watcher has, by default, a C<void *data> member that you can read
3579 or modify at any time: libev will completely ignore it. This can be used
3580 to associate arbitrary data with your watcher. If you need more data and
3581 don't want to allocate memory separately and store a pointer to it in that
3582 data member, you can also "subclass" the watcher type and provide your own
3583 data:
3585 struct my_io
3586 {
3587 ev_io io;
3588 int otherfd;
3589 void *somedata;
3590 struct whatever *mostinteresting;
3591 };
3593 ...
3594 struct my_io w;
3595 ev_io_init (&, my_cb, fd, EV_READ);
3597 And since your callback will be called with a pointer to the watcher, you
3598 can cast it back to your own type:
3600 static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
3601 {
3602 struct my_io *w = (struct my_io *)w_;
3603 ...
3604 }
3606 More interesting and less C-conformant ways of casting your callback
3607 function type instead have been omitted.
3611 Another common scenario is to use some data structure with multiple
3612 embedded watchers, in effect creating your own watcher that combines
3613 multiple libev event sources into one "super-watcher":
3615 struct my_biggy
3616 {
3617 int some_data;
3618 ev_timer t1;
3619 ev_timer t2;
3620 }
3622 In this case getting the pointer to C<my_biggy> is a bit more
3623 complicated: Either you store the address of your C<my_biggy> struct in
3624 the C<data> member of the watcher (for woozies or C++ coders), or you need
3625 to use some pointer arithmetic using C<offsetof> inside your watchers (for
3626 real programmers):
3628 #include <stddef.h>
3630 static void
3631 t1_cb (EV_P_ ev_timer *w, int revents)
3632 {
3633 struct my_biggy big = (struct my_biggy *)
3634 (((char *)w) - offsetof (struct my_biggy, t1));
3635 }
3637 static void
3638 t2_cb (EV_P_ ev_timer *w, int revents)
3639 {
3640 struct my_biggy big = (struct my_biggy *)
3641 (((char *)w) - offsetof (struct my_biggy, t2));
3642 }
3646 Often you have structures like this in event-based programs:
3648 callback ()
3649 {
3650 free (request);
3651 }
3653 request = start_new_request (..., callback);
3655 The intent is to start some "lengthy" operation. The C<request> could be
3656 used to cancel the operation, or do other things with it.
3658 It's not uncommon to have code paths in C<start_new_request> that
3659 immediately invoke the callback, for example, to report errors. Or you add
3660 some caching layer that finds that it can skip the lengthy aspects of the
3661 operation and simply invoke the callback with the result.
3663 The problem here is that this will happen I<before> C<start_new_request>
3664 has returned, so C<request> is not set.
3666 Even if you pass the request by some safer means to the callback, you
3667 might want to do something to the request after starting it, such as
3668 canceling it, which probably isn't working so well when the callback has
3669 already been invoked.
3671 A common way around all these issues is to make sure that
3672 C<start_new_request> I<always> returns before the callback is invoked. If
3673 C<start_new_request> immediately knows the result, it can artificially
3674 delay invoking the callback by using a C<prepare> or C<idle> watcher for
3675 example, or more sneakily, by reusing an existing (stopped) watcher and
3676 pushing it into the pending queue:
3678 ev_set_cb (watcher, callback);
3679 ev_feed_event (EV_A_ watcher, 0);
3681 This way, C<start_new_request> can safely return before the callback is
3682 invoked, while not delaying callback invocation too much.
3686 Often (especially in GUI toolkits) there are places where you have
3687 I<modal> interaction, which is most easily implemented by recursively
3688 invoking C<ev_run>.
3690 This brings the problem of exiting - a callback might want to finish the
3691 main C<ev_run> call, but not the nested one (e.g. user clicked "Quit", but
3692 a modal "Are you sure?" dialog is still waiting), or just the nested one
3693 and not the main one (e.g. user clocked "Ok" in a modal dialog), or some
3694 other combination: In these cases, a simple C<ev_break> will not work.
3696 The solution is to maintain "break this loop" variable for each C<ev_run>
3697 invocation, and use a loop around C<ev_run> until the condition is
3698 triggered, using C<EVRUN_ONCE>:
3700 // main loop
3701 int exit_main_loop = 0;
3703 while (!exit_main_loop)
3704 ev_run (EV_DEFAULT_ EVRUN_ONCE);
3706 // in a modal watcher
3707 int exit_nested_loop = 0;
3709 while (!exit_nested_loop)
3710 ev_run (EV_A_ EVRUN_ONCE);
3712 To exit from any of these loops, just set the corresponding exit variable:
3714 // exit modal loop
3715 exit_nested_loop = 1;
3717 // exit main program, after modal loop is finished
3718 exit_main_loop = 1;
3720 // exit both
3721 exit_main_loop = exit_nested_loop = 1;
3725 Here is a fictitious example of how to run an event loop in a different
3726 thread from where callbacks are being invoked and watchers are
3727 created/added/removed.
3729 For a real-world example, see the C<EV::Loop::Async> perl module,
3730 which uses exactly this technique (which is suited for many high-level
3731 languages).
3733 The example uses a pthread mutex to protect the loop data, a condition
3734 variable to wait for callback invocations, an async watcher to notify the
3735 event loop thread and an unspecified mechanism to wake up the main thread.
3737 First, you need to associate some data with the event loop:
3739 typedef struct {
3740 mutex_t lock; /* global loop lock */
3741 ev_async async_w;
3742 thread_t tid;
3743 cond_t invoke_cv;
3744 } userdata;
3746 void prepare_loop (EV_P)
3747 {
3748 // for simplicity, we use a static userdata struct.
3749 static userdata u;
3751 ev_async_init (&u->async_w, async_cb);
3752 ev_async_start (EV_A_ &u->async_w);
3754 pthread_mutex_init (&u->lock, 0);
3755 pthread_cond_init (&u->invoke_cv, 0);
3757 // now associate this with the loop
3758 ev_set_userdata (EV_A_ u);
3759 ev_set_invoke_pending_cb (EV_A_ l_invoke);
3760 ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
3762 // then create the thread running ev_run
3763 pthread_create (&u->tid, 0, l_run, EV_A);
3764 }
3766 The callback for the C<ev_async> watcher does nothing: the watcher is used
3767 solely to wake up the event loop so it takes notice of any new watchers
3768 that might have been added:
3770 static void
3771 async_cb (EV_P_ ev_async *w, int revents)
3772 {
3773 // just used for the side effects
3774 }
3776 The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
3777 protecting the loop data, respectively.
3779 static void
3780 l_release (EV_P)
3781 {
3782 userdata *u = ev_userdata (EV_A);
3783 pthread_mutex_unlock (&u->lock);
3784 }
3786 static void
3787 l_acquire (EV_P)
3788 {
3789 userdata *u = ev_userdata (EV_A);
3790 pthread_mutex_lock (&u->lock);
3791 }
3793 The event loop thread first acquires the mutex, and then jumps straight
3794 into C<ev_run>:
3796 void *
3797 l_run (void *thr_arg)
3798 {
3799 struct ev_loop *loop = (struct ev_loop *)thr_arg;
3801 l_acquire (EV_A);
3802 pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
3803 ev_run (EV_A_ 0);
3804 l_release (EV_A);
3806 return 0;
3807 }
3809 Instead of invoking all pending watchers, the C<l_invoke> callback will
3810 signal the main thread via some unspecified mechanism (signals? pipe
3811 writes? C<Async::Interrupt>?) and then waits until all pending watchers
3812 have been called (in a while loop because a) spurious wakeups are possible
3813 and b) skipping inter-thread-communication when there are no pending
3814 watchers is very beneficial):
3816 static void
3817 l_invoke (EV_P)
3818 {
3819 userdata *u = ev_userdata (EV_A);
3821 while (ev_pending_count (EV_A))
3822 {
3823 wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
3824 pthread_cond_wait (&u->invoke_cv, &u->lock);
3825 }
3826 }
3828 Now, whenever the main thread gets told to invoke pending watchers, it
3829 will grab the lock, call C<ev_invoke_pending> and then signal the loop
3830 thread to continue:
3832 static void
3833 real_invoke_pending (EV_P)
3834 {
3835 userdata *u = ev_userdata (EV_A);
3837 pthread_mutex_lock (&u->lock);
3838 ev_invoke_pending (EV_A);
3839 pthread_cond_signal (&u->invoke_cv);
3840 pthread_mutex_unlock (&u->lock);
3841 }
3843 Whenever you want to start/stop a watcher or do other modifications to an
3844 event loop, you will now have to lock:
3846 ev_timer timeout_watcher;
3847 userdata *u = ev_userdata (EV_A);
3849 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
3851 pthread_mutex_lock (&u->lock);
3852 ev_timer_start (EV_A_ &timeout_watcher);
3853 ev_async_send (EV_A_ &u->async_w);
3854 pthread_mutex_unlock (&u->lock);
3856 Note that sending the C<ev_async> watcher is required because otherwise
3857 an event loop currently blocking in the kernel will have no knowledge
3858 about the newly added timer. By waking up the loop it will pick up any new
3859 watchers in the next event loop iteration.
3863 While the overhead of a callback that e.g. schedules a thread is small, it
3864 is still an overhead. If you embed libev, and your main usage is with some
3865 kind of threads or coroutines, you might want to customise libev so that
3866 doesn't need callbacks anymore.
3868 Imagine you have coroutines that you can switch to using a function
3869 C<switch_to (coro)>, that libev runs in a coroutine called C<libev_coro>
3870 and that due to some magic, the currently active coroutine is stored in a
3871 global called C<current_coro>. Then you can build your own "wait for libev
3872 event" primitive by changing C<EV_CB_DECLARE> and C<EV_CB_INVOKE> (note
3873 the differing C<;> conventions):
3875 #define EV_CB_DECLARE(type) struct my_coro *cb;
3876 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
3878 That means instead of having a C callback function, you store the
3879 coroutine to switch to in each watcher, and instead of having libev call
3880 your callback, you instead have it switch to that coroutine.
3882 A coroutine might now wait for an event with a function called
3883 C<wait_for_event>. (the watcher needs to be started, as always, but it doesn't
3884 matter when, or whether the watcher is active or not when this function is
3885 called):
3887 void
3888 wait_for_event (ev_watcher *w)
3889 {
3890 ev_set_cb (w, current_coro);
3891 switch_to (libev_coro);
3892 }
3894 That basically suspends the coroutine inside C<wait_for_event> and
3895 continues the libev coroutine, which, when appropriate, switches back to
3896 this or any other coroutine.
3898 You can do similar tricks if you have, say, threads with an event queue -
3899 instead of storing a coroutine, you store the queue object and instead of
3900 switching to a coroutine, you push the watcher onto the queue and notify
3901 any waiters.
3903 To embed libev, see L</EMBEDDING>, but in short, it's easiest to create two
3904 files, F<my_ev.h> and F<my_ev.c> that include the respective libev files:
3906 // my_ev.h
3907 #define EV_CB_DECLARE(type) struct my_coro *cb;
3908 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
3909 #include "../libev/ev.h"
3911 // my_ev.c
3912 #define EV_H "my_ev.h"
3913 #include "../libev/ev.c"
3915 And then use F<my_ev.h> when you would normally use F<ev.h>, and compile
3916 F<my_ev.c> into your project. When properly specifying include paths, you
3917 can even use F<ev.h> as header file name directly.
3922 Libev offers a compatibility emulation layer for libevent. It cannot
3923 emulate the internals of libevent, so here are some usage hints:
3925 =over 4
3927 =item * Only the libevent-1.4.1-beta API is being emulated.
3929 This was the newest libevent version available when libev was implemented,
3930 and is still mostly unchanged in 2010.
3932 =item * Use it by including <event.h>, as usual.
3934 =item * The following members are fully supported: ev_base, ev_callback,
3935 ev_arg, ev_fd, ev_res, ev_events.
3937 =item * Avoid using ev_flags and the EVLIST_*-macros, while it is
3938 maintained by libev, it does not work exactly the same way as in libevent (consider
3939 it a private API).
3941 =item * Priorities are not currently supported. Initialising priorities
3942 will fail and all watchers will have the same priority, even though there
3943 is an ev_pri field.
3945 =item * In libevent, the last base created gets the signals, in libev, the
3946 base that registered the signal gets the signals.
3948 =item * Other members are not supported.
3950 =item * The libev emulation is I<not> ABI compatible to libevent, you need
3951 to use the libev header file and library.
3953 =back
3955 =head1 C++ SUPPORT
3957 =head2 C API
3959 The normal C API should work fine when used from C++: both ev.h and the
3960 libev sources can be compiled as C++. Therefore, code that uses the C API
3961 will work fine.
3963 Proper exception specifications might have to be added to callbacks passed
3964 to libev: exceptions may be thrown only from watcher callbacks, all
3965 other callbacks (allocator, syserr, loop acquire/release and periodic
3966 reschedule callbacks) must not throw exceptions, and might need a C<throw
3967 ()> specification. If you have code that needs to be compiled as both C
3968 and C++ you can use the C<EV_THROW> macro for this:
3970 static void
3971 fatal_error (const char *msg) EV_THROW
3972 {
3973 perror (msg);
3974 abort ();
3975 }
3977 ...
3978 ev_set_syserr_cb (fatal_error);
3980 The only API functions that can currently throw exceptions are C<ev_run>,
3981 C<ev_invoke>, C<ev_invoke_pending> and C<ev_loop_destroy> (the latter
3982 because it runs cleanup watchers).
3984 Throwing exceptions in watcher callbacks is only supported if libev itself
3985 is compiled with a C++ compiler or your C and C++ environments allow
3986 throwing exceptions through C libraries (most do).
3988 =head2 C++ API
3990 Libev comes with some simplistic wrapper classes for C++ that mainly allow
3991 you to use some convenience methods to start/stop watchers and also change
3992 the callback model to a model using method callbacks on objects.
3994 To use it,
3996 #include <ev++.h>
3998 This automatically includes F<ev.h> and puts all of its definitions (many
3999 of them macros) into the global namespace. All C++ specific things are
4000 put into the C<ev> namespace. It should support all the same embedding
4001 options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
4003 Care has been taken to keep the overhead low. The only data member the C++
4004 classes add (compared to plain C-style watchers) is the event loop pointer
4005 that the watcher is associated with (or no additional members at all if
4006 you disable C<EV_MULTIPLICITY> when embedding libev).
4008 Currently, functions, static and non-static member functions and classes
4009 with C<operator ()> can be used as callbacks. Other types should be easy
4010 to add as long as they only need one additional pointer for context. If
4011 you need support for other types of functors please contact the author
4012 (preferably after implementing it).
4014 For all this to work, your C++ compiler either has to use the same calling
4015 conventions as your C compiler (for static member functions), or you have
4016 to embed libev and compile libev itself as C++.
4018 Here is a list of things available in the C<ev> namespace:
4020 =over 4
4022 =item C<ev::READ>, C<ev::WRITE> etc.
4024 These are just enum values with the same values as the C<EV_READ> etc.
4025 macros from F<ev.h>.
4027 =item C<ev::tstamp>, C<ev::now>
4029 Aliases to the same types/functions as with the C<ev_> prefix.
4031 =item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
4033 For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
4034 the same name in the C<ev> namespace, with the exception of C<ev_signal>
4035 which is called C<ev::sig> to avoid clashes with the C<signal> macro
4036 defined by many implementations.
4038 All of those classes have these methods:
4040 =over 4
4042 =item ev::TYPE::TYPE ()
4044 =item ev::TYPE::TYPE (loop)
4046 =item ev::TYPE::~TYPE
4048 The constructor (optionally) takes an event loop to associate the watcher
4049 with. If it is omitted, it will use C<EV_DEFAULT>.
4051 The constructor calls C<ev_init> for you, which means you have to call the
4052 C<set> method before starting it.
4054 It will not set a callback, however: You have to call the templated C<set>
4055 method to set a callback before you can start the watcher.
4057 (The reason why you have to use a method is a limitation in C++ which does
4058 not allow explicit template arguments for constructors).
4060 The destructor automatically stops the watcher if it is active.
4062 =item w->set<class, &class::method> (object *)
4064 This method sets the callback method to call. The method has to have a
4065 signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
4066 first argument and the C<revents> as second. The object must be given as
4067 parameter and is stored in the C<data> member of the watcher.
4069 This method synthesizes efficient thunking code to call your method from
4070 the C callback that libev requires. If your compiler can inline your
4071 callback (i.e. it is visible to it at the place of the C<set> call and
4072 your compiler is good :), then the method will be fully inlined into the
4073 thunking function, making it as fast as a direct C callback.
4075 Example: simple class declaration and watcher initialisation
4077 struct myclass
4078 {
4079 void io_cb (ev::io &w, int revents) { }
4080 }
4082 myclass obj;
4083 ev::io iow;
4084 iow.set <myclass, &myclass::io_cb> (&obj);
4086 =item w->set (object *)
4088 This is a variation of a method callback - leaving out the method to call
4089 will default the method to C<operator ()>, which makes it possible to use
4090 functor objects without having to manually specify the C<operator ()> all
4091 the time. Incidentally, you can then also leave out the template argument
4092 list.
4094 The C<operator ()> method prototype must be C<void operator ()(watcher &w,
4095 int revents)>.
4097 See the method-C<set> above for more details.
4099 Example: use a functor object as callback.
4101 struct myfunctor
4102 {
4103 void operator() (ev::io &w, int revents)
4104 {
4105 ...
4106 }
4107 }
4109 myfunctor f;
4111 ev::io w;
4112 w.set (&f);
4114 =item w->set<function> (void *data = 0)
4116 Also sets a callback, but uses a static method or plain function as
4117 callback. The optional C<data> argument will be stored in the watcher's
4118 C<data> member and is free for you to use.
4120 The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
4122 See the method-C<set> above for more details.
4124 Example: Use a plain function as callback.
4126 static void io_cb (ev::io &w, int revents) { }
4127 iow.set <io_cb> ();
4129 =item w->set (loop)
4131 Associates a different C<struct ev_loop> with this watcher. You can only
4132 do this when the watcher is inactive (and not pending either).
4134 =item w->set ([arguments])
4136 Basically the same as C<ev_TYPE_set> (except for C<ev::embed> watchers>),
4137 with the same arguments. Either this method or a suitable start method
4138 must be called at least once. Unlike the C counterpart, an active watcher
4139 gets automatically stopped and restarted when reconfiguring it with this
4140 method.
4142 For C<ev::embed> watchers this method is called C<set_embed>, to avoid
4143 clashing with the C<set (loop)> method.
4145 =item w->start ()
4147 Starts the watcher. Note that there is no C<loop> argument, as the
4148 constructor already stores the event loop.
4150 =item w->start ([arguments])
4152 Instead of calling C<set> and C<start> methods separately, it is often
4153 convenient to wrap them in one call. Uses the same type of arguments as
4154 the configure C<set> method of the watcher.
4156 =item w->stop ()
4158 Stops the watcher if it is active. Again, no C<loop> argument.
4160 =item w->again () (C<ev::timer>, C<ev::periodic> only)
4162 For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
4163 C<ev_TYPE_again> function.
4165 =item w->sweep () (C<ev::embed> only)
4167 Invokes C<ev_embed_sweep>.
4169 =item w->update () (C<ev::stat> only)
4171 Invokes C<ev_stat_stat>.
4173 =back
4175 =back
4177 Example: Define a class with two I/O and idle watchers, start the I/O
4178 watchers in the constructor.
4180 class myclass
4181 {
4182 ev::io io ; void io_cb (ev::io &w, int revents);
4183 ev::io io2 ; void io2_cb (ev::io &w, int revents);
4184 ev::idle idle; void idle_cb (ev::idle &w, int revents);
4186 myclass (int fd)
4187 {
4188 io .set <myclass, &myclass::io_cb > (this);
4189 io2 .set <myclass, &myclass::io2_cb > (this);
4190 idle.set <myclass, &myclass::idle_cb> (this);
4192 io.set (fd, ev::WRITE); // configure the watcher
4193 io.start (); // start it whenever convenient
4195 io2.start (fd, ev::READ); // set + start in one call
4196 }
4197 };
4202 Libev does not offer other language bindings itself, but bindings for a
4203 number of languages exist in the form of third-party packages. If you know
4204 any interesting language binding in addition to the ones listed here, drop
4205 me a note.
4207 =over 4
4209 =item Perl
4211 The EV module implements the full libev API and is actually used to test
4212 libev. EV is developed together with libev. Apart from the EV core module,
4213 there are additional modules that implement libev-compatible interfaces
4214 to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
4215 C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
4216 and C<EV::Glib>).
4218 It can be found and installed via CPAN, its homepage is at
4219 L<>.
4221 =item Python
4223 Python bindings can be found at L<>. It
4224 seems to be quite complete and well-documented.
4226 =item Ruby
4228 Tony Arcieri has written a ruby extension that offers access to a subset
4229 of the libev API and adds file handle abstractions, asynchronous DNS and
4230 more on top of it. It can be found via gem servers. Its homepage is at
4231 L<>.
4233 Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
4234 makes rev work even on mingw.
4236 =item Haskell
4238 A haskell binding to libev is available at
4239 L<>.
4241 =item D
4243 Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
4244 be found at L<>.
4246 =item Ocaml
4248 Erkki Seppala has written Ocaml bindings for libev, to be found at
4249 L<>.
4251 =item Lua
4253 Brian Maher has written a partial interface to libev for lua (at the
4254 time of this writing, only C<ev_io> and C<ev_timer>), to be found at
4255 L<>.
4257 =item Javascript
4259 Node.js (L<>) uses libev as the underlying event library.
4261 =item Others
4263 There are others, and I stopped counting.
4265 =back
4268 =head1 MACRO MAGIC
4270 Libev can be compiled with a variety of options, the most fundamental
4271 of which is C<EV_MULTIPLICITY>. This option determines whether (most)
4272 functions and callbacks have an initial C<struct ev_loop *> argument.
4274 To make it easier to write programs that cope with either variant, the
4275 following macros are defined:
4277 =over 4
4279 =item C<EV_A>, C<EV_A_>
4281 This provides the loop I<argument> for functions, if one is required ("ev
4282 loop argument"). The C<EV_A> form is used when this is the sole argument,
4283 C<EV_A_> is used when other arguments are following. Example:
4285 ev_unref (EV_A);
4286 ev_timer_add (EV_A_ watcher);
4287 ev_run (EV_A_ 0);
4289 It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
4290 which is often provided by the following macro.
4292 =item C<EV_P>, C<EV_P_>
4294 This provides the loop I<parameter> for functions, if one is required ("ev
4295 loop parameter"). The C<EV_P> form is used when this is the sole parameter,
4296 C<EV_P_> is used when other parameters are following. Example:
4298 // this is how ev_unref is being declared
4299 static void ev_unref (EV_P);
4301 // this is how you can declare your typical callback
4302 static void cb (EV_P_ ev_timer *w, int revents)
4304 It declares a parameter C<loop> of type C<struct ev_loop *>, quite
4305 suitable for use with C<EV_A>.
4307 =item C<EV_DEFAULT>, C<EV_DEFAULT_>
4309 Similar to the other two macros, this gives you the value of the default
4310 loop, if multiple loops are supported ("ev loop default"). The default loop
4311 will be initialised if it isn't already initialised.
4313 For non-multiplicity builds, these macros do nothing, so you always have
4314 to initialise the loop somewhere.
4318 Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
4319 default loop has been initialised (C<UC> == unchecked). Their behaviour
4320 is undefined when the default loop has not been initialised by a previous
4321 execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
4323 It is often prudent to use C<EV_DEFAULT> when initialising the first
4324 watcher in a function but use C<EV_DEFAULT_UC> afterwards.
4326 =back
4328 Example: Declare and initialise a check watcher, utilising the above
4329 macros so it will work regardless of whether multiple loops are supported
4330 or not.
4332 static void
4333 check_cb (EV_P_ ev_timer *w, int revents)
4334 {
4335 ev_check_stop (EV_A_ w);
4336 }
4338 ev_check check;
4339 ev_check_init (&check, check_cb);
4340 ev_check_start (EV_DEFAULT_ &check);
4341 ev_run (EV_DEFAULT_ 0);
4343 =head1 EMBEDDING
4345 Libev can (and often is) directly embedded into host
4346 applications. Examples of applications that embed it include the Deliantra
4347 Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
4348 and rxvt-unicode.
4350 The goal is to enable you to just copy the necessary files into your
4351 source directory without having to change even a single line in them, so
4352 you can easily upgrade by simply copying (or having a checked-out copy of
4353 libev somewhere in your source tree).
4355 =head2 FILESETS
4357 Depending on what features you need you need to include one or more sets of files
4358 in your application.
4360 =head3 CORE EVENT LOOP
4362 To include only the libev core (all the C<ev_*> functions), with manual
4363 configuration (no autoconf):
4365 #define EV_STANDALONE 1
4366 #include "ev.c"
4368 This will automatically include F<ev.h>, too, and should be done in a
4369 single C source file only to provide the function implementations. To use
4370 it, do the same for F<ev.h> in all files wishing to use this API (best
4371 done by writing a wrapper around F<ev.h> that you can include instead and
4372 where you can put other configuration options):
4374 #define EV_STANDALONE 1
4375 #include "ev.h"
4377 Both header files and implementation files can be compiled with a C++
4378 compiler (at least, that's a stated goal, and breakage will be treated
4379 as a bug).
4381 You need the following files in your source tree, or in a directory
4382 in your include path (e.g. in libev/ when using -Ilibev):
4384 ev.h
4385 ev.c
4386 ev_vars.h
4387 ev_wrap.h
4389 ev_win32.c required on win32 platforms only
4391 ev_select.c only when select backend is enabled (which is enabled by default)
4392 ev_poll.c only when poll backend is enabled (disabled by default)
4393 ev_epoll.c only when the epoll backend is enabled (disabled by default)
4394 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
4395 ev_port.c only when the solaris port backend is enabled (disabled by default)
4397 F<ev.c> includes the backend files directly when enabled, so you only need
4398 to compile this single file.
4402 To include the libevent compatibility API, also include:
4404 #include "event.c"
4406 in the file including F<ev.c>, and:
4408 #include "event.h"
4410 in the files that want to use the libevent API. This also includes F<ev.h>.
4412 You need the following additional files for this:
4414 event.h
4415 event.c
4419 Instead of using C<EV_STANDALONE=1> and providing your configuration in
4420 whatever way you want, you can also C<m4_include([libev.m4])> in your
4421 F<> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
4422 include F<config.h> and configure itself accordingly.
4424 For this of course you need the m4 file:
4426 libev.m4
4430 Libev can be configured via a variety of preprocessor symbols you have to
4431 define before including (or compiling) any of its files. The default in
4432 the absence of autoconf is documented for every option.
4434 Symbols marked with "(h)" do not change the ABI, and can have different
4435 values when compiling libev vs. including F<ev.h>, so it is permissible
4436 to redefine them before including F<ev.h> without breaking compatibility
4437 to a compiled library. All other symbols change the ABI, which means all
4438 users of libev and the libev code itself must be compiled with compatible
4439 settings.
4441 =over 4
4443 =item EV_COMPAT3 (h)
4445 Backwards compatibility is a major concern for libev. This is why this
4446 release of libev comes with wrappers for the functions and symbols that
4447 have been renamed between libev version 3 and 4.
4449 You can disable these wrappers (to test compatibility with future
4450 versions) by defining C<EV_COMPAT3> to C<0> when compiling your
4451 sources. This has the additional advantage that you can drop the C<struct>
4452 from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
4453 typedef in that case.
4455 In some future version, the default for C<EV_COMPAT3> will become C<0>,
4456 and in some even more future version the compatibility code will be
4457 removed completely.
4459 =item EV_STANDALONE (h)
4461 Must always be C<1> if you do not use autoconf configuration, which
4462 keeps libev from including F<config.h>, and it also defines dummy
4463 implementations for some libevent functions (such as logging, which is not
4464 supported). It will also not define any of the structs usually found in
4465 F<event.h> that are not directly supported by the libev core alone.
4467 In standalone mode, libev will still try to automatically deduce the
4468 configuration, but has to be more conservative.
4470 =item EV_USE_FLOOR
4472 If defined to be C<1>, libev will use the C<floor ()> function for its
4473 periodic reschedule calculations, otherwise libev will fall back on a
4474 portable (slower) implementation. If you enable this, you usually have to
4475 link against libm or something equivalent. Enabling this when the C<floor>
4476 function is not available will fail, so the safe default is to not enable
4477 this.
4481 If defined to be C<1>, libev will try to detect the availability of the
4482 monotonic clock option at both compile time and runtime. Otherwise no
4483 use of the monotonic clock option will be attempted. If you enable this,
4484 you usually have to link against librt or something similar. Enabling it
4485 when the functionality isn't available is safe, though, although you have
4486 to make sure you link against any libraries where the C<clock_gettime>
4487 function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
4489 =item EV_USE_REALTIME
4491 If defined to be C<1>, libev will try to detect the availability of the
4492 real-time clock option at compile time (and assume its availability
4493 at runtime if successful). Otherwise no use of the real-time clock
4494 option will be attempted. This effectively replaces C<gettimeofday>
4495 by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
4496 correctness. See the note about libraries in the description of
4497 C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
4502 If defined to be C<1>, libev will try to use a direct syscall instead
4503 of calling the system-provided C<clock_gettime> function. This option
4504 exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
4505 unconditionally pulls in C<libpthread>, slowing down single-threaded
4506 programs needlessly. Using a direct syscall is slightly slower (in
4507 theory), because no optimised vdso implementation can be used, but avoids
4508 the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
4509 higher, as it simplifies linking (no need for C<-lrt>).
4513 If defined to be C<1>, libev will assume that C<nanosleep ()> is available
4514 and will use it for delays. Otherwise it will use C<select ()>.
4516 =item EV_USE_EVENTFD
4518 If defined to be C<1>, then libev will assume that C<eventfd ()> is
4519 available and will probe for kernel support at runtime. This will improve
4520 C<ev_signal> and C<ev_async> performance and reduce resource consumption.
4521 If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
4522 2.7 or newer, otherwise disabled.
4524 =item EV_USE_SELECT
4526 If undefined or defined to be C<1>, libev will compile in support for the
4527 C<select>(2) backend. No attempt at auto-detection will be done: if no
4528 other method takes over, select will be it. Otherwise the select backend
4529 will not be compiled in.
4533 If defined to C<1>, then the select backend will use the system C<fd_set>
4534 structure. This is useful if libev doesn't compile due to a missing
4535 C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
4536 on exotic systems. This usually limits the range of file descriptors to
4537 some low limit such as 1024 or might have other limitations (winsocket
4538 only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
4539 configures the maximum size of the C<fd_set>.
4543 When defined to C<1>, the select backend will assume that
4544 select/socket/connect etc. don't understand file descriptors but
4545 wants osf handles on win32 (this is the case when the select to
4546 be used is the winsock select). This means that it will call
4547 C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
4548 it is assumed that all these functions actually work on fds, even
4549 on win32. Should not be defined on non-win32 platforms.
4551 =item EV_FD_TO_WIN32_HANDLE(fd)
4553 If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
4554 file descriptors to socket handles. When not defining this symbol (the
4555 default), then libev will call C<_get_osfhandle>, which is usually
4556 correct. In some cases, programs use their own file descriptor management,
4557 in which case they can provide this function to map fds to socket handles.
4559 =item EV_WIN32_HANDLE_TO_FD(handle)
4561 If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
4562 using the standard C<_open_osfhandle> function. For programs implementing
4563 their own fd to handle mapping, overwriting this function makes it easier
4564 to do so. This can be done by defining this macro to an appropriate value.
4566 =item EV_WIN32_CLOSE_FD(fd)
4568 If programs implement their own fd to handle mapping on win32, then this
4569 macro can be used to override the C<close> function, useful to unregister
4570 file descriptors again. Note that the replacement function has to close
4571 the underlying OS handle.
4575 If defined to be C<1>, libev will use C<WSASocket> to create its internal
4576 communication socket, which works better in some environments. Otherwise,
4577 the normal C<socket> function will be used, which works better in other
4578 environments.
4580 =item EV_USE_POLL
4582 If defined to be C<1>, libev will compile in support for the C<poll>(2)
4583 backend. Otherwise it will be enabled on non-win32 platforms. It
4584 takes precedence over select.
4586 =item EV_USE_EPOLL
4588 If defined to be C<1>, libev will compile in support for the Linux
4589 C<epoll>(7) backend. Its availability will be detected at runtime,
4590 otherwise another method will be used as fallback. This is the preferred
4591 backend for GNU/Linux systems. If undefined, it will be enabled if the
4592 headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
4594 =item EV_USE_KQUEUE
4596 If defined to be C<1>, libev will compile in support for the BSD style
4597 C<kqueue>(2) backend. Its actual availability will be detected at runtime,
4598 otherwise another method will be used as fallback. This is the preferred
4599 backend for BSD and BSD-like systems, although on most BSDs kqueue only
4600 supports some types of fds correctly (the only platform we found that
4601 supports ptys for example was NetBSD), so kqueue might be compiled in, but
4602 not be used unless explicitly requested. The best way to use it is to find
4603 out whether kqueue supports your type of fd properly and use an embedded
4604 kqueue loop.
4606 =item EV_USE_PORT
4608 If defined to be C<1>, libev will compile in support for the Solaris
4609 10 port style backend. Its availability will be detected at runtime,
4610 otherwise another method will be used as fallback. This is the preferred
4611 backend for Solaris 10 systems.
4613 =item EV_USE_DEVPOLL
4615 Reserved for future expansion, works like the USE symbols above.
4617 =item EV_USE_INOTIFY
4619 If defined to be C<1>, libev will compile in support for the Linux inotify
4620 interface to speed up C<ev_stat> watchers. Its actual availability will
4621 be detected at runtime. If undefined, it will be enabled if the headers
4622 indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
4624 =item EV_NO_SMP
4626 If defined to be C<1>, libev will assume that memory is always coherent
4627 between threads, that is, threads can be used, but threads never run on
4628 different cpus (or different cpu cores). This reduces dependencies
4629 and makes libev faster.
4631 =item EV_NO_THREADS
4633 If defined to be C<1>, libev will assume that it will never be called from
4634 different threads (that includes signal handlers), which is a stronger
4635 assumption than C<EV_NO_SMP>, above. This reduces dependencies and makes
4636 libev faster.
4638 =item EV_ATOMIC_T
4640 Libev requires an integer type (suitable for storing C<0> or C<1>) whose
4641 access is atomic with respect to other threads or signal contexts. No
4642 such type is easily found in the C language, so you can provide your own
4643 type that you know is safe for your purposes. It is used both for signal
4644 handler "locking" as well as for signal and thread safety in C<ev_async>
4645 watchers.
4647 In the absence of this define, libev will use C<sig_atomic_t volatile>
4648 (from F<signal.h>), which is usually good enough on most platforms.
4650 =item EV_H (h)
4652 The name of the F<ev.h> header file used to include it. The default if
4653 undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
4654 used to virtually rename the F<ev.h> header file in case of conflicts.
4656 =item EV_CONFIG_H (h)
4658 If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
4659 F<ev.c>'s idea of where to find the F<config.h> file, similarly to
4660 C<EV_H>, above.
4662 =item EV_EVENT_H (h)
4664 Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
4665 of how the F<event.h> header can be found, the default is C<"event.h">.
4667 =item EV_PROTOTYPES (h)
4669 If defined to be C<0>, then F<ev.h> will not define any function
4670 prototypes, but still define all the structs and other symbols. This is
4671 occasionally useful if you want to provide your own wrapper functions
4672 around libev functions.
4676 If undefined or defined to C<1>, then all event-loop-specific functions
4677 will have the C<struct ev_loop *> as first argument, and you can create
4678 additional independent event loops. Otherwise there will be no support
4679 for multiple event loops and there is no first event loop pointer
4680 argument. Instead, all functions act on the single default loop.
4682 Note that C<EV_DEFAULT> and C<EV_DEFAULT_> will no longer provide a
4683 default loop when multiplicity is switched off - you always have to
4684 initialise the loop manually in this case.
4686 =item EV_MINPRI
4688 =item EV_MAXPRI
4690 The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
4691 C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
4692 provide for more priorities by overriding those symbols (usually defined
4693 to be C<-2> and C<2>, respectively).
4695 When doing priority-based operations, libev usually has to linearly search
4696 all the priorities, so having many of them (hundreds) uses a lot of space
4697 and time, so using the defaults of five priorities (-2 .. +2) is usually
4698 fine.
4700 If your embedding application does not need any priorities, defining these
4701 both to C<0> will save some memory and CPU.
4707 If undefined or defined to be C<1> (and the platform supports it), then
4708 the respective watcher type is supported. If defined to be C<0>, then it
4709 is not. Disabling watcher types mainly saves code size.
4711 =item EV_FEATURES
4713 If you need to shave off some kilobytes of code at the expense of some
4714 speed (but with the full API), you can define this symbol to request
4715 certain subsets of functionality. The default is to enable all features
4716 that can be enabled on the platform.
4718 A typical way to use this symbol is to define it to C<0> (or to a bitset
4719 with some broad features you want) and then selectively re-enable
4720 additional parts you want, for example if you want everything minimal,
4721 but multiple event loop support, async and child watchers and the poll
4722 backend, use this:
4724 #define EV_FEATURES 0
4725 #define EV_MULTIPLICITY 1
4726 #define EV_USE_POLL 1
4727 #define EV_CHILD_ENABLE 1
4728 #define EV_ASYNC_ENABLE 1
4730 The actual value is a bitset, it can be a combination of the following
4731 values (by default, all of these are enabled):
4733 =over 4
4735 =item C<1> - faster/larger code
4737 Use larger code to speed up some operations.
4739 Currently this is used to override some inlining decisions (enlarging the
4740 code size by roughly 30% on amd64).
4742 When optimising for size, use of compiler flags such as C<-Os> with
4743 gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
4744 assertions.
4746 The default is off when C<__OPTIMIZE_SIZE__> is defined by your compiler
4747 (e.g. gcc with C<-Os>).
4749 =item C<2> - faster/larger data structures
4751 Replaces the small 2-heap for timer management by a faster 4-heap, larger
4752 hash table sizes and so on. This will usually further increase code size
4753 and can additionally have an effect on the size of data structures at
4754 runtime.
4756 The default is off when C<__OPTIMIZE_SIZE__> is defined by your compiler
4757 (e.g. gcc with C<-Os>).
4759 =item C<4> - full API configuration
4761 This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
4762 enables multiplicity (C<EV_MULTIPLICITY>=1).
4764 =item C<8> - full API
4766 This enables a lot of the "lesser used" API functions. See C<ev.h> for
4767 details on which parts of the API are still available without this
4768 feature, and do not complain if this subset changes over time.
4770 =item C<16> - enable all optional watcher types
4772 Enables all optional watcher types. If you want to selectively enable
4773 only some watcher types other than I/O and timers (e.g. prepare,
4774 embed, async, child...) you can enable them manually by defining
4775 C<EV_watchertype_ENABLE> to C<1> instead.
4777 =item C<32> - enable all backends
4779 This enables all backends - without this feature, you need to enable at
4780 least one backend manually (C<EV_USE_SELECT> is a good choice).
4782 =item C<64> - enable OS-specific "helper" APIs
4784 Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
4785 default.
4787 =back
4789 Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
4790 reduces the compiled size of libev from 24.7Kb code/2.8Kb data to 6.5Kb
4791 code/0.3Kb data on my GNU/Linux amd64 system, while still giving you I/O
4792 watchers, timers and monotonic clock support.
4794 With an intelligent-enough linker (gcc+binutils are intelligent enough
4795 when you use C<-Wl,--gc-sections -ffunction-sections>) functions unused by
4796 your program might be left out as well - a binary starting a timer and an
4797 I/O watcher then might come out at only 5Kb.
4799 =item EV_API_STATIC
4801 If this symbol is defined (by default it is not), then all identifiers
4802 will have static linkage. This means that libev will not export any
4803 identifiers, and you cannot link against libev anymore. This can be useful
4804 when you embed libev, only want to use libev functions in a single file,
4805 and do not want its identifiers to be visible.
4807 To use this, define C<EV_API_STATIC> and include F<ev.c> in the file that
4808 wants to use libev.
4810 This option only works when libev is compiled with a C compiler, as C++
4811 doesn't support the required declaration syntax.
4813 =item EV_AVOID_STDIO
4815 If this is set to C<1> at compiletime, then libev will avoid using stdio
4816 functions (printf, scanf, perror etc.). This will increase the code size
4817 somewhat, but if your program doesn't otherwise depend on stdio and your
4818 libc allows it, this avoids linking in the stdio library which is quite
4819 big.
4821 Note that error messages might become less precise when this option is
4822 enabled.
4824 =item EV_NSIG
4826 The highest supported signal number, +1 (or, the number of
4827 signals): Normally, libev tries to deduce the maximum number of signals
4828 automatically, but sometimes this fails, in which case it can be
4829 specified. Also, using a lower number than detected (C<32> should be
4830 good for about any system in existence) can save some memory, as libev
4831 statically allocates some 12-24 bytes per signal number.
4833 =item EV_PID_HASHSIZE
4835 C<ev_child> watchers use a small hash table to distribute workload by
4836 pid. The default size is C<16> (or C<1> with C<EV_FEATURES> disabled),
4837 usually more than enough. If you need to manage thousands of children you
4838 might want to increase this value (I<must> be a power of two).
4842 C<ev_stat> watchers use a small hash table to distribute workload by
4843 inotify watch id. The default size is C<16> (or C<1> with C<EV_FEATURES>
4844 disabled), usually more than enough. If you need to manage thousands of
4845 C<ev_stat> watchers you might want to increase this value (I<must> be a
4846 power of two).
4848 =item EV_USE_4HEAP
4850 Heaps are not very cache-efficient. To improve the cache-efficiency of the
4851 timer and periodics heaps, libev uses a 4-heap when this symbol is defined
4852 to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
4853 faster performance with many (thousands) of watchers.
4855 The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
4856 will be C<0>.
4858 =item EV_HEAP_CACHE_AT
4860 Heaps are not very cache-efficient. To improve the cache-efficiency of the
4861 timer and periodics heaps, libev can cache the timestamp (I<at>) within
4862 the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
4863 which uses 8-12 bytes more per watcher and a few hundred bytes more code,
4864 but avoids random read accesses on heap changes. This improves performance
4865 noticeably with many (hundreds) of watchers.
4867 The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
4868 will be C<0>.
4870 =item EV_VERIFY
4872 Controls how much internal verification (see C<ev_verify ()>) will
4873 be done: If set to C<0>, no internal verification code will be compiled
4874 in. If set to C<1>, then verification code will be compiled in, but not
4875 called. If set to C<2>, then the internal verification code will be
4876 called once per loop, which can slow down libev. If set to C<3>, then the
4877 verification code will be called very frequently, which will slow down
4878 libev considerably.
4880 The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
4881 will be C<0>.
4883 =item EV_COMMON
4885 By default, all watchers have a C<void *data> member. By redefining
4886 this macro to something else you can include more and other types of
4887 members. You have to define it each time you include one of the files,
4888 though, and it must be identical each time.
4890 For example, the perl EV module uses something like this:
4892 #define EV_COMMON \
4893 SV *self; /* contains this struct */ \
4894 SV *cb_sv, *fh /* note no trailing ";" */
4896 =item EV_CB_DECLARE (type)
4898 =item EV_CB_INVOKE (watcher, revents)
4900 =item ev_set_cb (ev, cb)
4902 Can be used to change the callback member declaration in each watcher,
4903 and the way callbacks are invoked and set. Must expand to a struct member
4904 definition and a statement, respectively. See the F<ev.h> header file for
4905 their default definitions. One possible use for overriding these is to
4906 avoid the C<struct ev_loop *> as first argument in all cases, or to use
4907 method calls instead of plain function calls in C++.
4909 =back
4913 If you need to re-export the API (e.g. via a DLL) and you need a list of
4914 exported symbols, you can use the provided F<Symbol.*> files which list
4915 all public symbols, one per line:
4917 Symbols.ev for libev proper
4918 Symbols.event for the libevent emulation
4920 This can also be used to rename all public symbols to avoid clashes with
4921 multiple versions of libev linked together (which is obviously bad in
4922 itself, but sometimes it is inconvenient to avoid this).
4924 A sed command like this will create wrapper C<#define>'s that you need to
4925 include before including F<ev.h>:
4927 <Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
4929 This would create a file F<wrap.h> which essentially looks like this:
4931 #define ev_backend myprefix_ev_backend
4932 #define ev_check_start myprefix_ev_check_start
4933 #define ev_check_stop myprefix_ev_check_stop
4934 ...
4936 =head2 EXAMPLES
4938 For a real-world example of a program the includes libev
4939 verbatim, you can have a look at the EV perl module
4940 (L<>). It has the libev files in
4941 the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
4942 interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
4943 will be compiled. It is pretty complex because it provides its own header
4944 file.
4946 The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
4947 that everybody includes and which overrides some configure choices:
4949 #define EV_FEATURES 8
4950 #define EV_USE_SELECT 1
4951 #define EV_PREPARE_ENABLE 1
4952 #define EV_IDLE_ENABLE 1
4953 #define EV_SIGNAL_ENABLE 1
4954 #define EV_CHILD_ENABLE 1
4955 #define EV_USE_STDEXCEPT 0
4956 #define EV_CONFIG_H <config.h>
4958 #include "ev++.h"
4960 And a F<ev_cpp.C> implementation file that contains libev proper and is compiled: